diff --git a/.github/scripts/pytest_md_summary.py b/.github/scripts/pytest_md_summary.py
new file mode 100644
index 00000000000..ac9b8c91b2b
--- /dev/null
+++ b/.github/scripts/pytest_md_summary.py
@@ -0,0 +1,59 @@
+# Copyright (c) 2024 Intel Corporation
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""
+This script generates a summary table in Markdown format from an XML report generated by pytest.
+
+Usage in GitHub workflow:
+    - name: Test Summary
+    if: ${{ !cancelled() }}
+    run: |
+        python .github/scripts/generate_examples_summary.py pytest-results.xml >> $GITHUB_STEP_SUMMARY
+"""
+
+import sys
+import xml.etree.ElementTree as ET
+
+# Load the XML report generated by pytest
+xml_file = sys.argv[1]
+
+try:
+    tree = ET.parse(xml_file)
+except FileNotFoundError:
+    sys.exit(1)
+
+root = tree.getroot()
+
+# Build the summary table in Markdown format
+table_lines = []
+table_lines.append("| Test Name | Status | Time | Message |")
+table_lines.append("|:----------|:------:|-----:|:--------|")
+
+# Iterate over test cases
+for testcase in root.findall(".//testcase"):
+    test_name = testcase.get("name")
+    time_duration = float(testcase.get("time", "0"))
+    message = ""
+    if testcase.find("failure") is not None:
+        status = "$${\color{red}Failed}$$"
+        message = testcase.find("failure").get("message", "")
+    elif testcase.find("error") is not None:
+        status = "$${\color{red}Error}$$"
+    elif testcase.find("skipped") is not None:
+        status = "$${\color{orange}Skipped}$$"
+        message = testcase.find("skipped").get("message", "")
+    else:
+        status = "$${\color{green}Ok}$$"
+
+    # Append each row to the table
+    table_lines.append(f"| {test_name} | {status} | {time_duration:.0f} | {message} |")
+
+print("\n".join(table_lines))
diff --git a/.github/workflows/examples.yml b/.github/workflows/examples.yml
index 150ec493839..b5bf5ff6e7f 100644
--- a/.github/workflows/examples.yml
+++ b/.github/workflows/examples.yml
@@ -13,6 +13,10 @@ on:
         description: 'Pytest arguments'
         default: ''
 
+concurrency:
+  group: test-examples-${{ github.workflow }}-${{ github.ref }}-${{ github.event.inputs.pytest_args || '' }}-${{github.event.inputs.pull_request_number || ''}}
+  cancel-in-progress: false
+
 jobs:
   examples-cpu:
     name: Test exmaples CPU [${{ matrix.group }}/4]
@@ -48,19 +52,19 @@ jobs:
         run: pip list
       - name: Run examples test scope
         run: |
-          python -m pytest -ras tests/cross_fw/examples \
-            --junit-xml=pytest-results-${{ matrix.group }}.xml \
+          set +e
+          python -m pytest -s -ra tests/cross_fw/examples \
+            --junit-xml=pytest-results.xml \
             --durations-path=tests/cross_fw/examples/.test_durations \
             --splitting-algorithm=least_duration \
             --splits 4 \
             --group ${{ matrix.group }} \
             ${{ github.event.inputs.pytest_args || '' }}
+          ret=$?
+          [ $ret -eq 5 ] && [ -n "${{ github.event.inputs.pytest_args || '' }}" ]  && exit 0 || exit $ret
         env:
           TQDM_DISABLE: 1
-      - name: Upload artifact
-        uses: actions/upload-artifact@v4
+      - name: Test Summary
         if: ${{ !cancelled() }}
-        with:
-          name: pytest-results-${{ matrix.group }}
-          path: pytest-results-${{ matrix.group }}.xml
-          overwrite: True
+        run: |
+            python .github/scripts/pytest_md_summary.py pytest-results.xml >> $GITHUB_STEP_SUMMARY
diff --git a/.gitignore b/.gitignore
index 9751a56d366..1328d5a0416 100644
--- a/.gitignore
+++ b/.gitignore
@@ -129,6 +129,7 @@ examples/post_training_quantization/openvino/yolov8/yolov8n*
 examples/post_training_quantization/openvino/yolov8_quantize_with_accuracy_control/yolov8n*
 examples/**/runs/**
 examples/**/results/**
+examples/llm_compression/openvino/tiny_llama_find_hyperparams/statistics
 compressed_graph.dot
 original_graph.dot
 datasets/**
diff --git a/docs/usage/post_training_compression/weights_compression/Usage.md b/docs/usage/post_training_compression/weights_compression/Usage.md
index 31beb7dcfeb..1c5cd1257a7 100644
--- a/docs/usage/post_training_compression/weights_compression/Usage.md
+++ b/docs/usage/post_training_compression/weights_compression/Usage.md
@@ -22,7 +22,7 @@ The Weights Compression algorithm is aimed at compressing the weights of the mod
 ### Supported modes
 
 By default, weights are compressed asymmetrically to 8-bit integer data type - "INT8_ASYM" mode.
-OpenVINO backend also supports 4 modes of mixed precision weight quantization with a 4-bit data type as a primary precision - INT4_SYM, INT4_ASYM, NF4, E2M1. The primary precision in case of INT4_SYM mode is signed 4-bit integer and weights are quantized to it [symmetrically](/docs/usage/training_time_compression/other_algorithms/LegacyQuantization.md#symmetric-quantization) without zero point. In case of INT4_ASYM mode - unsigned 4-bit integer and weight are quantized to it [asymmetrically](/docs/usage/training_time_compression/other_algorithms/LegacyQuantization.md#asymmetric-quantization) with a typical non-fixed zero point. In case of NF4 mode - [nf4](https://arxiv.org/pdf/2305.14314v1.pdf) data type without zero point. In case of E2M1 mode - [e2m1](https://www.opencompute.org/documents/ocp-microscaling-formats-mx-v1-0-spec-final-pdf) data type without zero point and has 8bit [E8M0](https://www.opencompute.org/documents/ocp-microscaling-formats-mx-v1-0-spec-final-pdf) scale.
+OpenVINO backend also supports 4 modes of mixed precision weight quantization with a 4-bit data type as a primary precision - INT4_SYM, INT4_ASYM, NF4, E2M1. The primary precision in case of INT4_SYM mode is signed 4-bit integer and weights are quantized to it [symmetrically](/docs/usage/training_time_compression/other_algorithms/LegacyQuantization.md#symmetric-quantization) without zero point. In case of INT4_ASYM mode - unsigned 4-bit integer and weight are quantized to it [asymmetrically](/docs/usage/training_time_compression/other_algorithms/LegacyQuantization.md#asymmetric-quantization) with a typical non-fixed zero point. In case of NF4 mode - [nf4](https://arxiv.org/pdf/2305.14314v1.pdf) data type without zero point. In case of E2M1 mode - [e2m1](https://arxiv.org/pdf/2310.10537) data type without zero point and has 8bit [E8M0](https://arxiv.org/pdf/2310.10537) scale.
 All 4-bit modes have a grouped quantization support, when small group of weights (e.g. 128) in the channel dimension share quantization parameters (scale).
 All embeddings, convolutions and last linear layers are always compressed to a backup mode, which is "INT8_ASYM", by default. To quantize embeddings and last linear layers to 4-bit, use `all_layers=True`.
 Percent of the rest layers compressed to 4-bit can be configured by "ratio" parameter. E.g. ratio=0.9 means 90% of layers compressed to the corresponding 4-bit data type and the rest to a backup mode. OpenVINO backend supports 3 backup modes: INT8_SYM, INT8_ASYM, and NONE, which retains the original floating-point precision of the model weights. Backup mode is supported only for mixed-precision weight quantization.
diff --git a/examples/llm_compression/openvino/tiny_llama_find_hyperparams/main.py b/examples/llm_compression/openvino/tiny_llama_find_hyperparams/main.py
index 081e99125b4..5799f649aa4 100644
--- a/examples/llm_compression/openvino/tiny_llama_find_hyperparams/main.py
+++ b/examples/llm_compression/openvino/tiny_llama_find_hyperparams/main.py
@@ -31,6 +31,7 @@
 
 ROOT = Path(__file__).parent.resolve()
 MODEL_PATH = ROOT / "compressed_model.xml"
+STATISTICS_PATH = ROOT / "statistics"
 
 COMPRESSION_MODE = nncf.parameters.CompressWeightsMode.INT4_SYM
 MAX_DROP = 0.2
@@ -64,7 +65,7 @@ def compress_model(
         group_size=group_size,
         awq=awq,
         sensitivity_metric=nncf.parameters.SensitivityMetric.MAX_ACTIVATION_VARIANCE,
-        advanced_parameters=AdvancedCompressionParameters(statistics_path="statistics"),
+        advanced_parameters=AdvancedCompressionParameters(statistics_path=STATISTICS_PATH),
     )
     return optimized_ov_model
 
diff --git a/nncf/experimental/common/tensor_statistics/collectors.py b/nncf/experimental/common/tensor_statistics/collectors.py
index bfb304e75c9..c7e19193fcc 100644
--- a/nncf/experimental/common/tensor_statistics/collectors.py
+++ b/nncf/experimental/common/tensor_statistics/collectors.py
@@ -465,6 +465,21 @@ def _reduce_out_of_place(self, x: List[Tensor]) -> List[Tensor]:
         return [fns.mean(x, reduction_axes, keepdims=self._keepdims)]
 
 
+class MeanVarianceReducer(TensorReducerBase):
+    def _reduce_out_of_place(self, x: List[TensorType]) -> List[TensorType]:
+        raise NotImplementedError()
+
+
+class MaxVarianceReducer(TensorReducerBase):
+    def _reduce_out_of_place(self, x: List[TensorType]) -> List[TensorType]:
+        raise NotImplementedError()
+
+
+class MeanAbsMaxReducer(TensorReducerBase):
+    def _reduce_out_of_place(self, x: List[TensorType]) -> List[TensorType]:
+        raise NotImplementedError()
+
+
 class QuantileReducerBase(TensorReducerBase):
     def __init__(
         self,
diff --git a/nncf/experimental/torch/fx/model_transformer.py b/nncf/experimental/torch/fx/model_transformer.py
index 21c4d40604b..4e0f6f64499 100644
--- a/nncf/experimental/torch/fx/model_transformer.py
+++ b/nncf/experimental/torch/fx/model_transformer.py
@@ -84,6 +84,9 @@ def _traverse_graph(
                 continue
 
             visited.add(in_node.name)
+            # Any constant is a stop op during the traversing procedure.
+            if in_node.op == "get_attr":
+                continue
             input_nodes.extend(in_node.all_input_nodes)
             input_nodes.extend(list(in_node.users))
 
diff --git a/nncf/experimental/torch/fx/nncf_graph_builder.py b/nncf/experimental/torch/fx/nncf_graph_builder.py
index 4599c649b56..c005073b941 100644
--- a/nncf/experimental/torch/fx/nncf_graph_builder.py
+++ b/nncf/experimental/torch/fx/nncf_graph_builder.py
@@ -187,7 +187,9 @@ def get_edge_params(
         if source_node.op in ("get_attr",):
             tensor_shape = tuple(get_tensor_constant_from_node(source_node, model).shape)
         elif "val" in source_node.meta:
-            if source_nncf_node.metatype is om.PTBatchNormMetatype:
+            if source_nncf_node.metatype is om.PTBatchNormMetatype and isinstance(
+                source_node.meta["val"], (tuple, list)
+            ):
                 tensor = source_node.meta["val"][0]
             elif source_nncf_node.metatype in [om.PTSplitMetatype, om.PTMaxMetatype, om.PTMinMetatype]:
                 tensor = source_node.meta["val"][output_idx]
diff --git a/nncf/experimental/torch/fx/quantization/quantize_model.py b/nncf/experimental/torch/fx/quantization/quantize_model.py
index 277ab50edb3..bdfe839994c 100644
--- a/nncf/experimental/torch/fx/quantization/quantize_model.py
+++ b/nncf/experimental/torch/fx/quantization/quantize_model.py
@@ -31,7 +31,6 @@
 from nncf.experimental.torch.fx.transformations import compress_post_quantize_transformation
 from nncf.experimental.torch.fx.transformations import fq_weights_transformation
 from nncf.experimental.torch.fx.transformations import revert_quantization_transformations
-from nncf.experimental.torch.fx.transformations import shared_constants_unification_transformation
 from nncf.parameters import BackupMode
 from nncf.parameters import CompressWeightsMode
 from nncf.parameters import ModelType
@@ -158,7 +157,6 @@ def compress_weights_impl(
         backup_mode,
         advanced_parameters,
     )
-    shared_constants_unification_transformation(model)
     graph = NNCFGraphFactory.create(model)
     compressed_model = compression_algorithm.apply(model, graph, dataset=dataset)
     compressed_model = GraphModule(compressed_model, compressed_model.graph)
diff --git a/nncf/experimental/torch/fx/transformations.py b/nncf/experimental/torch/fx/transformations.py
index da277d03836..c695220c981 100644
--- a/nncf/experimental/torch/fx/transformations.py
+++ b/nncf/experimental/torch/fx/transformations.py
@@ -187,28 +187,6 @@ def bias_update_transformation(model: torch.fx.GraphModule):
     return bias_update_transformation
 
 
-def shared_constants_unification_transformation(model: torch.fx.GraphModule):
-    """
-    checks FX graph for shared constants and eliminates redundant
-    shared constant while keeping only the first instance of the constant node.
-    This unification transformation is cruicial since the current algorithms(min_max, solver, BC, etc.)
-    for torch fx do not utilize the is_shared attribute of nodes for shared constants.
-
-    :param model: Target Torch FX GraphModule
-    """
-    prev_targets = {}
-
-    for source_node in model.graph.nodes:
-        dist_node = list(source_node.users)
-        if source_node.target in prev_targets and source_node.op in ("get_attr",):
-            dist_node[0].replace_input_with(source_node, prev_targets[source_node.target])
-        else:
-            prev_targets[source_node.target] = source_node
-
-    model.graph.eliminate_dead_code()
-    model.recompile()
-
-
 def constant_update_transformation_builder(
     node: NNCFNode, value: torch.Tensor, input_port_id: int = 1
 ) -> TransformationFNType:
@@ -541,6 +519,7 @@ def _is_supported_batch_norm_for_training(node: torch.fx.Node):
     Return True if the given node refers to an aten batch norm op QAT supports.
     """
     supported_ops = [
+        torch.ops.aten.batch_norm.default,
         torch.ops.aten._native_batch_norm_legit.default,
         torch.ops.aten.cudnn_batch_norm.default,
         torch.ops.aten.miopen_batch_norm.default,
@@ -807,7 +786,6 @@ def apply_quantization_transformations(model: torch.fx.GraphModule) -> None:
     fuse_conv_bn(model)
     separate_conv_and_bias(model)
     separate_linear_and_bias(model)
-    shared_constants_unification_transformation(model)
 
 
 def fold_constant_except_qdq(model: torch.fx.GraphModule):
diff --git a/nncf/openvino/statistics/collectors.py b/nncf/openvino/statistics/collectors.py
index ed63ea9ea6f..44a6ed606d0 100644
--- a/nncf/openvino/statistics/collectors.py
+++ b/nncf/openvino/statistics/collectors.py
@@ -9,17 +9,19 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import List, Optional
+from typing import Optional
 
-from nncf.common.tensor import TensorType
 from nncf.experimental.common.tensor_statistics.collectors import AbsMaxReducer
 from nncf.experimental.common.tensor_statistics.collectors import AbsQuantileReducer
 from nncf.experimental.common.tensor_statistics.collectors import BatchMeanReducer
 from nncf.experimental.common.tensor_statistics.collectors import InplaceInsertionFNType
 from nncf.experimental.common.tensor_statistics.collectors import MaxReducer
+from nncf.experimental.common.tensor_statistics.collectors import MaxVarianceReducer
+from nncf.experimental.common.tensor_statistics.collectors import MeanAbsMaxReducer
 from nncf.experimental.common.tensor_statistics.collectors import MeanAggregator
 from nncf.experimental.common.tensor_statistics.collectors import MeanPerChReducer
 from nncf.experimental.common.tensor_statistics.collectors import MeanReducer
+from nncf.experimental.common.tensor_statistics.collectors import MeanVarianceReducer
 from nncf.experimental.common.tensor_statistics.collectors import MinReducer
 from nncf.experimental.common.tensor_statistics.collectors import NoopAggregator
 from nncf.experimental.common.tensor_statistics.collectors import QuantileReducer
@@ -27,7 +29,6 @@
 from nncf.experimental.common.tensor_statistics.collectors import ShapeAggregator
 from nncf.experimental.common.tensor_statistics.collectors import ShapeReducer
 from nncf.experimental.common.tensor_statistics.collectors import TensorCollector
-from nncf.experimental.common.tensor_statistics.collectors import TensorReducerBase
 from nncf.experimental.common.tensor_statistics.statistics import MeanTensorStatistic
 from nncf.experimental.common.tensor_statistics.statistics import RawTensorStatistic
 from nncf.openvino.graph.node_utils import get_inplace_batch_mean_op
@@ -66,26 +67,17 @@ def get_inplace_fn(self):
         return get_inplace_mean_op(self._reduction_axes)
 
 
-class OVMeanVarianceReducer(TensorReducerBase):
-    def _reduce_out_of_place(self, x: List[TensorType]) -> List[TensorType]:
-        raise NotImplementedError()
-
+class OVMeanVarianceReducer(MeanVarianceReducer):
     def get_inplace_fn(self):
         return get_inplace_mean_var_op(self._reduction_axes)
 
 
-class OVMaxVarianceReducer(TensorReducerBase):
-    def _reduce_out_of_place(self, x: List[TensorType]) -> List[TensorType]:
-        raise NotImplementedError()
-
+class OVMaxVarianceReducer(MaxVarianceReducer):
     def get_inplace_fn(self):
         return get_inplace_max_var_op(self._reduction_axes)
 
 
-class OVMeanAbsMaxReducer(TensorReducerBase):
-    def _reduce_out_of_place(self, x: List[TensorType]) -> List[TensorType]:
-        raise NotImplementedError()
-
+class OVMeanAbsMaxReducer(MeanAbsMaxReducer):
     def get_inplace_fn(self):
         return get_inplace_mean_max_op(self._reduction_axes, True)
 
diff --git a/nncf/quantization/algorithms/weight_compression/mixed_precision.py b/nncf/quantization/algorithms/weight_compression/mixed_precision.py
index a96c09fcb19..f2fbae67a5c 100644
--- a/nncf/quantization/algorithms/weight_compression/mixed_precision.py
+++ b/nncf/quantization/algorithms/weight_compression/mixed_precision.py
@@ -281,7 +281,7 @@ def get_statistic_points(
         return statistic_container
 
     @abstractmethod
-    def _get_statistic_collector():
+    def _get_statistic_collector(self):
         """
         Get statistic collector
         """
@@ -360,7 +360,7 @@ def _calc_weight_sensitivity(
         return fns.linalg.norm(decompressed_weight - weight, ord="fro").item()
 
     def _get_statistic_collector(self):
-        return self._backend_entity.hawq_statistic_collector()
+        return self._backend_entity.hawq_statistic_collector(self._subset_size)
 
 
 @MIXED_PRECISION_CRITERIA.register(SensitivityMetric.MEAN_ACTIVATION_VARIANCE)
diff --git a/tests/cross_fw/test_templates/test_weights_compression_backends.py b/tests/cross_fw/test_templates/test_weights_compression_backends.py
new file mode 100644
index 00000000000..fb521fd0726
--- /dev/null
+++ b/tests/cross_fw/test_templates/test_weights_compression_backends.py
@@ -0,0 +1,77 @@
+# Copyright (c) 2024 Intel Corporation
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from abc import abstractmethod
+
+import pytest
+
+from nncf.experimental.common.tensor_statistics.collectors import HAWQAggregator
+from nncf.experimental.common.tensor_statistics.collectors import MaxVarianceReducer
+from nncf.experimental.common.tensor_statistics.collectors import MeanAbsMaxReducer
+from nncf.experimental.common.tensor_statistics.collectors import MeanAggregator
+from nncf.experimental.common.tensor_statistics.collectors import MeanVarianceReducer
+from nncf.experimental.common.tensor_statistics.collectors import NoopReducer
+from nncf.experimental.common.tensor_statistics.collectors import TensorCollector
+
+
+class TemplateTestMixedPrecisionAlgoBackend:
+    @abstractmethod
+    def get_hawq_with_backend(self, subset_size: int):
+        """Returns a HAWQ instance of the algorithm."""
+
+    @abstractmethod
+    def get_mean_variance_with_backend(self, subset_size: int):
+        """Returns a Mean Variance instance of the algorithm."""
+
+    @abstractmethod
+    def get_max_variance_with_backend(self, subset_size: int):
+        """Returns a Max Variance instance of the algorithm."""
+
+    @abstractmethod
+    def get_mean_max_with_backend(self, subset_size: int):
+        """Returns a Mean Max instance of the algorithm."""
+
+    def check_aggregator(self, collector: TensorCollector, expected_aggregator_type, subset_size: int):
+        assert len(collector.aggregators) == 1, "Collector should have exactly one aggregator."
+        _, aggregator = collector.aggregators.popitem()
+        assert isinstance(
+            aggregator, expected_aggregator_type
+        ), f"Expected aggregator of type {expected_aggregator_type.__name__}, got {type(aggregator).__name__}."
+        assert aggregator.num_samples == subset_size, "Aggregator num_samples does not match the provided subset size."
+
+    def check_reducer(self, collector: TensorCollector, expected_reducer_type):
+        assert len(collector.reducers) == 1
+        reducer = collector.reducers.pop()
+        assert isinstance(
+            reducer, expected_reducer_type
+        ), f"Expected reducer of type {expected_reducer_type.__name__}, got {type(reducer).__name__}."
+
+    @pytest.mark.parametrize("subset_size", [1, 10, None])
+    @pytest.mark.parametrize(
+        "algo_func, aggregator_type, reducer_type",
+        [
+            ("get_hawq_with_backend", HAWQAggregator, NoopReducer),
+            ("get_mean_variance_with_backend", MeanAggregator, MeanVarianceReducer),
+            ("get_max_variance_with_backend", MeanAggregator, MaxVarianceReducer),
+            ("get_mean_max_with_backend", MeanAggregator, MeanAbsMaxReducer),
+        ],
+    )
+    def test_statistic_collector(self, subset_size, algo_func, aggregator_type, reducer_type):
+        """Test function to validate statistic collectors."""
+        algo = getattr(self, algo_func)(subset_size)
+        collector = algo._get_statistic_collector()
+
+        # Verify the collector instance and properties
+        assert isinstance(collector, TensorCollector), "Collector is not an instance of TensorCollector."
+
+        # Validate the aggregator and reducer types
+        self.check_aggregator(collector, aggregator_type, subset_size)
+        self.check_reducer(collector, reducer_type)
diff --git a/tests/openvino/native/quantization/test_weights_compression.py b/tests/openvino/native/quantization/test_weights_compression.py
index 46f70266691..541b8a386e5 100644
--- a/tests/openvino/native/quantization/test_weights_compression.py
+++ b/tests/openvino/native/quantization/test_weights_compression.py
@@ -899,12 +899,54 @@ def test_compression_for_different_dtypes(activation_dtype, weight_dtype):
 )
 @pytest.mark.parametrize(
     ("compression_args", "multiplier_of_calls"),
-    (
-        (dict(mode=CompressWeightsMode.INT4_ASYM, ratio=1), 0),  # data-free, no reducers
-        (dict(mode=CompressWeightsMode.INT4_ASYM, ratio=0.5), 1),  # 1 reducer for mixed precision
-        (dict(mode=CompressWeightsMode.INT4_ASYM, ratio=1, awq=True), 2),  # mean & shape reducer for AWQ
-        (dict(mode=CompressWeightsMode.INT4_ASYM, ratio=0.5, awq=True), 3),  # 2 - for AWQ + 1 - for Mixed Precision
-    ),
+    [
+        ({"mode": CompressWeightsMode.INT4_ASYM, "ratio": 1}, 0),  # data-free, no reducers
+        ({"mode": CompressWeightsMode.INT4_ASYM, "ratio": 1, "awq": True}, 2),  # mean & shape reducer for AWQ
+        (
+            {"mode": CompressWeightsMode.INT4_ASYM, "ratio": 0.5, "awq": True},
+            3,
+        ),  # 2 - for AWQ + 1 - for Mixed Precision
+        (
+            {
+                "mode": CompressWeightsMode.INT4_ASYM,
+                "ratio": 0.5,
+                "sensitivity_metric": nncf.SensitivityMetric.HESSIAN_INPUT_ACTIVATION,
+            },
+            1,
+        ),  # 1 reducer for mixed precision
+        (
+            {
+                "mode": CompressWeightsMode.INT4_ASYM,
+                "ratio": 0.5,
+                "sensitivity_metric": nncf.SensitivityMetric.MEAN_ACTIVATION_VARIANCE,
+            },
+            1,
+        ),  # 1 reducer for mixed precision
+        (
+            {
+                "mode": CompressWeightsMode.INT4_ASYM,
+                "ratio": 0.5,
+                "sensitivity_metric": nncf.SensitivityMetric.MAX_ACTIVATION_VARIANCE,
+            },
+            1,
+        ),  # 1 reducer for mixed precision
+        (
+            {
+                "mode": CompressWeightsMode.INT4_ASYM,
+                "ratio": 0.5,
+                "sensitivity_metric": nncf.SensitivityMetric.MEAN_ACTIVATION_MAGNITUDE,
+            },
+            1,
+        ),  # 1 reducer for mixed precision
+        (
+            {
+                "mode": CompressWeightsMode.INT4_ASYM,
+                "ratio": 0.5,
+                "sensitivity_metric": nncf.SensitivityMetric.WEIGHT_QUANTIZATION_ERROR,
+            },
+            0,
+        ),  # 0 - data-free method
+    ],
 )
 def test_number_of_reduced_statistics_for_subset_size(
     mocker, dataset_size, subset_size, ref_size, compression_args, multiplier_of_calls
diff --git a/tests/openvino/native/quantization/test_weights_compression_backends.py b/tests/openvino/native/quantization/test_weights_compression_backends.py
new file mode 100644
index 00000000000..ab0271a6a4f
--- /dev/null
+++ b/tests/openvino/native/quantization/test_weights_compression_backends.py
@@ -0,0 +1,39 @@
+# Copyright (c) 2024 Intel Corporation
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#      http://www.apache.org/licenses/LICENSE-2.0
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+from nncf.quantization.algorithms.weight_compression.mixed_precision import HAWQCriterion
+from nncf.quantization.algorithms.weight_compression.mixed_precision import MaxVarianceCriterion
+from nncf.quantization.algorithms.weight_compression.mixed_precision import MeanMaxCriterion
+from nncf.quantization.algorithms.weight_compression.mixed_precision import MeanVarianceCriterion
+from nncf.quantization.algorithms.weight_compression.openvino_backend import OVMixedPrecisionAlgoBackend
+from tests.cross_fw.test_templates.test_weights_compression_backends import TemplateTestMixedPrecisionAlgoBackend
+from tests.openvino.native.models import IdentityMatmul
+
+
+class TestOVMixedPrecisionAlgoBackend(TemplateTestMixedPrecisionAlgoBackend):
+    def get_hawq_with_backend(self, subset_size):
+        hawq = HAWQCriterion(None, None, subset_size=subset_size)
+        hawq._backend_entity = OVMixedPrecisionAlgoBackend(IdentityMatmul().ov_model)
+        return hawq
+
+    def get_mean_variance_with_backend(self, subset_size: int):
+        mean_variance = MeanVarianceCriterion(None, None, subset_size=subset_size)
+        mean_variance._backend_entity = OVMixedPrecisionAlgoBackend(IdentityMatmul().ov_model)
+        return mean_variance
+
+    def get_max_variance_with_backend(self, subset_size: int):
+        max_variance = MaxVarianceCriterion(None, None, subset_size=subset_size)
+        max_variance._backend_entity = OVMixedPrecisionAlgoBackend(IdentityMatmul().ov_model)
+        return max_variance
+
+    def get_mean_max_with_backend(self, subset_size: int):
+        mean_max_variance = MeanMaxCriterion(None, None, subset_size=subset_size)
+        mean_max_variance._backend_entity = OVMixedPrecisionAlgoBackend(IdentityMatmul().ov_model)
+        return mean_max_variance
diff --git a/tests/post_training/data/wc_reference_data.yaml b/tests/post_training/data/wc_reference_data.yaml
index b3c2cc44d20..8fbc14a4396 100644
--- a/tests/post_training/data/wc_reference_data.yaml
+++ b/tests/post_training/data/wc_reference_data.yaml
@@ -36,11 +36,11 @@ tinyllama_data_aware_lora_stateful_backend_OV:
   num_int4: 94
   num_int8: 500
 tinyllama_NF4_scale_estimation_stateful_per_channel_backend_OV:
-  metric_value: 0.87942
+  metric_value: 0.87132
   num_int4: 11
   num_int8: 290
   metrics_xfail_reason: "Issue-148819"
 tinyllama_awq_backup_mode_none_backend_OV:
-  metric_value: 0.84793
+  metric_value: 0.85679
   num_int4: 208
   num_int8: 0
diff --git a/tests/tensorflow/sota_checkpoints_eval.json b/tests/tensorflow/sota_checkpoints_eval.json
index 0272cf0a11b..ceb65f510ab 100644
--- a/tests/tensorflow/sota_checkpoints_eval.json
+++ b/tests/tensorflow/sota_checkpoints_eval.json
@@ -320,7 +320,9 @@
                     "model_description": "YOLO v4",
                     "batch_per_gpu": 15,
                     "target_tf": 47.07,
-                    "target_ov": 48.46
+                    "target_ov": 48.46,
+                    "diff_target_ov_min": -0.02,
+                    "diff_target_ov_max": 0.02
                 },
                 "yolo_v4_coco_int8": {
                     "config": "examples/tensorflow/object_detection/configs/quantization/yolo_v4_coco_int8.json",
diff --git a/tests/torch/conftest.py b/tests/torch/conftest.py
index c2e524bd0f5..fd3bc9937b5 100644
--- a/tests/torch/conftest.py
+++ b/tests/torch/conftest.py
@@ -54,6 +54,14 @@ def pytest_addoption(parser: Parser):
         "reference .dot files will be regenerated "
         "using the current state of the repository.",
     )
+    parser.addoption(
+        "--regen-json",
+        action="store_true",
+        default=False,
+        help="If specified, the "
+        "reference .json files will be regenerated "
+        "using the current state of the repository.",
+    )
     parser.addoption(
         "--torch-home", type=str, default=None, help="Path to cached test models, downloaded by torchvision"
     )
@@ -116,9 +124,9 @@ def pytest_addoption(parser: Parser):
 
 
 def pytest_configure(config: Config):
-    regen_dot = config.getoption("--regen-dot", False)
-    if regen_dot:
-        os.environ["NNCF_TEST_REGEN_DOT"] = "1"
+    for regen_option in ["dot", "json"]:
+        if config.getoption(f"--regen-{regen_option}", False):
+            os.environ[f"NNCF_TEST_REGEN_{regen_option.upper()}"] = "1"
 
 
 @pytest.fixture(scope="module")
diff --git a/tests/torch/data/reference_graphs/fx/extracted/ConvolutionWithAllConstantInputsModelconv2d_conv2d.dot b/tests/torch/data/reference_graphs/fx/extracted/ConvolutionWithAllConstantInputsModelconv2d_conv2d.dot
index e8e439e2f3c..2132db4c1bc 100644
--- a/tests/torch/data/reference_graphs/fx/extracted/ConvolutionWithAllConstantInputsModelconv2d_conv2d.dot
+++ b/tests/torch/data/reference_graphs/fx/extracted/ConvolutionWithAllConstantInputsModelconv2d_conv2d.dot
@@ -1,10 +1,10 @@
 strict digraph  {
-"0 _param_constant0" [id=0, type=get_attr];
+"0 _conv_w" [id=0, type=get_attr];
 "1 add" [id=1, type=add];
 "2 conv2d_input" [id=2, type=input];
 "3 conv2d" [id=3, type=conv2d];
 "4 output" [id=4, type=output];
-"0 _param_constant0" -> "1 add"  [label="(1, 1, 1, 1)", style=solid];
+"0 _conv_w" -> "1 add"  [label="(1, 1, 1, 1)", style=solid];
 "1 add" -> "3 conv2d"  [label="(1, 1, 1, 1)", style=solid];
 "2 conv2d_input" -> "3 conv2d"  [label=None, style=solid];
 "3 conv2d" -> "4 output"  [label="(1, 1, 1, 1)", style=solid];
diff --git a/tests/torch/data/reference_graphs/fx/extracted/ConvolutionWithNotTensorBiasModelconv2d_conv2d.dot b/tests/torch/data/reference_graphs/fx/extracted/ConvolutionWithNotTensorBiasModelconv2d_conv2d.dot
index d3197df425c..8fcafac75d0 100644
--- a/tests/torch/data/reference_graphs/fx/extracted/ConvolutionWithNotTensorBiasModelconv2d_conv2d.dot
+++ b/tests/torch/data/reference_graphs/fx/extracted/ConvolutionWithNotTensorBiasModelconv2d_conv2d.dot
@@ -1,10 +1,10 @@
 strict digraph  {
-"0 _param_constant0" [id=0, type=get_attr];
+"0 _conv_w" [id=0, type=get_attr];
 "1 add" [id=1, type=add];
 "2 conv2d_input" [id=2, type=input];
 "3 conv2d" [id=3, type=conv2d];
 "4 output" [id=4, type=output];
-"0 _param_constant0" -> "1 add"  [label="(1, 1, 1, 1)", style=solid];
+"0 _conv_w" -> "1 add"  [label="(1, 1, 1, 1)", style=solid];
 "1 add" -> "3 conv2d"  [label="(1, 1, 1, 1)", style=solid];
 "2 conv2d_input" -> "3 conv2d"  [label=None, style=solid];
 "3 conv2d" -> "4 output"  [label="(1, 1, 3, 3)", style=solid];
diff --git a/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_1_add__1.dot b/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_1_add__1.dot
index 129ebbbfe63..a646adf8f37 100644
--- a/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_1_add__1.dot
+++ b/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_1_add__1.dot
@@ -1,15 +1,15 @@
 strict digraph  {
-"0 _param_constant2" [id=0, type=get_attr];
-"1 _param_constant3" [id=1, type=get_attr];
-"2 conv2d_1_input" [id=2, type=input];
-"3 conv2d_1" [id=3, type=conv2d];
-"4 _tensor_constant0_1" [id=4, type=get_attr];
+"0 conv_b_weight" [id=0, type=get_attr];
+"1 conv_b_bias" [id=1, type=get_attr];
+"2 bias" [id=2, type=get_attr];
+"3 conv2d_1_input" [id=3, type=input];
+"4 conv2d_1" [id=4, type=conv2d];
 "5 add__1" [id=5, type=add_];
 "6 output" [id=6, type=output];
-"0 _param_constant2" -> "3 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"1 _param_constant3" -> "3 conv2d_1"  [label="(3,)", style=solid];
-"2 conv2d_1_input" -> "3 conv2d_1"  [label=None, style=solid];
-"3 conv2d_1" -> "5 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"4 _tensor_constant0_1" -> "5 add__1"  [label="(1,)", style=solid];
+"0 conv_b_weight" -> "4 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_b_bias" -> "4 conv2d_1"  [label="(3,)", style=solid];
+"2 bias" -> "5 add__1"  [label="(1,)", style=solid];
+"3 conv2d_1_input" -> "4 conv2d_1"  [label=None, style=solid];
+"4 conv2d_1" -> "5 add__1"  [label="(1, 3, 3, 3)", style=solid];
 "5 add__1" -> "6 output"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_add__1.dot b/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_add__1.dot
index 75d148598cb..f22a6293e89 100644
--- a/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_add__1.dot
+++ b/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_add__1.dot
@@ -1,38 +1,36 @@
 strict digraph  {
-"0 _param_constant0" [id=0, type=get_attr];
-"1 _param_constant1" [id=1, type=get_attr];
-"2 conv2d_input" [id=2, type=input];
-"3 conv2d" [id=3, type=conv2d];
-"4 _param_constant2" [id=4, type=get_attr];
-"5 _param_constant3" [id=5, type=get_attr];
-"6 conv2d_1" [id=6, type=conv2d];
-"7 _tensor_constant0" [id=7, type=get_attr];
-"8 add_" [id=8, type=add_];
-"9 _tensor_constant0_1" [id=9, type=get_attr];
-"10 add__1" [id=10, type=add_];
-"11 add" [id=11, type=add];
-"12 _param_constant4" [id=12, type=get_attr];
-"13 _param_constant5" [id=13, type=get_attr];
-"14 conv2d_2" [id=14, type=conv2d];
-"15 _tensor_constant0_2" [id=15, type=get_attr];
-"16 add_1" [id=16, type=add];
-"17 output" [id=17, type=output];
-"0 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"1 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"2 conv2d_input" -> "3 conv2d"  [label=None, style=solid];
-"3 conv2d" -> "6 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"3 conv2d" -> "8 add_"  [label="(1, 3, 3, 3)", style=solid];
-"4 _param_constant2" -> "6 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"5 _param_constant3" -> "6 conv2d_1"  [label="(3,)", style=solid];
-"6 conv2d_1" -> "10 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"7 _tensor_constant0" -> "8 add_"  [label="(1,)", style=solid];
-"8 add_" -> "11 add"  [label="(1, 3, 3, 3)", style=solid];
-"9 _tensor_constant0_1" -> "10 add__1"  [label="(1,)", style=solid];
-"10 add__1" -> "11 add"  [label="(1, 3, 3, 3)", style=solid];
-"10 add__1" -> "17 output"  [label="(1, 3, 3, 3)", style=solid];
-"11 add" -> "14 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"12 _param_constant4" -> "14 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"13 _param_constant5" -> "14 conv2d_2"  [label="(3,)", style=solid];
-"14 conv2d_2" -> "16 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"15 _tensor_constant0_2" -> "16 add_1"  [label="(1,)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 conv2d_input" [id=7, type=input];
+"8 conv2d" [id=8, type=conv2d];
+"9 conv2d_1" [id=9, type=conv2d];
+"10 add_" [id=10, type=add_];
+"11 add__1" [id=11, type=add_];
+"12 add" [id=12, type=add];
+"13 conv2d_2" [id=13, type=conv2d];
+"14 add_1" [id=14, type=add];
+"15 output" [id=15, type=output];
+"0 conv_a_weight" -> "8 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "8 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "9 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "9 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "13 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "13 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "10 add_"  [label="(1,)", style=solid];
+"6 bias" -> "11 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "14 add_1"  [label="(1,)", style=solid];
+"7 conv2d_input" -> "8 conv2d"  [label=None, style=solid];
+"8 conv2d" -> "9 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d" -> "10 add_"  [label="(1, 3, 3, 3)", style=solid];
+"9 conv2d_1" -> "11 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"10 add_" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
+"11 add__1" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
+"11 add__1" -> "15 output"  [label="(1, 3, 3, 3)", style=solid];
+"12 add" -> "13 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"13 conv2d_2" -> "14 add_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_add__add.dot b/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_add__add.dot
index 8f7520bb87d..2b81cd6e8cf 100644
--- a/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_add__add.dot
+++ b/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_add__add.dot
@@ -1,29 +1,28 @@
 strict digraph  {
-"0 _param_constant0" [id=0, type=get_attr];
-"1 _param_constant1" [id=1, type=get_attr];
-"2 conv2d_input" [id=2, type=input];
-"3 conv2d" [id=3, type=conv2d];
-"4 _param_constant2" [id=4, type=get_attr];
-"5 _param_constant3" [id=5, type=get_attr];
-"6 conv2d_1" [id=6, type=conv2d];
-"7 _tensor_constant0" [id=7, type=get_attr];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 bias" [id=4, type=get_attr];
+"5 conv2d_input" [id=5, type=input];
+"6 conv2d" [id=6, type=conv2d];
+"7 conv2d_1" [id=7, type=conv2d];
 "8 add_" [id=8, type=add_];
-"9 _tensor_constant0_1" [id=9, type=get_attr];
-"10 add__1" [id=10, type=add_];
-"11 add" [id=11, type=add];
-"12 output" [id=12, type=output];
-"0 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"1 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"2 conv2d_input" -> "3 conv2d"  [label=None, style=solid];
-"3 conv2d" -> "6 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"3 conv2d" -> "8 add_"  [label="(1, 3, 3, 3)", style=solid];
-"4 _param_constant2" -> "6 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"5 _param_constant3" -> "6 conv2d_1"  [label="(3,)", style=solid];
-"6 conv2d_1" -> "10 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"7 _tensor_constant0" -> "8 add_"  [label="(1,)", style=solid];
-"8 add_" -> "11 add"  [label="(1, 3, 3, 3)", style=solid];
-"8 add_" -> "12 output"  [label="(1, 3, 3, 3)", style=solid];
-"9 _tensor_constant0_1" -> "10 add__1"  [label="(1,)", style=solid];
-"10 add__1" -> "11 add"  [label="(1, 3, 3, 3)", style=solid];
-"11 add" -> "12 output"  [label="(1, 3, 3, 3)", style=solid];
+"9 add__1" [id=9, type=add_];
+"10 add" [id=10, type=add];
+"11 output" [id=11, type=output];
+"0 conv_a_weight" -> "6 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "6 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "7 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "7 conv2d_1"  [label="(3,)", style=solid];
+"4 bias" -> "8 add_"  [label="(1,)", style=solid];
+"4 bias" -> "9 add__1"  [label="(1,)", style=solid];
+"5 conv2d_input" -> "6 conv2d"  [label=None, style=solid];
+"6 conv2d" -> "7 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"6 conv2d" -> "8 add_"  [label="(1, 3, 3, 3)", style=solid];
+"7 conv2d_1" -> "9 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"8 add_" -> "10 add"  [label="(1, 3, 3, 3)", style=solid];
+"8 add_" -> "11 output"  [label="(1, 3, 3, 3)", style=solid];
+"9 add__1" -> "10 add"  [label="(1, 3, 3, 3)", style=solid];
+"10 add" -> "11 output"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_conv2d_1_add__add__1.dot b/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_conv2d_1_add__add__1.dot
index 5774489bd49..5b73eb3f774 100644
--- a/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_conv2d_1_add__add__1.dot
+++ b/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_conv2d_1_add__add__1.dot
@@ -1,27 +1,26 @@
 strict digraph  {
-"0 _param_constant0" [id=0, type=get_attr];
-"1 _param_constant1" [id=1, type=get_attr];
-"2 conv2d_input" [id=2, type=input];
-"3 conv2d" [id=3, type=conv2d];
-"4 _param_constant2" [id=4, type=get_attr];
-"5 _param_constant3" [id=5, type=get_attr];
-"6 conv2d_1_input" [id=6, type=input];
-"7 conv2d_1" [id=7, type=conv2d];
-"8 _tensor_constant0" [id=8, type=get_attr];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 bias" [id=4, type=get_attr];
+"5 conv2d_input" [id=5, type=input];
+"6 conv2d" [id=6, type=conv2d];
+"7 conv2d_1_input" [id=7, type=input];
+"8 conv2d_1" [id=8, type=conv2d];
 "9 add_" [id=9, type=add_];
-"10 _tensor_constant0_1" [id=10, type=get_attr];
-"11 add__1" [id=11, type=add_];
-"12 output" [id=12, type=output];
-"0 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"1 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"2 conv2d_input" -> "3 conv2d"  [label=None, style=solid];
-"3 conv2d" -> "9 add_"  [label="(1, 3, 3, 3)", style=solid];
-"4 _param_constant2" -> "7 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"5 _param_constant3" -> "7 conv2d_1"  [label="(3,)", style=solid];
-"6 conv2d_1_input" -> "7 conv2d_1"  [label=None, style=solid];
-"7 conv2d_1" -> "11 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"8 _tensor_constant0" -> "9 add_"  [label="(1,)", style=solid];
-"9 add_" -> "12 output"  [label="(1, 3, 3, 3)", style=solid];
-"10 _tensor_constant0_1" -> "11 add__1"  [label="(1,)", style=solid];
-"11 add__1" -> "12 output"  [label="(1, 3, 3, 3)", style=solid];
+"10 add__1" [id=10, type=add_];
+"11 output" [id=11, type=output];
+"0 conv_a_weight" -> "6 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "6 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "8 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "8 conv2d_1"  [label="(3,)", style=solid];
+"4 bias" -> "9 add_"  [label="(1,)", style=solid];
+"4 bias" -> "10 add__1"  [label="(1,)", style=solid];
+"5 conv2d_input" -> "6 conv2d"  [label=None, style=solid];
+"6 conv2d" -> "9 add_"  [label="(1, 3, 3, 3)", style=solid];
+"7 conv2d_1_input" -> "8 conv2d_1"  [label=None, style=solid];
+"8 conv2d_1" -> "10 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"9 add_" -> "11 output"  [label="(1, 3, 3, 3)", style=solid];
+"10 add__1" -> "11 output"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_conv2d_2.dot b/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_conv2d_2.dot
index c2c42b388a6..bc231147555 100644
--- a/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_conv2d_2.dot
+++ b/tests/torch/data/reference_graphs/fx/extracted/MultiBranchesConnectedModelconv2d_conv2d_2.dot
@@ -1,34 +1,33 @@
 strict digraph  {
-"0 _param_constant0" [id=0, type=get_attr];
-"1 _param_constant1" [id=1, type=get_attr];
-"2 conv2d_input" [id=2, type=input];
-"3 conv2d" [id=3, type=conv2d];
-"4 _param_constant2" [id=4, type=get_attr];
-"5 _param_constant3" [id=5, type=get_attr];
-"6 conv2d_1" [id=6, type=conv2d];
-"7 _tensor_constant0" [id=7, type=get_attr];
-"8 add_" [id=8, type=add_];
-"9 _tensor_constant0_1" [id=9, type=get_attr];
-"10 add__1" [id=10, type=add_];
-"11 add" [id=11, type=add];
-"12 _param_constant4" [id=12, type=get_attr];
-"13 _param_constant5" [id=13, type=get_attr];
-"14 conv2d_2" [id=14, type=conv2d];
-"15 output" [id=15, type=output];
-"0 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"1 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"2 conv2d_input" -> "3 conv2d"  [label=None, style=solid];
-"3 conv2d" -> "6 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"3 conv2d" -> "8 add_"  [label="(1, 3, 3, 3)", style=solid];
-"4 _param_constant2" -> "6 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"5 _param_constant3" -> "6 conv2d_1"  [label="(3,)", style=solid];
-"6 conv2d_1" -> "10 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"7 _tensor_constant0" -> "8 add_"  [label="(1,)", style=solid];
-"8 add_" -> "11 add"  [label="(1, 3, 3, 3)", style=solid];
-"9 _tensor_constant0_1" -> "10 add__1"  [label="(1,)", style=solid];
-"10 add__1" -> "11 add"  [label="(1, 3, 3, 3)", style=solid];
-"11 add" -> "14 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"12 _param_constant4" -> "14 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"13 _param_constant5" -> "14 conv2d_2"  [label="(3,)", style=solid];
-"14 conv2d_2" -> "15 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 conv2d_input" [id=7, type=input];
+"8 conv2d" [id=8, type=conv2d];
+"9 conv2d_1" [id=9, type=conv2d];
+"10 add_" [id=10, type=add_];
+"11 add__1" [id=11, type=add_];
+"12 add" [id=12, type=add];
+"13 conv2d_2" [id=13, type=conv2d];
+"14 output" [id=14, type=output];
+"0 conv_a_weight" -> "8 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "8 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "9 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "9 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "13 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "13 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "10 add_"  [label="(1,)", style=solid];
+"6 bias" -> "11 add__1"  [label="(1,)", style=solid];
+"7 conv2d_input" -> "8 conv2d"  [label=None, style=solid];
+"8 conv2d" -> "9 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d" -> "10 add_"  [label="(1, 3, 3, 3)", style=solid];
+"9 conv2d_1" -> "11 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"10 add_" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
+"11 add__1" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
+"12 add" -> "13 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"13 conv2d_2" -> "14 output"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/mobilenet_v3_small.dot b/tests/torch/data/reference_graphs/fx/mobilenet_v3_small.dot
index 11ecae8985d..7d74485f281 100644
--- a/tests/torch/data/reference_graphs/fx/mobilenet_v3_small.dot
+++ b/tests/torch/data/reference_graphs/fx/mobilenet_v3_small.dot
@@ -1,992 +1,788 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 conv2d" [id=2, type=conv2d];
-"3 empty" [id=3, type=empty];
-"4 _param_constant1" [id=4, type=get_attr];
-"5 _param_constant2" [id=5, type=get_attr];
-"6 _tensor_constant0" [id=6, type=get_attr];
-"7 _tensor_constant1" [id=7, type=get_attr];
-"8 _native_batch_norm_legit_no_training" [id=8, type=_native_batch_norm_legit_no_training];
-"9 getitem" [id=9, type=__getitem__];
-"10 getitem_1" [id=10, type=__getitem__];
-"11 getitem_2" [id=11, type=__getitem__];
-"12 hardswish_" [id=12, type=hardswish_];
-"13 _param_constant3" [id=13, type=get_attr];
-"14 conv2d_1" [id=14, type=conv2d];
-"15 empty_1" [id=15, type=empty];
-"16 _param_constant4" [id=16, type=get_attr];
-"17 _param_constant5" [id=17, type=get_attr];
-"18 _tensor_constant2" [id=18, type=get_attr];
-"19 _tensor_constant3" [id=19, type=get_attr];
-"20 _native_batch_norm_legit_no_training_1" [id=20, type=_native_batch_norm_legit_no_training];
-"21 getitem_3" [id=21, type=__getitem__];
-"22 getitem_4" [id=22, type=__getitem__];
-"23 getitem_5" [id=23, type=__getitem__];
-"24 relu_" [id=24, type=relu_];
-"25 adaptive_avg_pool2d" [id=25, type=adaptive_avg_pool2d];
-"26 _param_constant6" [id=26, type=get_attr];
-"27 _param_constant7" [id=27, type=get_attr];
-"28 conv2d_2" [id=28, type=conv2d];
-"29 relu" [id=29, type=relu];
-"30 _param_constant8" [id=30, type=get_attr];
-"31 _param_constant9" [id=31, type=get_attr];
-"32 conv2d_3" [id=32, type=conv2d];
-"33 hardsigmoid" [id=33, type=hardsigmoid];
-"34 mul" [id=34, type=mul];
-"35 _param_constant10" [id=35, type=get_attr];
-"36 conv2d_4" [id=36, type=conv2d];
-"37 empty_2" [id=37, type=empty];
-"38 _param_constant11" [id=38, type=get_attr];
-"39 _param_constant12" [id=39, type=get_attr];
-"40 _tensor_constant4" [id=40, type=get_attr];
-"41 _tensor_constant5" [id=41, type=get_attr];
-"42 _native_batch_norm_legit_no_training_2" [id=42, type=_native_batch_norm_legit_no_training];
-"43 getitem_6" [id=43, type=__getitem__];
-"44 getitem_7" [id=44, type=__getitem__];
-"45 getitem_8" [id=45, type=__getitem__];
-"46 _param_constant13" [id=46, type=get_attr];
-"47 conv2d_5" [id=47, type=conv2d];
-"48 empty_3" [id=48, type=empty];
-"49 _param_constant14" [id=49, type=get_attr];
-"50 _param_constant15" [id=50, type=get_attr];
-"51 _tensor_constant6" [id=51, type=get_attr];
-"52 _tensor_constant7" [id=52, type=get_attr];
-"53 _native_batch_norm_legit_no_training_3" [id=53, type=_native_batch_norm_legit_no_training];
-"54 getitem_9" [id=54, type=__getitem__];
-"55 getitem_10" [id=55, type=__getitem__];
-"56 getitem_11" [id=56, type=__getitem__];
-"57 relu__1" [id=57, type=relu_];
-"58 _param_constant16" [id=58, type=get_attr];
-"59 conv2d_6" [id=59, type=conv2d];
-"60 empty_4" [id=60, type=empty];
-"61 _param_constant17" [id=61, type=get_attr];
-"62 _param_constant18" [id=62, type=get_attr];
-"63 _tensor_constant8" [id=63, type=get_attr];
-"64 _tensor_constant9" [id=64, type=get_attr];
-"65 _native_batch_norm_legit_no_training_4" [id=65, type=_native_batch_norm_legit_no_training];
-"66 getitem_12" [id=66, type=__getitem__];
-"67 getitem_13" [id=67, type=__getitem__];
-"68 getitem_14" [id=68, type=__getitem__];
-"69 relu__2" [id=69, type=relu_];
-"70 _param_constant19" [id=70, type=get_attr];
-"71 conv2d_7" [id=71, type=conv2d];
-"72 empty_5" [id=72, type=empty];
-"73 _param_constant20" [id=73, type=get_attr];
-"74 _param_constant21" [id=74, type=get_attr];
-"75 _tensor_constant10" [id=75, type=get_attr];
-"76 _tensor_constant11" [id=76, type=get_attr];
-"77 _native_batch_norm_legit_no_training_5" [id=77, type=_native_batch_norm_legit_no_training];
-"78 getitem_15" [id=78, type=__getitem__];
-"79 getitem_16" [id=79, type=__getitem__];
-"80 getitem_17" [id=80, type=__getitem__];
-"81 _param_constant22" [id=81, type=get_attr];
-"82 conv2d_8" [id=82, type=conv2d];
-"83 empty_6" [id=83, type=empty];
-"84 _param_constant23" [id=84, type=get_attr];
-"85 _param_constant24" [id=85, type=get_attr];
-"86 _tensor_constant12" [id=86, type=get_attr];
-"87 _tensor_constant13" [id=87, type=get_attr];
-"88 _native_batch_norm_legit_no_training_6" [id=88, type=_native_batch_norm_legit_no_training];
-"89 getitem_18" [id=89, type=__getitem__];
-"90 getitem_19" [id=90, type=__getitem__];
-"91 getitem_20" [id=91, type=__getitem__];
-"92 relu__3" [id=92, type=relu_];
-"93 _param_constant25" [id=93, type=get_attr];
-"94 conv2d_9" [id=94, type=conv2d];
-"95 empty_7" [id=95, type=empty];
-"96 _param_constant26" [id=96, type=get_attr];
-"97 _param_constant27" [id=97, type=get_attr];
-"98 _tensor_constant14" [id=98, type=get_attr];
-"99 _tensor_constant15" [id=99, type=get_attr];
-"100 _native_batch_norm_legit_no_training_7" [id=100, type=_native_batch_norm_legit_no_training];
-"101 getitem_21" [id=101, type=__getitem__];
-"102 getitem_22" [id=102, type=__getitem__];
-"103 getitem_23" [id=103, type=__getitem__];
-"104 relu__4" [id=104, type=relu_];
-"105 _param_constant28" [id=105, type=get_attr];
-"106 conv2d_10" [id=106, type=conv2d];
-"107 empty_8" [id=107, type=empty];
-"108 _param_constant29" [id=108, type=get_attr];
-"109 _param_constant30" [id=109, type=get_attr];
-"110 _tensor_constant16" [id=110, type=get_attr];
-"111 _tensor_constant17" [id=111, type=get_attr];
-"112 _native_batch_norm_legit_no_training_8" [id=112, type=_native_batch_norm_legit_no_training];
-"113 getitem_24" [id=113, type=__getitem__];
-"114 getitem_25" [id=114, type=__getitem__];
-"115 getitem_26" [id=115, type=__getitem__];
-"116 add_" [id=116, type=add_];
-"117 _param_constant31" [id=117, type=get_attr];
-"118 conv2d_11" [id=118, type=conv2d];
-"119 empty_9" [id=119, type=empty];
-"120 _param_constant32" [id=120, type=get_attr];
-"121 _param_constant33" [id=121, type=get_attr];
-"122 _tensor_constant18" [id=122, type=get_attr];
-"123 _tensor_constant19" [id=123, type=get_attr];
-"124 _native_batch_norm_legit_no_training_9" [id=124, type=_native_batch_norm_legit_no_training];
-"125 getitem_27" [id=125, type=__getitem__];
-"126 getitem_28" [id=126, type=__getitem__];
-"127 getitem_29" [id=127, type=__getitem__];
-"128 hardswish__1" [id=128, type=hardswish_];
-"129 _param_constant34" [id=129, type=get_attr];
-"130 conv2d_12" [id=130, type=conv2d];
-"131 empty_10" [id=131, type=empty];
-"132 _param_constant35" [id=132, type=get_attr];
-"133 _param_constant36" [id=133, type=get_attr];
-"134 _tensor_constant20" [id=134, type=get_attr];
-"135 _tensor_constant21" [id=135, type=get_attr];
-"136 _native_batch_norm_legit_no_training_10" [id=136, type=_native_batch_norm_legit_no_training];
-"137 getitem_30" [id=137, type=__getitem__];
-"138 getitem_31" [id=138, type=__getitem__];
-"139 getitem_32" [id=139, type=__getitem__];
-"140 hardswish__2" [id=140, type=hardswish_];
-"141 adaptive_avg_pool2d_1" [id=141, type=adaptive_avg_pool2d];
-"142 _param_constant37" [id=142, type=get_attr];
-"143 _param_constant38" [id=143, type=get_attr];
-"144 conv2d_13" [id=144, type=conv2d];
-"145 relu_1" [id=145, type=relu];
-"146 _param_constant39" [id=146, type=get_attr];
-"147 _param_constant40" [id=147, type=get_attr];
-"148 conv2d_14" [id=148, type=conv2d];
-"149 hardsigmoid_1" [id=149, type=hardsigmoid];
-"150 mul_1" [id=150, type=mul];
-"151 _param_constant41" [id=151, type=get_attr];
-"152 conv2d_15" [id=152, type=conv2d];
-"153 empty_11" [id=153, type=empty];
-"154 _param_constant42" [id=154, type=get_attr];
-"155 _param_constant43" [id=155, type=get_attr];
-"156 _tensor_constant22" [id=156, type=get_attr];
-"157 _tensor_constant23" [id=157, type=get_attr];
-"158 _native_batch_norm_legit_no_training_11" [id=158, type=_native_batch_norm_legit_no_training];
-"159 getitem_33" [id=159, type=__getitem__];
-"160 getitem_34" [id=160, type=__getitem__];
-"161 getitem_35" [id=161, type=__getitem__];
-"162 _param_constant44" [id=162, type=get_attr];
-"163 conv2d_16" [id=163, type=conv2d];
-"164 empty_12" [id=164, type=empty];
-"165 _param_constant45" [id=165, type=get_attr];
-"166 _param_constant46" [id=166, type=get_attr];
-"167 _tensor_constant24" [id=167, type=get_attr];
-"168 _tensor_constant25" [id=168, type=get_attr];
-"169 _native_batch_norm_legit_no_training_12" [id=169, type=_native_batch_norm_legit_no_training];
-"170 getitem_36" [id=170, type=__getitem__];
-"171 getitem_37" [id=171, type=__getitem__];
-"172 getitem_38" [id=172, type=__getitem__];
-"173 hardswish__3" [id=173, type=hardswish_];
-"174 _param_constant47" [id=174, type=get_attr];
-"175 conv2d_17" [id=175, type=conv2d];
-"176 empty_13" [id=176, type=empty];
-"177 _param_constant48" [id=177, type=get_attr];
-"178 _param_constant49" [id=178, type=get_attr];
-"179 _tensor_constant26" [id=179, type=get_attr];
-"180 _tensor_constant27" [id=180, type=get_attr];
-"181 _native_batch_norm_legit_no_training_13" [id=181, type=_native_batch_norm_legit_no_training];
-"182 getitem_39" [id=182, type=__getitem__];
-"183 getitem_40" [id=183, type=__getitem__];
-"184 getitem_41" [id=184, type=__getitem__];
-"185 hardswish__4" [id=185, type=hardswish_];
-"186 adaptive_avg_pool2d_2" [id=186, type=adaptive_avg_pool2d];
-"187 _param_constant50" [id=187, type=get_attr];
-"188 _param_constant51" [id=188, type=get_attr];
-"189 conv2d_18" [id=189, type=conv2d];
-"190 relu_2" [id=190, type=relu];
-"191 _param_constant52" [id=191, type=get_attr];
-"192 _param_constant53" [id=192, type=get_attr];
-"193 conv2d_19" [id=193, type=conv2d];
-"194 hardsigmoid_2" [id=194, type=hardsigmoid];
-"195 mul_2" [id=195, type=mul];
-"196 _param_constant54" [id=196, type=get_attr];
-"197 conv2d_20" [id=197, type=conv2d];
-"198 empty_14" [id=198, type=empty];
-"199 _param_constant55" [id=199, type=get_attr];
-"200 _param_constant56" [id=200, type=get_attr];
-"201 _tensor_constant28" [id=201, type=get_attr];
-"202 _tensor_constant29" [id=202, type=get_attr];
-"203 _native_batch_norm_legit_no_training_14" [id=203, type=_native_batch_norm_legit_no_training];
-"204 getitem_42" [id=204, type=__getitem__];
-"205 getitem_43" [id=205, type=__getitem__];
-"206 getitem_44" [id=206, type=__getitem__];
-"207 add__1" [id=207, type=add_];
-"208 _param_constant57" [id=208, type=get_attr];
-"209 conv2d_21" [id=209, type=conv2d];
-"210 empty_15" [id=210, type=empty];
-"211 _param_constant58" [id=211, type=get_attr];
-"212 _param_constant59" [id=212, type=get_attr];
-"213 _tensor_constant30" [id=213, type=get_attr];
-"214 _tensor_constant31" [id=214, type=get_attr];
-"215 _native_batch_norm_legit_no_training_15" [id=215, type=_native_batch_norm_legit_no_training];
-"216 getitem_45" [id=216, type=__getitem__];
-"217 getitem_46" [id=217, type=__getitem__];
-"218 getitem_47" [id=218, type=__getitem__];
-"219 hardswish__5" [id=219, type=hardswish_];
-"220 _param_constant60" [id=220, type=get_attr];
-"221 conv2d_22" [id=221, type=conv2d];
-"222 empty_16" [id=222, type=empty];
-"223 _param_constant61" [id=223, type=get_attr];
-"224 _param_constant62" [id=224, type=get_attr];
-"225 _tensor_constant32" [id=225, type=get_attr];
-"226 _tensor_constant33" [id=226, type=get_attr];
-"227 _native_batch_norm_legit_no_training_16" [id=227, type=_native_batch_norm_legit_no_training];
-"228 getitem_48" [id=228, type=__getitem__];
-"229 getitem_49" [id=229, type=__getitem__];
-"230 getitem_50" [id=230, type=__getitem__];
-"231 hardswish__6" [id=231, type=hardswish_];
-"232 adaptive_avg_pool2d_3" [id=232, type=adaptive_avg_pool2d];
-"233 _param_constant63" [id=233, type=get_attr];
-"234 _param_constant64" [id=234, type=get_attr];
-"235 conv2d_23" [id=235, type=conv2d];
-"236 relu_3" [id=236, type=relu];
-"237 _param_constant65" [id=237, type=get_attr];
-"238 _param_constant66" [id=238, type=get_attr];
-"239 conv2d_24" [id=239, type=conv2d];
-"240 hardsigmoid_3" [id=240, type=hardsigmoid];
-"241 mul_3" [id=241, type=mul];
-"242 _param_constant67" [id=242, type=get_attr];
-"243 conv2d_25" [id=243, type=conv2d];
-"244 empty_17" [id=244, type=empty];
-"245 _param_constant68" [id=245, type=get_attr];
-"246 _param_constant69" [id=246, type=get_attr];
-"247 _tensor_constant34" [id=247, type=get_attr];
-"248 _tensor_constant35" [id=248, type=get_attr];
-"249 _native_batch_norm_legit_no_training_17" [id=249, type=_native_batch_norm_legit_no_training];
-"250 getitem_51" [id=250, type=__getitem__];
-"251 getitem_52" [id=251, type=__getitem__];
-"252 getitem_53" [id=252, type=__getitem__];
-"253 add__2" [id=253, type=add_];
-"254 _param_constant70" [id=254, type=get_attr];
-"255 conv2d_26" [id=255, type=conv2d];
-"256 empty_18" [id=256, type=empty];
-"257 _param_constant71" [id=257, type=get_attr];
-"258 _param_constant72" [id=258, type=get_attr];
-"259 _tensor_constant36" [id=259, type=get_attr];
-"260 _tensor_constant37" [id=260, type=get_attr];
-"261 _native_batch_norm_legit_no_training_18" [id=261, type=_native_batch_norm_legit_no_training];
-"262 getitem_54" [id=262, type=__getitem__];
-"263 getitem_55" [id=263, type=__getitem__];
-"264 getitem_56" [id=264, type=__getitem__];
-"265 hardswish__7" [id=265, type=hardswish_];
-"266 _param_constant73" [id=266, type=get_attr];
-"267 conv2d_27" [id=267, type=conv2d];
-"268 empty_19" [id=268, type=empty];
-"269 _param_constant74" [id=269, type=get_attr];
-"270 _param_constant75" [id=270, type=get_attr];
-"271 _tensor_constant38" [id=271, type=get_attr];
-"272 _tensor_constant39" [id=272, type=get_attr];
-"273 _native_batch_norm_legit_no_training_19" [id=273, type=_native_batch_norm_legit_no_training];
-"274 getitem_57" [id=274, type=__getitem__];
-"275 getitem_58" [id=275, type=__getitem__];
-"276 getitem_59" [id=276, type=__getitem__];
-"277 hardswish__8" [id=277, type=hardswish_];
-"278 adaptive_avg_pool2d_4" [id=278, type=adaptive_avg_pool2d];
-"279 _param_constant76" [id=279, type=get_attr];
-"280 _param_constant77" [id=280, type=get_attr];
-"281 conv2d_28" [id=281, type=conv2d];
-"282 relu_4" [id=282, type=relu];
-"283 _param_constant78" [id=283, type=get_attr];
-"284 _param_constant79" [id=284, type=get_attr];
-"285 conv2d_29" [id=285, type=conv2d];
-"286 hardsigmoid_4" [id=286, type=hardsigmoid];
-"287 mul_4" [id=287, type=mul];
-"288 _param_constant80" [id=288, type=get_attr];
-"289 conv2d_30" [id=289, type=conv2d];
-"290 empty_20" [id=290, type=empty];
-"291 _param_constant81" [id=291, type=get_attr];
-"292 _param_constant82" [id=292, type=get_attr];
-"293 _tensor_constant40" [id=293, type=get_attr];
-"294 _tensor_constant41" [id=294, type=get_attr];
-"295 _native_batch_norm_legit_no_training_20" [id=295, type=_native_batch_norm_legit_no_training];
-"296 getitem_60" [id=296, type=__getitem__];
-"297 getitem_61" [id=297, type=__getitem__];
-"298 getitem_62" [id=298, type=__getitem__];
-"299 _param_constant83" [id=299, type=get_attr];
-"300 conv2d_31" [id=300, type=conv2d];
-"301 empty_21" [id=301, type=empty];
-"302 _param_constant84" [id=302, type=get_attr];
-"303 _param_constant85" [id=303, type=get_attr];
-"304 _tensor_constant42" [id=304, type=get_attr];
-"305 _tensor_constant43" [id=305, type=get_attr];
-"306 _native_batch_norm_legit_no_training_21" [id=306, type=_native_batch_norm_legit_no_training];
-"307 getitem_63" [id=307, type=__getitem__];
-"308 getitem_64" [id=308, type=__getitem__];
-"309 getitem_65" [id=309, type=__getitem__];
-"310 hardswish__9" [id=310, type=hardswish_];
-"311 _param_constant86" [id=311, type=get_attr];
-"312 conv2d_32" [id=312, type=conv2d];
-"313 empty_22" [id=313, type=empty];
-"314 _param_constant87" [id=314, type=get_attr];
-"315 _param_constant88" [id=315, type=get_attr];
-"316 _tensor_constant44" [id=316, type=get_attr];
-"317 _tensor_constant45" [id=317, type=get_attr];
-"318 _native_batch_norm_legit_no_training_22" [id=318, type=_native_batch_norm_legit_no_training];
-"319 getitem_66" [id=319, type=__getitem__];
-"320 getitem_67" [id=320, type=__getitem__];
-"321 getitem_68" [id=321, type=__getitem__];
-"322 hardswish__10" [id=322, type=hardswish_];
-"323 adaptive_avg_pool2d_5" [id=323, type=adaptive_avg_pool2d];
-"324 _param_constant89" [id=324, type=get_attr];
-"325 _param_constant90" [id=325, type=get_attr];
-"326 conv2d_33" [id=326, type=conv2d];
-"327 relu_5" [id=327, type=relu];
-"328 _param_constant91" [id=328, type=get_attr];
-"329 _param_constant92" [id=329, type=get_attr];
-"330 conv2d_34" [id=330, type=conv2d];
-"331 hardsigmoid_5" [id=331, type=hardsigmoid];
-"332 mul_5" [id=332, type=mul];
-"333 _param_constant93" [id=333, type=get_attr];
-"334 conv2d_35" [id=334, type=conv2d];
-"335 empty_23" [id=335, type=empty];
-"336 _param_constant94" [id=336, type=get_attr];
-"337 _param_constant95" [id=337, type=get_attr];
-"338 _tensor_constant46" [id=338, type=get_attr];
-"339 _tensor_constant47" [id=339, type=get_attr];
-"340 _native_batch_norm_legit_no_training_23" [id=340, type=_native_batch_norm_legit_no_training];
-"341 getitem_69" [id=341, type=__getitem__];
-"342 getitem_70" [id=342, type=__getitem__];
-"343 getitem_71" [id=343, type=__getitem__];
-"344 add__3" [id=344, type=add_];
-"345 _param_constant96" [id=345, type=get_attr];
-"346 conv2d_36" [id=346, type=conv2d];
-"347 empty_24" [id=347, type=empty];
-"348 _param_constant97" [id=348, type=get_attr];
-"349 _param_constant98" [id=349, type=get_attr];
-"350 _tensor_constant48" [id=350, type=get_attr];
-"351 _tensor_constant49" [id=351, type=get_attr];
-"352 _native_batch_norm_legit_no_training_24" [id=352, type=_native_batch_norm_legit_no_training];
-"353 getitem_72" [id=353, type=__getitem__];
-"354 getitem_73" [id=354, type=__getitem__];
-"355 getitem_74" [id=355, type=__getitem__];
-"356 hardswish__11" [id=356, type=hardswish_];
-"357 _param_constant99" [id=357, type=get_attr];
-"358 conv2d_37" [id=358, type=conv2d];
-"359 empty_25" [id=359, type=empty];
-"360 _param_constant100" [id=360, type=get_attr];
-"361 _param_constant101" [id=361, type=get_attr];
-"362 _tensor_constant50" [id=362, type=get_attr];
-"363 _tensor_constant51" [id=363, type=get_attr];
-"364 _native_batch_norm_legit_no_training_25" [id=364, type=_native_batch_norm_legit_no_training];
-"365 getitem_75" [id=365, type=__getitem__];
-"366 getitem_76" [id=366, type=__getitem__];
-"367 getitem_77" [id=367, type=__getitem__];
-"368 hardswish__12" [id=368, type=hardswish_];
-"369 adaptive_avg_pool2d_6" [id=369, type=adaptive_avg_pool2d];
-"370 _param_constant102" [id=370, type=get_attr];
-"371 _param_constant103" [id=371, type=get_attr];
-"372 conv2d_38" [id=372, type=conv2d];
-"373 relu_6" [id=373, type=relu];
-"374 _param_constant104" [id=374, type=get_attr];
-"375 _param_constant105" [id=375, type=get_attr];
-"376 conv2d_39" [id=376, type=conv2d];
-"377 hardsigmoid_6" [id=377, type=hardsigmoid];
-"378 mul_6" [id=378, type=mul];
-"379 _param_constant106" [id=379, type=get_attr];
-"380 conv2d_40" [id=380, type=conv2d];
-"381 empty_26" [id=381, type=empty];
-"382 _param_constant107" [id=382, type=get_attr];
-"383 _param_constant108" [id=383, type=get_attr];
-"384 _tensor_constant52" [id=384, type=get_attr];
-"385 _tensor_constant53" [id=385, type=get_attr];
-"386 _native_batch_norm_legit_no_training_26" [id=386, type=_native_batch_norm_legit_no_training];
-"387 getitem_78" [id=387, type=__getitem__];
-"388 getitem_79" [id=388, type=__getitem__];
-"389 getitem_80" [id=389, type=__getitem__];
-"390 _param_constant109" [id=390, type=get_attr];
-"391 conv2d_41" [id=391, type=conv2d];
-"392 empty_27" [id=392, type=empty];
-"393 _param_constant110" [id=393, type=get_attr];
-"394 _param_constant111" [id=394, type=get_attr];
-"395 _tensor_constant54" [id=395, type=get_attr];
-"396 _tensor_constant55" [id=396, type=get_attr];
-"397 _native_batch_norm_legit_no_training_27" [id=397, type=_native_batch_norm_legit_no_training];
-"398 getitem_81" [id=398, type=__getitem__];
-"399 getitem_82" [id=399, type=__getitem__];
-"400 getitem_83" [id=400, type=__getitem__];
-"401 hardswish__13" [id=401, type=hardswish_];
-"402 _param_constant112" [id=402, type=get_attr];
-"403 conv2d_42" [id=403, type=conv2d];
-"404 empty_28" [id=404, type=empty];
-"405 _param_constant113" [id=405, type=get_attr];
-"406 _param_constant114" [id=406, type=get_attr];
-"407 _tensor_constant56" [id=407, type=get_attr];
-"408 _tensor_constant57" [id=408, type=get_attr];
-"409 _native_batch_norm_legit_no_training_28" [id=409, type=_native_batch_norm_legit_no_training];
-"410 getitem_84" [id=410, type=__getitem__];
-"411 getitem_85" [id=411, type=__getitem__];
-"412 getitem_86" [id=412, type=__getitem__];
-"413 hardswish__14" [id=413, type=hardswish_];
-"414 adaptive_avg_pool2d_7" [id=414, type=adaptive_avg_pool2d];
-"415 _param_constant115" [id=415, type=get_attr];
-"416 _param_constant116" [id=416, type=get_attr];
-"417 conv2d_43" [id=417, type=conv2d];
-"418 relu_7" [id=418, type=relu];
-"419 _param_constant117" [id=419, type=get_attr];
-"420 _param_constant118" [id=420, type=get_attr];
-"421 conv2d_44" [id=421, type=conv2d];
-"422 hardsigmoid_7" [id=422, type=hardsigmoid];
-"423 mul_7" [id=423, type=mul];
-"424 _param_constant119" [id=424, type=get_attr];
-"425 conv2d_45" [id=425, type=conv2d];
-"426 empty_29" [id=426, type=empty];
-"427 _param_constant120" [id=427, type=get_attr];
-"428 _param_constant121" [id=428, type=get_attr];
-"429 _tensor_constant58" [id=429, type=get_attr];
-"430 _tensor_constant59" [id=430, type=get_attr];
-"431 _native_batch_norm_legit_no_training_29" [id=431, type=_native_batch_norm_legit_no_training];
-"432 getitem_87" [id=432, type=__getitem__];
-"433 getitem_88" [id=433, type=__getitem__];
-"434 getitem_89" [id=434, type=__getitem__];
-"435 add__4" [id=435, type=add_];
-"436 _param_constant122" [id=436, type=get_attr];
-"437 conv2d_46" [id=437, type=conv2d];
-"438 empty_30" [id=438, type=empty];
-"439 _param_constant123" [id=439, type=get_attr];
-"440 _param_constant124" [id=440, type=get_attr];
-"441 _tensor_constant60" [id=441, type=get_attr];
-"442 _tensor_constant61" [id=442, type=get_attr];
-"443 _native_batch_norm_legit_no_training_30" [id=443, type=_native_batch_norm_legit_no_training];
-"444 getitem_90" [id=444, type=__getitem__];
-"445 getitem_91" [id=445, type=__getitem__];
-"446 getitem_92" [id=446, type=__getitem__];
-"447 hardswish__15" [id=447, type=hardswish_];
-"448 _param_constant125" [id=448, type=get_attr];
-"449 conv2d_47" [id=449, type=conv2d];
-"450 empty_31" [id=450, type=empty];
-"451 _param_constant126" [id=451, type=get_attr];
-"452 _param_constant127" [id=452, type=get_attr];
-"453 _tensor_constant62" [id=453, type=get_attr];
-"454 _tensor_constant63" [id=454, type=get_attr];
-"455 _native_batch_norm_legit_no_training_31" [id=455, type=_native_batch_norm_legit_no_training];
-"456 getitem_93" [id=456, type=__getitem__];
-"457 getitem_94" [id=457, type=__getitem__];
-"458 getitem_95" [id=458, type=__getitem__];
-"459 hardswish__16" [id=459, type=hardswish_];
-"460 adaptive_avg_pool2d_8" [id=460, type=adaptive_avg_pool2d];
-"461 _param_constant128" [id=461, type=get_attr];
-"462 _param_constant129" [id=462, type=get_attr];
-"463 conv2d_48" [id=463, type=conv2d];
-"464 relu_8" [id=464, type=relu];
-"465 _param_constant130" [id=465, type=get_attr];
-"466 _param_constant131" [id=466, type=get_attr];
-"467 conv2d_49" [id=467, type=conv2d];
-"468 hardsigmoid_8" [id=468, type=hardsigmoid];
-"469 mul_8" [id=469, type=mul];
-"470 _param_constant132" [id=470, type=get_attr];
-"471 conv2d_50" [id=471, type=conv2d];
-"472 empty_32" [id=472, type=empty];
-"473 _param_constant133" [id=473, type=get_attr];
-"474 _param_constant134" [id=474, type=get_attr];
-"475 _tensor_constant64" [id=475, type=get_attr];
-"476 _tensor_constant65" [id=476, type=get_attr];
-"477 _native_batch_norm_legit_no_training_32" [id=477, type=_native_batch_norm_legit_no_training];
-"478 getitem_96" [id=478, type=__getitem__];
-"479 getitem_97" [id=479, type=__getitem__];
-"480 getitem_98" [id=480, type=__getitem__];
-"481 add__5" [id=481, type=add_];
-"482 _param_constant135" [id=482, type=get_attr];
-"483 conv2d_51" [id=483, type=conv2d];
-"484 empty_33" [id=484, type=empty];
-"485 _param_constant136" [id=485, type=get_attr];
-"486 _param_constant137" [id=486, type=get_attr];
-"487 _tensor_constant66" [id=487, type=get_attr];
-"488 _tensor_constant67" [id=488, type=get_attr];
-"489 _native_batch_norm_legit_no_training_33" [id=489, type=_native_batch_norm_legit_no_training];
-"490 getitem_99" [id=490, type=__getitem__];
-"491 getitem_100" [id=491, type=__getitem__];
-"492 getitem_101" [id=492, type=__getitem__];
-"493 hardswish__17" [id=493, type=hardswish_];
-"494 adaptive_avg_pool2d_9" [id=494, type=adaptive_avg_pool2d];
-"495 flatten" [id=495, type=flatten];
-"496 _param_constant138" [id=496, type=get_attr];
-"497 _param_constant139" [id=497, type=get_attr];
-"498 linear" [id=498, type=linear];
-"499 hardswish__18" [id=499, type=hardswish_];
-"500 dropout_" [id=500, type=dropout_];
-"501 _param_constant140" [id=501, type=get_attr];
-"502 _param_constant141" [id=502, type=get_attr];
-"503 linear_1" [id=503, type=linear];
-"504 output" [id=504, type=output];
-"0 arg0_1" -> "2 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"1 _param_constant0" -> "2 conv2d"  [label="(16, 3, 3, 3)", style=solid];
-"2 conv2d" -> "8 _native_batch_norm_legit_no_training"  [label="(1, 16, 112, 112)", style=solid];
-"4 _param_constant1" -> "8 _native_batch_norm_legit_no_training"  [label="(16,)", style=solid];
-"5 _param_constant2" -> "8 _native_batch_norm_legit_no_training"  [label="(16,)", style=solid];
-"6 _tensor_constant0" -> "8 _native_batch_norm_legit_no_training"  [label="(16,)", style=solid];
-"7 _tensor_constant1" -> "8 _native_batch_norm_legit_no_training"  [label="(16,)", style=solid];
-"8 _native_batch_norm_legit_no_training" -> "9 getitem"  [label="(1, 16, 112, 112)", style=solid];
-"8 _native_batch_norm_legit_no_training" -> "10 getitem_1"  [label="(1, 16, 112, 112)", style=solid];
-"8 _native_batch_norm_legit_no_training" -> "11 getitem_2"  [label="(1, 16, 112, 112)", style=solid];
-"9 getitem" -> "12 hardswish_"  [label="(1, 16, 112, 112)", style=solid];
-"12 hardswish_" -> "14 conv2d_1"  [label="(1, 16, 112, 112)", style=solid];
-"13 _param_constant3" -> "14 conv2d_1"  [label="(16, 1, 3, 3)", style=solid];
-"14 conv2d_1" -> "20 _native_batch_norm_legit_no_training_1"  [label="(1, 16, 56, 56)", style=solid];
-"16 _param_constant4" -> "20 _native_batch_norm_legit_no_training_1"  [label="(16,)", style=solid];
-"17 _param_constant5" -> "20 _native_batch_norm_legit_no_training_1"  [label="(16,)", style=solid];
-"18 _tensor_constant2" -> "20 _native_batch_norm_legit_no_training_1"  [label="(16,)", style=solid];
-"19 _tensor_constant3" -> "20 _native_batch_norm_legit_no_training_1"  [label="(16,)", style=solid];
-"20 _native_batch_norm_legit_no_training_1" -> "21 getitem_3"  [label="(1, 16, 56, 56)", style=solid];
-"20 _native_batch_norm_legit_no_training_1" -> "22 getitem_4"  [label="(1, 16, 56, 56)", style=solid];
-"20 _native_batch_norm_legit_no_training_1" -> "23 getitem_5"  [label="(1, 16, 56, 56)", style=solid];
-"21 getitem_3" -> "24 relu_"  [label="(1, 16, 56, 56)", style=solid];
-"24 relu_" -> "25 adaptive_avg_pool2d"  [label="(1, 16, 56, 56)", style=solid];
-"24 relu_" -> "34 mul"  [label="(1, 16, 56, 56)", style=solid];
-"25 adaptive_avg_pool2d" -> "28 conv2d_2"  [label="(1, 16, 1, 1)", style=solid];
-"26 _param_constant6" -> "28 conv2d_2"  [label="(8, 16, 1, 1)", style=solid];
-"27 _param_constant7" -> "28 conv2d_2"  [label="(8,)", style=solid];
-"28 conv2d_2" -> "29 relu"  [label="(1, 8, 1, 1)", style=solid];
-"29 relu" -> "32 conv2d_3"  [label="(1, 8, 1, 1)", style=solid];
-"30 _param_constant8" -> "32 conv2d_3"  [label="(16, 8, 1, 1)", style=solid];
-"31 _param_constant9" -> "32 conv2d_3"  [label="(16,)", style=solid];
-"32 conv2d_3" -> "33 hardsigmoid"  [label="(1, 16, 1, 1)", style=solid];
-"33 hardsigmoid" -> "34 mul"  [label="(1, 16, 1, 1)", style=solid];
-"34 mul" -> "36 conv2d_4"  [label="(1, 16, 56, 56)", style=solid];
-"35 _param_constant10" -> "36 conv2d_4"  [label="(16, 16, 1, 1)", style=solid];
-"36 conv2d_4" -> "42 _native_batch_norm_legit_no_training_2"  [label="(1, 16, 56, 56)", style=solid];
-"38 _param_constant11" -> "42 _native_batch_norm_legit_no_training_2"  [label="(16,)", style=solid];
-"39 _param_constant12" -> "42 _native_batch_norm_legit_no_training_2"  [label="(16,)", style=solid];
-"40 _tensor_constant4" -> "42 _native_batch_norm_legit_no_training_2"  [label="(16,)", style=solid];
-"41 _tensor_constant5" -> "42 _native_batch_norm_legit_no_training_2"  [label="(16,)", style=solid];
-"42 _native_batch_norm_legit_no_training_2" -> "43 getitem_6"  [label="(1, 16, 56, 56)", style=solid];
-"42 _native_batch_norm_legit_no_training_2" -> "44 getitem_7"  [label="(1, 16, 56, 56)", style=solid];
-"42 _native_batch_norm_legit_no_training_2" -> "45 getitem_8"  [label="(1, 16, 56, 56)", style=solid];
-"43 getitem_6" -> "47 conv2d_5"  [label="(1, 16, 56, 56)", style=solid];
-"46 _param_constant13" -> "47 conv2d_5"  [label="(72, 16, 1, 1)", style=solid];
-"47 conv2d_5" -> "53 _native_batch_norm_legit_no_training_3"  [label="(1, 72, 56, 56)", style=solid];
-"49 _param_constant14" -> "53 _native_batch_norm_legit_no_training_3"  [label="(72,)", style=solid];
-"50 _param_constant15" -> "53 _native_batch_norm_legit_no_training_3"  [label="(72,)", style=solid];
-"51 _tensor_constant6" -> "53 _native_batch_norm_legit_no_training_3"  [label="(72,)", style=solid];
-"52 _tensor_constant7" -> "53 _native_batch_norm_legit_no_training_3"  [label="(72,)", style=solid];
-"53 _native_batch_norm_legit_no_training_3" -> "54 getitem_9"  [label="(1, 72, 56, 56)", style=solid];
-"53 _native_batch_norm_legit_no_training_3" -> "55 getitem_10"  [label="(1, 72, 56, 56)", style=solid];
-"53 _native_batch_norm_legit_no_training_3" -> "56 getitem_11"  [label="(1, 72, 56, 56)", style=solid];
-"54 getitem_9" -> "57 relu__1"  [label="(1, 72, 56, 56)", style=solid];
-"57 relu__1" -> "59 conv2d_6"  [label="(1, 72, 56, 56)", style=solid];
-"58 _param_constant16" -> "59 conv2d_6"  [label="(72, 1, 3, 3)", style=solid];
-"59 conv2d_6" -> "65 _native_batch_norm_legit_no_training_4"  [label="(1, 72, 28, 28)", style=solid];
-"61 _param_constant17" -> "65 _native_batch_norm_legit_no_training_4"  [label="(72,)", style=solid];
-"62 _param_constant18" -> "65 _native_batch_norm_legit_no_training_4"  [label="(72,)", style=solid];
-"63 _tensor_constant8" -> "65 _native_batch_norm_legit_no_training_4"  [label="(72,)", style=solid];
-"64 _tensor_constant9" -> "65 _native_batch_norm_legit_no_training_4"  [label="(72,)", style=solid];
-"65 _native_batch_norm_legit_no_training_4" -> "66 getitem_12"  [label="(1, 72, 28, 28)", style=solid];
-"65 _native_batch_norm_legit_no_training_4" -> "67 getitem_13"  [label="(1, 72, 28, 28)", style=solid];
-"65 _native_batch_norm_legit_no_training_4" -> "68 getitem_14"  [label="(1, 72, 28, 28)", style=solid];
-"66 getitem_12" -> "69 relu__2"  [label="(1, 72, 28, 28)", style=solid];
-"69 relu__2" -> "71 conv2d_7"  [label="(1, 72, 28, 28)", style=solid];
-"70 _param_constant19" -> "71 conv2d_7"  [label="(24, 72, 1, 1)", style=solid];
-"71 conv2d_7" -> "77 _native_batch_norm_legit_no_training_5"  [label="(1, 24, 28, 28)", style=solid];
-"73 _param_constant20" -> "77 _native_batch_norm_legit_no_training_5"  [label="(24,)", style=solid];
-"74 _param_constant21" -> "77 _native_batch_norm_legit_no_training_5"  [label="(24,)", style=solid];
-"75 _tensor_constant10" -> "77 _native_batch_norm_legit_no_training_5"  [label="(24,)", style=solid];
-"76 _tensor_constant11" -> "77 _native_batch_norm_legit_no_training_5"  [label="(24,)", style=solid];
-"77 _native_batch_norm_legit_no_training_5" -> "78 getitem_15"  [label="(1, 24, 28, 28)", style=solid];
-"77 _native_batch_norm_legit_no_training_5" -> "79 getitem_16"  [label="(1, 24, 28, 28)", style=solid];
-"77 _native_batch_norm_legit_no_training_5" -> "80 getitem_17"  [label="(1, 24, 28, 28)", style=solid];
-"78 getitem_15" -> "82 conv2d_8"  [label="(1, 24, 28, 28)", style=solid];
-"78 getitem_15" -> "116 add_"  [label="(1, 24, 28, 28)", style=solid];
-"81 _param_constant22" -> "82 conv2d_8"  [label="(88, 24, 1, 1)", style=solid];
-"82 conv2d_8" -> "88 _native_batch_norm_legit_no_training_6"  [label="(1, 88, 28, 28)", style=solid];
-"84 _param_constant23" -> "88 _native_batch_norm_legit_no_training_6"  [label="(88,)", style=solid];
-"85 _param_constant24" -> "88 _native_batch_norm_legit_no_training_6"  [label="(88,)", style=solid];
-"86 _tensor_constant12" -> "88 _native_batch_norm_legit_no_training_6"  [label="(88,)", style=solid];
-"87 _tensor_constant13" -> "88 _native_batch_norm_legit_no_training_6"  [label="(88,)", style=solid];
-"88 _native_batch_norm_legit_no_training_6" -> "89 getitem_18"  [label="(1, 88, 28, 28)", style=solid];
-"88 _native_batch_norm_legit_no_training_6" -> "90 getitem_19"  [label="(1, 88, 28, 28)", style=solid];
-"88 _native_batch_norm_legit_no_training_6" -> "91 getitem_20"  [label="(1, 88, 28, 28)", style=solid];
-"89 getitem_18" -> "92 relu__3"  [label="(1, 88, 28, 28)", style=solid];
-"92 relu__3" -> "94 conv2d_9"  [label="(1, 88, 28, 28)", style=solid];
-"93 _param_constant25" -> "94 conv2d_9"  [label="(88, 1, 3, 3)", style=solid];
-"94 conv2d_9" -> "100 _native_batch_norm_legit_no_training_7"  [label="(1, 88, 28, 28)", style=solid];
-"96 _param_constant26" -> "100 _native_batch_norm_legit_no_training_7"  [label="(88,)", style=solid];
-"97 _param_constant27" -> "100 _native_batch_norm_legit_no_training_7"  [label="(88,)", style=solid];
-"98 _tensor_constant14" -> "100 _native_batch_norm_legit_no_training_7"  [label="(88,)", style=solid];
-"99 _tensor_constant15" -> "100 _native_batch_norm_legit_no_training_7"  [label="(88,)", style=solid];
-"100 _native_batch_norm_legit_no_training_7" -> "101 getitem_21"  [label="(1, 88, 28, 28)", style=solid];
-"100 _native_batch_norm_legit_no_training_7" -> "102 getitem_22"  [label="(1, 88, 28, 28)", style=solid];
-"100 _native_batch_norm_legit_no_training_7" -> "103 getitem_23"  [label="(1, 88, 28, 28)", style=solid];
-"101 getitem_21" -> "104 relu__4"  [label="(1, 88, 28, 28)", style=solid];
-"104 relu__4" -> "106 conv2d_10"  [label="(1, 88, 28, 28)", style=solid];
-"105 _param_constant28" -> "106 conv2d_10"  [label="(24, 88, 1, 1)", style=solid];
-"106 conv2d_10" -> "112 _native_batch_norm_legit_no_training_8"  [label="(1, 24, 28, 28)", style=solid];
-"108 _param_constant29" -> "112 _native_batch_norm_legit_no_training_8"  [label="(24,)", style=solid];
-"109 _param_constant30" -> "112 _native_batch_norm_legit_no_training_8"  [label="(24,)", style=solid];
-"110 _tensor_constant16" -> "112 _native_batch_norm_legit_no_training_8"  [label="(24,)", style=solid];
-"111 _tensor_constant17" -> "112 _native_batch_norm_legit_no_training_8"  [label="(24,)", style=solid];
-"112 _native_batch_norm_legit_no_training_8" -> "113 getitem_24"  [label="(1, 24, 28, 28)", style=solid];
-"112 _native_batch_norm_legit_no_training_8" -> "114 getitem_25"  [label="(1, 24, 28, 28)", style=solid];
-"112 _native_batch_norm_legit_no_training_8" -> "115 getitem_26"  [label="(1, 24, 28, 28)", style=solid];
-"113 getitem_24" -> "116 add_"  [label="(1, 24, 28, 28)", style=solid];
-"116 add_" -> "118 conv2d_11"  [label="(1, 24, 28, 28)", style=solid];
-"117 _param_constant31" -> "118 conv2d_11"  [label="(96, 24, 1, 1)", style=solid];
-"118 conv2d_11" -> "124 _native_batch_norm_legit_no_training_9"  [label="(1, 96, 28, 28)", style=solid];
-"120 _param_constant32" -> "124 _native_batch_norm_legit_no_training_9"  [label="(96,)", style=solid];
-"121 _param_constant33" -> "124 _native_batch_norm_legit_no_training_9"  [label="(96,)", style=solid];
-"122 _tensor_constant18" -> "124 _native_batch_norm_legit_no_training_9"  [label="(96,)", style=solid];
-"123 _tensor_constant19" -> "124 _native_batch_norm_legit_no_training_9"  [label="(96,)", style=solid];
-"124 _native_batch_norm_legit_no_training_9" -> "125 getitem_27"  [label="(1, 96, 28, 28)", style=solid];
-"124 _native_batch_norm_legit_no_training_9" -> "126 getitem_28"  [label="(1, 96, 28, 28)", style=solid];
-"124 _native_batch_norm_legit_no_training_9" -> "127 getitem_29"  [label="(1, 96, 28, 28)", style=solid];
-"125 getitem_27" -> "128 hardswish__1"  [label="(1, 96, 28, 28)", style=solid];
-"128 hardswish__1" -> "130 conv2d_12"  [label="(1, 96, 28, 28)", style=solid];
-"129 _param_constant34" -> "130 conv2d_12"  [label="(96, 1, 5, 5)", style=solid];
-"130 conv2d_12" -> "136 _native_batch_norm_legit_no_training_10"  [label="(1, 96, 14, 14)", style=solid];
-"132 _param_constant35" -> "136 _native_batch_norm_legit_no_training_10"  [label="(96,)", style=solid];
-"133 _param_constant36" -> "136 _native_batch_norm_legit_no_training_10"  [label="(96,)", style=solid];
-"134 _tensor_constant20" -> "136 _native_batch_norm_legit_no_training_10"  [label="(96,)", style=solid];
-"135 _tensor_constant21" -> "136 _native_batch_norm_legit_no_training_10"  [label="(96,)", style=solid];
-"136 _native_batch_norm_legit_no_training_10" -> "137 getitem_30"  [label="(1, 96, 14, 14)", style=solid];
-"136 _native_batch_norm_legit_no_training_10" -> "138 getitem_31"  [label="(1, 96, 14, 14)", style=solid];
-"136 _native_batch_norm_legit_no_training_10" -> "139 getitem_32"  [label="(1, 96, 14, 14)", style=solid];
-"137 getitem_30" -> "140 hardswish__2"  [label="(1, 96, 14, 14)", style=solid];
-"140 hardswish__2" -> "141 adaptive_avg_pool2d_1"  [label="(1, 96, 14, 14)", style=solid];
-"140 hardswish__2" -> "150 mul_1"  [label="(1, 96, 14, 14)", style=solid];
-"141 adaptive_avg_pool2d_1" -> "144 conv2d_13"  [label="(1, 96, 1, 1)", style=solid];
-"142 _param_constant37" -> "144 conv2d_13"  [label="(24, 96, 1, 1)", style=solid];
-"143 _param_constant38" -> "144 conv2d_13"  [label="(24,)", style=solid];
-"144 conv2d_13" -> "145 relu_1"  [label="(1, 24, 1, 1)", style=solid];
-"145 relu_1" -> "148 conv2d_14"  [label="(1, 24, 1, 1)", style=solid];
-"146 _param_constant39" -> "148 conv2d_14"  [label="(96, 24, 1, 1)", style=solid];
-"147 _param_constant40" -> "148 conv2d_14"  [label="(96,)", style=solid];
-"148 conv2d_14" -> "149 hardsigmoid_1"  [label="(1, 96, 1, 1)", style=solid];
-"149 hardsigmoid_1" -> "150 mul_1"  [label="(1, 96, 1, 1)", style=solid];
-"150 mul_1" -> "152 conv2d_15"  [label="(1, 96, 14, 14)", style=solid];
-"151 _param_constant41" -> "152 conv2d_15"  [label="(40, 96, 1, 1)", style=solid];
-"152 conv2d_15" -> "158 _native_batch_norm_legit_no_training_11"  [label="(1, 40, 14, 14)", style=solid];
-"154 _param_constant42" -> "158 _native_batch_norm_legit_no_training_11"  [label="(40,)", style=solid];
-"155 _param_constant43" -> "158 _native_batch_norm_legit_no_training_11"  [label="(40,)", style=solid];
-"156 _tensor_constant22" -> "158 _native_batch_norm_legit_no_training_11"  [label="(40,)", style=solid];
-"157 _tensor_constant23" -> "158 _native_batch_norm_legit_no_training_11"  [label="(40,)", style=solid];
-"158 _native_batch_norm_legit_no_training_11" -> "159 getitem_33"  [label="(1, 40, 14, 14)", style=solid];
-"158 _native_batch_norm_legit_no_training_11" -> "160 getitem_34"  [label="(1, 40, 14, 14)", style=solid];
-"158 _native_batch_norm_legit_no_training_11" -> "161 getitem_35"  [label="(1, 40, 14, 14)", style=solid];
-"159 getitem_33" -> "163 conv2d_16"  [label="(1, 40, 14, 14)", style=solid];
-"159 getitem_33" -> "207 add__1"  [label="(1, 40, 14, 14)", style=solid];
-"162 _param_constant44" -> "163 conv2d_16"  [label="(240, 40, 1, 1)", style=solid];
-"163 conv2d_16" -> "169 _native_batch_norm_legit_no_training_12"  [label="(1, 240, 14, 14)", style=solid];
-"165 _param_constant45" -> "169 _native_batch_norm_legit_no_training_12"  [label="(240,)", style=solid];
-"166 _param_constant46" -> "169 _native_batch_norm_legit_no_training_12"  [label="(240,)", style=solid];
-"167 _tensor_constant24" -> "169 _native_batch_norm_legit_no_training_12"  [label="(240,)", style=solid];
-"168 _tensor_constant25" -> "169 _native_batch_norm_legit_no_training_12"  [label="(240,)", style=solid];
-"169 _native_batch_norm_legit_no_training_12" -> "170 getitem_36"  [label="(1, 240, 14, 14)", style=solid];
-"169 _native_batch_norm_legit_no_training_12" -> "171 getitem_37"  [label="(1, 240, 14, 14)", style=solid];
-"169 _native_batch_norm_legit_no_training_12" -> "172 getitem_38"  [label="(1, 240, 14, 14)", style=solid];
-"170 getitem_36" -> "173 hardswish__3"  [label="(1, 240, 14, 14)", style=solid];
-"173 hardswish__3" -> "175 conv2d_17"  [label="(1, 240, 14, 14)", style=solid];
-"174 _param_constant47" -> "175 conv2d_17"  [label="(240, 1, 5, 5)", style=solid];
-"175 conv2d_17" -> "181 _native_batch_norm_legit_no_training_13"  [label="(1, 240, 14, 14)", style=solid];
-"177 _param_constant48" -> "181 _native_batch_norm_legit_no_training_13"  [label="(240,)", style=solid];
-"178 _param_constant49" -> "181 _native_batch_norm_legit_no_training_13"  [label="(240,)", style=solid];
-"179 _tensor_constant26" -> "181 _native_batch_norm_legit_no_training_13"  [label="(240,)", style=solid];
-"180 _tensor_constant27" -> "181 _native_batch_norm_legit_no_training_13"  [label="(240,)", style=solid];
-"181 _native_batch_norm_legit_no_training_13" -> "182 getitem_39"  [label="(1, 240, 14, 14)", style=solid];
-"181 _native_batch_norm_legit_no_training_13" -> "183 getitem_40"  [label="(1, 240, 14, 14)", style=solid];
-"181 _native_batch_norm_legit_no_training_13" -> "184 getitem_41"  [label="(1, 240, 14, 14)", style=solid];
-"182 getitem_39" -> "185 hardswish__4"  [label="(1, 240, 14, 14)", style=solid];
-"185 hardswish__4" -> "186 adaptive_avg_pool2d_2"  [label="(1, 240, 14, 14)", style=solid];
-"185 hardswish__4" -> "195 mul_2"  [label="(1, 240, 14, 14)", style=solid];
-"186 adaptive_avg_pool2d_2" -> "189 conv2d_18"  [label="(1, 240, 1, 1)", style=solid];
-"187 _param_constant50" -> "189 conv2d_18"  [label="(64, 240, 1, 1)", style=solid];
-"188 _param_constant51" -> "189 conv2d_18"  [label="(64,)", style=solid];
-"189 conv2d_18" -> "190 relu_2"  [label="(1, 64, 1, 1)", style=solid];
-"190 relu_2" -> "193 conv2d_19"  [label="(1, 64, 1, 1)", style=solid];
-"191 _param_constant52" -> "193 conv2d_19"  [label="(240, 64, 1, 1)", style=solid];
-"192 _param_constant53" -> "193 conv2d_19"  [label="(240,)", style=solid];
-"193 conv2d_19" -> "194 hardsigmoid_2"  [label="(1, 240, 1, 1)", style=solid];
-"194 hardsigmoid_2" -> "195 mul_2"  [label="(1, 240, 1, 1)", style=solid];
-"195 mul_2" -> "197 conv2d_20"  [label="(1, 240, 14, 14)", style=solid];
-"196 _param_constant54" -> "197 conv2d_20"  [label="(40, 240, 1, 1)", style=solid];
-"197 conv2d_20" -> "203 _native_batch_norm_legit_no_training_14"  [label="(1, 40, 14, 14)", style=solid];
-"199 _param_constant55" -> "203 _native_batch_norm_legit_no_training_14"  [label="(40,)", style=solid];
-"200 _param_constant56" -> "203 _native_batch_norm_legit_no_training_14"  [label="(40,)", style=solid];
-"201 _tensor_constant28" -> "203 _native_batch_norm_legit_no_training_14"  [label="(40,)", style=solid];
-"202 _tensor_constant29" -> "203 _native_batch_norm_legit_no_training_14"  [label="(40,)", style=solid];
-"203 _native_batch_norm_legit_no_training_14" -> "204 getitem_42"  [label="(1, 40, 14, 14)", style=solid];
-"203 _native_batch_norm_legit_no_training_14" -> "205 getitem_43"  [label="(1, 40, 14, 14)", style=solid];
-"203 _native_batch_norm_legit_no_training_14" -> "206 getitem_44"  [label="(1, 40, 14, 14)", style=solid];
-"204 getitem_42" -> "207 add__1"  [label="(1, 40, 14, 14)", style=solid];
-"207 add__1" -> "209 conv2d_21"  [label="(1, 40, 14, 14)", style=solid];
-"207 add__1" -> "253 add__2"  [label="(1, 40, 14, 14)", style=solid];
-"208 _param_constant57" -> "209 conv2d_21"  [label="(240, 40, 1, 1)", style=solid];
-"209 conv2d_21" -> "215 _native_batch_norm_legit_no_training_15"  [label="(1, 240, 14, 14)", style=solid];
-"211 _param_constant58" -> "215 _native_batch_norm_legit_no_training_15"  [label="(240,)", style=solid];
-"212 _param_constant59" -> "215 _native_batch_norm_legit_no_training_15"  [label="(240,)", style=solid];
-"213 _tensor_constant30" -> "215 _native_batch_norm_legit_no_training_15"  [label="(240,)", style=solid];
-"214 _tensor_constant31" -> "215 _native_batch_norm_legit_no_training_15"  [label="(240,)", style=solid];
-"215 _native_batch_norm_legit_no_training_15" -> "216 getitem_45"  [label="(1, 240, 14, 14)", style=solid];
-"215 _native_batch_norm_legit_no_training_15" -> "217 getitem_46"  [label="(1, 240, 14, 14)", style=solid];
-"215 _native_batch_norm_legit_no_training_15" -> "218 getitem_47"  [label="(1, 240, 14, 14)", style=solid];
-"216 getitem_45" -> "219 hardswish__5"  [label="(1, 240, 14, 14)", style=solid];
-"219 hardswish__5" -> "221 conv2d_22"  [label="(1, 240, 14, 14)", style=solid];
-"220 _param_constant60" -> "221 conv2d_22"  [label="(240, 1, 5, 5)", style=solid];
-"221 conv2d_22" -> "227 _native_batch_norm_legit_no_training_16"  [label="(1, 240, 14, 14)", style=solid];
-"223 _param_constant61" -> "227 _native_batch_norm_legit_no_training_16"  [label="(240,)", style=solid];
-"224 _param_constant62" -> "227 _native_batch_norm_legit_no_training_16"  [label="(240,)", style=solid];
-"225 _tensor_constant32" -> "227 _native_batch_norm_legit_no_training_16"  [label="(240,)", style=solid];
-"226 _tensor_constant33" -> "227 _native_batch_norm_legit_no_training_16"  [label="(240,)", style=solid];
-"227 _native_batch_norm_legit_no_training_16" -> "228 getitem_48"  [label="(1, 240, 14, 14)", style=solid];
-"227 _native_batch_norm_legit_no_training_16" -> "229 getitem_49"  [label="(1, 240, 14, 14)", style=solid];
-"227 _native_batch_norm_legit_no_training_16" -> "230 getitem_50"  [label="(1, 240, 14, 14)", style=solid];
-"228 getitem_48" -> "231 hardswish__6"  [label="(1, 240, 14, 14)", style=solid];
-"231 hardswish__6" -> "232 adaptive_avg_pool2d_3"  [label="(1, 240, 14, 14)", style=solid];
-"231 hardswish__6" -> "241 mul_3"  [label="(1, 240, 14, 14)", style=solid];
-"232 adaptive_avg_pool2d_3" -> "235 conv2d_23"  [label="(1, 240, 1, 1)", style=solid];
-"233 _param_constant63" -> "235 conv2d_23"  [label="(64, 240, 1, 1)", style=solid];
-"234 _param_constant64" -> "235 conv2d_23"  [label="(64,)", style=solid];
-"235 conv2d_23" -> "236 relu_3"  [label="(1, 64, 1, 1)", style=solid];
-"236 relu_3" -> "239 conv2d_24"  [label="(1, 64, 1, 1)", style=solid];
-"237 _param_constant65" -> "239 conv2d_24"  [label="(240, 64, 1, 1)", style=solid];
-"238 _param_constant66" -> "239 conv2d_24"  [label="(240,)", style=solid];
-"239 conv2d_24" -> "240 hardsigmoid_3"  [label="(1, 240, 1, 1)", style=solid];
-"240 hardsigmoid_3" -> "241 mul_3"  [label="(1, 240, 1, 1)", style=solid];
-"241 mul_3" -> "243 conv2d_25"  [label="(1, 240, 14, 14)", style=solid];
-"242 _param_constant67" -> "243 conv2d_25"  [label="(40, 240, 1, 1)", style=solid];
-"243 conv2d_25" -> "249 _native_batch_norm_legit_no_training_17"  [label="(1, 40, 14, 14)", style=solid];
-"245 _param_constant68" -> "249 _native_batch_norm_legit_no_training_17"  [label="(40,)", style=solid];
-"246 _param_constant69" -> "249 _native_batch_norm_legit_no_training_17"  [label="(40,)", style=solid];
-"247 _tensor_constant34" -> "249 _native_batch_norm_legit_no_training_17"  [label="(40,)", style=solid];
-"248 _tensor_constant35" -> "249 _native_batch_norm_legit_no_training_17"  [label="(40,)", style=solid];
-"249 _native_batch_norm_legit_no_training_17" -> "250 getitem_51"  [label="(1, 40, 14, 14)", style=solid];
-"249 _native_batch_norm_legit_no_training_17" -> "251 getitem_52"  [label="(1, 40, 14, 14)", style=solid];
-"249 _native_batch_norm_legit_no_training_17" -> "252 getitem_53"  [label="(1, 40, 14, 14)", style=solid];
-"250 getitem_51" -> "253 add__2"  [label="(1, 40, 14, 14)", style=solid];
-"253 add__2" -> "255 conv2d_26"  [label="(1, 40, 14, 14)", style=solid];
-"254 _param_constant70" -> "255 conv2d_26"  [label="(120, 40, 1, 1)", style=solid];
-"255 conv2d_26" -> "261 _native_batch_norm_legit_no_training_18"  [label="(1, 120, 14, 14)", style=solid];
-"257 _param_constant71" -> "261 _native_batch_norm_legit_no_training_18"  [label="(120,)", style=solid];
-"258 _param_constant72" -> "261 _native_batch_norm_legit_no_training_18"  [label="(120,)", style=solid];
-"259 _tensor_constant36" -> "261 _native_batch_norm_legit_no_training_18"  [label="(120,)", style=solid];
-"260 _tensor_constant37" -> "261 _native_batch_norm_legit_no_training_18"  [label="(120,)", style=solid];
-"261 _native_batch_norm_legit_no_training_18" -> "262 getitem_54"  [label="(1, 120, 14, 14)", style=solid];
-"261 _native_batch_norm_legit_no_training_18" -> "263 getitem_55"  [label="(1, 120, 14, 14)", style=solid];
-"261 _native_batch_norm_legit_no_training_18" -> "264 getitem_56"  [label="(1, 120, 14, 14)", style=solid];
-"262 getitem_54" -> "265 hardswish__7"  [label="(1, 120, 14, 14)", style=solid];
-"265 hardswish__7" -> "267 conv2d_27"  [label="(1, 120, 14, 14)", style=solid];
-"266 _param_constant73" -> "267 conv2d_27"  [label="(120, 1, 5, 5)", style=solid];
-"267 conv2d_27" -> "273 _native_batch_norm_legit_no_training_19"  [label="(1, 120, 14, 14)", style=solid];
-"269 _param_constant74" -> "273 _native_batch_norm_legit_no_training_19"  [label="(120,)", style=solid];
-"270 _param_constant75" -> "273 _native_batch_norm_legit_no_training_19"  [label="(120,)", style=solid];
-"271 _tensor_constant38" -> "273 _native_batch_norm_legit_no_training_19"  [label="(120,)", style=solid];
-"272 _tensor_constant39" -> "273 _native_batch_norm_legit_no_training_19"  [label="(120,)", style=solid];
-"273 _native_batch_norm_legit_no_training_19" -> "274 getitem_57"  [label="(1, 120, 14, 14)", style=solid];
-"273 _native_batch_norm_legit_no_training_19" -> "275 getitem_58"  [label="(1, 120, 14, 14)", style=solid];
-"273 _native_batch_norm_legit_no_training_19" -> "276 getitem_59"  [label="(1, 120, 14, 14)", style=solid];
-"274 getitem_57" -> "277 hardswish__8"  [label="(1, 120, 14, 14)", style=solid];
-"277 hardswish__8" -> "278 adaptive_avg_pool2d_4"  [label="(1, 120, 14, 14)", style=solid];
-"277 hardswish__8" -> "287 mul_4"  [label="(1, 120, 14, 14)", style=solid];
-"278 adaptive_avg_pool2d_4" -> "281 conv2d_28"  [label="(1, 120, 1, 1)", style=solid];
-"279 _param_constant76" -> "281 conv2d_28"  [label="(32, 120, 1, 1)", style=solid];
-"280 _param_constant77" -> "281 conv2d_28"  [label="(32,)", style=solid];
-"281 conv2d_28" -> "282 relu_4"  [label="(1, 32, 1, 1)", style=solid];
-"282 relu_4" -> "285 conv2d_29"  [label="(1, 32, 1, 1)", style=solid];
-"283 _param_constant78" -> "285 conv2d_29"  [label="(120, 32, 1, 1)", style=solid];
-"284 _param_constant79" -> "285 conv2d_29"  [label="(120,)", style=solid];
-"285 conv2d_29" -> "286 hardsigmoid_4"  [label="(1, 120, 1, 1)", style=solid];
-"286 hardsigmoid_4" -> "287 mul_4"  [label="(1, 120, 1, 1)", style=solid];
-"287 mul_4" -> "289 conv2d_30"  [label="(1, 120, 14, 14)", style=solid];
-"288 _param_constant80" -> "289 conv2d_30"  [label="(48, 120, 1, 1)", style=solid];
-"289 conv2d_30" -> "295 _native_batch_norm_legit_no_training_20"  [label="(1, 48, 14, 14)", style=solid];
-"291 _param_constant81" -> "295 _native_batch_norm_legit_no_training_20"  [label="(48,)", style=solid];
-"292 _param_constant82" -> "295 _native_batch_norm_legit_no_training_20"  [label="(48,)", style=solid];
-"293 _tensor_constant40" -> "295 _native_batch_norm_legit_no_training_20"  [label="(48,)", style=solid];
-"294 _tensor_constant41" -> "295 _native_batch_norm_legit_no_training_20"  [label="(48,)", style=solid];
-"295 _native_batch_norm_legit_no_training_20" -> "296 getitem_60"  [label="(1, 48, 14, 14)", style=solid];
-"295 _native_batch_norm_legit_no_training_20" -> "297 getitem_61"  [label="(1, 48, 14, 14)", style=solid];
-"295 _native_batch_norm_legit_no_training_20" -> "298 getitem_62"  [label="(1, 48, 14, 14)", style=solid];
-"296 getitem_60" -> "300 conv2d_31"  [label="(1, 48, 14, 14)", style=solid];
-"296 getitem_60" -> "344 add__3"  [label="(1, 48, 14, 14)", style=solid];
-"299 _param_constant83" -> "300 conv2d_31"  [label="(144, 48, 1, 1)", style=solid];
-"300 conv2d_31" -> "306 _native_batch_norm_legit_no_training_21"  [label="(1, 144, 14, 14)", style=solid];
-"302 _param_constant84" -> "306 _native_batch_norm_legit_no_training_21"  [label="(144,)", style=solid];
-"303 _param_constant85" -> "306 _native_batch_norm_legit_no_training_21"  [label="(144,)", style=solid];
-"304 _tensor_constant42" -> "306 _native_batch_norm_legit_no_training_21"  [label="(144,)", style=solid];
-"305 _tensor_constant43" -> "306 _native_batch_norm_legit_no_training_21"  [label="(144,)", style=solid];
-"306 _native_batch_norm_legit_no_training_21" -> "307 getitem_63"  [label="(1, 144, 14, 14)", style=solid];
-"306 _native_batch_norm_legit_no_training_21" -> "308 getitem_64"  [label="(1, 144, 14, 14)", style=solid];
-"306 _native_batch_norm_legit_no_training_21" -> "309 getitem_65"  [label="(1, 144, 14, 14)", style=solid];
-"307 getitem_63" -> "310 hardswish__9"  [label="(1, 144, 14, 14)", style=solid];
-"310 hardswish__9" -> "312 conv2d_32"  [label="(1, 144, 14, 14)", style=solid];
-"311 _param_constant86" -> "312 conv2d_32"  [label="(144, 1, 5, 5)", style=solid];
-"312 conv2d_32" -> "318 _native_batch_norm_legit_no_training_22"  [label="(1, 144, 14, 14)", style=solid];
-"314 _param_constant87" -> "318 _native_batch_norm_legit_no_training_22"  [label="(144,)", style=solid];
-"315 _param_constant88" -> "318 _native_batch_norm_legit_no_training_22"  [label="(144,)", style=solid];
-"316 _tensor_constant44" -> "318 _native_batch_norm_legit_no_training_22"  [label="(144,)", style=solid];
-"317 _tensor_constant45" -> "318 _native_batch_norm_legit_no_training_22"  [label="(144,)", style=solid];
-"318 _native_batch_norm_legit_no_training_22" -> "319 getitem_66"  [label="(1, 144, 14, 14)", style=solid];
-"318 _native_batch_norm_legit_no_training_22" -> "320 getitem_67"  [label="(1, 144, 14, 14)", style=solid];
-"318 _native_batch_norm_legit_no_training_22" -> "321 getitem_68"  [label="(1, 144, 14, 14)", style=solid];
-"319 getitem_66" -> "322 hardswish__10"  [label="(1, 144, 14, 14)", style=solid];
-"322 hardswish__10" -> "323 adaptive_avg_pool2d_5"  [label="(1, 144, 14, 14)", style=solid];
-"322 hardswish__10" -> "332 mul_5"  [label="(1, 144, 14, 14)", style=solid];
-"323 adaptive_avg_pool2d_5" -> "326 conv2d_33"  [label="(1, 144, 1, 1)", style=solid];
-"324 _param_constant89" -> "326 conv2d_33"  [label="(40, 144, 1, 1)", style=solid];
-"325 _param_constant90" -> "326 conv2d_33"  [label="(40,)", style=solid];
-"326 conv2d_33" -> "327 relu_5"  [label="(1, 40, 1, 1)", style=solid];
-"327 relu_5" -> "330 conv2d_34"  [label="(1, 40, 1, 1)", style=solid];
-"328 _param_constant91" -> "330 conv2d_34"  [label="(144, 40, 1, 1)", style=solid];
-"329 _param_constant92" -> "330 conv2d_34"  [label="(144,)", style=solid];
-"330 conv2d_34" -> "331 hardsigmoid_5"  [label="(1, 144, 1, 1)", style=solid];
-"331 hardsigmoid_5" -> "332 mul_5"  [label="(1, 144, 1, 1)", style=solid];
-"332 mul_5" -> "334 conv2d_35"  [label="(1, 144, 14, 14)", style=solid];
-"333 _param_constant93" -> "334 conv2d_35"  [label="(48, 144, 1, 1)", style=solid];
-"334 conv2d_35" -> "340 _native_batch_norm_legit_no_training_23"  [label="(1, 48, 14, 14)", style=solid];
-"336 _param_constant94" -> "340 _native_batch_norm_legit_no_training_23"  [label="(48,)", style=solid];
-"337 _param_constant95" -> "340 _native_batch_norm_legit_no_training_23"  [label="(48,)", style=solid];
-"338 _tensor_constant46" -> "340 _native_batch_norm_legit_no_training_23"  [label="(48,)", style=solid];
-"339 _tensor_constant47" -> "340 _native_batch_norm_legit_no_training_23"  [label="(48,)", style=solid];
-"340 _native_batch_norm_legit_no_training_23" -> "341 getitem_69"  [label="(1, 48, 14, 14)", style=solid];
-"340 _native_batch_norm_legit_no_training_23" -> "342 getitem_70"  [label="(1, 48, 14, 14)", style=solid];
-"340 _native_batch_norm_legit_no_training_23" -> "343 getitem_71"  [label="(1, 48, 14, 14)", style=solid];
-"341 getitem_69" -> "344 add__3"  [label="(1, 48, 14, 14)", style=solid];
-"344 add__3" -> "346 conv2d_36"  [label="(1, 48, 14, 14)", style=solid];
-"345 _param_constant96" -> "346 conv2d_36"  [label="(288, 48, 1, 1)", style=solid];
-"346 conv2d_36" -> "352 _native_batch_norm_legit_no_training_24"  [label="(1, 288, 14, 14)", style=solid];
-"348 _param_constant97" -> "352 _native_batch_norm_legit_no_training_24"  [label="(288,)", style=solid];
-"349 _param_constant98" -> "352 _native_batch_norm_legit_no_training_24"  [label="(288,)", style=solid];
-"350 _tensor_constant48" -> "352 _native_batch_norm_legit_no_training_24"  [label="(288,)", style=solid];
-"351 _tensor_constant49" -> "352 _native_batch_norm_legit_no_training_24"  [label="(288,)", style=solid];
-"352 _native_batch_norm_legit_no_training_24" -> "353 getitem_72"  [label="(1, 288, 14, 14)", style=solid];
-"352 _native_batch_norm_legit_no_training_24" -> "354 getitem_73"  [label="(1, 288, 14, 14)", style=solid];
-"352 _native_batch_norm_legit_no_training_24" -> "355 getitem_74"  [label="(1, 288, 14, 14)", style=solid];
-"353 getitem_72" -> "356 hardswish__11"  [label="(1, 288, 14, 14)", style=solid];
-"356 hardswish__11" -> "358 conv2d_37"  [label="(1, 288, 14, 14)", style=solid];
-"357 _param_constant99" -> "358 conv2d_37"  [label="(288, 1, 5, 5)", style=solid];
-"358 conv2d_37" -> "364 _native_batch_norm_legit_no_training_25"  [label="(1, 288, 7, 7)", style=solid];
-"360 _param_constant100" -> "364 _native_batch_norm_legit_no_training_25"  [label="(288,)", style=solid];
-"361 _param_constant101" -> "364 _native_batch_norm_legit_no_training_25"  [label="(288,)", style=solid];
-"362 _tensor_constant50" -> "364 _native_batch_norm_legit_no_training_25"  [label="(288,)", style=solid];
-"363 _tensor_constant51" -> "364 _native_batch_norm_legit_no_training_25"  [label="(288,)", style=solid];
-"364 _native_batch_norm_legit_no_training_25" -> "365 getitem_75"  [label="(1, 288, 7, 7)", style=solid];
-"364 _native_batch_norm_legit_no_training_25" -> "366 getitem_76"  [label="(1, 288, 7, 7)", style=solid];
-"364 _native_batch_norm_legit_no_training_25" -> "367 getitem_77"  [label="(1, 288, 7, 7)", style=solid];
-"365 getitem_75" -> "368 hardswish__12"  [label="(1, 288, 7, 7)", style=solid];
-"368 hardswish__12" -> "369 adaptive_avg_pool2d_6"  [label="(1, 288, 7, 7)", style=solid];
-"368 hardswish__12" -> "378 mul_6"  [label="(1, 288, 7, 7)", style=solid];
-"369 adaptive_avg_pool2d_6" -> "372 conv2d_38"  [label="(1, 288, 1, 1)", style=solid];
-"370 _param_constant102" -> "372 conv2d_38"  [label="(72, 288, 1, 1)", style=solid];
-"371 _param_constant103" -> "372 conv2d_38"  [label="(72,)", style=solid];
-"372 conv2d_38" -> "373 relu_6"  [label="(1, 72, 1, 1)", style=solid];
-"373 relu_6" -> "376 conv2d_39"  [label="(1, 72, 1, 1)", style=solid];
-"374 _param_constant104" -> "376 conv2d_39"  [label="(288, 72, 1, 1)", style=solid];
-"375 _param_constant105" -> "376 conv2d_39"  [label="(288,)", style=solid];
-"376 conv2d_39" -> "377 hardsigmoid_6"  [label="(1, 288, 1, 1)", style=solid];
-"377 hardsigmoid_6" -> "378 mul_6"  [label="(1, 288, 1, 1)", style=solid];
-"378 mul_6" -> "380 conv2d_40"  [label="(1, 288, 7, 7)", style=solid];
-"379 _param_constant106" -> "380 conv2d_40"  [label="(96, 288, 1, 1)", style=solid];
-"380 conv2d_40" -> "386 _native_batch_norm_legit_no_training_26"  [label="(1, 96, 7, 7)", style=solid];
-"382 _param_constant107" -> "386 _native_batch_norm_legit_no_training_26"  [label="(96,)", style=solid];
-"383 _param_constant108" -> "386 _native_batch_norm_legit_no_training_26"  [label="(96,)", style=solid];
-"384 _tensor_constant52" -> "386 _native_batch_norm_legit_no_training_26"  [label="(96,)", style=solid];
-"385 _tensor_constant53" -> "386 _native_batch_norm_legit_no_training_26"  [label="(96,)", style=solid];
-"386 _native_batch_norm_legit_no_training_26" -> "387 getitem_78"  [label="(1, 96, 7, 7)", style=solid];
-"386 _native_batch_norm_legit_no_training_26" -> "388 getitem_79"  [label="(1, 96, 7, 7)", style=solid];
-"386 _native_batch_norm_legit_no_training_26" -> "389 getitem_80"  [label="(1, 96, 7, 7)", style=solid];
-"387 getitem_78" -> "391 conv2d_41"  [label="(1, 96, 7, 7)", style=solid];
-"387 getitem_78" -> "435 add__4"  [label="(1, 96, 7, 7)", style=solid];
-"390 _param_constant109" -> "391 conv2d_41"  [label="(576, 96, 1, 1)", style=solid];
-"391 conv2d_41" -> "397 _native_batch_norm_legit_no_training_27"  [label="(1, 576, 7, 7)", style=solid];
-"393 _param_constant110" -> "397 _native_batch_norm_legit_no_training_27"  [label="(576,)", style=solid];
-"394 _param_constant111" -> "397 _native_batch_norm_legit_no_training_27"  [label="(576,)", style=solid];
-"395 _tensor_constant54" -> "397 _native_batch_norm_legit_no_training_27"  [label="(576,)", style=solid];
-"396 _tensor_constant55" -> "397 _native_batch_norm_legit_no_training_27"  [label="(576,)", style=solid];
-"397 _native_batch_norm_legit_no_training_27" -> "398 getitem_81"  [label="(1, 576, 7, 7)", style=solid];
-"397 _native_batch_norm_legit_no_training_27" -> "399 getitem_82"  [label="(1, 576, 7, 7)", style=solid];
-"397 _native_batch_norm_legit_no_training_27" -> "400 getitem_83"  [label="(1, 576, 7, 7)", style=solid];
-"398 getitem_81" -> "401 hardswish__13"  [label="(1, 576, 7, 7)", style=solid];
-"401 hardswish__13" -> "403 conv2d_42"  [label="(1, 576, 7, 7)", style=solid];
-"402 _param_constant112" -> "403 conv2d_42"  [label="(576, 1, 5, 5)", style=solid];
-"403 conv2d_42" -> "409 _native_batch_norm_legit_no_training_28"  [label="(1, 576, 7, 7)", style=solid];
-"405 _param_constant113" -> "409 _native_batch_norm_legit_no_training_28"  [label="(576,)", style=solid];
-"406 _param_constant114" -> "409 _native_batch_norm_legit_no_training_28"  [label="(576,)", style=solid];
-"407 _tensor_constant56" -> "409 _native_batch_norm_legit_no_training_28"  [label="(576,)", style=solid];
-"408 _tensor_constant57" -> "409 _native_batch_norm_legit_no_training_28"  [label="(576,)", style=solid];
-"409 _native_batch_norm_legit_no_training_28" -> "410 getitem_84"  [label="(1, 576, 7, 7)", style=solid];
-"409 _native_batch_norm_legit_no_training_28" -> "411 getitem_85"  [label="(1, 576, 7, 7)", style=solid];
-"409 _native_batch_norm_legit_no_training_28" -> "412 getitem_86"  [label="(1, 576, 7, 7)", style=solid];
-"410 getitem_84" -> "413 hardswish__14"  [label="(1, 576, 7, 7)", style=solid];
-"413 hardswish__14" -> "414 adaptive_avg_pool2d_7"  [label="(1, 576, 7, 7)", style=solid];
-"413 hardswish__14" -> "423 mul_7"  [label="(1, 576, 7, 7)", style=solid];
-"414 adaptive_avg_pool2d_7" -> "417 conv2d_43"  [label="(1, 576, 1, 1)", style=solid];
-"415 _param_constant115" -> "417 conv2d_43"  [label="(144, 576, 1, 1)", style=solid];
-"416 _param_constant116" -> "417 conv2d_43"  [label="(144,)", style=solid];
-"417 conv2d_43" -> "418 relu_7"  [label="(1, 144, 1, 1)", style=solid];
-"418 relu_7" -> "421 conv2d_44"  [label="(1, 144, 1, 1)", style=solid];
-"419 _param_constant117" -> "421 conv2d_44"  [label="(576, 144, 1, 1)", style=solid];
-"420 _param_constant118" -> "421 conv2d_44"  [label="(576,)", style=solid];
-"421 conv2d_44" -> "422 hardsigmoid_7"  [label="(1, 576, 1, 1)", style=solid];
-"422 hardsigmoid_7" -> "423 mul_7"  [label="(1, 576, 1, 1)", style=solid];
-"423 mul_7" -> "425 conv2d_45"  [label="(1, 576, 7, 7)", style=solid];
-"424 _param_constant119" -> "425 conv2d_45"  [label="(96, 576, 1, 1)", style=solid];
-"425 conv2d_45" -> "431 _native_batch_norm_legit_no_training_29"  [label="(1, 96, 7, 7)", style=solid];
-"427 _param_constant120" -> "431 _native_batch_norm_legit_no_training_29"  [label="(96,)", style=solid];
-"428 _param_constant121" -> "431 _native_batch_norm_legit_no_training_29"  [label="(96,)", style=solid];
-"429 _tensor_constant58" -> "431 _native_batch_norm_legit_no_training_29"  [label="(96,)", style=solid];
-"430 _tensor_constant59" -> "431 _native_batch_norm_legit_no_training_29"  [label="(96,)", style=solid];
-"431 _native_batch_norm_legit_no_training_29" -> "432 getitem_87"  [label="(1, 96, 7, 7)", style=solid];
-"431 _native_batch_norm_legit_no_training_29" -> "433 getitem_88"  [label="(1, 96, 7, 7)", style=solid];
-"431 _native_batch_norm_legit_no_training_29" -> "434 getitem_89"  [label="(1, 96, 7, 7)", style=solid];
-"432 getitem_87" -> "435 add__4"  [label="(1, 96, 7, 7)", style=solid];
-"435 add__4" -> "437 conv2d_46"  [label="(1, 96, 7, 7)", style=solid];
-"435 add__4" -> "481 add__5"  [label="(1, 96, 7, 7)", style=solid];
-"436 _param_constant122" -> "437 conv2d_46"  [label="(576, 96, 1, 1)", style=solid];
-"437 conv2d_46" -> "443 _native_batch_norm_legit_no_training_30"  [label="(1, 576, 7, 7)", style=solid];
-"439 _param_constant123" -> "443 _native_batch_norm_legit_no_training_30"  [label="(576,)", style=solid];
-"440 _param_constant124" -> "443 _native_batch_norm_legit_no_training_30"  [label="(576,)", style=solid];
-"441 _tensor_constant60" -> "443 _native_batch_norm_legit_no_training_30"  [label="(576,)", style=solid];
-"442 _tensor_constant61" -> "443 _native_batch_norm_legit_no_training_30"  [label="(576,)", style=solid];
-"443 _native_batch_norm_legit_no_training_30" -> "444 getitem_90"  [label="(1, 576, 7, 7)", style=solid];
-"443 _native_batch_norm_legit_no_training_30" -> "445 getitem_91"  [label="(1, 576, 7, 7)", style=solid];
-"443 _native_batch_norm_legit_no_training_30" -> "446 getitem_92"  [label="(1, 576, 7, 7)", style=solid];
-"444 getitem_90" -> "447 hardswish__15"  [label="(1, 576, 7, 7)", style=solid];
-"447 hardswish__15" -> "449 conv2d_47"  [label="(1, 576, 7, 7)", style=solid];
-"448 _param_constant125" -> "449 conv2d_47"  [label="(576, 1, 5, 5)", style=solid];
-"449 conv2d_47" -> "455 _native_batch_norm_legit_no_training_31"  [label="(1, 576, 7, 7)", style=solid];
-"451 _param_constant126" -> "455 _native_batch_norm_legit_no_training_31"  [label="(576,)", style=solid];
-"452 _param_constant127" -> "455 _native_batch_norm_legit_no_training_31"  [label="(576,)", style=solid];
-"453 _tensor_constant62" -> "455 _native_batch_norm_legit_no_training_31"  [label="(576,)", style=solid];
-"454 _tensor_constant63" -> "455 _native_batch_norm_legit_no_training_31"  [label="(576,)", style=solid];
-"455 _native_batch_norm_legit_no_training_31" -> "456 getitem_93"  [label="(1, 576, 7, 7)", style=solid];
-"455 _native_batch_norm_legit_no_training_31" -> "457 getitem_94"  [label="(1, 576, 7, 7)", style=solid];
-"455 _native_batch_norm_legit_no_training_31" -> "458 getitem_95"  [label="(1, 576, 7, 7)", style=solid];
-"456 getitem_93" -> "459 hardswish__16"  [label="(1, 576, 7, 7)", style=solid];
-"459 hardswish__16" -> "460 adaptive_avg_pool2d_8"  [label="(1, 576, 7, 7)", style=solid];
-"459 hardswish__16" -> "469 mul_8"  [label="(1, 576, 7, 7)", style=solid];
-"460 adaptive_avg_pool2d_8" -> "463 conv2d_48"  [label="(1, 576, 1, 1)", style=solid];
-"461 _param_constant128" -> "463 conv2d_48"  [label="(144, 576, 1, 1)", style=solid];
-"462 _param_constant129" -> "463 conv2d_48"  [label="(144,)", style=solid];
-"463 conv2d_48" -> "464 relu_8"  [label="(1, 144, 1, 1)", style=solid];
-"464 relu_8" -> "467 conv2d_49"  [label="(1, 144, 1, 1)", style=solid];
-"465 _param_constant130" -> "467 conv2d_49"  [label="(576, 144, 1, 1)", style=solid];
-"466 _param_constant131" -> "467 conv2d_49"  [label="(576,)", style=solid];
-"467 conv2d_49" -> "468 hardsigmoid_8"  [label="(1, 576, 1, 1)", style=solid];
-"468 hardsigmoid_8" -> "469 mul_8"  [label="(1, 576, 1, 1)", style=solid];
-"469 mul_8" -> "471 conv2d_50"  [label="(1, 576, 7, 7)", style=solid];
-"470 _param_constant132" -> "471 conv2d_50"  [label="(96, 576, 1, 1)", style=solid];
-"471 conv2d_50" -> "477 _native_batch_norm_legit_no_training_32"  [label="(1, 96, 7, 7)", style=solid];
-"473 _param_constant133" -> "477 _native_batch_norm_legit_no_training_32"  [label="(96,)", style=solid];
-"474 _param_constant134" -> "477 _native_batch_norm_legit_no_training_32"  [label="(96,)", style=solid];
-"475 _tensor_constant64" -> "477 _native_batch_norm_legit_no_training_32"  [label="(96,)", style=solid];
-"476 _tensor_constant65" -> "477 _native_batch_norm_legit_no_training_32"  [label="(96,)", style=solid];
-"477 _native_batch_norm_legit_no_training_32" -> "478 getitem_96"  [label="(1, 96, 7, 7)", style=solid];
-"477 _native_batch_norm_legit_no_training_32" -> "479 getitem_97"  [label="(1, 96, 7, 7)", style=solid];
-"477 _native_batch_norm_legit_no_training_32" -> "480 getitem_98"  [label="(1, 96, 7, 7)", style=solid];
-"478 getitem_96" -> "481 add__5"  [label="(1, 96, 7, 7)", style=solid];
-"481 add__5" -> "483 conv2d_51"  [label="(1, 96, 7, 7)", style=solid];
-"482 _param_constant135" -> "483 conv2d_51"  [label="(576, 96, 1, 1)", style=solid];
-"483 conv2d_51" -> "489 _native_batch_norm_legit_no_training_33"  [label="(1, 576, 7, 7)", style=solid];
-"485 _param_constant136" -> "489 _native_batch_norm_legit_no_training_33"  [label="(576,)", style=solid];
-"486 _param_constant137" -> "489 _native_batch_norm_legit_no_training_33"  [label="(576,)", style=solid];
-"487 _tensor_constant66" -> "489 _native_batch_norm_legit_no_training_33"  [label="(576,)", style=solid];
-"488 _tensor_constant67" -> "489 _native_batch_norm_legit_no_training_33"  [label="(576,)", style=solid];
-"489 _native_batch_norm_legit_no_training_33" -> "490 getitem_99"  [label="(1, 576, 7, 7)", style=solid];
-"489 _native_batch_norm_legit_no_training_33" -> "491 getitem_100"  [label="(1, 576, 7, 7)", style=solid];
-"489 _native_batch_norm_legit_no_training_33" -> "492 getitem_101"  [label="(1, 576, 7, 7)", style=solid];
-"490 getitem_99" -> "493 hardswish__17"  [label="(1, 576, 7, 7)", style=solid];
-"493 hardswish__17" -> "494 adaptive_avg_pool2d_9"  [label="(1, 576, 7, 7)", style=solid];
-"494 adaptive_avg_pool2d_9" -> "495 flatten"  [label="(1, 576, 1, 1)", style=solid];
-"495 flatten" -> "498 linear"  [label="(1, 576)", style=solid];
-"496 _param_constant138" -> "498 linear"  [label="(1024, 576)", style=solid];
-"497 _param_constant139" -> "498 linear"  [label="(1024,)", style=solid];
-"498 linear" -> "499 hardswish__18"  [label="(1, 1024)", style=solid];
-"499 hardswish__18" -> "500 dropout_"  [label="(1, 1024)", style=solid];
-"500 dropout_" -> "503 linear_1"  [label="(1, 1024)", style=solid];
-"501 _param_constant140" -> "503 linear_1"  [label="(1000, 1024)", style=solid];
-"502 _param_constant141" -> "503 linear_1"  [label="(1000,)", style=solid];
-"503 linear_1" -> "504 output"  [label="(1, 1000)", style=solid];
+"0 features_0_0_weight" [id=0, type=get_attr];
+"1 features_0_1_weight" [id=1, type=get_attr];
+"2 features_0_1_bias" [id=2, type=get_attr];
+"3 features_1_block_0_0_weight" [id=3, type=get_attr];
+"4 features_1_block_0_1_weight" [id=4, type=get_attr];
+"5 features_1_block_0_1_bias" [id=5, type=get_attr];
+"6 features_1_block_1_fc1_weight" [id=6, type=get_attr];
+"7 features_1_block_1_fc1_bias" [id=7, type=get_attr];
+"8 features_1_block_1_fc2_weight" [id=8, type=get_attr];
+"9 features_1_block_1_fc2_bias" [id=9, type=get_attr];
+"10 features_1_block_2_0_weight" [id=10, type=get_attr];
+"11 features_1_block_2_1_weight" [id=11, type=get_attr];
+"12 features_1_block_2_1_bias" [id=12, type=get_attr];
+"13 features_2_block_0_0_weight" [id=13, type=get_attr];
+"14 features_2_block_0_1_weight" [id=14, type=get_attr];
+"15 features_2_block_0_1_bias" [id=15, type=get_attr];
+"16 features_2_block_1_0_weight" [id=16, type=get_attr];
+"17 features_2_block_1_1_weight" [id=17, type=get_attr];
+"18 features_2_block_1_1_bias" [id=18, type=get_attr];
+"19 features_2_block_2_0_weight" [id=19, type=get_attr];
+"20 features_2_block_2_1_weight" [id=20, type=get_attr];
+"21 features_2_block_2_1_bias" [id=21, type=get_attr];
+"22 features_3_block_0_0_weight" [id=22, type=get_attr];
+"23 features_3_block_0_1_weight" [id=23, type=get_attr];
+"24 features_3_block_0_1_bias" [id=24, type=get_attr];
+"25 features_3_block_1_0_weight" [id=25, type=get_attr];
+"26 features_3_block_1_1_weight" [id=26, type=get_attr];
+"27 features_3_block_1_1_bias" [id=27, type=get_attr];
+"28 features_3_block_2_0_weight" [id=28, type=get_attr];
+"29 features_3_block_2_1_weight" [id=29, type=get_attr];
+"30 features_3_block_2_1_bias" [id=30, type=get_attr];
+"31 features_4_block_0_0_weight" [id=31, type=get_attr];
+"32 features_4_block_0_1_weight" [id=32, type=get_attr];
+"33 features_4_block_0_1_bias" [id=33, type=get_attr];
+"34 features_4_block_1_0_weight" [id=34, type=get_attr];
+"35 features_4_block_1_1_weight" [id=35, type=get_attr];
+"36 features_4_block_1_1_bias" [id=36, type=get_attr];
+"37 features_4_block_2_fc1_weight" [id=37, type=get_attr];
+"38 features_4_block_2_fc1_bias" [id=38, type=get_attr];
+"39 features_4_block_2_fc2_weight" [id=39, type=get_attr];
+"40 features_4_block_2_fc2_bias" [id=40, type=get_attr];
+"41 features_4_block_3_0_weight" [id=41, type=get_attr];
+"42 features_4_block_3_1_weight" [id=42, type=get_attr];
+"43 features_4_block_3_1_bias" [id=43, type=get_attr];
+"44 features_5_block_0_0_weight" [id=44, type=get_attr];
+"45 features_5_block_0_1_weight" [id=45, type=get_attr];
+"46 features_5_block_0_1_bias" [id=46, type=get_attr];
+"47 features_5_block_1_0_weight" [id=47, type=get_attr];
+"48 features_5_block_1_1_weight" [id=48, type=get_attr];
+"49 features_5_block_1_1_bias" [id=49, type=get_attr];
+"50 features_5_block_2_fc1_weight" [id=50, type=get_attr];
+"51 features_5_block_2_fc1_bias" [id=51, type=get_attr];
+"52 features_5_block_2_fc2_weight" [id=52, type=get_attr];
+"53 features_5_block_2_fc2_bias" [id=53, type=get_attr];
+"54 features_5_block_3_0_weight" [id=54, type=get_attr];
+"55 features_5_block_3_1_weight" [id=55, type=get_attr];
+"56 features_5_block_3_1_bias" [id=56, type=get_attr];
+"57 features_6_block_0_0_weight" [id=57, type=get_attr];
+"58 features_6_block_0_1_weight" [id=58, type=get_attr];
+"59 features_6_block_0_1_bias" [id=59, type=get_attr];
+"60 features_6_block_1_0_weight" [id=60, type=get_attr];
+"61 features_6_block_1_1_weight" [id=61, type=get_attr];
+"62 features_6_block_1_1_bias" [id=62, type=get_attr];
+"63 features_6_block_2_fc1_weight" [id=63, type=get_attr];
+"64 features_6_block_2_fc1_bias" [id=64, type=get_attr];
+"65 features_6_block_2_fc2_weight" [id=65, type=get_attr];
+"66 features_6_block_2_fc2_bias" [id=66, type=get_attr];
+"67 features_6_block_3_0_weight" [id=67, type=get_attr];
+"68 features_6_block_3_1_weight" [id=68, type=get_attr];
+"69 features_6_block_3_1_bias" [id=69, type=get_attr];
+"70 features_7_block_0_0_weight" [id=70, type=get_attr];
+"71 features_7_block_0_1_weight" [id=71, type=get_attr];
+"72 features_7_block_0_1_bias" [id=72, type=get_attr];
+"73 features_7_block_1_0_weight" [id=73, type=get_attr];
+"74 features_7_block_1_1_weight" [id=74, type=get_attr];
+"75 features_7_block_1_1_bias" [id=75, type=get_attr];
+"76 features_7_block_2_fc1_weight" [id=76, type=get_attr];
+"77 features_7_block_2_fc1_bias" [id=77, type=get_attr];
+"78 features_7_block_2_fc2_weight" [id=78, type=get_attr];
+"79 features_7_block_2_fc2_bias" [id=79, type=get_attr];
+"80 features_7_block_3_0_weight" [id=80, type=get_attr];
+"81 features_7_block_3_1_weight" [id=81, type=get_attr];
+"82 features_7_block_3_1_bias" [id=82, type=get_attr];
+"83 features_8_block_0_0_weight" [id=83, type=get_attr];
+"84 features_8_block_0_1_weight" [id=84, type=get_attr];
+"85 features_8_block_0_1_bias" [id=85, type=get_attr];
+"86 features_8_block_1_0_weight" [id=86, type=get_attr];
+"87 features_8_block_1_1_weight" [id=87, type=get_attr];
+"88 features_8_block_1_1_bias" [id=88, type=get_attr];
+"89 features_8_block_2_fc1_weight" [id=89, type=get_attr];
+"90 features_8_block_2_fc1_bias" [id=90, type=get_attr];
+"91 features_8_block_2_fc2_weight" [id=91, type=get_attr];
+"92 features_8_block_2_fc2_bias" [id=92, type=get_attr];
+"93 features_8_block_3_0_weight" [id=93, type=get_attr];
+"94 features_8_block_3_1_weight" [id=94, type=get_attr];
+"95 features_8_block_3_1_bias" [id=95, type=get_attr];
+"96 features_9_block_0_0_weight" [id=96, type=get_attr];
+"97 features_9_block_0_1_weight" [id=97, type=get_attr];
+"98 features_9_block_0_1_bias" [id=98, type=get_attr];
+"99 features_9_block_1_0_weight" [id=99, type=get_attr];
+"100 features_9_block_1_1_weight" [id=100, type=get_attr];
+"101 features_9_block_1_1_bias" [id=101, type=get_attr];
+"102 features_9_block_2_fc1_weight" [id=102, type=get_attr];
+"103 features_9_block_2_fc1_bias" [id=103, type=get_attr];
+"104 features_9_block_2_fc2_weight" [id=104, type=get_attr];
+"105 features_9_block_2_fc2_bias" [id=105, type=get_attr];
+"106 features_9_block_3_0_weight" [id=106, type=get_attr];
+"107 features_9_block_3_1_weight" [id=107, type=get_attr];
+"108 features_9_block_3_1_bias" [id=108, type=get_attr];
+"109 features_10_block_0_0_weight" [id=109, type=get_attr];
+"110 features_10_block_0_1_weight" [id=110, type=get_attr];
+"111 features_10_block_0_1_bias" [id=111, type=get_attr];
+"112 features_10_block_1_0_weight" [id=112, type=get_attr];
+"113 features_10_block_1_1_weight" [id=113, type=get_attr];
+"114 features_10_block_1_1_bias" [id=114, type=get_attr];
+"115 features_10_block_2_fc1_weight" [id=115, type=get_attr];
+"116 features_10_block_2_fc1_bias" [id=116, type=get_attr];
+"117 features_10_block_2_fc2_weight" [id=117, type=get_attr];
+"118 features_10_block_2_fc2_bias" [id=118, type=get_attr];
+"119 features_10_block_3_0_weight" [id=119, type=get_attr];
+"120 features_10_block_3_1_weight" [id=120, type=get_attr];
+"121 features_10_block_3_1_bias" [id=121, type=get_attr];
+"122 features_11_block_0_0_weight" [id=122, type=get_attr];
+"123 features_11_block_0_1_weight" [id=123, type=get_attr];
+"124 features_11_block_0_1_bias" [id=124, type=get_attr];
+"125 features_11_block_1_0_weight" [id=125, type=get_attr];
+"126 features_11_block_1_1_weight" [id=126, type=get_attr];
+"127 features_11_block_1_1_bias" [id=127, type=get_attr];
+"128 features_11_block_2_fc1_weight" [id=128, type=get_attr];
+"129 features_11_block_2_fc1_bias" [id=129, type=get_attr];
+"130 features_11_block_2_fc2_weight" [id=130, type=get_attr];
+"131 features_11_block_2_fc2_bias" [id=131, type=get_attr];
+"132 features_11_block_3_0_weight" [id=132, type=get_attr];
+"133 features_11_block_3_1_weight" [id=133, type=get_attr];
+"134 features_11_block_3_1_bias" [id=134, type=get_attr];
+"135 features_12_0_weight" [id=135, type=get_attr];
+"136 features_12_1_weight" [id=136, type=get_attr];
+"137 features_12_1_bias" [id=137, type=get_attr];
+"138 classifier_0_weight" [id=138, type=get_attr];
+"139 classifier_0_bias" [id=139, type=get_attr];
+"140 classifier_3_weight" [id=140, type=get_attr];
+"141 classifier_3_bias" [id=141, type=get_attr];
+"142 features_0_1_running_mean" [id=142, type=get_attr];
+"143 features_0_1_running_var" [id=143, type=get_attr];
+"144 features_0_1_num_batches_tracked" [id=144, type=get_attr];
+"145 features_1_block_0_1_running_mean" [id=145, type=get_attr];
+"146 features_1_block_0_1_running_var" [id=146, type=get_attr];
+"147 features_1_block_0_1_num_batches_tracked" [id=147, type=get_attr];
+"148 features_1_block_2_1_running_mean" [id=148, type=get_attr];
+"149 features_1_block_2_1_running_var" [id=149, type=get_attr];
+"150 features_1_block_2_1_num_batches_tracked" [id=150, type=get_attr];
+"151 features_2_block_0_1_running_mean" [id=151, type=get_attr];
+"152 features_2_block_0_1_running_var" [id=152, type=get_attr];
+"153 features_2_block_0_1_num_batches_tracked" [id=153, type=get_attr];
+"154 features_2_block_1_1_running_mean" [id=154, type=get_attr];
+"155 features_2_block_1_1_running_var" [id=155, type=get_attr];
+"156 features_2_block_1_1_num_batches_tracked" [id=156, type=get_attr];
+"157 features_2_block_2_1_running_mean" [id=157, type=get_attr];
+"158 features_2_block_2_1_running_var" [id=158, type=get_attr];
+"159 features_2_block_2_1_num_batches_tracked" [id=159, type=get_attr];
+"160 features_3_block_0_1_running_mean" [id=160, type=get_attr];
+"161 features_3_block_0_1_running_var" [id=161, type=get_attr];
+"162 features_3_block_0_1_num_batches_tracked" [id=162, type=get_attr];
+"163 features_3_block_1_1_running_mean" [id=163, type=get_attr];
+"164 features_3_block_1_1_running_var" [id=164, type=get_attr];
+"165 features_3_block_1_1_num_batches_tracked" [id=165, type=get_attr];
+"166 features_3_block_2_1_running_mean" [id=166, type=get_attr];
+"167 features_3_block_2_1_running_var" [id=167, type=get_attr];
+"168 features_3_block_2_1_num_batches_tracked" [id=168, type=get_attr];
+"169 features_4_block_0_1_running_mean" [id=169, type=get_attr];
+"170 features_4_block_0_1_running_var" [id=170, type=get_attr];
+"171 features_4_block_0_1_num_batches_tracked" [id=171, type=get_attr];
+"172 features_4_block_1_1_running_mean" [id=172, type=get_attr];
+"173 features_4_block_1_1_running_var" [id=173, type=get_attr];
+"174 features_4_block_1_1_num_batches_tracked" [id=174, type=get_attr];
+"175 features_4_block_3_1_running_mean" [id=175, type=get_attr];
+"176 features_4_block_3_1_running_var" [id=176, type=get_attr];
+"177 features_4_block_3_1_num_batches_tracked" [id=177, type=get_attr];
+"178 features_5_block_0_1_running_mean" [id=178, type=get_attr];
+"179 features_5_block_0_1_running_var" [id=179, type=get_attr];
+"180 features_5_block_0_1_num_batches_tracked" [id=180, type=get_attr];
+"181 features_5_block_1_1_running_mean" [id=181, type=get_attr];
+"182 features_5_block_1_1_running_var" [id=182, type=get_attr];
+"183 features_5_block_1_1_num_batches_tracked" [id=183, type=get_attr];
+"184 features_5_block_3_1_running_mean" [id=184, type=get_attr];
+"185 features_5_block_3_1_running_var" [id=185, type=get_attr];
+"186 features_5_block_3_1_num_batches_tracked" [id=186, type=get_attr];
+"187 features_6_block_0_1_running_mean" [id=187, type=get_attr];
+"188 features_6_block_0_1_running_var" [id=188, type=get_attr];
+"189 features_6_block_0_1_num_batches_tracked" [id=189, type=get_attr];
+"190 features_6_block_1_1_running_mean" [id=190, type=get_attr];
+"191 features_6_block_1_1_running_var" [id=191, type=get_attr];
+"192 features_6_block_1_1_num_batches_tracked" [id=192, type=get_attr];
+"193 features_6_block_3_1_running_mean" [id=193, type=get_attr];
+"194 features_6_block_3_1_running_var" [id=194, type=get_attr];
+"195 features_6_block_3_1_num_batches_tracked" [id=195, type=get_attr];
+"196 features_7_block_0_1_running_mean" [id=196, type=get_attr];
+"197 features_7_block_0_1_running_var" [id=197, type=get_attr];
+"198 features_7_block_0_1_num_batches_tracked" [id=198, type=get_attr];
+"199 features_7_block_1_1_running_mean" [id=199, type=get_attr];
+"200 features_7_block_1_1_running_var" [id=200, type=get_attr];
+"201 features_7_block_1_1_num_batches_tracked" [id=201, type=get_attr];
+"202 features_7_block_3_1_running_mean" [id=202, type=get_attr];
+"203 features_7_block_3_1_running_var" [id=203, type=get_attr];
+"204 features_7_block_3_1_num_batches_tracked" [id=204, type=get_attr];
+"205 features_8_block_0_1_running_mean" [id=205, type=get_attr];
+"206 features_8_block_0_1_running_var" [id=206, type=get_attr];
+"207 features_8_block_0_1_num_batches_tracked" [id=207, type=get_attr];
+"208 features_8_block_1_1_running_mean" [id=208, type=get_attr];
+"209 features_8_block_1_1_running_var" [id=209, type=get_attr];
+"210 features_8_block_1_1_num_batches_tracked" [id=210, type=get_attr];
+"211 features_8_block_3_1_running_mean" [id=211, type=get_attr];
+"212 features_8_block_3_1_running_var" [id=212, type=get_attr];
+"213 features_8_block_3_1_num_batches_tracked" [id=213, type=get_attr];
+"214 features_9_block_0_1_running_mean" [id=214, type=get_attr];
+"215 features_9_block_0_1_running_var" [id=215, type=get_attr];
+"216 features_9_block_0_1_num_batches_tracked" [id=216, type=get_attr];
+"217 features_9_block_1_1_running_mean" [id=217, type=get_attr];
+"218 features_9_block_1_1_running_var" [id=218, type=get_attr];
+"219 features_9_block_1_1_num_batches_tracked" [id=219, type=get_attr];
+"220 features_9_block_3_1_running_mean" [id=220, type=get_attr];
+"221 features_9_block_3_1_running_var" [id=221, type=get_attr];
+"222 features_9_block_3_1_num_batches_tracked" [id=222, type=get_attr];
+"223 features_10_block_0_1_running_mean" [id=223, type=get_attr];
+"224 features_10_block_0_1_running_var" [id=224, type=get_attr];
+"225 features_10_block_0_1_num_batches_tracked" [id=225, type=get_attr];
+"226 features_10_block_1_1_running_mean" [id=226, type=get_attr];
+"227 features_10_block_1_1_running_var" [id=227, type=get_attr];
+"228 features_10_block_1_1_num_batches_tracked" [id=228, type=get_attr];
+"229 features_10_block_3_1_running_mean" [id=229, type=get_attr];
+"230 features_10_block_3_1_running_var" [id=230, type=get_attr];
+"231 features_10_block_3_1_num_batches_tracked" [id=231, type=get_attr];
+"232 features_11_block_0_1_running_mean" [id=232, type=get_attr];
+"233 features_11_block_0_1_running_var" [id=233, type=get_attr];
+"234 features_11_block_0_1_num_batches_tracked" [id=234, type=get_attr];
+"235 features_11_block_1_1_running_mean" [id=235, type=get_attr];
+"236 features_11_block_1_1_running_var" [id=236, type=get_attr];
+"237 features_11_block_1_1_num_batches_tracked" [id=237, type=get_attr];
+"238 features_11_block_3_1_running_mean" [id=238, type=get_attr];
+"239 features_11_block_3_1_running_var" [id=239, type=get_attr];
+"240 features_11_block_3_1_num_batches_tracked" [id=240, type=get_attr];
+"241 features_12_1_running_mean" [id=241, type=get_attr];
+"242 features_12_1_running_var" [id=242, type=get_attr];
+"243 features_12_1_num_batches_tracked" [id=243, type=get_attr];
+"244 x" [id=244, type=input];
+"245 conv2d" [id=245, type=conv2d];
+"246 batch_norm" [id=246, type=batch_norm];
+"247 hardswish_" [id=247, type=hardswish_];
+"248 conv2d_1" [id=248, type=conv2d];
+"249 batch_norm_1" [id=249, type=batch_norm];
+"250 relu_" [id=250, type=relu_];
+"251 adaptive_avg_pool2d" [id=251, type=adaptive_avg_pool2d];
+"252 conv2d_2" [id=252, type=conv2d];
+"253 relu" [id=253, type=relu];
+"254 conv2d_3" [id=254, type=conv2d];
+"255 hardsigmoid" [id=255, type=hardsigmoid];
+"256 mul" [id=256, type=mul];
+"257 conv2d_4" [id=257, type=conv2d];
+"258 batch_norm_2" [id=258, type=batch_norm];
+"259 conv2d_5" [id=259, type=conv2d];
+"260 batch_norm_3" [id=260, type=batch_norm];
+"261 relu__1" [id=261, type=relu_];
+"262 conv2d_6" [id=262, type=conv2d];
+"263 batch_norm_4" [id=263, type=batch_norm];
+"264 relu__2" [id=264, type=relu_];
+"265 conv2d_7" [id=265, type=conv2d];
+"266 batch_norm_5" [id=266, type=batch_norm];
+"267 conv2d_8" [id=267, type=conv2d];
+"268 batch_norm_6" [id=268, type=batch_norm];
+"269 relu__3" [id=269, type=relu_];
+"270 conv2d_9" [id=270, type=conv2d];
+"271 batch_norm_7" [id=271, type=batch_norm];
+"272 relu__4" [id=272, type=relu_];
+"273 conv2d_10" [id=273, type=conv2d];
+"274 batch_norm_8" [id=274, type=batch_norm];
+"275 add_" [id=275, type=add_];
+"276 conv2d_11" [id=276, type=conv2d];
+"277 batch_norm_9" [id=277, type=batch_norm];
+"278 hardswish__1" [id=278, type=hardswish_];
+"279 conv2d_12" [id=279, type=conv2d];
+"280 batch_norm_10" [id=280, type=batch_norm];
+"281 hardswish__2" [id=281, type=hardswish_];
+"282 adaptive_avg_pool2d_1" [id=282, type=adaptive_avg_pool2d];
+"283 conv2d_13" [id=283, type=conv2d];
+"284 relu_1" [id=284, type=relu];
+"285 conv2d_14" [id=285, type=conv2d];
+"286 hardsigmoid_1" [id=286, type=hardsigmoid];
+"287 mul_1" [id=287, type=mul];
+"288 conv2d_15" [id=288, type=conv2d];
+"289 batch_norm_11" [id=289, type=batch_norm];
+"290 conv2d_16" [id=290, type=conv2d];
+"291 batch_norm_12" [id=291, type=batch_norm];
+"292 hardswish__3" [id=292, type=hardswish_];
+"293 conv2d_17" [id=293, type=conv2d];
+"294 batch_norm_13" [id=294, type=batch_norm];
+"295 hardswish__4" [id=295, type=hardswish_];
+"296 adaptive_avg_pool2d_2" [id=296, type=adaptive_avg_pool2d];
+"297 conv2d_18" [id=297, type=conv2d];
+"298 relu_2" [id=298, type=relu];
+"299 conv2d_19" [id=299, type=conv2d];
+"300 hardsigmoid_2" [id=300, type=hardsigmoid];
+"301 mul_2" [id=301, type=mul];
+"302 conv2d_20" [id=302, type=conv2d];
+"303 batch_norm_14" [id=303, type=batch_norm];
+"304 add__1" [id=304, type=add_];
+"305 conv2d_21" [id=305, type=conv2d];
+"306 batch_norm_15" [id=306, type=batch_norm];
+"307 hardswish__5" [id=307, type=hardswish_];
+"308 conv2d_22" [id=308, type=conv2d];
+"309 batch_norm_16" [id=309, type=batch_norm];
+"310 hardswish__6" [id=310, type=hardswish_];
+"311 adaptive_avg_pool2d_3" [id=311, type=adaptive_avg_pool2d];
+"312 conv2d_23" [id=312, type=conv2d];
+"313 relu_3" [id=313, type=relu];
+"314 conv2d_24" [id=314, type=conv2d];
+"315 hardsigmoid_3" [id=315, type=hardsigmoid];
+"316 mul_3" [id=316, type=mul];
+"317 conv2d_25" [id=317, type=conv2d];
+"318 batch_norm_17" [id=318, type=batch_norm];
+"319 add__2" [id=319, type=add_];
+"320 conv2d_26" [id=320, type=conv2d];
+"321 batch_norm_18" [id=321, type=batch_norm];
+"322 hardswish__7" [id=322, type=hardswish_];
+"323 conv2d_27" [id=323, type=conv2d];
+"324 batch_norm_19" [id=324, type=batch_norm];
+"325 hardswish__8" [id=325, type=hardswish_];
+"326 adaptive_avg_pool2d_4" [id=326, type=adaptive_avg_pool2d];
+"327 conv2d_28" [id=327, type=conv2d];
+"328 relu_4" [id=328, type=relu];
+"329 conv2d_29" [id=329, type=conv2d];
+"330 hardsigmoid_4" [id=330, type=hardsigmoid];
+"331 mul_4" [id=331, type=mul];
+"332 conv2d_30" [id=332, type=conv2d];
+"333 batch_norm_20" [id=333, type=batch_norm];
+"334 conv2d_31" [id=334, type=conv2d];
+"335 batch_norm_21" [id=335, type=batch_norm];
+"336 hardswish__9" [id=336, type=hardswish_];
+"337 conv2d_32" [id=337, type=conv2d];
+"338 batch_norm_22" [id=338, type=batch_norm];
+"339 hardswish__10" [id=339, type=hardswish_];
+"340 adaptive_avg_pool2d_5" [id=340, type=adaptive_avg_pool2d];
+"341 conv2d_33" [id=341, type=conv2d];
+"342 relu_5" [id=342, type=relu];
+"343 conv2d_34" [id=343, type=conv2d];
+"344 hardsigmoid_5" [id=344, type=hardsigmoid];
+"345 mul_5" [id=345, type=mul];
+"346 conv2d_35" [id=346, type=conv2d];
+"347 batch_norm_23" [id=347, type=batch_norm];
+"348 add__3" [id=348, type=add_];
+"349 conv2d_36" [id=349, type=conv2d];
+"350 batch_norm_24" [id=350, type=batch_norm];
+"351 hardswish__11" [id=351, type=hardswish_];
+"352 conv2d_37" [id=352, type=conv2d];
+"353 batch_norm_25" [id=353, type=batch_norm];
+"354 hardswish__12" [id=354, type=hardswish_];
+"355 adaptive_avg_pool2d_6" [id=355, type=adaptive_avg_pool2d];
+"356 conv2d_38" [id=356, type=conv2d];
+"357 relu_6" [id=357, type=relu];
+"358 conv2d_39" [id=358, type=conv2d];
+"359 hardsigmoid_6" [id=359, type=hardsigmoid];
+"360 mul_6" [id=360, type=mul];
+"361 conv2d_40" [id=361, type=conv2d];
+"362 batch_norm_26" [id=362, type=batch_norm];
+"363 conv2d_41" [id=363, type=conv2d];
+"364 batch_norm_27" [id=364, type=batch_norm];
+"365 hardswish__13" [id=365, type=hardswish_];
+"366 conv2d_42" [id=366, type=conv2d];
+"367 batch_norm_28" [id=367, type=batch_norm];
+"368 hardswish__14" [id=368, type=hardswish_];
+"369 adaptive_avg_pool2d_7" [id=369, type=adaptive_avg_pool2d];
+"370 conv2d_43" [id=370, type=conv2d];
+"371 relu_7" [id=371, type=relu];
+"372 conv2d_44" [id=372, type=conv2d];
+"373 hardsigmoid_7" [id=373, type=hardsigmoid];
+"374 mul_7" [id=374, type=mul];
+"375 conv2d_45" [id=375, type=conv2d];
+"376 batch_norm_29" [id=376, type=batch_norm];
+"377 add__4" [id=377, type=add_];
+"378 conv2d_46" [id=378, type=conv2d];
+"379 batch_norm_30" [id=379, type=batch_norm];
+"380 hardswish__15" [id=380, type=hardswish_];
+"381 conv2d_47" [id=381, type=conv2d];
+"382 batch_norm_31" [id=382, type=batch_norm];
+"383 hardswish__16" [id=383, type=hardswish_];
+"384 adaptive_avg_pool2d_8" [id=384, type=adaptive_avg_pool2d];
+"385 conv2d_48" [id=385, type=conv2d];
+"386 relu_8" [id=386, type=relu];
+"387 conv2d_49" [id=387, type=conv2d];
+"388 hardsigmoid_8" [id=388, type=hardsigmoid];
+"389 mul_8" [id=389, type=mul];
+"390 conv2d_50" [id=390, type=conv2d];
+"391 batch_norm_32" [id=391, type=batch_norm];
+"392 add__5" [id=392, type=add_];
+"393 conv2d_51" [id=393, type=conv2d];
+"394 batch_norm_33" [id=394, type=batch_norm];
+"395 hardswish__17" [id=395, type=hardswish_];
+"396 adaptive_avg_pool2d_9" [id=396, type=adaptive_avg_pool2d];
+"397 flatten" [id=397, type=flatten];
+"398 linear" [id=398, type=linear];
+"399 hardswish__18" [id=399, type=hardswish_];
+"400 dropout_" [id=400, type=dropout_];
+"401 linear_1" [id=401, type=linear];
+"402 output_1" [id=402, type=output];
+"0 features_0_0_weight" -> "245 conv2d"  [label="(16, 3, 3, 3)", style=solid];
+"1 features_0_1_weight" -> "246 batch_norm"  [label="(16,)", style=solid];
+"2 features_0_1_bias" -> "246 batch_norm"  [label="(16,)", style=solid];
+"3 features_1_block_0_0_weight" -> "248 conv2d_1"  [label="(16, 1, 3, 3)", style=solid];
+"4 features_1_block_0_1_weight" -> "249 batch_norm_1"  [label="(16,)", style=solid];
+"5 features_1_block_0_1_bias" -> "249 batch_norm_1"  [label="(16,)", style=solid];
+"6 features_1_block_1_fc1_weight" -> "252 conv2d_2"  [label="(8, 16, 1, 1)", style=solid];
+"7 features_1_block_1_fc1_bias" -> "252 conv2d_2"  [label="(8,)", style=solid];
+"8 features_1_block_1_fc2_weight" -> "254 conv2d_3"  [label="(16, 8, 1, 1)", style=solid];
+"9 features_1_block_1_fc2_bias" -> "254 conv2d_3"  [label="(16,)", style=solid];
+"10 features_1_block_2_0_weight" -> "257 conv2d_4"  [label="(16, 16, 1, 1)", style=solid];
+"11 features_1_block_2_1_weight" -> "258 batch_norm_2"  [label="(16,)", style=solid];
+"12 features_1_block_2_1_bias" -> "258 batch_norm_2"  [label="(16,)", style=solid];
+"13 features_2_block_0_0_weight" -> "259 conv2d_5"  [label="(72, 16, 1, 1)", style=solid];
+"14 features_2_block_0_1_weight" -> "260 batch_norm_3"  [label="(72,)", style=solid];
+"15 features_2_block_0_1_bias" -> "260 batch_norm_3"  [label="(72,)", style=solid];
+"16 features_2_block_1_0_weight" -> "262 conv2d_6"  [label="(72, 1, 3, 3)", style=solid];
+"17 features_2_block_1_1_weight" -> "263 batch_norm_4"  [label="(72,)", style=solid];
+"18 features_2_block_1_1_bias" -> "263 batch_norm_4"  [label="(72,)", style=solid];
+"19 features_2_block_2_0_weight" -> "265 conv2d_7"  [label="(24, 72, 1, 1)", style=solid];
+"20 features_2_block_2_1_weight" -> "266 batch_norm_5"  [label="(24,)", style=solid];
+"21 features_2_block_2_1_bias" -> "266 batch_norm_5"  [label="(24,)", style=solid];
+"22 features_3_block_0_0_weight" -> "267 conv2d_8"  [label="(88, 24, 1, 1)", style=solid];
+"23 features_3_block_0_1_weight" -> "268 batch_norm_6"  [label="(88,)", style=solid];
+"24 features_3_block_0_1_bias" -> "268 batch_norm_6"  [label="(88,)", style=solid];
+"25 features_3_block_1_0_weight" -> "270 conv2d_9"  [label="(88, 1, 3, 3)", style=solid];
+"26 features_3_block_1_1_weight" -> "271 batch_norm_7"  [label="(88,)", style=solid];
+"27 features_3_block_1_1_bias" -> "271 batch_norm_7"  [label="(88,)", style=solid];
+"28 features_3_block_2_0_weight" -> "273 conv2d_10"  [label="(24, 88, 1, 1)", style=solid];
+"29 features_3_block_2_1_weight" -> "274 batch_norm_8"  [label="(24,)", style=solid];
+"30 features_3_block_2_1_bias" -> "274 batch_norm_8"  [label="(24,)", style=solid];
+"31 features_4_block_0_0_weight" -> "276 conv2d_11"  [label="(96, 24, 1, 1)", style=solid];
+"32 features_4_block_0_1_weight" -> "277 batch_norm_9"  [label="(96,)", style=solid];
+"33 features_4_block_0_1_bias" -> "277 batch_norm_9"  [label="(96,)", style=solid];
+"34 features_4_block_1_0_weight" -> "279 conv2d_12"  [label="(96, 1, 5, 5)", style=solid];
+"35 features_4_block_1_1_weight" -> "280 batch_norm_10"  [label="(96,)", style=solid];
+"36 features_4_block_1_1_bias" -> "280 batch_norm_10"  [label="(96,)", style=solid];
+"37 features_4_block_2_fc1_weight" -> "283 conv2d_13"  [label="(24, 96, 1, 1)", style=solid];
+"38 features_4_block_2_fc1_bias" -> "283 conv2d_13"  [label="(24,)", style=solid];
+"39 features_4_block_2_fc2_weight" -> "285 conv2d_14"  [label="(96, 24, 1, 1)", style=solid];
+"40 features_4_block_2_fc2_bias" -> "285 conv2d_14"  [label="(96,)", style=solid];
+"41 features_4_block_3_0_weight" -> "288 conv2d_15"  [label="(40, 96, 1, 1)", style=solid];
+"42 features_4_block_3_1_weight" -> "289 batch_norm_11"  [label="(40,)", style=solid];
+"43 features_4_block_3_1_bias" -> "289 batch_norm_11"  [label="(40,)", style=solid];
+"44 features_5_block_0_0_weight" -> "290 conv2d_16"  [label="(240, 40, 1, 1)", style=solid];
+"45 features_5_block_0_1_weight" -> "291 batch_norm_12"  [label="(240,)", style=solid];
+"46 features_5_block_0_1_bias" -> "291 batch_norm_12"  [label="(240,)", style=solid];
+"47 features_5_block_1_0_weight" -> "293 conv2d_17"  [label="(240, 1, 5, 5)", style=solid];
+"48 features_5_block_1_1_weight" -> "294 batch_norm_13"  [label="(240,)", style=solid];
+"49 features_5_block_1_1_bias" -> "294 batch_norm_13"  [label="(240,)", style=solid];
+"50 features_5_block_2_fc1_weight" -> "297 conv2d_18"  [label="(64, 240, 1, 1)", style=solid];
+"51 features_5_block_2_fc1_bias" -> "297 conv2d_18"  [label="(64,)", style=solid];
+"52 features_5_block_2_fc2_weight" -> "299 conv2d_19"  [label="(240, 64, 1, 1)", style=solid];
+"53 features_5_block_2_fc2_bias" -> "299 conv2d_19"  [label="(240,)", style=solid];
+"54 features_5_block_3_0_weight" -> "302 conv2d_20"  [label="(40, 240, 1, 1)", style=solid];
+"55 features_5_block_3_1_weight" -> "303 batch_norm_14"  [label="(40,)", style=solid];
+"56 features_5_block_3_1_bias" -> "303 batch_norm_14"  [label="(40,)", style=solid];
+"57 features_6_block_0_0_weight" -> "305 conv2d_21"  [label="(240, 40, 1, 1)", style=solid];
+"58 features_6_block_0_1_weight" -> "306 batch_norm_15"  [label="(240,)", style=solid];
+"59 features_6_block_0_1_bias" -> "306 batch_norm_15"  [label="(240,)", style=solid];
+"60 features_6_block_1_0_weight" -> "308 conv2d_22"  [label="(240, 1, 5, 5)", style=solid];
+"61 features_6_block_1_1_weight" -> "309 batch_norm_16"  [label="(240,)", style=solid];
+"62 features_6_block_1_1_bias" -> "309 batch_norm_16"  [label="(240,)", style=solid];
+"63 features_6_block_2_fc1_weight" -> "312 conv2d_23"  [label="(64, 240, 1, 1)", style=solid];
+"64 features_6_block_2_fc1_bias" -> "312 conv2d_23"  [label="(64,)", style=solid];
+"65 features_6_block_2_fc2_weight" -> "314 conv2d_24"  [label="(240, 64, 1, 1)", style=solid];
+"66 features_6_block_2_fc2_bias" -> "314 conv2d_24"  [label="(240,)", style=solid];
+"67 features_6_block_3_0_weight" -> "317 conv2d_25"  [label="(40, 240, 1, 1)", style=solid];
+"68 features_6_block_3_1_weight" -> "318 batch_norm_17"  [label="(40,)", style=solid];
+"69 features_6_block_3_1_bias" -> "318 batch_norm_17"  [label="(40,)", style=solid];
+"70 features_7_block_0_0_weight" -> "320 conv2d_26"  [label="(120, 40, 1, 1)", style=solid];
+"71 features_7_block_0_1_weight" -> "321 batch_norm_18"  [label="(120,)", style=solid];
+"72 features_7_block_0_1_bias" -> "321 batch_norm_18"  [label="(120,)", style=solid];
+"73 features_7_block_1_0_weight" -> "323 conv2d_27"  [label="(120, 1, 5, 5)", style=solid];
+"74 features_7_block_1_1_weight" -> "324 batch_norm_19"  [label="(120,)", style=solid];
+"75 features_7_block_1_1_bias" -> "324 batch_norm_19"  [label="(120,)", style=solid];
+"76 features_7_block_2_fc1_weight" -> "327 conv2d_28"  [label="(32, 120, 1, 1)", style=solid];
+"77 features_7_block_2_fc1_bias" -> "327 conv2d_28"  [label="(32,)", style=solid];
+"78 features_7_block_2_fc2_weight" -> "329 conv2d_29"  [label="(120, 32, 1, 1)", style=solid];
+"79 features_7_block_2_fc2_bias" -> "329 conv2d_29"  [label="(120,)", style=solid];
+"80 features_7_block_3_0_weight" -> "332 conv2d_30"  [label="(48, 120, 1, 1)", style=solid];
+"81 features_7_block_3_1_weight" -> "333 batch_norm_20"  [label="(48,)", style=solid];
+"82 features_7_block_3_1_bias" -> "333 batch_norm_20"  [label="(48,)", style=solid];
+"83 features_8_block_0_0_weight" -> "334 conv2d_31"  [label="(144, 48, 1, 1)", style=solid];
+"84 features_8_block_0_1_weight" -> "335 batch_norm_21"  [label="(144,)", style=solid];
+"85 features_8_block_0_1_bias" -> "335 batch_norm_21"  [label="(144,)", style=solid];
+"86 features_8_block_1_0_weight" -> "337 conv2d_32"  [label="(144, 1, 5, 5)", style=solid];
+"87 features_8_block_1_1_weight" -> "338 batch_norm_22"  [label="(144,)", style=solid];
+"88 features_8_block_1_1_bias" -> "338 batch_norm_22"  [label="(144,)", style=solid];
+"89 features_8_block_2_fc1_weight" -> "341 conv2d_33"  [label="(40, 144, 1, 1)", style=solid];
+"90 features_8_block_2_fc1_bias" -> "341 conv2d_33"  [label="(40,)", style=solid];
+"91 features_8_block_2_fc2_weight" -> "343 conv2d_34"  [label="(144, 40, 1, 1)", style=solid];
+"92 features_8_block_2_fc2_bias" -> "343 conv2d_34"  [label="(144,)", style=solid];
+"93 features_8_block_3_0_weight" -> "346 conv2d_35"  [label="(48, 144, 1, 1)", style=solid];
+"94 features_8_block_3_1_weight" -> "347 batch_norm_23"  [label="(48,)", style=solid];
+"95 features_8_block_3_1_bias" -> "347 batch_norm_23"  [label="(48,)", style=solid];
+"96 features_9_block_0_0_weight" -> "349 conv2d_36"  [label="(288, 48, 1, 1)", style=solid];
+"97 features_9_block_0_1_weight" -> "350 batch_norm_24"  [label="(288,)", style=solid];
+"98 features_9_block_0_1_bias" -> "350 batch_norm_24"  [label="(288,)", style=solid];
+"99 features_9_block_1_0_weight" -> "352 conv2d_37"  [label="(288, 1, 5, 5)", style=solid];
+"100 features_9_block_1_1_weight" -> "353 batch_norm_25"  [label="(288,)", style=solid];
+"101 features_9_block_1_1_bias" -> "353 batch_norm_25"  [label="(288,)", style=solid];
+"102 features_9_block_2_fc1_weight" -> "356 conv2d_38"  [label="(72, 288, 1, 1)", style=solid];
+"103 features_9_block_2_fc1_bias" -> "356 conv2d_38"  [label="(72,)", style=solid];
+"104 features_9_block_2_fc2_weight" -> "358 conv2d_39"  [label="(288, 72, 1, 1)", style=solid];
+"105 features_9_block_2_fc2_bias" -> "358 conv2d_39"  [label="(288,)", style=solid];
+"106 features_9_block_3_0_weight" -> "361 conv2d_40"  [label="(96, 288, 1, 1)", style=solid];
+"107 features_9_block_3_1_weight" -> "362 batch_norm_26"  [label="(96,)", style=solid];
+"108 features_9_block_3_1_bias" -> "362 batch_norm_26"  [label="(96,)", style=solid];
+"109 features_10_block_0_0_weight" -> "363 conv2d_41"  [label="(576, 96, 1, 1)", style=solid];
+"110 features_10_block_0_1_weight" -> "364 batch_norm_27"  [label="(576,)", style=solid];
+"111 features_10_block_0_1_bias" -> "364 batch_norm_27"  [label="(576,)", style=solid];
+"112 features_10_block_1_0_weight" -> "366 conv2d_42"  [label="(576, 1, 5, 5)", style=solid];
+"113 features_10_block_1_1_weight" -> "367 batch_norm_28"  [label="(576,)", style=solid];
+"114 features_10_block_1_1_bias" -> "367 batch_norm_28"  [label="(576,)", style=solid];
+"115 features_10_block_2_fc1_weight" -> "370 conv2d_43"  [label="(144, 576, 1, 1)", style=solid];
+"116 features_10_block_2_fc1_bias" -> "370 conv2d_43"  [label="(144,)", style=solid];
+"117 features_10_block_2_fc2_weight" -> "372 conv2d_44"  [label="(576, 144, 1, 1)", style=solid];
+"118 features_10_block_2_fc2_bias" -> "372 conv2d_44"  [label="(576,)", style=solid];
+"119 features_10_block_3_0_weight" -> "375 conv2d_45"  [label="(96, 576, 1, 1)", style=solid];
+"120 features_10_block_3_1_weight" -> "376 batch_norm_29"  [label="(96,)", style=solid];
+"121 features_10_block_3_1_bias" -> "376 batch_norm_29"  [label="(96,)", style=solid];
+"122 features_11_block_0_0_weight" -> "378 conv2d_46"  [label="(576, 96, 1, 1)", style=solid];
+"123 features_11_block_0_1_weight" -> "379 batch_norm_30"  [label="(576,)", style=solid];
+"124 features_11_block_0_1_bias" -> "379 batch_norm_30"  [label="(576,)", style=solid];
+"125 features_11_block_1_0_weight" -> "381 conv2d_47"  [label="(576, 1, 5, 5)", style=solid];
+"126 features_11_block_1_1_weight" -> "382 batch_norm_31"  [label="(576,)", style=solid];
+"127 features_11_block_1_1_bias" -> "382 batch_norm_31"  [label="(576,)", style=solid];
+"128 features_11_block_2_fc1_weight" -> "385 conv2d_48"  [label="(144, 576, 1, 1)", style=solid];
+"129 features_11_block_2_fc1_bias" -> "385 conv2d_48"  [label="(144,)", style=solid];
+"130 features_11_block_2_fc2_weight" -> "387 conv2d_49"  [label="(576, 144, 1, 1)", style=solid];
+"131 features_11_block_2_fc2_bias" -> "387 conv2d_49"  [label="(576,)", style=solid];
+"132 features_11_block_3_0_weight" -> "390 conv2d_50"  [label="(96, 576, 1, 1)", style=solid];
+"133 features_11_block_3_1_weight" -> "391 batch_norm_32"  [label="(96,)", style=solid];
+"134 features_11_block_3_1_bias" -> "391 batch_norm_32"  [label="(96,)", style=solid];
+"135 features_12_0_weight" -> "393 conv2d_51"  [label="(576, 96, 1, 1)", style=solid];
+"136 features_12_1_weight" -> "394 batch_norm_33"  [label="(576,)", style=solid];
+"137 features_12_1_bias" -> "394 batch_norm_33"  [label="(576,)", style=solid];
+"138 classifier_0_weight" -> "398 linear"  [label="(1024, 576)", style=solid];
+"139 classifier_0_bias" -> "398 linear"  [label="(1024,)", style=solid];
+"140 classifier_3_weight" -> "401 linear_1"  [label="(1000, 1024)", style=solid];
+"141 classifier_3_bias" -> "401 linear_1"  [label="(1000,)", style=solid];
+"142 features_0_1_running_mean" -> "246 batch_norm"  [label="(16,)", style=solid];
+"143 features_0_1_running_var" -> "246 batch_norm"  [label="(16,)", style=solid];
+"145 features_1_block_0_1_running_mean" -> "249 batch_norm_1"  [label="(16,)", style=solid];
+"146 features_1_block_0_1_running_var" -> "249 batch_norm_1"  [label="(16,)", style=solid];
+"148 features_1_block_2_1_running_mean" -> "258 batch_norm_2"  [label="(16,)", style=solid];
+"149 features_1_block_2_1_running_var" -> "258 batch_norm_2"  [label="(16,)", style=solid];
+"151 features_2_block_0_1_running_mean" -> "260 batch_norm_3"  [label="(72,)", style=solid];
+"152 features_2_block_0_1_running_var" -> "260 batch_norm_3"  [label="(72,)", style=solid];
+"154 features_2_block_1_1_running_mean" -> "263 batch_norm_4"  [label="(72,)", style=solid];
+"155 features_2_block_1_1_running_var" -> "263 batch_norm_4"  [label="(72,)", style=solid];
+"157 features_2_block_2_1_running_mean" -> "266 batch_norm_5"  [label="(24,)", style=solid];
+"158 features_2_block_2_1_running_var" -> "266 batch_norm_5"  [label="(24,)", style=solid];
+"160 features_3_block_0_1_running_mean" -> "268 batch_norm_6"  [label="(88,)", style=solid];
+"161 features_3_block_0_1_running_var" -> "268 batch_norm_6"  [label="(88,)", style=solid];
+"163 features_3_block_1_1_running_mean" -> "271 batch_norm_7"  [label="(88,)", style=solid];
+"164 features_3_block_1_1_running_var" -> "271 batch_norm_7"  [label="(88,)", style=solid];
+"166 features_3_block_2_1_running_mean" -> "274 batch_norm_8"  [label="(24,)", style=solid];
+"167 features_3_block_2_1_running_var" -> "274 batch_norm_8"  [label="(24,)", style=solid];
+"169 features_4_block_0_1_running_mean" -> "277 batch_norm_9"  [label="(96,)", style=solid];
+"170 features_4_block_0_1_running_var" -> "277 batch_norm_9"  [label="(96,)", style=solid];
+"172 features_4_block_1_1_running_mean" -> "280 batch_norm_10"  [label="(96,)", style=solid];
+"173 features_4_block_1_1_running_var" -> "280 batch_norm_10"  [label="(96,)", style=solid];
+"175 features_4_block_3_1_running_mean" -> "289 batch_norm_11"  [label="(40,)", style=solid];
+"176 features_4_block_3_1_running_var" -> "289 batch_norm_11"  [label="(40,)", style=solid];
+"178 features_5_block_0_1_running_mean" -> "291 batch_norm_12"  [label="(240,)", style=solid];
+"179 features_5_block_0_1_running_var" -> "291 batch_norm_12"  [label="(240,)", style=solid];
+"181 features_5_block_1_1_running_mean" -> "294 batch_norm_13"  [label="(240,)", style=solid];
+"182 features_5_block_1_1_running_var" -> "294 batch_norm_13"  [label="(240,)", style=solid];
+"184 features_5_block_3_1_running_mean" -> "303 batch_norm_14"  [label="(40,)", style=solid];
+"185 features_5_block_3_1_running_var" -> "303 batch_norm_14"  [label="(40,)", style=solid];
+"187 features_6_block_0_1_running_mean" -> "306 batch_norm_15"  [label="(240,)", style=solid];
+"188 features_6_block_0_1_running_var" -> "306 batch_norm_15"  [label="(240,)", style=solid];
+"190 features_6_block_1_1_running_mean" -> "309 batch_norm_16"  [label="(240,)", style=solid];
+"191 features_6_block_1_1_running_var" -> "309 batch_norm_16"  [label="(240,)", style=solid];
+"193 features_6_block_3_1_running_mean" -> "318 batch_norm_17"  [label="(40,)", style=solid];
+"194 features_6_block_3_1_running_var" -> "318 batch_norm_17"  [label="(40,)", style=solid];
+"196 features_7_block_0_1_running_mean" -> "321 batch_norm_18"  [label="(120,)", style=solid];
+"197 features_7_block_0_1_running_var" -> "321 batch_norm_18"  [label="(120,)", style=solid];
+"199 features_7_block_1_1_running_mean" -> "324 batch_norm_19"  [label="(120,)", style=solid];
+"200 features_7_block_1_1_running_var" -> "324 batch_norm_19"  [label="(120,)", style=solid];
+"202 features_7_block_3_1_running_mean" -> "333 batch_norm_20"  [label="(48,)", style=solid];
+"203 features_7_block_3_1_running_var" -> "333 batch_norm_20"  [label="(48,)", style=solid];
+"205 features_8_block_0_1_running_mean" -> "335 batch_norm_21"  [label="(144,)", style=solid];
+"206 features_8_block_0_1_running_var" -> "335 batch_norm_21"  [label="(144,)", style=solid];
+"208 features_8_block_1_1_running_mean" -> "338 batch_norm_22"  [label="(144,)", style=solid];
+"209 features_8_block_1_1_running_var" -> "338 batch_norm_22"  [label="(144,)", style=solid];
+"211 features_8_block_3_1_running_mean" -> "347 batch_norm_23"  [label="(48,)", style=solid];
+"212 features_8_block_3_1_running_var" -> "347 batch_norm_23"  [label="(48,)", style=solid];
+"214 features_9_block_0_1_running_mean" -> "350 batch_norm_24"  [label="(288,)", style=solid];
+"215 features_9_block_0_1_running_var" -> "350 batch_norm_24"  [label="(288,)", style=solid];
+"217 features_9_block_1_1_running_mean" -> "353 batch_norm_25"  [label="(288,)", style=solid];
+"218 features_9_block_1_1_running_var" -> "353 batch_norm_25"  [label="(288,)", style=solid];
+"220 features_9_block_3_1_running_mean" -> "362 batch_norm_26"  [label="(96,)", style=solid];
+"221 features_9_block_3_1_running_var" -> "362 batch_norm_26"  [label="(96,)", style=solid];
+"223 features_10_block_0_1_running_mean" -> "364 batch_norm_27"  [label="(576,)", style=solid];
+"224 features_10_block_0_1_running_var" -> "364 batch_norm_27"  [label="(576,)", style=solid];
+"226 features_10_block_1_1_running_mean" -> "367 batch_norm_28"  [label="(576,)", style=solid];
+"227 features_10_block_1_1_running_var" -> "367 batch_norm_28"  [label="(576,)", style=solid];
+"229 features_10_block_3_1_running_mean" -> "376 batch_norm_29"  [label="(96,)", style=solid];
+"230 features_10_block_3_1_running_var" -> "376 batch_norm_29"  [label="(96,)", style=solid];
+"232 features_11_block_0_1_running_mean" -> "379 batch_norm_30"  [label="(576,)", style=solid];
+"233 features_11_block_0_1_running_var" -> "379 batch_norm_30"  [label="(576,)", style=solid];
+"235 features_11_block_1_1_running_mean" -> "382 batch_norm_31"  [label="(576,)", style=solid];
+"236 features_11_block_1_1_running_var" -> "382 batch_norm_31"  [label="(576,)", style=solid];
+"238 features_11_block_3_1_running_mean" -> "391 batch_norm_32"  [label="(96,)", style=solid];
+"239 features_11_block_3_1_running_var" -> "391 batch_norm_32"  [label="(96,)", style=solid];
+"241 features_12_1_running_mean" -> "394 batch_norm_33"  [label="(576,)", style=solid];
+"242 features_12_1_running_var" -> "394 batch_norm_33"  [label="(576,)", style=solid];
+"244 x" -> "245 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"245 conv2d" -> "246 batch_norm"  [label="(1, 16, 112, 112)", style=solid];
+"246 batch_norm" -> "247 hardswish_"  [label="(1, 16, 112, 112)", style=solid];
+"247 hardswish_" -> "248 conv2d_1"  [label="(1, 16, 112, 112)", style=solid];
+"248 conv2d_1" -> "249 batch_norm_1"  [label="(1, 16, 56, 56)", style=solid];
+"249 batch_norm_1" -> "250 relu_"  [label="(1, 16, 56, 56)", style=solid];
+"250 relu_" -> "251 adaptive_avg_pool2d"  [label="(1, 16, 56, 56)", style=solid];
+"250 relu_" -> "256 mul"  [label="(1, 16, 56, 56)", style=solid];
+"251 adaptive_avg_pool2d" -> "252 conv2d_2"  [label="(1, 16, 1, 1)", style=solid];
+"252 conv2d_2" -> "253 relu"  [label="(1, 8, 1, 1)", style=solid];
+"253 relu" -> "254 conv2d_3"  [label="(1, 8, 1, 1)", style=solid];
+"254 conv2d_3" -> "255 hardsigmoid"  [label="(1, 16, 1, 1)", style=solid];
+"255 hardsigmoid" -> "256 mul"  [label="(1, 16, 1, 1)", style=solid];
+"256 mul" -> "257 conv2d_4"  [label="(1, 16, 56, 56)", style=solid];
+"257 conv2d_4" -> "258 batch_norm_2"  [label="(1, 16, 56, 56)", style=solid];
+"258 batch_norm_2" -> "259 conv2d_5"  [label="(1, 16, 56, 56)", style=solid];
+"259 conv2d_5" -> "260 batch_norm_3"  [label="(1, 72, 56, 56)", style=solid];
+"260 batch_norm_3" -> "261 relu__1"  [label="(1, 72, 56, 56)", style=solid];
+"261 relu__1" -> "262 conv2d_6"  [label="(1, 72, 56, 56)", style=solid];
+"262 conv2d_6" -> "263 batch_norm_4"  [label="(1, 72, 28, 28)", style=solid];
+"263 batch_norm_4" -> "264 relu__2"  [label="(1, 72, 28, 28)", style=solid];
+"264 relu__2" -> "265 conv2d_7"  [label="(1, 72, 28, 28)", style=solid];
+"265 conv2d_7" -> "266 batch_norm_5"  [label="(1, 24, 28, 28)", style=solid];
+"266 batch_norm_5" -> "267 conv2d_8"  [label="(1, 24, 28, 28)", style=solid];
+"266 batch_norm_5" -> "275 add_"  [label="(1, 24, 28, 28)", style=solid];
+"267 conv2d_8" -> "268 batch_norm_6"  [label="(1, 88, 28, 28)", style=solid];
+"268 batch_norm_6" -> "269 relu__3"  [label="(1, 88, 28, 28)", style=solid];
+"269 relu__3" -> "270 conv2d_9"  [label="(1, 88, 28, 28)", style=solid];
+"270 conv2d_9" -> "271 batch_norm_7"  [label="(1, 88, 28, 28)", style=solid];
+"271 batch_norm_7" -> "272 relu__4"  [label="(1, 88, 28, 28)", style=solid];
+"272 relu__4" -> "273 conv2d_10"  [label="(1, 88, 28, 28)", style=solid];
+"273 conv2d_10" -> "274 batch_norm_8"  [label="(1, 24, 28, 28)", style=solid];
+"274 batch_norm_8" -> "275 add_"  [label="(1, 24, 28, 28)", style=solid];
+"275 add_" -> "276 conv2d_11"  [label="(1, 24, 28, 28)", style=solid];
+"276 conv2d_11" -> "277 batch_norm_9"  [label="(1, 96, 28, 28)", style=solid];
+"277 batch_norm_9" -> "278 hardswish__1"  [label="(1, 96, 28, 28)", style=solid];
+"278 hardswish__1" -> "279 conv2d_12"  [label="(1, 96, 28, 28)", style=solid];
+"279 conv2d_12" -> "280 batch_norm_10"  [label="(1, 96, 14, 14)", style=solid];
+"280 batch_norm_10" -> "281 hardswish__2"  [label="(1, 96, 14, 14)", style=solid];
+"281 hardswish__2" -> "282 adaptive_avg_pool2d_1"  [label="(1, 96, 14, 14)", style=solid];
+"281 hardswish__2" -> "287 mul_1"  [label="(1, 96, 14, 14)", style=solid];
+"282 adaptive_avg_pool2d_1" -> "283 conv2d_13"  [label="(1, 96, 1, 1)", style=solid];
+"283 conv2d_13" -> "284 relu_1"  [label="(1, 24, 1, 1)", style=solid];
+"284 relu_1" -> "285 conv2d_14"  [label="(1, 24, 1, 1)", style=solid];
+"285 conv2d_14" -> "286 hardsigmoid_1"  [label="(1, 96, 1, 1)", style=solid];
+"286 hardsigmoid_1" -> "287 mul_1"  [label="(1, 96, 1, 1)", style=solid];
+"287 mul_1" -> "288 conv2d_15"  [label="(1, 96, 14, 14)", style=solid];
+"288 conv2d_15" -> "289 batch_norm_11"  [label="(1, 40, 14, 14)", style=solid];
+"289 batch_norm_11" -> "290 conv2d_16"  [label="(1, 40, 14, 14)", style=solid];
+"289 batch_norm_11" -> "304 add__1"  [label="(1, 40, 14, 14)", style=solid];
+"290 conv2d_16" -> "291 batch_norm_12"  [label="(1, 240, 14, 14)", style=solid];
+"291 batch_norm_12" -> "292 hardswish__3"  [label="(1, 240, 14, 14)", style=solid];
+"292 hardswish__3" -> "293 conv2d_17"  [label="(1, 240, 14, 14)", style=solid];
+"293 conv2d_17" -> "294 batch_norm_13"  [label="(1, 240, 14, 14)", style=solid];
+"294 batch_norm_13" -> "295 hardswish__4"  [label="(1, 240, 14, 14)", style=solid];
+"295 hardswish__4" -> "296 adaptive_avg_pool2d_2"  [label="(1, 240, 14, 14)", style=solid];
+"295 hardswish__4" -> "301 mul_2"  [label="(1, 240, 14, 14)", style=solid];
+"296 adaptive_avg_pool2d_2" -> "297 conv2d_18"  [label="(1, 240, 1, 1)", style=solid];
+"297 conv2d_18" -> "298 relu_2"  [label="(1, 64, 1, 1)", style=solid];
+"298 relu_2" -> "299 conv2d_19"  [label="(1, 64, 1, 1)", style=solid];
+"299 conv2d_19" -> "300 hardsigmoid_2"  [label="(1, 240, 1, 1)", style=solid];
+"300 hardsigmoid_2" -> "301 mul_2"  [label="(1, 240, 1, 1)", style=solid];
+"301 mul_2" -> "302 conv2d_20"  [label="(1, 240, 14, 14)", style=solid];
+"302 conv2d_20" -> "303 batch_norm_14"  [label="(1, 40, 14, 14)", style=solid];
+"303 batch_norm_14" -> "304 add__1"  [label="(1, 40, 14, 14)", style=solid];
+"304 add__1" -> "305 conv2d_21"  [label="(1, 40, 14, 14)", style=solid];
+"304 add__1" -> "319 add__2"  [label="(1, 40, 14, 14)", style=solid];
+"305 conv2d_21" -> "306 batch_norm_15"  [label="(1, 240, 14, 14)", style=solid];
+"306 batch_norm_15" -> "307 hardswish__5"  [label="(1, 240, 14, 14)", style=solid];
+"307 hardswish__5" -> "308 conv2d_22"  [label="(1, 240, 14, 14)", style=solid];
+"308 conv2d_22" -> "309 batch_norm_16"  [label="(1, 240, 14, 14)", style=solid];
+"309 batch_norm_16" -> "310 hardswish__6"  [label="(1, 240, 14, 14)", style=solid];
+"310 hardswish__6" -> "311 adaptive_avg_pool2d_3"  [label="(1, 240, 14, 14)", style=solid];
+"310 hardswish__6" -> "316 mul_3"  [label="(1, 240, 14, 14)", style=solid];
+"311 adaptive_avg_pool2d_3" -> "312 conv2d_23"  [label="(1, 240, 1, 1)", style=solid];
+"312 conv2d_23" -> "313 relu_3"  [label="(1, 64, 1, 1)", style=solid];
+"313 relu_3" -> "314 conv2d_24"  [label="(1, 64, 1, 1)", style=solid];
+"314 conv2d_24" -> "315 hardsigmoid_3"  [label="(1, 240, 1, 1)", style=solid];
+"315 hardsigmoid_3" -> "316 mul_3"  [label="(1, 240, 1, 1)", style=solid];
+"316 mul_3" -> "317 conv2d_25"  [label="(1, 240, 14, 14)", style=solid];
+"317 conv2d_25" -> "318 batch_norm_17"  [label="(1, 40, 14, 14)", style=solid];
+"318 batch_norm_17" -> "319 add__2"  [label="(1, 40, 14, 14)", style=solid];
+"319 add__2" -> "320 conv2d_26"  [label="(1, 40, 14, 14)", style=solid];
+"320 conv2d_26" -> "321 batch_norm_18"  [label="(1, 120, 14, 14)", style=solid];
+"321 batch_norm_18" -> "322 hardswish__7"  [label="(1, 120, 14, 14)", style=solid];
+"322 hardswish__7" -> "323 conv2d_27"  [label="(1, 120, 14, 14)", style=solid];
+"323 conv2d_27" -> "324 batch_norm_19"  [label="(1, 120, 14, 14)", style=solid];
+"324 batch_norm_19" -> "325 hardswish__8"  [label="(1, 120, 14, 14)", style=solid];
+"325 hardswish__8" -> "326 adaptive_avg_pool2d_4"  [label="(1, 120, 14, 14)", style=solid];
+"325 hardswish__8" -> "331 mul_4"  [label="(1, 120, 14, 14)", style=solid];
+"326 adaptive_avg_pool2d_4" -> "327 conv2d_28"  [label="(1, 120, 1, 1)", style=solid];
+"327 conv2d_28" -> "328 relu_4"  [label="(1, 32, 1, 1)", style=solid];
+"328 relu_4" -> "329 conv2d_29"  [label="(1, 32, 1, 1)", style=solid];
+"329 conv2d_29" -> "330 hardsigmoid_4"  [label="(1, 120, 1, 1)", style=solid];
+"330 hardsigmoid_4" -> "331 mul_4"  [label="(1, 120, 1, 1)", style=solid];
+"331 mul_4" -> "332 conv2d_30"  [label="(1, 120, 14, 14)", style=solid];
+"332 conv2d_30" -> "333 batch_norm_20"  [label="(1, 48, 14, 14)", style=solid];
+"333 batch_norm_20" -> "334 conv2d_31"  [label="(1, 48, 14, 14)", style=solid];
+"333 batch_norm_20" -> "348 add__3"  [label="(1, 48, 14, 14)", style=solid];
+"334 conv2d_31" -> "335 batch_norm_21"  [label="(1, 144, 14, 14)", style=solid];
+"335 batch_norm_21" -> "336 hardswish__9"  [label="(1, 144, 14, 14)", style=solid];
+"336 hardswish__9" -> "337 conv2d_32"  [label="(1, 144, 14, 14)", style=solid];
+"337 conv2d_32" -> "338 batch_norm_22"  [label="(1, 144, 14, 14)", style=solid];
+"338 batch_norm_22" -> "339 hardswish__10"  [label="(1, 144, 14, 14)", style=solid];
+"339 hardswish__10" -> "340 adaptive_avg_pool2d_5"  [label="(1, 144, 14, 14)", style=solid];
+"339 hardswish__10" -> "345 mul_5"  [label="(1, 144, 14, 14)", style=solid];
+"340 adaptive_avg_pool2d_5" -> "341 conv2d_33"  [label="(1, 144, 1, 1)", style=solid];
+"341 conv2d_33" -> "342 relu_5"  [label="(1, 40, 1, 1)", style=solid];
+"342 relu_5" -> "343 conv2d_34"  [label="(1, 40, 1, 1)", style=solid];
+"343 conv2d_34" -> "344 hardsigmoid_5"  [label="(1, 144, 1, 1)", style=solid];
+"344 hardsigmoid_5" -> "345 mul_5"  [label="(1, 144, 1, 1)", style=solid];
+"345 mul_5" -> "346 conv2d_35"  [label="(1, 144, 14, 14)", style=solid];
+"346 conv2d_35" -> "347 batch_norm_23"  [label="(1, 48, 14, 14)", style=solid];
+"347 batch_norm_23" -> "348 add__3"  [label="(1, 48, 14, 14)", style=solid];
+"348 add__3" -> "349 conv2d_36"  [label="(1, 48, 14, 14)", style=solid];
+"349 conv2d_36" -> "350 batch_norm_24"  [label="(1, 288, 14, 14)", style=solid];
+"350 batch_norm_24" -> "351 hardswish__11"  [label="(1, 288, 14, 14)", style=solid];
+"351 hardswish__11" -> "352 conv2d_37"  [label="(1, 288, 14, 14)", style=solid];
+"352 conv2d_37" -> "353 batch_norm_25"  [label="(1, 288, 7, 7)", style=solid];
+"353 batch_norm_25" -> "354 hardswish__12"  [label="(1, 288, 7, 7)", style=solid];
+"354 hardswish__12" -> "355 adaptive_avg_pool2d_6"  [label="(1, 288, 7, 7)", style=solid];
+"354 hardswish__12" -> "360 mul_6"  [label="(1, 288, 7, 7)", style=solid];
+"355 adaptive_avg_pool2d_6" -> "356 conv2d_38"  [label="(1, 288, 1, 1)", style=solid];
+"356 conv2d_38" -> "357 relu_6"  [label="(1, 72, 1, 1)", style=solid];
+"357 relu_6" -> "358 conv2d_39"  [label="(1, 72, 1, 1)", style=solid];
+"358 conv2d_39" -> "359 hardsigmoid_6"  [label="(1, 288, 1, 1)", style=solid];
+"359 hardsigmoid_6" -> "360 mul_6"  [label="(1, 288, 1, 1)", style=solid];
+"360 mul_6" -> "361 conv2d_40"  [label="(1, 288, 7, 7)", style=solid];
+"361 conv2d_40" -> "362 batch_norm_26"  [label="(1, 96, 7, 7)", style=solid];
+"362 batch_norm_26" -> "363 conv2d_41"  [label="(1, 96, 7, 7)", style=solid];
+"362 batch_norm_26" -> "377 add__4"  [label="(1, 96, 7, 7)", style=solid];
+"363 conv2d_41" -> "364 batch_norm_27"  [label="(1, 576, 7, 7)", style=solid];
+"364 batch_norm_27" -> "365 hardswish__13"  [label="(1, 576, 7, 7)", style=solid];
+"365 hardswish__13" -> "366 conv2d_42"  [label="(1, 576, 7, 7)", style=solid];
+"366 conv2d_42" -> "367 batch_norm_28"  [label="(1, 576, 7, 7)", style=solid];
+"367 batch_norm_28" -> "368 hardswish__14"  [label="(1, 576, 7, 7)", style=solid];
+"368 hardswish__14" -> "369 adaptive_avg_pool2d_7"  [label="(1, 576, 7, 7)", style=solid];
+"368 hardswish__14" -> "374 mul_7"  [label="(1, 576, 7, 7)", style=solid];
+"369 adaptive_avg_pool2d_7" -> "370 conv2d_43"  [label="(1, 576, 1, 1)", style=solid];
+"370 conv2d_43" -> "371 relu_7"  [label="(1, 144, 1, 1)", style=solid];
+"371 relu_7" -> "372 conv2d_44"  [label="(1, 144, 1, 1)", style=solid];
+"372 conv2d_44" -> "373 hardsigmoid_7"  [label="(1, 576, 1, 1)", style=solid];
+"373 hardsigmoid_7" -> "374 mul_7"  [label="(1, 576, 1, 1)", style=solid];
+"374 mul_7" -> "375 conv2d_45"  [label="(1, 576, 7, 7)", style=solid];
+"375 conv2d_45" -> "376 batch_norm_29"  [label="(1, 96, 7, 7)", style=solid];
+"376 batch_norm_29" -> "377 add__4"  [label="(1, 96, 7, 7)", style=solid];
+"377 add__4" -> "378 conv2d_46"  [label="(1, 96, 7, 7)", style=solid];
+"377 add__4" -> "392 add__5"  [label="(1, 96, 7, 7)", style=solid];
+"378 conv2d_46" -> "379 batch_norm_30"  [label="(1, 576, 7, 7)", style=solid];
+"379 batch_norm_30" -> "380 hardswish__15"  [label="(1, 576, 7, 7)", style=solid];
+"380 hardswish__15" -> "381 conv2d_47"  [label="(1, 576, 7, 7)", style=solid];
+"381 conv2d_47" -> "382 batch_norm_31"  [label="(1, 576, 7, 7)", style=solid];
+"382 batch_norm_31" -> "383 hardswish__16"  [label="(1, 576, 7, 7)", style=solid];
+"383 hardswish__16" -> "384 adaptive_avg_pool2d_8"  [label="(1, 576, 7, 7)", style=solid];
+"383 hardswish__16" -> "389 mul_8"  [label="(1, 576, 7, 7)", style=solid];
+"384 adaptive_avg_pool2d_8" -> "385 conv2d_48"  [label="(1, 576, 1, 1)", style=solid];
+"385 conv2d_48" -> "386 relu_8"  [label="(1, 144, 1, 1)", style=solid];
+"386 relu_8" -> "387 conv2d_49"  [label="(1, 144, 1, 1)", style=solid];
+"387 conv2d_49" -> "388 hardsigmoid_8"  [label="(1, 576, 1, 1)", style=solid];
+"388 hardsigmoid_8" -> "389 mul_8"  [label="(1, 576, 1, 1)", style=solid];
+"389 mul_8" -> "390 conv2d_50"  [label="(1, 576, 7, 7)", style=solid];
+"390 conv2d_50" -> "391 batch_norm_32"  [label="(1, 96, 7, 7)", style=solid];
+"391 batch_norm_32" -> "392 add__5"  [label="(1, 96, 7, 7)", style=solid];
+"392 add__5" -> "393 conv2d_51"  [label="(1, 96, 7, 7)", style=solid];
+"393 conv2d_51" -> "394 batch_norm_33"  [label="(1, 576, 7, 7)", style=solid];
+"394 batch_norm_33" -> "395 hardswish__17"  [label="(1, 576, 7, 7)", style=solid];
+"395 hardswish__17" -> "396 adaptive_avg_pool2d_9"  [label="(1, 576, 7, 7)", style=solid];
+"396 adaptive_avg_pool2d_9" -> "397 flatten"  [label="(1, 576, 1, 1)", style=solid];
+"397 flatten" -> "398 linear"  [label="(1, 576)", style=solid];
+"398 linear" -> "399 hardswish__18"  [label="(1, 1024)", style=solid];
+"399 hardswish__18" -> "400 dropout_"  [label="(1, 1024)", style=solid];
+"400 dropout_" -> "401 linear_1"  [label="(1, 1024)", style=solid];
+"401 linear_1" -> "402 output_1"  [label="(1, 1000)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/mobilenet_v3_small.dot b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/mobilenet_v3_small.dot
index ef626e650f2..47c6390f606 100644
--- a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/mobilenet_v3_small.dot
+++ b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/mobilenet_v3_small.dot
@@ -1,488 +1,488 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 quantize_per_tensor_default" [id=1, type=quantize_per_tensor];
-"2 dequantize_per_tensor_default" [id=2, type=dequantize_per_tensor];
-"3 features_0_0_weight_bias_0_0" [id=3, type=get_attr];
-"4 scale_updated_constant0" [id=4, type=get_attr];
-"5 compressed_weight_updated_constant0" [id=5, type=get_attr];
-"6 mul_tensor" [id=6, type=mul];
-"7 zero_point_updated_constant0" [id=7, type=get_attr];
-"8 sub_tensor" [id=8, type=sub];
-"9 conv2d" [id=9, type=conv2d];
-"10 hardswish__scale_0" [id=10, type=get_attr];
-"11 hardswish__zero_point_0" [id=11, type=get_attr];
-"12 hardswish_" [id=12, type=hardswish_];
-"13 quantize_per_channel_default_1" [id=13, type=quantize_per_channel];
-"14 dequantize_per_channel_default_1" [id=14, type=dequantize_per_channel];
-"15 features_1_block_0_0_weight_bias_0_0" [id=15, type=get_attr];
-"16 scale_updated_constant1" [id=16, type=get_attr];
-"17 compressed_weight_updated_constant1" [id=17, type=get_attr];
-"18 mul_tensor_1" [id=18, type=mul];
-"19 zero_point_updated_constant1" [id=19, type=get_attr];
-"20 sub_tensor_1" [id=20, type=sub];
-"21 conv2d_1" [id=21, type=conv2d];
-"22 relu_" [id=22, type=relu_];
-"23 quantize_per_tensor_default_1" [id=23, type=quantize_per_tensor];
-"24 dequantize_per_tensor_default_2" [id=24, type=dequantize_per_tensor];
-"25 dequantize_per_tensor_default_1" [id=25, type=dequantize_per_tensor];
-"26 adaptive_avg_pool2d" [id=26, type=adaptive_avg_pool2d];
-"27 _param_constant6" [id=27, type=get_attr];
-"28 _param_constant7_0_0" [id=28, type=get_attr];
-"29 conv2d_2" [id=29, type=conv2d];
-"30 relu" [id=30, type=relu];
-"31 _param_constant8" [id=31, type=get_attr];
-"32 _param_constant9_0_0" [id=32, type=get_attr];
-"33 conv2d_3" [id=33, type=conv2d];
-"34 hardsigmoid" [id=34, type=hardsigmoid];
-"35 quantize_per_tensor_default_2" [id=35, type=quantize_per_tensor];
-"36 dequantize_per_tensor_default_3" [id=36, type=dequantize_per_tensor];
-"37 mul" [id=37, type=mul];
-"38 quantize_per_tensor_default_3" [id=38, type=quantize_per_tensor];
-"39 dequantize_per_tensor_default_4" [id=39, type=dequantize_per_tensor];
-"40 features_1_block_2_0_weight_bias_0_0" [id=40, type=get_attr];
-"41 scale_updated_constant2" [id=41, type=get_attr];
-"42 compressed_weight_updated_constant2" [id=42, type=get_attr];
-"43 mul_tensor_2" [id=43, type=mul];
-"44 zero_point_updated_constant2" [id=44, type=get_attr];
-"45 sub_tensor_2" [id=45, type=sub];
-"46 conv2d_4" [id=46, type=conv2d];
-"47 quantize_per_tensor_default_4" [id=47, type=quantize_per_tensor];
-"48 dequantize_per_tensor_default_5" [id=48, type=dequantize_per_tensor];
-"49 features_2_block_0_0_weight_bias_0_0" [id=49, type=get_attr];
-"50 scale_updated_constant3" [id=50, type=get_attr];
-"51 compressed_weight_updated_constant3" [id=51, type=get_attr];
-"52 mul_tensor_3" [id=52, type=mul];
-"53 zero_point_updated_constant3" [id=53, type=get_attr];
-"54 sub_tensor_3" [id=54, type=sub];
-"55 conv2d_5" [id=55, type=conv2d];
-"56 relu__1_scale_0" [id=56, type=get_attr];
-"57 relu__1_zero_point_0" [id=57, type=get_attr];
-"58 relu__1" [id=58, type=relu_];
-"59 quantize_per_channel_default_5" [id=59, type=quantize_per_channel];
-"60 dequantize_per_channel_default_5" [id=60, type=dequantize_per_channel];
-"61 features_2_block_1_0_weight_bias_0_0" [id=61, type=get_attr];
-"62 scale_updated_constant4" [id=62, type=get_attr];
-"63 compressed_weight_updated_constant4" [id=63, type=get_attr];
-"64 mul_tensor_4" [id=64, type=mul];
-"65 zero_point_updated_constant4" [id=65, type=get_attr];
-"66 sub_tensor_4" [id=66, type=sub];
-"67 conv2d_6" [id=67, type=conv2d];
-"68 relu__2" [id=68, type=relu_];
-"69 quantize_per_tensor_default_5" [id=69, type=quantize_per_tensor];
-"70 dequantize_per_tensor_default_6" [id=70, type=dequantize_per_tensor];
-"71 features_2_block_2_0_weight_bias_0_0" [id=71, type=get_attr];
-"72 scale_updated_constant5" [id=72, type=get_attr];
-"73 compressed_weight_updated_constant5" [id=73, type=get_attr];
-"74 mul_tensor_5" [id=74, type=mul];
-"75 zero_point_updated_constant5" [id=75, type=get_attr];
-"76 sub_tensor_5" [id=76, type=sub];
-"77 conv2d_7" [id=77, type=conv2d];
-"78 quantize_per_tensor_default_6" [id=78, type=quantize_per_tensor];
-"79 dequantize_per_tensor_default_8" [id=79, type=dequantize_per_tensor];
-"80 dequantize_per_tensor_default_7" [id=80, type=dequantize_per_tensor];
-"81 features_3_block_0_0_weight_bias_0_0" [id=81, type=get_attr];
-"82 scale_updated_constant6" [id=82, type=get_attr];
-"83 compressed_weight_updated_constant6" [id=83, type=get_attr];
-"84 mul_tensor_6" [id=84, type=mul];
-"85 zero_point_updated_constant6" [id=85, type=get_attr];
-"86 sub_tensor_6" [id=86, type=sub];
-"87 conv2d_8" [id=87, type=conv2d];
-"88 relu__3_scale_0" [id=88, type=get_attr];
-"89 relu__3_zero_point_0" [id=89, type=get_attr];
-"90 relu__3" [id=90, type=relu_];
-"91 quantize_per_channel_default_9" [id=91, type=quantize_per_channel];
-"92 dequantize_per_channel_default_9" [id=92, type=dequantize_per_channel];
-"93 features_3_block_1_0_weight_bias_0_0" [id=93, type=get_attr];
-"94 scale_updated_constant7" [id=94, type=get_attr];
-"95 compressed_weight_updated_constant7" [id=95, type=get_attr];
-"96 mul_tensor_7" [id=96, type=mul];
-"97 zero_point_updated_constant7" [id=97, type=get_attr];
-"98 sub_tensor_7" [id=98, type=sub];
-"99 conv2d_9" [id=99, type=conv2d];
-"100 relu__4" [id=100, type=relu_];
-"101 quantize_per_tensor_default_7" [id=101, type=quantize_per_tensor];
-"102 dequantize_per_tensor_default_9" [id=102, type=dequantize_per_tensor];
-"103 features_3_block_2_0_weight_bias_0_0" [id=103, type=get_attr];
-"104 scale_updated_constant8" [id=104, type=get_attr];
-"105 compressed_weight_updated_constant8" [id=105, type=get_attr];
-"106 mul_tensor_8" [id=106, type=mul];
-"107 zero_point_updated_constant8" [id=107, type=get_attr];
-"108 sub_tensor_8" [id=108, type=sub];
-"109 conv2d_10" [id=109, type=conv2d];
-"110 quantize_per_tensor_default_8" [id=110, type=quantize_per_tensor];
-"111 dequantize_per_tensor_default_10" [id=111, type=dequantize_per_tensor];
-"112 add_" [id=112, type=add_];
-"113 quantize_per_tensor_default_9" [id=113, type=quantize_per_tensor];
-"114 dequantize_per_tensor_default_11" [id=114, type=dequantize_per_tensor];
-"115 features_4_block_0_0_weight_bias_0_0" [id=115, type=get_attr];
-"116 scale_updated_constant9" [id=116, type=get_attr];
-"117 compressed_weight_updated_constant9" [id=117, type=get_attr];
-"118 mul_tensor_9" [id=118, type=mul];
-"119 zero_point_updated_constant9" [id=119, type=get_attr];
-"120 sub_tensor_9" [id=120, type=sub];
-"121 conv2d_11" [id=121, type=conv2d];
-"122 hardswish__1_scale_0" [id=122, type=get_attr];
-"123 hardswish__1_zero_point_0" [id=123, type=get_attr];
-"124 hardswish__1" [id=124, type=hardswish_];
-"125 quantize_per_channel_default_13" [id=125, type=quantize_per_channel];
-"126 dequantize_per_channel_default_13" [id=126, type=dequantize_per_channel];
-"127 features_4_block_1_0_weight_bias_0_0" [id=127, type=get_attr];
-"128 scale_updated_constant10" [id=128, type=get_attr];
-"129 compressed_weight_updated_constant10" [id=129, type=get_attr];
-"130 mul_tensor_10" [id=130, type=mul];
-"131 zero_point_updated_constant10" [id=131, type=get_attr];
-"132 sub_tensor_10" [id=132, type=sub];
-"133 conv2d_12" [id=133, type=conv2d];
-"134 hardswish__2" [id=134, type=hardswish_];
-"135 quantize_per_tensor_default_10" [id=135, type=quantize_per_tensor];
-"136 dequantize_per_tensor_default_13" [id=136, type=dequantize_per_tensor];
-"137 dequantize_per_tensor_default_12" [id=137, type=dequantize_per_tensor];
-"138 adaptive_avg_pool2d_1" [id=138, type=adaptive_avg_pool2d];
-"139 _param_constant37" [id=139, type=get_attr];
-"140 _param_constant38_0_0" [id=140, type=get_attr];
-"141 conv2d_13" [id=141, type=conv2d];
-"142 relu_1" [id=142, type=relu];
-"143 _param_constant39" [id=143, type=get_attr];
-"144 _param_constant40_0_0" [id=144, type=get_attr];
-"145 conv2d_14" [id=145, type=conv2d];
-"146 hardsigmoid_1" [id=146, type=hardsigmoid];
-"147 quantize_per_tensor_default_11" [id=147, type=quantize_per_tensor];
-"148 dequantize_per_tensor_default_14" [id=148, type=dequantize_per_tensor];
-"149 mul_1" [id=149, type=mul];
-"150 quantize_per_tensor_default_12" [id=150, type=quantize_per_tensor];
-"151 dequantize_per_tensor_default_15" [id=151, type=dequantize_per_tensor];
-"152 features_4_block_3_0_weight_bias_0_0" [id=152, type=get_attr];
-"153 scale_updated_constant11" [id=153, type=get_attr];
-"154 compressed_weight_updated_constant11" [id=154, type=get_attr];
-"155 mul_tensor_11" [id=155, type=mul];
-"156 zero_point_updated_constant11" [id=156, type=get_attr];
-"157 sub_tensor_11" [id=157, type=sub];
-"158 conv2d_15" [id=158, type=conv2d];
-"159 quantize_per_tensor_default_13" [id=159, type=quantize_per_tensor];
-"160 dequantize_per_tensor_default_17" [id=160, type=dequantize_per_tensor];
-"161 dequantize_per_tensor_default_16" [id=161, type=dequantize_per_tensor];
-"162 features_5_block_0_0_weight_bias_0_0" [id=162, type=get_attr];
-"163 scale_updated_constant12" [id=163, type=get_attr];
-"164 compressed_weight_updated_constant12" [id=164, type=get_attr];
-"165 mul_tensor_12" [id=165, type=mul];
-"166 zero_point_updated_constant12" [id=166, type=get_attr];
-"167 sub_tensor_12" [id=167, type=sub];
-"168 conv2d_16" [id=168, type=conv2d];
-"169 hardswish__3_scale_0" [id=169, type=get_attr];
-"170 hardswish__3_zero_point_0" [id=170, type=get_attr];
-"171 hardswish__3" [id=171, type=hardswish_];
-"172 quantize_per_channel_default_17" [id=172, type=quantize_per_channel];
-"173 dequantize_per_channel_default_17" [id=173, type=dequantize_per_channel];
-"174 features_5_block_1_0_weight_bias_0_0" [id=174, type=get_attr];
-"175 scale_updated_constant13" [id=175, type=get_attr];
-"176 compressed_weight_updated_constant13" [id=176, type=get_attr];
-"177 mul_tensor_13" [id=177, type=mul];
-"178 zero_point_updated_constant13" [id=178, type=get_attr];
-"179 sub_tensor_13" [id=179, type=sub];
-"180 conv2d_17" [id=180, type=conv2d];
-"181 hardswish__4" [id=181, type=hardswish_];
-"182 quantize_per_tensor_default_14" [id=182, type=quantize_per_tensor];
-"183 dequantize_per_tensor_default_19" [id=183, type=dequantize_per_tensor];
-"184 dequantize_per_tensor_default_18" [id=184, type=dequantize_per_tensor];
-"185 adaptive_avg_pool2d_2" [id=185, type=adaptive_avg_pool2d];
-"186 _param_constant50" [id=186, type=get_attr];
-"187 _param_constant51_0_0" [id=187, type=get_attr];
-"188 conv2d_18" [id=188, type=conv2d];
-"189 relu_2" [id=189, type=relu];
-"190 _param_constant52" [id=190, type=get_attr];
-"191 _param_constant53_0_0" [id=191, type=get_attr];
-"192 conv2d_19" [id=192, type=conv2d];
-"193 hardsigmoid_2" [id=193, type=hardsigmoid];
-"194 quantize_per_tensor_default_15" [id=194, type=quantize_per_tensor];
-"195 dequantize_per_tensor_default_20" [id=195, type=dequantize_per_tensor];
-"196 mul_2" [id=196, type=mul];
-"197 quantize_per_tensor_default_16" [id=197, type=quantize_per_tensor];
-"198 dequantize_per_tensor_default_21" [id=198, type=dequantize_per_tensor];
-"199 features_5_block_3_0_weight_bias_0_0" [id=199, type=get_attr];
-"200 scale_updated_constant14" [id=200, type=get_attr];
-"201 compressed_weight_updated_constant14" [id=201, type=get_attr];
-"202 mul_tensor_14" [id=202, type=mul];
-"203 zero_point_updated_constant14" [id=203, type=get_attr];
-"204 sub_tensor_14" [id=204, type=sub];
-"205 conv2d_20" [id=205, type=conv2d];
-"206 quantize_per_tensor_default_17" [id=206, type=quantize_per_tensor];
-"207 dequantize_per_tensor_default_22" [id=207, type=dequantize_per_tensor];
-"208 add__1" [id=208, type=add_];
-"209 quantize_per_tensor_default_18" [id=209, type=quantize_per_tensor];
-"210 dequantize_per_tensor_default_24" [id=210, type=dequantize_per_tensor];
-"211 dequantize_per_tensor_default_23" [id=211, type=dequantize_per_tensor];
-"212 features_6_block_0_0_weight_bias_0_0" [id=212, type=get_attr];
-"213 scale_updated_constant15" [id=213, type=get_attr];
-"214 compressed_weight_updated_constant15" [id=214, type=get_attr];
-"215 mul_tensor_15" [id=215, type=mul];
-"216 zero_point_updated_constant15" [id=216, type=get_attr];
-"217 sub_tensor_15" [id=217, type=sub];
-"218 conv2d_21" [id=218, type=conv2d];
-"219 hardswish__5_scale_0" [id=219, type=get_attr];
-"220 hardswish__5_zero_point_0" [id=220, type=get_attr];
-"221 hardswish__5" [id=221, type=hardswish_];
-"222 quantize_per_channel_default_21" [id=222, type=quantize_per_channel];
-"223 dequantize_per_channel_default_21" [id=223, type=dequantize_per_channel];
-"224 features_6_block_1_0_weight_bias_0_0" [id=224, type=get_attr];
-"225 scale_updated_constant16" [id=225, type=get_attr];
-"226 compressed_weight_updated_constant16" [id=226, type=get_attr];
-"227 mul_tensor_16" [id=227, type=mul];
-"228 zero_point_updated_constant16" [id=228, type=get_attr];
-"229 sub_tensor_16" [id=229, type=sub];
-"230 conv2d_22" [id=230, type=conv2d];
-"231 hardswish__6" [id=231, type=hardswish_];
-"232 quantize_per_tensor_default_19" [id=232, type=quantize_per_tensor];
-"233 dequantize_per_tensor_default_26" [id=233, type=dequantize_per_tensor];
-"234 dequantize_per_tensor_default_25" [id=234, type=dequantize_per_tensor];
-"235 adaptive_avg_pool2d_3" [id=235, type=adaptive_avg_pool2d];
-"236 _param_constant63" [id=236, type=get_attr];
-"237 _param_constant64_0_0" [id=237, type=get_attr];
-"238 conv2d_23" [id=238, type=conv2d];
-"239 relu_3" [id=239, type=relu];
-"240 _param_constant65" [id=240, type=get_attr];
-"241 _param_constant66_0_0" [id=241, type=get_attr];
-"242 conv2d_24" [id=242, type=conv2d];
-"243 hardsigmoid_3" [id=243, type=hardsigmoid];
-"244 quantize_per_tensor_default_20" [id=244, type=quantize_per_tensor];
-"245 dequantize_per_tensor_default_27" [id=245, type=dequantize_per_tensor];
-"246 mul_3" [id=246, type=mul];
-"247 quantize_per_tensor_default_21" [id=247, type=quantize_per_tensor];
-"248 dequantize_per_tensor_default_28" [id=248, type=dequantize_per_tensor];
-"249 features_6_block_3_0_weight_bias_0_0" [id=249, type=get_attr];
-"250 scale_updated_constant17" [id=250, type=get_attr];
-"251 compressed_weight_updated_constant17" [id=251, type=get_attr];
-"252 mul_tensor_17" [id=252, type=mul];
-"253 zero_point_updated_constant17" [id=253, type=get_attr];
-"254 sub_tensor_17" [id=254, type=sub];
-"255 conv2d_25" [id=255, type=conv2d];
-"256 quantize_per_tensor_default_22" [id=256, type=quantize_per_tensor];
-"257 dequantize_per_tensor_default_29" [id=257, type=dequantize_per_tensor];
-"258 add__2" [id=258, type=add_];
-"259 quantize_per_tensor_default_23" [id=259, type=quantize_per_tensor];
-"260 dequantize_per_tensor_default_30" [id=260, type=dequantize_per_tensor];
-"261 features_7_block_0_0_weight_bias_0_0" [id=261, type=get_attr];
-"262 scale_updated_constant18" [id=262, type=get_attr];
-"263 compressed_weight_updated_constant18" [id=263, type=get_attr];
-"264 mul_tensor_18" [id=264, type=mul];
-"265 zero_point_updated_constant18" [id=265, type=get_attr];
-"266 sub_tensor_18" [id=266, type=sub];
-"267 conv2d_26" [id=267, type=conv2d];
-"268 hardswish__7_scale_0" [id=268, type=get_attr];
-"269 hardswish__7_zero_point_0" [id=269, type=get_attr];
-"270 hardswish__7" [id=270, type=hardswish_];
-"271 quantize_per_channel_default_25" [id=271, type=quantize_per_channel];
-"272 dequantize_per_channel_default_25" [id=272, type=dequantize_per_channel];
-"273 features_7_block_1_0_weight_bias_0_0" [id=273, type=get_attr];
-"274 scale_updated_constant19" [id=274, type=get_attr];
-"275 compressed_weight_updated_constant19" [id=275, type=get_attr];
-"276 mul_tensor_19" [id=276, type=mul];
-"277 zero_point_updated_constant19" [id=277, type=get_attr];
-"278 sub_tensor_19" [id=278, type=sub];
-"279 conv2d_27" [id=279, type=conv2d];
-"280 hardswish__8" [id=280, type=hardswish_];
-"281 quantize_per_tensor_default_24" [id=281, type=quantize_per_tensor];
-"282 dequantize_per_tensor_default_32" [id=282, type=dequantize_per_tensor];
-"283 dequantize_per_tensor_default_31" [id=283, type=dequantize_per_tensor];
-"284 adaptive_avg_pool2d_4" [id=284, type=adaptive_avg_pool2d];
-"285 _param_constant76" [id=285, type=get_attr];
-"286 _param_constant77_0_0" [id=286, type=get_attr];
-"287 conv2d_28" [id=287, type=conv2d];
-"288 relu_4" [id=288, type=relu];
-"289 _param_constant78" [id=289, type=get_attr];
-"290 _param_constant79_0_0" [id=290, type=get_attr];
-"291 conv2d_29" [id=291, type=conv2d];
-"292 hardsigmoid_4" [id=292, type=hardsigmoid];
-"293 quantize_per_tensor_default_25" [id=293, type=quantize_per_tensor];
-"294 dequantize_per_tensor_default_33" [id=294, type=dequantize_per_tensor];
-"295 mul_4" [id=295, type=mul];
-"296 quantize_per_tensor_default_26" [id=296, type=quantize_per_tensor];
-"297 dequantize_per_tensor_default_34" [id=297, type=dequantize_per_tensor];
-"298 features_7_block_3_0_weight_bias_0_0" [id=298, type=get_attr];
-"299 scale_updated_constant20" [id=299, type=get_attr];
-"300 compressed_weight_updated_constant20" [id=300, type=get_attr];
-"301 mul_tensor_20" [id=301, type=mul];
-"302 zero_point_updated_constant20" [id=302, type=get_attr];
-"303 sub_tensor_20" [id=303, type=sub];
-"304 conv2d_30" [id=304, type=conv2d];
-"305 quantize_per_tensor_default_27" [id=305, type=quantize_per_tensor];
-"306 dequantize_per_tensor_default_36" [id=306, type=dequantize_per_tensor];
-"307 dequantize_per_tensor_default_35" [id=307, type=dequantize_per_tensor];
-"308 features_8_block_0_0_weight_bias_0_0" [id=308, type=get_attr];
-"309 scale_updated_constant21" [id=309, type=get_attr];
-"310 compressed_weight_updated_constant21" [id=310, type=get_attr];
-"311 mul_tensor_21" [id=311, type=mul];
-"312 zero_point_updated_constant21" [id=312, type=get_attr];
-"313 sub_tensor_21" [id=313, type=sub];
-"314 conv2d_31" [id=314, type=conv2d];
-"315 hardswish__9_scale_0" [id=315, type=get_attr];
-"316 hardswish__9_zero_point_0" [id=316, type=get_attr];
-"317 hardswish__9" [id=317, type=hardswish_];
-"318 quantize_per_channel_default_29" [id=318, type=quantize_per_channel];
-"319 dequantize_per_channel_default_29" [id=319, type=dequantize_per_channel];
-"320 features_8_block_1_0_weight_bias_0_0" [id=320, type=get_attr];
-"321 scale_updated_constant22" [id=321, type=get_attr];
-"322 compressed_weight_updated_constant22" [id=322, type=get_attr];
-"323 mul_tensor_22" [id=323, type=mul];
-"324 zero_point_updated_constant22" [id=324, type=get_attr];
-"325 sub_tensor_22" [id=325, type=sub];
-"326 conv2d_32" [id=326, type=conv2d];
-"327 hardswish__10" [id=327, type=hardswish_];
-"328 quantize_per_tensor_default_28" [id=328, type=quantize_per_tensor];
-"329 dequantize_per_tensor_default_38" [id=329, type=dequantize_per_tensor];
-"330 dequantize_per_tensor_default_37" [id=330, type=dequantize_per_tensor];
-"331 adaptive_avg_pool2d_5" [id=331, type=adaptive_avg_pool2d];
-"332 _param_constant89" [id=332, type=get_attr];
-"333 _param_constant90_0_0" [id=333, type=get_attr];
-"334 conv2d_33" [id=334, type=conv2d];
-"335 relu_5" [id=335, type=relu];
-"336 _param_constant91" [id=336, type=get_attr];
-"337 _param_constant92_0_0" [id=337, type=get_attr];
-"338 conv2d_34" [id=338, type=conv2d];
-"339 hardsigmoid_5" [id=339, type=hardsigmoid];
-"340 quantize_per_tensor_default_29" [id=340, type=quantize_per_tensor];
-"341 dequantize_per_tensor_default_39" [id=341, type=dequantize_per_tensor];
-"342 mul_5" [id=342, type=mul];
-"343 quantize_per_tensor_default_30" [id=343, type=quantize_per_tensor];
-"344 dequantize_per_tensor_default_40" [id=344, type=dequantize_per_tensor];
-"345 features_8_block_3_0_weight_bias_0_0" [id=345, type=get_attr];
-"346 scale_updated_constant23" [id=346, type=get_attr];
-"347 compressed_weight_updated_constant23" [id=347, type=get_attr];
-"348 mul_tensor_23" [id=348, type=mul];
-"349 zero_point_updated_constant23" [id=349, type=get_attr];
-"350 sub_tensor_23" [id=350, type=sub];
-"351 conv2d_35" [id=351, type=conv2d];
-"352 quantize_per_tensor_default_31" [id=352, type=quantize_per_tensor];
-"353 dequantize_per_tensor_default_41" [id=353, type=dequantize_per_tensor];
-"354 add__3" [id=354, type=add_];
-"355 quantize_per_tensor_default_32" [id=355, type=quantize_per_tensor];
-"356 dequantize_per_tensor_default_42" [id=356, type=dequantize_per_tensor];
-"357 features_9_block_0_0_weight_bias_0_0" [id=357, type=get_attr];
-"358 scale_updated_constant24" [id=358, type=get_attr];
-"359 compressed_weight_updated_constant24" [id=359, type=get_attr];
-"360 mul_tensor_24" [id=360, type=mul];
-"361 zero_point_updated_constant24" [id=361, type=get_attr];
-"362 sub_tensor_24" [id=362, type=sub];
-"363 conv2d_36" [id=363, type=conv2d];
-"364 hardswish__11_scale_0" [id=364, type=get_attr];
-"365 hardswish__11_zero_point_0" [id=365, type=get_attr];
-"366 hardswish__11" [id=366, type=hardswish_];
-"367 quantize_per_channel_default_33" [id=367, type=quantize_per_channel];
-"368 dequantize_per_channel_default_33" [id=368, type=dequantize_per_channel];
-"369 features_9_block_1_0_weight_bias_0_0" [id=369, type=get_attr];
-"370 scale_updated_constant25" [id=370, type=get_attr];
-"371 compressed_weight_updated_constant25" [id=371, type=get_attr];
-"372 mul_tensor_25" [id=372, type=mul];
-"373 zero_point_updated_constant25" [id=373, type=get_attr];
-"374 sub_tensor_25" [id=374, type=sub];
-"375 conv2d_37" [id=375, type=conv2d];
-"376 hardswish__12" [id=376, type=hardswish_];
-"377 quantize_per_tensor_default_33" [id=377, type=quantize_per_tensor];
-"378 dequantize_per_tensor_default_44" [id=378, type=dequantize_per_tensor];
-"379 dequantize_per_tensor_default_43" [id=379, type=dequantize_per_tensor];
-"380 adaptive_avg_pool2d_6" [id=380, type=adaptive_avg_pool2d];
-"381 _param_constant102" [id=381, type=get_attr];
-"382 _param_constant103_0_0" [id=382, type=get_attr];
-"383 conv2d_38" [id=383, type=conv2d];
-"384 relu_6" [id=384, type=relu];
-"385 _param_constant104" [id=385, type=get_attr];
-"386 _param_constant105_0_0" [id=386, type=get_attr];
-"387 conv2d_39" [id=387, type=conv2d];
-"388 hardsigmoid_6" [id=388, type=hardsigmoid];
-"389 quantize_per_tensor_default_34" [id=389, type=quantize_per_tensor];
-"390 dequantize_per_tensor_default_45" [id=390, type=dequantize_per_tensor];
-"391 mul_6" [id=391, type=mul];
-"392 quantize_per_tensor_default_35" [id=392, type=quantize_per_tensor];
-"393 dequantize_per_tensor_default_46" [id=393, type=dequantize_per_tensor];
-"394 features_9_block_3_0_weight_bias_0_0" [id=394, type=get_attr];
-"395 scale_updated_constant26" [id=395, type=get_attr];
-"396 compressed_weight_updated_constant26" [id=396, type=get_attr];
-"397 mul_tensor_26" [id=397, type=mul];
-"398 zero_point_updated_constant26" [id=398, type=get_attr];
-"399 sub_tensor_26" [id=399, type=sub];
-"400 conv2d_40" [id=400, type=conv2d];
-"401 quantize_per_tensor_default_36" [id=401, type=quantize_per_tensor];
-"402 dequantize_per_tensor_default_48" [id=402, type=dequantize_per_tensor];
-"403 dequantize_per_tensor_default_47" [id=403, type=dequantize_per_tensor];
-"404 features_10_block_0_0_weight_bias_0_0" [id=404, type=get_attr];
-"405 scale_updated_constant27" [id=405, type=get_attr];
-"406 compressed_weight_updated_constant27" [id=406, type=get_attr];
-"407 mul_tensor_27" [id=407, type=mul];
-"408 zero_point_updated_constant27" [id=408, type=get_attr];
-"409 sub_tensor_27" [id=409, type=sub];
-"410 conv2d_41" [id=410, type=conv2d];
-"411 hardswish__13_scale_0" [id=411, type=get_attr];
-"412 hardswish__13_zero_point_0" [id=412, type=get_attr];
-"413 hardswish__13" [id=413, type=hardswish_];
-"414 quantize_per_channel_default_37" [id=414, type=quantize_per_channel];
-"415 dequantize_per_channel_default_37" [id=415, type=dequantize_per_channel];
-"416 features_10_block_1_0_weight_bias_0_0" [id=416, type=get_attr];
-"417 scale_updated_constant28" [id=417, type=get_attr];
-"418 compressed_weight_updated_constant28" [id=418, type=get_attr];
-"419 mul_tensor_28" [id=419, type=mul];
-"420 zero_point_updated_constant28" [id=420, type=get_attr];
-"421 sub_tensor_28" [id=421, type=sub];
-"422 conv2d_42" [id=422, type=conv2d];
-"423 hardswish__14" [id=423, type=hardswish_];
-"424 quantize_per_tensor_default_37" [id=424, type=quantize_per_tensor];
-"425 dequantize_per_tensor_default_50" [id=425, type=dequantize_per_tensor];
-"426 dequantize_per_tensor_default_49" [id=426, type=dequantize_per_tensor];
-"427 adaptive_avg_pool2d_7" [id=427, type=adaptive_avg_pool2d];
-"428 _param_constant115" [id=428, type=get_attr];
-"429 _param_constant116_0_0" [id=429, type=get_attr];
-"430 conv2d_43" [id=430, type=conv2d];
-"431 relu_7" [id=431, type=relu];
-"432 _param_constant117" [id=432, type=get_attr];
-"433 _param_constant118_0_0" [id=433, type=get_attr];
-"434 conv2d_44" [id=434, type=conv2d];
-"435 hardsigmoid_7" [id=435, type=hardsigmoid];
-"436 quantize_per_tensor_default_38" [id=436, type=quantize_per_tensor];
-"437 dequantize_per_tensor_default_51" [id=437, type=dequantize_per_tensor];
-"438 mul_7" [id=438, type=mul];
-"439 quantize_per_tensor_default_39" [id=439, type=quantize_per_tensor];
-"440 dequantize_per_tensor_default_52" [id=440, type=dequantize_per_tensor];
-"441 features_10_block_3_0_weight_bias_0_0" [id=441, type=get_attr];
-"442 scale_updated_constant29" [id=442, type=get_attr];
-"443 compressed_weight_updated_constant29" [id=443, type=get_attr];
-"444 mul_tensor_29" [id=444, type=mul];
-"445 zero_point_updated_constant29" [id=445, type=get_attr];
-"446 sub_tensor_29" [id=446, type=sub];
-"447 conv2d_45" [id=447, type=conv2d];
-"448 quantize_per_tensor_default_40" [id=448, type=quantize_per_tensor];
-"449 dequantize_per_tensor_default_53" [id=449, type=dequantize_per_tensor];
-"450 add__4" [id=450, type=add_];
-"451 quantize_per_tensor_default_41" [id=451, type=quantize_per_tensor];
-"452 dequantize_per_tensor_default_55" [id=452, type=dequantize_per_tensor];
-"453 dequantize_per_tensor_default_54" [id=453, type=dequantize_per_tensor];
-"454 features_11_block_0_0_weight_bias_0_0" [id=454, type=get_attr];
-"455 scale_updated_constant30" [id=455, type=get_attr];
-"456 compressed_weight_updated_constant30" [id=456, type=get_attr];
-"457 mul_tensor_30" [id=457, type=mul];
-"458 zero_point_updated_constant30" [id=458, type=get_attr];
-"459 sub_tensor_30" [id=459, type=sub];
-"460 conv2d_46" [id=460, type=conv2d];
-"461 hardswish__15_scale_0" [id=461, type=get_attr];
-"462 hardswish__15_zero_point_0" [id=462, type=get_attr];
-"463 hardswish__15" [id=463, type=hardswish_];
-"464 quantize_per_channel_default_41" [id=464, type=quantize_per_channel];
-"465 dequantize_per_channel_default_41" [id=465, type=dequantize_per_channel];
-"466 features_11_block_1_0_weight_bias_0_0" [id=466, type=get_attr];
-"467 scale_updated_constant31" [id=467, type=get_attr];
-"468 compressed_weight_updated_constant31" [id=468, type=get_attr];
-"469 mul_tensor_31" [id=469, type=mul];
-"470 zero_point_updated_constant31" [id=470, type=get_attr];
-"471 sub_tensor_31" [id=471, type=sub];
-"472 conv2d_47" [id=472, type=conv2d];
-"473 hardswish__16" [id=473, type=hardswish_];
-"474 quantize_per_tensor_default_42" [id=474, type=quantize_per_tensor];
-"475 dequantize_per_tensor_default_57" [id=475, type=dequantize_per_tensor];
-"476 dequantize_per_tensor_default_56" [id=476, type=dequantize_per_tensor];
-"477 adaptive_avg_pool2d_8" [id=477, type=adaptive_avg_pool2d];
-"478 _param_constant128" [id=478, type=get_attr];
-"479 _param_constant129_0_0" [id=479, type=get_attr];
-"480 conv2d_48" [id=480, type=conv2d];
-"481 relu_8" [id=481, type=relu];
-"482 _param_constant130" [id=482, type=get_attr];
-"483 _param_constant131_0_0" [id=483, type=get_attr];
+"0 features_1_block_1_fc1_weight" [id=0, type=get_attr];
+"1 features_1_block_1_fc2_weight" [id=1, type=get_attr];
+"2 features_4_block_2_fc1_weight" [id=2, type=get_attr];
+"3 features_4_block_2_fc2_weight" [id=3, type=get_attr];
+"4 features_5_block_2_fc1_weight" [id=4, type=get_attr];
+"5 features_5_block_2_fc2_weight" [id=5, type=get_attr];
+"6 features_6_block_2_fc1_weight" [id=6, type=get_attr];
+"7 features_6_block_2_fc2_weight" [id=7, type=get_attr];
+"8 features_7_block_2_fc1_weight" [id=8, type=get_attr];
+"9 features_7_block_2_fc2_weight" [id=9, type=get_attr];
+"10 features_8_block_2_fc1_weight" [id=10, type=get_attr];
+"11 features_8_block_2_fc2_weight" [id=11, type=get_attr];
+"12 features_9_block_2_fc1_weight" [id=12, type=get_attr];
+"13 features_9_block_2_fc2_weight" [id=13, type=get_attr];
+"14 features_10_block_2_fc1_weight" [id=14, type=get_attr];
+"15 features_10_block_2_fc2_weight" [id=15, type=get_attr];
+"16 features_11_block_2_fc1_weight" [id=16, type=get_attr];
+"17 features_11_block_2_fc2_weight" [id=17, type=get_attr];
+"18 x" [id=18, type=input];
+"19 quantize_per_tensor_default" [id=19, type=quantize_per_tensor];
+"20 dequantize_per_tensor_default" [id=20, type=dequantize_per_tensor];
+"21 features_0_0_weight_bias_0_0" [id=21, type=get_attr];
+"22 scale_updated_constant0" [id=22, type=get_attr];
+"23 compressed_weight_updated_constant0" [id=23, type=get_attr];
+"24 mul_tensor" [id=24, type=mul];
+"25 zero_point_updated_constant0" [id=25, type=get_attr];
+"26 sub_tensor" [id=26, type=sub];
+"27 conv2d" [id=27, type=conv2d];
+"28 hardswish__scale_0" [id=28, type=get_attr];
+"29 hardswish__zero_point_0" [id=29, type=get_attr];
+"30 hardswish_" [id=30, type=hardswish_];
+"31 quantize_per_channel_default_1" [id=31, type=quantize_per_channel];
+"32 dequantize_per_channel_default_1" [id=32, type=dequantize_per_channel];
+"33 features_1_block_0_0_weight_bias_0_0" [id=33, type=get_attr];
+"34 scale_updated_constant1" [id=34, type=get_attr];
+"35 compressed_weight_updated_constant1" [id=35, type=get_attr];
+"36 mul_tensor_1" [id=36, type=mul];
+"37 zero_point_updated_constant1" [id=37, type=get_attr];
+"38 sub_tensor_1" [id=38, type=sub];
+"39 conv2d_1" [id=39, type=conv2d];
+"40 relu_" [id=40, type=relu_];
+"41 quantize_per_tensor_default_1" [id=41, type=quantize_per_tensor];
+"42 dequantize_per_tensor_default_2" [id=42, type=dequantize_per_tensor];
+"43 dequantize_per_tensor_default_1" [id=43, type=dequantize_per_tensor];
+"44 adaptive_avg_pool2d" [id=44, type=adaptive_avg_pool2d];
+"45 features_1_block_1_fc1_bias_0_0" [id=45, type=get_attr];
+"46 conv2d_2" [id=46, type=conv2d];
+"47 relu" [id=47, type=relu];
+"48 features_1_block_1_fc2_bias_0_0" [id=48, type=get_attr];
+"49 conv2d_3" [id=49, type=conv2d];
+"50 hardsigmoid" [id=50, type=hardsigmoid];
+"51 quantize_per_tensor_default_2" [id=51, type=quantize_per_tensor];
+"52 dequantize_per_tensor_default_3" [id=52, type=dequantize_per_tensor];
+"53 mul" [id=53, type=mul];
+"54 quantize_per_tensor_default_3" [id=54, type=quantize_per_tensor];
+"55 dequantize_per_tensor_default_4" [id=55, type=dequantize_per_tensor];
+"56 features_1_block_2_0_weight_bias_0_0" [id=56, type=get_attr];
+"57 scale_updated_constant2" [id=57, type=get_attr];
+"58 compressed_weight_updated_constant2" [id=58, type=get_attr];
+"59 mul_tensor_2" [id=59, type=mul];
+"60 zero_point_updated_constant2" [id=60, type=get_attr];
+"61 sub_tensor_2" [id=61, type=sub];
+"62 conv2d_4" [id=62, type=conv2d];
+"63 quantize_per_tensor_default_4" [id=63, type=quantize_per_tensor];
+"64 dequantize_per_tensor_default_5" [id=64, type=dequantize_per_tensor];
+"65 features_2_block_0_0_weight_bias_0_0" [id=65, type=get_attr];
+"66 scale_updated_constant3" [id=66, type=get_attr];
+"67 compressed_weight_updated_constant3" [id=67, type=get_attr];
+"68 mul_tensor_3" [id=68, type=mul];
+"69 zero_point_updated_constant3" [id=69, type=get_attr];
+"70 sub_tensor_3" [id=70, type=sub];
+"71 conv2d_5" [id=71, type=conv2d];
+"72 relu__1_scale_0" [id=72, type=get_attr];
+"73 relu__1_zero_point_0" [id=73, type=get_attr];
+"74 relu__1" [id=74, type=relu_];
+"75 quantize_per_channel_default_5" [id=75, type=quantize_per_channel];
+"76 dequantize_per_channel_default_5" [id=76, type=dequantize_per_channel];
+"77 features_2_block_1_0_weight_bias_0_0" [id=77, type=get_attr];
+"78 scale_updated_constant4" [id=78, type=get_attr];
+"79 compressed_weight_updated_constant4" [id=79, type=get_attr];
+"80 mul_tensor_4" [id=80, type=mul];
+"81 zero_point_updated_constant4" [id=81, type=get_attr];
+"82 sub_tensor_4" [id=82, type=sub];
+"83 conv2d_6" [id=83, type=conv2d];
+"84 relu__2" [id=84, type=relu_];
+"85 quantize_per_tensor_default_5" [id=85, type=quantize_per_tensor];
+"86 dequantize_per_tensor_default_6" [id=86, type=dequantize_per_tensor];
+"87 features_2_block_2_0_weight_bias_0_0" [id=87, type=get_attr];
+"88 scale_updated_constant5" [id=88, type=get_attr];
+"89 compressed_weight_updated_constant5" [id=89, type=get_attr];
+"90 mul_tensor_5" [id=90, type=mul];
+"91 zero_point_updated_constant5" [id=91, type=get_attr];
+"92 sub_tensor_5" [id=92, type=sub];
+"93 conv2d_7" [id=93, type=conv2d];
+"94 quantize_per_tensor_default_6" [id=94, type=quantize_per_tensor];
+"95 dequantize_per_tensor_default_8" [id=95, type=dequantize_per_tensor];
+"96 dequantize_per_tensor_default_7" [id=96, type=dequantize_per_tensor];
+"97 features_3_block_0_0_weight_bias_0_0" [id=97, type=get_attr];
+"98 scale_updated_constant6" [id=98, type=get_attr];
+"99 compressed_weight_updated_constant6" [id=99, type=get_attr];
+"100 mul_tensor_6" [id=100, type=mul];
+"101 zero_point_updated_constant6" [id=101, type=get_attr];
+"102 sub_tensor_6" [id=102, type=sub];
+"103 conv2d_8" [id=103, type=conv2d];
+"104 relu__3_scale_0" [id=104, type=get_attr];
+"105 relu__3_zero_point_0" [id=105, type=get_attr];
+"106 relu__3" [id=106, type=relu_];
+"107 quantize_per_channel_default_9" [id=107, type=quantize_per_channel];
+"108 dequantize_per_channel_default_9" [id=108, type=dequantize_per_channel];
+"109 features_3_block_1_0_weight_bias_0_0" [id=109, type=get_attr];
+"110 scale_updated_constant7" [id=110, type=get_attr];
+"111 compressed_weight_updated_constant7" [id=111, type=get_attr];
+"112 mul_tensor_7" [id=112, type=mul];
+"113 zero_point_updated_constant7" [id=113, type=get_attr];
+"114 sub_tensor_7" [id=114, type=sub];
+"115 conv2d_9" [id=115, type=conv2d];
+"116 relu__4" [id=116, type=relu_];
+"117 quantize_per_tensor_default_7" [id=117, type=quantize_per_tensor];
+"118 dequantize_per_tensor_default_9" [id=118, type=dequantize_per_tensor];
+"119 features_3_block_2_0_weight_bias_0_0" [id=119, type=get_attr];
+"120 scale_updated_constant8" [id=120, type=get_attr];
+"121 compressed_weight_updated_constant8" [id=121, type=get_attr];
+"122 mul_tensor_8" [id=122, type=mul];
+"123 zero_point_updated_constant8" [id=123, type=get_attr];
+"124 sub_tensor_8" [id=124, type=sub];
+"125 conv2d_10" [id=125, type=conv2d];
+"126 quantize_per_tensor_default_8" [id=126, type=quantize_per_tensor];
+"127 dequantize_per_tensor_default_10" [id=127, type=dequantize_per_tensor];
+"128 add_" [id=128, type=add_];
+"129 quantize_per_tensor_default_9" [id=129, type=quantize_per_tensor];
+"130 dequantize_per_tensor_default_11" [id=130, type=dequantize_per_tensor];
+"131 features_4_block_0_0_weight_bias_0_0" [id=131, type=get_attr];
+"132 scale_updated_constant9" [id=132, type=get_attr];
+"133 compressed_weight_updated_constant9" [id=133, type=get_attr];
+"134 mul_tensor_9" [id=134, type=mul];
+"135 zero_point_updated_constant9" [id=135, type=get_attr];
+"136 sub_tensor_9" [id=136, type=sub];
+"137 conv2d_11" [id=137, type=conv2d];
+"138 hardswish__1_scale_0" [id=138, type=get_attr];
+"139 hardswish__1_zero_point_0" [id=139, type=get_attr];
+"140 hardswish__1" [id=140, type=hardswish_];
+"141 quantize_per_channel_default_13" [id=141, type=quantize_per_channel];
+"142 dequantize_per_channel_default_13" [id=142, type=dequantize_per_channel];
+"143 features_4_block_1_0_weight_bias_0_0" [id=143, type=get_attr];
+"144 scale_updated_constant10" [id=144, type=get_attr];
+"145 compressed_weight_updated_constant10" [id=145, type=get_attr];
+"146 mul_tensor_10" [id=146, type=mul];
+"147 zero_point_updated_constant10" [id=147, type=get_attr];
+"148 sub_tensor_10" [id=148, type=sub];
+"149 conv2d_12" [id=149, type=conv2d];
+"150 hardswish__2" [id=150, type=hardswish_];
+"151 quantize_per_tensor_default_10" [id=151, type=quantize_per_tensor];
+"152 dequantize_per_tensor_default_13" [id=152, type=dequantize_per_tensor];
+"153 dequantize_per_tensor_default_12" [id=153, type=dequantize_per_tensor];
+"154 adaptive_avg_pool2d_1" [id=154, type=adaptive_avg_pool2d];
+"155 features_4_block_2_fc1_bias_0_0" [id=155, type=get_attr];
+"156 conv2d_13" [id=156, type=conv2d];
+"157 relu_1" [id=157, type=relu];
+"158 features_4_block_2_fc2_bias_0_0" [id=158, type=get_attr];
+"159 conv2d_14" [id=159, type=conv2d];
+"160 hardsigmoid_1" [id=160, type=hardsigmoid];
+"161 quantize_per_tensor_default_11" [id=161, type=quantize_per_tensor];
+"162 dequantize_per_tensor_default_14" [id=162, type=dequantize_per_tensor];
+"163 mul_1" [id=163, type=mul];
+"164 quantize_per_tensor_default_12" [id=164, type=quantize_per_tensor];
+"165 dequantize_per_tensor_default_15" [id=165, type=dequantize_per_tensor];
+"166 features_4_block_3_0_weight_bias_0_0" [id=166, type=get_attr];
+"167 scale_updated_constant11" [id=167, type=get_attr];
+"168 compressed_weight_updated_constant11" [id=168, type=get_attr];
+"169 mul_tensor_11" [id=169, type=mul];
+"170 zero_point_updated_constant11" [id=170, type=get_attr];
+"171 sub_tensor_11" [id=171, type=sub];
+"172 conv2d_15" [id=172, type=conv2d];
+"173 quantize_per_tensor_default_13" [id=173, type=quantize_per_tensor];
+"174 dequantize_per_tensor_default_17" [id=174, type=dequantize_per_tensor];
+"175 dequantize_per_tensor_default_16" [id=175, type=dequantize_per_tensor];
+"176 features_5_block_0_0_weight_bias_0_0" [id=176, type=get_attr];
+"177 scale_updated_constant12" [id=177, type=get_attr];
+"178 compressed_weight_updated_constant12" [id=178, type=get_attr];
+"179 mul_tensor_12" [id=179, type=mul];
+"180 zero_point_updated_constant12" [id=180, type=get_attr];
+"181 sub_tensor_12" [id=181, type=sub];
+"182 conv2d_16" [id=182, type=conv2d];
+"183 hardswish__3_scale_0" [id=183, type=get_attr];
+"184 hardswish__3_zero_point_0" [id=184, type=get_attr];
+"185 hardswish__3" [id=185, type=hardswish_];
+"186 quantize_per_channel_default_17" [id=186, type=quantize_per_channel];
+"187 dequantize_per_channel_default_17" [id=187, type=dequantize_per_channel];
+"188 features_5_block_1_0_weight_bias_0_0" [id=188, type=get_attr];
+"189 scale_updated_constant13" [id=189, type=get_attr];
+"190 compressed_weight_updated_constant13" [id=190, type=get_attr];
+"191 mul_tensor_13" [id=191, type=mul];
+"192 zero_point_updated_constant13" [id=192, type=get_attr];
+"193 sub_tensor_13" [id=193, type=sub];
+"194 conv2d_17" [id=194, type=conv2d];
+"195 hardswish__4" [id=195, type=hardswish_];
+"196 quantize_per_tensor_default_14" [id=196, type=quantize_per_tensor];
+"197 dequantize_per_tensor_default_19" [id=197, type=dequantize_per_tensor];
+"198 dequantize_per_tensor_default_18" [id=198, type=dequantize_per_tensor];
+"199 adaptive_avg_pool2d_2" [id=199, type=adaptive_avg_pool2d];
+"200 features_5_block_2_fc1_bias_0_0" [id=200, type=get_attr];
+"201 conv2d_18" [id=201, type=conv2d];
+"202 relu_2" [id=202, type=relu];
+"203 features_5_block_2_fc2_bias_0_0" [id=203, type=get_attr];
+"204 conv2d_19" [id=204, type=conv2d];
+"205 hardsigmoid_2" [id=205, type=hardsigmoid];
+"206 quantize_per_tensor_default_15" [id=206, type=quantize_per_tensor];
+"207 dequantize_per_tensor_default_20" [id=207, type=dequantize_per_tensor];
+"208 mul_2" [id=208, type=mul];
+"209 quantize_per_tensor_default_16" [id=209, type=quantize_per_tensor];
+"210 dequantize_per_tensor_default_21" [id=210, type=dequantize_per_tensor];
+"211 features_5_block_3_0_weight_bias_0_0" [id=211, type=get_attr];
+"212 scale_updated_constant14" [id=212, type=get_attr];
+"213 compressed_weight_updated_constant14" [id=213, type=get_attr];
+"214 mul_tensor_14" [id=214, type=mul];
+"215 zero_point_updated_constant14" [id=215, type=get_attr];
+"216 sub_tensor_14" [id=216, type=sub];
+"217 conv2d_20" [id=217, type=conv2d];
+"218 quantize_per_tensor_default_17" [id=218, type=quantize_per_tensor];
+"219 dequantize_per_tensor_default_22" [id=219, type=dequantize_per_tensor];
+"220 add__1" [id=220, type=add_];
+"221 quantize_per_tensor_default_18" [id=221, type=quantize_per_tensor];
+"222 dequantize_per_tensor_default_24" [id=222, type=dequantize_per_tensor];
+"223 dequantize_per_tensor_default_23" [id=223, type=dequantize_per_tensor];
+"224 features_6_block_0_0_weight_bias_0_0" [id=224, type=get_attr];
+"225 scale_updated_constant15" [id=225, type=get_attr];
+"226 compressed_weight_updated_constant15" [id=226, type=get_attr];
+"227 mul_tensor_15" [id=227, type=mul];
+"228 zero_point_updated_constant15" [id=228, type=get_attr];
+"229 sub_tensor_15" [id=229, type=sub];
+"230 conv2d_21" [id=230, type=conv2d];
+"231 hardswish__5_scale_0" [id=231, type=get_attr];
+"232 hardswish__5_zero_point_0" [id=232, type=get_attr];
+"233 hardswish__5" [id=233, type=hardswish_];
+"234 quantize_per_channel_default_21" [id=234, type=quantize_per_channel];
+"235 dequantize_per_channel_default_21" [id=235, type=dequantize_per_channel];
+"236 features_6_block_1_0_weight_bias_0_0" [id=236, type=get_attr];
+"237 scale_updated_constant16" [id=237, type=get_attr];
+"238 compressed_weight_updated_constant16" [id=238, type=get_attr];
+"239 mul_tensor_16" [id=239, type=mul];
+"240 zero_point_updated_constant16" [id=240, type=get_attr];
+"241 sub_tensor_16" [id=241, type=sub];
+"242 conv2d_22" [id=242, type=conv2d];
+"243 hardswish__6" [id=243, type=hardswish_];
+"244 quantize_per_tensor_default_19" [id=244, type=quantize_per_tensor];
+"245 dequantize_per_tensor_default_26" [id=245, type=dequantize_per_tensor];
+"246 dequantize_per_tensor_default_25" [id=246, type=dequantize_per_tensor];
+"247 adaptive_avg_pool2d_3" [id=247, type=adaptive_avg_pool2d];
+"248 features_6_block_2_fc1_bias_0_0" [id=248, type=get_attr];
+"249 conv2d_23" [id=249, type=conv2d];
+"250 relu_3" [id=250, type=relu];
+"251 features_6_block_2_fc2_bias_0_0" [id=251, type=get_attr];
+"252 conv2d_24" [id=252, type=conv2d];
+"253 hardsigmoid_3" [id=253, type=hardsigmoid];
+"254 quantize_per_tensor_default_20" [id=254, type=quantize_per_tensor];
+"255 dequantize_per_tensor_default_27" [id=255, type=dequantize_per_tensor];
+"256 mul_3" [id=256, type=mul];
+"257 quantize_per_tensor_default_21" [id=257, type=quantize_per_tensor];
+"258 dequantize_per_tensor_default_28" [id=258, type=dequantize_per_tensor];
+"259 features_6_block_3_0_weight_bias_0_0" [id=259, type=get_attr];
+"260 scale_updated_constant17" [id=260, type=get_attr];
+"261 compressed_weight_updated_constant17" [id=261, type=get_attr];
+"262 mul_tensor_17" [id=262, type=mul];
+"263 zero_point_updated_constant17" [id=263, type=get_attr];
+"264 sub_tensor_17" [id=264, type=sub];
+"265 conv2d_25" [id=265, type=conv2d];
+"266 quantize_per_tensor_default_22" [id=266, type=quantize_per_tensor];
+"267 dequantize_per_tensor_default_29" [id=267, type=dequantize_per_tensor];
+"268 add__2" [id=268, type=add_];
+"269 quantize_per_tensor_default_23" [id=269, type=quantize_per_tensor];
+"270 dequantize_per_tensor_default_30" [id=270, type=dequantize_per_tensor];
+"271 features_7_block_0_0_weight_bias_0_0" [id=271, type=get_attr];
+"272 scale_updated_constant18" [id=272, type=get_attr];
+"273 compressed_weight_updated_constant18" [id=273, type=get_attr];
+"274 mul_tensor_18" [id=274, type=mul];
+"275 zero_point_updated_constant18" [id=275, type=get_attr];
+"276 sub_tensor_18" [id=276, type=sub];
+"277 conv2d_26" [id=277, type=conv2d];
+"278 hardswish__7_scale_0" [id=278, type=get_attr];
+"279 hardswish__7_zero_point_0" [id=279, type=get_attr];
+"280 hardswish__7" [id=280, type=hardswish_];
+"281 quantize_per_channel_default_25" [id=281, type=quantize_per_channel];
+"282 dequantize_per_channel_default_25" [id=282, type=dequantize_per_channel];
+"283 features_7_block_1_0_weight_bias_0_0" [id=283, type=get_attr];
+"284 scale_updated_constant19" [id=284, type=get_attr];
+"285 compressed_weight_updated_constant19" [id=285, type=get_attr];
+"286 mul_tensor_19" [id=286, type=mul];
+"287 zero_point_updated_constant19" [id=287, type=get_attr];
+"288 sub_tensor_19" [id=288, type=sub];
+"289 conv2d_27" [id=289, type=conv2d];
+"290 hardswish__8" [id=290, type=hardswish_];
+"291 quantize_per_tensor_default_24" [id=291, type=quantize_per_tensor];
+"292 dequantize_per_tensor_default_32" [id=292, type=dequantize_per_tensor];
+"293 dequantize_per_tensor_default_31" [id=293, type=dequantize_per_tensor];
+"294 adaptive_avg_pool2d_4" [id=294, type=adaptive_avg_pool2d];
+"295 features_7_block_2_fc1_bias_0_0" [id=295, type=get_attr];
+"296 conv2d_28" [id=296, type=conv2d];
+"297 relu_4" [id=297, type=relu];
+"298 features_7_block_2_fc2_bias_0_0" [id=298, type=get_attr];
+"299 conv2d_29" [id=299, type=conv2d];
+"300 hardsigmoid_4" [id=300, type=hardsigmoid];
+"301 quantize_per_tensor_default_25" [id=301, type=quantize_per_tensor];
+"302 dequantize_per_tensor_default_33" [id=302, type=dequantize_per_tensor];
+"303 mul_4" [id=303, type=mul];
+"304 quantize_per_tensor_default_26" [id=304, type=quantize_per_tensor];
+"305 dequantize_per_tensor_default_34" [id=305, type=dequantize_per_tensor];
+"306 features_7_block_3_0_weight_bias_0_0" [id=306, type=get_attr];
+"307 scale_updated_constant20" [id=307, type=get_attr];
+"308 compressed_weight_updated_constant20" [id=308, type=get_attr];
+"309 mul_tensor_20" [id=309, type=mul];
+"310 zero_point_updated_constant20" [id=310, type=get_attr];
+"311 sub_tensor_20" [id=311, type=sub];
+"312 conv2d_30" [id=312, type=conv2d];
+"313 quantize_per_tensor_default_27" [id=313, type=quantize_per_tensor];
+"314 dequantize_per_tensor_default_36" [id=314, type=dequantize_per_tensor];
+"315 dequantize_per_tensor_default_35" [id=315, type=dequantize_per_tensor];
+"316 features_8_block_0_0_weight_bias_0_0" [id=316, type=get_attr];
+"317 scale_updated_constant21" [id=317, type=get_attr];
+"318 compressed_weight_updated_constant21" [id=318, type=get_attr];
+"319 mul_tensor_21" [id=319, type=mul];
+"320 zero_point_updated_constant21" [id=320, type=get_attr];
+"321 sub_tensor_21" [id=321, type=sub];
+"322 conv2d_31" [id=322, type=conv2d];
+"323 hardswish__9_scale_0" [id=323, type=get_attr];
+"324 hardswish__9_zero_point_0" [id=324, type=get_attr];
+"325 hardswish__9" [id=325, type=hardswish_];
+"326 quantize_per_channel_default_29" [id=326, type=quantize_per_channel];
+"327 dequantize_per_channel_default_29" [id=327, type=dequantize_per_channel];
+"328 features_8_block_1_0_weight_bias_0_0" [id=328, type=get_attr];
+"329 scale_updated_constant22" [id=329, type=get_attr];
+"330 compressed_weight_updated_constant22" [id=330, type=get_attr];
+"331 mul_tensor_22" [id=331, type=mul];
+"332 zero_point_updated_constant22" [id=332, type=get_attr];
+"333 sub_tensor_22" [id=333, type=sub];
+"334 conv2d_32" [id=334, type=conv2d];
+"335 hardswish__10" [id=335, type=hardswish_];
+"336 quantize_per_tensor_default_28" [id=336, type=quantize_per_tensor];
+"337 dequantize_per_tensor_default_38" [id=337, type=dequantize_per_tensor];
+"338 dequantize_per_tensor_default_37" [id=338, type=dequantize_per_tensor];
+"339 adaptive_avg_pool2d_5" [id=339, type=adaptive_avg_pool2d];
+"340 features_8_block_2_fc1_bias_0_0" [id=340, type=get_attr];
+"341 conv2d_33" [id=341, type=conv2d];
+"342 relu_5" [id=342, type=relu];
+"343 features_8_block_2_fc2_bias_0_0" [id=343, type=get_attr];
+"344 conv2d_34" [id=344, type=conv2d];
+"345 hardsigmoid_5" [id=345, type=hardsigmoid];
+"346 quantize_per_tensor_default_29" [id=346, type=quantize_per_tensor];
+"347 dequantize_per_tensor_default_39" [id=347, type=dequantize_per_tensor];
+"348 mul_5" [id=348, type=mul];
+"349 quantize_per_tensor_default_30" [id=349, type=quantize_per_tensor];
+"350 dequantize_per_tensor_default_40" [id=350, type=dequantize_per_tensor];
+"351 features_8_block_3_0_weight_bias_0_0" [id=351, type=get_attr];
+"352 scale_updated_constant23" [id=352, type=get_attr];
+"353 compressed_weight_updated_constant23" [id=353, type=get_attr];
+"354 mul_tensor_23" [id=354, type=mul];
+"355 zero_point_updated_constant23" [id=355, type=get_attr];
+"356 sub_tensor_23" [id=356, type=sub];
+"357 conv2d_35" [id=357, type=conv2d];
+"358 quantize_per_tensor_default_31" [id=358, type=quantize_per_tensor];
+"359 dequantize_per_tensor_default_41" [id=359, type=dequantize_per_tensor];
+"360 add__3" [id=360, type=add_];
+"361 quantize_per_tensor_default_32" [id=361, type=quantize_per_tensor];
+"362 dequantize_per_tensor_default_42" [id=362, type=dequantize_per_tensor];
+"363 features_9_block_0_0_weight_bias_0_0" [id=363, type=get_attr];
+"364 scale_updated_constant24" [id=364, type=get_attr];
+"365 compressed_weight_updated_constant24" [id=365, type=get_attr];
+"366 mul_tensor_24" [id=366, type=mul];
+"367 zero_point_updated_constant24" [id=367, type=get_attr];
+"368 sub_tensor_24" [id=368, type=sub];
+"369 conv2d_36" [id=369, type=conv2d];
+"370 hardswish__11_scale_0" [id=370, type=get_attr];
+"371 hardswish__11_zero_point_0" [id=371, type=get_attr];
+"372 hardswish__11" [id=372, type=hardswish_];
+"373 quantize_per_channel_default_33" [id=373, type=quantize_per_channel];
+"374 dequantize_per_channel_default_33" [id=374, type=dequantize_per_channel];
+"375 features_9_block_1_0_weight_bias_0_0" [id=375, type=get_attr];
+"376 scale_updated_constant25" [id=376, type=get_attr];
+"377 compressed_weight_updated_constant25" [id=377, type=get_attr];
+"378 mul_tensor_25" [id=378, type=mul];
+"379 zero_point_updated_constant25" [id=379, type=get_attr];
+"380 sub_tensor_25" [id=380, type=sub];
+"381 conv2d_37" [id=381, type=conv2d];
+"382 hardswish__12" [id=382, type=hardswish_];
+"383 quantize_per_tensor_default_33" [id=383, type=quantize_per_tensor];
+"384 dequantize_per_tensor_default_44" [id=384, type=dequantize_per_tensor];
+"385 dequantize_per_tensor_default_43" [id=385, type=dequantize_per_tensor];
+"386 adaptive_avg_pool2d_6" [id=386, type=adaptive_avg_pool2d];
+"387 features_9_block_2_fc1_bias_0_0" [id=387, type=get_attr];
+"388 conv2d_38" [id=388, type=conv2d];
+"389 relu_6" [id=389, type=relu];
+"390 features_9_block_2_fc2_bias_0_0" [id=390, type=get_attr];
+"391 conv2d_39" [id=391, type=conv2d];
+"392 hardsigmoid_6" [id=392, type=hardsigmoid];
+"393 quantize_per_tensor_default_34" [id=393, type=quantize_per_tensor];
+"394 dequantize_per_tensor_default_45" [id=394, type=dequantize_per_tensor];
+"395 mul_6" [id=395, type=mul];
+"396 quantize_per_tensor_default_35" [id=396, type=quantize_per_tensor];
+"397 dequantize_per_tensor_default_46" [id=397, type=dequantize_per_tensor];
+"398 features_9_block_3_0_weight_bias_0_0" [id=398, type=get_attr];
+"399 scale_updated_constant26" [id=399, type=get_attr];
+"400 compressed_weight_updated_constant26" [id=400, type=get_attr];
+"401 mul_tensor_26" [id=401, type=mul];
+"402 zero_point_updated_constant26" [id=402, type=get_attr];
+"403 sub_tensor_26" [id=403, type=sub];
+"404 conv2d_40" [id=404, type=conv2d];
+"405 quantize_per_tensor_default_36" [id=405, type=quantize_per_tensor];
+"406 dequantize_per_tensor_default_48" [id=406, type=dequantize_per_tensor];
+"407 dequantize_per_tensor_default_47" [id=407, type=dequantize_per_tensor];
+"408 features_10_block_0_0_weight_bias_0_0" [id=408, type=get_attr];
+"409 scale_updated_constant27" [id=409, type=get_attr];
+"410 compressed_weight_updated_constant27" [id=410, type=get_attr];
+"411 mul_tensor_27" [id=411, type=mul];
+"412 zero_point_updated_constant27" [id=412, type=get_attr];
+"413 sub_tensor_27" [id=413, type=sub];
+"414 conv2d_41" [id=414, type=conv2d];
+"415 hardswish__13_scale_0" [id=415, type=get_attr];
+"416 hardswish__13_zero_point_0" [id=416, type=get_attr];
+"417 hardswish__13" [id=417, type=hardswish_];
+"418 quantize_per_channel_default_37" [id=418, type=quantize_per_channel];
+"419 dequantize_per_channel_default_37" [id=419, type=dequantize_per_channel];
+"420 features_10_block_1_0_weight_bias_0_0" [id=420, type=get_attr];
+"421 scale_updated_constant28" [id=421, type=get_attr];
+"422 compressed_weight_updated_constant28" [id=422, type=get_attr];
+"423 mul_tensor_28" [id=423, type=mul];
+"424 zero_point_updated_constant28" [id=424, type=get_attr];
+"425 sub_tensor_28" [id=425, type=sub];
+"426 conv2d_42" [id=426, type=conv2d];
+"427 hardswish__14" [id=427, type=hardswish_];
+"428 quantize_per_tensor_default_37" [id=428, type=quantize_per_tensor];
+"429 dequantize_per_tensor_default_50" [id=429, type=dequantize_per_tensor];
+"430 dequantize_per_tensor_default_49" [id=430, type=dequantize_per_tensor];
+"431 adaptive_avg_pool2d_7" [id=431, type=adaptive_avg_pool2d];
+"432 features_10_block_2_fc1_bias_0_0" [id=432, type=get_attr];
+"433 conv2d_43" [id=433, type=conv2d];
+"434 relu_7" [id=434, type=relu];
+"435 features_10_block_2_fc2_bias_0_0" [id=435, type=get_attr];
+"436 conv2d_44" [id=436, type=conv2d];
+"437 hardsigmoid_7" [id=437, type=hardsigmoid];
+"438 quantize_per_tensor_default_38" [id=438, type=quantize_per_tensor];
+"439 dequantize_per_tensor_default_51" [id=439, type=dequantize_per_tensor];
+"440 mul_7" [id=440, type=mul];
+"441 quantize_per_tensor_default_39" [id=441, type=quantize_per_tensor];
+"442 dequantize_per_tensor_default_52" [id=442, type=dequantize_per_tensor];
+"443 features_10_block_3_0_weight_bias_0_0" [id=443, type=get_attr];
+"444 scale_updated_constant29" [id=444, type=get_attr];
+"445 compressed_weight_updated_constant29" [id=445, type=get_attr];
+"446 mul_tensor_29" [id=446, type=mul];
+"447 zero_point_updated_constant29" [id=447, type=get_attr];
+"448 sub_tensor_29" [id=448, type=sub];
+"449 conv2d_45" [id=449, type=conv2d];
+"450 quantize_per_tensor_default_40" [id=450, type=quantize_per_tensor];
+"451 dequantize_per_tensor_default_53" [id=451, type=dequantize_per_tensor];
+"452 add__4" [id=452, type=add_];
+"453 quantize_per_tensor_default_41" [id=453, type=quantize_per_tensor];
+"454 dequantize_per_tensor_default_55" [id=454, type=dequantize_per_tensor];
+"455 dequantize_per_tensor_default_54" [id=455, type=dequantize_per_tensor];
+"456 features_11_block_0_0_weight_bias_0_0" [id=456, type=get_attr];
+"457 scale_updated_constant30" [id=457, type=get_attr];
+"458 compressed_weight_updated_constant30" [id=458, type=get_attr];
+"459 mul_tensor_30" [id=459, type=mul];
+"460 zero_point_updated_constant30" [id=460, type=get_attr];
+"461 sub_tensor_30" [id=461, type=sub];
+"462 conv2d_46" [id=462, type=conv2d];
+"463 hardswish__15_scale_0" [id=463, type=get_attr];
+"464 hardswish__15_zero_point_0" [id=464, type=get_attr];
+"465 hardswish__15" [id=465, type=hardswish_];
+"466 quantize_per_channel_default_41" [id=466, type=quantize_per_channel];
+"467 dequantize_per_channel_default_41" [id=467, type=dequantize_per_channel];
+"468 features_11_block_1_0_weight_bias_0_0" [id=468, type=get_attr];
+"469 scale_updated_constant31" [id=469, type=get_attr];
+"470 compressed_weight_updated_constant31" [id=470, type=get_attr];
+"471 mul_tensor_31" [id=471, type=mul];
+"472 zero_point_updated_constant31" [id=472, type=get_attr];
+"473 sub_tensor_31" [id=473, type=sub];
+"474 conv2d_47" [id=474, type=conv2d];
+"475 hardswish__16" [id=475, type=hardswish_];
+"476 quantize_per_tensor_default_42" [id=476, type=quantize_per_tensor];
+"477 dequantize_per_tensor_default_57" [id=477, type=dequantize_per_tensor];
+"478 dequantize_per_tensor_default_56" [id=478, type=dequantize_per_tensor];
+"479 adaptive_avg_pool2d_8" [id=479, type=adaptive_avg_pool2d];
+"480 features_11_block_2_fc1_bias_0_0" [id=480, type=get_attr];
+"481 conv2d_48" [id=481, type=conv2d];
+"482 relu_8" [id=482, type=relu];
+"483 features_11_block_2_fc2_bias_0_0" [id=483, type=get_attr];
 "484 conv2d_49" [id=484, type=conv2d];
 "485 hardsigmoid_8" [id=485, type=hardsigmoid];
 "486 quantize_per_tensor_default_43" [id=486, type=quantize_per_tensor];
@@ -516,7 +516,7 @@ strict digraph  {
 "514 quantize_per_tensor_default_48" [id=514, type=quantize_per_tensor];
 "515 dequantize_per_tensor_default_63" [id=515, type=dequantize_per_tensor];
 "516 flatten" [id=516, type=flatten];
-"517 _param_constant139_0_0" [id=517, type=get_attr];
+"517 classifier_0_bias_0_0" [id=517, type=get_attr];
 "518 scale_updated_constant34" [id=518, type=get_attr];
 "519 compressed_weight_updated_constant34" [id=519, type=get_attr];
 "520 mul_tensor_34" [id=520, type=mul];
@@ -527,7 +527,7 @@ strict digraph  {
 "525 quantize_per_tensor_default_49" [id=525, type=quantize_per_tensor];
 "526 dequantize_per_tensor_default_64" [id=526, type=dequantize_per_tensor];
 "527 dropout_" [id=527, type=dropout_];
-"528 _param_constant141_0_0" [id=528, type=get_attr];
+"528 classifier_3_bias_0_0" [id=528, type=get_attr];
 "529 scale_updated_constant35" [id=529, type=get_attr];
 "530 compressed_weight_updated_constant35" [id=530, type=get_attr];
 "531 mul_tensor_35" [id=531, type=mul];
@@ -535,527 +535,527 @@ strict digraph  {
 "533 sub_tensor_35" [id=533, type=sub];
 "534 linear_1" [id=534, type=linear];
 "535 output" [id=535, type=output];
-"0 arg0_1" -> "1 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"1 quantize_per_tensor_default" -> "2 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"2 dequantize_per_tensor_default" -> "9 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"3 features_0_0_weight_bias_0_0" -> "9 conv2d"  [label="(16,)", style=solid];
-"4 scale_updated_constant0" -> "6 mul_tensor"  [label="(16, 1, 1, 1)", style=solid];
-"5 compressed_weight_updated_constant0" -> "6 mul_tensor"  [label="(16, 3, 3, 3)", style=solid];
-"6 mul_tensor" -> "8 sub_tensor"  [label="(16, 3, 3, 3)", style=solid];
-"7 zero_point_updated_constant0" -> "8 sub_tensor"  [label="(16, 1, 1, 1)", style=solid];
-"8 sub_tensor" -> "9 conv2d"  [label="(16, 3, 3, 3)", style=solid];
-"9 conv2d" -> "12 hardswish_"  [label="(1, 16, 112, 112)", style=solid];
-"10 hardswish__scale_0" -> "13 quantize_per_channel_default_1"  [label="(16,)", style=solid];
-"10 hardswish__scale_0" -> "14 dequantize_per_channel_default_1"  [label="(16,)", style=solid];
-"11 hardswish__zero_point_0" -> "13 quantize_per_channel_default_1"  [label="(16,)", style=solid];
-"11 hardswish__zero_point_0" -> "14 dequantize_per_channel_default_1"  [label="(16,)", style=solid];
-"12 hardswish_" -> "13 quantize_per_channel_default_1"  [label="(1, 16, 112, 112)", style=solid];
-"13 quantize_per_channel_default_1" -> "14 dequantize_per_channel_default_1"  [label="(1, 16, 112, 112)", style=solid];
-"14 dequantize_per_channel_default_1" -> "21 conv2d_1"  [label="(1, 16, 112, 112)", style=solid];
-"15 features_1_block_0_0_weight_bias_0_0" -> "21 conv2d_1"  [label="(16,)", style=solid];
-"16 scale_updated_constant1" -> "18 mul_tensor_1"  [label="(16, 1, 1, 1)", style=solid];
-"17 compressed_weight_updated_constant1" -> "18 mul_tensor_1"  [label="(16, 1, 3, 3)", style=solid];
-"18 mul_tensor_1" -> "20 sub_tensor_1"  [label="(16, 1, 3, 3)", style=solid];
-"19 zero_point_updated_constant1" -> "20 sub_tensor_1"  [label="(16, 1, 1, 1)", style=solid];
-"20 sub_tensor_1" -> "21 conv2d_1"  [label="(16, 1, 3, 3)", style=solid];
-"21 conv2d_1" -> "22 relu_"  [label="(1, 16, 56, 56)", style=solid];
-"22 relu_" -> "23 quantize_per_tensor_default_1"  [label="(1, 16, 56, 56)", style=solid];
-"23 quantize_per_tensor_default_1" -> "24 dequantize_per_tensor_default_2"  [label="(1, 16, 56, 56)", style=solid];
-"23 quantize_per_tensor_default_1" -> "25 dequantize_per_tensor_default_1"  [label="(1, 16, 56, 56)", style=solid];
-"24 dequantize_per_tensor_default_2" -> "37 mul"  [label="(1, 16, 56, 56)", style=solid];
-"25 dequantize_per_tensor_default_1" -> "26 adaptive_avg_pool2d"  [label="(1, 16, 56, 56)", style=solid];
-"26 adaptive_avg_pool2d" -> "29 conv2d_2"  [label="(1, 16, 1, 1)", style=solid];
-"27 _param_constant6" -> "29 conv2d_2"  [label="(8, 16, 1, 1)", style=solid];
-"28 _param_constant7_0_0" -> "29 conv2d_2"  [label="(8,)", style=solid];
-"29 conv2d_2" -> "30 relu"  [label="(1, 8, 1, 1)", style=solid];
-"30 relu" -> "33 conv2d_3"  [label="(1, 8, 1, 1)", style=solid];
-"31 _param_constant8" -> "33 conv2d_3"  [label="(16, 8, 1, 1)", style=solid];
-"32 _param_constant9_0_0" -> "33 conv2d_3"  [label="(16,)", style=solid];
-"33 conv2d_3" -> "34 hardsigmoid"  [label="(1, 16, 1, 1)", style=solid];
-"34 hardsigmoid" -> "35 quantize_per_tensor_default_2"  [label="(1, 16, 1, 1)", style=solid];
-"35 quantize_per_tensor_default_2" -> "36 dequantize_per_tensor_default_3"  [label="(1, 16, 1, 1)", style=solid];
-"36 dequantize_per_tensor_default_3" -> "37 mul"  [label="(1, 16, 1, 1)", style=solid];
-"37 mul" -> "38 quantize_per_tensor_default_3"  [label="(1, 16, 56, 56)", style=solid];
-"38 quantize_per_tensor_default_3" -> "39 dequantize_per_tensor_default_4"  [label="(1, 16, 56, 56)", style=solid];
-"39 dequantize_per_tensor_default_4" -> "46 conv2d_4"  [label="(1, 16, 56, 56)", style=solid];
-"40 features_1_block_2_0_weight_bias_0_0" -> "46 conv2d_4"  [label="(16,)", style=solid];
-"41 scale_updated_constant2" -> "43 mul_tensor_2"  [label="(16, 1, 1, 1)", style=solid];
-"42 compressed_weight_updated_constant2" -> "43 mul_tensor_2"  [label="(16, 16, 1, 1)", style=solid];
-"43 mul_tensor_2" -> "45 sub_tensor_2"  [label="(16, 16, 1, 1)", style=solid];
-"44 zero_point_updated_constant2" -> "45 sub_tensor_2"  [label="(16, 1, 1, 1)", style=solid];
-"45 sub_tensor_2" -> "46 conv2d_4"  [label="(16, 16, 1, 1)", style=solid];
-"46 conv2d_4" -> "47 quantize_per_tensor_default_4"  [label="(1, 16, 56, 56)", style=solid];
-"47 quantize_per_tensor_default_4" -> "48 dequantize_per_tensor_default_5"  [label="(1, 16, 56, 56)", style=solid];
-"48 dequantize_per_tensor_default_5" -> "55 conv2d_5"  [label="(1, 16, 56, 56)", style=solid];
-"49 features_2_block_0_0_weight_bias_0_0" -> "55 conv2d_5"  [label="(72,)", style=solid];
-"50 scale_updated_constant3" -> "52 mul_tensor_3"  [label="(72, 1, 1, 1)", style=solid];
-"51 compressed_weight_updated_constant3" -> "52 mul_tensor_3"  [label="(72, 16, 1, 1)", style=solid];
-"52 mul_tensor_3" -> "54 sub_tensor_3"  [label="(72, 16, 1, 1)", style=solid];
-"53 zero_point_updated_constant3" -> "54 sub_tensor_3"  [label="(72, 1, 1, 1)", style=solid];
-"54 sub_tensor_3" -> "55 conv2d_5"  [label="(72, 16, 1, 1)", style=solid];
-"55 conv2d_5" -> "58 relu__1"  [label="(1, 72, 56, 56)", style=solid];
-"56 relu__1_scale_0" -> "59 quantize_per_channel_default_5"  [label="(72,)", style=solid];
-"56 relu__1_scale_0" -> "60 dequantize_per_channel_default_5"  [label="(72,)", style=solid];
-"57 relu__1_zero_point_0" -> "59 quantize_per_channel_default_5"  [label="(72,)", style=solid];
-"57 relu__1_zero_point_0" -> "60 dequantize_per_channel_default_5"  [label="(72,)", style=solid];
-"58 relu__1" -> "59 quantize_per_channel_default_5"  [label="(1, 72, 56, 56)", style=solid];
-"59 quantize_per_channel_default_5" -> "60 dequantize_per_channel_default_5"  [label="(1, 72, 56, 56)", style=solid];
-"60 dequantize_per_channel_default_5" -> "67 conv2d_6"  [label="(1, 72, 56, 56)", style=solid];
-"61 features_2_block_1_0_weight_bias_0_0" -> "67 conv2d_6"  [label="(72,)", style=solid];
-"62 scale_updated_constant4" -> "64 mul_tensor_4"  [label="(72, 1, 1, 1)", style=solid];
-"63 compressed_weight_updated_constant4" -> "64 mul_tensor_4"  [label="(72, 1, 3, 3)", style=solid];
-"64 mul_tensor_4" -> "66 sub_tensor_4"  [label="(72, 1, 3, 3)", style=solid];
-"65 zero_point_updated_constant4" -> "66 sub_tensor_4"  [label="(72, 1, 1, 1)", style=solid];
-"66 sub_tensor_4" -> "67 conv2d_6"  [label="(72, 1, 3, 3)", style=solid];
-"67 conv2d_6" -> "68 relu__2"  [label="(1, 72, 28, 28)", style=solid];
-"68 relu__2" -> "69 quantize_per_tensor_default_5"  [label="(1, 72, 28, 28)", style=solid];
-"69 quantize_per_tensor_default_5" -> "70 dequantize_per_tensor_default_6"  [label="(1, 72, 28, 28)", style=solid];
-"70 dequantize_per_tensor_default_6" -> "77 conv2d_7"  [label="(1, 72, 28, 28)", style=solid];
-"71 features_2_block_2_0_weight_bias_0_0" -> "77 conv2d_7"  [label="(24,)", style=solid];
-"72 scale_updated_constant5" -> "74 mul_tensor_5"  [label="(24, 1, 1, 1)", style=solid];
-"73 compressed_weight_updated_constant5" -> "74 mul_tensor_5"  [label="(24, 72, 1, 1)", style=solid];
-"74 mul_tensor_5" -> "76 sub_tensor_5"  [label="(24, 72, 1, 1)", style=solid];
-"75 zero_point_updated_constant5" -> "76 sub_tensor_5"  [label="(24, 1, 1, 1)", style=solid];
-"76 sub_tensor_5" -> "77 conv2d_7"  [label="(24, 72, 1, 1)", style=solid];
-"77 conv2d_7" -> "78 quantize_per_tensor_default_6"  [label="(1, 24, 28, 28)", style=solid];
-"78 quantize_per_tensor_default_6" -> "79 dequantize_per_tensor_default_8"  [label="(1, 24, 28, 28)", style=solid];
-"78 quantize_per_tensor_default_6" -> "80 dequantize_per_tensor_default_7"  [label="(1, 24, 28, 28)", style=solid];
-"79 dequantize_per_tensor_default_8" -> "112 add_"  [label="(1, 24, 28, 28)", style=solid];
-"80 dequantize_per_tensor_default_7" -> "87 conv2d_8"  [label="(1, 24, 28, 28)", style=solid];
-"81 features_3_block_0_0_weight_bias_0_0" -> "87 conv2d_8"  [label="(88,)", style=solid];
-"82 scale_updated_constant6" -> "84 mul_tensor_6"  [label="(88, 1, 1, 1)", style=solid];
-"83 compressed_weight_updated_constant6" -> "84 mul_tensor_6"  [label="(88, 24, 1, 1)", style=solid];
-"84 mul_tensor_6" -> "86 sub_tensor_6"  [label="(88, 24, 1, 1)", style=solid];
-"85 zero_point_updated_constant6" -> "86 sub_tensor_6"  [label="(88, 1, 1, 1)", style=solid];
-"86 sub_tensor_6" -> "87 conv2d_8"  [label="(88, 24, 1, 1)", style=solid];
-"87 conv2d_8" -> "90 relu__3"  [label="(1, 88, 28, 28)", style=solid];
-"88 relu__3_scale_0" -> "91 quantize_per_channel_default_9"  [label="(88,)", style=solid];
-"88 relu__3_scale_0" -> "92 dequantize_per_channel_default_9"  [label="(88,)", style=solid];
-"89 relu__3_zero_point_0" -> "91 quantize_per_channel_default_9"  [label="(88,)", style=solid];
-"89 relu__3_zero_point_0" -> "92 dequantize_per_channel_default_9"  [label="(88,)", style=solid];
-"90 relu__3" -> "91 quantize_per_channel_default_9"  [label="(1, 88, 28, 28)", style=solid];
-"91 quantize_per_channel_default_9" -> "92 dequantize_per_channel_default_9"  [label="(1, 88, 28, 28)", style=solid];
-"92 dequantize_per_channel_default_9" -> "99 conv2d_9"  [label="(1, 88, 28, 28)", style=solid];
-"93 features_3_block_1_0_weight_bias_0_0" -> "99 conv2d_9"  [label="(88,)", style=solid];
-"94 scale_updated_constant7" -> "96 mul_tensor_7"  [label="(88, 1, 1, 1)", style=solid];
-"95 compressed_weight_updated_constant7" -> "96 mul_tensor_7"  [label="(88, 1, 3, 3)", style=solid];
-"96 mul_tensor_7" -> "98 sub_tensor_7"  [label="(88, 1, 3, 3)", style=solid];
-"97 zero_point_updated_constant7" -> "98 sub_tensor_7"  [label="(88, 1, 1, 1)", style=solid];
-"98 sub_tensor_7" -> "99 conv2d_9"  [label="(88, 1, 3, 3)", style=solid];
-"99 conv2d_9" -> "100 relu__4"  [label="(1, 88, 28, 28)", style=solid];
-"100 relu__4" -> "101 quantize_per_tensor_default_7"  [label="(1, 88, 28, 28)", style=solid];
-"101 quantize_per_tensor_default_7" -> "102 dequantize_per_tensor_default_9"  [label="(1, 88, 28, 28)", style=solid];
-"102 dequantize_per_tensor_default_9" -> "109 conv2d_10"  [label="(1, 88, 28, 28)", style=solid];
-"103 features_3_block_2_0_weight_bias_0_0" -> "109 conv2d_10"  [label="(24,)", style=solid];
-"104 scale_updated_constant8" -> "106 mul_tensor_8"  [label="(24, 1, 1, 1)", style=solid];
-"105 compressed_weight_updated_constant8" -> "106 mul_tensor_8"  [label="(24, 88, 1, 1)", style=solid];
-"106 mul_tensor_8" -> "108 sub_tensor_8"  [label="(24, 88, 1, 1)", style=solid];
-"107 zero_point_updated_constant8" -> "108 sub_tensor_8"  [label="(24, 1, 1, 1)", style=solid];
-"108 sub_tensor_8" -> "109 conv2d_10"  [label="(24, 88, 1, 1)", style=solid];
-"109 conv2d_10" -> "110 quantize_per_tensor_default_8"  [label="(1, 24, 28, 28)", style=solid];
-"110 quantize_per_tensor_default_8" -> "111 dequantize_per_tensor_default_10"  [label="(1, 24, 28, 28)", style=solid];
-"111 dequantize_per_tensor_default_10" -> "112 add_"  [label="(1, 24, 28, 28)", style=solid];
-"112 add_" -> "113 quantize_per_tensor_default_9"  [label="(1, 24, 28, 28)", style=solid];
-"113 quantize_per_tensor_default_9" -> "114 dequantize_per_tensor_default_11"  [label="(1, 24, 28, 28)", style=solid];
-"114 dequantize_per_tensor_default_11" -> "121 conv2d_11"  [label="(1, 24, 28, 28)", style=solid];
-"115 features_4_block_0_0_weight_bias_0_0" -> "121 conv2d_11"  [label="(96,)", style=solid];
-"116 scale_updated_constant9" -> "118 mul_tensor_9"  [label="(96, 1, 1, 1)", style=solid];
-"117 compressed_weight_updated_constant9" -> "118 mul_tensor_9"  [label="(96, 24, 1, 1)", style=solid];
-"118 mul_tensor_9" -> "120 sub_tensor_9"  [label="(96, 24, 1, 1)", style=solid];
-"119 zero_point_updated_constant9" -> "120 sub_tensor_9"  [label="(96, 1, 1, 1)", style=solid];
-"120 sub_tensor_9" -> "121 conv2d_11"  [label="(96, 24, 1, 1)", style=solid];
-"121 conv2d_11" -> "124 hardswish__1"  [label="(1, 96, 28, 28)", style=solid];
-"122 hardswish__1_scale_0" -> "125 quantize_per_channel_default_13"  [label="(96,)", style=solid];
-"122 hardswish__1_scale_0" -> "126 dequantize_per_channel_default_13"  [label="(96,)", style=solid];
-"123 hardswish__1_zero_point_0" -> "125 quantize_per_channel_default_13"  [label="(96,)", style=solid];
-"123 hardswish__1_zero_point_0" -> "126 dequantize_per_channel_default_13"  [label="(96,)", style=solid];
-"124 hardswish__1" -> "125 quantize_per_channel_default_13"  [label="(1, 96, 28, 28)", style=solid];
-"125 quantize_per_channel_default_13" -> "126 dequantize_per_channel_default_13"  [label="(1, 96, 28, 28)", style=solid];
-"126 dequantize_per_channel_default_13" -> "133 conv2d_12"  [label="(1, 96, 28, 28)", style=solid];
-"127 features_4_block_1_0_weight_bias_0_0" -> "133 conv2d_12"  [label="(96,)", style=solid];
-"128 scale_updated_constant10" -> "130 mul_tensor_10"  [label="(96, 1, 1, 1)", style=solid];
-"129 compressed_weight_updated_constant10" -> "130 mul_tensor_10"  [label="(96, 1, 5, 5)", style=solid];
-"130 mul_tensor_10" -> "132 sub_tensor_10"  [label="(96, 1, 5, 5)", style=solid];
-"131 zero_point_updated_constant10" -> "132 sub_tensor_10"  [label="(96, 1, 1, 1)", style=solid];
-"132 sub_tensor_10" -> "133 conv2d_12"  [label="(96, 1, 5, 5)", style=solid];
-"133 conv2d_12" -> "134 hardswish__2"  [label="(1, 96, 14, 14)", style=solid];
-"134 hardswish__2" -> "135 quantize_per_tensor_default_10"  [label="(1, 96, 14, 14)", style=solid];
-"135 quantize_per_tensor_default_10" -> "136 dequantize_per_tensor_default_13"  [label="(1, 96, 14, 14)", style=solid];
-"135 quantize_per_tensor_default_10" -> "137 dequantize_per_tensor_default_12"  [label="(1, 96, 14, 14)", style=solid];
-"136 dequantize_per_tensor_default_13" -> "149 mul_1"  [label="(1, 96, 14, 14)", style=solid];
-"137 dequantize_per_tensor_default_12" -> "138 adaptive_avg_pool2d_1"  [label="(1, 96, 14, 14)", style=solid];
-"138 adaptive_avg_pool2d_1" -> "141 conv2d_13"  [label="(1, 96, 1, 1)", style=solid];
-"139 _param_constant37" -> "141 conv2d_13"  [label="(24, 96, 1, 1)", style=solid];
-"140 _param_constant38_0_0" -> "141 conv2d_13"  [label="(24,)", style=solid];
-"141 conv2d_13" -> "142 relu_1"  [label="(1, 24, 1, 1)", style=solid];
-"142 relu_1" -> "145 conv2d_14"  [label="(1, 24, 1, 1)", style=solid];
-"143 _param_constant39" -> "145 conv2d_14"  [label="(96, 24, 1, 1)", style=solid];
-"144 _param_constant40_0_0" -> "145 conv2d_14"  [label="(96,)", style=solid];
-"145 conv2d_14" -> "146 hardsigmoid_1"  [label="(1, 96, 1, 1)", style=solid];
-"146 hardsigmoid_1" -> "147 quantize_per_tensor_default_11"  [label="(1, 96, 1, 1)", style=solid];
-"147 quantize_per_tensor_default_11" -> "148 dequantize_per_tensor_default_14"  [label="(1, 96, 1, 1)", style=solid];
-"148 dequantize_per_tensor_default_14" -> "149 mul_1"  [label="(1, 96, 1, 1)", style=solid];
-"149 mul_1" -> "150 quantize_per_tensor_default_12"  [label="(1, 96, 14, 14)", style=solid];
-"150 quantize_per_tensor_default_12" -> "151 dequantize_per_tensor_default_15"  [label="(1, 96, 14, 14)", style=solid];
-"151 dequantize_per_tensor_default_15" -> "158 conv2d_15"  [label="(1, 96, 14, 14)", style=solid];
-"152 features_4_block_3_0_weight_bias_0_0" -> "158 conv2d_15"  [label="(40,)", style=solid];
-"153 scale_updated_constant11" -> "155 mul_tensor_11"  [label="(40, 1, 1, 1)", style=solid];
-"154 compressed_weight_updated_constant11" -> "155 mul_tensor_11"  [label="(40, 96, 1, 1)", style=solid];
-"155 mul_tensor_11" -> "157 sub_tensor_11"  [label="(40, 96, 1, 1)", style=solid];
-"156 zero_point_updated_constant11" -> "157 sub_tensor_11"  [label="(40, 1, 1, 1)", style=solid];
-"157 sub_tensor_11" -> "158 conv2d_15"  [label="(40, 96, 1, 1)", style=solid];
-"158 conv2d_15" -> "159 quantize_per_tensor_default_13"  [label="(1, 40, 14, 14)", style=solid];
-"159 quantize_per_tensor_default_13" -> "160 dequantize_per_tensor_default_17"  [label="(1, 40, 14, 14)", style=solid];
-"159 quantize_per_tensor_default_13" -> "161 dequantize_per_tensor_default_16"  [label="(1, 40, 14, 14)", style=solid];
-"160 dequantize_per_tensor_default_17" -> "208 add__1"  [label="(1, 40, 14, 14)", style=solid];
-"161 dequantize_per_tensor_default_16" -> "168 conv2d_16"  [label="(1, 40, 14, 14)", style=solid];
-"162 features_5_block_0_0_weight_bias_0_0" -> "168 conv2d_16"  [label="(240,)", style=solid];
-"163 scale_updated_constant12" -> "165 mul_tensor_12"  [label="(240, 1, 1, 1)", style=solid];
-"164 compressed_weight_updated_constant12" -> "165 mul_tensor_12"  [label="(240, 40, 1, 1)", style=solid];
-"165 mul_tensor_12" -> "167 sub_tensor_12"  [label="(240, 40, 1, 1)", style=solid];
-"166 zero_point_updated_constant12" -> "167 sub_tensor_12"  [label="(240, 1, 1, 1)", style=solid];
-"167 sub_tensor_12" -> "168 conv2d_16"  [label="(240, 40, 1, 1)", style=solid];
-"168 conv2d_16" -> "171 hardswish__3"  [label="(1, 240, 14, 14)", style=solid];
-"169 hardswish__3_scale_0" -> "172 quantize_per_channel_default_17"  [label="(240,)", style=solid];
-"169 hardswish__3_scale_0" -> "173 dequantize_per_channel_default_17"  [label="(240,)", style=solid];
-"170 hardswish__3_zero_point_0" -> "172 quantize_per_channel_default_17"  [label="(240,)", style=solid];
-"170 hardswish__3_zero_point_0" -> "173 dequantize_per_channel_default_17"  [label="(240,)", style=solid];
-"171 hardswish__3" -> "172 quantize_per_channel_default_17"  [label="(1, 240, 14, 14)", style=solid];
-"172 quantize_per_channel_default_17" -> "173 dequantize_per_channel_default_17"  [label="(1, 240, 14, 14)", style=solid];
-"173 dequantize_per_channel_default_17" -> "180 conv2d_17"  [label="(1, 240, 14, 14)", style=solid];
-"174 features_5_block_1_0_weight_bias_0_0" -> "180 conv2d_17"  [label="(240,)", style=solid];
-"175 scale_updated_constant13" -> "177 mul_tensor_13"  [label="(240, 1, 1, 1)", style=solid];
-"176 compressed_weight_updated_constant13" -> "177 mul_tensor_13"  [label="(240, 1, 5, 5)", style=solid];
-"177 mul_tensor_13" -> "179 sub_tensor_13"  [label="(240, 1, 5, 5)", style=solid];
-"178 zero_point_updated_constant13" -> "179 sub_tensor_13"  [label="(240, 1, 1, 1)", style=solid];
-"179 sub_tensor_13" -> "180 conv2d_17"  [label="(240, 1, 5, 5)", style=solid];
-"180 conv2d_17" -> "181 hardswish__4"  [label="(1, 240, 14, 14)", style=solid];
-"181 hardswish__4" -> "182 quantize_per_tensor_default_14"  [label="(1, 240, 14, 14)", style=solid];
-"182 quantize_per_tensor_default_14" -> "183 dequantize_per_tensor_default_19"  [label="(1, 240, 14, 14)", style=solid];
-"182 quantize_per_tensor_default_14" -> "184 dequantize_per_tensor_default_18"  [label="(1, 240, 14, 14)", style=solid];
-"183 dequantize_per_tensor_default_19" -> "196 mul_2"  [label="(1, 240, 14, 14)", style=solid];
-"184 dequantize_per_tensor_default_18" -> "185 adaptive_avg_pool2d_2"  [label="(1, 240, 14, 14)", style=solid];
-"185 adaptive_avg_pool2d_2" -> "188 conv2d_18"  [label="(1, 240, 1, 1)", style=solid];
-"186 _param_constant50" -> "188 conv2d_18"  [label="(64, 240, 1, 1)", style=solid];
-"187 _param_constant51_0_0" -> "188 conv2d_18"  [label="(64,)", style=solid];
-"188 conv2d_18" -> "189 relu_2"  [label="(1, 64, 1, 1)", style=solid];
-"189 relu_2" -> "192 conv2d_19"  [label="(1, 64, 1, 1)", style=solid];
-"190 _param_constant52" -> "192 conv2d_19"  [label="(240, 64, 1, 1)", style=solid];
-"191 _param_constant53_0_0" -> "192 conv2d_19"  [label="(240,)", style=solid];
-"192 conv2d_19" -> "193 hardsigmoid_2"  [label="(1, 240, 1, 1)", style=solid];
-"193 hardsigmoid_2" -> "194 quantize_per_tensor_default_15"  [label="(1, 240, 1, 1)", style=solid];
-"194 quantize_per_tensor_default_15" -> "195 dequantize_per_tensor_default_20"  [label="(1, 240, 1, 1)", style=solid];
-"195 dequantize_per_tensor_default_20" -> "196 mul_2"  [label="(1, 240, 1, 1)", style=solid];
-"196 mul_2" -> "197 quantize_per_tensor_default_16"  [label="(1, 240, 14, 14)", style=solid];
-"197 quantize_per_tensor_default_16" -> "198 dequantize_per_tensor_default_21"  [label="(1, 240, 14, 14)", style=solid];
-"198 dequantize_per_tensor_default_21" -> "205 conv2d_20"  [label="(1, 240, 14, 14)", style=solid];
-"199 features_5_block_3_0_weight_bias_0_0" -> "205 conv2d_20"  [label="(40,)", style=solid];
-"200 scale_updated_constant14" -> "202 mul_tensor_14"  [label="(40, 1, 1, 1)", style=solid];
-"201 compressed_weight_updated_constant14" -> "202 mul_tensor_14"  [label="(40, 240, 1, 1)", style=solid];
-"202 mul_tensor_14" -> "204 sub_tensor_14"  [label="(40, 240, 1, 1)", style=solid];
-"203 zero_point_updated_constant14" -> "204 sub_tensor_14"  [label="(40, 1, 1, 1)", style=solid];
-"204 sub_tensor_14" -> "205 conv2d_20"  [label="(40, 240, 1, 1)", style=solid];
-"205 conv2d_20" -> "206 quantize_per_tensor_default_17"  [label="(1, 40, 14, 14)", style=solid];
-"206 quantize_per_tensor_default_17" -> "207 dequantize_per_tensor_default_22"  [label="(1, 40, 14, 14)", style=solid];
-"207 dequantize_per_tensor_default_22" -> "208 add__1"  [label="(1, 40, 14, 14)", style=solid];
-"208 add__1" -> "209 quantize_per_tensor_default_18"  [label="(1, 40, 14, 14)", style=solid];
-"209 quantize_per_tensor_default_18" -> "210 dequantize_per_tensor_default_24"  [label="(1, 40, 14, 14)", style=solid];
-"209 quantize_per_tensor_default_18" -> "211 dequantize_per_tensor_default_23"  [label="(1, 40, 14, 14)", style=solid];
-"210 dequantize_per_tensor_default_24" -> "258 add__2"  [label="(1, 40, 14, 14)", style=solid];
-"211 dequantize_per_tensor_default_23" -> "218 conv2d_21"  [label="(1, 40, 14, 14)", style=solid];
-"212 features_6_block_0_0_weight_bias_0_0" -> "218 conv2d_21"  [label="(240,)", style=solid];
-"213 scale_updated_constant15" -> "215 mul_tensor_15"  [label="(240, 1, 1, 1)", style=solid];
-"214 compressed_weight_updated_constant15" -> "215 mul_tensor_15"  [label="(240, 40, 1, 1)", style=solid];
-"215 mul_tensor_15" -> "217 sub_tensor_15"  [label="(240, 40, 1, 1)", style=solid];
-"216 zero_point_updated_constant15" -> "217 sub_tensor_15"  [label="(240, 1, 1, 1)", style=solid];
-"217 sub_tensor_15" -> "218 conv2d_21"  [label="(240, 40, 1, 1)", style=solid];
-"218 conv2d_21" -> "221 hardswish__5"  [label="(1, 240, 14, 14)", style=solid];
-"219 hardswish__5_scale_0" -> "222 quantize_per_channel_default_21"  [label="(240,)", style=solid];
-"219 hardswish__5_scale_0" -> "223 dequantize_per_channel_default_21"  [label="(240,)", style=solid];
-"220 hardswish__5_zero_point_0" -> "222 quantize_per_channel_default_21"  [label="(240,)", style=solid];
-"220 hardswish__5_zero_point_0" -> "223 dequantize_per_channel_default_21"  [label="(240,)", style=solid];
-"221 hardswish__5" -> "222 quantize_per_channel_default_21"  [label="(1, 240, 14, 14)", style=solid];
-"222 quantize_per_channel_default_21" -> "223 dequantize_per_channel_default_21"  [label="(1, 240, 14, 14)", style=solid];
-"223 dequantize_per_channel_default_21" -> "230 conv2d_22"  [label="(1, 240, 14, 14)", style=solid];
-"224 features_6_block_1_0_weight_bias_0_0" -> "230 conv2d_22"  [label="(240,)", style=solid];
-"225 scale_updated_constant16" -> "227 mul_tensor_16"  [label="(240, 1, 1, 1)", style=solid];
-"226 compressed_weight_updated_constant16" -> "227 mul_tensor_16"  [label="(240, 1, 5, 5)", style=solid];
-"227 mul_tensor_16" -> "229 sub_tensor_16"  [label="(240, 1, 5, 5)", style=solid];
-"228 zero_point_updated_constant16" -> "229 sub_tensor_16"  [label="(240, 1, 1, 1)", style=solid];
-"229 sub_tensor_16" -> "230 conv2d_22"  [label="(240, 1, 5, 5)", style=solid];
-"230 conv2d_22" -> "231 hardswish__6"  [label="(1, 240, 14, 14)", style=solid];
-"231 hardswish__6" -> "232 quantize_per_tensor_default_19"  [label="(1, 240, 14, 14)", style=solid];
-"232 quantize_per_tensor_default_19" -> "233 dequantize_per_tensor_default_26"  [label="(1, 240, 14, 14)", style=solid];
-"232 quantize_per_tensor_default_19" -> "234 dequantize_per_tensor_default_25"  [label="(1, 240, 14, 14)", style=solid];
-"233 dequantize_per_tensor_default_26" -> "246 mul_3"  [label="(1, 240, 14, 14)", style=solid];
-"234 dequantize_per_tensor_default_25" -> "235 adaptive_avg_pool2d_3"  [label="(1, 240, 14, 14)", style=solid];
-"235 adaptive_avg_pool2d_3" -> "238 conv2d_23"  [label="(1, 240, 1, 1)", style=solid];
-"236 _param_constant63" -> "238 conv2d_23"  [label="(64, 240, 1, 1)", style=solid];
-"237 _param_constant64_0_0" -> "238 conv2d_23"  [label="(64,)", style=solid];
-"238 conv2d_23" -> "239 relu_3"  [label="(1, 64, 1, 1)", style=solid];
-"239 relu_3" -> "242 conv2d_24"  [label="(1, 64, 1, 1)", style=solid];
-"240 _param_constant65" -> "242 conv2d_24"  [label="(240, 64, 1, 1)", style=solid];
-"241 _param_constant66_0_0" -> "242 conv2d_24"  [label="(240,)", style=solid];
-"242 conv2d_24" -> "243 hardsigmoid_3"  [label="(1, 240, 1, 1)", style=solid];
-"243 hardsigmoid_3" -> "244 quantize_per_tensor_default_20"  [label="(1, 240, 1, 1)", style=solid];
-"244 quantize_per_tensor_default_20" -> "245 dequantize_per_tensor_default_27"  [label="(1, 240, 1, 1)", style=solid];
-"245 dequantize_per_tensor_default_27" -> "246 mul_3"  [label="(1, 240, 1, 1)", style=solid];
-"246 mul_3" -> "247 quantize_per_tensor_default_21"  [label="(1, 240, 14, 14)", style=solid];
-"247 quantize_per_tensor_default_21" -> "248 dequantize_per_tensor_default_28"  [label="(1, 240, 14, 14)", style=solid];
-"248 dequantize_per_tensor_default_28" -> "255 conv2d_25"  [label="(1, 240, 14, 14)", style=solid];
-"249 features_6_block_3_0_weight_bias_0_0" -> "255 conv2d_25"  [label="(40,)", style=solid];
-"250 scale_updated_constant17" -> "252 mul_tensor_17"  [label="(40, 1, 1, 1)", style=solid];
-"251 compressed_weight_updated_constant17" -> "252 mul_tensor_17"  [label="(40, 240, 1, 1)", style=solid];
-"252 mul_tensor_17" -> "254 sub_tensor_17"  [label="(40, 240, 1, 1)", style=solid];
-"253 zero_point_updated_constant17" -> "254 sub_tensor_17"  [label="(40, 1, 1, 1)", style=solid];
-"254 sub_tensor_17" -> "255 conv2d_25"  [label="(40, 240, 1, 1)", style=solid];
-"255 conv2d_25" -> "256 quantize_per_tensor_default_22"  [label="(1, 40, 14, 14)", style=solid];
-"256 quantize_per_tensor_default_22" -> "257 dequantize_per_tensor_default_29"  [label="(1, 40, 14, 14)", style=solid];
-"257 dequantize_per_tensor_default_29" -> "258 add__2"  [label="(1, 40, 14, 14)", style=solid];
-"258 add__2" -> "259 quantize_per_tensor_default_23"  [label="(1, 40, 14, 14)", style=solid];
-"259 quantize_per_tensor_default_23" -> "260 dequantize_per_tensor_default_30"  [label="(1, 40, 14, 14)", style=solid];
-"260 dequantize_per_tensor_default_30" -> "267 conv2d_26"  [label="(1, 40, 14, 14)", style=solid];
-"261 features_7_block_0_0_weight_bias_0_0" -> "267 conv2d_26"  [label="(120,)", style=solid];
-"262 scale_updated_constant18" -> "264 mul_tensor_18"  [label="(120, 1, 1, 1)", style=solid];
-"263 compressed_weight_updated_constant18" -> "264 mul_tensor_18"  [label="(120, 40, 1, 1)", style=solid];
-"264 mul_tensor_18" -> "266 sub_tensor_18"  [label="(120, 40, 1, 1)", style=solid];
-"265 zero_point_updated_constant18" -> "266 sub_tensor_18"  [label="(120, 1, 1, 1)", style=solid];
-"266 sub_tensor_18" -> "267 conv2d_26"  [label="(120, 40, 1, 1)", style=solid];
-"267 conv2d_26" -> "270 hardswish__7"  [label="(1, 120, 14, 14)", style=solid];
-"268 hardswish__7_scale_0" -> "271 quantize_per_channel_default_25"  [label="(120,)", style=solid];
-"268 hardswish__7_scale_0" -> "272 dequantize_per_channel_default_25"  [label="(120,)", style=solid];
-"269 hardswish__7_zero_point_0" -> "271 quantize_per_channel_default_25"  [label="(120,)", style=solid];
-"269 hardswish__7_zero_point_0" -> "272 dequantize_per_channel_default_25"  [label="(120,)", style=solid];
-"270 hardswish__7" -> "271 quantize_per_channel_default_25"  [label="(1, 120, 14, 14)", style=solid];
-"271 quantize_per_channel_default_25" -> "272 dequantize_per_channel_default_25"  [label="(1, 120, 14, 14)", style=solid];
-"272 dequantize_per_channel_default_25" -> "279 conv2d_27"  [label="(1, 120, 14, 14)", style=solid];
-"273 features_7_block_1_0_weight_bias_0_0" -> "279 conv2d_27"  [label="(120,)", style=solid];
-"274 scale_updated_constant19" -> "276 mul_tensor_19"  [label="(120, 1, 1, 1)", style=solid];
-"275 compressed_weight_updated_constant19" -> "276 mul_tensor_19"  [label="(120, 1, 5, 5)", style=solid];
-"276 mul_tensor_19" -> "278 sub_tensor_19"  [label="(120, 1, 5, 5)", style=solid];
-"277 zero_point_updated_constant19" -> "278 sub_tensor_19"  [label="(120, 1, 1, 1)", style=solid];
-"278 sub_tensor_19" -> "279 conv2d_27"  [label="(120, 1, 5, 5)", style=solid];
-"279 conv2d_27" -> "280 hardswish__8"  [label="(1, 120, 14, 14)", style=solid];
-"280 hardswish__8" -> "281 quantize_per_tensor_default_24"  [label="(1, 120, 14, 14)", style=solid];
-"281 quantize_per_tensor_default_24" -> "282 dequantize_per_tensor_default_32"  [label="(1, 120, 14, 14)", style=solid];
-"281 quantize_per_tensor_default_24" -> "283 dequantize_per_tensor_default_31"  [label="(1, 120, 14, 14)", style=solid];
-"282 dequantize_per_tensor_default_32" -> "295 mul_4"  [label="(1, 120, 14, 14)", style=solid];
-"283 dequantize_per_tensor_default_31" -> "284 adaptive_avg_pool2d_4"  [label="(1, 120, 14, 14)", style=solid];
-"284 adaptive_avg_pool2d_4" -> "287 conv2d_28"  [label="(1, 120, 1, 1)", style=solid];
-"285 _param_constant76" -> "287 conv2d_28"  [label="(32, 120, 1, 1)", style=solid];
-"286 _param_constant77_0_0" -> "287 conv2d_28"  [label="(32,)", style=solid];
-"287 conv2d_28" -> "288 relu_4"  [label="(1, 32, 1, 1)", style=solid];
-"288 relu_4" -> "291 conv2d_29"  [label="(1, 32, 1, 1)", style=solid];
-"289 _param_constant78" -> "291 conv2d_29"  [label="(120, 32, 1, 1)", style=solid];
-"290 _param_constant79_0_0" -> "291 conv2d_29"  [label="(120,)", style=solid];
-"291 conv2d_29" -> "292 hardsigmoid_4"  [label="(1, 120, 1, 1)", style=solid];
-"292 hardsigmoid_4" -> "293 quantize_per_tensor_default_25"  [label="(1, 120, 1, 1)", style=solid];
-"293 quantize_per_tensor_default_25" -> "294 dequantize_per_tensor_default_33"  [label="(1, 120, 1, 1)", style=solid];
-"294 dequantize_per_tensor_default_33" -> "295 mul_4"  [label="(1, 120, 1, 1)", style=solid];
-"295 mul_4" -> "296 quantize_per_tensor_default_26"  [label="(1, 120, 14, 14)", style=solid];
-"296 quantize_per_tensor_default_26" -> "297 dequantize_per_tensor_default_34"  [label="(1, 120, 14, 14)", style=solid];
-"297 dequantize_per_tensor_default_34" -> "304 conv2d_30"  [label="(1, 120, 14, 14)", style=solid];
-"298 features_7_block_3_0_weight_bias_0_0" -> "304 conv2d_30"  [label="(48,)", style=solid];
-"299 scale_updated_constant20" -> "301 mul_tensor_20"  [label="(48, 1, 1, 1)", style=solid];
-"300 compressed_weight_updated_constant20" -> "301 mul_tensor_20"  [label="(48, 120, 1, 1)", style=solid];
-"301 mul_tensor_20" -> "303 sub_tensor_20"  [label="(48, 120, 1, 1)", style=solid];
-"302 zero_point_updated_constant20" -> "303 sub_tensor_20"  [label="(48, 1, 1, 1)", style=solid];
-"303 sub_tensor_20" -> "304 conv2d_30"  [label="(48, 120, 1, 1)", style=solid];
-"304 conv2d_30" -> "305 quantize_per_tensor_default_27"  [label="(1, 48, 14, 14)", style=solid];
-"305 quantize_per_tensor_default_27" -> "306 dequantize_per_tensor_default_36"  [label="(1, 48, 14, 14)", style=solid];
-"305 quantize_per_tensor_default_27" -> "307 dequantize_per_tensor_default_35"  [label="(1, 48, 14, 14)", style=solid];
-"306 dequantize_per_tensor_default_36" -> "354 add__3"  [label="(1, 48, 14, 14)", style=solid];
-"307 dequantize_per_tensor_default_35" -> "314 conv2d_31"  [label="(1, 48, 14, 14)", style=solid];
-"308 features_8_block_0_0_weight_bias_0_0" -> "314 conv2d_31"  [label="(144,)", style=solid];
-"309 scale_updated_constant21" -> "311 mul_tensor_21"  [label="(144, 1, 1, 1)", style=solid];
-"310 compressed_weight_updated_constant21" -> "311 mul_tensor_21"  [label="(144, 48, 1, 1)", style=solid];
-"311 mul_tensor_21" -> "313 sub_tensor_21"  [label="(144, 48, 1, 1)", style=solid];
-"312 zero_point_updated_constant21" -> "313 sub_tensor_21"  [label="(144, 1, 1, 1)", style=solid];
-"313 sub_tensor_21" -> "314 conv2d_31"  [label="(144, 48, 1, 1)", style=solid];
-"314 conv2d_31" -> "317 hardswish__9"  [label="(1, 144, 14, 14)", style=solid];
-"315 hardswish__9_scale_0" -> "318 quantize_per_channel_default_29"  [label="(144,)", style=solid];
-"315 hardswish__9_scale_0" -> "319 dequantize_per_channel_default_29"  [label="(144,)", style=solid];
-"316 hardswish__9_zero_point_0" -> "318 quantize_per_channel_default_29"  [label="(144,)", style=solid];
-"316 hardswish__9_zero_point_0" -> "319 dequantize_per_channel_default_29"  [label="(144,)", style=solid];
-"317 hardswish__9" -> "318 quantize_per_channel_default_29"  [label="(1, 144, 14, 14)", style=solid];
-"318 quantize_per_channel_default_29" -> "319 dequantize_per_channel_default_29"  [label="(1, 144, 14, 14)", style=solid];
-"319 dequantize_per_channel_default_29" -> "326 conv2d_32"  [label="(1, 144, 14, 14)", style=solid];
-"320 features_8_block_1_0_weight_bias_0_0" -> "326 conv2d_32"  [label="(144,)", style=solid];
-"321 scale_updated_constant22" -> "323 mul_tensor_22"  [label="(144, 1, 1, 1)", style=solid];
-"322 compressed_weight_updated_constant22" -> "323 mul_tensor_22"  [label="(144, 1, 5, 5)", style=solid];
-"323 mul_tensor_22" -> "325 sub_tensor_22"  [label="(144, 1, 5, 5)", style=solid];
-"324 zero_point_updated_constant22" -> "325 sub_tensor_22"  [label="(144, 1, 1, 1)", style=solid];
-"325 sub_tensor_22" -> "326 conv2d_32"  [label="(144, 1, 5, 5)", style=solid];
-"326 conv2d_32" -> "327 hardswish__10"  [label="(1, 144, 14, 14)", style=solid];
-"327 hardswish__10" -> "328 quantize_per_tensor_default_28"  [label="(1, 144, 14, 14)", style=solid];
-"328 quantize_per_tensor_default_28" -> "329 dequantize_per_tensor_default_38"  [label="(1, 144, 14, 14)", style=solid];
-"328 quantize_per_tensor_default_28" -> "330 dequantize_per_tensor_default_37"  [label="(1, 144, 14, 14)", style=solid];
-"329 dequantize_per_tensor_default_38" -> "342 mul_5"  [label="(1, 144, 14, 14)", style=solid];
-"330 dequantize_per_tensor_default_37" -> "331 adaptive_avg_pool2d_5"  [label="(1, 144, 14, 14)", style=solid];
-"331 adaptive_avg_pool2d_5" -> "334 conv2d_33"  [label="(1, 144, 1, 1)", style=solid];
-"332 _param_constant89" -> "334 conv2d_33"  [label="(40, 144, 1, 1)", style=solid];
-"333 _param_constant90_0_0" -> "334 conv2d_33"  [label="(40,)", style=solid];
-"334 conv2d_33" -> "335 relu_5"  [label="(1, 40, 1, 1)", style=solid];
-"335 relu_5" -> "338 conv2d_34"  [label="(1, 40, 1, 1)", style=solid];
-"336 _param_constant91" -> "338 conv2d_34"  [label="(144, 40, 1, 1)", style=solid];
-"337 _param_constant92_0_0" -> "338 conv2d_34"  [label="(144,)", style=solid];
-"338 conv2d_34" -> "339 hardsigmoid_5"  [label="(1, 144, 1, 1)", style=solid];
-"339 hardsigmoid_5" -> "340 quantize_per_tensor_default_29"  [label="(1, 144, 1, 1)", style=solid];
-"340 quantize_per_tensor_default_29" -> "341 dequantize_per_tensor_default_39"  [label="(1, 144, 1, 1)", style=solid];
-"341 dequantize_per_tensor_default_39" -> "342 mul_5"  [label="(1, 144, 1, 1)", style=solid];
-"342 mul_5" -> "343 quantize_per_tensor_default_30"  [label="(1, 144, 14, 14)", style=solid];
-"343 quantize_per_tensor_default_30" -> "344 dequantize_per_tensor_default_40"  [label="(1, 144, 14, 14)", style=solid];
-"344 dequantize_per_tensor_default_40" -> "351 conv2d_35"  [label="(1, 144, 14, 14)", style=solid];
-"345 features_8_block_3_0_weight_bias_0_0" -> "351 conv2d_35"  [label="(48,)", style=solid];
-"346 scale_updated_constant23" -> "348 mul_tensor_23"  [label="(48, 1, 1, 1)", style=solid];
-"347 compressed_weight_updated_constant23" -> "348 mul_tensor_23"  [label="(48, 144, 1, 1)", style=solid];
-"348 mul_tensor_23" -> "350 sub_tensor_23"  [label="(48, 144, 1, 1)", style=solid];
-"349 zero_point_updated_constant23" -> "350 sub_tensor_23"  [label="(48, 1, 1, 1)", style=solid];
-"350 sub_tensor_23" -> "351 conv2d_35"  [label="(48, 144, 1, 1)", style=solid];
-"351 conv2d_35" -> "352 quantize_per_tensor_default_31"  [label="(1, 48, 14, 14)", style=solid];
-"352 quantize_per_tensor_default_31" -> "353 dequantize_per_tensor_default_41"  [label="(1, 48, 14, 14)", style=solid];
-"353 dequantize_per_tensor_default_41" -> "354 add__3"  [label="(1, 48, 14, 14)", style=solid];
-"354 add__3" -> "355 quantize_per_tensor_default_32"  [label="(1, 48, 14, 14)", style=solid];
-"355 quantize_per_tensor_default_32" -> "356 dequantize_per_tensor_default_42"  [label="(1, 48, 14, 14)", style=solid];
-"356 dequantize_per_tensor_default_42" -> "363 conv2d_36"  [label="(1, 48, 14, 14)", style=solid];
-"357 features_9_block_0_0_weight_bias_0_0" -> "363 conv2d_36"  [label="(288,)", style=solid];
-"358 scale_updated_constant24" -> "360 mul_tensor_24"  [label="(288, 1, 1, 1)", style=solid];
-"359 compressed_weight_updated_constant24" -> "360 mul_tensor_24"  [label="(288, 48, 1, 1)", style=solid];
-"360 mul_tensor_24" -> "362 sub_tensor_24"  [label="(288, 48, 1, 1)", style=solid];
-"361 zero_point_updated_constant24" -> "362 sub_tensor_24"  [label="(288, 1, 1, 1)", style=solid];
-"362 sub_tensor_24" -> "363 conv2d_36"  [label="(288, 48, 1, 1)", style=solid];
-"363 conv2d_36" -> "366 hardswish__11"  [label="(1, 288, 14, 14)", style=solid];
-"364 hardswish__11_scale_0" -> "367 quantize_per_channel_default_33"  [label="(288,)", style=solid];
-"364 hardswish__11_scale_0" -> "368 dequantize_per_channel_default_33"  [label="(288,)", style=solid];
-"365 hardswish__11_zero_point_0" -> "367 quantize_per_channel_default_33"  [label="(288,)", style=solid];
-"365 hardswish__11_zero_point_0" -> "368 dequantize_per_channel_default_33"  [label="(288,)", style=solid];
-"366 hardswish__11" -> "367 quantize_per_channel_default_33"  [label="(1, 288, 14, 14)", style=solid];
-"367 quantize_per_channel_default_33" -> "368 dequantize_per_channel_default_33"  [label="(1, 288, 14, 14)", style=solid];
-"368 dequantize_per_channel_default_33" -> "375 conv2d_37"  [label="(1, 288, 14, 14)", style=solid];
-"369 features_9_block_1_0_weight_bias_0_0" -> "375 conv2d_37"  [label="(288,)", style=solid];
-"370 scale_updated_constant25" -> "372 mul_tensor_25"  [label="(288, 1, 1, 1)", style=solid];
-"371 compressed_weight_updated_constant25" -> "372 mul_tensor_25"  [label="(288, 1, 5, 5)", style=solid];
-"372 mul_tensor_25" -> "374 sub_tensor_25"  [label="(288, 1, 5, 5)", style=solid];
-"373 zero_point_updated_constant25" -> "374 sub_tensor_25"  [label="(288, 1, 1, 1)", style=solid];
-"374 sub_tensor_25" -> "375 conv2d_37"  [label="(288, 1, 5, 5)", style=solid];
-"375 conv2d_37" -> "376 hardswish__12"  [label="(1, 288, 7, 7)", style=solid];
-"376 hardswish__12" -> "377 quantize_per_tensor_default_33"  [label="(1, 288, 7, 7)", style=solid];
-"377 quantize_per_tensor_default_33" -> "378 dequantize_per_tensor_default_44"  [label="(1, 288, 7, 7)", style=solid];
-"377 quantize_per_tensor_default_33" -> "379 dequantize_per_tensor_default_43"  [label="(1, 288, 7, 7)", style=solid];
-"378 dequantize_per_tensor_default_44" -> "391 mul_6"  [label="(1, 288, 7, 7)", style=solid];
-"379 dequantize_per_tensor_default_43" -> "380 adaptive_avg_pool2d_6"  [label="(1, 288, 7, 7)", style=solid];
-"380 adaptive_avg_pool2d_6" -> "383 conv2d_38"  [label="(1, 288, 1, 1)", style=solid];
-"381 _param_constant102" -> "383 conv2d_38"  [label="(72, 288, 1, 1)", style=solid];
-"382 _param_constant103_0_0" -> "383 conv2d_38"  [label="(72,)", style=solid];
-"383 conv2d_38" -> "384 relu_6"  [label="(1, 72, 1, 1)", style=solid];
-"384 relu_6" -> "387 conv2d_39"  [label="(1, 72, 1, 1)", style=solid];
-"385 _param_constant104" -> "387 conv2d_39"  [label="(288, 72, 1, 1)", style=solid];
-"386 _param_constant105_0_0" -> "387 conv2d_39"  [label="(288,)", style=solid];
-"387 conv2d_39" -> "388 hardsigmoid_6"  [label="(1, 288, 1, 1)", style=solid];
-"388 hardsigmoid_6" -> "389 quantize_per_tensor_default_34"  [label="(1, 288, 1, 1)", style=solid];
-"389 quantize_per_tensor_default_34" -> "390 dequantize_per_tensor_default_45"  [label="(1, 288, 1, 1)", style=solid];
-"390 dequantize_per_tensor_default_45" -> "391 mul_6"  [label="(1, 288, 1, 1)", style=solid];
-"391 mul_6" -> "392 quantize_per_tensor_default_35"  [label="(1, 288, 7, 7)", style=solid];
-"392 quantize_per_tensor_default_35" -> "393 dequantize_per_tensor_default_46"  [label="(1, 288, 7, 7)", style=solid];
-"393 dequantize_per_tensor_default_46" -> "400 conv2d_40"  [label="(1, 288, 7, 7)", style=solid];
-"394 features_9_block_3_0_weight_bias_0_0" -> "400 conv2d_40"  [label="(96,)", style=solid];
-"395 scale_updated_constant26" -> "397 mul_tensor_26"  [label="(96, 1, 1, 1)", style=solid];
-"396 compressed_weight_updated_constant26" -> "397 mul_tensor_26"  [label="(96, 288, 1, 1)", style=solid];
-"397 mul_tensor_26" -> "399 sub_tensor_26"  [label="(96, 288, 1, 1)", style=solid];
-"398 zero_point_updated_constant26" -> "399 sub_tensor_26"  [label="(96, 1, 1, 1)", style=solid];
-"399 sub_tensor_26" -> "400 conv2d_40"  [label="(96, 288, 1, 1)", style=solid];
-"400 conv2d_40" -> "401 quantize_per_tensor_default_36"  [label="(1, 96, 7, 7)", style=solid];
-"401 quantize_per_tensor_default_36" -> "402 dequantize_per_tensor_default_48"  [label="(1, 96, 7, 7)", style=solid];
-"401 quantize_per_tensor_default_36" -> "403 dequantize_per_tensor_default_47"  [label="(1, 96, 7, 7)", style=solid];
-"402 dequantize_per_tensor_default_48" -> "450 add__4"  [label="(1, 96, 7, 7)", style=solid];
-"403 dequantize_per_tensor_default_47" -> "410 conv2d_41"  [label="(1, 96, 7, 7)", style=solid];
-"404 features_10_block_0_0_weight_bias_0_0" -> "410 conv2d_41"  [label="(576,)", style=solid];
-"405 scale_updated_constant27" -> "407 mul_tensor_27"  [label="(576, 1, 1, 1)", style=solid];
-"406 compressed_weight_updated_constant27" -> "407 mul_tensor_27"  [label="(576, 96, 1, 1)", style=solid];
-"407 mul_tensor_27" -> "409 sub_tensor_27"  [label="(576, 96, 1, 1)", style=solid];
-"408 zero_point_updated_constant27" -> "409 sub_tensor_27"  [label="(576, 1, 1, 1)", style=solid];
-"409 sub_tensor_27" -> "410 conv2d_41"  [label="(576, 96, 1, 1)", style=solid];
-"410 conv2d_41" -> "413 hardswish__13"  [label="(1, 576, 7, 7)", style=solid];
-"411 hardswish__13_scale_0" -> "414 quantize_per_channel_default_37"  [label="(576,)", style=solid];
-"411 hardswish__13_scale_0" -> "415 dequantize_per_channel_default_37"  [label="(576,)", style=solid];
-"412 hardswish__13_zero_point_0" -> "414 quantize_per_channel_default_37"  [label="(576,)", style=solid];
-"412 hardswish__13_zero_point_0" -> "415 dequantize_per_channel_default_37"  [label="(576,)", style=solid];
-"413 hardswish__13" -> "414 quantize_per_channel_default_37"  [label="(1, 576, 7, 7)", style=solid];
-"414 quantize_per_channel_default_37" -> "415 dequantize_per_channel_default_37"  [label="(1, 576, 7, 7)", style=solid];
-"415 dequantize_per_channel_default_37" -> "422 conv2d_42"  [label="(1, 576, 7, 7)", style=solid];
-"416 features_10_block_1_0_weight_bias_0_0" -> "422 conv2d_42"  [label="(576,)", style=solid];
-"417 scale_updated_constant28" -> "419 mul_tensor_28"  [label="(576, 1, 1, 1)", style=solid];
-"418 compressed_weight_updated_constant28" -> "419 mul_tensor_28"  [label="(576, 1, 5, 5)", style=solid];
-"419 mul_tensor_28" -> "421 sub_tensor_28"  [label="(576, 1, 5, 5)", style=solid];
-"420 zero_point_updated_constant28" -> "421 sub_tensor_28"  [label="(576, 1, 1, 1)", style=solid];
-"421 sub_tensor_28" -> "422 conv2d_42"  [label="(576, 1, 5, 5)", style=solid];
-"422 conv2d_42" -> "423 hardswish__14"  [label="(1, 576, 7, 7)", style=solid];
-"423 hardswish__14" -> "424 quantize_per_tensor_default_37"  [label="(1, 576, 7, 7)", style=solid];
-"424 quantize_per_tensor_default_37" -> "425 dequantize_per_tensor_default_50"  [label="(1, 576, 7, 7)", style=solid];
-"424 quantize_per_tensor_default_37" -> "426 dequantize_per_tensor_default_49"  [label="(1, 576, 7, 7)", style=solid];
-"425 dequantize_per_tensor_default_50" -> "438 mul_7"  [label="(1, 576, 7, 7)", style=solid];
-"426 dequantize_per_tensor_default_49" -> "427 adaptive_avg_pool2d_7"  [label="(1, 576, 7, 7)", style=solid];
-"427 adaptive_avg_pool2d_7" -> "430 conv2d_43"  [label="(1, 576, 1, 1)", style=solid];
-"428 _param_constant115" -> "430 conv2d_43"  [label="(144, 576, 1, 1)", style=solid];
-"429 _param_constant116_0_0" -> "430 conv2d_43"  [label="(144,)", style=solid];
-"430 conv2d_43" -> "431 relu_7"  [label="(1, 144, 1, 1)", style=solid];
-"431 relu_7" -> "434 conv2d_44"  [label="(1, 144, 1, 1)", style=solid];
-"432 _param_constant117" -> "434 conv2d_44"  [label="(576, 144, 1, 1)", style=solid];
-"433 _param_constant118_0_0" -> "434 conv2d_44"  [label="(576,)", style=solid];
-"434 conv2d_44" -> "435 hardsigmoid_7"  [label="(1, 576, 1, 1)", style=solid];
-"435 hardsigmoid_7" -> "436 quantize_per_tensor_default_38"  [label="(1, 576, 1, 1)", style=solid];
-"436 quantize_per_tensor_default_38" -> "437 dequantize_per_tensor_default_51"  [label="(1, 576, 1, 1)", style=solid];
-"437 dequantize_per_tensor_default_51" -> "438 mul_7"  [label="(1, 576, 1, 1)", style=solid];
-"438 mul_7" -> "439 quantize_per_tensor_default_39"  [label="(1, 576, 7, 7)", style=solid];
-"439 quantize_per_tensor_default_39" -> "440 dequantize_per_tensor_default_52"  [label="(1, 576, 7, 7)", style=solid];
-"440 dequantize_per_tensor_default_52" -> "447 conv2d_45"  [label="(1, 576, 7, 7)", style=solid];
-"441 features_10_block_3_0_weight_bias_0_0" -> "447 conv2d_45"  [label="(96,)", style=solid];
-"442 scale_updated_constant29" -> "444 mul_tensor_29"  [label="(96, 1, 1, 1)", style=solid];
-"443 compressed_weight_updated_constant29" -> "444 mul_tensor_29"  [label="(96, 576, 1, 1)", style=solid];
-"444 mul_tensor_29" -> "446 sub_tensor_29"  [label="(96, 576, 1, 1)", style=solid];
-"445 zero_point_updated_constant29" -> "446 sub_tensor_29"  [label="(96, 1, 1, 1)", style=solid];
-"446 sub_tensor_29" -> "447 conv2d_45"  [label="(96, 576, 1, 1)", style=solid];
-"447 conv2d_45" -> "448 quantize_per_tensor_default_40"  [label="(1, 96, 7, 7)", style=solid];
-"448 quantize_per_tensor_default_40" -> "449 dequantize_per_tensor_default_53"  [label="(1, 96, 7, 7)", style=solid];
-"449 dequantize_per_tensor_default_53" -> "450 add__4"  [label="(1, 96, 7, 7)", style=solid];
-"450 add__4" -> "451 quantize_per_tensor_default_41"  [label="(1, 96, 7, 7)", style=solid];
-"451 quantize_per_tensor_default_41" -> "452 dequantize_per_tensor_default_55"  [label="(1, 96, 7, 7)", style=solid];
-"451 quantize_per_tensor_default_41" -> "453 dequantize_per_tensor_default_54"  [label="(1, 96, 7, 7)", style=solid];
-"452 dequantize_per_tensor_default_55" -> "500 add__5"  [label="(1, 96, 7, 7)", style=solid];
-"453 dequantize_per_tensor_default_54" -> "460 conv2d_46"  [label="(1, 96, 7, 7)", style=solid];
-"454 features_11_block_0_0_weight_bias_0_0" -> "460 conv2d_46"  [label="(576,)", style=solid];
-"455 scale_updated_constant30" -> "457 mul_tensor_30"  [label="(576, 1, 1, 1)", style=solid];
-"456 compressed_weight_updated_constant30" -> "457 mul_tensor_30"  [label="(576, 96, 1, 1)", style=solid];
-"457 mul_tensor_30" -> "459 sub_tensor_30"  [label="(576, 96, 1, 1)", style=solid];
-"458 zero_point_updated_constant30" -> "459 sub_tensor_30"  [label="(576, 1, 1, 1)", style=solid];
-"459 sub_tensor_30" -> "460 conv2d_46"  [label="(576, 96, 1, 1)", style=solid];
-"460 conv2d_46" -> "463 hardswish__15"  [label="(1, 576, 7, 7)", style=solid];
-"461 hardswish__15_scale_0" -> "464 quantize_per_channel_default_41"  [label="(576,)", style=solid];
-"461 hardswish__15_scale_0" -> "465 dequantize_per_channel_default_41"  [label="(576,)", style=solid];
-"462 hardswish__15_zero_point_0" -> "464 quantize_per_channel_default_41"  [label="(576,)", style=solid];
-"462 hardswish__15_zero_point_0" -> "465 dequantize_per_channel_default_41"  [label="(576,)", style=solid];
-"463 hardswish__15" -> "464 quantize_per_channel_default_41"  [label="(1, 576, 7, 7)", style=solid];
-"464 quantize_per_channel_default_41" -> "465 dequantize_per_channel_default_41"  [label="(1, 576, 7, 7)", style=solid];
-"465 dequantize_per_channel_default_41" -> "472 conv2d_47"  [label="(1, 576, 7, 7)", style=solid];
-"466 features_11_block_1_0_weight_bias_0_0" -> "472 conv2d_47"  [label="(576,)", style=solid];
-"467 scale_updated_constant31" -> "469 mul_tensor_31"  [label="(576, 1, 1, 1)", style=solid];
-"468 compressed_weight_updated_constant31" -> "469 mul_tensor_31"  [label="(576, 1, 5, 5)", style=solid];
-"469 mul_tensor_31" -> "471 sub_tensor_31"  [label="(576, 1, 5, 5)", style=solid];
-"470 zero_point_updated_constant31" -> "471 sub_tensor_31"  [label="(576, 1, 1, 1)", style=solid];
-"471 sub_tensor_31" -> "472 conv2d_47"  [label="(576, 1, 5, 5)", style=solid];
-"472 conv2d_47" -> "473 hardswish__16"  [label="(1, 576, 7, 7)", style=solid];
-"473 hardswish__16" -> "474 quantize_per_tensor_default_42"  [label="(1, 576, 7, 7)", style=solid];
-"474 quantize_per_tensor_default_42" -> "475 dequantize_per_tensor_default_57"  [label="(1, 576, 7, 7)", style=solid];
-"474 quantize_per_tensor_default_42" -> "476 dequantize_per_tensor_default_56"  [label="(1, 576, 7, 7)", style=solid];
-"475 dequantize_per_tensor_default_57" -> "488 mul_8"  [label="(1, 576, 7, 7)", style=solid];
-"476 dequantize_per_tensor_default_56" -> "477 adaptive_avg_pool2d_8"  [label="(1, 576, 7, 7)", style=solid];
-"477 adaptive_avg_pool2d_8" -> "480 conv2d_48"  [label="(1, 576, 1, 1)", style=solid];
-"478 _param_constant128" -> "480 conv2d_48"  [label="(144, 576, 1, 1)", style=solid];
-"479 _param_constant129_0_0" -> "480 conv2d_48"  [label="(144,)", style=solid];
-"480 conv2d_48" -> "481 relu_8"  [label="(1, 144, 1, 1)", style=solid];
-"481 relu_8" -> "484 conv2d_49"  [label="(1, 144, 1, 1)", style=solid];
-"482 _param_constant130" -> "484 conv2d_49"  [label="(576, 144, 1, 1)", style=solid];
-"483 _param_constant131_0_0" -> "484 conv2d_49"  [label="(576,)", style=solid];
+"0 features_1_block_1_fc1_weight" -> "46 conv2d_2"  [label="(8, 16, 1, 1)", style=solid];
+"1 features_1_block_1_fc2_weight" -> "49 conv2d_3"  [label="(16, 8, 1, 1)", style=solid];
+"2 features_4_block_2_fc1_weight" -> "156 conv2d_13"  [label="(24, 96, 1, 1)", style=solid];
+"3 features_4_block_2_fc2_weight" -> "159 conv2d_14"  [label="(96, 24, 1, 1)", style=solid];
+"4 features_5_block_2_fc1_weight" -> "201 conv2d_18"  [label="(64, 240, 1, 1)", style=solid];
+"5 features_5_block_2_fc2_weight" -> "204 conv2d_19"  [label="(240, 64, 1, 1)", style=solid];
+"6 features_6_block_2_fc1_weight" -> "249 conv2d_23"  [label="(64, 240, 1, 1)", style=solid];
+"7 features_6_block_2_fc2_weight" -> "252 conv2d_24"  [label="(240, 64, 1, 1)", style=solid];
+"8 features_7_block_2_fc1_weight" -> "296 conv2d_28"  [label="(32, 120, 1, 1)", style=solid];
+"9 features_7_block_2_fc2_weight" -> "299 conv2d_29"  [label="(120, 32, 1, 1)", style=solid];
+"10 features_8_block_2_fc1_weight" -> "341 conv2d_33"  [label="(40, 144, 1, 1)", style=solid];
+"11 features_8_block_2_fc2_weight" -> "344 conv2d_34"  [label="(144, 40, 1, 1)", style=solid];
+"12 features_9_block_2_fc1_weight" -> "388 conv2d_38"  [label="(72, 288, 1, 1)", style=solid];
+"13 features_9_block_2_fc2_weight" -> "391 conv2d_39"  [label="(288, 72, 1, 1)", style=solid];
+"14 features_10_block_2_fc1_weight" -> "433 conv2d_43"  [label="(144, 576, 1, 1)", style=solid];
+"15 features_10_block_2_fc2_weight" -> "436 conv2d_44"  [label="(576, 144, 1, 1)", style=solid];
+"16 features_11_block_2_fc1_weight" -> "481 conv2d_48"  [label="(144, 576, 1, 1)", style=solid];
+"17 features_11_block_2_fc2_weight" -> "484 conv2d_49"  [label="(576, 144, 1, 1)", style=solid];
+"18 x" -> "19 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"19 quantize_per_tensor_default" -> "20 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"20 dequantize_per_tensor_default" -> "27 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"21 features_0_0_weight_bias_0_0" -> "27 conv2d"  [label="(16,)", style=solid];
+"22 scale_updated_constant0" -> "24 mul_tensor"  [label="(16, 1, 1, 1)", style=solid];
+"23 compressed_weight_updated_constant0" -> "24 mul_tensor"  [label="(16, 3, 3, 3)", style=solid];
+"24 mul_tensor" -> "26 sub_tensor"  [label="(16, 3, 3, 3)", style=solid];
+"25 zero_point_updated_constant0" -> "26 sub_tensor"  [label="(16, 1, 1, 1)", style=solid];
+"26 sub_tensor" -> "27 conv2d"  [label="(16, 3, 3, 3)", style=solid];
+"27 conv2d" -> "30 hardswish_"  [label="(1, 16, 112, 112)", style=solid];
+"28 hardswish__scale_0" -> "31 quantize_per_channel_default_1"  [label="(16,)", style=solid];
+"28 hardswish__scale_0" -> "32 dequantize_per_channel_default_1"  [label="(16,)", style=solid];
+"29 hardswish__zero_point_0" -> "31 quantize_per_channel_default_1"  [label="(16,)", style=solid];
+"29 hardswish__zero_point_0" -> "32 dequantize_per_channel_default_1"  [label="(16,)", style=solid];
+"30 hardswish_" -> "31 quantize_per_channel_default_1"  [label="(1, 16, 112, 112)", style=solid];
+"31 quantize_per_channel_default_1" -> "32 dequantize_per_channel_default_1"  [label="(1, 16, 112, 112)", style=solid];
+"32 dequantize_per_channel_default_1" -> "39 conv2d_1"  [label="(1, 16, 112, 112)", style=solid];
+"33 features_1_block_0_0_weight_bias_0_0" -> "39 conv2d_1"  [label="(16,)", style=solid];
+"34 scale_updated_constant1" -> "36 mul_tensor_1"  [label="(16, 1, 1, 1)", style=solid];
+"35 compressed_weight_updated_constant1" -> "36 mul_tensor_1"  [label="(16, 1, 3, 3)", style=solid];
+"36 mul_tensor_1" -> "38 sub_tensor_1"  [label="(16, 1, 3, 3)", style=solid];
+"37 zero_point_updated_constant1" -> "38 sub_tensor_1"  [label="(16, 1, 1, 1)", style=solid];
+"38 sub_tensor_1" -> "39 conv2d_1"  [label="(16, 1, 3, 3)", style=solid];
+"39 conv2d_1" -> "40 relu_"  [label="(1, 16, 56, 56)", style=solid];
+"40 relu_" -> "41 quantize_per_tensor_default_1"  [label="(1, 16, 56, 56)", style=solid];
+"41 quantize_per_tensor_default_1" -> "42 dequantize_per_tensor_default_2"  [label="(1, 16, 56, 56)", style=solid];
+"41 quantize_per_tensor_default_1" -> "43 dequantize_per_tensor_default_1"  [label="(1, 16, 56, 56)", style=solid];
+"42 dequantize_per_tensor_default_2" -> "53 mul"  [label="(1, 16, 56, 56)", style=solid];
+"43 dequantize_per_tensor_default_1" -> "44 adaptive_avg_pool2d"  [label="(1, 16, 56, 56)", style=solid];
+"44 adaptive_avg_pool2d" -> "46 conv2d_2"  [label="(1, 16, 1, 1)", style=solid];
+"45 features_1_block_1_fc1_bias_0_0" -> "46 conv2d_2"  [label="(8,)", style=solid];
+"46 conv2d_2" -> "47 relu"  [label="(1, 8, 1, 1)", style=solid];
+"47 relu" -> "49 conv2d_3"  [label="(1, 8, 1, 1)", style=solid];
+"48 features_1_block_1_fc2_bias_0_0" -> "49 conv2d_3"  [label="(16,)", style=solid];
+"49 conv2d_3" -> "50 hardsigmoid"  [label="(1, 16, 1, 1)", style=solid];
+"50 hardsigmoid" -> "51 quantize_per_tensor_default_2"  [label="(1, 16, 1, 1)", style=solid];
+"51 quantize_per_tensor_default_2" -> "52 dequantize_per_tensor_default_3"  [label="(1, 16, 1, 1)", style=solid];
+"52 dequantize_per_tensor_default_3" -> "53 mul"  [label="(1, 16, 1, 1)", style=solid];
+"53 mul" -> "54 quantize_per_tensor_default_3"  [label="(1, 16, 56, 56)", style=solid];
+"54 quantize_per_tensor_default_3" -> "55 dequantize_per_tensor_default_4"  [label="(1, 16, 56, 56)", style=solid];
+"55 dequantize_per_tensor_default_4" -> "62 conv2d_4"  [label="(1, 16, 56, 56)", style=solid];
+"56 features_1_block_2_0_weight_bias_0_0" -> "62 conv2d_4"  [label="(16,)", style=solid];
+"57 scale_updated_constant2" -> "59 mul_tensor_2"  [label="(16, 1, 1, 1)", style=solid];
+"58 compressed_weight_updated_constant2" -> "59 mul_tensor_2"  [label="(16, 16, 1, 1)", style=solid];
+"59 mul_tensor_2" -> "61 sub_tensor_2"  [label="(16, 16, 1, 1)", style=solid];
+"60 zero_point_updated_constant2" -> "61 sub_tensor_2"  [label="(16, 1, 1, 1)", style=solid];
+"61 sub_tensor_2" -> "62 conv2d_4"  [label="(16, 16, 1, 1)", style=solid];
+"62 conv2d_4" -> "63 quantize_per_tensor_default_4"  [label="(1, 16, 56, 56)", style=solid];
+"63 quantize_per_tensor_default_4" -> "64 dequantize_per_tensor_default_5"  [label="(1, 16, 56, 56)", style=solid];
+"64 dequantize_per_tensor_default_5" -> "71 conv2d_5"  [label="(1, 16, 56, 56)", style=solid];
+"65 features_2_block_0_0_weight_bias_0_0" -> "71 conv2d_5"  [label="(72,)", style=solid];
+"66 scale_updated_constant3" -> "68 mul_tensor_3"  [label="(72, 1, 1, 1)", style=solid];
+"67 compressed_weight_updated_constant3" -> "68 mul_tensor_3"  [label="(72, 16, 1, 1)", style=solid];
+"68 mul_tensor_3" -> "70 sub_tensor_3"  [label="(72, 16, 1, 1)", style=solid];
+"69 zero_point_updated_constant3" -> "70 sub_tensor_3"  [label="(72, 1, 1, 1)", style=solid];
+"70 sub_tensor_3" -> "71 conv2d_5"  [label="(72, 16, 1, 1)", style=solid];
+"71 conv2d_5" -> "74 relu__1"  [label="(1, 72, 56, 56)", style=solid];
+"72 relu__1_scale_0" -> "75 quantize_per_channel_default_5"  [label="(72,)", style=solid];
+"72 relu__1_scale_0" -> "76 dequantize_per_channel_default_5"  [label="(72,)", style=solid];
+"73 relu__1_zero_point_0" -> "75 quantize_per_channel_default_5"  [label="(72,)", style=solid];
+"73 relu__1_zero_point_0" -> "76 dequantize_per_channel_default_5"  [label="(72,)", style=solid];
+"74 relu__1" -> "75 quantize_per_channel_default_5"  [label="(1, 72, 56, 56)", style=solid];
+"75 quantize_per_channel_default_5" -> "76 dequantize_per_channel_default_5"  [label="(1, 72, 56, 56)", style=solid];
+"76 dequantize_per_channel_default_5" -> "83 conv2d_6"  [label="(1, 72, 56, 56)", style=solid];
+"77 features_2_block_1_0_weight_bias_0_0" -> "83 conv2d_6"  [label="(72,)", style=solid];
+"78 scale_updated_constant4" -> "80 mul_tensor_4"  [label="(72, 1, 1, 1)", style=solid];
+"79 compressed_weight_updated_constant4" -> "80 mul_tensor_4"  [label="(72, 1, 3, 3)", style=solid];
+"80 mul_tensor_4" -> "82 sub_tensor_4"  [label="(72, 1, 3, 3)", style=solid];
+"81 zero_point_updated_constant4" -> "82 sub_tensor_4"  [label="(72, 1, 1, 1)", style=solid];
+"82 sub_tensor_4" -> "83 conv2d_6"  [label="(72, 1, 3, 3)", style=solid];
+"83 conv2d_6" -> "84 relu__2"  [label="(1, 72, 28, 28)", style=solid];
+"84 relu__2" -> "85 quantize_per_tensor_default_5"  [label="(1, 72, 28, 28)", style=solid];
+"85 quantize_per_tensor_default_5" -> "86 dequantize_per_tensor_default_6"  [label="(1, 72, 28, 28)", style=solid];
+"86 dequantize_per_tensor_default_6" -> "93 conv2d_7"  [label="(1, 72, 28, 28)", style=solid];
+"87 features_2_block_2_0_weight_bias_0_0" -> "93 conv2d_7"  [label="(24,)", style=solid];
+"88 scale_updated_constant5" -> "90 mul_tensor_5"  [label="(24, 1, 1, 1)", style=solid];
+"89 compressed_weight_updated_constant5" -> "90 mul_tensor_5"  [label="(24, 72, 1, 1)", style=solid];
+"90 mul_tensor_5" -> "92 sub_tensor_5"  [label="(24, 72, 1, 1)", style=solid];
+"91 zero_point_updated_constant5" -> "92 sub_tensor_5"  [label="(24, 1, 1, 1)", style=solid];
+"92 sub_tensor_5" -> "93 conv2d_7"  [label="(24, 72, 1, 1)", style=solid];
+"93 conv2d_7" -> "94 quantize_per_tensor_default_6"  [label="(1, 24, 28, 28)", style=solid];
+"94 quantize_per_tensor_default_6" -> "95 dequantize_per_tensor_default_8"  [label="(1, 24, 28, 28)", style=solid];
+"94 quantize_per_tensor_default_6" -> "96 dequantize_per_tensor_default_7"  [label="(1, 24, 28, 28)", style=solid];
+"95 dequantize_per_tensor_default_8" -> "128 add_"  [label="(1, 24, 28, 28)", style=solid];
+"96 dequantize_per_tensor_default_7" -> "103 conv2d_8"  [label="(1, 24, 28, 28)", style=solid];
+"97 features_3_block_0_0_weight_bias_0_0" -> "103 conv2d_8"  [label="(88,)", style=solid];
+"98 scale_updated_constant6" -> "100 mul_tensor_6"  [label="(88, 1, 1, 1)", style=solid];
+"99 compressed_weight_updated_constant6" -> "100 mul_tensor_6"  [label="(88, 24, 1, 1)", style=solid];
+"100 mul_tensor_6" -> "102 sub_tensor_6"  [label="(88, 24, 1, 1)", style=solid];
+"101 zero_point_updated_constant6" -> "102 sub_tensor_6"  [label="(88, 1, 1, 1)", style=solid];
+"102 sub_tensor_6" -> "103 conv2d_8"  [label="(88, 24, 1, 1)", style=solid];
+"103 conv2d_8" -> "106 relu__3"  [label="(1, 88, 28, 28)", style=solid];
+"104 relu__3_scale_0" -> "107 quantize_per_channel_default_9"  [label="(88,)", style=solid];
+"104 relu__3_scale_0" -> "108 dequantize_per_channel_default_9"  [label="(88,)", style=solid];
+"105 relu__3_zero_point_0" -> "107 quantize_per_channel_default_9"  [label="(88,)", style=solid];
+"105 relu__3_zero_point_0" -> "108 dequantize_per_channel_default_9"  [label="(88,)", style=solid];
+"106 relu__3" -> "107 quantize_per_channel_default_9"  [label="(1, 88, 28, 28)", style=solid];
+"107 quantize_per_channel_default_9" -> "108 dequantize_per_channel_default_9"  [label="(1, 88, 28, 28)", style=solid];
+"108 dequantize_per_channel_default_9" -> "115 conv2d_9"  [label="(1, 88, 28, 28)", style=solid];
+"109 features_3_block_1_0_weight_bias_0_0" -> "115 conv2d_9"  [label="(88,)", style=solid];
+"110 scale_updated_constant7" -> "112 mul_tensor_7"  [label="(88, 1, 1, 1)", style=solid];
+"111 compressed_weight_updated_constant7" -> "112 mul_tensor_7"  [label="(88, 1, 3, 3)", style=solid];
+"112 mul_tensor_7" -> "114 sub_tensor_7"  [label="(88, 1, 3, 3)", style=solid];
+"113 zero_point_updated_constant7" -> "114 sub_tensor_7"  [label="(88, 1, 1, 1)", style=solid];
+"114 sub_tensor_7" -> "115 conv2d_9"  [label="(88, 1, 3, 3)", style=solid];
+"115 conv2d_9" -> "116 relu__4"  [label="(1, 88, 28, 28)", style=solid];
+"116 relu__4" -> "117 quantize_per_tensor_default_7"  [label="(1, 88, 28, 28)", style=solid];
+"117 quantize_per_tensor_default_7" -> "118 dequantize_per_tensor_default_9"  [label="(1, 88, 28, 28)", style=solid];
+"118 dequantize_per_tensor_default_9" -> "125 conv2d_10"  [label="(1, 88, 28, 28)", style=solid];
+"119 features_3_block_2_0_weight_bias_0_0" -> "125 conv2d_10"  [label="(24,)", style=solid];
+"120 scale_updated_constant8" -> "122 mul_tensor_8"  [label="(24, 1, 1, 1)", style=solid];
+"121 compressed_weight_updated_constant8" -> "122 mul_tensor_8"  [label="(24, 88, 1, 1)", style=solid];
+"122 mul_tensor_8" -> "124 sub_tensor_8"  [label="(24, 88, 1, 1)", style=solid];
+"123 zero_point_updated_constant8" -> "124 sub_tensor_8"  [label="(24, 1, 1, 1)", style=solid];
+"124 sub_tensor_8" -> "125 conv2d_10"  [label="(24, 88, 1, 1)", style=solid];
+"125 conv2d_10" -> "126 quantize_per_tensor_default_8"  [label="(1, 24, 28, 28)", style=solid];
+"126 quantize_per_tensor_default_8" -> "127 dequantize_per_tensor_default_10"  [label="(1, 24, 28, 28)", style=solid];
+"127 dequantize_per_tensor_default_10" -> "128 add_"  [label="(1, 24, 28, 28)", style=solid];
+"128 add_" -> "129 quantize_per_tensor_default_9"  [label="(1, 24, 28, 28)", style=solid];
+"129 quantize_per_tensor_default_9" -> "130 dequantize_per_tensor_default_11"  [label="(1, 24, 28, 28)", style=solid];
+"130 dequantize_per_tensor_default_11" -> "137 conv2d_11"  [label="(1, 24, 28, 28)", style=solid];
+"131 features_4_block_0_0_weight_bias_0_0" -> "137 conv2d_11"  [label="(96,)", style=solid];
+"132 scale_updated_constant9" -> "134 mul_tensor_9"  [label="(96, 1, 1, 1)", style=solid];
+"133 compressed_weight_updated_constant9" -> "134 mul_tensor_9"  [label="(96, 24, 1, 1)", style=solid];
+"134 mul_tensor_9" -> "136 sub_tensor_9"  [label="(96, 24, 1, 1)", style=solid];
+"135 zero_point_updated_constant9" -> "136 sub_tensor_9"  [label="(96, 1, 1, 1)", style=solid];
+"136 sub_tensor_9" -> "137 conv2d_11"  [label="(96, 24, 1, 1)", style=solid];
+"137 conv2d_11" -> "140 hardswish__1"  [label="(1, 96, 28, 28)", style=solid];
+"138 hardswish__1_scale_0" -> "141 quantize_per_channel_default_13"  [label="(96,)", style=solid];
+"138 hardswish__1_scale_0" -> "142 dequantize_per_channel_default_13"  [label="(96,)", style=solid];
+"139 hardswish__1_zero_point_0" -> "141 quantize_per_channel_default_13"  [label="(96,)", style=solid];
+"139 hardswish__1_zero_point_0" -> "142 dequantize_per_channel_default_13"  [label="(96,)", style=solid];
+"140 hardswish__1" -> "141 quantize_per_channel_default_13"  [label="(1, 96, 28, 28)", style=solid];
+"141 quantize_per_channel_default_13" -> "142 dequantize_per_channel_default_13"  [label="(1, 96, 28, 28)", style=solid];
+"142 dequantize_per_channel_default_13" -> "149 conv2d_12"  [label="(1, 96, 28, 28)", style=solid];
+"143 features_4_block_1_0_weight_bias_0_0" -> "149 conv2d_12"  [label="(96,)", style=solid];
+"144 scale_updated_constant10" -> "146 mul_tensor_10"  [label="(96, 1, 1, 1)", style=solid];
+"145 compressed_weight_updated_constant10" -> "146 mul_tensor_10"  [label="(96, 1, 5, 5)", style=solid];
+"146 mul_tensor_10" -> "148 sub_tensor_10"  [label="(96, 1, 5, 5)", style=solid];
+"147 zero_point_updated_constant10" -> "148 sub_tensor_10"  [label="(96, 1, 1, 1)", style=solid];
+"148 sub_tensor_10" -> "149 conv2d_12"  [label="(96, 1, 5, 5)", style=solid];
+"149 conv2d_12" -> "150 hardswish__2"  [label="(1, 96, 14, 14)", style=solid];
+"150 hardswish__2" -> "151 quantize_per_tensor_default_10"  [label="(1, 96, 14, 14)", style=solid];
+"151 quantize_per_tensor_default_10" -> "152 dequantize_per_tensor_default_13"  [label="(1, 96, 14, 14)", style=solid];
+"151 quantize_per_tensor_default_10" -> "153 dequantize_per_tensor_default_12"  [label="(1, 96, 14, 14)", style=solid];
+"152 dequantize_per_tensor_default_13" -> "163 mul_1"  [label="(1, 96, 14, 14)", style=solid];
+"153 dequantize_per_tensor_default_12" -> "154 adaptive_avg_pool2d_1"  [label="(1, 96, 14, 14)", style=solid];
+"154 adaptive_avg_pool2d_1" -> "156 conv2d_13"  [label="(1, 96, 1, 1)", style=solid];
+"155 features_4_block_2_fc1_bias_0_0" -> "156 conv2d_13"  [label="(24,)", style=solid];
+"156 conv2d_13" -> "157 relu_1"  [label="(1, 24, 1, 1)", style=solid];
+"157 relu_1" -> "159 conv2d_14"  [label="(1, 24, 1, 1)", style=solid];
+"158 features_4_block_2_fc2_bias_0_0" -> "159 conv2d_14"  [label="(96,)", style=solid];
+"159 conv2d_14" -> "160 hardsigmoid_1"  [label="(1, 96, 1, 1)", style=solid];
+"160 hardsigmoid_1" -> "161 quantize_per_tensor_default_11"  [label="(1, 96, 1, 1)", style=solid];
+"161 quantize_per_tensor_default_11" -> "162 dequantize_per_tensor_default_14"  [label="(1, 96, 1, 1)", style=solid];
+"162 dequantize_per_tensor_default_14" -> "163 mul_1"  [label="(1, 96, 1, 1)", style=solid];
+"163 mul_1" -> "164 quantize_per_tensor_default_12"  [label="(1, 96, 14, 14)", style=solid];
+"164 quantize_per_tensor_default_12" -> "165 dequantize_per_tensor_default_15"  [label="(1, 96, 14, 14)", style=solid];
+"165 dequantize_per_tensor_default_15" -> "172 conv2d_15"  [label="(1, 96, 14, 14)", style=solid];
+"166 features_4_block_3_0_weight_bias_0_0" -> "172 conv2d_15"  [label="(40,)", style=solid];
+"167 scale_updated_constant11" -> "169 mul_tensor_11"  [label="(40, 1, 1, 1)", style=solid];
+"168 compressed_weight_updated_constant11" -> "169 mul_tensor_11"  [label="(40, 96, 1, 1)", style=solid];
+"169 mul_tensor_11" -> "171 sub_tensor_11"  [label="(40, 96, 1, 1)", style=solid];
+"170 zero_point_updated_constant11" -> "171 sub_tensor_11"  [label="(40, 1, 1, 1)", style=solid];
+"171 sub_tensor_11" -> "172 conv2d_15"  [label="(40, 96, 1, 1)", style=solid];
+"172 conv2d_15" -> "173 quantize_per_tensor_default_13"  [label="(1, 40, 14, 14)", style=solid];
+"173 quantize_per_tensor_default_13" -> "174 dequantize_per_tensor_default_17"  [label="(1, 40, 14, 14)", style=solid];
+"173 quantize_per_tensor_default_13" -> "175 dequantize_per_tensor_default_16"  [label="(1, 40, 14, 14)", style=solid];
+"174 dequantize_per_tensor_default_17" -> "220 add__1"  [label="(1, 40, 14, 14)", style=solid];
+"175 dequantize_per_tensor_default_16" -> "182 conv2d_16"  [label="(1, 40, 14, 14)", style=solid];
+"176 features_5_block_0_0_weight_bias_0_0" -> "182 conv2d_16"  [label="(240,)", style=solid];
+"177 scale_updated_constant12" -> "179 mul_tensor_12"  [label="(240, 1, 1, 1)", style=solid];
+"178 compressed_weight_updated_constant12" -> "179 mul_tensor_12"  [label="(240, 40, 1, 1)", style=solid];
+"179 mul_tensor_12" -> "181 sub_tensor_12"  [label="(240, 40, 1, 1)", style=solid];
+"180 zero_point_updated_constant12" -> "181 sub_tensor_12"  [label="(240, 1, 1, 1)", style=solid];
+"181 sub_tensor_12" -> "182 conv2d_16"  [label="(240, 40, 1, 1)", style=solid];
+"182 conv2d_16" -> "185 hardswish__3"  [label="(1, 240, 14, 14)", style=solid];
+"183 hardswish__3_scale_0" -> "186 quantize_per_channel_default_17"  [label="(240,)", style=solid];
+"183 hardswish__3_scale_0" -> "187 dequantize_per_channel_default_17"  [label="(240,)", style=solid];
+"184 hardswish__3_zero_point_0" -> "186 quantize_per_channel_default_17"  [label="(240,)", style=solid];
+"184 hardswish__3_zero_point_0" -> "187 dequantize_per_channel_default_17"  [label="(240,)", style=solid];
+"185 hardswish__3" -> "186 quantize_per_channel_default_17"  [label="(1, 240, 14, 14)", style=solid];
+"186 quantize_per_channel_default_17" -> "187 dequantize_per_channel_default_17"  [label="(1, 240, 14, 14)", style=solid];
+"187 dequantize_per_channel_default_17" -> "194 conv2d_17"  [label="(1, 240, 14, 14)", style=solid];
+"188 features_5_block_1_0_weight_bias_0_0" -> "194 conv2d_17"  [label="(240,)", style=solid];
+"189 scale_updated_constant13" -> "191 mul_tensor_13"  [label="(240, 1, 1, 1)", style=solid];
+"190 compressed_weight_updated_constant13" -> "191 mul_tensor_13"  [label="(240, 1, 5, 5)", style=solid];
+"191 mul_tensor_13" -> "193 sub_tensor_13"  [label="(240, 1, 5, 5)", style=solid];
+"192 zero_point_updated_constant13" -> "193 sub_tensor_13"  [label="(240, 1, 1, 1)", style=solid];
+"193 sub_tensor_13" -> "194 conv2d_17"  [label="(240, 1, 5, 5)", style=solid];
+"194 conv2d_17" -> "195 hardswish__4"  [label="(1, 240, 14, 14)", style=solid];
+"195 hardswish__4" -> "196 quantize_per_tensor_default_14"  [label="(1, 240, 14, 14)", style=solid];
+"196 quantize_per_tensor_default_14" -> "197 dequantize_per_tensor_default_19"  [label="(1, 240, 14, 14)", style=solid];
+"196 quantize_per_tensor_default_14" -> "198 dequantize_per_tensor_default_18"  [label="(1, 240, 14, 14)", style=solid];
+"197 dequantize_per_tensor_default_19" -> "208 mul_2"  [label="(1, 240, 14, 14)", style=solid];
+"198 dequantize_per_tensor_default_18" -> "199 adaptive_avg_pool2d_2"  [label="(1, 240, 14, 14)", style=solid];
+"199 adaptive_avg_pool2d_2" -> "201 conv2d_18"  [label="(1, 240, 1, 1)", style=solid];
+"200 features_5_block_2_fc1_bias_0_0" -> "201 conv2d_18"  [label="(64,)", style=solid];
+"201 conv2d_18" -> "202 relu_2"  [label="(1, 64, 1, 1)", style=solid];
+"202 relu_2" -> "204 conv2d_19"  [label="(1, 64, 1, 1)", style=solid];
+"203 features_5_block_2_fc2_bias_0_0" -> "204 conv2d_19"  [label="(240,)", style=solid];
+"204 conv2d_19" -> "205 hardsigmoid_2"  [label="(1, 240, 1, 1)", style=solid];
+"205 hardsigmoid_2" -> "206 quantize_per_tensor_default_15"  [label="(1, 240, 1, 1)", style=solid];
+"206 quantize_per_tensor_default_15" -> "207 dequantize_per_tensor_default_20"  [label="(1, 240, 1, 1)", style=solid];
+"207 dequantize_per_tensor_default_20" -> "208 mul_2"  [label="(1, 240, 1, 1)", style=solid];
+"208 mul_2" -> "209 quantize_per_tensor_default_16"  [label="(1, 240, 14, 14)", style=solid];
+"209 quantize_per_tensor_default_16" -> "210 dequantize_per_tensor_default_21"  [label="(1, 240, 14, 14)", style=solid];
+"210 dequantize_per_tensor_default_21" -> "217 conv2d_20"  [label="(1, 240, 14, 14)", style=solid];
+"211 features_5_block_3_0_weight_bias_0_0" -> "217 conv2d_20"  [label="(40,)", style=solid];
+"212 scale_updated_constant14" -> "214 mul_tensor_14"  [label="(40, 1, 1, 1)", style=solid];
+"213 compressed_weight_updated_constant14" -> "214 mul_tensor_14"  [label="(40, 240, 1, 1)", style=solid];
+"214 mul_tensor_14" -> "216 sub_tensor_14"  [label="(40, 240, 1, 1)", style=solid];
+"215 zero_point_updated_constant14" -> "216 sub_tensor_14"  [label="(40, 1, 1, 1)", style=solid];
+"216 sub_tensor_14" -> "217 conv2d_20"  [label="(40, 240, 1, 1)", style=solid];
+"217 conv2d_20" -> "218 quantize_per_tensor_default_17"  [label="(1, 40, 14, 14)", style=solid];
+"218 quantize_per_tensor_default_17" -> "219 dequantize_per_tensor_default_22"  [label="(1, 40, 14, 14)", style=solid];
+"219 dequantize_per_tensor_default_22" -> "220 add__1"  [label="(1, 40, 14, 14)", style=solid];
+"220 add__1" -> "221 quantize_per_tensor_default_18"  [label="(1, 40, 14, 14)", style=solid];
+"221 quantize_per_tensor_default_18" -> "222 dequantize_per_tensor_default_24"  [label="(1, 40, 14, 14)", style=solid];
+"221 quantize_per_tensor_default_18" -> "223 dequantize_per_tensor_default_23"  [label="(1, 40, 14, 14)", style=solid];
+"222 dequantize_per_tensor_default_24" -> "268 add__2"  [label="(1, 40, 14, 14)", style=solid];
+"223 dequantize_per_tensor_default_23" -> "230 conv2d_21"  [label="(1, 40, 14, 14)", style=solid];
+"224 features_6_block_0_0_weight_bias_0_0" -> "230 conv2d_21"  [label="(240,)", style=solid];
+"225 scale_updated_constant15" -> "227 mul_tensor_15"  [label="(240, 1, 1, 1)", style=solid];
+"226 compressed_weight_updated_constant15" -> "227 mul_tensor_15"  [label="(240, 40, 1, 1)", style=solid];
+"227 mul_tensor_15" -> "229 sub_tensor_15"  [label="(240, 40, 1, 1)", style=solid];
+"228 zero_point_updated_constant15" -> "229 sub_tensor_15"  [label="(240, 1, 1, 1)", style=solid];
+"229 sub_tensor_15" -> "230 conv2d_21"  [label="(240, 40, 1, 1)", style=solid];
+"230 conv2d_21" -> "233 hardswish__5"  [label="(1, 240, 14, 14)", style=solid];
+"231 hardswish__5_scale_0" -> "234 quantize_per_channel_default_21"  [label="(240,)", style=solid];
+"231 hardswish__5_scale_0" -> "235 dequantize_per_channel_default_21"  [label="(240,)", style=solid];
+"232 hardswish__5_zero_point_0" -> "234 quantize_per_channel_default_21"  [label="(240,)", style=solid];
+"232 hardswish__5_zero_point_0" -> "235 dequantize_per_channel_default_21"  [label="(240,)", style=solid];
+"233 hardswish__5" -> "234 quantize_per_channel_default_21"  [label="(1, 240, 14, 14)", style=solid];
+"234 quantize_per_channel_default_21" -> "235 dequantize_per_channel_default_21"  [label="(1, 240, 14, 14)", style=solid];
+"235 dequantize_per_channel_default_21" -> "242 conv2d_22"  [label="(1, 240, 14, 14)", style=solid];
+"236 features_6_block_1_0_weight_bias_0_0" -> "242 conv2d_22"  [label="(240,)", style=solid];
+"237 scale_updated_constant16" -> "239 mul_tensor_16"  [label="(240, 1, 1, 1)", style=solid];
+"238 compressed_weight_updated_constant16" -> "239 mul_tensor_16"  [label="(240, 1, 5, 5)", style=solid];
+"239 mul_tensor_16" -> "241 sub_tensor_16"  [label="(240, 1, 5, 5)", style=solid];
+"240 zero_point_updated_constant16" -> "241 sub_tensor_16"  [label="(240, 1, 1, 1)", style=solid];
+"241 sub_tensor_16" -> "242 conv2d_22"  [label="(240, 1, 5, 5)", style=solid];
+"242 conv2d_22" -> "243 hardswish__6"  [label="(1, 240, 14, 14)", style=solid];
+"243 hardswish__6" -> "244 quantize_per_tensor_default_19"  [label="(1, 240, 14, 14)", style=solid];
+"244 quantize_per_tensor_default_19" -> "245 dequantize_per_tensor_default_26"  [label="(1, 240, 14, 14)", style=solid];
+"244 quantize_per_tensor_default_19" -> "246 dequantize_per_tensor_default_25"  [label="(1, 240, 14, 14)", style=solid];
+"245 dequantize_per_tensor_default_26" -> "256 mul_3"  [label="(1, 240, 14, 14)", style=solid];
+"246 dequantize_per_tensor_default_25" -> "247 adaptive_avg_pool2d_3"  [label="(1, 240, 14, 14)", style=solid];
+"247 adaptive_avg_pool2d_3" -> "249 conv2d_23"  [label="(1, 240, 1, 1)", style=solid];
+"248 features_6_block_2_fc1_bias_0_0" -> "249 conv2d_23"  [label="(64,)", style=solid];
+"249 conv2d_23" -> "250 relu_3"  [label="(1, 64, 1, 1)", style=solid];
+"250 relu_3" -> "252 conv2d_24"  [label="(1, 64, 1, 1)", style=solid];
+"251 features_6_block_2_fc2_bias_0_0" -> "252 conv2d_24"  [label="(240,)", style=solid];
+"252 conv2d_24" -> "253 hardsigmoid_3"  [label="(1, 240, 1, 1)", style=solid];
+"253 hardsigmoid_3" -> "254 quantize_per_tensor_default_20"  [label="(1, 240, 1, 1)", style=solid];
+"254 quantize_per_tensor_default_20" -> "255 dequantize_per_tensor_default_27"  [label="(1, 240, 1, 1)", style=solid];
+"255 dequantize_per_tensor_default_27" -> "256 mul_3"  [label="(1, 240, 1, 1)", style=solid];
+"256 mul_3" -> "257 quantize_per_tensor_default_21"  [label="(1, 240, 14, 14)", style=solid];
+"257 quantize_per_tensor_default_21" -> "258 dequantize_per_tensor_default_28"  [label="(1, 240, 14, 14)", style=solid];
+"258 dequantize_per_tensor_default_28" -> "265 conv2d_25"  [label="(1, 240, 14, 14)", style=solid];
+"259 features_6_block_3_0_weight_bias_0_0" -> "265 conv2d_25"  [label="(40,)", style=solid];
+"260 scale_updated_constant17" -> "262 mul_tensor_17"  [label="(40, 1, 1, 1)", style=solid];
+"261 compressed_weight_updated_constant17" -> "262 mul_tensor_17"  [label="(40, 240, 1, 1)", style=solid];
+"262 mul_tensor_17" -> "264 sub_tensor_17"  [label="(40, 240, 1, 1)", style=solid];
+"263 zero_point_updated_constant17" -> "264 sub_tensor_17"  [label="(40, 1, 1, 1)", style=solid];
+"264 sub_tensor_17" -> "265 conv2d_25"  [label="(40, 240, 1, 1)", style=solid];
+"265 conv2d_25" -> "266 quantize_per_tensor_default_22"  [label="(1, 40, 14, 14)", style=solid];
+"266 quantize_per_tensor_default_22" -> "267 dequantize_per_tensor_default_29"  [label="(1, 40, 14, 14)", style=solid];
+"267 dequantize_per_tensor_default_29" -> "268 add__2"  [label="(1, 40, 14, 14)", style=solid];
+"268 add__2" -> "269 quantize_per_tensor_default_23"  [label="(1, 40, 14, 14)", style=solid];
+"269 quantize_per_tensor_default_23" -> "270 dequantize_per_tensor_default_30"  [label="(1, 40, 14, 14)", style=solid];
+"270 dequantize_per_tensor_default_30" -> "277 conv2d_26"  [label="(1, 40, 14, 14)", style=solid];
+"271 features_7_block_0_0_weight_bias_0_0" -> "277 conv2d_26"  [label="(120,)", style=solid];
+"272 scale_updated_constant18" -> "274 mul_tensor_18"  [label="(120, 1, 1, 1)", style=solid];
+"273 compressed_weight_updated_constant18" -> "274 mul_tensor_18"  [label="(120, 40, 1, 1)", style=solid];
+"274 mul_tensor_18" -> "276 sub_tensor_18"  [label="(120, 40, 1, 1)", style=solid];
+"275 zero_point_updated_constant18" -> "276 sub_tensor_18"  [label="(120, 1, 1, 1)", style=solid];
+"276 sub_tensor_18" -> "277 conv2d_26"  [label="(120, 40, 1, 1)", style=solid];
+"277 conv2d_26" -> "280 hardswish__7"  [label="(1, 120, 14, 14)", style=solid];
+"278 hardswish__7_scale_0" -> "281 quantize_per_channel_default_25"  [label="(120,)", style=solid];
+"278 hardswish__7_scale_0" -> "282 dequantize_per_channel_default_25"  [label="(120,)", style=solid];
+"279 hardswish__7_zero_point_0" -> "281 quantize_per_channel_default_25"  [label="(120,)", style=solid];
+"279 hardswish__7_zero_point_0" -> "282 dequantize_per_channel_default_25"  [label="(120,)", style=solid];
+"280 hardswish__7" -> "281 quantize_per_channel_default_25"  [label="(1, 120, 14, 14)", style=solid];
+"281 quantize_per_channel_default_25" -> "282 dequantize_per_channel_default_25"  [label="(1, 120, 14, 14)", style=solid];
+"282 dequantize_per_channel_default_25" -> "289 conv2d_27"  [label="(1, 120, 14, 14)", style=solid];
+"283 features_7_block_1_0_weight_bias_0_0" -> "289 conv2d_27"  [label="(120,)", style=solid];
+"284 scale_updated_constant19" -> "286 mul_tensor_19"  [label="(120, 1, 1, 1)", style=solid];
+"285 compressed_weight_updated_constant19" -> "286 mul_tensor_19"  [label="(120, 1, 5, 5)", style=solid];
+"286 mul_tensor_19" -> "288 sub_tensor_19"  [label="(120, 1, 5, 5)", style=solid];
+"287 zero_point_updated_constant19" -> "288 sub_tensor_19"  [label="(120, 1, 1, 1)", style=solid];
+"288 sub_tensor_19" -> "289 conv2d_27"  [label="(120, 1, 5, 5)", style=solid];
+"289 conv2d_27" -> "290 hardswish__8"  [label="(1, 120, 14, 14)", style=solid];
+"290 hardswish__8" -> "291 quantize_per_tensor_default_24"  [label="(1, 120, 14, 14)", style=solid];
+"291 quantize_per_tensor_default_24" -> "292 dequantize_per_tensor_default_32"  [label="(1, 120, 14, 14)", style=solid];
+"291 quantize_per_tensor_default_24" -> "293 dequantize_per_tensor_default_31"  [label="(1, 120, 14, 14)", style=solid];
+"292 dequantize_per_tensor_default_32" -> "303 mul_4"  [label="(1, 120, 14, 14)", style=solid];
+"293 dequantize_per_tensor_default_31" -> "294 adaptive_avg_pool2d_4"  [label="(1, 120, 14, 14)", style=solid];
+"294 adaptive_avg_pool2d_4" -> "296 conv2d_28"  [label="(1, 120, 1, 1)", style=solid];
+"295 features_7_block_2_fc1_bias_0_0" -> "296 conv2d_28"  [label="(32,)", style=solid];
+"296 conv2d_28" -> "297 relu_4"  [label="(1, 32, 1, 1)", style=solid];
+"297 relu_4" -> "299 conv2d_29"  [label="(1, 32, 1, 1)", style=solid];
+"298 features_7_block_2_fc2_bias_0_0" -> "299 conv2d_29"  [label="(120,)", style=solid];
+"299 conv2d_29" -> "300 hardsigmoid_4"  [label="(1, 120, 1, 1)", style=solid];
+"300 hardsigmoid_4" -> "301 quantize_per_tensor_default_25"  [label="(1, 120, 1, 1)", style=solid];
+"301 quantize_per_tensor_default_25" -> "302 dequantize_per_tensor_default_33"  [label="(1, 120, 1, 1)", style=solid];
+"302 dequantize_per_tensor_default_33" -> "303 mul_4"  [label="(1, 120, 1, 1)", style=solid];
+"303 mul_4" -> "304 quantize_per_tensor_default_26"  [label="(1, 120, 14, 14)", style=solid];
+"304 quantize_per_tensor_default_26" -> "305 dequantize_per_tensor_default_34"  [label="(1, 120, 14, 14)", style=solid];
+"305 dequantize_per_tensor_default_34" -> "312 conv2d_30"  [label="(1, 120, 14, 14)", style=solid];
+"306 features_7_block_3_0_weight_bias_0_0" -> "312 conv2d_30"  [label="(48,)", style=solid];
+"307 scale_updated_constant20" -> "309 mul_tensor_20"  [label="(48, 1, 1, 1)", style=solid];
+"308 compressed_weight_updated_constant20" -> "309 mul_tensor_20"  [label="(48, 120, 1, 1)", style=solid];
+"309 mul_tensor_20" -> "311 sub_tensor_20"  [label="(48, 120, 1, 1)", style=solid];
+"310 zero_point_updated_constant20" -> "311 sub_tensor_20"  [label="(48, 1, 1, 1)", style=solid];
+"311 sub_tensor_20" -> "312 conv2d_30"  [label="(48, 120, 1, 1)", style=solid];
+"312 conv2d_30" -> "313 quantize_per_tensor_default_27"  [label="(1, 48, 14, 14)", style=solid];
+"313 quantize_per_tensor_default_27" -> "314 dequantize_per_tensor_default_36"  [label="(1, 48, 14, 14)", style=solid];
+"313 quantize_per_tensor_default_27" -> "315 dequantize_per_tensor_default_35"  [label="(1, 48, 14, 14)", style=solid];
+"314 dequantize_per_tensor_default_36" -> "360 add__3"  [label="(1, 48, 14, 14)", style=solid];
+"315 dequantize_per_tensor_default_35" -> "322 conv2d_31"  [label="(1, 48, 14, 14)", style=solid];
+"316 features_8_block_0_0_weight_bias_0_0" -> "322 conv2d_31"  [label="(144,)", style=solid];
+"317 scale_updated_constant21" -> "319 mul_tensor_21"  [label="(144, 1, 1, 1)", style=solid];
+"318 compressed_weight_updated_constant21" -> "319 mul_tensor_21"  [label="(144, 48, 1, 1)", style=solid];
+"319 mul_tensor_21" -> "321 sub_tensor_21"  [label="(144, 48, 1, 1)", style=solid];
+"320 zero_point_updated_constant21" -> "321 sub_tensor_21"  [label="(144, 1, 1, 1)", style=solid];
+"321 sub_tensor_21" -> "322 conv2d_31"  [label="(144, 48, 1, 1)", style=solid];
+"322 conv2d_31" -> "325 hardswish__9"  [label="(1, 144, 14, 14)", style=solid];
+"323 hardswish__9_scale_0" -> "326 quantize_per_channel_default_29"  [label="(144,)", style=solid];
+"323 hardswish__9_scale_0" -> "327 dequantize_per_channel_default_29"  [label="(144,)", style=solid];
+"324 hardswish__9_zero_point_0" -> "326 quantize_per_channel_default_29"  [label="(144,)", style=solid];
+"324 hardswish__9_zero_point_0" -> "327 dequantize_per_channel_default_29"  [label="(144,)", style=solid];
+"325 hardswish__9" -> "326 quantize_per_channel_default_29"  [label="(1, 144, 14, 14)", style=solid];
+"326 quantize_per_channel_default_29" -> "327 dequantize_per_channel_default_29"  [label="(1, 144, 14, 14)", style=solid];
+"327 dequantize_per_channel_default_29" -> "334 conv2d_32"  [label="(1, 144, 14, 14)", style=solid];
+"328 features_8_block_1_0_weight_bias_0_0" -> "334 conv2d_32"  [label="(144,)", style=solid];
+"329 scale_updated_constant22" -> "331 mul_tensor_22"  [label="(144, 1, 1, 1)", style=solid];
+"330 compressed_weight_updated_constant22" -> "331 mul_tensor_22"  [label="(144, 1, 5, 5)", style=solid];
+"331 mul_tensor_22" -> "333 sub_tensor_22"  [label="(144, 1, 5, 5)", style=solid];
+"332 zero_point_updated_constant22" -> "333 sub_tensor_22"  [label="(144, 1, 1, 1)", style=solid];
+"333 sub_tensor_22" -> "334 conv2d_32"  [label="(144, 1, 5, 5)", style=solid];
+"334 conv2d_32" -> "335 hardswish__10"  [label="(1, 144, 14, 14)", style=solid];
+"335 hardswish__10" -> "336 quantize_per_tensor_default_28"  [label="(1, 144, 14, 14)", style=solid];
+"336 quantize_per_tensor_default_28" -> "337 dequantize_per_tensor_default_38"  [label="(1, 144, 14, 14)", style=solid];
+"336 quantize_per_tensor_default_28" -> "338 dequantize_per_tensor_default_37"  [label="(1, 144, 14, 14)", style=solid];
+"337 dequantize_per_tensor_default_38" -> "348 mul_5"  [label="(1, 144, 14, 14)", style=solid];
+"338 dequantize_per_tensor_default_37" -> "339 adaptive_avg_pool2d_5"  [label="(1, 144, 14, 14)", style=solid];
+"339 adaptive_avg_pool2d_5" -> "341 conv2d_33"  [label="(1, 144, 1, 1)", style=solid];
+"340 features_8_block_2_fc1_bias_0_0" -> "341 conv2d_33"  [label="(40,)", style=solid];
+"341 conv2d_33" -> "342 relu_5"  [label="(1, 40, 1, 1)", style=solid];
+"342 relu_5" -> "344 conv2d_34"  [label="(1, 40, 1, 1)", style=solid];
+"343 features_8_block_2_fc2_bias_0_0" -> "344 conv2d_34"  [label="(144,)", style=solid];
+"344 conv2d_34" -> "345 hardsigmoid_5"  [label="(1, 144, 1, 1)", style=solid];
+"345 hardsigmoid_5" -> "346 quantize_per_tensor_default_29"  [label="(1, 144, 1, 1)", style=solid];
+"346 quantize_per_tensor_default_29" -> "347 dequantize_per_tensor_default_39"  [label="(1, 144, 1, 1)", style=solid];
+"347 dequantize_per_tensor_default_39" -> "348 mul_5"  [label="(1, 144, 1, 1)", style=solid];
+"348 mul_5" -> "349 quantize_per_tensor_default_30"  [label="(1, 144, 14, 14)", style=solid];
+"349 quantize_per_tensor_default_30" -> "350 dequantize_per_tensor_default_40"  [label="(1, 144, 14, 14)", style=solid];
+"350 dequantize_per_tensor_default_40" -> "357 conv2d_35"  [label="(1, 144, 14, 14)", style=solid];
+"351 features_8_block_3_0_weight_bias_0_0" -> "357 conv2d_35"  [label="(48,)", style=solid];
+"352 scale_updated_constant23" -> "354 mul_tensor_23"  [label="(48, 1, 1, 1)", style=solid];
+"353 compressed_weight_updated_constant23" -> "354 mul_tensor_23"  [label="(48, 144, 1, 1)", style=solid];
+"354 mul_tensor_23" -> "356 sub_tensor_23"  [label="(48, 144, 1, 1)", style=solid];
+"355 zero_point_updated_constant23" -> "356 sub_tensor_23"  [label="(48, 1, 1, 1)", style=solid];
+"356 sub_tensor_23" -> "357 conv2d_35"  [label="(48, 144, 1, 1)", style=solid];
+"357 conv2d_35" -> "358 quantize_per_tensor_default_31"  [label="(1, 48, 14, 14)", style=solid];
+"358 quantize_per_tensor_default_31" -> "359 dequantize_per_tensor_default_41"  [label="(1, 48, 14, 14)", style=solid];
+"359 dequantize_per_tensor_default_41" -> "360 add__3"  [label="(1, 48, 14, 14)", style=solid];
+"360 add__3" -> "361 quantize_per_tensor_default_32"  [label="(1, 48, 14, 14)", style=solid];
+"361 quantize_per_tensor_default_32" -> "362 dequantize_per_tensor_default_42"  [label="(1, 48, 14, 14)", style=solid];
+"362 dequantize_per_tensor_default_42" -> "369 conv2d_36"  [label="(1, 48, 14, 14)", style=solid];
+"363 features_9_block_0_0_weight_bias_0_0" -> "369 conv2d_36"  [label="(288,)", style=solid];
+"364 scale_updated_constant24" -> "366 mul_tensor_24"  [label="(288, 1, 1, 1)", style=solid];
+"365 compressed_weight_updated_constant24" -> "366 mul_tensor_24"  [label="(288, 48, 1, 1)", style=solid];
+"366 mul_tensor_24" -> "368 sub_tensor_24"  [label="(288, 48, 1, 1)", style=solid];
+"367 zero_point_updated_constant24" -> "368 sub_tensor_24"  [label="(288, 1, 1, 1)", style=solid];
+"368 sub_tensor_24" -> "369 conv2d_36"  [label="(288, 48, 1, 1)", style=solid];
+"369 conv2d_36" -> "372 hardswish__11"  [label="(1, 288, 14, 14)", style=solid];
+"370 hardswish__11_scale_0" -> "373 quantize_per_channel_default_33"  [label="(288,)", style=solid];
+"370 hardswish__11_scale_0" -> "374 dequantize_per_channel_default_33"  [label="(288,)", style=solid];
+"371 hardswish__11_zero_point_0" -> "373 quantize_per_channel_default_33"  [label="(288,)", style=solid];
+"371 hardswish__11_zero_point_0" -> "374 dequantize_per_channel_default_33"  [label="(288,)", style=solid];
+"372 hardswish__11" -> "373 quantize_per_channel_default_33"  [label="(1, 288, 14, 14)", style=solid];
+"373 quantize_per_channel_default_33" -> "374 dequantize_per_channel_default_33"  [label="(1, 288, 14, 14)", style=solid];
+"374 dequantize_per_channel_default_33" -> "381 conv2d_37"  [label="(1, 288, 14, 14)", style=solid];
+"375 features_9_block_1_0_weight_bias_0_0" -> "381 conv2d_37"  [label="(288,)", style=solid];
+"376 scale_updated_constant25" -> "378 mul_tensor_25"  [label="(288, 1, 1, 1)", style=solid];
+"377 compressed_weight_updated_constant25" -> "378 mul_tensor_25"  [label="(288, 1, 5, 5)", style=solid];
+"378 mul_tensor_25" -> "380 sub_tensor_25"  [label="(288, 1, 5, 5)", style=solid];
+"379 zero_point_updated_constant25" -> "380 sub_tensor_25"  [label="(288, 1, 1, 1)", style=solid];
+"380 sub_tensor_25" -> "381 conv2d_37"  [label="(288, 1, 5, 5)", style=solid];
+"381 conv2d_37" -> "382 hardswish__12"  [label="(1, 288, 7, 7)", style=solid];
+"382 hardswish__12" -> "383 quantize_per_tensor_default_33"  [label="(1, 288, 7, 7)", style=solid];
+"383 quantize_per_tensor_default_33" -> "384 dequantize_per_tensor_default_44"  [label="(1, 288, 7, 7)", style=solid];
+"383 quantize_per_tensor_default_33" -> "385 dequantize_per_tensor_default_43"  [label="(1, 288, 7, 7)", style=solid];
+"384 dequantize_per_tensor_default_44" -> "395 mul_6"  [label="(1, 288, 7, 7)", style=solid];
+"385 dequantize_per_tensor_default_43" -> "386 adaptive_avg_pool2d_6"  [label="(1, 288, 7, 7)", style=solid];
+"386 adaptive_avg_pool2d_6" -> "388 conv2d_38"  [label="(1, 288, 1, 1)", style=solid];
+"387 features_9_block_2_fc1_bias_0_0" -> "388 conv2d_38"  [label="(72,)", style=solid];
+"388 conv2d_38" -> "389 relu_6"  [label="(1, 72, 1, 1)", style=solid];
+"389 relu_6" -> "391 conv2d_39"  [label="(1, 72, 1, 1)", style=solid];
+"390 features_9_block_2_fc2_bias_0_0" -> "391 conv2d_39"  [label="(288,)", style=solid];
+"391 conv2d_39" -> "392 hardsigmoid_6"  [label="(1, 288, 1, 1)", style=solid];
+"392 hardsigmoid_6" -> "393 quantize_per_tensor_default_34"  [label="(1, 288, 1, 1)", style=solid];
+"393 quantize_per_tensor_default_34" -> "394 dequantize_per_tensor_default_45"  [label="(1, 288, 1, 1)", style=solid];
+"394 dequantize_per_tensor_default_45" -> "395 mul_6"  [label="(1, 288, 1, 1)", style=solid];
+"395 mul_6" -> "396 quantize_per_tensor_default_35"  [label="(1, 288, 7, 7)", style=solid];
+"396 quantize_per_tensor_default_35" -> "397 dequantize_per_tensor_default_46"  [label="(1, 288, 7, 7)", style=solid];
+"397 dequantize_per_tensor_default_46" -> "404 conv2d_40"  [label="(1, 288, 7, 7)", style=solid];
+"398 features_9_block_3_0_weight_bias_0_0" -> "404 conv2d_40"  [label="(96,)", style=solid];
+"399 scale_updated_constant26" -> "401 mul_tensor_26"  [label="(96, 1, 1, 1)", style=solid];
+"400 compressed_weight_updated_constant26" -> "401 mul_tensor_26"  [label="(96, 288, 1, 1)", style=solid];
+"401 mul_tensor_26" -> "403 sub_tensor_26"  [label="(96, 288, 1, 1)", style=solid];
+"402 zero_point_updated_constant26" -> "403 sub_tensor_26"  [label="(96, 1, 1, 1)", style=solid];
+"403 sub_tensor_26" -> "404 conv2d_40"  [label="(96, 288, 1, 1)", style=solid];
+"404 conv2d_40" -> "405 quantize_per_tensor_default_36"  [label="(1, 96, 7, 7)", style=solid];
+"405 quantize_per_tensor_default_36" -> "406 dequantize_per_tensor_default_48"  [label="(1, 96, 7, 7)", style=solid];
+"405 quantize_per_tensor_default_36" -> "407 dequantize_per_tensor_default_47"  [label="(1, 96, 7, 7)", style=solid];
+"406 dequantize_per_tensor_default_48" -> "452 add__4"  [label="(1, 96, 7, 7)", style=solid];
+"407 dequantize_per_tensor_default_47" -> "414 conv2d_41"  [label="(1, 96, 7, 7)", style=solid];
+"408 features_10_block_0_0_weight_bias_0_0" -> "414 conv2d_41"  [label="(576,)", style=solid];
+"409 scale_updated_constant27" -> "411 mul_tensor_27"  [label="(576, 1, 1, 1)", style=solid];
+"410 compressed_weight_updated_constant27" -> "411 mul_tensor_27"  [label="(576, 96, 1, 1)", style=solid];
+"411 mul_tensor_27" -> "413 sub_tensor_27"  [label="(576, 96, 1, 1)", style=solid];
+"412 zero_point_updated_constant27" -> "413 sub_tensor_27"  [label="(576, 1, 1, 1)", style=solid];
+"413 sub_tensor_27" -> "414 conv2d_41"  [label="(576, 96, 1, 1)", style=solid];
+"414 conv2d_41" -> "417 hardswish__13"  [label="(1, 576, 7, 7)", style=solid];
+"415 hardswish__13_scale_0" -> "418 quantize_per_channel_default_37"  [label="(576,)", style=solid];
+"415 hardswish__13_scale_0" -> "419 dequantize_per_channel_default_37"  [label="(576,)", style=solid];
+"416 hardswish__13_zero_point_0" -> "418 quantize_per_channel_default_37"  [label="(576,)", style=solid];
+"416 hardswish__13_zero_point_0" -> "419 dequantize_per_channel_default_37"  [label="(576,)", style=solid];
+"417 hardswish__13" -> "418 quantize_per_channel_default_37"  [label="(1, 576, 7, 7)", style=solid];
+"418 quantize_per_channel_default_37" -> "419 dequantize_per_channel_default_37"  [label="(1, 576, 7, 7)", style=solid];
+"419 dequantize_per_channel_default_37" -> "426 conv2d_42"  [label="(1, 576, 7, 7)", style=solid];
+"420 features_10_block_1_0_weight_bias_0_0" -> "426 conv2d_42"  [label="(576,)", style=solid];
+"421 scale_updated_constant28" -> "423 mul_tensor_28"  [label="(576, 1, 1, 1)", style=solid];
+"422 compressed_weight_updated_constant28" -> "423 mul_tensor_28"  [label="(576, 1, 5, 5)", style=solid];
+"423 mul_tensor_28" -> "425 sub_tensor_28"  [label="(576, 1, 5, 5)", style=solid];
+"424 zero_point_updated_constant28" -> "425 sub_tensor_28"  [label="(576, 1, 1, 1)", style=solid];
+"425 sub_tensor_28" -> "426 conv2d_42"  [label="(576, 1, 5, 5)", style=solid];
+"426 conv2d_42" -> "427 hardswish__14"  [label="(1, 576, 7, 7)", style=solid];
+"427 hardswish__14" -> "428 quantize_per_tensor_default_37"  [label="(1, 576, 7, 7)", style=solid];
+"428 quantize_per_tensor_default_37" -> "429 dequantize_per_tensor_default_50"  [label="(1, 576, 7, 7)", style=solid];
+"428 quantize_per_tensor_default_37" -> "430 dequantize_per_tensor_default_49"  [label="(1, 576, 7, 7)", style=solid];
+"429 dequantize_per_tensor_default_50" -> "440 mul_7"  [label="(1, 576, 7, 7)", style=solid];
+"430 dequantize_per_tensor_default_49" -> "431 adaptive_avg_pool2d_7"  [label="(1, 576, 7, 7)", style=solid];
+"431 adaptive_avg_pool2d_7" -> "433 conv2d_43"  [label="(1, 576, 1, 1)", style=solid];
+"432 features_10_block_2_fc1_bias_0_0" -> "433 conv2d_43"  [label="(144,)", style=solid];
+"433 conv2d_43" -> "434 relu_7"  [label="(1, 144, 1, 1)", style=solid];
+"434 relu_7" -> "436 conv2d_44"  [label="(1, 144, 1, 1)", style=solid];
+"435 features_10_block_2_fc2_bias_0_0" -> "436 conv2d_44"  [label="(576,)", style=solid];
+"436 conv2d_44" -> "437 hardsigmoid_7"  [label="(1, 576, 1, 1)", style=solid];
+"437 hardsigmoid_7" -> "438 quantize_per_tensor_default_38"  [label="(1, 576, 1, 1)", style=solid];
+"438 quantize_per_tensor_default_38" -> "439 dequantize_per_tensor_default_51"  [label="(1, 576, 1, 1)", style=solid];
+"439 dequantize_per_tensor_default_51" -> "440 mul_7"  [label="(1, 576, 1, 1)", style=solid];
+"440 mul_7" -> "441 quantize_per_tensor_default_39"  [label="(1, 576, 7, 7)", style=solid];
+"441 quantize_per_tensor_default_39" -> "442 dequantize_per_tensor_default_52"  [label="(1, 576, 7, 7)", style=solid];
+"442 dequantize_per_tensor_default_52" -> "449 conv2d_45"  [label="(1, 576, 7, 7)", style=solid];
+"443 features_10_block_3_0_weight_bias_0_0" -> "449 conv2d_45"  [label="(96,)", style=solid];
+"444 scale_updated_constant29" -> "446 mul_tensor_29"  [label="(96, 1, 1, 1)", style=solid];
+"445 compressed_weight_updated_constant29" -> "446 mul_tensor_29"  [label="(96, 576, 1, 1)", style=solid];
+"446 mul_tensor_29" -> "448 sub_tensor_29"  [label="(96, 576, 1, 1)", style=solid];
+"447 zero_point_updated_constant29" -> "448 sub_tensor_29"  [label="(96, 1, 1, 1)", style=solid];
+"448 sub_tensor_29" -> "449 conv2d_45"  [label="(96, 576, 1, 1)", style=solid];
+"449 conv2d_45" -> "450 quantize_per_tensor_default_40"  [label="(1, 96, 7, 7)", style=solid];
+"450 quantize_per_tensor_default_40" -> "451 dequantize_per_tensor_default_53"  [label="(1, 96, 7, 7)", style=solid];
+"451 dequantize_per_tensor_default_53" -> "452 add__4"  [label="(1, 96, 7, 7)", style=solid];
+"452 add__4" -> "453 quantize_per_tensor_default_41"  [label="(1, 96, 7, 7)", style=solid];
+"453 quantize_per_tensor_default_41" -> "454 dequantize_per_tensor_default_55"  [label="(1, 96, 7, 7)", style=solid];
+"453 quantize_per_tensor_default_41" -> "455 dequantize_per_tensor_default_54"  [label="(1, 96, 7, 7)", style=solid];
+"454 dequantize_per_tensor_default_55" -> "500 add__5"  [label="(1, 96, 7, 7)", style=solid];
+"455 dequantize_per_tensor_default_54" -> "462 conv2d_46"  [label="(1, 96, 7, 7)", style=solid];
+"456 features_11_block_0_0_weight_bias_0_0" -> "462 conv2d_46"  [label="(576,)", style=solid];
+"457 scale_updated_constant30" -> "459 mul_tensor_30"  [label="(576, 1, 1, 1)", style=solid];
+"458 compressed_weight_updated_constant30" -> "459 mul_tensor_30"  [label="(576, 96, 1, 1)", style=solid];
+"459 mul_tensor_30" -> "461 sub_tensor_30"  [label="(576, 96, 1, 1)", style=solid];
+"460 zero_point_updated_constant30" -> "461 sub_tensor_30"  [label="(576, 1, 1, 1)", style=solid];
+"461 sub_tensor_30" -> "462 conv2d_46"  [label="(576, 96, 1, 1)", style=solid];
+"462 conv2d_46" -> "465 hardswish__15"  [label="(1, 576, 7, 7)", style=solid];
+"463 hardswish__15_scale_0" -> "466 quantize_per_channel_default_41"  [label="(576,)", style=solid];
+"463 hardswish__15_scale_0" -> "467 dequantize_per_channel_default_41"  [label="(576,)", style=solid];
+"464 hardswish__15_zero_point_0" -> "466 quantize_per_channel_default_41"  [label="(576,)", style=solid];
+"464 hardswish__15_zero_point_0" -> "467 dequantize_per_channel_default_41"  [label="(576,)", style=solid];
+"465 hardswish__15" -> "466 quantize_per_channel_default_41"  [label="(1, 576, 7, 7)", style=solid];
+"466 quantize_per_channel_default_41" -> "467 dequantize_per_channel_default_41"  [label="(1, 576, 7, 7)", style=solid];
+"467 dequantize_per_channel_default_41" -> "474 conv2d_47"  [label="(1, 576, 7, 7)", style=solid];
+"468 features_11_block_1_0_weight_bias_0_0" -> "474 conv2d_47"  [label="(576,)", style=solid];
+"469 scale_updated_constant31" -> "471 mul_tensor_31"  [label="(576, 1, 1, 1)", style=solid];
+"470 compressed_weight_updated_constant31" -> "471 mul_tensor_31"  [label="(576, 1, 5, 5)", style=solid];
+"471 mul_tensor_31" -> "473 sub_tensor_31"  [label="(576, 1, 5, 5)", style=solid];
+"472 zero_point_updated_constant31" -> "473 sub_tensor_31"  [label="(576, 1, 1, 1)", style=solid];
+"473 sub_tensor_31" -> "474 conv2d_47"  [label="(576, 1, 5, 5)", style=solid];
+"474 conv2d_47" -> "475 hardswish__16"  [label="(1, 576, 7, 7)", style=solid];
+"475 hardswish__16" -> "476 quantize_per_tensor_default_42"  [label="(1, 576, 7, 7)", style=solid];
+"476 quantize_per_tensor_default_42" -> "477 dequantize_per_tensor_default_57"  [label="(1, 576, 7, 7)", style=solid];
+"476 quantize_per_tensor_default_42" -> "478 dequantize_per_tensor_default_56"  [label="(1, 576, 7, 7)", style=solid];
+"477 dequantize_per_tensor_default_57" -> "488 mul_8"  [label="(1, 576, 7, 7)", style=solid];
+"478 dequantize_per_tensor_default_56" -> "479 adaptive_avg_pool2d_8"  [label="(1, 576, 7, 7)", style=solid];
+"479 adaptive_avg_pool2d_8" -> "481 conv2d_48"  [label="(1, 576, 1, 1)", style=solid];
+"480 features_11_block_2_fc1_bias_0_0" -> "481 conv2d_48"  [label="(144,)", style=solid];
+"481 conv2d_48" -> "482 relu_8"  [label="(1, 144, 1, 1)", style=solid];
+"482 relu_8" -> "484 conv2d_49"  [label="(1, 144, 1, 1)", style=solid];
+"483 features_11_block_2_fc2_bias_0_0" -> "484 conv2d_49"  [label="(576,)", style=solid];
 "484 conv2d_49" -> "485 hardsigmoid_8"  [label="(1, 576, 1, 1)", style=solid];
 "485 hardsigmoid_8" -> "486 quantize_per_tensor_default_43"  [label="(1, 576, 1, 1)", style=solid];
 "486 quantize_per_tensor_default_43" -> "487 dequantize_per_tensor_default_58"  [label="(1, 576, 1, 1)", style=solid];
@@ -1089,7 +1089,7 @@ strict digraph  {
 "514 quantize_per_tensor_default_48" -> "515 dequantize_per_tensor_default_63"  [label="(1, 576, 1, 1)", style=solid];
 "515 dequantize_per_tensor_default_63" -> "516 flatten"  [label="(1, 576, 1, 1)", style=solid];
 "516 flatten" -> "523 linear"  [label="(1, 576)", style=solid];
-"517 _param_constant139_0_0" -> "523 linear"  [label="(1024,)", style=solid];
+"517 classifier_0_bias_0_0" -> "523 linear"  [label="(1024,)", style=solid];
 "518 scale_updated_constant34" -> "520 mul_tensor_34"  [label="(1024, 1)", style=solid];
 "519 compressed_weight_updated_constant34" -> "520 mul_tensor_34"  [label="(1024, 576)", style=solid];
 "520 mul_tensor_34" -> "522 sub_tensor_34"  [label="(1024, 576)", style=solid];
@@ -1100,7 +1100,7 @@ strict digraph  {
 "525 quantize_per_tensor_default_49" -> "526 dequantize_per_tensor_default_64"  [label="(1, 1024)", style=solid];
 "526 dequantize_per_tensor_default_64" -> "527 dropout_"  [label="(1, 1024)", style=solid];
 "527 dropout_" -> "534 linear_1"  [label="(1, 1024)", style=solid];
-"528 _param_constant141_0_0" -> "534 linear_1"  [label="(1000,)", style=solid];
+"528 classifier_3_bias_0_0" -> "534 linear_1"  [label="(1000,)", style=solid];
 "529 scale_updated_constant35" -> "531 mul_tensor_35"  [label="(1000, 1)", style=solid];
 "530 compressed_weight_updated_constant35" -> "531 mul_tensor_35"  [label="(1000, 1024)", style=solid];
 "531 mul_tensor_35" -> "533 sub_tensor_35"  [label="(1000, 1024)", style=solid];
diff --git a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/resnet18.dot b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/resnet18.dot
index ea00d5f290e..27ed6dbd832 100644
--- a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/resnet18.dot
+++ b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/resnet18.dot
@@ -1,5 +1,5 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
+"0 x" [id=0, type=input];
 "1 quantize_per_tensor_default" [id=1, type=quantize_per_tensor];
 "2 dequantize_per_tensor_default" [id=2, type=dequantize_per_tensor];
 "3 conv1_weight_bias_0_0" [id=3, type=get_attr];
@@ -235,7 +235,7 @@ strict digraph  {
 "233 quantize_per_tensor_default_29" [id=233, type=quantize_per_tensor];
 "234 dequantize_per_tensor_default_36" [id=234, type=dequantize_per_tensor];
 "235 flatten" [id=235, type=flatten];
-"236 _param_constant61_0_0" [id=236, type=get_attr];
+"236 fc_bias_0_0" [id=236, type=get_attr];
 "237 scale_updated_constant20" [id=237, type=get_attr];
 "238 compressed_weight_updated_constant20" [id=238, type=get_attr];
 "239 mul_tensor_20" [id=239, type=mul];
@@ -243,7 +243,7 @@ strict digraph  {
 "241 sub_tensor_20" [id=241, type=sub];
 "242 linear" [id=242, type=linear];
 "243 output" [id=243, type=output];
-"0 arg0_1" -> "1 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"0 x" -> "1 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
 "1 quantize_per_tensor_default" -> "2 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
 "2 dequantize_per_tensor_default" -> "9 conv2d"  [label="(1, 3, 224, 224)", style=solid];
 "3 conv1_weight_bias_0_0" -> "9 conv2d"  [label="(64,)", style=solid];
@@ -487,7 +487,7 @@ strict digraph  {
 "233 quantize_per_tensor_default_29" -> "234 dequantize_per_tensor_default_36"  [label="(1, 512, 1, 1)", style=solid];
 "234 dequantize_per_tensor_default_36" -> "235 flatten"  [label="(1, 512, 1, 1)", style=solid];
 "235 flatten" -> "242 linear"  [label="(1, 512)", style=solid];
-"236 _param_constant61_0_0" -> "242 linear"  [label="(1000,)", style=solid];
+"236 fc_bias_0_0" -> "242 linear"  [label="(1000,)", style=solid];
 "237 scale_updated_constant20" -> "239 mul_tensor_20"  [label="(1000, 1)", style=solid];
 "238 compressed_weight_updated_constant20" -> "239 mul_tensor_20"  [label="(1000, 512)", style=solid];
 "239 mul_tensor_20" -> "241 sub_tensor_20"  [label="(1000, 512)", style=solid];
diff --git a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/swin_v2_s.dot b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/swin_v2_s.dot
index b29d498e86f..74e0e2c334c 100644
--- a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/swin_v2_s.dot
+++ b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/swin_v2_s.dot
@@ -1,2498 +1,2498 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 quantize_per_tensor_default" [id=1, type=quantize_per_tensor];
-"2 dequantize_per_tensor_default" [id=2, type=dequantize_per_tensor];
-"3 _param_constant1_0_0" [id=3, type=get_attr];
-"4 scale_updated_constant0" [id=4, type=get_attr];
-"5 compressed_weight_updated_constant0" [id=5, type=get_attr];
-"6 mul_tensor" [id=6, type=mul];
-"7 zero_point_updated_constant0" [id=7, type=get_attr];
-"8 sub_tensor" [id=8, type=sub];
-"9 conv2d" [id=9, type=conv2d];
-"10 permute" [id=10, type=permute];
-"11 _param_constant2" [id=11, type=get_attr];
-"12 _param_constant3" [id=12, type=get_attr];
-"13 layer_norm" [id=13, type=layer_norm];
-"14 _frozen_param0" [id=14, type=get_attr];
-"15 pad" [id=15, type=pad];
-"16 view_2" [id=16, type=view];
-"17 permute_2" [id=17, type=permute];
-"18 reshape" [id=18, type=reshape];
-"19 reshape_0_0_nncf_smooth_quant_0" [id=19, type=call_module];
-"20 quantize_per_tensor_default_1" [id=20, type=quantize_per_tensor];
-"21 dequantize_per_tensor_default_1" [id=21, type=dequantize_per_tensor];
-"22 _frozen_param1_0_0" [id=22, type=get_attr];
-"23 scale_updated_constant1" [id=23, type=get_attr];
-"24 compressed_weight_updated_constant1" [id=24, type=get_attr];
-"25 mul_tensor_1" [id=25, type=mul];
-"26 zero_point_updated_constant1" [id=26, type=get_attr];
-"27 sub_tensor_1" [id=27, type=sub];
-"28 linear_2" [id=28, type=linear];
-"29 reshape_1" [id=29, type=reshape];
-"30 permute_3" [id=30, type=permute];
-"31 select" [id=31, type=select];
-"32 select_1" [id=32, type=select];
-"33 select_2" [id=33, type=select];
-"34 linalg_vector_norm" [id=34, type=linalg_vector_norm];
-"35 clamp_min" [id=35, type=clamp_min];
-"36 expand_as" [id=36, type=expand_as];
-"37 div" [id=37, type=div];
-"38 quantize_per_tensor_default_2" [id=38, type=quantize_per_tensor];
-"39 dequantize_per_tensor_default_2" [id=39, type=dequantize_per_tensor];
-"40 linalg_vector_norm_1" [id=40, type=linalg_vector_norm];
-"41 clamp_min_1" [id=41, type=clamp_min];
-"42 expand_as_1" [id=42, type=expand_as];
-"43 div_1" [id=43, type=div];
-"44 quantize_per_tensor_default_3" [id=44, type=quantize_per_tensor];
-"45 dequantize_per_tensor_default_3" [id=45, type=dequantize_per_tensor];
-"46 transpose" [id=46, type=transpose];
-"47 matmul" [id=47, type=matmul];
-"48 _frozen_param3" [id=48, type=get_attr];
-"49 mul_1" [id=49, type=mul];
-"50 add" [id=50, type=add];
-"51 softmax" [id=51, type=softmax];
-"52 dropout" [id=52, type=dropout];
-"53 matmul_1" [id=53, type=matmul];
-"54 transpose_1" [id=54, type=transpose];
-"55 reshape_2" [id=55, type=reshape];
-"56 reshape_2_0_0_nncf_smooth_quant_0" [id=56, type=call_module];
-"57 quantize_per_tensor_default_4" [id=57, type=quantize_per_tensor];
-"58 dequantize_per_tensor_default_4" [id=58, type=dequantize_per_tensor];
-"59 _param_constant11_0_0" [id=59, type=get_attr];
-"60 scale_updated_constant2" [id=60, type=get_attr];
-"61 compressed_weight_updated_constant2" [id=61, type=get_attr];
-"62 mul_tensor_2" [id=62, type=mul];
-"63 zero_point_updated_constant2" [id=63, type=get_attr];
-"64 sub_tensor_2" [id=64, type=sub];
-"65 linear_3" [id=65, type=linear];
-"66 dropout_1" [id=66, type=dropout];
-"67 view_3" [id=67, type=view];
-"68 permute_4" [id=68, type=permute];
-"69 reshape_3" [id=69, type=reshape];
-"70 slice_2" [id=70, type=slice];
-"71 slice_3" [id=71, type=slice];
-"72 _param_constant12" [id=72, type=get_attr];
-"73 _param_constant13" [id=73, type=get_attr];
-"74 layer_norm_1" [id=74, type=layer_norm];
-"75 add_1" [id=75, type=add];
-"76 add_1_0_0_nncf_smooth_quant_0" [id=76, type=call_module];
-"77 quantize_per_tensor_default_5" [id=77, type=quantize_per_tensor];
-"78 dequantize_per_tensor_default_5" [id=78, type=dequantize_per_tensor];
-"79 _param_constant15_0_0" [id=79, type=get_attr];
-"80 scale_updated_constant3" [id=80, type=get_attr];
-"81 compressed_weight_updated_constant3" [id=81, type=get_attr];
-"82 mul_tensor_3" [id=82, type=mul];
-"83 zero_point_updated_constant3" [id=83, type=get_attr];
-"84 sub_tensor_3" [id=84, type=sub];
-"85 linear_4" [id=85, type=linear];
-"86 gelu" [id=86, type=gelu];
-"87 dropout_2" [id=87, type=dropout];
-"88 dropout_2_0_0_nncf_smooth_quant_0" [id=88, type=call_module];
-"89 quantize_per_tensor_default_6" [id=89, type=quantize_per_tensor];
-"90 dequantize_per_tensor_default_6" [id=90, type=dequantize_per_tensor];
-"91 _param_constant17_0_0" [id=91, type=get_attr];
-"92 scale_updated_constant4" [id=92, type=get_attr];
-"93 compressed_weight_updated_constant4" [id=93, type=get_attr];
-"94 mul_tensor_4" [id=94, type=mul];
-"95 zero_point_updated_constant4" [id=95, type=get_attr];
-"96 sub_tensor_4" [id=96, type=sub];
-"97 linear_5" [id=97, type=linear];
-"98 dropout_3" [id=98, type=dropout];
-"99 _param_constant18" [id=99, type=get_attr];
-"100 _param_constant19" [id=100, type=get_attr];
-"101 layer_norm_2" [id=101, type=layer_norm];
-"102 add_2" [id=102, type=add];
-"103 _frozen_param4" [id=103, type=get_attr];
-"104 pad_1" [id=104, type=pad];
-"105 roll" [id=105, type=roll];
-"106 view_6" [id=106, type=view];
-"107 permute_6" [id=107, type=permute];
-"108 reshape_4" [id=108, type=reshape];
-"109 reshape_4_0_0_nncf_smooth_quant_0" [id=109, type=call_module];
-"110 quantize_per_tensor_default_7" [id=110, type=quantize_per_tensor];
-"111 dequantize_per_tensor_default_7" [id=111, type=dequantize_per_tensor];
-"112 _frozen_param5_0_0" [id=112, type=get_attr];
-"113 scale_updated_constant5" [id=113, type=get_attr];
-"114 compressed_weight_updated_constant5" [id=114, type=get_attr];
-"115 mul_tensor_5" [id=115, type=mul];
-"116 zero_point_updated_constant5" [id=116, type=get_attr];
-"117 sub_tensor_5" [id=117, type=sub];
-"118 linear_8" [id=118, type=linear];
-"119 reshape_5" [id=119, type=reshape];
-"120 permute_7" [id=120, type=permute];
-"121 select_3" [id=121, type=select];
-"122 select_4" [id=122, type=select];
-"123 select_5" [id=123, type=select];
-"124 linalg_vector_norm_2" [id=124, type=linalg_vector_norm];
-"125 clamp_min_2" [id=125, type=clamp_min];
-"126 expand_as_2" [id=126, type=expand_as];
-"127 div_2" [id=127, type=div];
-"128 quantize_per_tensor_default_8" [id=128, type=quantize_per_tensor];
-"129 dequantize_per_tensor_default_8" [id=129, type=dequantize_per_tensor];
-"130 linalg_vector_norm_3" [id=130, type=linalg_vector_norm];
-"131 clamp_min_3" [id=131, type=clamp_min];
-"132 expand_as_3" [id=132, type=expand_as];
-"133 div_3" [id=133, type=div];
-"134 quantize_per_tensor_default_9" [id=134, type=quantize_per_tensor];
-"135 dequantize_per_tensor_default_9" [id=135, type=dequantize_per_tensor];
-"136 transpose_2" [id=136, type=transpose];
-"137 matmul_2" [id=137, type=matmul];
-"138 _frozen_param7" [id=138, type=get_attr];
-"139 mul_3" [id=139, type=mul];
-"140 add_3" [id=140, type=add];
-"141 new_zeros" [id=141, type=new_zeros];
-"142 view_7" [id=142, type=view];
-"143 permute_8" [id=143, type=permute];
-"144 reshape_6" [id=144, type=reshape];
-"145 unsqueeze_2" [id=145, type=unsqueeze];
-"146 unsqueeze_3" [id=146, type=unsqueeze];
-"147 sub" [id=147, type=sub];
-"148 ne" [id=148, type=ne];
-"149 masked_fill" [id=149, type=masked_fill];
-"150 eq" [id=150, type=eq];
-"151 masked_fill_1" [id=151, type=masked_fill];
-"152 view_8" [id=152, type=view];
-"153 unsqueeze_4" [id=153, type=unsqueeze];
-"154 unsqueeze_5" [id=154, type=unsqueeze];
-"155 add_4" [id=155, type=add];
-"156 view_9" [id=156, type=view];
-"157 softmax_1" [id=157, type=softmax];
-"158 dropout_4" [id=158, type=dropout];
-"159 matmul_3" [id=159, type=matmul];
-"160 transpose_3" [id=160, type=transpose];
-"161 reshape_7" [id=161, type=reshape];
-"162 reshape_7_0_0_nncf_smooth_quant_0" [id=162, type=call_module];
-"163 quantize_per_tensor_default_10" [id=163, type=quantize_per_tensor];
-"164 dequantize_per_tensor_default_10" [id=164, type=dequantize_per_tensor];
-"165 _param_constant27_0_0" [id=165, type=get_attr];
-"166 scale_updated_constant6" [id=166, type=get_attr];
-"167 compressed_weight_updated_constant6" [id=167, type=get_attr];
-"168 mul_tensor_6" [id=168, type=mul];
-"169 zero_point_updated_constant6" [id=169, type=get_attr];
-"170 sub_tensor_6" [id=170, type=sub];
-"171 linear_9" [id=171, type=linear];
-"172 dropout_5" [id=172, type=dropout];
-"173 view_10" [id=173, type=view];
-"174 permute_9" [id=174, type=permute];
-"175 reshape_8" [id=175, type=reshape];
-"176 roll_1" [id=176, type=roll];
-"177 slice_23" [id=177, type=slice];
-"178 slice_24" [id=178, type=slice];
-"179 _param_constant28" [id=179, type=get_attr];
-"180 _param_constant29" [id=180, type=get_attr];
-"181 layer_norm_3" [id=181, type=layer_norm];
-"182 add_5" [id=182, type=add];
-"183 add_5_0_0_nncf_smooth_quant_0" [id=183, type=call_module];
-"184 quantize_per_tensor_default_11" [id=184, type=quantize_per_tensor];
-"185 dequantize_per_tensor_default_11" [id=185, type=dequantize_per_tensor];
-"186 _param_constant31_0_0" [id=186, type=get_attr];
-"187 scale_updated_constant7" [id=187, type=get_attr];
-"188 compressed_weight_updated_constant7" [id=188, type=get_attr];
-"189 mul_tensor_7" [id=189, type=mul];
-"190 zero_point_updated_constant7" [id=190, type=get_attr];
-"191 sub_tensor_7" [id=191, type=sub];
-"192 linear_10" [id=192, type=linear];
-"193 gelu_1" [id=193, type=gelu];
-"194 dropout_6" [id=194, type=dropout];
-"195 dropout_6_0_0_nncf_smooth_quant_0" [id=195, type=call_module];
-"196 quantize_per_tensor_default_12" [id=196, type=quantize_per_tensor];
-"197 dequantize_per_tensor_default_12" [id=197, type=dequantize_per_tensor];
-"198 _param_constant33_0_0" [id=198, type=get_attr];
-"199 scale_updated_constant8" [id=199, type=get_attr];
-"200 compressed_weight_updated_constant8" [id=200, type=get_attr];
-"201 mul_tensor_8" [id=201, type=mul];
-"202 zero_point_updated_constant8" [id=202, type=get_attr];
-"203 sub_tensor_8" [id=203, type=sub];
-"204 linear_11" [id=204, type=linear];
-"205 dropout_7" [id=205, type=dropout];
-"206 _param_constant34" [id=206, type=get_attr];
-"207 _param_constant35" [id=207, type=get_attr];
-"208 layer_norm_4" [id=208, type=layer_norm];
-"209 add_6" [id=209, type=add];
-"210 pad_2" [id=210, type=pad];
-"211 slice_25" [id=211, type=slice];
-"212 slice_26" [id=212, type=slice];
-"213 slice_27" [id=213, type=slice];
-"214 slice_28" [id=214, type=slice];
-"215 slice_29" [id=215, type=slice];
-"216 slice_30" [id=216, type=slice];
-"217 slice_31" [id=217, type=slice];
-"218 slice_32" [id=218, type=slice];
-"219 slice_33" [id=219, type=slice];
-"220 slice_34" [id=220, type=slice];
-"221 slice_35" [id=221, type=slice];
-"222 slice_36" [id=222, type=slice];
-"223 cat" [id=223, type=cat];
-"224 cat_0_0_nncf_smooth_quant_0" [id=224, type=call_module];
-"225 quantize_per_tensor_default_13" [id=225, type=quantize_per_tensor];
-"226 dequantize_per_tensor_default_13" [id=226, type=dequantize_per_tensor];
-"227 scale_updated_constant9" [id=227, type=get_attr];
-"228 compressed_weight_updated_constant9" [id=228, type=get_attr];
-"229 mul_tensor_9" [id=229, type=mul];
-"230 zero_point_updated_constant9" [id=230, type=get_attr];
-"231 sub_tensor_9" [id=231, type=sub];
-"232 linear_12" [id=232, type=linear];
-"233 _param_constant37" [id=233, type=get_attr];
-"234 _param_constant38" [id=234, type=get_attr];
-"235 layer_norm_5" [id=235, type=layer_norm];
-"236 _frozen_param17" [id=236, type=get_attr];
-"237 pad_3" [id=237, type=pad];
-"238 view_13" [id=238, type=view];
-"239 permute_11" [id=239, type=permute];
-"240 reshape_9" [id=240, type=reshape];
-"241 reshape_9_0_0_nncf_smooth_quant_0" [id=241, type=call_module];
-"242 quantize_per_tensor_default_14" [id=242, type=quantize_per_tensor];
-"243 dequantize_per_tensor_default_14" [id=243, type=dequantize_per_tensor];
-"244 _frozen_param18_0_0" [id=244, type=get_attr];
-"245 scale_updated_constant10" [id=245, type=get_attr];
-"246 compressed_weight_updated_constant10" [id=246, type=get_attr];
-"247 mul_tensor_10" [id=247, type=mul];
-"248 zero_point_updated_constant10" [id=248, type=get_attr];
-"249 sub_tensor_10" [id=249, type=sub];
-"250 linear_15" [id=250, type=linear];
-"251 reshape_10" [id=251, type=reshape];
-"252 permute_12" [id=252, type=permute];
-"253 select_6" [id=253, type=select];
-"254 select_7" [id=254, type=select];
-"255 select_8" [id=255, type=select];
-"256 linalg_vector_norm_4" [id=256, type=linalg_vector_norm];
-"257 clamp_min_4" [id=257, type=clamp_min];
-"258 expand_as_4" [id=258, type=expand_as];
-"259 div_4" [id=259, type=div];
-"260 quantize_per_tensor_default_15" [id=260, type=quantize_per_tensor];
-"261 dequantize_per_tensor_default_15" [id=261, type=dequantize_per_tensor];
-"262 linalg_vector_norm_5" [id=262, type=linalg_vector_norm];
-"263 clamp_min_5" [id=263, type=clamp_min];
-"264 expand_as_5" [id=264, type=expand_as];
-"265 div_5" [id=265, type=div];
-"266 quantize_per_tensor_default_16" [id=266, type=quantize_per_tensor];
-"267 dequantize_per_tensor_default_16" [id=267, type=dequantize_per_tensor];
-"268 transpose_4" [id=268, type=transpose];
-"269 matmul_4" [id=269, type=matmul];
-"270 _frozen_param20" [id=270, type=get_attr];
-"271 mul_5" [id=271, type=mul];
-"272 add_7" [id=272, type=add];
-"273 softmax_2" [id=273, type=softmax];
-"274 dropout_8" [id=274, type=dropout];
-"275 matmul_5" [id=275, type=matmul];
-"276 transpose_5" [id=276, type=transpose];
-"277 reshape_11" [id=277, type=reshape];
-"278 reshape_11_0_0_nncf_smooth_quant_0" [id=278, type=call_module];
-"279 quantize_per_tensor_default_17" [id=279, type=quantize_per_tensor];
-"280 dequantize_per_tensor_default_17" [id=280, type=dequantize_per_tensor];
-"281 _param_constant46_0_0" [id=281, type=get_attr];
-"282 scale_updated_constant11" [id=282, type=get_attr];
-"283 compressed_weight_updated_constant11" [id=283, type=get_attr];
-"284 mul_tensor_11" [id=284, type=mul];
-"285 zero_point_updated_constant11" [id=285, type=get_attr];
-"286 sub_tensor_11" [id=286, type=sub];
-"287 linear_16" [id=287, type=linear];
-"288 dropout_9" [id=288, type=dropout];
-"289 view_14" [id=289, type=view];
-"290 permute_13" [id=290, type=permute];
-"291 reshape_12" [id=291, type=reshape];
-"292 slice_38" [id=292, type=slice];
-"293 slice_39" [id=293, type=slice];
-"294 slice_40" [id=294, type=slice];
-"295 slice_41" [id=295, type=slice];
-"296 contiguous_3" [id=296, type=contiguous];
-"297 _param_constant47" [id=297, type=get_attr];
-"298 _param_constant48" [id=298, type=get_attr];
-"299 layer_norm_6" [id=299, type=layer_norm];
-"300 add_8" [id=300, type=add];
-"301 add_8_0_0_nncf_smooth_quant_0" [id=301, type=call_module];
-"302 quantize_per_tensor_default_18" [id=302, type=quantize_per_tensor];
-"303 dequantize_per_tensor_default_18" [id=303, type=dequantize_per_tensor];
-"304 _param_constant50_0_0" [id=304, type=get_attr];
-"305 scale_updated_constant12" [id=305, type=get_attr];
-"306 compressed_weight_updated_constant12" [id=306, type=get_attr];
-"307 mul_tensor_12" [id=307, type=mul];
-"308 zero_point_updated_constant12" [id=308, type=get_attr];
-"309 sub_tensor_12" [id=309, type=sub];
-"310 linear_17" [id=310, type=linear];
-"311 gelu_2" [id=311, type=gelu];
-"312 dropout_10" [id=312, type=dropout];
-"313 dropout_10_0_0_nncf_smooth_quant_0" [id=313, type=call_module];
-"314 quantize_per_tensor_default_19" [id=314, type=quantize_per_tensor];
-"315 dequantize_per_tensor_default_19" [id=315, type=dequantize_per_tensor];
-"316 _param_constant52_0_0" [id=316, type=get_attr];
-"317 scale_updated_constant13" [id=317, type=get_attr];
-"318 compressed_weight_updated_constant13" [id=318, type=get_attr];
-"319 mul_tensor_13" [id=319, type=mul];
-"320 zero_point_updated_constant13" [id=320, type=get_attr];
-"321 sub_tensor_13" [id=321, type=sub];
-"322 linear_18" [id=322, type=linear];
-"323 dropout_11" [id=323, type=dropout];
-"324 _param_constant53" [id=324, type=get_attr];
-"325 _param_constant54" [id=325, type=get_attr];
-"326 layer_norm_7" [id=326, type=layer_norm];
-"327 add_9" [id=327, type=add];
-"328 _frozen_param21" [id=328, type=get_attr];
-"329 pad_4" [id=329, type=pad];
-"330 roll_2" [id=330, type=roll];
-"331 view_17" [id=331, type=view];
-"332 permute_15" [id=332, type=permute];
-"333 reshape_13" [id=333, type=reshape];
-"334 reshape_13_0_0_nncf_smooth_quant_0" [id=334, type=call_module];
-"335 quantize_per_tensor_default_20" [id=335, type=quantize_per_tensor];
-"336 dequantize_per_tensor_default_20" [id=336, type=dequantize_per_tensor];
-"337 _frozen_param22_0_0" [id=337, type=get_attr];
-"338 scale_updated_constant14" [id=338, type=get_attr];
-"339 compressed_weight_updated_constant14" [id=339, type=get_attr];
-"340 mul_tensor_14" [id=340, type=mul];
-"341 zero_point_updated_constant14" [id=341, type=get_attr];
-"342 sub_tensor_14" [id=342, type=sub];
-"343 linear_21" [id=343, type=linear];
-"344 reshape_14" [id=344, type=reshape];
-"345 permute_16" [id=345, type=permute];
-"346 select_9" [id=346, type=select];
-"347 select_10" [id=347, type=select];
-"348 select_11" [id=348, type=select];
-"349 linalg_vector_norm_6" [id=349, type=linalg_vector_norm];
-"350 clamp_min_6" [id=350, type=clamp_min];
-"351 expand_as_6" [id=351, type=expand_as];
-"352 div_6" [id=352, type=div];
-"353 quantize_per_tensor_default_21" [id=353, type=quantize_per_tensor];
-"354 dequantize_per_tensor_default_21" [id=354, type=dequantize_per_tensor];
-"355 linalg_vector_norm_7" [id=355, type=linalg_vector_norm];
-"356 clamp_min_7" [id=356, type=clamp_min];
-"357 expand_as_7" [id=357, type=expand_as];
-"358 div_7" [id=358, type=div];
-"359 quantize_per_tensor_default_22" [id=359, type=quantize_per_tensor];
-"360 dequantize_per_tensor_default_22" [id=360, type=dequantize_per_tensor];
-"361 transpose_6" [id=361, type=transpose];
-"362 matmul_6" [id=362, type=matmul];
-"363 _frozen_param24" [id=363, type=get_attr];
-"364 mul_7" [id=364, type=mul];
-"365 add_10" [id=365, type=add];
-"366 new_zeros_1" [id=366, type=new_zeros];
-"367 view_18" [id=367, type=view];
-"368 permute_17" [id=368, type=permute];
-"369 reshape_15" [id=369, type=reshape];
-"370 unsqueeze_8" [id=370, type=unsqueeze];
-"371 unsqueeze_9" [id=371, type=unsqueeze];
-"372 sub_1" [id=372, type=sub];
-"373 ne_1" [id=373, type=ne];
-"374 masked_fill_2" [id=374, type=masked_fill];
-"375 eq_1" [id=375, type=eq];
-"376 masked_fill_3" [id=376, type=masked_fill];
-"377 view_19" [id=377, type=view];
-"378 unsqueeze_10" [id=378, type=unsqueeze];
-"379 unsqueeze_11" [id=379, type=unsqueeze];
-"380 add_11" [id=380, type=add];
-"381 view_20" [id=381, type=view];
-"382 softmax_3" [id=382, type=softmax];
-"383 dropout_12" [id=383, type=dropout];
-"384 matmul_7" [id=384, type=matmul];
-"385 transpose_7" [id=385, type=transpose];
-"386 reshape_16" [id=386, type=reshape];
-"387 reshape_16_0_0_nncf_smooth_quant_0" [id=387, type=call_module];
-"388 quantize_per_tensor_default_23" [id=388, type=quantize_per_tensor];
-"389 dequantize_per_tensor_default_23" [id=389, type=dequantize_per_tensor];
-"390 _param_constant62_0_0" [id=390, type=get_attr];
-"391 scale_updated_constant15" [id=391, type=get_attr];
-"392 compressed_weight_updated_constant15" [id=392, type=get_attr];
-"393 mul_tensor_15" [id=393, type=mul];
-"394 zero_point_updated_constant15" [id=394, type=get_attr];
-"395 sub_tensor_15" [id=395, type=sub];
-"396 linear_22" [id=396, type=linear];
-"397 dropout_13" [id=397, type=dropout];
-"398 view_21" [id=398, type=view];
-"399 permute_18" [id=399, type=permute];
-"400 reshape_17" [id=400, type=reshape];
-"401 roll_3" [id=401, type=roll];
-"402 slice_61" [id=402, type=slice];
-"403 slice_62" [id=403, type=slice];
-"404 slice_63" [id=404, type=slice];
-"405 slice_64" [id=405, type=slice];
-"406 contiguous_5" [id=406, type=contiguous];
-"407 _param_constant63" [id=407, type=get_attr];
-"408 _param_constant64" [id=408, type=get_attr];
-"409 layer_norm_8" [id=409, type=layer_norm];
-"410 add_12" [id=410, type=add];
-"411 add_12_0_0_nncf_smooth_quant_0" [id=411, type=call_module];
-"412 quantize_per_tensor_default_24" [id=412, type=quantize_per_tensor];
-"413 dequantize_per_tensor_default_24" [id=413, type=dequantize_per_tensor];
-"414 _param_constant66_0_0" [id=414, type=get_attr];
-"415 scale_updated_constant16" [id=415, type=get_attr];
-"416 compressed_weight_updated_constant16" [id=416, type=get_attr];
-"417 mul_tensor_16" [id=417, type=mul];
-"418 zero_point_updated_constant16" [id=418, type=get_attr];
-"419 sub_tensor_16" [id=419, type=sub];
-"420 linear_23" [id=420, type=linear];
-"421 gelu_3" [id=421, type=gelu];
-"422 dropout_14" [id=422, type=dropout];
-"423 dropout_14_0_0_nncf_smooth_quant_0" [id=423, type=call_module];
-"424 quantize_per_tensor_default_25" [id=424, type=quantize_per_tensor];
-"425 dequantize_per_tensor_default_25" [id=425, type=dequantize_per_tensor];
-"426 _param_constant68_0_0" [id=426, type=get_attr];
-"427 scale_updated_constant17" [id=427, type=get_attr];
-"428 compressed_weight_updated_constant17" [id=428, type=get_attr];
-"429 mul_tensor_17" [id=429, type=mul];
-"430 zero_point_updated_constant17" [id=430, type=get_attr];
-"431 sub_tensor_17" [id=431, type=sub];
-"432 linear_24" [id=432, type=linear];
-"433 dropout_15" [id=433, type=dropout];
-"434 _param_constant69" [id=434, type=get_attr];
-"435 _param_constant70" [id=435, type=get_attr];
-"436 layer_norm_9" [id=436, type=layer_norm];
-"437 add_13" [id=437, type=add];
-"438 pad_5" [id=438, type=pad];
-"439 slice_65" [id=439, type=slice];
-"440 slice_66" [id=440, type=slice];
-"441 slice_67" [id=441, type=slice];
-"442 slice_68" [id=442, type=slice];
-"443 slice_69" [id=443, type=slice];
-"444 slice_70" [id=444, type=slice];
-"445 slice_71" [id=445, type=slice];
-"446 slice_72" [id=446, type=slice];
-"447 slice_73" [id=447, type=slice];
-"448 slice_74" [id=448, type=slice];
-"449 slice_75" [id=449, type=slice];
-"450 slice_76" [id=450, type=slice];
-"451 cat_1" [id=451, type=cat];
-"452 cat_1_0_0_nncf_smooth_quant_0" [id=452, type=call_module];
-"453 quantize_per_tensor_default_26" [id=453, type=quantize_per_tensor];
-"454 dequantize_per_tensor_default_26" [id=454, type=dequantize_per_tensor];
-"455 scale_updated_constant18" [id=455, type=get_attr];
-"456 compressed_weight_updated_constant18" [id=456, type=get_attr];
-"457 mul_tensor_18" [id=457, type=mul];
-"458 zero_point_updated_constant18" [id=458, type=get_attr];
-"459 sub_tensor_18" [id=459, type=sub];
-"460 linear_25" [id=460, type=linear];
-"461 _param_constant72" [id=461, type=get_attr];
-"462 _param_constant73" [id=462, type=get_attr];
-"463 layer_norm_10" [id=463, type=layer_norm];
-"464 _frozen_param34" [id=464, type=get_attr];
-"465 pad_6" [id=465, type=pad];
-"466 view_24" [id=466, type=view];
-"467 permute_20" [id=467, type=permute];
-"468 reshape_18" [id=468, type=reshape];
-"469 reshape_18_0_0_nncf_smooth_quant_0" [id=469, type=call_module];
-"470 quantize_per_tensor_default_27" [id=470, type=quantize_per_tensor];
-"471 dequantize_per_tensor_default_27" [id=471, type=dequantize_per_tensor];
-"472 _frozen_param35_0_0" [id=472, type=get_attr];
-"473 scale_updated_constant19" [id=473, type=get_attr];
-"474 compressed_weight_updated_constant19" [id=474, type=get_attr];
-"475 mul_tensor_19" [id=475, type=mul];
-"476 zero_point_updated_constant19" [id=476, type=get_attr];
-"477 sub_tensor_19" [id=477, type=sub];
-"478 linear_28" [id=478, type=linear];
-"479 reshape_19" [id=479, type=reshape];
-"480 permute_21" [id=480, type=permute];
-"481 select_12" [id=481, type=select];
-"482 select_13" [id=482, type=select];
-"483 select_14" [id=483, type=select];
-"484 linalg_vector_norm_8" [id=484, type=linalg_vector_norm];
-"485 clamp_min_8" [id=485, type=clamp_min];
-"486 expand_as_8" [id=486, type=expand_as];
-"487 div_8" [id=487, type=div];
-"488 quantize_per_tensor_default_28" [id=488, type=quantize_per_tensor];
-"489 dequantize_per_tensor_default_28" [id=489, type=dequantize_per_tensor];
-"490 linalg_vector_norm_9" [id=490, type=linalg_vector_norm];
-"491 clamp_min_9" [id=491, type=clamp_min];
-"492 expand_as_9" [id=492, type=expand_as];
-"493 div_9" [id=493, type=div];
-"494 quantize_per_tensor_default_29" [id=494, type=quantize_per_tensor];
-"495 dequantize_per_tensor_default_29" [id=495, type=dequantize_per_tensor];
-"496 transpose_8" [id=496, type=transpose];
-"497 matmul_8" [id=497, type=matmul];
-"498 _frozen_param37" [id=498, type=get_attr];
-"499 mul_9" [id=499, type=mul];
-"500 add_14" [id=500, type=add];
-"501 softmax_4" [id=501, type=softmax];
-"502 dropout_16" [id=502, type=dropout];
-"503 matmul_9" [id=503, type=matmul];
-"504 transpose_9" [id=504, type=transpose];
-"505 reshape_20" [id=505, type=reshape];
-"506 reshape_20_0_0_nncf_smooth_quant_0" [id=506, type=call_module];
-"507 quantize_per_tensor_default_30" [id=507, type=quantize_per_tensor];
-"508 dequantize_per_tensor_default_30" [id=508, type=dequantize_per_tensor];
-"509 _param_constant81_0_0" [id=509, type=get_attr];
-"510 scale_updated_constant20" [id=510, type=get_attr];
-"511 compressed_weight_updated_constant20" [id=511, type=get_attr];
-"512 mul_tensor_20" [id=512, type=mul];
-"513 zero_point_updated_constant20" [id=513, type=get_attr];
-"514 sub_tensor_20" [id=514, type=sub];
-"515 linear_29" [id=515, type=linear];
-"516 dropout_17" [id=516, type=dropout];
-"517 view_25" [id=517, type=view];
-"518 permute_22" [id=518, type=permute];
-"519 reshape_21" [id=519, type=reshape];
-"520 slice_78" [id=520, type=slice];
-"521 slice_79" [id=521, type=slice];
-"522 slice_80" [id=522, type=slice];
-"523 slice_81" [id=523, type=slice];
-"524 contiguous_7" [id=524, type=contiguous];
-"525 _param_constant82" [id=525, type=get_attr];
-"526 _param_constant83" [id=526, type=get_attr];
-"527 layer_norm_11" [id=527, type=layer_norm];
-"528 add_15" [id=528, type=add];
-"529 add_15_0_0_nncf_smooth_quant_0" [id=529, type=call_module];
-"530 quantize_per_tensor_default_31" [id=530, type=quantize_per_tensor];
-"531 dequantize_per_tensor_default_31" [id=531, type=dequantize_per_tensor];
-"532 _param_constant85_0_0" [id=532, type=get_attr];
-"533 scale_updated_constant21" [id=533, type=get_attr];
-"534 compressed_weight_updated_constant21" [id=534, type=get_attr];
-"535 mul_tensor_21" [id=535, type=mul];
-"536 zero_point_updated_constant21" [id=536, type=get_attr];
-"537 sub_tensor_21" [id=537, type=sub];
-"538 linear_30" [id=538, type=linear];
-"539 gelu_4" [id=539, type=gelu];
-"540 dropout_18" [id=540, type=dropout];
-"541 dropout_18_0_0_nncf_smooth_quant_0" [id=541, type=call_module];
-"542 quantize_per_tensor_default_32" [id=542, type=quantize_per_tensor];
-"543 dequantize_per_tensor_default_32" [id=543, type=dequantize_per_tensor];
-"544 _param_constant87_0_0" [id=544, type=get_attr];
-"545 scale_updated_constant22" [id=545, type=get_attr];
-"546 compressed_weight_updated_constant22" [id=546, type=get_attr];
-"547 mul_tensor_22" [id=547, type=mul];
-"548 zero_point_updated_constant22" [id=548, type=get_attr];
-"549 sub_tensor_22" [id=549, type=sub];
-"550 linear_31" [id=550, type=linear];
-"551 dropout_19" [id=551, type=dropout];
-"552 _param_constant88" [id=552, type=get_attr];
-"553 _param_constant89" [id=553, type=get_attr];
-"554 layer_norm_12" [id=554, type=layer_norm];
-"555 add_16" [id=555, type=add];
-"556 _frozen_param38" [id=556, type=get_attr];
-"557 pad_7" [id=557, type=pad];
-"558 roll_4" [id=558, type=roll];
-"559 view_28" [id=559, type=view];
-"560 permute_24" [id=560, type=permute];
-"561 reshape_22" [id=561, type=reshape];
-"562 reshape_22_0_0_nncf_smooth_quant_0" [id=562, type=call_module];
-"563 quantize_per_tensor_default_33" [id=563, type=quantize_per_tensor];
-"564 dequantize_per_tensor_default_33" [id=564, type=dequantize_per_tensor];
-"565 _frozen_param39_0_0" [id=565, type=get_attr];
-"566 scale_updated_constant23" [id=566, type=get_attr];
-"567 compressed_weight_updated_constant23" [id=567, type=get_attr];
-"568 mul_tensor_23" [id=568, type=mul];
-"569 zero_point_updated_constant23" [id=569, type=get_attr];
-"570 sub_tensor_23" [id=570, type=sub];
-"571 linear_34" [id=571, type=linear];
-"572 reshape_23" [id=572, type=reshape];
-"573 permute_25" [id=573, type=permute];
-"574 select_15" [id=574, type=select];
-"575 select_16" [id=575, type=select];
-"576 select_17" [id=576, type=select];
-"577 linalg_vector_norm_10" [id=577, type=linalg_vector_norm];
-"578 clamp_min_10" [id=578, type=clamp_min];
-"579 expand_as_10" [id=579, type=expand_as];
-"580 div_10" [id=580, type=div];
-"581 quantize_per_tensor_default_34" [id=581, type=quantize_per_tensor];
-"582 dequantize_per_tensor_default_34" [id=582, type=dequantize_per_tensor];
-"583 linalg_vector_norm_11" [id=583, type=linalg_vector_norm];
-"584 clamp_min_11" [id=584, type=clamp_min];
-"585 expand_as_11" [id=585, type=expand_as];
-"586 div_11" [id=586, type=div];
-"587 quantize_per_tensor_default_35" [id=587, type=quantize_per_tensor];
-"588 dequantize_per_tensor_default_35" [id=588, type=dequantize_per_tensor];
-"589 transpose_10" [id=589, type=transpose];
-"590 matmul_10" [id=590, type=matmul];
-"591 _frozen_param41" [id=591, type=get_attr];
-"592 mul_11" [id=592, type=mul];
-"593 add_17" [id=593, type=add];
-"594 new_zeros_2" [id=594, type=new_zeros];
-"595 view_29" [id=595, type=view];
-"596 permute_26" [id=596, type=permute];
-"597 reshape_24" [id=597, type=reshape];
-"598 unsqueeze_14" [id=598, type=unsqueeze];
-"599 unsqueeze_15" [id=599, type=unsqueeze];
-"600 sub_2" [id=600, type=sub];
-"601 ne_2" [id=601, type=ne];
-"602 masked_fill_4" [id=602, type=masked_fill];
-"603 eq_2" [id=603, type=eq];
-"604 masked_fill_5" [id=604, type=masked_fill];
-"605 view_30" [id=605, type=view];
-"606 unsqueeze_16" [id=606, type=unsqueeze];
-"607 unsqueeze_17" [id=607, type=unsqueeze];
-"608 add_18" [id=608, type=add];
-"609 view_31" [id=609, type=view];
-"610 softmax_5" [id=610, type=softmax];
-"611 dropout_20" [id=611, type=dropout];
-"612 matmul_11" [id=612, type=matmul];
-"613 transpose_11" [id=613, type=transpose];
-"614 reshape_25" [id=614, type=reshape];
-"615 reshape_25_0_0_nncf_smooth_quant_0" [id=615, type=call_module];
-"616 quantize_per_tensor_default_36" [id=616, type=quantize_per_tensor];
-"617 dequantize_per_tensor_default_36" [id=617, type=dequantize_per_tensor];
-"618 _param_constant97_0_0" [id=618, type=get_attr];
-"619 scale_updated_constant24" [id=619, type=get_attr];
-"620 compressed_weight_updated_constant24" [id=620, type=get_attr];
-"621 mul_tensor_24" [id=621, type=mul];
-"622 zero_point_updated_constant24" [id=622, type=get_attr];
-"623 sub_tensor_24" [id=623, type=sub];
-"624 linear_35" [id=624, type=linear];
-"625 dropout_21" [id=625, type=dropout];
-"626 view_32" [id=626, type=view];
-"627 permute_27" [id=627, type=permute];
-"628 reshape_26" [id=628, type=reshape];
-"629 roll_5" [id=629, type=roll];
-"630 slice_101" [id=630, type=slice];
-"631 slice_102" [id=631, type=slice];
-"632 slice_103" [id=632, type=slice];
-"633 slice_104" [id=633, type=slice];
-"634 contiguous_9" [id=634, type=contiguous];
-"635 _param_constant98" [id=635, type=get_attr];
-"636 _param_constant99" [id=636, type=get_attr];
-"637 layer_norm_13" [id=637, type=layer_norm];
-"638 add_19" [id=638, type=add];
-"639 add_19_0_0_nncf_smooth_quant_0" [id=639, type=call_module];
-"640 quantize_per_tensor_default_37" [id=640, type=quantize_per_tensor];
-"641 dequantize_per_tensor_default_37" [id=641, type=dequantize_per_tensor];
-"642 _param_constant101_0_0" [id=642, type=get_attr];
-"643 scale_updated_constant25" [id=643, type=get_attr];
-"644 compressed_weight_updated_constant25" [id=644, type=get_attr];
-"645 mul_tensor_25" [id=645, type=mul];
-"646 zero_point_updated_constant25" [id=646, type=get_attr];
-"647 sub_tensor_25" [id=647, type=sub];
-"648 linear_36" [id=648, type=linear];
-"649 gelu_5" [id=649, type=gelu];
-"650 dropout_22" [id=650, type=dropout];
-"651 dropout_22_0_0_nncf_smooth_quant_0" [id=651, type=call_module];
-"652 quantize_per_tensor_default_38" [id=652, type=quantize_per_tensor];
-"653 dequantize_per_tensor_default_38" [id=653, type=dequantize_per_tensor];
-"654 _param_constant103_0_0" [id=654, type=get_attr];
-"655 scale_updated_constant26" [id=655, type=get_attr];
-"656 compressed_weight_updated_constant26" [id=656, type=get_attr];
-"657 mul_tensor_26" [id=657, type=mul];
-"658 zero_point_updated_constant26" [id=658, type=get_attr];
-"659 sub_tensor_26" [id=659, type=sub];
-"660 linear_37" [id=660, type=linear];
-"661 dropout_23" [id=661, type=dropout];
-"662 _param_constant104" [id=662, type=get_attr];
-"663 _param_constant105" [id=663, type=get_attr];
-"664 layer_norm_14" [id=664, type=layer_norm];
-"665 add_20" [id=665, type=add];
-"666 _frozen_param51" [id=666, type=get_attr];
-"667 pad_8" [id=667, type=pad];
-"668 view_35" [id=668, type=view];
-"669 permute_29" [id=669, type=permute];
-"670 reshape_27" [id=670, type=reshape];
-"671 reshape_27_0_0_nncf_smooth_quant_0" [id=671, type=call_module];
-"672 quantize_per_tensor_default_39" [id=672, type=quantize_per_tensor];
-"673 dequantize_per_tensor_default_39" [id=673, type=dequantize_per_tensor];
-"674 _frozen_param52_0_0" [id=674, type=get_attr];
-"675 scale_updated_constant27" [id=675, type=get_attr];
-"676 compressed_weight_updated_constant27" [id=676, type=get_attr];
-"677 mul_tensor_27" [id=677, type=mul];
-"678 zero_point_updated_constant27" [id=678, type=get_attr];
-"679 sub_tensor_27" [id=679, type=sub];
-"680 linear_40" [id=680, type=linear];
-"681 reshape_28" [id=681, type=reshape];
-"682 permute_30" [id=682, type=permute];
-"683 select_18" [id=683, type=select];
-"684 select_19" [id=684, type=select];
-"685 select_20" [id=685, type=select];
-"686 linalg_vector_norm_12" [id=686, type=linalg_vector_norm];
-"687 clamp_min_12" [id=687, type=clamp_min];
-"688 expand_as_12" [id=688, type=expand_as];
-"689 div_12" [id=689, type=div];
-"690 quantize_per_tensor_default_40" [id=690, type=quantize_per_tensor];
-"691 dequantize_per_tensor_default_40" [id=691, type=dequantize_per_tensor];
-"692 linalg_vector_norm_13" [id=692, type=linalg_vector_norm];
-"693 clamp_min_13" [id=693, type=clamp_min];
-"694 expand_as_13" [id=694, type=expand_as];
-"695 div_13" [id=695, type=div];
-"696 quantize_per_tensor_default_41" [id=696, type=quantize_per_tensor];
-"697 dequantize_per_tensor_default_41" [id=697, type=dequantize_per_tensor];
-"698 transpose_12" [id=698, type=transpose];
-"699 matmul_12" [id=699, type=matmul];
-"700 _frozen_param54" [id=700, type=get_attr];
-"701 mul_13" [id=701, type=mul];
-"702 add_21" [id=702, type=add];
-"703 softmax_6" [id=703, type=softmax];
-"704 dropout_24" [id=704, type=dropout];
-"705 matmul_13" [id=705, type=matmul];
-"706 transpose_13" [id=706, type=transpose];
-"707 reshape_29" [id=707, type=reshape];
-"708 reshape_29_0_0_nncf_smooth_quant_0" [id=708, type=call_module];
-"709 quantize_per_tensor_default_42" [id=709, type=quantize_per_tensor];
-"710 dequantize_per_tensor_default_42" [id=710, type=dequantize_per_tensor];
-"711 _param_constant113_0_0" [id=711, type=get_attr];
-"712 scale_updated_constant28" [id=712, type=get_attr];
-"713 compressed_weight_updated_constant28" [id=713, type=get_attr];
-"714 mul_tensor_28" [id=714, type=mul];
-"715 zero_point_updated_constant28" [id=715, type=get_attr];
-"716 sub_tensor_28" [id=716, type=sub];
-"717 linear_41" [id=717, type=linear];
-"718 dropout_25" [id=718, type=dropout];
-"719 view_36" [id=719, type=view];
-"720 permute_31" [id=720, type=permute];
-"721 reshape_30" [id=721, type=reshape];
-"722 slice_106" [id=722, type=slice];
-"723 slice_107" [id=723, type=slice];
-"724 slice_108" [id=724, type=slice];
-"725 slice_109" [id=725, type=slice];
-"726 contiguous_11" [id=726, type=contiguous];
-"727 _param_constant114" [id=727, type=get_attr];
-"728 _param_constant115" [id=728, type=get_attr];
-"729 layer_norm_15" [id=729, type=layer_norm];
-"730 add_22" [id=730, type=add];
-"731 add_22_0_0_nncf_smooth_quant_0" [id=731, type=call_module];
-"732 quantize_per_tensor_default_43" [id=732, type=quantize_per_tensor];
-"733 dequantize_per_tensor_default_43" [id=733, type=dequantize_per_tensor];
-"734 _param_constant117_0_0" [id=734, type=get_attr];
-"735 scale_updated_constant29" [id=735, type=get_attr];
-"736 compressed_weight_updated_constant29" [id=736, type=get_attr];
-"737 mul_tensor_29" [id=737, type=mul];
-"738 zero_point_updated_constant29" [id=738, type=get_attr];
-"739 sub_tensor_29" [id=739, type=sub];
-"740 linear_42" [id=740, type=linear];
-"741 gelu_6" [id=741, type=gelu];
-"742 dropout_26" [id=742, type=dropout];
-"743 dropout_26_0_0_nncf_smooth_quant_0" [id=743, type=call_module];
-"744 quantize_per_tensor_default_44" [id=744, type=quantize_per_tensor];
-"745 dequantize_per_tensor_default_44" [id=745, type=dequantize_per_tensor];
-"746 _param_constant119_0_0" [id=746, type=get_attr];
-"747 scale_updated_constant30" [id=747, type=get_attr];
-"748 compressed_weight_updated_constant30" [id=748, type=get_attr];
-"749 mul_tensor_30" [id=749, type=mul];
-"750 zero_point_updated_constant30" [id=750, type=get_attr];
-"751 sub_tensor_30" [id=751, type=sub];
-"752 linear_43" [id=752, type=linear];
-"753 dropout_27" [id=753, type=dropout];
-"754 _param_constant120" [id=754, type=get_attr];
-"755 _param_constant121" [id=755, type=get_attr];
-"756 layer_norm_16" [id=756, type=layer_norm];
-"757 add_23" [id=757, type=add];
-"758 _frozen_param55" [id=758, type=get_attr];
-"759 pad_9" [id=759, type=pad];
-"760 roll_6" [id=760, type=roll];
-"761 view_39" [id=761, type=view];
-"762 permute_33" [id=762, type=permute];
-"763 reshape_31" [id=763, type=reshape];
-"764 reshape_31_0_0_nncf_smooth_quant_0" [id=764, type=call_module];
-"765 quantize_per_tensor_default_45" [id=765, type=quantize_per_tensor];
-"766 dequantize_per_tensor_default_45" [id=766, type=dequantize_per_tensor];
-"767 _frozen_param56_0_0" [id=767, type=get_attr];
-"768 scale_updated_constant31" [id=768, type=get_attr];
-"769 compressed_weight_updated_constant31" [id=769, type=get_attr];
-"770 mul_tensor_31" [id=770, type=mul];
-"771 zero_point_updated_constant31" [id=771, type=get_attr];
-"772 sub_tensor_31" [id=772, type=sub];
-"773 linear_46" [id=773, type=linear];
-"774 reshape_32" [id=774, type=reshape];
-"775 permute_34" [id=775, type=permute];
-"776 select_21" [id=776, type=select];
-"777 select_22" [id=777, type=select];
-"778 select_23" [id=778, type=select];
-"779 linalg_vector_norm_14" [id=779, type=linalg_vector_norm];
-"780 clamp_min_14" [id=780, type=clamp_min];
-"781 expand_as_14" [id=781, type=expand_as];
-"782 div_14" [id=782, type=div];
-"783 quantize_per_tensor_default_46" [id=783, type=quantize_per_tensor];
-"784 dequantize_per_tensor_default_46" [id=784, type=dequantize_per_tensor];
-"785 linalg_vector_norm_15" [id=785, type=linalg_vector_norm];
-"786 clamp_min_15" [id=786, type=clamp_min];
-"787 expand_as_15" [id=787, type=expand_as];
-"788 div_15" [id=788, type=div];
-"789 quantize_per_tensor_default_47" [id=789, type=quantize_per_tensor];
-"790 dequantize_per_tensor_default_47" [id=790, type=dequantize_per_tensor];
-"791 transpose_14" [id=791, type=transpose];
-"792 matmul_14" [id=792, type=matmul];
-"793 _frozen_param58" [id=793, type=get_attr];
-"794 mul_15" [id=794, type=mul];
-"795 add_24" [id=795, type=add];
-"796 new_zeros_3" [id=796, type=new_zeros];
-"797 view_40" [id=797, type=view];
-"798 permute_35" [id=798, type=permute];
-"799 reshape_33" [id=799, type=reshape];
-"800 unsqueeze_20" [id=800, type=unsqueeze];
-"801 unsqueeze_21" [id=801, type=unsqueeze];
-"802 sub_3" [id=802, type=sub];
-"803 ne_3" [id=803, type=ne];
-"804 masked_fill_6" [id=804, type=masked_fill];
-"805 eq_3" [id=805, type=eq];
-"806 masked_fill_7" [id=806, type=masked_fill];
-"807 view_41" [id=807, type=view];
-"808 unsqueeze_22" [id=808, type=unsqueeze];
-"809 unsqueeze_23" [id=809, type=unsqueeze];
-"810 add_25" [id=810, type=add];
-"811 view_42" [id=811, type=view];
-"812 softmax_7" [id=812, type=softmax];
-"813 dropout_28" [id=813, type=dropout];
-"814 matmul_15" [id=814, type=matmul];
-"815 transpose_15" [id=815, type=transpose];
-"816 reshape_34" [id=816, type=reshape];
-"817 reshape_34_0_0_nncf_smooth_quant_0" [id=817, type=call_module];
-"818 quantize_per_tensor_default_48" [id=818, type=quantize_per_tensor];
-"819 dequantize_per_tensor_default_48" [id=819, type=dequantize_per_tensor];
-"820 _param_constant129_0_0" [id=820, type=get_attr];
-"821 scale_updated_constant32" [id=821, type=get_attr];
-"822 compressed_weight_updated_constant32" [id=822, type=get_attr];
-"823 mul_tensor_32" [id=823, type=mul];
-"824 zero_point_updated_constant32" [id=824, type=get_attr];
-"825 sub_tensor_32" [id=825, type=sub];
-"826 linear_47" [id=826, type=linear];
-"827 dropout_29" [id=827, type=dropout];
-"828 view_43" [id=828, type=view];
-"829 permute_36" [id=829, type=permute];
-"830 reshape_35" [id=830, type=reshape];
-"831 roll_7" [id=831, type=roll];
-"832 slice_129" [id=832, type=slice];
-"833 slice_130" [id=833, type=slice];
-"834 slice_131" [id=834, type=slice];
-"835 slice_132" [id=835, type=slice];
-"836 contiguous_13" [id=836, type=contiguous];
-"837 _param_constant130" [id=837, type=get_attr];
-"838 _param_constant131" [id=838, type=get_attr];
-"839 layer_norm_17" [id=839, type=layer_norm];
-"840 add_26" [id=840, type=add];
-"841 add_26_0_0_nncf_smooth_quant_0" [id=841, type=call_module];
-"842 quantize_per_tensor_default_49" [id=842, type=quantize_per_tensor];
-"843 dequantize_per_tensor_default_49" [id=843, type=dequantize_per_tensor];
-"844 _param_constant133_0_0" [id=844, type=get_attr];
-"845 scale_updated_constant33" [id=845, type=get_attr];
-"846 compressed_weight_updated_constant33" [id=846, type=get_attr];
-"847 mul_tensor_33" [id=847, type=mul];
-"848 zero_point_updated_constant33" [id=848, type=get_attr];
-"849 sub_tensor_33" [id=849, type=sub];
-"850 linear_48" [id=850, type=linear];
-"851 gelu_7" [id=851, type=gelu];
-"852 dropout_30" [id=852, type=dropout];
-"853 dropout_30_0_0_nncf_smooth_quant_0" [id=853, type=call_module];
-"854 quantize_per_tensor_default_50" [id=854, type=quantize_per_tensor];
-"855 dequantize_per_tensor_default_50" [id=855, type=dequantize_per_tensor];
-"856 _param_constant135_0_0" [id=856, type=get_attr];
-"857 scale_updated_constant34" [id=857, type=get_attr];
-"858 compressed_weight_updated_constant34" [id=858, type=get_attr];
-"859 mul_tensor_34" [id=859, type=mul];
-"860 zero_point_updated_constant34" [id=860, type=get_attr];
-"861 sub_tensor_34" [id=861, type=sub];
-"862 linear_49" [id=862, type=linear];
-"863 dropout_31" [id=863, type=dropout];
-"864 _param_constant136" [id=864, type=get_attr];
-"865 _param_constant137" [id=865, type=get_attr];
-"866 layer_norm_18" [id=866, type=layer_norm];
-"867 add_27" [id=867, type=add];
-"868 _frozen_param68" [id=868, type=get_attr];
-"869 pad_10" [id=869, type=pad];
-"870 view_46" [id=870, type=view];
-"871 permute_38" [id=871, type=permute];
-"872 reshape_36" [id=872, type=reshape];
-"873 reshape_36_0_0_nncf_smooth_quant_0" [id=873, type=call_module];
-"874 quantize_per_tensor_default_51" [id=874, type=quantize_per_tensor];
-"875 dequantize_per_tensor_default_51" [id=875, type=dequantize_per_tensor];
-"876 _frozen_param69_0_0" [id=876, type=get_attr];
-"877 scale_updated_constant35" [id=877, type=get_attr];
-"878 compressed_weight_updated_constant35" [id=878, type=get_attr];
-"879 mul_tensor_35" [id=879, type=mul];
-"880 zero_point_updated_constant35" [id=880, type=get_attr];
-"881 sub_tensor_35" [id=881, type=sub];
-"882 linear_52" [id=882, type=linear];
-"883 reshape_37" [id=883, type=reshape];
-"884 permute_39" [id=884, type=permute];
-"885 select_24" [id=885, type=select];
-"886 select_25" [id=886, type=select];
-"887 select_26" [id=887, type=select];
-"888 linalg_vector_norm_16" [id=888, type=linalg_vector_norm];
-"889 clamp_min_16" [id=889, type=clamp_min];
-"890 expand_as_16" [id=890, type=expand_as];
-"891 div_16" [id=891, type=div];
-"892 quantize_per_tensor_default_52" [id=892, type=quantize_per_tensor];
-"893 dequantize_per_tensor_default_52" [id=893, type=dequantize_per_tensor];
-"894 linalg_vector_norm_17" [id=894, type=linalg_vector_norm];
-"895 clamp_min_17" [id=895, type=clamp_min];
-"896 expand_as_17" [id=896, type=expand_as];
-"897 div_17" [id=897, type=div];
-"898 quantize_per_tensor_default_53" [id=898, type=quantize_per_tensor];
-"899 dequantize_per_tensor_default_53" [id=899, type=dequantize_per_tensor];
-"900 transpose_16" [id=900, type=transpose];
-"901 matmul_16" [id=901, type=matmul];
-"902 _frozen_param71" [id=902, type=get_attr];
-"903 mul_17" [id=903, type=mul];
-"904 add_28" [id=904, type=add];
-"905 softmax_8" [id=905, type=softmax];
-"906 dropout_32" [id=906, type=dropout];
-"907 matmul_17" [id=907, type=matmul];
-"908 transpose_17" [id=908, type=transpose];
-"909 reshape_38" [id=909, type=reshape];
-"910 reshape_38_0_0_nncf_smooth_quant_0" [id=910, type=call_module];
-"911 quantize_per_tensor_default_54" [id=911, type=quantize_per_tensor];
-"912 dequantize_per_tensor_default_54" [id=912, type=dequantize_per_tensor];
-"913 _param_constant145_0_0" [id=913, type=get_attr];
-"914 scale_updated_constant36" [id=914, type=get_attr];
-"915 compressed_weight_updated_constant36" [id=915, type=get_attr];
-"916 mul_tensor_36" [id=916, type=mul];
-"917 zero_point_updated_constant36" [id=917, type=get_attr];
-"918 sub_tensor_36" [id=918, type=sub];
-"919 linear_53" [id=919, type=linear];
-"920 dropout_33" [id=920, type=dropout];
-"921 view_47" [id=921, type=view];
-"922 permute_40" [id=922, type=permute];
-"923 reshape_39" [id=923, type=reshape];
-"924 slice_134" [id=924, type=slice];
-"925 slice_135" [id=925, type=slice];
-"926 slice_136" [id=926, type=slice];
-"927 slice_137" [id=927, type=slice];
-"928 contiguous_15" [id=928, type=contiguous];
-"929 _param_constant146" [id=929, type=get_attr];
-"930 _param_constant147" [id=930, type=get_attr];
-"931 layer_norm_19" [id=931, type=layer_norm];
-"932 add_29" [id=932, type=add];
-"933 add_29_0_0_nncf_smooth_quant_0" [id=933, type=call_module];
-"934 quantize_per_tensor_default_55" [id=934, type=quantize_per_tensor];
-"935 dequantize_per_tensor_default_55" [id=935, type=dequantize_per_tensor];
-"936 _param_constant149_0_0" [id=936, type=get_attr];
-"937 scale_updated_constant37" [id=937, type=get_attr];
-"938 compressed_weight_updated_constant37" [id=938, type=get_attr];
-"939 mul_tensor_37" [id=939, type=mul];
-"940 zero_point_updated_constant37" [id=940, type=get_attr];
-"941 sub_tensor_37" [id=941, type=sub];
-"942 linear_54" [id=942, type=linear];
-"943 gelu_8" [id=943, type=gelu];
-"944 dropout_34" [id=944, type=dropout];
-"945 dropout_34_0_0_nncf_smooth_quant_0" [id=945, type=call_module];
-"946 quantize_per_tensor_default_56" [id=946, type=quantize_per_tensor];
-"947 dequantize_per_tensor_default_56" [id=947, type=dequantize_per_tensor];
-"948 _param_constant151_0_0" [id=948, type=get_attr];
-"949 scale_updated_constant38" [id=949, type=get_attr];
-"950 compressed_weight_updated_constant38" [id=950, type=get_attr];
-"951 mul_tensor_38" [id=951, type=mul];
-"952 zero_point_updated_constant38" [id=952, type=get_attr];
-"953 sub_tensor_38" [id=953, type=sub];
-"954 linear_55" [id=954, type=linear];
-"955 dropout_35" [id=955, type=dropout];
-"956 _param_constant152" [id=956, type=get_attr];
-"957 _param_constant153" [id=957, type=get_attr];
-"958 layer_norm_20" [id=958, type=layer_norm];
-"959 add_30" [id=959, type=add];
-"960 _frozen_param72" [id=960, type=get_attr];
-"961 pad_11" [id=961, type=pad];
-"962 roll_8" [id=962, type=roll];
-"963 view_50" [id=963, type=view];
-"964 permute_42" [id=964, type=permute];
-"965 reshape_40" [id=965, type=reshape];
-"966 reshape_40_0_0_nncf_smooth_quant_0" [id=966, type=call_module];
-"967 quantize_per_tensor_default_57" [id=967, type=quantize_per_tensor];
-"968 dequantize_per_tensor_default_57" [id=968, type=dequantize_per_tensor];
-"969 _frozen_param73_0_0" [id=969, type=get_attr];
-"970 scale_updated_constant39" [id=970, type=get_attr];
-"971 compressed_weight_updated_constant39" [id=971, type=get_attr];
-"972 mul_tensor_39" [id=972, type=mul];
-"973 zero_point_updated_constant39" [id=973, type=get_attr];
-"974 sub_tensor_39" [id=974, type=sub];
-"975 linear_58" [id=975, type=linear];
-"976 reshape_41" [id=976, type=reshape];
-"977 permute_43" [id=977, type=permute];
-"978 select_27" [id=978, type=select];
-"979 select_28" [id=979, type=select];
-"980 select_29" [id=980, type=select];
-"981 linalg_vector_norm_18" [id=981, type=linalg_vector_norm];
-"982 clamp_min_18" [id=982, type=clamp_min];
-"983 expand_as_18" [id=983, type=expand_as];
-"984 div_18" [id=984, type=div];
-"985 quantize_per_tensor_default_58" [id=985, type=quantize_per_tensor];
-"986 dequantize_per_tensor_default_58" [id=986, type=dequantize_per_tensor];
-"987 linalg_vector_norm_19" [id=987, type=linalg_vector_norm];
-"988 clamp_min_19" [id=988, type=clamp_min];
-"989 expand_as_19" [id=989, type=expand_as];
-"990 div_19" [id=990, type=div];
-"991 quantize_per_tensor_default_59" [id=991, type=quantize_per_tensor];
-"992 dequantize_per_tensor_default_59" [id=992, type=dequantize_per_tensor];
-"993 transpose_18" [id=993, type=transpose];
-"994 matmul_18" [id=994, type=matmul];
-"995 _frozen_param75" [id=995, type=get_attr];
-"996 mul_19" [id=996, type=mul];
-"997 add_31" [id=997, type=add];
-"998 new_zeros_4" [id=998, type=new_zeros];
-"999 view_51" [id=999, type=view];
-"1000 permute_44" [id=1000, type=permute];
-"1001 reshape_42" [id=1001, type=reshape];
-"1002 unsqueeze_26" [id=1002, type=unsqueeze];
-"1003 unsqueeze_27" [id=1003, type=unsqueeze];
-"1004 sub_4" [id=1004, type=sub];
-"1005 ne_4" [id=1005, type=ne];
-"1006 masked_fill_8" [id=1006, type=masked_fill];
-"1007 eq_4" [id=1007, type=eq];
-"1008 masked_fill_9" [id=1008, type=masked_fill];
-"1009 view_52" [id=1009, type=view];
-"1010 unsqueeze_28" [id=1010, type=unsqueeze];
-"1011 unsqueeze_29" [id=1011, type=unsqueeze];
-"1012 add_32" [id=1012, type=add];
-"1013 view_53" [id=1013, type=view];
-"1014 softmax_9" [id=1014, type=softmax];
-"1015 dropout_36" [id=1015, type=dropout];
-"1016 matmul_19" [id=1016, type=matmul];
-"1017 transpose_19" [id=1017, type=transpose];
-"1018 reshape_43" [id=1018, type=reshape];
-"1019 reshape_43_0_0_nncf_smooth_quant_0" [id=1019, type=call_module];
-"1020 quantize_per_tensor_default_60" [id=1020, type=quantize_per_tensor];
-"1021 dequantize_per_tensor_default_60" [id=1021, type=dequantize_per_tensor];
-"1022 _param_constant161_0_0" [id=1022, type=get_attr];
-"1023 scale_updated_constant40" [id=1023, type=get_attr];
-"1024 compressed_weight_updated_constant40" [id=1024, type=get_attr];
-"1025 mul_tensor_40" [id=1025, type=mul];
-"1026 zero_point_updated_constant40" [id=1026, type=get_attr];
-"1027 sub_tensor_40" [id=1027, type=sub];
-"1028 linear_59" [id=1028, type=linear];
-"1029 dropout_37" [id=1029, type=dropout];
-"1030 view_54" [id=1030, type=view];
-"1031 permute_45" [id=1031, type=permute];
-"1032 reshape_44" [id=1032, type=reshape];
-"1033 roll_9" [id=1033, type=roll];
-"1034 slice_157" [id=1034, type=slice];
-"1035 slice_158" [id=1035, type=slice];
-"1036 slice_159" [id=1036, type=slice];
-"1037 slice_160" [id=1037, type=slice];
-"1038 contiguous_17" [id=1038, type=contiguous];
-"1039 _param_constant162" [id=1039, type=get_attr];
-"1040 _param_constant163" [id=1040, type=get_attr];
-"1041 layer_norm_21" [id=1041, type=layer_norm];
-"1042 add_33" [id=1042, type=add];
-"1043 add_33_0_0_nncf_smooth_quant_0" [id=1043, type=call_module];
-"1044 quantize_per_tensor_default_61" [id=1044, type=quantize_per_tensor];
-"1045 dequantize_per_tensor_default_61" [id=1045, type=dequantize_per_tensor];
-"1046 _param_constant165_0_0" [id=1046, type=get_attr];
-"1047 scale_updated_constant41" [id=1047, type=get_attr];
-"1048 compressed_weight_updated_constant41" [id=1048, type=get_attr];
-"1049 mul_tensor_41" [id=1049, type=mul];
-"1050 zero_point_updated_constant41" [id=1050, type=get_attr];
-"1051 sub_tensor_41" [id=1051, type=sub];
-"1052 linear_60" [id=1052, type=linear];
-"1053 gelu_9" [id=1053, type=gelu];
-"1054 dropout_38" [id=1054, type=dropout];
-"1055 dropout_38_0_0_nncf_smooth_quant_0" [id=1055, type=call_module];
-"1056 quantize_per_tensor_default_62" [id=1056, type=quantize_per_tensor];
-"1057 dequantize_per_tensor_default_62" [id=1057, type=dequantize_per_tensor];
-"1058 _param_constant167_0_0" [id=1058, type=get_attr];
-"1059 scale_updated_constant42" [id=1059, type=get_attr];
-"1060 compressed_weight_updated_constant42" [id=1060, type=get_attr];
-"1061 mul_tensor_42" [id=1061, type=mul];
-"1062 zero_point_updated_constant42" [id=1062, type=get_attr];
-"1063 sub_tensor_42" [id=1063, type=sub];
-"1064 linear_61" [id=1064, type=linear];
-"1065 dropout_39" [id=1065, type=dropout];
-"1066 _param_constant168" [id=1066, type=get_attr];
-"1067 _param_constant169" [id=1067, type=get_attr];
-"1068 layer_norm_22" [id=1068, type=layer_norm];
-"1069 add_34" [id=1069, type=add];
-"1070 _frozen_param85" [id=1070, type=get_attr];
-"1071 pad_12" [id=1071, type=pad];
-"1072 view_57" [id=1072, type=view];
-"1073 permute_47" [id=1073, type=permute];
-"1074 reshape_45" [id=1074, type=reshape];
-"1075 reshape_45_0_0_nncf_smooth_quant_0" [id=1075, type=call_module];
-"1076 quantize_per_tensor_default_63" [id=1076, type=quantize_per_tensor];
-"1077 dequantize_per_tensor_default_63" [id=1077, type=dequantize_per_tensor];
-"1078 _frozen_param86_0_0" [id=1078, type=get_attr];
-"1079 scale_updated_constant43" [id=1079, type=get_attr];
-"1080 compressed_weight_updated_constant43" [id=1080, type=get_attr];
-"1081 mul_tensor_43" [id=1081, type=mul];
-"1082 zero_point_updated_constant43" [id=1082, type=get_attr];
-"1083 sub_tensor_43" [id=1083, type=sub];
-"1084 linear_64" [id=1084, type=linear];
-"1085 reshape_46" [id=1085, type=reshape];
-"1086 permute_48" [id=1086, type=permute];
-"1087 select_30" [id=1087, type=select];
-"1088 select_31" [id=1088, type=select];
-"1089 select_32" [id=1089, type=select];
-"1090 linalg_vector_norm_20" [id=1090, type=linalg_vector_norm];
-"1091 clamp_min_20" [id=1091, type=clamp_min];
-"1092 expand_as_20" [id=1092, type=expand_as];
-"1093 div_20" [id=1093, type=div];
-"1094 quantize_per_tensor_default_64" [id=1094, type=quantize_per_tensor];
-"1095 dequantize_per_tensor_default_64" [id=1095, type=dequantize_per_tensor];
-"1096 linalg_vector_norm_21" [id=1096, type=linalg_vector_norm];
-"1097 clamp_min_21" [id=1097, type=clamp_min];
-"1098 expand_as_21" [id=1098, type=expand_as];
-"1099 div_21" [id=1099, type=div];
-"1100 quantize_per_tensor_default_65" [id=1100, type=quantize_per_tensor];
-"1101 dequantize_per_tensor_default_65" [id=1101, type=dequantize_per_tensor];
-"1102 transpose_20" [id=1102, type=transpose];
-"1103 matmul_20" [id=1103, type=matmul];
-"1104 _frozen_param88" [id=1104, type=get_attr];
-"1105 mul_21" [id=1105, type=mul];
-"1106 add_35" [id=1106, type=add];
-"1107 softmax_10" [id=1107, type=softmax];
-"1108 dropout_40" [id=1108, type=dropout];
-"1109 matmul_21" [id=1109, type=matmul];
-"1110 transpose_21" [id=1110, type=transpose];
-"1111 reshape_47" [id=1111, type=reshape];
-"1112 reshape_47_0_0_nncf_smooth_quant_0" [id=1112, type=call_module];
-"1113 quantize_per_tensor_default_66" [id=1113, type=quantize_per_tensor];
-"1114 dequantize_per_tensor_default_66" [id=1114, type=dequantize_per_tensor];
-"1115 _param_constant177_0_0" [id=1115, type=get_attr];
-"1116 scale_updated_constant44" [id=1116, type=get_attr];
-"1117 compressed_weight_updated_constant44" [id=1117, type=get_attr];
-"1118 mul_tensor_44" [id=1118, type=mul];
-"1119 zero_point_updated_constant44" [id=1119, type=get_attr];
-"1120 sub_tensor_44" [id=1120, type=sub];
-"1121 linear_65" [id=1121, type=linear];
-"1122 dropout_41" [id=1122, type=dropout];
-"1123 view_58" [id=1123, type=view];
-"1124 permute_49" [id=1124, type=permute];
-"1125 reshape_48" [id=1125, type=reshape];
-"1126 slice_162" [id=1126, type=slice];
-"1127 slice_163" [id=1127, type=slice];
-"1128 slice_164" [id=1128, type=slice];
-"1129 slice_165" [id=1129, type=slice];
-"1130 contiguous_19" [id=1130, type=contiguous];
-"1131 _param_constant178" [id=1131, type=get_attr];
-"1132 _param_constant179" [id=1132, type=get_attr];
-"1133 layer_norm_23" [id=1133, type=layer_norm];
-"1134 add_36" [id=1134, type=add];
-"1135 add_36_0_0_nncf_smooth_quant_0" [id=1135, type=call_module];
-"1136 quantize_per_tensor_default_67" [id=1136, type=quantize_per_tensor];
-"1137 dequantize_per_tensor_default_67" [id=1137, type=dequantize_per_tensor];
-"1138 _param_constant181_0_0" [id=1138, type=get_attr];
-"1139 scale_updated_constant45" [id=1139, type=get_attr];
-"1140 compressed_weight_updated_constant45" [id=1140, type=get_attr];
-"1141 mul_tensor_45" [id=1141, type=mul];
-"1142 zero_point_updated_constant45" [id=1142, type=get_attr];
-"1143 sub_tensor_45" [id=1143, type=sub];
-"1144 linear_66" [id=1144, type=linear];
-"1145 gelu_10" [id=1145, type=gelu];
-"1146 dropout_42" [id=1146, type=dropout];
-"1147 dropout_42_0_0_nncf_smooth_quant_0" [id=1147, type=call_module];
-"1148 quantize_per_tensor_default_68" [id=1148, type=quantize_per_tensor];
-"1149 dequantize_per_tensor_default_68" [id=1149, type=dequantize_per_tensor];
-"1150 _param_constant183_0_0" [id=1150, type=get_attr];
-"1151 scale_updated_constant46" [id=1151, type=get_attr];
-"1152 compressed_weight_updated_constant46" [id=1152, type=get_attr];
-"1153 mul_tensor_46" [id=1153, type=mul];
-"1154 zero_point_updated_constant46" [id=1154, type=get_attr];
-"1155 sub_tensor_46" [id=1155, type=sub];
-"1156 linear_67" [id=1156, type=linear];
-"1157 dropout_43" [id=1157, type=dropout];
-"1158 _param_constant184" [id=1158, type=get_attr];
-"1159 _param_constant185" [id=1159, type=get_attr];
-"1160 layer_norm_24" [id=1160, type=layer_norm];
-"1161 add_37" [id=1161, type=add];
-"1162 _frozen_param89" [id=1162, type=get_attr];
-"1163 pad_13" [id=1163, type=pad];
-"1164 roll_10" [id=1164, type=roll];
-"1165 view_61" [id=1165, type=view];
-"1166 permute_51" [id=1166, type=permute];
-"1167 reshape_49" [id=1167, type=reshape];
-"1168 reshape_49_0_0_nncf_smooth_quant_0" [id=1168, type=call_module];
-"1169 quantize_per_tensor_default_69" [id=1169, type=quantize_per_tensor];
-"1170 dequantize_per_tensor_default_69" [id=1170, type=dequantize_per_tensor];
-"1171 _frozen_param90_0_0" [id=1171, type=get_attr];
-"1172 scale_updated_constant47" [id=1172, type=get_attr];
-"1173 compressed_weight_updated_constant47" [id=1173, type=get_attr];
-"1174 mul_tensor_47" [id=1174, type=mul];
-"1175 zero_point_updated_constant47" [id=1175, type=get_attr];
-"1176 sub_tensor_47" [id=1176, type=sub];
-"1177 linear_70" [id=1177, type=linear];
-"1178 reshape_50" [id=1178, type=reshape];
-"1179 permute_52" [id=1179, type=permute];
-"1180 select_33" [id=1180, type=select];
-"1181 select_34" [id=1181, type=select];
-"1182 select_35" [id=1182, type=select];
-"1183 linalg_vector_norm_22" [id=1183, type=linalg_vector_norm];
-"1184 clamp_min_22" [id=1184, type=clamp_min];
-"1185 expand_as_22" [id=1185, type=expand_as];
-"1186 div_22" [id=1186, type=div];
-"1187 quantize_per_tensor_default_70" [id=1187, type=quantize_per_tensor];
-"1188 dequantize_per_tensor_default_70" [id=1188, type=dequantize_per_tensor];
-"1189 linalg_vector_norm_23" [id=1189, type=linalg_vector_norm];
-"1190 clamp_min_23" [id=1190, type=clamp_min];
-"1191 expand_as_23" [id=1191, type=expand_as];
-"1192 div_23" [id=1192, type=div];
-"1193 quantize_per_tensor_default_71" [id=1193, type=quantize_per_tensor];
-"1194 dequantize_per_tensor_default_71" [id=1194, type=dequantize_per_tensor];
-"1195 transpose_22" [id=1195, type=transpose];
-"1196 matmul_22" [id=1196, type=matmul];
-"1197 _frozen_param92" [id=1197, type=get_attr];
-"1198 mul_23" [id=1198, type=mul];
-"1199 add_38" [id=1199, type=add];
-"1200 new_zeros_5" [id=1200, type=new_zeros];
-"1201 view_62" [id=1201, type=view];
-"1202 permute_53" [id=1202, type=permute];
-"1203 reshape_51" [id=1203, type=reshape];
-"1204 unsqueeze_32" [id=1204, type=unsqueeze];
-"1205 unsqueeze_33" [id=1205, type=unsqueeze];
-"1206 sub_5" [id=1206, type=sub];
-"1207 ne_5" [id=1207, type=ne];
-"1208 masked_fill_10" [id=1208, type=masked_fill];
-"1209 eq_5" [id=1209, type=eq];
-"1210 masked_fill_11" [id=1210, type=masked_fill];
-"1211 view_63" [id=1211, type=view];
-"1212 unsqueeze_34" [id=1212, type=unsqueeze];
-"1213 unsqueeze_35" [id=1213, type=unsqueeze];
-"1214 add_39" [id=1214, type=add];
-"1215 view_64" [id=1215, type=view];
-"1216 softmax_11" [id=1216, type=softmax];
-"1217 dropout_44" [id=1217, type=dropout];
-"1218 matmul_23" [id=1218, type=matmul];
-"1219 transpose_23" [id=1219, type=transpose];
-"1220 reshape_52" [id=1220, type=reshape];
-"1221 reshape_52_0_0_nncf_smooth_quant_0" [id=1221, type=call_module];
-"1222 quantize_per_tensor_default_72" [id=1222, type=quantize_per_tensor];
-"1223 dequantize_per_tensor_default_72" [id=1223, type=dequantize_per_tensor];
-"1224 _param_constant193_0_0" [id=1224, type=get_attr];
-"1225 scale_updated_constant48" [id=1225, type=get_attr];
-"1226 compressed_weight_updated_constant48" [id=1226, type=get_attr];
-"1227 mul_tensor_48" [id=1227, type=mul];
-"1228 zero_point_updated_constant48" [id=1228, type=get_attr];
-"1229 sub_tensor_48" [id=1229, type=sub];
-"1230 linear_71" [id=1230, type=linear];
-"1231 dropout_45" [id=1231, type=dropout];
-"1232 view_65" [id=1232, type=view];
-"1233 permute_54" [id=1233, type=permute];
-"1234 reshape_53" [id=1234, type=reshape];
-"1235 roll_11" [id=1235, type=roll];
-"1236 slice_185" [id=1236, type=slice];
-"1237 slice_186" [id=1237, type=slice];
-"1238 slice_187" [id=1238, type=slice];
-"1239 slice_188" [id=1239, type=slice];
-"1240 contiguous_21" [id=1240, type=contiguous];
-"1241 _param_constant194" [id=1241, type=get_attr];
-"1242 _param_constant195" [id=1242, type=get_attr];
-"1243 layer_norm_25" [id=1243, type=layer_norm];
-"1244 add_40" [id=1244, type=add];
-"1245 add_40_0_0_nncf_smooth_quant_0" [id=1245, type=call_module];
-"1246 quantize_per_tensor_default_73" [id=1246, type=quantize_per_tensor];
-"1247 dequantize_per_tensor_default_73" [id=1247, type=dequantize_per_tensor];
-"1248 _param_constant197_0_0" [id=1248, type=get_attr];
-"1249 scale_updated_constant49" [id=1249, type=get_attr];
-"1250 compressed_weight_updated_constant49" [id=1250, type=get_attr];
-"1251 mul_tensor_49" [id=1251, type=mul];
-"1252 zero_point_updated_constant49" [id=1252, type=get_attr];
-"1253 sub_tensor_49" [id=1253, type=sub];
-"1254 linear_72" [id=1254, type=linear];
-"1255 gelu_11" [id=1255, type=gelu];
-"1256 dropout_46" [id=1256, type=dropout];
-"1257 dropout_46_0_0_nncf_smooth_quant_0" [id=1257, type=call_module];
-"1258 quantize_per_tensor_default_74" [id=1258, type=quantize_per_tensor];
-"1259 dequantize_per_tensor_default_74" [id=1259, type=dequantize_per_tensor];
-"1260 _param_constant199_0_0" [id=1260, type=get_attr];
-"1261 scale_updated_constant50" [id=1261, type=get_attr];
-"1262 compressed_weight_updated_constant50" [id=1262, type=get_attr];
-"1263 mul_tensor_50" [id=1263, type=mul];
-"1264 zero_point_updated_constant50" [id=1264, type=get_attr];
-"1265 sub_tensor_50" [id=1265, type=sub];
-"1266 linear_73" [id=1266, type=linear];
-"1267 dropout_47" [id=1267, type=dropout];
-"1268 _param_constant200" [id=1268, type=get_attr];
-"1269 _param_constant201" [id=1269, type=get_attr];
-"1270 layer_norm_26" [id=1270, type=layer_norm];
-"1271 add_41" [id=1271, type=add];
-"1272 _frozen_param102" [id=1272, type=get_attr];
-"1273 pad_14" [id=1273, type=pad];
-"1274 view_68" [id=1274, type=view];
-"1275 permute_56" [id=1275, type=permute];
-"1276 reshape_54" [id=1276, type=reshape];
-"1277 reshape_54_0_0_nncf_smooth_quant_0" [id=1277, type=call_module];
-"1278 quantize_per_tensor_default_75" [id=1278, type=quantize_per_tensor];
-"1279 dequantize_per_tensor_default_75" [id=1279, type=dequantize_per_tensor];
-"1280 _frozen_param103_0_0" [id=1280, type=get_attr];
-"1281 scale_updated_constant51" [id=1281, type=get_attr];
-"1282 compressed_weight_updated_constant51" [id=1282, type=get_attr];
-"1283 mul_tensor_51" [id=1283, type=mul];
-"1284 zero_point_updated_constant51" [id=1284, type=get_attr];
-"1285 sub_tensor_51" [id=1285, type=sub];
-"1286 linear_76" [id=1286, type=linear];
-"1287 reshape_55" [id=1287, type=reshape];
-"1288 permute_57" [id=1288, type=permute];
-"1289 select_36" [id=1289, type=select];
-"1290 select_37" [id=1290, type=select];
-"1291 select_38" [id=1291, type=select];
-"1292 linalg_vector_norm_24" [id=1292, type=linalg_vector_norm];
-"1293 clamp_min_24" [id=1293, type=clamp_min];
-"1294 expand_as_24" [id=1294, type=expand_as];
-"1295 div_24" [id=1295, type=div];
-"1296 quantize_per_tensor_default_76" [id=1296, type=quantize_per_tensor];
-"1297 dequantize_per_tensor_default_76" [id=1297, type=dequantize_per_tensor];
-"1298 linalg_vector_norm_25" [id=1298, type=linalg_vector_norm];
-"1299 clamp_min_25" [id=1299, type=clamp_min];
-"1300 expand_as_25" [id=1300, type=expand_as];
-"1301 div_25" [id=1301, type=div];
-"1302 quantize_per_tensor_default_77" [id=1302, type=quantize_per_tensor];
-"1303 dequantize_per_tensor_default_77" [id=1303, type=dequantize_per_tensor];
-"1304 transpose_24" [id=1304, type=transpose];
-"1305 matmul_24" [id=1305, type=matmul];
-"1306 _frozen_param105" [id=1306, type=get_attr];
-"1307 mul_25" [id=1307, type=mul];
-"1308 add_42" [id=1308, type=add];
-"1309 softmax_12" [id=1309, type=softmax];
-"1310 dropout_48" [id=1310, type=dropout];
-"1311 matmul_25" [id=1311, type=matmul];
-"1312 transpose_25" [id=1312, type=transpose];
-"1313 reshape_56" [id=1313, type=reshape];
-"1314 reshape_56_0_0_nncf_smooth_quant_0" [id=1314, type=call_module];
-"1315 quantize_per_tensor_default_78" [id=1315, type=quantize_per_tensor];
-"1316 dequantize_per_tensor_default_78" [id=1316, type=dequantize_per_tensor];
-"1317 _param_constant209_0_0" [id=1317, type=get_attr];
-"1318 scale_updated_constant52" [id=1318, type=get_attr];
-"1319 compressed_weight_updated_constant52" [id=1319, type=get_attr];
-"1320 mul_tensor_52" [id=1320, type=mul];
-"1321 zero_point_updated_constant52" [id=1321, type=get_attr];
-"1322 sub_tensor_52" [id=1322, type=sub];
-"1323 linear_77" [id=1323, type=linear];
-"1324 dropout_49" [id=1324, type=dropout];
-"1325 view_69" [id=1325, type=view];
-"1326 permute_58" [id=1326, type=permute];
-"1327 reshape_57" [id=1327, type=reshape];
-"1328 slice_190" [id=1328, type=slice];
-"1329 slice_191" [id=1329, type=slice];
-"1330 slice_192" [id=1330, type=slice];
-"1331 slice_193" [id=1331, type=slice];
-"1332 contiguous_23" [id=1332, type=contiguous];
-"1333 _param_constant210" [id=1333, type=get_attr];
-"1334 _param_constant211" [id=1334, type=get_attr];
-"1335 layer_norm_27" [id=1335, type=layer_norm];
-"1336 add_43" [id=1336, type=add];
-"1337 add_43_0_0_nncf_smooth_quant_0" [id=1337, type=call_module];
-"1338 quantize_per_tensor_default_79" [id=1338, type=quantize_per_tensor];
-"1339 dequantize_per_tensor_default_79" [id=1339, type=dequantize_per_tensor];
-"1340 _param_constant213_0_0" [id=1340, type=get_attr];
-"1341 scale_updated_constant53" [id=1341, type=get_attr];
-"1342 compressed_weight_updated_constant53" [id=1342, type=get_attr];
-"1343 mul_tensor_53" [id=1343, type=mul];
-"1344 zero_point_updated_constant53" [id=1344, type=get_attr];
-"1345 sub_tensor_53" [id=1345, type=sub];
-"1346 linear_78" [id=1346, type=linear];
-"1347 gelu_12" [id=1347, type=gelu];
-"1348 dropout_50" [id=1348, type=dropout];
-"1349 dropout_50_0_0_nncf_smooth_quant_0" [id=1349, type=call_module];
-"1350 quantize_per_tensor_default_80" [id=1350, type=quantize_per_tensor];
-"1351 dequantize_per_tensor_default_80" [id=1351, type=dequantize_per_tensor];
-"1352 _param_constant215_0_0" [id=1352, type=get_attr];
-"1353 scale_updated_constant54" [id=1353, type=get_attr];
-"1354 compressed_weight_updated_constant54" [id=1354, type=get_attr];
-"1355 mul_tensor_54" [id=1355, type=mul];
-"1356 zero_point_updated_constant54" [id=1356, type=get_attr];
-"1357 sub_tensor_54" [id=1357, type=sub];
-"1358 linear_79" [id=1358, type=linear];
-"1359 dropout_51" [id=1359, type=dropout];
-"1360 _param_constant216" [id=1360, type=get_attr];
-"1361 _param_constant217" [id=1361, type=get_attr];
-"1362 layer_norm_28" [id=1362, type=layer_norm];
-"1363 add_44" [id=1363, type=add];
-"1364 _frozen_param106" [id=1364, type=get_attr];
-"1365 pad_15" [id=1365, type=pad];
-"1366 roll_12" [id=1366, type=roll];
-"1367 view_72" [id=1367, type=view];
-"1368 permute_60" [id=1368, type=permute];
-"1369 reshape_58" [id=1369, type=reshape];
-"1370 reshape_58_0_0_nncf_smooth_quant_0" [id=1370, type=call_module];
-"1371 quantize_per_tensor_default_81" [id=1371, type=quantize_per_tensor];
-"1372 dequantize_per_tensor_default_81" [id=1372, type=dequantize_per_tensor];
-"1373 _frozen_param107_0_0" [id=1373, type=get_attr];
-"1374 scale_updated_constant55" [id=1374, type=get_attr];
-"1375 compressed_weight_updated_constant55" [id=1375, type=get_attr];
-"1376 mul_tensor_55" [id=1376, type=mul];
-"1377 zero_point_updated_constant55" [id=1377, type=get_attr];
-"1378 sub_tensor_55" [id=1378, type=sub];
-"1379 linear_82" [id=1379, type=linear];
-"1380 reshape_59" [id=1380, type=reshape];
-"1381 permute_61" [id=1381, type=permute];
-"1382 select_39" [id=1382, type=select];
-"1383 select_40" [id=1383, type=select];
-"1384 select_41" [id=1384, type=select];
-"1385 linalg_vector_norm_26" [id=1385, type=linalg_vector_norm];
-"1386 clamp_min_26" [id=1386, type=clamp_min];
-"1387 expand_as_26" [id=1387, type=expand_as];
-"1388 div_26" [id=1388, type=div];
-"1389 quantize_per_tensor_default_82" [id=1389, type=quantize_per_tensor];
-"1390 dequantize_per_tensor_default_82" [id=1390, type=dequantize_per_tensor];
-"1391 linalg_vector_norm_27" [id=1391, type=linalg_vector_norm];
-"1392 clamp_min_27" [id=1392, type=clamp_min];
-"1393 expand_as_27" [id=1393, type=expand_as];
-"1394 div_27" [id=1394, type=div];
-"1395 quantize_per_tensor_default_83" [id=1395, type=quantize_per_tensor];
-"1396 dequantize_per_tensor_default_83" [id=1396, type=dequantize_per_tensor];
-"1397 transpose_26" [id=1397, type=transpose];
-"1398 matmul_26" [id=1398, type=matmul];
-"1399 _frozen_param109" [id=1399, type=get_attr];
-"1400 mul_27" [id=1400, type=mul];
-"1401 add_45" [id=1401, type=add];
-"1402 new_zeros_6" [id=1402, type=new_zeros];
-"1403 view_73" [id=1403, type=view];
-"1404 permute_62" [id=1404, type=permute];
-"1405 reshape_60" [id=1405, type=reshape];
-"1406 unsqueeze_38" [id=1406, type=unsqueeze];
-"1407 unsqueeze_39" [id=1407, type=unsqueeze];
-"1408 sub_6" [id=1408, type=sub];
-"1409 ne_6" [id=1409, type=ne];
-"1410 masked_fill_12" [id=1410, type=masked_fill];
-"1411 eq_6" [id=1411, type=eq];
-"1412 masked_fill_13" [id=1412, type=masked_fill];
-"1413 view_74" [id=1413, type=view];
-"1414 unsqueeze_40" [id=1414, type=unsqueeze];
-"1415 unsqueeze_41" [id=1415, type=unsqueeze];
-"1416 add_46" [id=1416, type=add];
-"1417 view_75" [id=1417, type=view];
-"1418 softmax_13" [id=1418, type=softmax];
-"1419 dropout_52" [id=1419, type=dropout];
-"1420 matmul_27" [id=1420, type=matmul];
-"1421 transpose_27" [id=1421, type=transpose];
-"1422 reshape_61" [id=1422, type=reshape];
-"1423 reshape_61_0_0_nncf_smooth_quant_0" [id=1423, type=call_module];
-"1424 quantize_per_tensor_default_84" [id=1424, type=quantize_per_tensor];
-"1425 dequantize_per_tensor_default_84" [id=1425, type=dequantize_per_tensor];
-"1426 _param_constant225_0_0" [id=1426, type=get_attr];
-"1427 scale_updated_constant56" [id=1427, type=get_attr];
-"1428 compressed_weight_updated_constant56" [id=1428, type=get_attr];
-"1429 mul_tensor_56" [id=1429, type=mul];
-"1430 zero_point_updated_constant56" [id=1430, type=get_attr];
-"1431 sub_tensor_56" [id=1431, type=sub];
-"1432 linear_83" [id=1432, type=linear];
-"1433 dropout_53" [id=1433, type=dropout];
-"1434 view_76" [id=1434, type=view];
-"1435 permute_63" [id=1435, type=permute];
-"1436 reshape_62" [id=1436, type=reshape];
-"1437 roll_13" [id=1437, type=roll];
-"1438 slice_213" [id=1438, type=slice];
-"1439 slice_214" [id=1439, type=slice];
-"1440 slice_215" [id=1440, type=slice];
-"1441 slice_216" [id=1441, type=slice];
-"1442 contiguous_25" [id=1442, type=contiguous];
-"1443 _param_constant226" [id=1443, type=get_attr];
-"1444 _param_constant227" [id=1444, type=get_attr];
-"1445 layer_norm_29" [id=1445, type=layer_norm];
-"1446 add_47" [id=1446, type=add];
-"1447 add_47_0_0_nncf_smooth_quant_0" [id=1447, type=call_module];
-"1448 quantize_per_tensor_default_85" [id=1448, type=quantize_per_tensor];
-"1449 dequantize_per_tensor_default_85" [id=1449, type=dequantize_per_tensor];
-"1450 _param_constant229_0_0" [id=1450, type=get_attr];
-"1451 scale_updated_constant57" [id=1451, type=get_attr];
-"1452 compressed_weight_updated_constant57" [id=1452, type=get_attr];
-"1453 mul_tensor_57" [id=1453, type=mul];
-"1454 zero_point_updated_constant57" [id=1454, type=get_attr];
-"1455 sub_tensor_57" [id=1455, type=sub];
-"1456 linear_84" [id=1456, type=linear];
-"1457 gelu_13" [id=1457, type=gelu];
-"1458 dropout_54" [id=1458, type=dropout];
-"1459 dropout_54_0_0_nncf_smooth_quant_0" [id=1459, type=call_module];
-"1460 quantize_per_tensor_default_86" [id=1460, type=quantize_per_tensor];
-"1461 dequantize_per_tensor_default_86" [id=1461, type=dequantize_per_tensor];
-"1462 _param_constant231_0_0" [id=1462, type=get_attr];
-"1463 scale_updated_constant58" [id=1463, type=get_attr];
-"1464 compressed_weight_updated_constant58" [id=1464, type=get_attr];
-"1465 mul_tensor_58" [id=1465, type=mul];
-"1466 zero_point_updated_constant58" [id=1466, type=get_attr];
-"1467 sub_tensor_58" [id=1467, type=sub];
-"1468 linear_85" [id=1468, type=linear];
-"1469 dropout_55" [id=1469, type=dropout];
-"1470 _param_constant232" [id=1470, type=get_attr];
-"1471 _param_constant233" [id=1471, type=get_attr];
-"1472 layer_norm_30" [id=1472, type=layer_norm];
-"1473 add_48" [id=1473, type=add];
-"1474 _frozen_param119" [id=1474, type=get_attr];
-"1475 pad_16" [id=1475, type=pad];
-"1476 view_79" [id=1476, type=view];
-"1477 permute_65" [id=1477, type=permute];
-"1478 reshape_63" [id=1478, type=reshape];
-"1479 reshape_63_0_0_nncf_smooth_quant_0" [id=1479, type=call_module];
-"1480 quantize_per_tensor_default_87" [id=1480, type=quantize_per_tensor];
-"1481 dequantize_per_tensor_default_87" [id=1481, type=dequantize_per_tensor];
-"1482 _frozen_param120_0_0" [id=1482, type=get_attr];
-"1483 scale_updated_constant59" [id=1483, type=get_attr];
-"1484 compressed_weight_updated_constant59" [id=1484, type=get_attr];
-"1485 mul_tensor_59" [id=1485, type=mul];
-"1486 zero_point_updated_constant59" [id=1486, type=get_attr];
-"1487 sub_tensor_59" [id=1487, type=sub];
-"1488 linear_88" [id=1488, type=linear];
-"1489 reshape_64" [id=1489, type=reshape];
-"1490 permute_66" [id=1490, type=permute];
-"1491 select_42" [id=1491, type=select];
-"1492 select_43" [id=1492, type=select];
-"1493 select_44" [id=1493, type=select];
-"1494 linalg_vector_norm_28" [id=1494, type=linalg_vector_norm];
-"1495 clamp_min_28" [id=1495, type=clamp_min];
-"1496 expand_as_28" [id=1496, type=expand_as];
-"1497 div_28" [id=1497, type=div];
-"1498 quantize_per_tensor_default_88" [id=1498, type=quantize_per_tensor];
-"1499 dequantize_per_tensor_default_88" [id=1499, type=dequantize_per_tensor];
-"1500 linalg_vector_norm_29" [id=1500, type=linalg_vector_norm];
-"1501 clamp_min_29" [id=1501, type=clamp_min];
-"1502 expand_as_29" [id=1502, type=expand_as];
-"1503 div_29" [id=1503, type=div];
-"1504 quantize_per_tensor_default_89" [id=1504, type=quantize_per_tensor];
-"1505 dequantize_per_tensor_default_89" [id=1505, type=dequantize_per_tensor];
-"1506 transpose_28" [id=1506, type=transpose];
-"1507 matmul_28" [id=1507, type=matmul];
-"1508 _frozen_param122" [id=1508, type=get_attr];
-"1509 mul_29" [id=1509, type=mul];
-"1510 add_49" [id=1510, type=add];
-"1511 softmax_14" [id=1511, type=softmax];
-"1512 dropout_56" [id=1512, type=dropout];
-"1513 matmul_29" [id=1513, type=matmul];
-"1514 transpose_29" [id=1514, type=transpose];
-"1515 reshape_65" [id=1515, type=reshape];
-"1516 reshape_65_0_0_nncf_smooth_quant_0" [id=1516, type=call_module];
-"1517 quantize_per_tensor_default_90" [id=1517, type=quantize_per_tensor];
-"1518 dequantize_per_tensor_default_90" [id=1518, type=dequantize_per_tensor];
-"1519 _param_constant241_0_0" [id=1519, type=get_attr];
-"1520 scale_updated_constant60" [id=1520, type=get_attr];
-"1521 compressed_weight_updated_constant60" [id=1521, type=get_attr];
-"1522 mul_tensor_60" [id=1522, type=mul];
-"1523 zero_point_updated_constant60" [id=1523, type=get_attr];
-"1524 sub_tensor_60" [id=1524, type=sub];
-"1525 linear_89" [id=1525, type=linear];
-"1526 dropout_57" [id=1526, type=dropout];
-"1527 view_80" [id=1527, type=view];
-"1528 permute_67" [id=1528, type=permute];
-"1529 reshape_66" [id=1529, type=reshape];
-"1530 slice_218" [id=1530, type=slice];
-"1531 slice_219" [id=1531, type=slice];
-"1532 slice_220" [id=1532, type=slice];
-"1533 slice_221" [id=1533, type=slice];
-"1534 contiguous_27" [id=1534, type=contiguous];
-"1535 _param_constant242" [id=1535, type=get_attr];
-"1536 _param_constant243" [id=1536, type=get_attr];
-"1537 layer_norm_31" [id=1537, type=layer_norm];
-"1538 add_50" [id=1538, type=add];
-"1539 add_50_0_0_nncf_smooth_quant_0" [id=1539, type=call_module];
-"1540 quantize_per_tensor_default_91" [id=1540, type=quantize_per_tensor];
-"1541 dequantize_per_tensor_default_91" [id=1541, type=dequantize_per_tensor];
-"1542 _param_constant245_0_0" [id=1542, type=get_attr];
-"1543 scale_updated_constant61" [id=1543, type=get_attr];
-"1544 compressed_weight_updated_constant61" [id=1544, type=get_attr];
-"1545 mul_tensor_61" [id=1545, type=mul];
-"1546 zero_point_updated_constant61" [id=1546, type=get_attr];
-"1547 sub_tensor_61" [id=1547, type=sub];
-"1548 linear_90" [id=1548, type=linear];
-"1549 gelu_14" [id=1549, type=gelu];
-"1550 dropout_58" [id=1550, type=dropout];
-"1551 dropout_58_0_0_nncf_smooth_quant_0" [id=1551, type=call_module];
-"1552 quantize_per_tensor_default_92" [id=1552, type=quantize_per_tensor];
-"1553 dequantize_per_tensor_default_92" [id=1553, type=dequantize_per_tensor];
-"1554 _param_constant247_0_0" [id=1554, type=get_attr];
-"1555 scale_updated_constant62" [id=1555, type=get_attr];
-"1556 compressed_weight_updated_constant62" [id=1556, type=get_attr];
-"1557 mul_tensor_62" [id=1557, type=mul];
-"1558 zero_point_updated_constant62" [id=1558, type=get_attr];
-"1559 sub_tensor_62" [id=1559, type=sub];
-"1560 linear_91" [id=1560, type=linear];
-"1561 dropout_59" [id=1561, type=dropout];
-"1562 _param_constant248" [id=1562, type=get_attr];
-"1563 _param_constant249" [id=1563, type=get_attr];
-"1564 layer_norm_32" [id=1564, type=layer_norm];
-"1565 add_51" [id=1565, type=add];
-"1566 _frozen_param123" [id=1566, type=get_attr];
-"1567 pad_17" [id=1567, type=pad];
-"1568 roll_14" [id=1568, type=roll];
-"1569 view_83" [id=1569, type=view];
-"1570 permute_69" [id=1570, type=permute];
-"1571 reshape_67" [id=1571, type=reshape];
-"1572 reshape_67_0_0_nncf_smooth_quant_0" [id=1572, type=call_module];
-"1573 quantize_per_tensor_default_93" [id=1573, type=quantize_per_tensor];
-"1574 dequantize_per_tensor_default_93" [id=1574, type=dequantize_per_tensor];
-"1575 _frozen_param124_0_0" [id=1575, type=get_attr];
-"1576 scale_updated_constant63" [id=1576, type=get_attr];
-"1577 compressed_weight_updated_constant63" [id=1577, type=get_attr];
-"1578 mul_tensor_63" [id=1578, type=mul];
-"1579 zero_point_updated_constant63" [id=1579, type=get_attr];
-"1580 sub_tensor_63" [id=1580, type=sub];
-"1581 linear_94" [id=1581, type=linear];
-"1582 reshape_68" [id=1582, type=reshape];
-"1583 permute_70" [id=1583, type=permute];
-"1584 select_45" [id=1584, type=select];
-"1585 select_46" [id=1585, type=select];
-"1586 select_47" [id=1586, type=select];
-"1587 linalg_vector_norm_30" [id=1587, type=linalg_vector_norm];
-"1588 clamp_min_30" [id=1588, type=clamp_min];
-"1589 expand_as_30" [id=1589, type=expand_as];
-"1590 div_30" [id=1590, type=div];
-"1591 quantize_per_tensor_default_94" [id=1591, type=quantize_per_tensor];
-"1592 dequantize_per_tensor_default_94" [id=1592, type=dequantize_per_tensor];
-"1593 linalg_vector_norm_31" [id=1593, type=linalg_vector_norm];
-"1594 clamp_min_31" [id=1594, type=clamp_min];
-"1595 expand_as_31" [id=1595, type=expand_as];
-"1596 div_31" [id=1596, type=div];
-"1597 quantize_per_tensor_default_95" [id=1597, type=quantize_per_tensor];
-"1598 dequantize_per_tensor_default_95" [id=1598, type=dequantize_per_tensor];
-"1599 transpose_30" [id=1599, type=transpose];
-"1600 matmul_30" [id=1600, type=matmul];
-"1601 _frozen_param126" [id=1601, type=get_attr];
-"1602 mul_31" [id=1602, type=mul];
-"1603 add_52" [id=1603, type=add];
-"1604 new_zeros_7" [id=1604, type=new_zeros];
-"1605 view_84" [id=1605, type=view];
-"1606 permute_71" [id=1606, type=permute];
-"1607 reshape_69" [id=1607, type=reshape];
-"1608 unsqueeze_44" [id=1608, type=unsqueeze];
-"1609 unsqueeze_45" [id=1609, type=unsqueeze];
-"1610 sub_7" [id=1610, type=sub];
-"1611 ne_7" [id=1611, type=ne];
-"1612 masked_fill_14" [id=1612, type=masked_fill];
-"1613 eq_7" [id=1613, type=eq];
-"1614 masked_fill_15" [id=1614, type=masked_fill];
-"1615 view_85" [id=1615, type=view];
-"1616 unsqueeze_46" [id=1616, type=unsqueeze];
-"1617 unsqueeze_47" [id=1617, type=unsqueeze];
-"1618 add_53" [id=1618, type=add];
-"1619 view_86" [id=1619, type=view];
-"1620 softmax_15" [id=1620, type=softmax];
-"1621 dropout_60" [id=1621, type=dropout];
-"1622 matmul_31" [id=1622, type=matmul];
-"1623 transpose_31" [id=1623, type=transpose];
-"1624 reshape_70" [id=1624, type=reshape];
-"1625 reshape_70_0_0_nncf_smooth_quant_0" [id=1625, type=call_module];
-"1626 quantize_per_tensor_default_96" [id=1626, type=quantize_per_tensor];
-"1627 dequantize_per_tensor_default_96" [id=1627, type=dequantize_per_tensor];
-"1628 _param_constant257_0_0" [id=1628, type=get_attr];
-"1629 scale_updated_constant64" [id=1629, type=get_attr];
-"1630 compressed_weight_updated_constant64" [id=1630, type=get_attr];
-"1631 mul_tensor_64" [id=1631, type=mul];
-"1632 zero_point_updated_constant64" [id=1632, type=get_attr];
-"1633 sub_tensor_64" [id=1633, type=sub];
-"1634 linear_95" [id=1634, type=linear];
-"1635 dropout_61" [id=1635, type=dropout];
-"1636 view_87" [id=1636, type=view];
-"1637 permute_72" [id=1637, type=permute];
-"1638 reshape_71" [id=1638, type=reshape];
-"1639 roll_15" [id=1639, type=roll];
-"1640 slice_241" [id=1640, type=slice];
-"1641 slice_242" [id=1641, type=slice];
-"1642 slice_243" [id=1642, type=slice];
-"1643 slice_244" [id=1643, type=slice];
-"1644 contiguous_29" [id=1644, type=contiguous];
-"1645 _param_constant258" [id=1645, type=get_attr];
-"1646 _param_constant259" [id=1646, type=get_attr];
-"1647 layer_norm_33" [id=1647, type=layer_norm];
-"1648 add_54" [id=1648, type=add];
-"1649 add_54_0_0_nncf_smooth_quant_0" [id=1649, type=call_module];
-"1650 quantize_per_tensor_default_97" [id=1650, type=quantize_per_tensor];
-"1651 dequantize_per_tensor_default_97" [id=1651, type=dequantize_per_tensor];
-"1652 _param_constant261_0_0" [id=1652, type=get_attr];
-"1653 scale_updated_constant65" [id=1653, type=get_attr];
-"1654 compressed_weight_updated_constant65" [id=1654, type=get_attr];
-"1655 mul_tensor_65" [id=1655, type=mul];
-"1656 zero_point_updated_constant65" [id=1656, type=get_attr];
-"1657 sub_tensor_65" [id=1657, type=sub];
-"1658 linear_96" [id=1658, type=linear];
-"1659 gelu_15" [id=1659, type=gelu];
-"1660 dropout_62" [id=1660, type=dropout];
-"1661 dropout_62_0_0_nncf_smooth_quant_0" [id=1661, type=call_module];
-"1662 quantize_per_tensor_default_98" [id=1662, type=quantize_per_tensor];
-"1663 dequantize_per_tensor_default_98" [id=1663, type=dequantize_per_tensor];
-"1664 _param_constant263_0_0" [id=1664, type=get_attr];
-"1665 scale_updated_constant66" [id=1665, type=get_attr];
-"1666 compressed_weight_updated_constant66" [id=1666, type=get_attr];
-"1667 mul_tensor_66" [id=1667, type=mul];
-"1668 zero_point_updated_constant66" [id=1668, type=get_attr];
-"1669 sub_tensor_66" [id=1669, type=sub];
-"1670 linear_97" [id=1670, type=linear];
-"1671 dropout_63" [id=1671, type=dropout];
-"1672 _param_constant264" [id=1672, type=get_attr];
-"1673 _param_constant265" [id=1673, type=get_attr];
-"1674 layer_norm_34" [id=1674, type=layer_norm];
-"1675 add_55" [id=1675, type=add];
-"1676 _frozen_param136" [id=1676, type=get_attr];
-"1677 pad_18" [id=1677, type=pad];
-"1678 view_90" [id=1678, type=view];
-"1679 permute_74" [id=1679, type=permute];
-"1680 reshape_72" [id=1680, type=reshape];
-"1681 reshape_72_0_0_nncf_smooth_quant_0" [id=1681, type=call_module];
-"1682 quantize_per_tensor_default_99" [id=1682, type=quantize_per_tensor];
-"1683 dequantize_per_tensor_default_99" [id=1683, type=dequantize_per_tensor];
-"1684 _frozen_param137_0_0" [id=1684, type=get_attr];
-"1685 scale_updated_constant67" [id=1685, type=get_attr];
-"1686 compressed_weight_updated_constant67" [id=1686, type=get_attr];
-"1687 mul_tensor_67" [id=1687, type=mul];
-"1688 zero_point_updated_constant67" [id=1688, type=get_attr];
-"1689 sub_tensor_67" [id=1689, type=sub];
-"1690 linear_100" [id=1690, type=linear];
-"1691 reshape_73" [id=1691, type=reshape];
-"1692 permute_75" [id=1692, type=permute];
-"1693 select_48" [id=1693, type=select];
-"1694 select_49" [id=1694, type=select];
-"1695 select_50" [id=1695, type=select];
-"1696 linalg_vector_norm_32" [id=1696, type=linalg_vector_norm];
-"1697 clamp_min_32" [id=1697, type=clamp_min];
-"1698 expand_as_32" [id=1698, type=expand_as];
-"1699 div_32" [id=1699, type=div];
-"1700 quantize_per_tensor_default_100" [id=1700, type=quantize_per_tensor];
-"1701 dequantize_per_tensor_default_100" [id=1701, type=dequantize_per_tensor];
-"1702 linalg_vector_norm_33" [id=1702, type=linalg_vector_norm];
-"1703 clamp_min_33" [id=1703, type=clamp_min];
-"1704 expand_as_33" [id=1704, type=expand_as];
-"1705 div_33" [id=1705, type=div];
-"1706 quantize_per_tensor_default_101" [id=1706, type=quantize_per_tensor];
-"1707 dequantize_per_tensor_default_101" [id=1707, type=dequantize_per_tensor];
-"1708 transpose_32" [id=1708, type=transpose];
-"1709 matmul_32" [id=1709, type=matmul];
-"1710 _frozen_param139" [id=1710, type=get_attr];
-"1711 mul_33" [id=1711, type=mul];
-"1712 add_56" [id=1712, type=add];
-"1713 softmax_16" [id=1713, type=softmax];
-"1714 dropout_64" [id=1714, type=dropout];
-"1715 matmul_33" [id=1715, type=matmul];
-"1716 transpose_33" [id=1716, type=transpose];
-"1717 reshape_74" [id=1717, type=reshape];
-"1718 reshape_74_0_0_nncf_smooth_quant_0" [id=1718, type=call_module];
-"1719 quantize_per_tensor_default_102" [id=1719, type=quantize_per_tensor];
-"1720 dequantize_per_tensor_default_102" [id=1720, type=dequantize_per_tensor];
-"1721 _param_constant273_0_0" [id=1721, type=get_attr];
-"1722 scale_updated_constant68" [id=1722, type=get_attr];
-"1723 compressed_weight_updated_constant68" [id=1723, type=get_attr];
-"1724 mul_tensor_68" [id=1724, type=mul];
-"1725 zero_point_updated_constant68" [id=1725, type=get_attr];
-"1726 sub_tensor_68" [id=1726, type=sub];
-"1727 linear_101" [id=1727, type=linear];
-"1728 dropout_65" [id=1728, type=dropout];
-"1729 view_91" [id=1729, type=view];
-"1730 permute_76" [id=1730, type=permute];
-"1731 reshape_75" [id=1731, type=reshape];
-"1732 slice_246" [id=1732, type=slice];
-"1733 slice_247" [id=1733, type=slice];
-"1734 slice_248" [id=1734, type=slice];
-"1735 slice_249" [id=1735, type=slice];
-"1736 contiguous_31" [id=1736, type=contiguous];
-"1737 _param_constant274" [id=1737, type=get_attr];
-"1738 _param_constant275" [id=1738, type=get_attr];
-"1739 layer_norm_35" [id=1739, type=layer_norm];
-"1740 add_57" [id=1740, type=add];
-"1741 add_57_0_0_nncf_smooth_quant_0" [id=1741, type=call_module];
-"1742 quantize_per_tensor_default_103" [id=1742, type=quantize_per_tensor];
-"1743 dequantize_per_tensor_default_103" [id=1743, type=dequantize_per_tensor];
-"1744 _param_constant277_0_0" [id=1744, type=get_attr];
-"1745 scale_updated_constant69" [id=1745, type=get_attr];
-"1746 compressed_weight_updated_constant69" [id=1746, type=get_attr];
-"1747 mul_tensor_69" [id=1747, type=mul];
-"1748 zero_point_updated_constant69" [id=1748, type=get_attr];
-"1749 sub_tensor_69" [id=1749, type=sub];
-"1750 linear_102" [id=1750, type=linear];
-"1751 gelu_16" [id=1751, type=gelu];
-"1752 dropout_66" [id=1752, type=dropout];
-"1753 dropout_66_0_0_nncf_smooth_quant_0" [id=1753, type=call_module];
-"1754 quantize_per_tensor_default_104" [id=1754, type=quantize_per_tensor];
-"1755 dequantize_per_tensor_default_104" [id=1755, type=dequantize_per_tensor];
-"1756 _param_constant279_0_0" [id=1756, type=get_attr];
-"1757 scale_updated_constant70" [id=1757, type=get_attr];
-"1758 compressed_weight_updated_constant70" [id=1758, type=get_attr];
-"1759 mul_tensor_70" [id=1759, type=mul];
-"1760 zero_point_updated_constant70" [id=1760, type=get_attr];
-"1761 sub_tensor_70" [id=1761, type=sub];
-"1762 linear_103" [id=1762, type=linear];
-"1763 dropout_67" [id=1763, type=dropout];
-"1764 _param_constant280" [id=1764, type=get_attr];
-"1765 _param_constant281" [id=1765, type=get_attr];
-"1766 layer_norm_36" [id=1766, type=layer_norm];
-"1767 add_58" [id=1767, type=add];
-"1768 _frozen_param140" [id=1768, type=get_attr];
-"1769 pad_19" [id=1769, type=pad];
-"1770 roll_16" [id=1770, type=roll];
-"1771 view_94" [id=1771, type=view];
-"1772 permute_78" [id=1772, type=permute];
-"1773 reshape_76" [id=1773, type=reshape];
-"1774 reshape_76_0_0_nncf_smooth_quant_0" [id=1774, type=call_module];
-"1775 quantize_per_tensor_default_105" [id=1775, type=quantize_per_tensor];
-"1776 dequantize_per_tensor_default_105" [id=1776, type=dequantize_per_tensor];
-"1777 _frozen_param141_0_0" [id=1777, type=get_attr];
-"1778 scale_updated_constant71" [id=1778, type=get_attr];
-"1779 compressed_weight_updated_constant71" [id=1779, type=get_attr];
-"1780 mul_tensor_71" [id=1780, type=mul];
-"1781 zero_point_updated_constant71" [id=1781, type=get_attr];
-"1782 sub_tensor_71" [id=1782, type=sub];
-"1783 linear_106" [id=1783, type=linear];
-"1784 reshape_77" [id=1784, type=reshape];
-"1785 permute_79" [id=1785, type=permute];
-"1786 select_51" [id=1786, type=select];
-"1787 select_52" [id=1787, type=select];
-"1788 select_53" [id=1788, type=select];
-"1789 linalg_vector_norm_34" [id=1789, type=linalg_vector_norm];
-"1790 clamp_min_34" [id=1790, type=clamp_min];
-"1791 expand_as_34" [id=1791, type=expand_as];
-"1792 div_34" [id=1792, type=div];
-"1793 quantize_per_tensor_default_106" [id=1793, type=quantize_per_tensor];
-"1794 dequantize_per_tensor_default_106" [id=1794, type=dequantize_per_tensor];
-"1795 linalg_vector_norm_35" [id=1795, type=linalg_vector_norm];
-"1796 clamp_min_35" [id=1796, type=clamp_min];
-"1797 expand_as_35" [id=1797, type=expand_as];
-"1798 div_35" [id=1798, type=div];
-"1799 quantize_per_tensor_default_107" [id=1799, type=quantize_per_tensor];
-"1800 dequantize_per_tensor_default_107" [id=1800, type=dequantize_per_tensor];
-"1801 transpose_34" [id=1801, type=transpose];
-"1802 matmul_34" [id=1802, type=matmul];
-"1803 _frozen_param143" [id=1803, type=get_attr];
-"1804 mul_35" [id=1804, type=mul];
-"1805 add_59" [id=1805, type=add];
-"1806 new_zeros_8" [id=1806, type=new_zeros];
-"1807 view_95" [id=1807, type=view];
-"1808 permute_80" [id=1808, type=permute];
-"1809 reshape_78" [id=1809, type=reshape];
-"1810 unsqueeze_50" [id=1810, type=unsqueeze];
-"1811 unsqueeze_51" [id=1811, type=unsqueeze];
-"1812 sub_8" [id=1812, type=sub];
-"1813 ne_8" [id=1813, type=ne];
-"1814 masked_fill_16" [id=1814, type=masked_fill];
-"1815 eq_8" [id=1815, type=eq];
-"1816 masked_fill_17" [id=1816, type=masked_fill];
-"1817 view_96" [id=1817, type=view];
-"1818 unsqueeze_52" [id=1818, type=unsqueeze];
-"1819 unsqueeze_53" [id=1819, type=unsqueeze];
-"1820 add_60" [id=1820, type=add];
-"1821 view_97" [id=1821, type=view];
-"1822 softmax_17" [id=1822, type=softmax];
-"1823 dropout_68" [id=1823, type=dropout];
-"1824 matmul_35" [id=1824, type=matmul];
-"1825 transpose_35" [id=1825, type=transpose];
-"1826 reshape_79" [id=1826, type=reshape];
-"1827 reshape_79_0_0_nncf_smooth_quant_0" [id=1827, type=call_module];
-"1828 quantize_per_tensor_default_108" [id=1828, type=quantize_per_tensor];
-"1829 dequantize_per_tensor_default_108" [id=1829, type=dequantize_per_tensor];
-"1830 _param_constant289_0_0" [id=1830, type=get_attr];
-"1831 scale_updated_constant72" [id=1831, type=get_attr];
-"1832 compressed_weight_updated_constant72" [id=1832, type=get_attr];
-"1833 mul_tensor_72" [id=1833, type=mul];
-"1834 zero_point_updated_constant72" [id=1834, type=get_attr];
-"1835 sub_tensor_72" [id=1835, type=sub];
-"1836 linear_107" [id=1836, type=linear];
-"1837 dropout_69" [id=1837, type=dropout];
-"1838 view_98" [id=1838, type=view];
-"1839 permute_81" [id=1839, type=permute];
-"1840 reshape_80" [id=1840, type=reshape];
-"1841 roll_17" [id=1841, type=roll];
-"1842 slice_269" [id=1842, type=slice];
-"1843 slice_270" [id=1843, type=slice];
-"1844 slice_271" [id=1844, type=slice];
-"1845 slice_272" [id=1845, type=slice];
-"1846 contiguous_33" [id=1846, type=contiguous];
-"1847 _param_constant290" [id=1847, type=get_attr];
-"1848 _param_constant291" [id=1848, type=get_attr];
-"1849 layer_norm_37" [id=1849, type=layer_norm];
-"1850 add_61" [id=1850, type=add];
-"1851 add_61_0_0_nncf_smooth_quant_0" [id=1851, type=call_module];
-"1852 quantize_per_tensor_default_109" [id=1852, type=quantize_per_tensor];
-"1853 dequantize_per_tensor_default_109" [id=1853, type=dequantize_per_tensor];
-"1854 _param_constant293_0_0" [id=1854, type=get_attr];
-"1855 scale_updated_constant73" [id=1855, type=get_attr];
-"1856 compressed_weight_updated_constant73" [id=1856, type=get_attr];
-"1857 mul_tensor_73" [id=1857, type=mul];
-"1858 zero_point_updated_constant73" [id=1858, type=get_attr];
-"1859 sub_tensor_73" [id=1859, type=sub];
-"1860 linear_108" [id=1860, type=linear];
-"1861 gelu_17" [id=1861, type=gelu];
-"1862 dropout_70" [id=1862, type=dropout];
-"1863 dropout_70_0_0_nncf_smooth_quant_0" [id=1863, type=call_module];
-"1864 quantize_per_tensor_default_110" [id=1864, type=quantize_per_tensor];
-"1865 dequantize_per_tensor_default_110" [id=1865, type=dequantize_per_tensor];
-"1866 _param_constant295_0_0" [id=1866, type=get_attr];
-"1867 scale_updated_constant74" [id=1867, type=get_attr];
-"1868 compressed_weight_updated_constant74" [id=1868, type=get_attr];
-"1869 mul_tensor_74" [id=1869, type=mul];
-"1870 zero_point_updated_constant74" [id=1870, type=get_attr];
-"1871 sub_tensor_74" [id=1871, type=sub];
-"1872 linear_109" [id=1872, type=linear];
-"1873 dropout_71" [id=1873, type=dropout];
-"1874 _param_constant296" [id=1874, type=get_attr];
-"1875 _param_constant297" [id=1875, type=get_attr];
-"1876 layer_norm_38" [id=1876, type=layer_norm];
-"1877 add_62" [id=1877, type=add];
-"1878 _frozen_param153" [id=1878, type=get_attr];
-"1879 pad_20" [id=1879, type=pad];
-"1880 view_101" [id=1880, type=view];
-"1881 permute_83" [id=1881, type=permute];
-"1882 reshape_81" [id=1882, type=reshape];
-"1883 reshape_81_0_0_nncf_smooth_quant_0" [id=1883, type=call_module];
-"1884 quantize_per_tensor_default_111" [id=1884, type=quantize_per_tensor];
-"1885 dequantize_per_tensor_default_111" [id=1885, type=dequantize_per_tensor];
-"1886 _frozen_param154_0_0" [id=1886, type=get_attr];
-"1887 scale_updated_constant75" [id=1887, type=get_attr];
-"1888 compressed_weight_updated_constant75" [id=1888, type=get_attr];
-"1889 mul_tensor_75" [id=1889, type=mul];
-"1890 zero_point_updated_constant75" [id=1890, type=get_attr];
-"1891 sub_tensor_75" [id=1891, type=sub];
-"1892 linear_112" [id=1892, type=linear];
-"1893 reshape_82" [id=1893, type=reshape];
-"1894 permute_84" [id=1894, type=permute];
-"1895 select_54" [id=1895, type=select];
-"1896 select_55" [id=1896, type=select];
-"1897 select_56" [id=1897, type=select];
-"1898 linalg_vector_norm_36" [id=1898, type=linalg_vector_norm];
-"1899 clamp_min_36" [id=1899, type=clamp_min];
-"1900 expand_as_36" [id=1900, type=expand_as];
-"1901 div_36" [id=1901, type=div];
-"1902 quantize_per_tensor_default_112" [id=1902, type=quantize_per_tensor];
-"1903 dequantize_per_tensor_default_112" [id=1903, type=dequantize_per_tensor];
-"1904 linalg_vector_norm_37" [id=1904, type=linalg_vector_norm];
-"1905 clamp_min_37" [id=1905, type=clamp_min];
-"1906 expand_as_37" [id=1906, type=expand_as];
-"1907 div_37" [id=1907, type=div];
-"1908 quantize_per_tensor_default_113" [id=1908, type=quantize_per_tensor];
-"1909 dequantize_per_tensor_default_113" [id=1909, type=dequantize_per_tensor];
-"1910 transpose_36" [id=1910, type=transpose];
-"1911 matmul_36" [id=1911, type=matmul];
-"1912 _frozen_param156" [id=1912, type=get_attr];
-"1913 mul_37" [id=1913, type=mul];
-"1914 add_63" [id=1914, type=add];
-"1915 softmax_18" [id=1915, type=softmax];
-"1916 dropout_72" [id=1916, type=dropout];
-"1917 matmul_37" [id=1917, type=matmul];
-"1918 transpose_37" [id=1918, type=transpose];
-"1919 reshape_83" [id=1919, type=reshape];
-"1920 reshape_83_0_0_nncf_smooth_quant_0" [id=1920, type=call_module];
-"1921 quantize_per_tensor_default_114" [id=1921, type=quantize_per_tensor];
-"1922 dequantize_per_tensor_default_114" [id=1922, type=dequantize_per_tensor];
-"1923 _param_constant305_0_0" [id=1923, type=get_attr];
-"1924 scale_updated_constant76" [id=1924, type=get_attr];
-"1925 compressed_weight_updated_constant76" [id=1925, type=get_attr];
-"1926 mul_tensor_76" [id=1926, type=mul];
-"1927 zero_point_updated_constant76" [id=1927, type=get_attr];
-"1928 sub_tensor_76" [id=1928, type=sub];
-"1929 linear_113" [id=1929, type=linear];
-"1930 dropout_73" [id=1930, type=dropout];
-"1931 view_102" [id=1931, type=view];
-"1932 permute_85" [id=1932, type=permute];
-"1933 reshape_84" [id=1933, type=reshape];
-"1934 slice_274" [id=1934, type=slice];
-"1935 slice_275" [id=1935, type=slice];
-"1936 slice_276" [id=1936, type=slice];
-"1937 slice_277" [id=1937, type=slice];
-"1938 contiguous_35" [id=1938, type=contiguous];
-"1939 _param_constant306" [id=1939, type=get_attr];
-"1940 _param_constant307" [id=1940, type=get_attr];
-"1941 layer_norm_39" [id=1941, type=layer_norm];
-"1942 add_64" [id=1942, type=add];
-"1943 add_64_0_0_nncf_smooth_quant_0" [id=1943, type=call_module];
-"1944 quantize_per_tensor_default_115" [id=1944, type=quantize_per_tensor];
-"1945 dequantize_per_tensor_default_115" [id=1945, type=dequantize_per_tensor];
-"1946 _param_constant309_0_0" [id=1946, type=get_attr];
-"1947 scale_updated_constant77" [id=1947, type=get_attr];
-"1948 compressed_weight_updated_constant77" [id=1948, type=get_attr];
-"1949 mul_tensor_77" [id=1949, type=mul];
-"1950 zero_point_updated_constant77" [id=1950, type=get_attr];
-"1951 sub_tensor_77" [id=1951, type=sub];
-"1952 linear_114" [id=1952, type=linear];
-"1953 gelu_18" [id=1953, type=gelu];
-"1954 dropout_74" [id=1954, type=dropout];
-"1955 dropout_74_0_0_nncf_smooth_quant_0" [id=1955, type=call_module];
-"1956 quantize_per_tensor_default_116" [id=1956, type=quantize_per_tensor];
-"1957 dequantize_per_tensor_default_116" [id=1957, type=dequantize_per_tensor];
-"1958 _param_constant311_0_0" [id=1958, type=get_attr];
-"1959 scale_updated_constant78" [id=1959, type=get_attr];
-"1960 compressed_weight_updated_constant78" [id=1960, type=get_attr];
-"1961 mul_tensor_78" [id=1961, type=mul];
-"1962 zero_point_updated_constant78" [id=1962, type=get_attr];
-"1963 sub_tensor_78" [id=1963, type=sub];
-"1964 linear_115" [id=1964, type=linear];
-"1965 dropout_75" [id=1965, type=dropout];
-"1966 _param_constant312" [id=1966, type=get_attr];
-"1967 _param_constant313" [id=1967, type=get_attr];
-"1968 layer_norm_40" [id=1968, type=layer_norm];
-"1969 add_65" [id=1969, type=add];
-"1970 _frozen_param157" [id=1970, type=get_attr];
-"1971 pad_21" [id=1971, type=pad];
-"1972 roll_18" [id=1972, type=roll];
-"1973 view_105" [id=1973, type=view];
-"1974 permute_87" [id=1974, type=permute];
-"1975 reshape_85" [id=1975, type=reshape];
-"1976 reshape_85_0_0_nncf_smooth_quant_0" [id=1976, type=call_module];
-"1977 quantize_per_tensor_default_117" [id=1977, type=quantize_per_tensor];
-"1978 dequantize_per_tensor_default_117" [id=1978, type=dequantize_per_tensor];
-"1979 _frozen_param158_0_0" [id=1979, type=get_attr];
-"1980 scale_updated_constant79" [id=1980, type=get_attr];
-"1981 compressed_weight_updated_constant79" [id=1981, type=get_attr];
-"1982 mul_tensor_79" [id=1982, type=mul];
-"1983 zero_point_updated_constant79" [id=1983, type=get_attr];
-"1984 sub_tensor_79" [id=1984, type=sub];
-"1985 linear_118" [id=1985, type=linear];
-"1986 reshape_86" [id=1986, type=reshape];
-"1987 permute_88" [id=1987, type=permute];
-"1988 select_57" [id=1988, type=select];
-"1989 select_58" [id=1989, type=select];
-"1990 select_59" [id=1990, type=select];
-"1991 linalg_vector_norm_38" [id=1991, type=linalg_vector_norm];
-"1992 clamp_min_38" [id=1992, type=clamp_min];
-"1993 expand_as_38" [id=1993, type=expand_as];
-"1994 div_38" [id=1994, type=div];
-"1995 quantize_per_tensor_default_118" [id=1995, type=quantize_per_tensor];
-"1996 dequantize_per_tensor_default_118" [id=1996, type=dequantize_per_tensor];
-"1997 linalg_vector_norm_39" [id=1997, type=linalg_vector_norm];
-"1998 clamp_min_39" [id=1998, type=clamp_min];
-"1999 expand_as_39" [id=1999, type=expand_as];
-"2000 div_39" [id=2000, type=div];
-"2001 quantize_per_tensor_default_119" [id=2001, type=quantize_per_tensor];
-"2002 dequantize_per_tensor_default_119" [id=2002, type=dequantize_per_tensor];
-"2003 transpose_38" [id=2003, type=transpose];
-"2004 matmul_38" [id=2004, type=matmul];
-"2005 _frozen_param160" [id=2005, type=get_attr];
-"2006 mul_39" [id=2006, type=mul];
-"2007 add_66" [id=2007, type=add];
-"2008 new_zeros_9" [id=2008, type=new_zeros];
-"2009 view_106" [id=2009, type=view];
-"2010 permute_89" [id=2010, type=permute];
-"2011 reshape_87" [id=2011, type=reshape];
-"2012 unsqueeze_56" [id=2012, type=unsqueeze];
-"2013 unsqueeze_57" [id=2013, type=unsqueeze];
-"2014 sub_9" [id=2014, type=sub];
-"2015 ne_9" [id=2015, type=ne];
-"2016 masked_fill_18" [id=2016, type=masked_fill];
-"2017 eq_9" [id=2017, type=eq];
-"2018 masked_fill_19" [id=2018, type=masked_fill];
-"2019 view_107" [id=2019, type=view];
-"2020 unsqueeze_58" [id=2020, type=unsqueeze];
-"2021 unsqueeze_59" [id=2021, type=unsqueeze];
-"2022 add_67" [id=2022, type=add];
-"2023 view_108" [id=2023, type=view];
-"2024 softmax_19" [id=2024, type=softmax];
-"2025 dropout_76" [id=2025, type=dropout];
-"2026 matmul_39" [id=2026, type=matmul];
-"2027 transpose_39" [id=2027, type=transpose];
-"2028 reshape_88" [id=2028, type=reshape];
-"2029 reshape_88_0_0_nncf_smooth_quant_0" [id=2029, type=call_module];
-"2030 quantize_per_tensor_default_120" [id=2030, type=quantize_per_tensor];
-"2031 dequantize_per_tensor_default_120" [id=2031, type=dequantize_per_tensor];
-"2032 _param_constant321_0_0" [id=2032, type=get_attr];
-"2033 scale_updated_constant80" [id=2033, type=get_attr];
-"2034 compressed_weight_updated_constant80" [id=2034, type=get_attr];
-"2035 mul_tensor_80" [id=2035, type=mul];
-"2036 zero_point_updated_constant80" [id=2036, type=get_attr];
-"2037 sub_tensor_80" [id=2037, type=sub];
-"2038 linear_119" [id=2038, type=linear];
-"2039 dropout_77" [id=2039, type=dropout];
-"2040 view_109" [id=2040, type=view];
-"2041 permute_90" [id=2041, type=permute];
-"2042 reshape_89" [id=2042, type=reshape];
-"2043 roll_19" [id=2043, type=roll];
-"2044 slice_297" [id=2044, type=slice];
-"2045 slice_298" [id=2045, type=slice];
-"2046 slice_299" [id=2046, type=slice];
-"2047 slice_300" [id=2047, type=slice];
-"2048 contiguous_37" [id=2048, type=contiguous];
-"2049 _param_constant322" [id=2049, type=get_attr];
-"2050 _param_constant323" [id=2050, type=get_attr];
-"2051 layer_norm_41" [id=2051, type=layer_norm];
-"2052 add_68" [id=2052, type=add];
-"2053 add_68_0_0_nncf_smooth_quant_0" [id=2053, type=call_module];
-"2054 quantize_per_tensor_default_121" [id=2054, type=quantize_per_tensor];
-"2055 dequantize_per_tensor_default_121" [id=2055, type=dequantize_per_tensor];
-"2056 _param_constant325_0_0" [id=2056, type=get_attr];
-"2057 scale_updated_constant81" [id=2057, type=get_attr];
-"2058 compressed_weight_updated_constant81" [id=2058, type=get_attr];
-"2059 mul_tensor_81" [id=2059, type=mul];
-"2060 zero_point_updated_constant81" [id=2060, type=get_attr];
-"2061 sub_tensor_81" [id=2061, type=sub];
-"2062 linear_120" [id=2062, type=linear];
-"2063 gelu_19" [id=2063, type=gelu];
-"2064 dropout_78" [id=2064, type=dropout];
-"2065 dropout_78_0_0_nncf_smooth_quant_0" [id=2065, type=call_module];
-"2066 quantize_per_tensor_default_122" [id=2066, type=quantize_per_tensor];
-"2067 dequantize_per_tensor_default_122" [id=2067, type=dequantize_per_tensor];
-"2068 _param_constant327_0_0" [id=2068, type=get_attr];
-"2069 scale_updated_constant82" [id=2069, type=get_attr];
-"2070 compressed_weight_updated_constant82" [id=2070, type=get_attr];
-"2071 mul_tensor_82" [id=2071, type=mul];
-"2072 zero_point_updated_constant82" [id=2072, type=get_attr];
-"2073 sub_tensor_82" [id=2073, type=sub];
-"2074 linear_121" [id=2074, type=linear];
-"2075 dropout_79" [id=2075, type=dropout];
-"2076 _param_constant328" [id=2076, type=get_attr];
-"2077 _param_constant329" [id=2077, type=get_attr];
-"2078 layer_norm_42" [id=2078, type=layer_norm];
-"2079 add_69" [id=2079, type=add];
-"2080 _frozen_param170" [id=2080, type=get_attr];
-"2081 pad_22" [id=2081, type=pad];
-"2082 view_112" [id=2082, type=view];
-"2083 permute_92" [id=2083, type=permute];
-"2084 reshape_90" [id=2084, type=reshape];
-"2085 reshape_90_0_0_nncf_smooth_quant_0" [id=2085, type=call_module];
-"2086 quantize_per_tensor_default_123" [id=2086, type=quantize_per_tensor];
-"2087 dequantize_per_tensor_default_123" [id=2087, type=dequantize_per_tensor];
-"2088 _frozen_param171_0_0" [id=2088, type=get_attr];
-"2089 scale_updated_constant83" [id=2089, type=get_attr];
-"2090 compressed_weight_updated_constant83" [id=2090, type=get_attr];
-"2091 mul_tensor_83" [id=2091, type=mul];
-"2092 zero_point_updated_constant83" [id=2092, type=get_attr];
-"2093 sub_tensor_83" [id=2093, type=sub];
-"2094 linear_124" [id=2094, type=linear];
-"2095 reshape_91" [id=2095, type=reshape];
-"2096 permute_93" [id=2096, type=permute];
-"2097 select_60" [id=2097, type=select];
-"2098 select_61" [id=2098, type=select];
-"2099 select_62" [id=2099, type=select];
-"2100 linalg_vector_norm_40" [id=2100, type=linalg_vector_norm];
-"2101 clamp_min_40" [id=2101, type=clamp_min];
-"2102 expand_as_40" [id=2102, type=expand_as];
-"2103 div_40" [id=2103, type=div];
-"2104 quantize_per_tensor_default_124" [id=2104, type=quantize_per_tensor];
-"2105 dequantize_per_tensor_default_124" [id=2105, type=dequantize_per_tensor];
-"2106 linalg_vector_norm_41" [id=2106, type=linalg_vector_norm];
-"2107 clamp_min_41" [id=2107, type=clamp_min];
-"2108 expand_as_41" [id=2108, type=expand_as];
-"2109 div_41" [id=2109, type=div];
-"2110 quantize_per_tensor_default_125" [id=2110, type=quantize_per_tensor];
-"2111 dequantize_per_tensor_default_125" [id=2111, type=dequantize_per_tensor];
-"2112 transpose_40" [id=2112, type=transpose];
-"2113 matmul_40" [id=2113, type=matmul];
-"2114 _frozen_param173" [id=2114, type=get_attr];
-"2115 mul_41" [id=2115, type=mul];
-"2116 add_70" [id=2116, type=add];
-"2117 softmax_20" [id=2117, type=softmax];
-"2118 dropout_80" [id=2118, type=dropout];
-"2119 matmul_41" [id=2119, type=matmul];
-"2120 transpose_41" [id=2120, type=transpose];
-"2121 reshape_92" [id=2121, type=reshape];
-"2122 reshape_92_0_0_nncf_smooth_quant_0" [id=2122, type=call_module];
-"2123 quantize_per_tensor_default_126" [id=2123, type=quantize_per_tensor];
-"2124 dequantize_per_tensor_default_126" [id=2124, type=dequantize_per_tensor];
-"2125 _param_constant337_0_0" [id=2125, type=get_attr];
-"2126 scale_updated_constant84" [id=2126, type=get_attr];
-"2127 compressed_weight_updated_constant84" [id=2127, type=get_attr];
-"2128 mul_tensor_84" [id=2128, type=mul];
-"2129 zero_point_updated_constant84" [id=2129, type=get_attr];
-"2130 sub_tensor_84" [id=2130, type=sub];
-"2131 linear_125" [id=2131, type=linear];
-"2132 dropout_81" [id=2132, type=dropout];
-"2133 view_113" [id=2133, type=view];
-"2134 permute_94" [id=2134, type=permute];
-"2135 reshape_93" [id=2135, type=reshape];
-"2136 slice_302" [id=2136, type=slice];
-"2137 slice_303" [id=2137, type=slice];
-"2138 slice_304" [id=2138, type=slice];
-"2139 slice_305" [id=2139, type=slice];
-"2140 contiguous_39" [id=2140, type=contiguous];
-"2141 _param_constant338" [id=2141, type=get_attr];
-"2142 _param_constant339" [id=2142, type=get_attr];
-"2143 layer_norm_43" [id=2143, type=layer_norm];
-"2144 add_71" [id=2144, type=add];
-"2145 add_71_0_0_nncf_smooth_quant_0" [id=2145, type=call_module];
-"2146 quantize_per_tensor_default_127" [id=2146, type=quantize_per_tensor];
-"2147 dequantize_per_tensor_default_127" [id=2147, type=dequantize_per_tensor];
-"2148 _param_constant341_0_0" [id=2148, type=get_attr];
-"2149 scale_updated_constant85" [id=2149, type=get_attr];
-"2150 compressed_weight_updated_constant85" [id=2150, type=get_attr];
-"2151 mul_tensor_85" [id=2151, type=mul];
-"2152 zero_point_updated_constant85" [id=2152, type=get_attr];
-"2153 sub_tensor_85" [id=2153, type=sub];
-"2154 linear_126" [id=2154, type=linear];
-"2155 gelu_20" [id=2155, type=gelu];
-"2156 dropout_82" [id=2156, type=dropout];
-"2157 dropout_82_0_0_nncf_smooth_quant_0" [id=2157, type=call_module];
-"2158 quantize_per_tensor_default_128" [id=2158, type=quantize_per_tensor];
-"2159 dequantize_per_tensor_default_128" [id=2159, type=dequantize_per_tensor];
-"2160 _param_constant343_0_0" [id=2160, type=get_attr];
-"2161 scale_updated_constant86" [id=2161, type=get_attr];
-"2162 compressed_weight_updated_constant86" [id=2162, type=get_attr];
-"2163 mul_tensor_86" [id=2163, type=mul];
-"2164 zero_point_updated_constant86" [id=2164, type=get_attr];
-"2165 sub_tensor_86" [id=2165, type=sub];
-"2166 linear_127" [id=2166, type=linear];
-"2167 dropout_83" [id=2167, type=dropout];
-"2168 _param_constant344" [id=2168, type=get_attr];
-"2169 _param_constant345" [id=2169, type=get_attr];
-"2170 layer_norm_44" [id=2170, type=layer_norm];
-"2171 add_72" [id=2171, type=add];
-"2172 _frozen_param174" [id=2172, type=get_attr];
-"2173 pad_23" [id=2173, type=pad];
-"2174 roll_20" [id=2174, type=roll];
-"2175 view_116" [id=2175, type=view];
-"2176 permute_96" [id=2176, type=permute];
-"2177 reshape_94" [id=2177, type=reshape];
-"2178 reshape_94_0_0_nncf_smooth_quant_0" [id=2178, type=call_module];
-"2179 quantize_per_tensor_default_129" [id=2179, type=quantize_per_tensor];
-"2180 dequantize_per_tensor_default_129" [id=2180, type=dequantize_per_tensor];
-"2181 _frozen_param175_0_0" [id=2181, type=get_attr];
-"2182 scale_updated_constant87" [id=2182, type=get_attr];
-"2183 compressed_weight_updated_constant87" [id=2183, type=get_attr];
-"2184 mul_tensor_87" [id=2184, type=mul];
-"2185 zero_point_updated_constant87" [id=2185, type=get_attr];
-"2186 sub_tensor_87" [id=2186, type=sub];
-"2187 linear_130" [id=2187, type=linear];
-"2188 reshape_95" [id=2188, type=reshape];
-"2189 permute_97" [id=2189, type=permute];
-"2190 select_63" [id=2190, type=select];
-"2191 select_64" [id=2191, type=select];
-"2192 select_65" [id=2192, type=select];
-"2193 linalg_vector_norm_42" [id=2193, type=linalg_vector_norm];
-"2194 clamp_min_42" [id=2194, type=clamp_min];
-"2195 expand_as_42" [id=2195, type=expand_as];
-"2196 div_42" [id=2196, type=div];
-"2197 quantize_per_tensor_default_130" [id=2197, type=quantize_per_tensor];
-"2198 dequantize_per_tensor_default_130" [id=2198, type=dequantize_per_tensor];
-"2199 linalg_vector_norm_43" [id=2199, type=linalg_vector_norm];
-"2200 clamp_min_43" [id=2200, type=clamp_min];
-"2201 expand_as_43" [id=2201, type=expand_as];
-"2202 div_43" [id=2202, type=div];
-"2203 quantize_per_tensor_default_131" [id=2203, type=quantize_per_tensor];
-"2204 dequantize_per_tensor_default_131" [id=2204, type=dequantize_per_tensor];
-"2205 transpose_42" [id=2205, type=transpose];
-"2206 matmul_42" [id=2206, type=matmul];
-"2207 _frozen_param177" [id=2207, type=get_attr];
-"2208 mul_43" [id=2208, type=mul];
-"2209 add_73" [id=2209, type=add];
-"2210 new_zeros_10" [id=2210, type=new_zeros];
-"2211 view_117" [id=2211, type=view];
-"2212 permute_98" [id=2212, type=permute];
-"2213 reshape_96" [id=2213, type=reshape];
-"2214 unsqueeze_62" [id=2214, type=unsqueeze];
-"2215 unsqueeze_63" [id=2215, type=unsqueeze];
-"2216 sub_10" [id=2216, type=sub];
-"2217 ne_10" [id=2217, type=ne];
-"2218 masked_fill_20" [id=2218, type=masked_fill];
-"2219 eq_10" [id=2219, type=eq];
-"2220 masked_fill_21" [id=2220, type=masked_fill];
-"2221 view_118" [id=2221, type=view];
-"2222 unsqueeze_64" [id=2222, type=unsqueeze];
-"2223 unsqueeze_65" [id=2223, type=unsqueeze];
-"2224 add_74" [id=2224, type=add];
-"2225 view_119" [id=2225, type=view];
-"2226 softmax_21" [id=2226, type=softmax];
-"2227 dropout_84" [id=2227, type=dropout];
-"2228 matmul_43" [id=2228, type=matmul];
-"2229 transpose_43" [id=2229, type=transpose];
-"2230 reshape_97" [id=2230, type=reshape];
-"2231 reshape_97_0_0_nncf_smooth_quant_0" [id=2231, type=call_module];
-"2232 quantize_per_tensor_default_132" [id=2232, type=quantize_per_tensor];
-"2233 dequantize_per_tensor_default_132" [id=2233, type=dequantize_per_tensor];
-"2234 _param_constant353_0_0" [id=2234, type=get_attr];
-"2235 scale_updated_constant88" [id=2235, type=get_attr];
-"2236 compressed_weight_updated_constant88" [id=2236, type=get_attr];
-"2237 mul_tensor_88" [id=2237, type=mul];
-"2238 zero_point_updated_constant88" [id=2238, type=get_attr];
-"2239 sub_tensor_88" [id=2239, type=sub];
-"2240 linear_131" [id=2240, type=linear];
-"2241 dropout_85" [id=2241, type=dropout];
-"2242 view_120" [id=2242, type=view];
-"2243 permute_99" [id=2243, type=permute];
-"2244 reshape_98" [id=2244, type=reshape];
-"2245 roll_21" [id=2245, type=roll];
-"2246 slice_325" [id=2246, type=slice];
-"2247 slice_326" [id=2247, type=slice];
-"2248 slice_327" [id=2248, type=slice];
-"2249 slice_328" [id=2249, type=slice];
-"2250 contiguous_41" [id=2250, type=contiguous];
-"2251 _param_constant354" [id=2251, type=get_attr];
-"2252 _param_constant355" [id=2252, type=get_attr];
-"2253 layer_norm_45" [id=2253, type=layer_norm];
-"2254 add_75" [id=2254, type=add];
-"2255 add_75_0_0_nncf_smooth_quant_0" [id=2255, type=call_module];
-"2256 quantize_per_tensor_default_133" [id=2256, type=quantize_per_tensor];
-"2257 dequantize_per_tensor_default_133" [id=2257, type=dequantize_per_tensor];
-"2258 _param_constant357_0_0" [id=2258, type=get_attr];
-"2259 scale_updated_constant89" [id=2259, type=get_attr];
-"2260 compressed_weight_updated_constant89" [id=2260, type=get_attr];
-"2261 mul_tensor_89" [id=2261, type=mul];
-"2262 zero_point_updated_constant89" [id=2262, type=get_attr];
-"2263 sub_tensor_89" [id=2263, type=sub];
-"2264 linear_132" [id=2264, type=linear];
-"2265 gelu_21" [id=2265, type=gelu];
-"2266 dropout_86" [id=2266, type=dropout];
-"2267 dropout_86_0_0_nncf_smooth_quant_0" [id=2267, type=call_module];
-"2268 quantize_per_tensor_default_134" [id=2268, type=quantize_per_tensor];
-"2269 dequantize_per_tensor_default_134" [id=2269, type=dequantize_per_tensor];
-"2270 _param_constant359_0_0" [id=2270, type=get_attr];
-"2271 scale_updated_constant90" [id=2271, type=get_attr];
-"2272 compressed_weight_updated_constant90" [id=2272, type=get_attr];
-"2273 mul_tensor_90" [id=2273, type=mul];
-"2274 zero_point_updated_constant90" [id=2274, type=get_attr];
-"2275 sub_tensor_90" [id=2275, type=sub];
-"2276 linear_133" [id=2276, type=linear];
-"2277 dropout_87" [id=2277, type=dropout];
-"2278 _param_constant360" [id=2278, type=get_attr];
-"2279 _param_constant361" [id=2279, type=get_attr];
-"2280 layer_norm_46" [id=2280, type=layer_norm];
-"2281 add_76" [id=2281, type=add];
-"2282 pad_24" [id=2282, type=pad];
-"2283 slice_329" [id=2283, type=slice];
-"2284 slice_330" [id=2284, type=slice];
-"2285 slice_331" [id=2285, type=slice];
-"2286 slice_332" [id=2286, type=slice];
-"2287 slice_333" [id=2287, type=slice];
-"2288 slice_334" [id=2288, type=slice];
-"2289 slice_335" [id=2289, type=slice];
-"2290 slice_336" [id=2290, type=slice];
-"2291 slice_337" [id=2291, type=slice];
-"2292 slice_338" [id=2292, type=slice];
-"2293 slice_339" [id=2293, type=slice];
-"2294 slice_340" [id=2294, type=slice];
-"2295 cat_2" [id=2295, type=cat];
-"2296 cat_2_0_0_nncf_smooth_quant_0" [id=2296, type=call_module];
-"2297 quantize_per_tensor_default_135" [id=2297, type=quantize_per_tensor];
-"2298 dequantize_per_tensor_default_135" [id=2298, type=dequantize_per_tensor];
-"2299 scale_updated_constant91" [id=2299, type=get_attr];
-"2300 compressed_weight_updated_constant91" [id=2300, type=get_attr];
-"2301 mul_tensor_91" [id=2301, type=mul];
-"2302 zero_point_updated_constant91" [id=2302, type=get_attr];
-"2303 sub_tensor_91" [id=2303, type=sub];
-"2304 linear_134" [id=2304, type=linear];
-"2305 _param_constant363" [id=2305, type=get_attr];
-"2306 _param_constant364" [id=2306, type=get_attr];
-"2307 layer_norm_47" [id=2307, type=layer_norm];
-"2308 _frozen_param187" [id=2308, type=get_attr];
-"2309 pad_25" [id=2309, type=pad];
-"2310 view_123" [id=2310, type=view];
-"2311 permute_101" [id=2311, type=permute];
-"2312 reshape_99" [id=2312, type=reshape];
-"2313 reshape_99_0_0_nncf_smooth_quant_0" [id=2313, type=call_module];
-"2314 quantize_per_tensor_default_136" [id=2314, type=quantize_per_tensor];
-"2315 dequantize_per_tensor_default_136" [id=2315, type=dequantize_per_tensor];
-"2316 _frozen_param188_0_0" [id=2316, type=get_attr];
-"2317 scale_updated_constant92" [id=2317, type=get_attr];
-"2318 compressed_weight_updated_constant92" [id=2318, type=get_attr];
-"2319 mul_tensor_92" [id=2319, type=mul];
-"2320 zero_point_updated_constant92" [id=2320, type=get_attr];
-"2321 sub_tensor_92" [id=2321, type=sub];
-"2322 linear_137" [id=2322, type=linear];
-"2323 reshape_100" [id=2323, type=reshape];
-"2324 permute_102" [id=2324, type=permute];
-"2325 select_66" [id=2325, type=select];
-"2326 select_67" [id=2326, type=select];
-"2327 select_68" [id=2327, type=select];
-"2328 linalg_vector_norm_44" [id=2328, type=linalg_vector_norm];
-"2329 clamp_min_44" [id=2329, type=clamp_min];
-"2330 expand_as_44" [id=2330, type=expand_as];
-"2331 div_44" [id=2331, type=div];
-"2332 quantize_per_tensor_default_137" [id=2332, type=quantize_per_tensor];
-"2333 dequantize_per_tensor_default_137" [id=2333, type=dequantize_per_tensor];
-"2334 linalg_vector_norm_45" [id=2334, type=linalg_vector_norm];
-"2335 clamp_min_45" [id=2335, type=clamp_min];
-"2336 expand_as_45" [id=2336, type=expand_as];
-"2337 div_45" [id=2337, type=div];
-"2338 quantize_per_tensor_default_138" [id=2338, type=quantize_per_tensor];
-"2339 dequantize_per_tensor_default_138" [id=2339, type=dequantize_per_tensor];
-"2340 transpose_44" [id=2340, type=transpose];
-"2341 matmul_44" [id=2341, type=matmul];
-"2342 _frozen_param190" [id=2342, type=get_attr];
-"2343 mul_45" [id=2343, type=mul];
-"2344 add_77" [id=2344, type=add];
-"2345 softmax_22" [id=2345, type=softmax];
-"2346 dropout_88" [id=2346, type=dropout];
-"2347 matmul_45" [id=2347, type=matmul];
-"2348 transpose_45" [id=2348, type=transpose];
-"2349 reshape_101" [id=2349, type=reshape];
-"2350 reshape_101_0_0_nncf_smooth_quant_0" [id=2350, type=call_module];
-"2351 quantize_per_tensor_default_139" [id=2351, type=quantize_per_tensor];
-"2352 dequantize_per_tensor_default_139" [id=2352, type=dequantize_per_tensor];
-"2353 _param_constant372_0_0" [id=2353, type=get_attr];
-"2354 scale_updated_constant93" [id=2354, type=get_attr];
-"2355 compressed_weight_updated_constant93" [id=2355, type=get_attr];
-"2356 mul_tensor_93" [id=2356, type=mul];
-"2357 zero_point_updated_constant93" [id=2357, type=get_attr];
-"2358 sub_tensor_93" [id=2358, type=sub];
-"2359 linear_138" [id=2359, type=linear];
-"2360 dropout_89" [id=2360, type=dropout];
-"2361 view_124" [id=2361, type=view];
-"2362 permute_103" [id=2362, type=permute];
-"2363 reshape_102" [id=2363, type=reshape];
-"2364 slice_342" [id=2364, type=slice];
-"2365 slice_343" [id=2365, type=slice];
-"2366 slice_344" [id=2366, type=slice];
-"2367 slice_345" [id=2367, type=slice];
-"2368 contiguous_43" [id=2368, type=contiguous];
-"2369 _param_constant373" [id=2369, type=get_attr];
-"2370 _param_constant374" [id=2370, type=get_attr];
-"2371 layer_norm_48" [id=2371, type=layer_norm];
-"2372 add_78" [id=2372, type=add];
-"2373 add_78_0_0_nncf_smooth_quant_0" [id=2373, type=call_module];
-"2374 quantize_per_tensor_default_140" [id=2374, type=quantize_per_tensor];
-"2375 dequantize_per_tensor_default_140" [id=2375, type=dequantize_per_tensor];
-"2376 _param_constant376_0_0" [id=2376, type=get_attr];
-"2377 scale_updated_constant94" [id=2377, type=get_attr];
-"2378 compressed_weight_updated_constant94" [id=2378, type=get_attr];
-"2379 mul_tensor_94" [id=2379, type=mul];
-"2380 zero_point_updated_constant94" [id=2380, type=get_attr];
-"2381 sub_tensor_94" [id=2381, type=sub];
-"2382 linear_139" [id=2382, type=linear];
-"2383 gelu_22" [id=2383, type=gelu];
-"2384 dropout_90" [id=2384, type=dropout];
-"2385 dropout_90_0_0_nncf_smooth_quant_0" [id=2385, type=call_module];
-"2386 quantize_per_tensor_default_141" [id=2386, type=quantize_per_tensor];
-"2387 dequantize_per_tensor_default_141" [id=2387, type=dequantize_per_tensor];
-"2388 _param_constant378_0_0" [id=2388, type=get_attr];
-"2389 scale_updated_constant95" [id=2389, type=get_attr];
-"2390 compressed_weight_updated_constant95" [id=2390, type=get_attr];
-"2391 mul_tensor_95" [id=2391, type=mul];
-"2392 zero_point_updated_constant95" [id=2392, type=get_attr];
-"2393 sub_tensor_95" [id=2393, type=sub];
-"2394 linear_140" [id=2394, type=linear];
-"2395 dropout_91" [id=2395, type=dropout];
-"2396 _param_constant379" [id=2396, type=get_attr];
-"2397 _param_constant380" [id=2397, type=get_attr];
-"2398 layer_norm_49" [id=2398, type=layer_norm];
-"2399 add_79" [id=2399, type=add];
-"2400 _frozen_param191" [id=2400, type=get_attr];
-"2401 pad_26" [id=2401, type=pad];
-"2402 view_127" [id=2402, type=view];
-"2403 permute_105" [id=2403, type=permute];
-"2404 reshape_103" [id=2404, type=reshape];
-"2405 reshape_103_0_0_nncf_smooth_quant_0" [id=2405, type=call_module];
-"2406 quantize_per_tensor_default_142" [id=2406, type=quantize_per_tensor];
-"2407 dequantize_per_tensor_default_142" [id=2407, type=dequantize_per_tensor];
-"2408 _frozen_param192_0_0" [id=2408, type=get_attr];
-"2409 scale_updated_constant96" [id=2409, type=get_attr];
-"2410 compressed_weight_updated_constant96" [id=2410, type=get_attr];
-"2411 mul_tensor_96" [id=2411, type=mul];
-"2412 zero_point_updated_constant96" [id=2412, type=get_attr];
-"2413 sub_tensor_96" [id=2413, type=sub];
-"2414 linear_143" [id=2414, type=linear];
-"2415 reshape_104" [id=2415, type=reshape];
-"2416 permute_106" [id=2416, type=permute];
-"2417 select_69" [id=2417, type=select];
-"2418 select_70" [id=2418, type=select];
-"2419 select_71" [id=2419, type=select];
-"2420 linalg_vector_norm_46" [id=2420, type=linalg_vector_norm];
-"2421 clamp_min_46" [id=2421, type=clamp_min];
-"2422 expand_as_46" [id=2422, type=expand_as];
-"2423 div_46" [id=2423, type=div];
-"2424 quantize_per_tensor_default_143" [id=2424, type=quantize_per_tensor];
-"2425 dequantize_per_tensor_default_143" [id=2425, type=dequantize_per_tensor];
-"2426 linalg_vector_norm_47" [id=2426, type=linalg_vector_norm];
-"2427 clamp_min_47" [id=2427, type=clamp_min];
-"2428 expand_as_47" [id=2428, type=expand_as];
-"2429 div_47" [id=2429, type=div];
-"2430 quantize_per_tensor_default_144" [id=2430, type=quantize_per_tensor];
-"2431 dequantize_per_tensor_default_144" [id=2431, type=dequantize_per_tensor];
-"2432 transpose_46" [id=2432, type=transpose];
-"2433 matmul_46" [id=2433, type=matmul];
-"2434 _frozen_param194" [id=2434, type=get_attr];
-"2435 mul_47" [id=2435, type=mul];
-"2436 add_80" [id=2436, type=add];
-"2437 softmax_23" [id=2437, type=softmax];
-"2438 dropout_92" [id=2438, type=dropout];
-"2439 matmul_47" [id=2439, type=matmul];
-"2440 transpose_47" [id=2440, type=transpose];
-"2441 reshape_105" [id=2441, type=reshape];
-"2442 reshape_105_0_0_nncf_smooth_quant_0" [id=2442, type=call_module];
-"2443 quantize_per_tensor_default_145" [id=2443, type=quantize_per_tensor];
-"2444 dequantize_per_tensor_default_145" [id=2444, type=dequantize_per_tensor];
-"2445 _param_constant388_0_0" [id=2445, type=get_attr];
-"2446 scale_updated_constant97" [id=2446, type=get_attr];
-"2447 compressed_weight_updated_constant97" [id=2447, type=get_attr];
-"2448 mul_tensor_97" [id=2448, type=mul];
-"2449 zero_point_updated_constant97" [id=2449, type=get_attr];
-"2450 sub_tensor_97" [id=2450, type=sub];
-"2451 linear_144" [id=2451, type=linear];
-"2452 dropout_93" [id=2452, type=dropout];
-"2453 view_128" [id=2453, type=view];
-"2454 permute_107" [id=2454, type=permute];
-"2455 reshape_106" [id=2455, type=reshape];
-"2456 slice_347" [id=2456, type=slice];
-"2457 slice_348" [id=2457, type=slice];
-"2458 slice_349" [id=2458, type=slice];
-"2459 slice_350" [id=2459, type=slice];
-"2460 contiguous_45" [id=2460, type=contiguous];
-"2461 _param_constant389" [id=2461, type=get_attr];
-"2462 _param_constant390" [id=2462, type=get_attr];
-"2463 layer_norm_50" [id=2463, type=layer_norm];
-"2464 add_81" [id=2464, type=add];
-"2465 add_81_0_0_nncf_smooth_quant_0" [id=2465, type=call_module];
-"2466 quantize_per_tensor_default_146" [id=2466, type=quantize_per_tensor];
-"2467 dequantize_per_tensor_default_146" [id=2467, type=dequantize_per_tensor];
-"2468 _param_constant392_0_0" [id=2468, type=get_attr];
-"2469 scale_updated_constant98" [id=2469, type=get_attr];
-"2470 compressed_weight_updated_constant98" [id=2470, type=get_attr];
-"2471 mul_tensor_98" [id=2471, type=mul];
-"2472 zero_point_updated_constant98" [id=2472, type=get_attr];
-"2473 sub_tensor_98" [id=2473, type=sub];
-"2474 linear_145" [id=2474, type=linear];
-"2475 gelu_23" [id=2475, type=gelu];
-"2476 dropout_94" [id=2476, type=dropout];
-"2477 dropout_94_0_0_nncf_smooth_quant_0" [id=2477, type=call_module];
-"2478 quantize_per_tensor_default_147" [id=2478, type=quantize_per_tensor];
-"2479 dequantize_per_tensor_default_147" [id=2479, type=dequantize_per_tensor];
-"2480 _param_constant394_0_0" [id=2480, type=get_attr];
-"2481 scale_updated_constant99" [id=2481, type=get_attr];
-"2482 compressed_weight_updated_constant99" [id=2482, type=get_attr];
-"2483 mul_tensor_99" [id=2483, type=mul];
-"2484 zero_point_updated_constant99" [id=2484, type=get_attr];
-"2485 sub_tensor_99" [id=2485, type=sub];
-"2486 linear_146" [id=2486, type=linear];
-"2487 dropout_95" [id=2487, type=dropout];
-"2488 _param_constant395" [id=2488, type=get_attr];
-"2489 _param_constant396" [id=2489, type=get_attr];
-"2490 layer_norm_51" [id=2490, type=layer_norm];
-"2491 add_82" [id=2491, type=add];
-"2492 _param_constant397" [id=2492, type=get_attr];
-"2493 _param_constant398" [id=2493, type=get_attr];
+"0 features_0_2_weight" [id=0, type=get_attr];
+"1 features_0_2_bias" [id=1, type=get_attr];
+"2 features_1_0_norm1_weight" [id=2, type=get_attr];
+"3 features_1_0_norm1_bias" [id=3, type=get_attr];
+"4 features_1_0_norm2_weight" [id=4, type=get_attr];
+"5 features_1_0_norm2_bias" [id=5, type=get_attr];
+"6 features_1_1_norm1_weight" [id=6, type=get_attr];
+"7 features_1_1_norm1_bias" [id=7, type=get_attr];
+"8 features_1_1_norm2_weight" [id=8, type=get_attr];
+"9 features_1_1_norm2_bias" [id=9, type=get_attr];
+"10 features_2_norm_weight" [id=10, type=get_attr];
+"11 features_2_norm_bias" [id=11, type=get_attr];
+"12 features_3_0_norm1_weight" [id=12, type=get_attr];
+"13 features_3_0_norm1_bias" [id=13, type=get_attr];
+"14 features_3_0_norm2_weight" [id=14, type=get_attr];
+"15 features_3_0_norm2_bias" [id=15, type=get_attr];
+"16 features_3_1_norm1_weight" [id=16, type=get_attr];
+"17 features_3_1_norm1_bias" [id=17, type=get_attr];
+"18 features_3_1_norm2_weight" [id=18, type=get_attr];
+"19 features_3_1_norm2_bias" [id=19, type=get_attr];
+"20 features_4_norm_weight" [id=20, type=get_attr];
+"21 features_4_norm_bias" [id=21, type=get_attr];
+"22 features_5_0_norm1_weight" [id=22, type=get_attr];
+"23 features_5_0_norm1_bias" [id=23, type=get_attr];
+"24 features_5_0_norm2_weight" [id=24, type=get_attr];
+"25 features_5_0_norm2_bias" [id=25, type=get_attr];
+"26 features_5_1_norm1_weight" [id=26, type=get_attr];
+"27 features_5_1_norm1_bias" [id=27, type=get_attr];
+"28 features_5_1_norm2_weight" [id=28, type=get_attr];
+"29 features_5_1_norm2_bias" [id=29, type=get_attr];
+"30 features_5_2_norm1_weight" [id=30, type=get_attr];
+"31 features_5_2_norm1_bias" [id=31, type=get_attr];
+"32 features_5_2_norm2_weight" [id=32, type=get_attr];
+"33 features_5_2_norm2_bias" [id=33, type=get_attr];
+"34 features_5_3_norm1_weight" [id=34, type=get_attr];
+"35 features_5_3_norm1_bias" [id=35, type=get_attr];
+"36 features_5_3_norm2_weight" [id=36, type=get_attr];
+"37 features_5_3_norm2_bias" [id=37, type=get_attr];
+"38 features_5_4_norm1_weight" [id=38, type=get_attr];
+"39 features_5_4_norm1_bias" [id=39, type=get_attr];
+"40 features_5_4_norm2_weight" [id=40, type=get_attr];
+"41 features_5_4_norm2_bias" [id=41, type=get_attr];
+"42 features_5_5_norm1_weight" [id=42, type=get_attr];
+"43 features_5_5_norm1_bias" [id=43, type=get_attr];
+"44 features_5_5_norm2_weight" [id=44, type=get_attr];
+"45 features_5_5_norm2_bias" [id=45, type=get_attr];
+"46 features_5_6_norm1_weight" [id=46, type=get_attr];
+"47 features_5_6_norm1_bias" [id=47, type=get_attr];
+"48 features_5_6_norm2_weight" [id=48, type=get_attr];
+"49 features_5_6_norm2_bias" [id=49, type=get_attr];
+"50 features_5_7_norm1_weight" [id=50, type=get_attr];
+"51 features_5_7_norm1_bias" [id=51, type=get_attr];
+"52 features_5_7_norm2_weight" [id=52, type=get_attr];
+"53 features_5_7_norm2_bias" [id=53, type=get_attr];
+"54 features_5_8_norm1_weight" [id=54, type=get_attr];
+"55 features_5_8_norm1_bias" [id=55, type=get_attr];
+"56 features_5_8_norm2_weight" [id=56, type=get_attr];
+"57 features_5_8_norm2_bias" [id=57, type=get_attr];
+"58 features_5_9_norm1_weight" [id=58, type=get_attr];
+"59 features_5_9_norm1_bias" [id=59, type=get_attr];
+"60 features_5_9_norm2_weight" [id=60, type=get_attr];
+"61 features_5_9_norm2_bias" [id=61, type=get_attr];
+"62 features_5_10_norm1_weight" [id=62, type=get_attr];
+"63 features_5_10_norm1_bias" [id=63, type=get_attr];
+"64 features_5_10_norm2_weight" [id=64, type=get_attr];
+"65 features_5_10_norm2_bias" [id=65, type=get_attr];
+"66 features_5_11_norm1_weight" [id=66, type=get_attr];
+"67 features_5_11_norm1_bias" [id=67, type=get_attr];
+"68 features_5_11_norm2_weight" [id=68, type=get_attr];
+"69 features_5_11_norm2_bias" [id=69, type=get_attr];
+"70 features_5_12_norm1_weight" [id=70, type=get_attr];
+"71 features_5_12_norm1_bias" [id=71, type=get_attr];
+"72 features_5_12_norm2_weight" [id=72, type=get_attr];
+"73 features_5_12_norm2_bias" [id=73, type=get_attr];
+"74 features_5_13_norm1_weight" [id=74, type=get_attr];
+"75 features_5_13_norm1_bias" [id=75, type=get_attr];
+"76 features_5_13_norm2_weight" [id=76, type=get_attr];
+"77 features_5_13_norm2_bias" [id=77, type=get_attr];
+"78 features_5_14_norm1_weight" [id=78, type=get_attr];
+"79 features_5_14_norm1_bias" [id=79, type=get_attr];
+"80 features_5_14_norm2_weight" [id=80, type=get_attr];
+"81 features_5_14_norm2_bias" [id=81, type=get_attr];
+"82 features_5_15_norm1_weight" [id=82, type=get_attr];
+"83 features_5_15_norm1_bias" [id=83, type=get_attr];
+"84 features_5_15_norm2_weight" [id=84, type=get_attr];
+"85 features_5_15_norm2_bias" [id=85, type=get_attr];
+"86 features_5_16_norm1_weight" [id=86, type=get_attr];
+"87 features_5_16_norm1_bias" [id=87, type=get_attr];
+"88 features_5_16_norm2_weight" [id=88, type=get_attr];
+"89 features_5_16_norm2_bias" [id=89, type=get_attr];
+"90 features_5_17_norm1_weight" [id=90, type=get_attr];
+"91 features_5_17_norm1_bias" [id=91, type=get_attr];
+"92 features_5_17_norm2_weight" [id=92, type=get_attr];
+"93 features_5_17_norm2_bias" [id=93, type=get_attr];
+"94 features_6_norm_weight" [id=94, type=get_attr];
+"95 features_6_norm_bias" [id=95, type=get_attr];
+"96 features_7_0_norm1_weight" [id=96, type=get_attr];
+"97 features_7_0_norm1_bias" [id=97, type=get_attr];
+"98 features_7_0_norm2_weight" [id=98, type=get_attr];
+"99 features_7_0_norm2_bias" [id=99, type=get_attr];
+"100 features_7_1_norm1_weight" [id=100, type=get_attr];
+"101 features_7_1_norm1_bias" [id=101, type=get_attr];
+"102 features_7_1_norm2_weight" [id=102, type=get_attr];
+"103 features_7_1_norm2_bias" [id=103, type=get_attr];
+"104 norm_weight" [id=104, type=get_attr];
+"105 norm_bias" [id=105, type=get_attr];
+"106 x" [id=106, type=input];
+"107 quantize_per_tensor_default" [id=107, type=quantize_per_tensor];
+"108 dequantize_per_tensor_default" [id=108, type=dequantize_per_tensor];
+"109 features_0_0_bias_0_0" [id=109, type=get_attr];
+"110 scale_updated_constant0" [id=110, type=get_attr];
+"111 compressed_weight_updated_constant0" [id=111, type=get_attr];
+"112 mul_tensor" [id=112, type=mul];
+"113 zero_point_updated_constant0" [id=113, type=get_attr];
+"114 sub_tensor" [id=114, type=sub];
+"115 conv2d" [id=115, type=conv2d];
+"116 permute" [id=116, type=permute];
+"117 layer_norm" [id=117, type=layer_norm];
+"118 _frozen_param0" [id=118, type=get_attr];
+"119 pad" [id=119, type=pad];
+"120 view_2" [id=120, type=view];
+"121 permute_2" [id=121, type=permute];
+"122 reshape" [id=122, type=reshape];
+"123 reshape_0_0_nncf_smooth_quant_0" [id=123, type=call_module];
+"124 quantize_per_tensor_default_1" [id=124, type=quantize_per_tensor];
+"125 dequantize_per_tensor_default_1" [id=125, type=dequantize_per_tensor];
+"126 _frozen_param1_0_0" [id=126, type=get_attr];
+"127 scale_updated_constant1" [id=127, type=get_attr];
+"128 compressed_weight_updated_constant1" [id=128, type=get_attr];
+"129 mul_tensor_1" [id=129, type=mul];
+"130 zero_point_updated_constant1" [id=130, type=get_attr];
+"131 sub_tensor_1" [id=131, type=sub];
+"132 linear_2" [id=132, type=linear];
+"133 reshape_1" [id=133, type=reshape];
+"134 permute_3" [id=134, type=permute];
+"135 select" [id=135, type=select];
+"136 select_1" [id=136, type=select];
+"137 select_2" [id=137, type=select];
+"138 linalg_vector_norm" [id=138, type=linalg_vector_norm];
+"139 clamp_min" [id=139, type=clamp_min];
+"140 expand_as" [id=140, type=expand_as];
+"141 div" [id=141, type=div];
+"142 quantize_per_tensor_default_2" [id=142, type=quantize_per_tensor];
+"143 dequantize_per_tensor_default_2" [id=143, type=dequantize_per_tensor];
+"144 linalg_vector_norm_1" [id=144, type=linalg_vector_norm];
+"145 clamp_min_1" [id=145, type=clamp_min];
+"146 expand_as_1" [id=146, type=expand_as];
+"147 div_1" [id=147, type=div];
+"148 quantize_per_tensor_default_3" [id=148, type=quantize_per_tensor];
+"149 dequantize_per_tensor_default_3" [id=149, type=dequantize_per_tensor];
+"150 transpose" [id=150, type=transpose];
+"151 matmul" [id=151, type=matmul];
+"152 _frozen_param3" [id=152, type=get_attr];
+"153 mul_1" [id=153, type=mul];
+"154 add" [id=154, type=add];
+"155 softmax" [id=155, type=softmax];
+"156 dropout" [id=156, type=dropout];
+"157 matmul_1" [id=157, type=matmul];
+"158 transpose_1" [id=158, type=transpose];
+"159 reshape_2" [id=159, type=reshape];
+"160 reshape_2_0_0_nncf_smooth_quant_0" [id=160, type=call_module];
+"161 quantize_per_tensor_default_4" [id=161, type=quantize_per_tensor];
+"162 dequantize_per_tensor_default_4" [id=162, type=dequantize_per_tensor];
+"163 features_1_0_attn_proj_bias_0_0" [id=163, type=get_attr];
+"164 scale_updated_constant2" [id=164, type=get_attr];
+"165 compressed_weight_updated_constant2" [id=165, type=get_attr];
+"166 mul_tensor_2" [id=166, type=mul];
+"167 zero_point_updated_constant2" [id=167, type=get_attr];
+"168 sub_tensor_2" [id=168, type=sub];
+"169 linear_3" [id=169, type=linear];
+"170 dropout_1" [id=170, type=dropout];
+"171 view_3" [id=171, type=view];
+"172 permute_4" [id=172, type=permute];
+"173 reshape_3" [id=173, type=reshape];
+"174 slice_2" [id=174, type=slice];
+"175 slice_3" [id=175, type=slice];
+"176 layer_norm_1" [id=176, type=layer_norm];
+"177 add_1" [id=177, type=add];
+"178 add_1_0_0_nncf_smooth_quant_0" [id=178, type=call_module];
+"179 quantize_per_tensor_default_5" [id=179, type=quantize_per_tensor];
+"180 dequantize_per_tensor_default_5" [id=180, type=dequantize_per_tensor];
+"181 features_1_0_mlp_0_bias_0_0" [id=181, type=get_attr];
+"182 scale_updated_constant3" [id=182, type=get_attr];
+"183 compressed_weight_updated_constant3" [id=183, type=get_attr];
+"184 mul_tensor_3" [id=184, type=mul];
+"185 zero_point_updated_constant3" [id=185, type=get_attr];
+"186 sub_tensor_3" [id=186, type=sub];
+"187 linear_4" [id=187, type=linear];
+"188 gelu" [id=188, type=gelu];
+"189 dropout_2" [id=189, type=dropout];
+"190 dropout_2_0_0_nncf_smooth_quant_0" [id=190, type=call_module];
+"191 quantize_per_tensor_default_6" [id=191, type=quantize_per_tensor];
+"192 dequantize_per_tensor_default_6" [id=192, type=dequantize_per_tensor];
+"193 features_1_0_mlp_3_bias_0_0" [id=193, type=get_attr];
+"194 scale_updated_constant4" [id=194, type=get_attr];
+"195 compressed_weight_updated_constant4" [id=195, type=get_attr];
+"196 mul_tensor_4" [id=196, type=mul];
+"197 zero_point_updated_constant4" [id=197, type=get_attr];
+"198 sub_tensor_4" [id=198, type=sub];
+"199 linear_5" [id=199, type=linear];
+"200 dropout_3" [id=200, type=dropout];
+"201 layer_norm_2" [id=201, type=layer_norm];
+"202 add_2" [id=202, type=add];
+"203 _frozen_param4" [id=203, type=get_attr];
+"204 pad_1" [id=204, type=pad];
+"205 roll" [id=205, type=roll];
+"206 view_6" [id=206, type=view];
+"207 permute_6" [id=207, type=permute];
+"208 reshape_4" [id=208, type=reshape];
+"209 reshape_4_0_0_nncf_smooth_quant_0" [id=209, type=call_module];
+"210 quantize_per_tensor_default_7" [id=210, type=quantize_per_tensor];
+"211 dequantize_per_tensor_default_7" [id=211, type=dequantize_per_tensor];
+"212 _frozen_param5_0_0" [id=212, type=get_attr];
+"213 scale_updated_constant5" [id=213, type=get_attr];
+"214 compressed_weight_updated_constant5" [id=214, type=get_attr];
+"215 mul_tensor_5" [id=215, type=mul];
+"216 zero_point_updated_constant5" [id=216, type=get_attr];
+"217 sub_tensor_5" [id=217, type=sub];
+"218 linear_8" [id=218, type=linear];
+"219 reshape_5" [id=219, type=reshape];
+"220 permute_7" [id=220, type=permute];
+"221 select_3" [id=221, type=select];
+"222 select_4" [id=222, type=select];
+"223 select_5" [id=223, type=select];
+"224 linalg_vector_norm_2" [id=224, type=linalg_vector_norm];
+"225 clamp_min_2" [id=225, type=clamp_min];
+"226 expand_as_2" [id=226, type=expand_as];
+"227 div_2" [id=227, type=div];
+"228 quantize_per_tensor_default_8" [id=228, type=quantize_per_tensor];
+"229 dequantize_per_tensor_default_8" [id=229, type=dequantize_per_tensor];
+"230 linalg_vector_norm_3" [id=230, type=linalg_vector_norm];
+"231 clamp_min_3" [id=231, type=clamp_min];
+"232 expand_as_3" [id=232, type=expand_as];
+"233 div_3" [id=233, type=div];
+"234 quantize_per_tensor_default_9" [id=234, type=quantize_per_tensor];
+"235 dequantize_per_tensor_default_9" [id=235, type=dequantize_per_tensor];
+"236 transpose_2" [id=236, type=transpose];
+"237 matmul_2" [id=237, type=matmul];
+"238 _frozen_param7" [id=238, type=get_attr];
+"239 mul_3" [id=239, type=mul];
+"240 add_3" [id=240, type=add];
+"241 new_zeros" [id=241, type=new_zeros];
+"242 view_7" [id=242, type=view];
+"243 permute_8" [id=243, type=permute];
+"244 reshape_6" [id=244, type=reshape];
+"245 unsqueeze_2" [id=245, type=unsqueeze];
+"246 unsqueeze_3" [id=246, type=unsqueeze];
+"247 sub" [id=247, type=sub];
+"248 ne" [id=248, type=ne];
+"249 masked_fill" [id=249, type=masked_fill];
+"250 eq" [id=250, type=eq];
+"251 masked_fill_1" [id=251, type=masked_fill];
+"252 view_8" [id=252, type=view];
+"253 unsqueeze_4" [id=253, type=unsqueeze];
+"254 unsqueeze_5" [id=254, type=unsqueeze];
+"255 add_4" [id=255, type=add];
+"256 view_9" [id=256, type=view];
+"257 softmax_1" [id=257, type=softmax];
+"258 dropout_4" [id=258, type=dropout];
+"259 matmul_3" [id=259, type=matmul];
+"260 transpose_3" [id=260, type=transpose];
+"261 reshape_7" [id=261, type=reshape];
+"262 reshape_7_0_0_nncf_smooth_quant_0" [id=262, type=call_module];
+"263 quantize_per_tensor_default_10" [id=263, type=quantize_per_tensor];
+"264 dequantize_per_tensor_default_10" [id=264, type=dequantize_per_tensor];
+"265 features_1_1_attn_proj_bias_0_0" [id=265, type=get_attr];
+"266 scale_updated_constant6" [id=266, type=get_attr];
+"267 compressed_weight_updated_constant6" [id=267, type=get_attr];
+"268 mul_tensor_6" [id=268, type=mul];
+"269 zero_point_updated_constant6" [id=269, type=get_attr];
+"270 sub_tensor_6" [id=270, type=sub];
+"271 linear_9" [id=271, type=linear];
+"272 dropout_5" [id=272, type=dropout];
+"273 view_10" [id=273, type=view];
+"274 permute_9" [id=274, type=permute];
+"275 reshape_8" [id=275, type=reshape];
+"276 roll_1" [id=276, type=roll];
+"277 slice_23" [id=277, type=slice];
+"278 slice_24" [id=278, type=slice];
+"279 layer_norm_3" [id=279, type=layer_norm];
+"280 add_5" [id=280, type=add];
+"281 add_5_0_0_nncf_smooth_quant_0" [id=281, type=call_module];
+"282 quantize_per_tensor_default_11" [id=282, type=quantize_per_tensor];
+"283 dequantize_per_tensor_default_11" [id=283, type=dequantize_per_tensor];
+"284 features_1_1_mlp_0_bias_0_0" [id=284, type=get_attr];
+"285 scale_updated_constant7" [id=285, type=get_attr];
+"286 compressed_weight_updated_constant7" [id=286, type=get_attr];
+"287 mul_tensor_7" [id=287, type=mul];
+"288 zero_point_updated_constant7" [id=288, type=get_attr];
+"289 sub_tensor_7" [id=289, type=sub];
+"290 linear_10" [id=290, type=linear];
+"291 gelu_1" [id=291, type=gelu];
+"292 dropout_6" [id=292, type=dropout];
+"293 dropout_6_0_0_nncf_smooth_quant_0" [id=293, type=call_module];
+"294 quantize_per_tensor_default_12" [id=294, type=quantize_per_tensor];
+"295 dequantize_per_tensor_default_12" [id=295, type=dequantize_per_tensor];
+"296 features_1_1_mlp_3_bias_0_0" [id=296, type=get_attr];
+"297 scale_updated_constant8" [id=297, type=get_attr];
+"298 compressed_weight_updated_constant8" [id=298, type=get_attr];
+"299 mul_tensor_8" [id=299, type=mul];
+"300 zero_point_updated_constant8" [id=300, type=get_attr];
+"301 sub_tensor_8" [id=301, type=sub];
+"302 linear_11" [id=302, type=linear];
+"303 dropout_7" [id=303, type=dropout];
+"304 layer_norm_4" [id=304, type=layer_norm];
+"305 add_6" [id=305, type=add];
+"306 pad_2" [id=306, type=pad];
+"307 slice_25" [id=307, type=slice];
+"308 slice_26" [id=308, type=slice];
+"309 slice_27" [id=309, type=slice];
+"310 slice_28" [id=310, type=slice];
+"311 slice_29" [id=311, type=slice];
+"312 slice_30" [id=312, type=slice];
+"313 slice_31" [id=313, type=slice];
+"314 slice_32" [id=314, type=slice];
+"315 slice_33" [id=315, type=slice];
+"316 slice_34" [id=316, type=slice];
+"317 slice_35" [id=317, type=slice];
+"318 slice_36" [id=318, type=slice];
+"319 cat" [id=319, type=cat];
+"320 cat_0_0_nncf_smooth_quant_0" [id=320, type=call_module];
+"321 quantize_per_tensor_default_13" [id=321, type=quantize_per_tensor];
+"322 dequantize_per_tensor_default_13" [id=322, type=dequantize_per_tensor];
+"323 scale_updated_constant9" [id=323, type=get_attr];
+"324 compressed_weight_updated_constant9" [id=324, type=get_attr];
+"325 mul_tensor_9" [id=325, type=mul];
+"326 zero_point_updated_constant9" [id=326, type=get_attr];
+"327 sub_tensor_9" [id=327, type=sub];
+"328 linear_12" [id=328, type=linear];
+"329 layer_norm_5" [id=329, type=layer_norm];
+"330 _frozen_param17" [id=330, type=get_attr];
+"331 pad_3" [id=331, type=pad];
+"332 view_13" [id=332, type=view];
+"333 permute_11" [id=333, type=permute];
+"334 reshape_9" [id=334, type=reshape];
+"335 reshape_9_0_0_nncf_smooth_quant_0" [id=335, type=call_module];
+"336 quantize_per_tensor_default_14" [id=336, type=quantize_per_tensor];
+"337 dequantize_per_tensor_default_14" [id=337, type=dequantize_per_tensor];
+"338 _frozen_param18_0_0" [id=338, type=get_attr];
+"339 scale_updated_constant10" [id=339, type=get_attr];
+"340 compressed_weight_updated_constant10" [id=340, type=get_attr];
+"341 mul_tensor_10" [id=341, type=mul];
+"342 zero_point_updated_constant10" [id=342, type=get_attr];
+"343 sub_tensor_10" [id=343, type=sub];
+"344 linear_15" [id=344, type=linear];
+"345 reshape_10" [id=345, type=reshape];
+"346 permute_12" [id=346, type=permute];
+"347 select_6" [id=347, type=select];
+"348 select_7" [id=348, type=select];
+"349 select_8" [id=349, type=select];
+"350 linalg_vector_norm_4" [id=350, type=linalg_vector_norm];
+"351 clamp_min_4" [id=351, type=clamp_min];
+"352 expand_as_4" [id=352, type=expand_as];
+"353 div_4" [id=353, type=div];
+"354 quantize_per_tensor_default_15" [id=354, type=quantize_per_tensor];
+"355 dequantize_per_tensor_default_15" [id=355, type=dequantize_per_tensor];
+"356 linalg_vector_norm_5" [id=356, type=linalg_vector_norm];
+"357 clamp_min_5" [id=357, type=clamp_min];
+"358 expand_as_5" [id=358, type=expand_as];
+"359 div_5" [id=359, type=div];
+"360 quantize_per_tensor_default_16" [id=360, type=quantize_per_tensor];
+"361 dequantize_per_tensor_default_16" [id=361, type=dequantize_per_tensor];
+"362 transpose_4" [id=362, type=transpose];
+"363 matmul_4" [id=363, type=matmul];
+"364 _frozen_param20" [id=364, type=get_attr];
+"365 mul_5" [id=365, type=mul];
+"366 add_7" [id=366, type=add];
+"367 softmax_2" [id=367, type=softmax];
+"368 dropout_8" [id=368, type=dropout];
+"369 matmul_5" [id=369, type=matmul];
+"370 transpose_5" [id=370, type=transpose];
+"371 reshape_11" [id=371, type=reshape];
+"372 reshape_11_0_0_nncf_smooth_quant_0" [id=372, type=call_module];
+"373 quantize_per_tensor_default_17" [id=373, type=quantize_per_tensor];
+"374 dequantize_per_tensor_default_17" [id=374, type=dequantize_per_tensor];
+"375 features_3_0_attn_proj_bias_0_0" [id=375, type=get_attr];
+"376 scale_updated_constant11" [id=376, type=get_attr];
+"377 compressed_weight_updated_constant11" [id=377, type=get_attr];
+"378 mul_tensor_11" [id=378, type=mul];
+"379 zero_point_updated_constant11" [id=379, type=get_attr];
+"380 sub_tensor_11" [id=380, type=sub];
+"381 linear_16" [id=381, type=linear];
+"382 dropout_9" [id=382, type=dropout];
+"383 view_14" [id=383, type=view];
+"384 permute_13" [id=384, type=permute];
+"385 reshape_12" [id=385, type=reshape];
+"386 slice_38" [id=386, type=slice];
+"387 slice_39" [id=387, type=slice];
+"388 slice_40" [id=388, type=slice];
+"389 slice_41" [id=389, type=slice];
+"390 contiguous_3" [id=390, type=contiguous];
+"391 layer_norm_6" [id=391, type=layer_norm];
+"392 add_8" [id=392, type=add];
+"393 add_8_0_0_nncf_smooth_quant_0" [id=393, type=call_module];
+"394 quantize_per_tensor_default_18" [id=394, type=quantize_per_tensor];
+"395 dequantize_per_tensor_default_18" [id=395, type=dequantize_per_tensor];
+"396 features_3_0_mlp_0_bias_0_0" [id=396, type=get_attr];
+"397 scale_updated_constant12" [id=397, type=get_attr];
+"398 compressed_weight_updated_constant12" [id=398, type=get_attr];
+"399 mul_tensor_12" [id=399, type=mul];
+"400 zero_point_updated_constant12" [id=400, type=get_attr];
+"401 sub_tensor_12" [id=401, type=sub];
+"402 linear_17" [id=402, type=linear];
+"403 gelu_2" [id=403, type=gelu];
+"404 dropout_10" [id=404, type=dropout];
+"405 dropout_10_0_0_nncf_smooth_quant_0" [id=405, type=call_module];
+"406 quantize_per_tensor_default_19" [id=406, type=quantize_per_tensor];
+"407 dequantize_per_tensor_default_19" [id=407, type=dequantize_per_tensor];
+"408 features_3_0_mlp_3_bias_0_0" [id=408, type=get_attr];
+"409 scale_updated_constant13" [id=409, type=get_attr];
+"410 compressed_weight_updated_constant13" [id=410, type=get_attr];
+"411 mul_tensor_13" [id=411, type=mul];
+"412 zero_point_updated_constant13" [id=412, type=get_attr];
+"413 sub_tensor_13" [id=413, type=sub];
+"414 linear_18" [id=414, type=linear];
+"415 dropout_11" [id=415, type=dropout];
+"416 layer_norm_7" [id=416, type=layer_norm];
+"417 add_9" [id=417, type=add];
+"418 _frozen_param21" [id=418, type=get_attr];
+"419 pad_4" [id=419, type=pad];
+"420 roll_2" [id=420, type=roll];
+"421 view_17" [id=421, type=view];
+"422 permute_15" [id=422, type=permute];
+"423 reshape_13" [id=423, type=reshape];
+"424 reshape_13_0_0_nncf_smooth_quant_0" [id=424, type=call_module];
+"425 quantize_per_tensor_default_20" [id=425, type=quantize_per_tensor];
+"426 dequantize_per_tensor_default_20" [id=426, type=dequantize_per_tensor];
+"427 _frozen_param22_0_0" [id=427, type=get_attr];
+"428 scale_updated_constant14" [id=428, type=get_attr];
+"429 compressed_weight_updated_constant14" [id=429, type=get_attr];
+"430 mul_tensor_14" [id=430, type=mul];
+"431 zero_point_updated_constant14" [id=431, type=get_attr];
+"432 sub_tensor_14" [id=432, type=sub];
+"433 linear_21" [id=433, type=linear];
+"434 reshape_14" [id=434, type=reshape];
+"435 permute_16" [id=435, type=permute];
+"436 select_9" [id=436, type=select];
+"437 select_10" [id=437, type=select];
+"438 select_11" [id=438, type=select];
+"439 linalg_vector_norm_6" [id=439, type=linalg_vector_norm];
+"440 clamp_min_6" [id=440, type=clamp_min];
+"441 expand_as_6" [id=441, type=expand_as];
+"442 div_6" [id=442, type=div];
+"443 quantize_per_tensor_default_21" [id=443, type=quantize_per_tensor];
+"444 dequantize_per_tensor_default_21" [id=444, type=dequantize_per_tensor];
+"445 linalg_vector_norm_7" [id=445, type=linalg_vector_norm];
+"446 clamp_min_7" [id=446, type=clamp_min];
+"447 expand_as_7" [id=447, type=expand_as];
+"448 div_7" [id=448, type=div];
+"449 quantize_per_tensor_default_22" [id=449, type=quantize_per_tensor];
+"450 dequantize_per_tensor_default_22" [id=450, type=dequantize_per_tensor];
+"451 transpose_6" [id=451, type=transpose];
+"452 matmul_6" [id=452, type=matmul];
+"453 _frozen_param24" [id=453, type=get_attr];
+"454 mul_7" [id=454, type=mul];
+"455 add_10" [id=455, type=add];
+"456 new_zeros_1" [id=456, type=new_zeros];
+"457 view_18" [id=457, type=view];
+"458 permute_17" [id=458, type=permute];
+"459 reshape_15" [id=459, type=reshape];
+"460 unsqueeze_8" [id=460, type=unsqueeze];
+"461 unsqueeze_9" [id=461, type=unsqueeze];
+"462 sub_1" [id=462, type=sub];
+"463 ne_1" [id=463, type=ne];
+"464 masked_fill_2" [id=464, type=masked_fill];
+"465 eq_1" [id=465, type=eq];
+"466 masked_fill_3" [id=466, type=masked_fill];
+"467 view_19" [id=467, type=view];
+"468 unsqueeze_10" [id=468, type=unsqueeze];
+"469 unsqueeze_11" [id=469, type=unsqueeze];
+"470 add_11" [id=470, type=add];
+"471 view_20" [id=471, type=view];
+"472 softmax_3" [id=472, type=softmax];
+"473 dropout_12" [id=473, type=dropout];
+"474 matmul_7" [id=474, type=matmul];
+"475 transpose_7" [id=475, type=transpose];
+"476 reshape_16" [id=476, type=reshape];
+"477 reshape_16_0_0_nncf_smooth_quant_0" [id=477, type=call_module];
+"478 quantize_per_tensor_default_23" [id=478, type=quantize_per_tensor];
+"479 dequantize_per_tensor_default_23" [id=479, type=dequantize_per_tensor];
+"480 features_3_1_attn_proj_bias_0_0" [id=480, type=get_attr];
+"481 scale_updated_constant15" [id=481, type=get_attr];
+"482 compressed_weight_updated_constant15" [id=482, type=get_attr];
+"483 mul_tensor_15" [id=483, type=mul];
+"484 zero_point_updated_constant15" [id=484, type=get_attr];
+"485 sub_tensor_15" [id=485, type=sub];
+"486 linear_22" [id=486, type=linear];
+"487 dropout_13" [id=487, type=dropout];
+"488 view_21" [id=488, type=view];
+"489 permute_18" [id=489, type=permute];
+"490 reshape_17" [id=490, type=reshape];
+"491 roll_3" [id=491, type=roll];
+"492 slice_61" [id=492, type=slice];
+"493 slice_62" [id=493, type=slice];
+"494 slice_63" [id=494, type=slice];
+"495 slice_64" [id=495, type=slice];
+"496 contiguous_5" [id=496, type=contiguous];
+"497 layer_norm_8" [id=497, type=layer_norm];
+"498 add_12" [id=498, type=add];
+"499 add_12_0_0_nncf_smooth_quant_0" [id=499, type=call_module];
+"500 quantize_per_tensor_default_24" [id=500, type=quantize_per_tensor];
+"501 dequantize_per_tensor_default_24" [id=501, type=dequantize_per_tensor];
+"502 features_3_1_mlp_0_bias_0_0" [id=502, type=get_attr];
+"503 scale_updated_constant16" [id=503, type=get_attr];
+"504 compressed_weight_updated_constant16" [id=504, type=get_attr];
+"505 mul_tensor_16" [id=505, type=mul];
+"506 zero_point_updated_constant16" [id=506, type=get_attr];
+"507 sub_tensor_16" [id=507, type=sub];
+"508 linear_23" [id=508, type=linear];
+"509 gelu_3" [id=509, type=gelu];
+"510 dropout_14" [id=510, type=dropout];
+"511 dropout_14_0_0_nncf_smooth_quant_0" [id=511, type=call_module];
+"512 quantize_per_tensor_default_25" [id=512, type=quantize_per_tensor];
+"513 dequantize_per_tensor_default_25" [id=513, type=dequantize_per_tensor];
+"514 features_3_1_mlp_3_bias_0_0" [id=514, type=get_attr];
+"515 scale_updated_constant17" [id=515, type=get_attr];
+"516 compressed_weight_updated_constant17" [id=516, type=get_attr];
+"517 mul_tensor_17" [id=517, type=mul];
+"518 zero_point_updated_constant17" [id=518, type=get_attr];
+"519 sub_tensor_17" [id=519, type=sub];
+"520 linear_24" [id=520, type=linear];
+"521 dropout_15" [id=521, type=dropout];
+"522 layer_norm_9" [id=522, type=layer_norm];
+"523 add_13" [id=523, type=add];
+"524 pad_5" [id=524, type=pad];
+"525 slice_65" [id=525, type=slice];
+"526 slice_66" [id=526, type=slice];
+"527 slice_67" [id=527, type=slice];
+"528 slice_68" [id=528, type=slice];
+"529 slice_69" [id=529, type=slice];
+"530 slice_70" [id=530, type=slice];
+"531 slice_71" [id=531, type=slice];
+"532 slice_72" [id=532, type=slice];
+"533 slice_73" [id=533, type=slice];
+"534 slice_74" [id=534, type=slice];
+"535 slice_75" [id=535, type=slice];
+"536 slice_76" [id=536, type=slice];
+"537 cat_1" [id=537, type=cat];
+"538 cat_1_0_0_nncf_smooth_quant_0" [id=538, type=call_module];
+"539 quantize_per_tensor_default_26" [id=539, type=quantize_per_tensor];
+"540 dequantize_per_tensor_default_26" [id=540, type=dequantize_per_tensor];
+"541 scale_updated_constant18" [id=541, type=get_attr];
+"542 compressed_weight_updated_constant18" [id=542, type=get_attr];
+"543 mul_tensor_18" [id=543, type=mul];
+"544 zero_point_updated_constant18" [id=544, type=get_attr];
+"545 sub_tensor_18" [id=545, type=sub];
+"546 linear_25" [id=546, type=linear];
+"547 layer_norm_10" [id=547, type=layer_norm];
+"548 _frozen_param34" [id=548, type=get_attr];
+"549 pad_6" [id=549, type=pad];
+"550 view_24" [id=550, type=view];
+"551 permute_20" [id=551, type=permute];
+"552 reshape_18" [id=552, type=reshape];
+"553 reshape_18_0_0_nncf_smooth_quant_0" [id=553, type=call_module];
+"554 quantize_per_tensor_default_27" [id=554, type=quantize_per_tensor];
+"555 dequantize_per_tensor_default_27" [id=555, type=dequantize_per_tensor];
+"556 _frozen_param35_0_0" [id=556, type=get_attr];
+"557 scale_updated_constant19" [id=557, type=get_attr];
+"558 compressed_weight_updated_constant19" [id=558, type=get_attr];
+"559 mul_tensor_19" [id=559, type=mul];
+"560 zero_point_updated_constant19" [id=560, type=get_attr];
+"561 sub_tensor_19" [id=561, type=sub];
+"562 linear_28" [id=562, type=linear];
+"563 reshape_19" [id=563, type=reshape];
+"564 permute_21" [id=564, type=permute];
+"565 select_12" [id=565, type=select];
+"566 select_13" [id=566, type=select];
+"567 select_14" [id=567, type=select];
+"568 linalg_vector_norm_8" [id=568, type=linalg_vector_norm];
+"569 clamp_min_8" [id=569, type=clamp_min];
+"570 expand_as_8" [id=570, type=expand_as];
+"571 div_8" [id=571, type=div];
+"572 quantize_per_tensor_default_28" [id=572, type=quantize_per_tensor];
+"573 dequantize_per_tensor_default_28" [id=573, type=dequantize_per_tensor];
+"574 linalg_vector_norm_9" [id=574, type=linalg_vector_norm];
+"575 clamp_min_9" [id=575, type=clamp_min];
+"576 expand_as_9" [id=576, type=expand_as];
+"577 div_9" [id=577, type=div];
+"578 quantize_per_tensor_default_29" [id=578, type=quantize_per_tensor];
+"579 dequantize_per_tensor_default_29" [id=579, type=dequantize_per_tensor];
+"580 transpose_8" [id=580, type=transpose];
+"581 matmul_8" [id=581, type=matmul];
+"582 _frozen_param37" [id=582, type=get_attr];
+"583 mul_9" [id=583, type=mul];
+"584 add_14" [id=584, type=add];
+"585 softmax_4" [id=585, type=softmax];
+"586 dropout_16" [id=586, type=dropout];
+"587 matmul_9" [id=587, type=matmul];
+"588 transpose_9" [id=588, type=transpose];
+"589 reshape_20" [id=589, type=reshape];
+"590 reshape_20_0_0_nncf_smooth_quant_0" [id=590, type=call_module];
+"591 quantize_per_tensor_default_30" [id=591, type=quantize_per_tensor];
+"592 dequantize_per_tensor_default_30" [id=592, type=dequantize_per_tensor];
+"593 features_5_0_attn_proj_bias_0_0" [id=593, type=get_attr];
+"594 scale_updated_constant20" [id=594, type=get_attr];
+"595 compressed_weight_updated_constant20" [id=595, type=get_attr];
+"596 mul_tensor_20" [id=596, type=mul];
+"597 zero_point_updated_constant20" [id=597, type=get_attr];
+"598 sub_tensor_20" [id=598, type=sub];
+"599 linear_29" [id=599, type=linear];
+"600 dropout_17" [id=600, type=dropout];
+"601 view_25" [id=601, type=view];
+"602 permute_22" [id=602, type=permute];
+"603 reshape_21" [id=603, type=reshape];
+"604 slice_78" [id=604, type=slice];
+"605 slice_79" [id=605, type=slice];
+"606 slice_80" [id=606, type=slice];
+"607 slice_81" [id=607, type=slice];
+"608 contiguous_7" [id=608, type=contiguous];
+"609 layer_norm_11" [id=609, type=layer_norm];
+"610 add_15" [id=610, type=add];
+"611 add_15_0_0_nncf_smooth_quant_0" [id=611, type=call_module];
+"612 quantize_per_tensor_default_31" [id=612, type=quantize_per_tensor];
+"613 dequantize_per_tensor_default_31" [id=613, type=dequantize_per_tensor];
+"614 features_5_0_mlp_0_bias_0_0" [id=614, type=get_attr];
+"615 scale_updated_constant21" [id=615, type=get_attr];
+"616 compressed_weight_updated_constant21" [id=616, type=get_attr];
+"617 mul_tensor_21" [id=617, type=mul];
+"618 zero_point_updated_constant21" [id=618, type=get_attr];
+"619 sub_tensor_21" [id=619, type=sub];
+"620 linear_30" [id=620, type=linear];
+"621 gelu_4" [id=621, type=gelu];
+"622 dropout_18" [id=622, type=dropout];
+"623 dropout_18_0_0_nncf_smooth_quant_0" [id=623, type=call_module];
+"624 quantize_per_tensor_default_32" [id=624, type=quantize_per_tensor];
+"625 dequantize_per_tensor_default_32" [id=625, type=dequantize_per_tensor];
+"626 features_5_0_mlp_3_bias_0_0" [id=626, type=get_attr];
+"627 scale_updated_constant22" [id=627, type=get_attr];
+"628 compressed_weight_updated_constant22" [id=628, type=get_attr];
+"629 mul_tensor_22" [id=629, type=mul];
+"630 zero_point_updated_constant22" [id=630, type=get_attr];
+"631 sub_tensor_22" [id=631, type=sub];
+"632 linear_31" [id=632, type=linear];
+"633 dropout_19" [id=633, type=dropout];
+"634 layer_norm_12" [id=634, type=layer_norm];
+"635 add_16" [id=635, type=add];
+"636 _frozen_param38" [id=636, type=get_attr];
+"637 pad_7" [id=637, type=pad];
+"638 roll_4" [id=638, type=roll];
+"639 view_28" [id=639, type=view];
+"640 permute_24" [id=640, type=permute];
+"641 reshape_22" [id=641, type=reshape];
+"642 reshape_22_0_0_nncf_smooth_quant_0" [id=642, type=call_module];
+"643 quantize_per_tensor_default_33" [id=643, type=quantize_per_tensor];
+"644 dequantize_per_tensor_default_33" [id=644, type=dequantize_per_tensor];
+"645 _frozen_param39_0_0" [id=645, type=get_attr];
+"646 scale_updated_constant23" [id=646, type=get_attr];
+"647 compressed_weight_updated_constant23" [id=647, type=get_attr];
+"648 mul_tensor_23" [id=648, type=mul];
+"649 zero_point_updated_constant23" [id=649, type=get_attr];
+"650 sub_tensor_23" [id=650, type=sub];
+"651 linear_34" [id=651, type=linear];
+"652 reshape_23" [id=652, type=reshape];
+"653 permute_25" [id=653, type=permute];
+"654 select_15" [id=654, type=select];
+"655 select_16" [id=655, type=select];
+"656 select_17" [id=656, type=select];
+"657 linalg_vector_norm_10" [id=657, type=linalg_vector_norm];
+"658 clamp_min_10" [id=658, type=clamp_min];
+"659 expand_as_10" [id=659, type=expand_as];
+"660 div_10" [id=660, type=div];
+"661 quantize_per_tensor_default_34" [id=661, type=quantize_per_tensor];
+"662 dequantize_per_tensor_default_34" [id=662, type=dequantize_per_tensor];
+"663 linalg_vector_norm_11" [id=663, type=linalg_vector_norm];
+"664 clamp_min_11" [id=664, type=clamp_min];
+"665 expand_as_11" [id=665, type=expand_as];
+"666 div_11" [id=666, type=div];
+"667 quantize_per_tensor_default_35" [id=667, type=quantize_per_tensor];
+"668 dequantize_per_tensor_default_35" [id=668, type=dequantize_per_tensor];
+"669 transpose_10" [id=669, type=transpose];
+"670 matmul_10" [id=670, type=matmul];
+"671 _frozen_param41" [id=671, type=get_attr];
+"672 mul_11" [id=672, type=mul];
+"673 add_17" [id=673, type=add];
+"674 new_zeros_2" [id=674, type=new_zeros];
+"675 view_29" [id=675, type=view];
+"676 permute_26" [id=676, type=permute];
+"677 reshape_24" [id=677, type=reshape];
+"678 unsqueeze_14" [id=678, type=unsqueeze];
+"679 unsqueeze_15" [id=679, type=unsqueeze];
+"680 sub_2" [id=680, type=sub];
+"681 ne_2" [id=681, type=ne];
+"682 masked_fill_4" [id=682, type=masked_fill];
+"683 eq_2" [id=683, type=eq];
+"684 masked_fill_5" [id=684, type=masked_fill];
+"685 view_30" [id=685, type=view];
+"686 unsqueeze_16" [id=686, type=unsqueeze];
+"687 unsqueeze_17" [id=687, type=unsqueeze];
+"688 add_18" [id=688, type=add];
+"689 view_31" [id=689, type=view];
+"690 softmax_5" [id=690, type=softmax];
+"691 dropout_20" [id=691, type=dropout];
+"692 matmul_11" [id=692, type=matmul];
+"693 transpose_11" [id=693, type=transpose];
+"694 reshape_25" [id=694, type=reshape];
+"695 reshape_25_0_0_nncf_smooth_quant_0" [id=695, type=call_module];
+"696 quantize_per_tensor_default_36" [id=696, type=quantize_per_tensor];
+"697 dequantize_per_tensor_default_36" [id=697, type=dequantize_per_tensor];
+"698 features_5_1_attn_proj_bias_0_0" [id=698, type=get_attr];
+"699 scale_updated_constant24" [id=699, type=get_attr];
+"700 compressed_weight_updated_constant24" [id=700, type=get_attr];
+"701 mul_tensor_24" [id=701, type=mul];
+"702 zero_point_updated_constant24" [id=702, type=get_attr];
+"703 sub_tensor_24" [id=703, type=sub];
+"704 linear_35" [id=704, type=linear];
+"705 dropout_21" [id=705, type=dropout];
+"706 view_32" [id=706, type=view];
+"707 permute_27" [id=707, type=permute];
+"708 reshape_26" [id=708, type=reshape];
+"709 roll_5" [id=709, type=roll];
+"710 slice_101" [id=710, type=slice];
+"711 slice_102" [id=711, type=slice];
+"712 slice_103" [id=712, type=slice];
+"713 slice_104" [id=713, type=slice];
+"714 contiguous_9" [id=714, type=contiguous];
+"715 layer_norm_13" [id=715, type=layer_norm];
+"716 add_19" [id=716, type=add];
+"717 add_19_0_0_nncf_smooth_quant_0" [id=717, type=call_module];
+"718 quantize_per_tensor_default_37" [id=718, type=quantize_per_tensor];
+"719 dequantize_per_tensor_default_37" [id=719, type=dequantize_per_tensor];
+"720 features_5_1_mlp_0_bias_0_0" [id=720, type=get_attr];
+"721 scale_updated_constant25" [id=721, type=get_attr];
+"722 compressed_weight_updated_constant25" [id=722, type=get_attr];
+"723 mul_tensor_25" [id=723, type=mul];
+"724 zero_point_updated_constant25" [id=724, type=get_attr];
+"725 sub_tensor_25" [id=725, type=sub];
+"726 linear_36" [id=726, type=linear];
+"727 gelu_5" [id=727, type=gelu];
+"728 dropout_22" [id=728, type=dropout];
+"729 dropout_22_0_0_nncf_smooth_quant_0" [id=729, type=call_module];
+"730 quantize_per_tensor_default_38" [id=730, type=quantize_per_tensor];
+"731 dequantize_per_tensor_default_38" [id=731, type=dequantize_per_tensor];
+"732 features_5_1_mlp_3_bias_0_0" [id=732, type=get_attr];
+"733 scale_updated_constant26" [id=733, type=get_attr];
+"734 compressed_weight_updated_constant26" [id=734, type=get_attr];
+"735 mul_tensor_26" [id=735, type=mul];
+"736 zero_point_updated_constant26" [id=736, type=get_attr];
+"737 sub_tensor_26" [id=737, type=sub];
+"738 linear_37" [id=738, type=linear];
+"739 dropout_23" [id=739, type=dropout];
+"740 layer_norm_14" [id=740, type=layer_norm];
+"741 add_20" [id=741, type=add];
+"742 _frozen_param51" [id=742, type=get_attr];
+"743 pad_8" [id=743, type=pad];
+"744 view_35" [id=744, type=view];
+"745 permute_29" [id=745, type=permute];
+"746 reshape_27" [id=746, type=reshape];
+"747 reshape_27_0_0_nncf_smooth_quant_0" [id=747, type=call_module];
+"748 quantize_per_tensor_default_39" [id=748, type=quantize_per_tensor];
+"749 dequantize_per_tensor_default_39" [id=749, type=dequantize_per_tensor];
+"750 _frozen_param52_0_0" [id=750, type=get_attr];
+"751 scale_updated_constant27" [id=751, type=get_attr];
+"752 compressed_weight_updated_constant27" [id=752, type=get_attr];
+"753 mul_tensor_27" [id=753, type=mul];
+"754 zero_point_updated_constant27" [id=754, type=get_attr];
+"755 sub_tensor_27" [id=755, type=sub];
+"756 linear_40" [id=756, type=linear];
+"757 reshape_28" [id=757, type=reshape];
+"758 permute_30" [id=758, type=permute];
+"759 select_18" [id=759, type=select];
+"760 select_19" [id=760, type=select];
+"761 select_20" [id=761, type=select];
+"762 linalg_vector_norm_12" [id=762, type=linalg_vector_norm];
+"763 clamp_min_12" [id=763, type=clamp_min];
+"764 expand_as_12" [id=764, type=expand_as];
+"765 div_12" [id=765, type=div];
+"766 quantize_per_tensor_default_40" [id=766, type=quantize_per_tensor];
+"767 dequantize_per_tensor_default_40" [id=767, type=dequantize_per_tensor];
+"768 linalg_vector_norm_13" [id=768, type=linalg_vector_norm];
+"769 clamp_min_13" [id=769, type=clamp_min];
+"770 expand_as_13" [id=770, type=expand_as];
+"771 div_13" [id=771, type=div];
+"772 quantize_per_tensor_default_41" [id=772, type=quantize_per_tensor];
+"773 dequantize_per_tensor_default_41" [id=773, type=dequantize_per_tensor];
+"774 transpose_12" [id=774, type=transpose];
+"775 matmul_12" [id=775, type=matmul];
+"776 _frozen_param54" [id=776, type=get_attr];
+"777 mul_13" [id=777, type=mul];
+"778 add_21" [id=778, type=add];
+"779 softmax_6" [id=779, type=softmax];
+"780 dropout_24" [id=780, type=dropout];
+"781 matmul_13" [id=781, type=matmul];
+"782 transpose_13" [id=782, type=transpose];
+"783 reshape_29" [id=783, type=reshape];
+"784 reshape_29_0_0_nncf_smooth_quant_0" [id=784, type=call_module];
+"785 quantize_per_tensor_default_42" [id=785, type=quantize_per_tensor];
+"786 dequantize_per_tensor_default_42" [id=786, type=dequantize_per_tensor];
+"787 features_5_2_attn_proj_bias_0_0" [id=787, type=get_attr];
+"788 scale_updated_constant28" [id=788, type=get_attr];
+"789 compressed_weight_updated_constant28" [id=789, type=get_attr];
+"790 mul_tensor_28" [id=790, type=mul];
+"791 zero_point_updated_constant28" [id=791, type=get_attr];
+"792 sub_tensor_28" [id=792, type=sub];
+"793 linear_41" [id=793, type=linear];
+"794 dropout_25" [id=794, type=dropout];
+"795 view_36" [id=795, type=view];
+"796 permute_31" [id=796, type=permute];
+"797 reshape_30" [id=797, type=reshape];
+"798 slice_106" [id=798, type=slice];
+"799 slice_107" [id=799, type=slice];
+"800 slice_108" [id=800, type=slice];
+"801 slice_109" [id=801, type=slice];
+"802 contiguous_11" [id=802, type=contiguous];
+"803 layer_norm_15" [id=803, type=layer_norm];
+"804 add_22" [id=804, type=add];
+"805 add_22_0_0_nncf_smooth_quant_0" [id=805, type=call_module];
+"806 quantize_per_tensor_default_43" [id=806, type=quantize_per_tensor];
+"807 dequantize_per_tensor_default_43" [id=807, type=dequantize_per_tensor];
+"808 features_5_2_mlp_0_bias_0_0" [id=808, type=get_attr];
+"809 scale_updated_constant29" [id=809, type=get_attr];
+"810 compressed_weight_updated_constant29" [id=810, type=get_attr];
+"811 mul_tensor_29" [id=811, type=mul];
+"812 zero_point_updated_constant29" [id=812, type=get_attr];
+"813 sub_tensor_29" [id=813, type=sub];
+"814 linear_42" [id=814, type=linear];
+"815 gelu_6" [id=815, type=gelu];
+"816 dropout_26" [id=816, type=dropout];
+"817 dropout_26_0_0_nncf_smooth_quant_0" [id=817, type=call_module];
+"818 quantize_per_tensor_default_44" [id=818, type=quantize_per_tensor];
+"819 dequantize_per_tensor_default_44" [id=819, type=dequantize_per_tensor];
+"820 features_5_2_mlp_3_bias_0_0" [id=820, type=get_attr];
+"821 scale_updated_constant30" [id=821, type=get_attr];
+"822 compressed_weight_updated_constant30" [id=822, type=get_attr];
+"823 mul_tensor_30" [id=823, type=mul];
+"824 zero_point_updated_constant30" [id=824, type=get_attr];
+"825 sub_tensor_30" [id=825, type=sub];
+"826 linear_43" [id=826, type=linear];
+"827 dropout_27" [id=827, type=dropout];
+"828 layer_norm_16" [id=828, type=layer_norm];
+"829 add_23" [id=829, type=add];
+"830 _frozen_param55" [id=830, type=get_attr];
+"831 pad_9" [id=831, type=pad];
+"832 roll_6" [id=832, type=roll];
+"833 view_39" [id=833, type=view];
+"834 permute_33" [id=834, type=permute];
+"835 reshape_31" [id=835, type=reshape];
+"836 reshape_31_0_0_nncf_smooth_quant_0" [id=836, type=call_module];
+"837 quantize_per_tensor_default_45" [id=837, type=quantize_per_tensor];
+"838 dequantize_per_tensor_default_45" [id=838, type=dequantize_per_tensor];
+"839 _frozen_param56_0_0" [id=839, type=get_attr];
+"840 scale_updated_constant31" [id=840, type=get_attr];
+"841 compressed_weight_updated_constant31" [id=841, type=get_attr];
+"842 mul_tensor_31" [id=842, type=mul];
+"843 zero_point_updated_constant31" [id=843, type=get_attr];
+"844 sub_tensor_31" [id=844, type=sub];
+"845 linear_46" [id=845, type=linear];
+"846 reshape_32" [id=846, type=reshape];
+"847 permute_34" [id=847, type=permute];
+"848 select_21" [id=848, type=select];
+"849 select_22" [id=849, type=select];
+"850 select_23" [id=850, type=select];
+"851 linalg_vector_norm_14" [id=851, type=linalg_vector_norm];
+"852 clamp_min_14" [id=852, type=clamp_min];
+"853 expand_as_14" [id=853, type=expand_as];
+"854 div_14" [id=854, type=div];
+"855 quantize_per_tensor_default_46" [id=855, type=quantize_per_tensor];
+"856 dequantize_per_tensor_default_46" [id=856, type=dequantize_per_tensor];
+"857 linalg_vector_norm_15" [id=857, type=linalg_vector_norm];
+"858 clamp_min_15" [id=858, type=clamp_min];
+"859 expand_as_15" [id=859, type=expand_as];
+"860 div_15" [id=860, type=div];
+"861 quantize_per_tensor_default_47" [id=861, type=quantize_per_tensor];
+"862 dequantize_per_tensor_default_47" [id=862, type=dequantize_per_tensor];
+"863 transpose_14" [id=863, type=transpose];
+"864 matmul_14" [id=864, type=matmul];
+"865 _frozen_param58" [id=865, type=get_attr];
+"866 mul_15" [id=866, type=mul];
+"867 add_24" [id=867, type=add];
+"868 new_zeros_3" [id=868, type=new_zeros];
+"869 view_40" [id=869, type=view];
+"870 permute_35" [id=870, type=permute];
+"871 reshape_33" [id=871, type=reshape];
+"872 unsqueeze_20" [id=872, type=unsqueeze];
+"873 unsqueeze_21" [id=873, type=unsqueeze];
+"874 sub_3" [id=874, type=sub];
+"875 ne_3" [id=875, type=ne];
+"876 masked_fill_6" [id=876, type=masked_fill];
+"877 eq_3" [id=877, type=eq];
+"878 masked_fill_7" [id=878, type=masked_fill];
+"879 view_41" [id=879, type=view];
+"880 unsqueeze_22" [id=880, type=unsqueeze];
+"881 unsqueeze_23" [id=881, type=unsqueeze];
+"882 add_25" [id=882, type=add];
+"883 view_42" [id=883, type=view];
+"884 softmax_7" [id=884, type=softmax];
+"885 dropout_28" [id=885, type=dropout];
+"886 matmul_15" [id=886, type=matmul];
+"887 transpose_15" [id=887, type=transpose];
+"888 reshape_34" [id=888, type=reshape];
+"889 reshape_34_0_0_nncf_smooth_quant_0" [id=889, type=call_module];
+"890 quantize_per_tensor_default_48" [id=890, type=quantize_per_tensor];
+"891 dequantize_per_tensor_default_48" [id=891, type=dequantize_per_tensor];
+"892 features_5_3_attn_proj_bias_0_0" [id=892, type=get_attr];
+"893 scale_updated_constant32" [id=893, type=get_attr];
+"894 compressed_weight_updated_constant32" [id=894, type=get_attr];
+"895 mul_tensor_32" [id=895, type=mul];
+"896 zero_point_updated_constant32" [id=896, type=get_attr];
+"897 sub_tensor_32" [id=897, type=sub];
+"898 linear_47" [id=898, type=linear];
+"899 dropout_29" [id=899, type=dropout];
+"900 view_43" [id=900, type=view];
+"901 permute_36" [id=901, type=permute];
+"902 reshape_35" [id=902, type=reshape];
+"903 roll_7" [id=903, type=roll];
+"904 slice_129" [id=904, type=slice];
+"905 slice_130" [id=905, type=slice];
+"906 slice_131" [id=906, type=slice];
+"907 slice_132" [id=907, type=slice];
+"908 contiguous_13" [id=908, type=contiguous];
+"909 layer_norm_17" [id=909, type=layer_norm];
+"910 add_26" [id=910, type=add];
+"911 add_26_0_0_nncf_smooth_quant_0" [id=911, type=call_module];
+"912 quantize_per_tensor_default_49" [id=912, type=quantize_per_tensor];
+"913 dequantize_per_tensor_default_49" [id=913, type=dequantize_per_tensor];
+"914 features_5_3_mlp_0_bias_0_0" [id=914, type=get_attr];
+"915 scale_updated_constant33" [id=915, type=get_attr];
+"916 compressed_weight_updated_constant33" [id=916, type=get_attr];
+"917 mul_tensor_33" [id=917, type=mul];
+"918 zero_point_updated_constant33" [id=918, type=get_attr];
+"919 sub_tensor_33" [id=919, type=sub];
+"920 linear_48" [id=920, type=linear];
+"921 gelu_7" [id=921, type=gelu];
+"922 dropout_30" [id=922, type=dropout];
+"923 dropout_30_0_0_nncf_smooth_quant_0" [id=923, type=call_module];
+"924 quantize_per_tensor_default_50" [id=924, type=quantize_per_tensor];
+"925 dequantize_per_tensor_default_50" [id=925, type=dequantize_per_tensor];
+"926 features_5_3_mlp_3_bias_0_0" [id=926, type=get_attr];
+"927 scale_updated_constant34" [id=927, type=get_attr];
+"928 compressed_weight_updated_constant34" [id=928, type=get_attr];
+"929 mul_tensor_34" [id=929, type=mul];
+"930 zero_point_updated_constant34" [id=930, type=get_attr];
+"931 sub_tensor_34" [id=931, type=sub];
+"932 linear_49" [id=932, type=linear];
+"933 dropout_31" [id=933, type=dropout];
+"934 layer_norm_18" [id=934, type=layer_norm];
+"935 add_27" [id=935, type=add];
+"936 _frozen_param68" [id=936, type=get_attr];
+"937 pad_10" [id=937, type=pad];
+"938 view_46" [id=938, type=view];
+"939 permute_38" [id=939, type=permute];
+"940 reshape_36" [id=940, type=reshape];
+"941 reshape_36_0_0_nncf_smooth_quant_0" [id=941, type=call_module];
+"942 quantize_per_tensor_default_51" [id=942, type=quantize_per_tensor];
+"943 dequantize_per_tensor_default_51" [id=943, type=dequantize_per_tensor];
+"944 _frozen_param69_0_0" [id=944, type=get_attr];
+"945 scale_updated_constant35" [id=945, type=get_attr];
+"946 compressed_weight_updated_constant35" [id=946, type=get_attr];
+"947 mul_tensor_35" [id=947, type=mul];
+"948 zero_point_updated_constant35" [id=948, type=get_attr];
+"949 sub_tensor_35" [id=949, type=sub];
+"950 linear_52" [id=950, type=linear];
+"951 reshape_37" [id=951, type=reshape];
+"952 permute_39" [id=952, type=permute];
+"953 select_24" [id=953, type=select];
+"954 select_25" [id=954, type=select];
+"955 select_26" [id=955, type=select];
+"956 linalg_vector_norm_16" [id=956, type=linalg_vector_norm];
+"957 clamp_min_16" [id=957, type=clamp_min];
+"958 expand_as_16" [id=958, type=expand_as];
+"959 div_16" [id=959, type=div];
+"960 quantize_per_tensor_default_52" [id=960, type=quantize_per_tensor];
+"961 dequantize_per_tensor_default_52" [id=961, type=dequantize_per_tensor];
+"962 linalg_vector_norm_17" [id=962, type=linalg_vector_norm];
+"963 clamp_min_17" [id=963, type=clamp_min];
+"964 expand_as_17" [id=964, type=expand_as];
+"965 div_17" [id=965, type=div];
+"966 quantize_per_tensor_default_53" [id=966, type=quantize_per_tensor];
+"967 dequantize_per_tensor_default_53" [id=967, type=dequantize_per_tensor];
+"968 transpose_16" [id=968, type=transpose];
+"969 matmul_16" [id=969, type=matmul];
+"970 _frozen_param71" [id=970, type=get_attr];
+"971 mul_17" [id=971, type=mul];
+"972 add_28" [id=972, type=add];
+"973 softmax_8" [id=973, type=softmax];
+"974 dropout_32" [id=974, type=dropout];
+"975 matmul_17" [id=975, type=matmul];
+"976 transpose_17" [id=976, type=transpose];
+"977 reshape_38" [id=977, type=reshape];
+"978 reshape_38_0_0_nncf_smooth_quant_0" [id=978, type=call_module];
+"979 quantize_per_tensor_default_54" [id=979, type=quantize_per_tensor];
+"980 dequantize_per_tensor_default_54" [id=980, type=dequantize_per_tensor];
+"981 features_5_4_attn_proj_bias_0_0" [id=981, type=get_attr];
+"982 scale_updated_constant36" [id=982, type=get_attr];
+"983 compressed_weight_updated_constant36" [id=983, type=get_attr];
+"984 mul_tensor_36" [id=984, type=mul];
+"985 zero_point_updated_constant36" [id=985, type=get_attr];
+"986 sub_tensor_36" [id=986, type=sub];
+"987 linear_53" [id=987, type=linear];
+"988 dropout_33" [id=988, type=dropout];
+"989 view_47" [id=989, type=view];
+"990 permute_40" [id=990, type=permute];
+"991 reshape_39" [id=991, type=reshape];
+"992 slice_134" [id=992, type=slice];
+"993 slice_135" [id=993, type=slice];
+"994 slice_136" [id=994, type=slice];
+"995 slice_137" [id=995, type=slice];
+"996 contiguous_15" [id=996, type=contiguous];
+"997 layer_norm_19" [id=997, type=layer_norm];
+"998 add_29" [id=998, type=add];
+"999 add_29_0_0_nncf_smooth_quant_0" [id=999, type=call_module];
+"1000 quantize_per_tensor_default_55" [id=1000, type=quantize_per_tensor];
+"1001 dequantize_per_tensor_default_55" [id=1001, type=dequantize_per_tensor];
+"1002 features_5_4_mlp_0_bias_0_0" [id=1002, type=get_attr];
+"1003 scale_updated_constant37" [id=1003, type=get_attr];
+"1004 compressed_weight_updated_constant37" [id=1004, type=get_attr];
+"1005 mul_tensor_37" [id=1005, type=mul];
+"1006 zero_point_updated_constant37" [id=1006, type=get_attr];
+"1007 sub_tensor_37" [id=1007, type=sub];
+"1008 linear_54" [id=1008, type=linear];
+"1009 gelu_8" [id=1009, type=gelu];
+"1010 dropout_34" [id=1010, type=dropout];
+"1011 dropout_34_0_0_nncf_smooth_quant_0" [id=1011, type=call_module];
+"1012 quantize_per_tensor_default_56" [id=1012, type=quantize_per_tensor];
+"1013 dequantize_per_tensor_default_56" [id=1013, type=dequantize_per_tensor];
+"1014 features_5_4_mlp_3_bias_0_0" [id=1014, type=get_attr];
+"1015 scale_updated_constant38" [id=1015, type=get_attr];
+"1016 compressed_weight_updated_constant38" [id=1016, type=get_attr];
+"1017 mul_tensor_38" [id=1017, type=mul];
+"1018 zero_point_updated_constant38" [id=1018, type=get_attr];
+"1019 sub_tensor_38" [id=1019, type=sub];
+"1020 linear_55" [id=1020, type=linear];
+"1021 dropout_35" [id=1021, type=dropout];
+"1022 layer_norm_20" [id=1022, type=layer_norm];
+"1023 add_30" [id=1023, type=add];
+"1024 _frozen_param72" [id=1024, type=get_attr];
+"1025 pad_11" [id=1025, type=pad];
+"1026 roll_8" [id=1026, type=roll];
+"1027 view_50" [id=1027, type=view];
+"1028 permute_42" [id=1028, type=permute];
+"1029 reshape_40" [id=1029, type=reshape];
+"1030 reshape_40_0_0_nncf_smooth_quant_0" [id=1030, type=call_module];
+"1031 quantize_per_tensor_default_57" [id=1031, type=quantize_per_tensor];
+"1032 dequantize_per_tensor_default_57" [id=1032, type=dequantize_per_tensor];
+"1033 _frozen_param73_0_0" [id=1033, type=get_attr];
+"1034 scale_updated_constant39" [id=1034, type=get_attr];
+"1035 compressed_weight_updated_constant39" [id=1035, type=get_attr];
+"1036 mul_tensor_39" [id=1036, type=mul];
+"1037 zero_point_updated_constant39" [id=1037, type=get_attr];
+"1038 sub_tensor_39" [id=1038, type=sub];
+"1039 linear_58" [id=1039, type=linear];
+"1040 reshape_41" [id=1040, type=reshape];
+"1041 permute_43" [id=1041, type=permute];
+"1042 select_27" [id=1042, type=select];
+"1043 select_28" [id=1043, type=select];
+"1044 select_29" [id=1044, type=select];
+"1045 linalg_vector_norm_18" [id=1045, type=linalg_vector_norm];
+"1046 clamp_min_18" [id=1046, type=clamp_min];
+"1047 expand_as_18" [id=1047, type=expand_as];
+"1048 div_18" [id=1048, type=div];
+"1049 quantize_per_tensor_default_58" [id=1049, type=quantize_per_tensor];
+"1050 dequantize_per_tensor_default_58" [id=1050, type=dequantize_per_tensor];
+"1051 linalg_vector_norm_19" [id=1051, type=linalg_vector_norm];
+"1052 clamp_min_19" [id=1052, type=clamp_min];
+"1053 expand_as_19" [id=1053, type=expand_as];
+"1054 div_19" [id=1054, type=div];
+"1055 quantize_per_tensor_default_59" [id=1055, type=quantize_per_tensor];
+"1056 dequantize_per_tensor_default_59" [id=1056, type=dequantize_per_tensor];
+"1057 transpose_18" [id=1057, type=transpose];
+"1058 matmul_18" [id=1058, type=matmul];
+"1059 _frozen_param75" [id=1059, type=get_attr];
+"1060 mul_19" [id=1060, type=mul];
+"1061 add_31" [id=1061, type=add];
+"1062 new_zeros_4" [id=1062, type=new_zeros];
+"1063 view_51" [id=1063, type=view];
+"1064 permute_44" [id=1064, type=permute];
+"1065 reshape_42" [id=1065, type=reshape];
+"1066 unsqueeze_26" [id=1066, type=unsqueeze];
+"1067 unsqueeze_27" [id=1067, type=unsqueeze];
+"1068 sub_4" [id=1068, type=sub];
+"1069 ne_4" [id=1069, type=ne];
+"1070 masked_fill_8" [id=1070, type=masked_fill];
+"1071 eq_4" [id=1071, type=eq];
+"1072 masked_fill_9" [id=1072, type=masked_fill];
+"1073 view_52" [id=1073, type=view];
+"1074 unsqueeze_28" [id=1074, type=unsqueeze];
+"1075 unsqueeze_29" [id=1075, type=unsqueeze];
+"1076 add_32" [id=1076, type=add];
+"1077 view_53" [id=1077, type=view];
+"1078 softmax_9" [id=1078, type=softmax];
+"1079 dropout_36" [id=1079, type=dropout];
+"1080 matmul_19" [id=1080, type=matmul];
+"1081 transpose_19" [id=1081, type=transpose];
+"1082 reshape_43" [id=1082, type=reshape];
+"1083 reshape_43_0_0_nncf_smooth_quant_0" [id=1083, type=call_module];
+"1084 quantize_per_tensor_default_60" [id=1084, type=quantize_per_tensor];
+"1085 dequantize_per_tensor_default_60" [id=1085, type=dequantize_per_tensor];
+"1086 features_5_5_attn_proj_bias_0_0" [id=1086, type=get_attr];
+"1087 scale_updated_constant40" [id=1087, type=get_attr];
+"1088 compressed_weight_updated_constant40" [id=1088, type=get_attr];
+"1089 mul_tensor_40" [id=1089, type=mul];
+"1090 zero_point_updated_constant40" [id=1090, type=get_attr];
+"1091 sub_tensor_40" [id=1091, type=sub];
+"1092 linear_59" [id=1092, type=linear];
+"1093 dropout_37" [id=1093, type=dropout];
+"1094 view_54" [id=1094, type=view];
+"1095 permute_45" [id=1095, type=permute];
+"1096 reshape_44" [id=1096, type=reshape];
+"1097 roll_9" [id=1097, type=roll];
+"1098 slice_157" [id=1098, type=slice];
+"1099 slice_158" [id=1099, type=slice];
+"1100 slice_159" [id=1100, type=slice];
+"1101 slice_160" [id=1101, type=slice];
+"1102 contiguous_17" [id=1102, type=contiguous];
+"1103 layer_norm_21" [id=1103, type=layer_norm];
+"1104 add_33" [id=1104, type=add];
+"1105 add_33_0_0_nncf_smooth_quant_0" [id=1105, type=call_module];
+"1106 quantize_per_tensor_default_61" [id=1106, type=quantize_per_tensor];
+"1107 dequantize_per_tensor_default_61" [id=1107, type=dequantize_per_tensor];
+"1108 features_5_5_mlp_0_bias_0_0" [id=1108, type=get_attr];
+"1109 scale_updated_constant41" [id=1109, type=get_attr];
+"1110 compressed_weight_updated_constant41" [id=1110, type=get_attr];
+"1111 mul_tensor_41" [id=1111, type=mul];
+"1112 zero_point_updated_constant41" [id=1112, type=get_attr];
+"1113 sub_tensor_41" [id=1113, type=sub];
+"1114 linear_60" [id=1114, type=linear];
+"1115 gelu_9" [id=1115, type=gelu];
+"1116 dropout_38" [id=1116, type=dropout];
+"1117 dropout_38_0_0_nncf_smooth_quant_0" [id=1117, type=call_module];
+"1118 quantize_per_tensor_default_62" [id=1118, type=quantize_per_tensor];
+"1119 dequantize_per_tensor_default_62" [id=1119, type=dequantize_per_tensor];
+"1120 features_5_5_mlp_3_bias_0_0" [id=1120, type=get_attr];
+"1121 scale_updated_constant42" [id=1121, type=get_attr];
+"1122 compressed_weight_updated_constant42" [id=1122, type=get_attr];
+"1123 mul_tensor_42" [id=1123, type=mul];
+"1124 zero_point_updated_constant42" [id=1124, type=get_attr];
+"1125 sub_tensor_42" [id=1125, type=sub];
+"1126 linear_61" [id=1126, type=linear];
+"1127 dropout_39" [id=1127, type=dropout];
+"1128 layer_norm_22" [id=1128, type=layer_norm];
+"1129 add_34" [id=1129, type=add];
+"1130 _frozen_param85" [id=1130, type=get_attr];
+"1131 pad_12" [id=1131, type=pad];
+"1132 view_57" [id=1132, type=view];
+"1133 permute_47" [id=1133, type=permute];
+"1134 reshape_45" [id=1134, type=reshape];
+"1135 reshape_45_0_0_nncf_smooth_quant_0" [id=1135, type=call_module];
+"1136 quantize_per_tensor_default_63" [id=1136, type=quantize_per_tensor];
+"1137 dequantize_per_tensor_default_63" [id=1137, type=dequantize_per_tensor];
+"1138 _frozen_param86_0_0" [id=1138, type=get_attr];
+"1139 scale_updated_constant43" [id=1139, type=get_attr];
+"1140 compressed_weight_updated_constant43" [id=1140, type=get_attr];
+"1141 mul_tensor_43" [id=1141, type=mul];
+"1142 zero_point_updated_constant43" [id=1142, type=get_attr];
+"1143 sub_tensor_43" [id=1143, type=sub];
+"1144 linear_64" [id=1144, type=linear];
+"1145 reshape_46" [id=1145, type=reshape];
+"1146 permute_48" [id=1146, type=permute];
+"1147 select_30" [id=1147, type=select];
+"1148 select_31" [id=1148, type=select];
+"1149 select_32" [id=1149, type=select];
+"1150 linalg_vector_norm_20" [id=1150, type=linalg_vector_norm];
+"1151 clamp_min_20" [id=1151, type=clamp_min];
+"1152 expand_as_20" [id=1152, type=expand_as];
+"1153 div_20" [id=1153, type=div];
+"1154 quantize_per_tensor_default_64" [id=1154, type=quantize_per_tensor];
+"1155 dequantize_per_tensor_default_64" [id=1155, type=dequantize_per_tensor];
+"1156 linalg_vector_norm_21" [id=1156, type=linalg_vector_norm];
+"1157 clamp_min_21" [id=1157, type=clamp_min];
+"1158 expand_as_21" [id=1158, type=expand_as];
+"1159 div_21" [id=1159, type=div];
+"1160 quantize_per_tensor_default_65" [id=1160, type=quantize_per_tensor];
+"1161 dequantize_per_tensor_default_65" [id=1161, type=dequantize_per_tensor];
+"1162 transpose_20" [id=1162, type=transpose];
+"1163 matmul_20" [id=1163, type=matmul];
+"1164 _frozen_param88" [id=1164, type=get_attr];
+"1165 mul_21" [id=1165, type=mul];
+"1166 add_35" [id=1166, type=add];
+"1167 softmax_10" [id=1167, type=softmax];
+"1168 dropout_40" [id=1168, type=dropout];
+"1169 matmul_21" [id=1169, type=matmul];
+"1170 transpose_21" [id=1170, type=transpose];
+"1171 reshape_47" [id=1171, type=reshape];
+"1172 reshape_47_0_0_nncf_smooth_quant_0" [id=1172, type=call_module];
+"1173 quantize_per_tensor_default_66" [id=1173, type=quantize_per_tensor];
+"1174 dequantize_per_tensor_default_66" [id=1174, type=dequantize_per_tensor];
+"1175 features_5_6_attn_proj_bias_0_0" [id=1175, type=get_attr];
+"1176 scale_updated_constant44" [id=1176, type=get_attr];
+"1177 compressed_weight_updated_constant44" [id=1177, type=get_attr];
+"1178 mul_tensor_44" [id=1178, type=mul];
+"1179 zero_point_updated_constant44" [id=1179, type=get_attr];
+"1180 sub_tensor_44" [id=1180, type=sub];
+"1181 linear_65" [id=1181, type=linear];
+"1182 dropout_41" [id=1182, type=dropout];
+"1183 view_58" [id=1183, type=view];
+"1184 permute_49" [id=1184, type=permute];
+"1185 reshape_48" [id=1185, type=reshape];
+"1186 slice_162" [id=1186, type=slice];
+"1187 slice_163" [id=1187, type=slice];
+"1188 slice_164" [id=1188, type=slice];
+"1189 slice_165" [id=1189, type=slice];
+"1190 contiguous_19" [id=1190, type=contiguous];
+"1191 layer_norm_23" [id=1191, type=layer_norm];
+"1192 add_36" [id=1192, type=add];
+"1193 add_36_0_0_nncf_smooth_quant_0" [id=1193, type=call_module];
+"1194 quantize_per_tensor_default_67" [id=1194, type=quantize_per_tensor];
+"1195 dequantize_per_tensor_default_67" [id=1195, type=dequantize_per_tensor];
+"1196 features_5_6_mlp_0_bias_0_0" [id=1196, type=get_attr];
+"1197 scale_updated_constant45" [id=1197, type=get_attr];
+"1198 compressed_weight_updated_constant45" [id=1198, type=get_attr];
+"1199 mul_tensor_45" [id=1199, type=mul];
+"1200 zero_point_updated_constant45" [id=1200, type=get_attr];
+"1201 sub_tensor_45" [id=1201, type=sub];
+"1202 linear_66" [id=1202, type=linear];
+"1203 gelu_10" [id=1203, type=gelu];
+"1204 dropout_42" [id=1204, type=dropout];
+"1205 dropout_42_0_0_nncf_smooth_quant_0" [id=1205, type=call_module];
+"1206 quantize_per_tensor_default_68" [id=1206, type=quantize_per_tensor];
+"1207 dequantize_per_tensor_default_68" [id=1207, type=dequantize_per_tensor];
+"1208 features_5_6_mlp_3_bias_0_0" [id=1208, type=get_attr];
+"1209 scale_updated_constant46" [id=1209, type=get_attr];
+"1210 compressed_weight_updated_constant46" [id=1210, type=get_attr];
+"1211 mul_tensor_46" [id=1211, type=mul];
+"1212 zero_point_updated_constant46" [id=1212, type=get_attr];
+"1213 sub_tensor_46" [id=1213, type=sub];
+"1214 linear_67" [id=1214, type=linear];
+"1215 dropout_43" [id=1215, type=dropout];
+"1216 layer_norm_24" [id=1216, type=layer_norm];
+"1217 add_37" [id=1217, type=add];
+"1218 _frozen_param89" [id=1218, type=get_attr];
+"1219 pad_13" [id=1219, type=pad];
+"1220 roll_10" [id=1220, type=roll];
+"1221 view_61" [id=1221, type=view];
+"1222 permute_51" [id=1222, type=permute];
+"1223 reshape_49" [id=1223, type=reshape];
+"1224 reshape_49_0_0_nncf_smooth_quant_0" [id=1224, type=call_module];
+"1225 quantize_per_tensor_default_69" [id=1225, type=quantize_per_tensor];
+"1226 dequantize_per_tensor_default_69" [id=1226, type=dequantize_per_tensor];
+"1227 _frozen_param90_0_0" [id=1227, type=get_attr];
+"1228 scale_updated_constant47" [id=1228, type=get_attr];
+"1229 compressed_weight_updated_constant47" [id=1229, type=get_attr];
+"1230 mul_tensor_47" [id=1230, type=mul];
+"1231 zero_point_updated_constant47" [id=1231, type=get_attr];
+"1232 sub_tensor_47" [id=1232, type=sub];
+"1233 linear_70" [id=1233, type=linear];
+"1234 reshape_50" [id=1234, type=reshape];
+"1235 permute_52" [id=1235, type=permute];
+"1236 select_33" [id=1236, type=select];
+"1237 select_34" [id=1237, type=select];
+"1238 select_35" [id=1238, type=select];
+"1239 linalg_vector_norm_22" [id=1239, type=linalg_vector_norm];
+"1240 clamp_min_22" [id=1240, type=clamp_min];
+"1241 expand_as_22" [id=1241, type=expand_as];
+"1242 div_22" [id=1242, type=div];
+"1243 quantize_per_tensor_default_70" [id=1243, type=quantize_per_tensor];
+"1244 dequantize_per_tensor_default_70" [id=1244, type=dequantize_per_tensor];
+"1245 linalg_vector_norm_23" [id=1245, type=linalg_vector_norm];
+"1246 clamp_min_23" [id=1246, type=clamp_min];
+"1247 expand_as_23" [id=1247, type=expand_as];
+"1248 div_23" [id=1248, type=div];
+"1249 quantize_per_tensor_default_71" [id=1249, type=quantize_per_tensor];
+"1250 dequantize_per_tensor_default_71" [id=1250, type=dequantize_per_tensor];
+"1251 transpose_22" [id=1251, type=transpose];
+"1252 matmul_22" [id=1252, type=matmul];
+"1253 _frozen_param92" [id=1253, type=get_attr];
+"1254 mul_23" [id=1254, type=mul];
+"1255 add_38" [id=1255, type=add];
+"1256 new_zeros_5" [id=1256, type=new_zeros];
+"1257 view_62" [id=1257, type=view];
+"1258 permute_53" [id=1258, type=permute];
+"1259 reshape_51" [id=1259, type=reshape];
+"1260 unsqueeze_32" [id=1260, type=unsqueeze];
+"1261 unsqueeze_33" [id=1261, type=unsqueeze];
+"1262 sub_5" [id=1262, type=sub];
+"1263 ne_5" [id=1263, type=ne];
+"1264 masked_fill_10" [id=1264, type=masked_fill];
+"1265 eq_5" [id=1265, type=eq];
+"1266 masked_fill_11" [id=1266, type=masked_fill];
+"1267 view_63" [id=1267, type=view];
+"1268 unsqueeze_34" [id=1268, type=unsqueeze];
+"1269 unsqueeze_35" [id=1269, type=unsqueeze];
+"1270 add_39" [id=1270, type=add];
+"1271 view_64" [id=1271, type=view];
+"1272 softmax_11" [id=1272, type=softmax];
+"1273 dropout_44" [id=1273, type=dropout];
+"1274 matmul_23" [id=1274, type=matmul];
+"1275 transpose_23" [id=1275, type=transpose];
+"1276 reshape_52" [id=1276, type=reshape];
+"1277 reshape_52_0_0_nncf_smooth_quant_0" [id=1277, type=call_module];
+"1278 quantize_per_tensor_default_72" [id=1278, type=quantize_per_tensor];
+"1279 dequantize_per_tensor_default_72" [id=1279, type=dequantize_per_tensor];
+"1280 features_5_7_attn_proj_bias_0_0" [id=1280, type=get_attr];
+"1281 scale_updated_constant48" [id=1281, type=get_attr];
+"1282 compressed_weight_updated_constant48" [id=1282, type=get_attr];
+"1283 mul_tensor_48" [id=1283, type=mul];
+"1284 zero_point_updated_constant48" [id=1284, type=get_attr];
+"1285 sub_tensor_48" [id=1285, type=sub];
+"1286 linear_71" [id=1286, type=linear];
+"1287 dropout_45" [id=1287, type=dropout];
+"1288 view_65" [id=1288, type=view];
+"1289 permute_54" [id=1289, type=permute];
+"1290 reshape_53" [id=1290, type=reshape];
+"1291 roll_11" [id=1291, type=roll];
+"1292 slice_185" [id=1292, type=slice];
+"1293 slice_186" [id=1293, type=slice];
+"1294 slice_187" [id=1294, type=slice];
+"1295 slice_188" [id=1295, type=slice];
+"1296 contiguous_21" [id=1296, type=contiguous];
+"1297 layer_norm_25" [id=1297, type=layer_norm];
+"1298 add_40" [id=1298, type=add];
+"1299 add_40_0_0_nncf_smooth_quant_0" [id=1299, type=call_module];
+"1300 quantize_per_tensor_default_73" [id=1300, type=quantize_per_tensor];
+"1301 dequantize_per_tensor_default_73" [id=1301, type=dequantize_per_tensor];
+"1302 features_5_7_mlp_0_bias_0_0" [id=1302, type=get_attr];
+"1303 scale_updated_constant49" [id=1303, type=get_attr];
+"1304 compressed_weight_updated_constant49" [id=1304, type=get_attr];
+"1305 mul_tensor_49" [id=1305, type=mul];
+"1306 zero_point_updated_constant49" [id=1306, type=get_attr];
+"1307 sub_tensor_49" [id=1307, type=sub];
+"1308 linear_72" [id=1308, type=linear];
+"1309 gelu_11" [id=1309, type=gelu];
+"1310 dropout_46" [id=1310, type=dropout];
+"1311 dropout_46_0_0_nncf_smooth_quant_0" [id=1311, type=call_module];
+"1312 quantize_per_tensor_default_74" [id=1312, type=quantize_per_tensor];
+"1313 dequantize_per_tensor_default_74" [id=1313, type=dequantize_per_tensor];
+"1314 features_5_7_mlp_3_bias_0_0" [id=1314, type=get_attr];
+"1315 scale_updated_constant50" [id=1315, type=get_attr];
+"1316 compressed_weight_updated_constant50" [id=1316, type=get_attr];
+"1317 mul_tensor_50" [id=1317, type=mul];
+"1318 zero_point_updated_constant50" [id=1318, type=get_attr];
+"1319 sub_tensor_50" [id=1319, type=sub];
+"1320 linear_73" [id=1320, type=linear];
+"1321 dropout_47" [id=1321, type=dropout];
+"1322 layer_norm_26" [id=1322, type=layer_norm];
+"1323 add_41" [id=1323, type=add];
+"1324 _frozen_param102" [id=1324, type=get_attr];
+"1325 pad_14" [id=1325, type=pad];
+"1326 view_68" [id=1326, type=view];
+"1327 permute_56" [id=1327, type=permute];
+"1328 reshape_54" [id=1328, type=reshape];
+"1329 reshape_54_0_0_nncf_smooth_quant_0" [id=1329, type=call_module];
+"1330 quantize_per_tensor_default_75" [id=1330, type=quantize_per_tensor];
+"1331 dequantize_per_tensor_default_75" [id=1331, type=dequantize_per_tensor];
+"1332 _frozen_param103_0_0" [id=1332, type=get_attr];
+"1333 scale_updated_constant51" [id=1333, type=get_attr];
+"1334 compressed_weight_updated_constant51" [id=1334, type=get_attr];
+"1335 mul_tensor_51" [id=1335, type=mul];
+"1336 zero_point_updated_constant51" [id=1336, type=get_attr];
+"1337 sub_tensor_51" [id=1337, type=sub];
+"1338 linear_76" [id=1338, type=linear];
+"1339 reshape_55" [id=1339, type=reshape];
+"1340 permute_57" [id=1340, type=permute];
+"1341 select_36" [id=1341, type=select];
+"1342 select_37" [id=1342, type=select];
+"1343 select_38" [id=1343, type=select];
+"1344 linalg_vector_norm_24" [id=1344, type=linalg_vector_norm];
+"1345 clamp_min_24" [id=1345, type=clamp_min];
+"1346 expand_as_24" [id=1346, type=expand_as];
+"1347 div_24" [id=1347, type=div];
+"1348 quantize_per_tensor_default_76" [id=1348, type=quantize_per_tensor];
+"1349 dequantize_per_tensor_default_76" [id=1349, type=dequantize_per_tensor];
+"1350 linalg_vector_norm_25" [id=1350, type=linalg_vector_norm];
+"1351 clamp_min_25" [id=1351, type=clamp_min];
+"1352 expand_as_25" [id=1352, type=expand_as];
+"1353 div_25" [id=1353, type=div];
+"1354 quantize_per_tensor_default_77" [id=1354, type=quantize_per_tensor];
+"1355 dequantize_per_tensor_default_77" [id=1355, type=dequantize_per_tensor];
+"1356 transpose_24" [id=1356, type=transpose];
+"1357 matmul_24" [id=1357, type=matmul];
+"1358 _frozen_param105" [id=1358, type=get_attr];
+"1359 mul_25" [id=1359, type=mul];
+"1360 add_42" [id=1360, type=add];
+"1361 softmax_12" [id=1361, type=softmax];
+"1362 dropout_48" [id=1362, type=dropout];
+"1363 matmul_25" [id=1363, type=matmul];
+"1364 transpose_25" [id=1364, type=transpose];
+"1365 reshape_56" [id=1365, type=reshape];
+"1366 reshape_56_0_0_nncf_smooth_quant_0" [id=1366, type=call_module];
+"1367 quantize_per_tensor_default_78" [id=1367, type=quantize_per_tensor];
+"1368 dequantize_per_tensor_default_78" [id=1368, type=dequantize_per_tensor];
+"1369 features_5_8_attn_proj_bias_0_0" [id=1369, type=get_attr];
+"1370 scale_updated_constant52" [id=1370, type=get_attr];
+"1371 compressed_weight_updated_constant52" [id=1371, type=get_attr];
+"1372 mul_tensor_52" [id=1372, type=mul];
+"1373 zero_point_updated_constant52" [id=1373, type=get_attr];
+"1374 sub_tensor_52" [id=1374, type=sub];
+"1375 linear_77" [id=1375, type=linear];
+"1376 dropout_49" [id=1376, type=dropout];
+"1377 view_69" [id=1377, type=view];
+"1378 permute_58" [id=1378, type=permute];
+"1379 reshape_57" [id=1379, type=reshape];
+"1380 slice_190" [id=1380, type=slice];
+"1381 slice_191" [id=1381, type=slice];
+"1382 slice_192" [id=1382, type=slice];
+"1383 slice_193" [id=1383, type=slice];
+"1384 contiguous_23" [id=1384, type=contiguous];
+"1385 layer_norm_27" [id=1385, type=layer_norm];
+"1386 add_43" [id=1386, type=add];
+"1387 add_43_0_0_nncf_smooth_quant_0" [id=1387, type=call_module];
+"1388 quantize_per_tensor_default_79" [id=1388, type=quantize_per_tensor];
+"1389 dequantize_per_tensor_default_79" [id=1389, type=dequantize_per_tensor];
+"1390 features_5_8_mlp_0_bias_0_0" [id=1390, type=get_attr];
+"1391 scale_updated_constant53" [id=1391, type=get_attr];
+"1392 compressed_weight_updated_constant53" [id=1392, type=get_attr];
+"1393 mul_tensor_53" [id=1393, type=mul];
+"1394 zero_point_updated_constant53" [id=1394, type=get_attr];
+"1395 sub_tensor_53" [id=1395, type=sub];
+"1396 linear_78" [id=1396, type=linear];
+"1397 gelu_12" [id=1397, type=gelu];
+"1398 dropout_50" [id=1398, type=dropout];
+"1399 dropout_50_0_0_nncf_smooth_quant_0" [id=1399, type=call_module];
+"1400 quantize_per_tensor_default_80" [id=1400, type=quantize_per_tensor];
+"1401 dequantize_per_tensor_default_80" [id=1401, type=dequantize_per_tensor];
+"1402 features_5_8_mlp_3_bias_0_0" [id=1402, type=get_attr];
+"1403 scale_updated_constant54" [id=1403, type=get_attr];
+"1404 compressed_weight_updated_constant54" [id=1404, type=get_attr];
+"1405 mul_tensor_54" [id=1405, type=mul];
+"1406 zero_point_updated_constant54" [id=1406, type=get_attr];
+"1407 sub_tensor_54" [id=1407, type=sub];
+"1408 linear_79" [id=1408, type=linear];
+"1409 dropout_51" [id=1409, type=dropout];
+"1410 layer_norm_28" [id=1410, type=layer_norm];
+"1411 add_44" [id=1411, type=add];
+"1412 _frozen_param106" [id=1412, type=get_attr];
+"1413 pad_15" [id=1413, type=pad];
+"1414 roll_12" [id=1414, type=roll];
+"1415 view_72" [id=1415, type=view];
+"1416 permute_60" [id=1416, type=permute];
+"1417 reshape_58" [id=1417, type=reshape];
+"1418 reshape_58_0_0_nncf_smooth_quant_0" [id=1418, type=call_module];
+"1419 quantize_per_tensor_default_81" [id=1419, type=quantize_per_tensor];
+"1420 dequantize_per_tensor_default_81" [id=1420, type=dequantize_per_tensor];
+"1421 _frozen_param107_0_0" [id=1421, type=get_attr];
+"1422 scale_updated_constant55" [id=1422, type=get_attr];
+"1423 compressed_weight_updated_constant55" [id=1423, type=get_attr];
+"1424 mul_tensor_55" [id=1424, type=mul];
+"1425 zero_point_updated_constant55" [id=1425, type=get_attr];
+"1426 sub_tensor_55" [id=1426, type=sub];
+"1427 linear_82" [id=1427, type=linear];
+"1428 reshape_59" [id=1428, type=reshape];
+"1429 permute_61" [id=1429, type=permute];
+"1430 select_39" [id=1430, type=select];
+"1431 select_40" [id=1431, type=select];
+"1432 select_41" [id=1432, type=select];
+"1433 linalg_vector_norm_26" [id=1433, type=linalg_vector_norm];
+"1434 clamp_min_26" [id=1434, type=clamp_min];
+"1435 expand_as_26" [id=1435, type=expand_as];
+"1436 div_26" [id=1436, type=div];
+"1437 quantize_per_tensor_default_82" [id=1437, type=quantize_per_tensor];
+"1438 dequantize_per_tensor_default_82" [id=1438, type=dequantize_per_tensor];
+"1439 linalg_vector_norm_27" [id=1439, type=linalg_vector_norm];
+"1440 clamp_min_27" [id=1440, type=clamp_min];
+"1441 expand_as_27" [id=1441, type=expand_as];
+"1442 div_27" [id=1442, type=div];
+"1443 quantize_per_tensor_default_83" [id=1443, type=quantize_per_tensor];
+"1444 dequantize_per_tensor_default_83" [id=1444, type=dequantize_per_tensor];
+"1445 transpose_26" [id=1445, type=transpose];
+"1446 matmul_26" [id=1446, type=matmul];
+"1447 _frozen_param109" [id=1447, type=get_attr];
+"1448 mul_27" [id=1448, type=mul];
+"1449 add_45" [id=1449, type=add];
+"1450 new_zeros_6" [id=1450, type=new_zeros];
+"1451 view_73" [id=1451, type=view];
+"1452 permute_62" [id=1452, type=permute];
+"1453 reshape_60" [id=1453, type=reshape];
+"1454 unsqueeze_38" [id=1454, type=unsqueeze];
+"1455 unsqueeze_39" [id=1455, type=unsqueeze];
+"1456 sub_6" [id=1456, type=sub];
+"1457 ne_6" [id=1457, type=ne];
+"1458 masked_fill_12" [id=1458, type=masked_fill];
+"1459 eq_6" [id=1459, type=eq];
+"1460 masked_fill_13" [id=1460, type=masked_fill];
+"1461 view_74" [id=1461, type=view];
+"1462 unsqueeze_40" [id=1462, type=unsqueeze];
+"1463 unsqueeze_41" [id=1463, type=unsqueeze];
+"1464 add_46" [id=1464, type=add];
+"1465 view_75" [id=1465, type=view];
+"1466 softmax_13" [id=1466, type=softmax];
+"1467 dropout_52" [id=1467, type=dropout];
+"1468 matmul_27" [id=1468, type=matmul];
+"1469 transpose_27" [id=1469, type=transpose];
+"1470 reshape_61" [id=1470, type=reshape];
+"1471 reshape_61_0_0_nncf_smooth_quant_0" [id=1471, type=call_module];
+"1472 quantize_per_tensor_default_84" [id=1472, type=quantize_per_tensor];
+"1473 dequantize_per_tensor_default_84" [id=1473, type=dequantize_per_tensor];
+"1474 features_5_9_attn_proj_bias_0_0" [id=1474, type=get_attr];
+"1475 scale_updated_constant56" [id=1475, type=get_attr];
+"1476 compressed_weight_updated_constant56" [id=1476, type=get_attr];
+"1477 mul_tensor_56" [id=1477, type=mul];
+"1478 zero_point_updated_constant56" [id=1478, type=get_attr];
+"1479 sub_tensor_56" [id=1479, type=sub];
+"1480 linear_83" [id=1480, type=linear];
+"1481 dropout_53" [id=1481, type=dropout];
+"1482 view_76" [id=1482, type=view];
+"1483 permute_63" [id=1483, type=permute];
+"1484 reshape_62" [id=1484, type=reshape];
+"1485 roll_13" [id=1485, type=roll];
+"1486 slice_213" [id=1486, type=slice];
+"1487 slice_214" [id=1487, type=slice];
+"1488 slice_215" [id=1488, type=slice];
+"1489 slice_216" [id=1489, type=slice];
+"1490 contiguous_25" [id=1490, type=contiguous];
+"1491 layer_norm_29" [id=1491, type=layer_norm];
+"1492 add_47" [id=1492, type=add];
+"1493 add_47_0_0_nncf_smooth_quant_0" [id=1493, type=call_module];
+"1494 quantize_per_tensor_default_85" [id=1494, type=quantize_per_tensor];
+"1495 dequantize_per_tensor_default_85" [id=1495, type=dequantize_per_tensor];
+"1496 features_5_9_mlp_0_bias_0_0" [id=1496, type=get_attr];
+"1497 scale_updated_constant57" [id=1497, type=get_attr];
+"1498 compressed_weight_updated_constant57" [id=1498, type=get_attr];
+"1499 mul_tensor_57" [id=1499, type=mul];
+"1500 zero_point_updated_constant57" [id=1500, type=get_attr];
+"1501 sub_tensor_57" [id=1501, type=sub];
+"1502 linear_84" [id=1502, type=linear];
+"1503 gelu_13" [id=1503, type=gelu];
+"1504 dropout_54" [id=1504, type=dropout];
+"1505 dropout_54_0_0_nncf_smooth_quant_0" [id=1505, type=call_module];
+"1506 quantize_per_tensor_default_86" [id=1506, type=quantize_per_tensor];
+"1507 dequantize_per_tensor_default_86" [id=1507, type=dequantize_per_tensor];
+"1508 features_5_9_mlp_3_bias_0_0" [id=1508, type=get_attr];
+"1509 scale_updated_constant58" [id=1509, type=get_attr];
+"1510 compressed_weight_updated_constant58" [id=1510, type=get_attr];
+"1511 mul_tensor_58" [id=1511, type=mul];
+"1512 zero_point_updated_constant58" [id=1512, type=get_attr];
+"1513 sub_tensor_58" [id=1513, type=sub];
+"1514 linear_85" [id=1514, type=linear];
+"1515 dropout_55" [id=1515, type=dropout];
+"1516 layer_norm_30" [id=1516, type=layer_norm];
+"1517 add_48" [id=1517, type=add];
+"1518 _frozen_param119" [id=1518, type=get_attr];
+"1519 pad_16" [id=1519, type=pad];
+"1520 view_79" [id=1520, type=view];
+"1521 permute_65" [id=1521, type=permute];
+"1522 reshape_63" [id=1522, type=reshape];
+"1523 reshape_63_0_0_nncf_smooth_quant_0" [id=1523, type=call_module];
+"1524 quantize_per_tensor_default_87" [id=1524, type=quantize_per_tensor];
+"1525 dequantize_per_tensor_default_87" [id=1525, type=dequantize_per_tensor];
+"1526 _frozen_param120_0_0" [id=1526, type=get_attr];
+"1527 scale_updated_constant59" [id=1527, type=get_attr];
+"1528 compressed_weight_updated_constant59" [id=1528, type=get_attr];
+"1529 mul_tensor_59" [id=1529, type=mul];
+"1530 zero_point_updated_constant59" [id=1530, type=get_attr];
+"1531 sub_tensor_59" [id=1531, type=sub];
+"1532 linear_88" [id=1532, type=linear];
+"1533 reshape_64" [id=1533, type=reshape];
+"1534 permute_66" [id=1534, type=permute];
+"1535 select_42" [id=1535, type=select];
+"1536 select_43" [id=1536, type=select];
+"1537 select_44" [id=1537, type=select];
+"1538 linalg_vector_norm_28" [id=1538, type=linalg_vector_norm];
+"1539 clamp_min_28" [id=1539, type=clamp_min];
+"1540 expand_as_28" [id=1540, type=expand_as];
+"1541 div_28" [id=1541, type=div];
+"1542 quantize_per_tensor_default_88" [id=1542, type=quantize_per_tensor];
+"1543 dequantize_per_tensor_default_88" [id=1543, type=dequantize_per_tensor];
+"1544 linalg_vector_norm_29" [id=1544, type=linalg_vector_norm];
+"1545 clamp_min_29" [id=1545, type=clamp_min];
+"1546 expand_as_29" [id=1546, type=expand_as];
+"1547 div_29" [id=1547, type=div];
+"1548 quantize_per_tensor_default_89" [id=1548, type=quantize_per_tensor];
+"1549 dequantize_per_tensor_default_89" [id=1549, type=dequantize_per_tensor];
+"1550 transpose_28" [id=1550, type=transpose];
+"1551 matmul_28" [id=1551, type=matmul];
+"1552 _frozen_param122" [id=1552, type=get_attr];
+"1553 mul_29" [id=1553, type=mul];
+"1554 add_49" [id=1554, type=add];
+"1555 softmax_14" [id=1555, type=softmax];
+"1556 dropout_56" [id=1556, type=dropout];
+"1557 matmul_29" [id=1557, type=matmul];
+"1558 transpose_29" [id=1558, type=transpose];
+"1559 reshape_65" [id=1559, type=reshape];
+"1560 reshape_65_0_0_nncf_smooth_quant_0" [id=1560, type=call_module];
+"1561 quantize_per_tensor_default_90" [id=1561, type=quantize_per_tensor];
+"1562 dequantize_per_tensor_default_90" [id=1562, type=dequantize_per_tensor];
+"1563 features_5_10_attn_proj_bias_0_0" [id=1563, type=get_attr];
+"1564 scale_updated_constant60" [id=1564, type=get_attr];
+"1565 compressed_weight_updated_constant60" [id=1565, type=get_attr];
+"1566 mul_tensor_60" [id=1566, type=mul];
+"1567 zero_point_updated_constant60" [id=1567, type=get_attr];
+"1568 sub_tensor_60" [id=1568, type=sub];
+"1569 linear_89" [id=1569, type=linear];
+"1570 dropout_57" [id=1570, type=dropout];
+"1571 view_80" [id=1571, type=view];
+"1572 permute_67" [id=1572, type=permute];
+"1573 reshape_66" [id=1573, type=reshape];
+"1574 slice_218" [id=1574, type=slice];
+"1575 slice_219" [id=1575, type=slice];
+"1576 slice_220" [id=1576, type=slice];
+"1577 slice_221" [id=1577, type=slice];
+"1578 contiguous_27" [id=1578, type=contiguous];
+"1579 layer_norm_31" [id=1579, type=layer_norm];
+"1580 add_50" [id=1580, type=add];
+"1581 add_50_0_0_nncf_smooth_quant_0" [id=1581, type=call_module];
+"1582 quantize_per_tensor_default_91" [id=1582, type=quantize_per_tensor];
+"1583 dequantize_per_tensor_default_91" [id=1583, type=dequantize_per_tensor];
+"1584 features_5_10_mlp_0_bias_0_0" [id=1584, type=get_attr];
+"1585 scale_updated_constant61" [id=1585, type=get_attr];
+"1586 compressed_weight_updated_constant61" [id=1586, type=get_attr];
+"1587 mul_tensor_61" [id=1587, type=mul];
+"1588 zero_point_updated_constant61" [id=1588, type=get_attr];
+"1589 sub_tensor_61" [id=1589, type=sub];
+"1590 linear_90" [id=1590, type=linear];
+"1591 gelu_14" [id=1591, type=gelu];
+"1592 dropout_58" [id=1592, type=dropout];
+"1593 dropout_58_0_0_nncf_smooth_quant_0" [id=1593, type=call_module];
+"1594 quantize_per_tensor_default_92" [id=1594, type=quantize_per_tensor];
+"1595 dequantize_per_tensor_default_92" [id=1595, type=dequantize_per_tensor];
+"1596 features_5_10_mlp_3_bias_0_0" [id=1596, type=get_attr];
+"1597 scale_updated_constant62" [id=1597, type=get_attr];
+"1598 compressed_weight_updated_constant62" [id=1598, type=get_attr];
+"1599 mul_tensor_62" [id=1599, type=mul];
+"1600 zero_point_updated_constant62" [id=1600, type=get_attr];
+"1601 sub_tensor_62" [id=1601, type=sub];
+"1602 linear_91" [id=1602, type=linear];
+"1603 dropout_59" [id=1603, type=dropout];
+"1604 layer_norm_32" [id=1604, type=layer_norm];
+"1605 add_51" [id=1605, type=add];
+"1606 _frozen_param123" [id=1606, type=get_attr];
+"1607 pad_17" [id=1607, type=pad];
+"1608 roll_14" [id=1608, type=roll];
+"1609 view_83" [id=1609, type=view];
+"1610 permute_69" [id=1610, type=permute];
+"1611 reshape_67" [id=1611, type=reshape];
+"1612 reshape_67_0_0_nncf_smooth_quant_0" [id=1612, type=call_module];
+"1613 quantize_per_tensor_default_93" [id=1613, type=quantize_per_tensor];
+"1614 dequantize_per_tensor_default_93" [id=1614, type=dequantize_per_tensor];
+"1615 _frozen_param124_0_0" [id=1615, type=get_attr];
+"1616 scale_updated_constant63" [id=1616, type=get_attr];
+"1617 compressed_weight_updated_constant63" [id=1617, type=get_attr];
+"1618 mul_tensor_63" [id=1618, type=mul];
+"1619 zero_point_updated_constant63" [id=1619, type=get_attr];
+"1620 sub_tensor_63" [id=1620, type=sub];
+"1621 linear_94" [id=1621, type=linear];
+"1622 reshape_68" [id=1622, type=reshape];
+"1623 permute_70" [id=1623, type=permute];
+"1624 select_45" [id=1624, type=select];
+"1625 select_46" [id=1625, type=select];
+"1626 select_47" [id=1626, type=select];
+"1627 linalg_vector_norm_30" [id=1627, type=linalg_vector_norm];
+"1628 clamp_min_30" [id=1628, type=clamp_min];
+"1629 expand_as_30" [id=1629, type=expand_as];
+"1630 div_30" [id=1630, type=div];
+"1631 quantize_per_tensor_default_94" [id=1631, type=quantize_per_tensor];
+"1632 dequantize_per_tensor_default_94" [id=1632, type=dequantize_per_tensor];
+"1633 linalg_vector_norm_31" [id=1633, type=linalg_vector_norm];
+"1634 clamp_min_31" [id=1634, type=clamp_min];
+"1635 expand_as_31" [id=1635, type=expand_as];
+"1636 div_31" [id=1636, type=div];
+"1637 quantize_per_tensor_default_95" [id=1637, type=quantize_per_tensor];
+"1638 dequantize_per_tensor_default_95" [id=1638, type=dequantize_per_tensor];
+"1639 transpose_30" [id=1639, type=transpose];
+"1640 matmul_30" [id=1640, type=matmul];
+"1641 _frozen_param126" [id=1641, type=get_attr];
+"1642 mul_31" [id=1642, type=mul];
+"1643 add_52" [id=1643, type=add];
+"1644 new_zeros_7" [id=1644, type=new_zeros];
+"1645 view_84" [id=1645, type=view];
+"1646 permute_71" [id=1646, type=permute];
+"1647 reshape_69" [id=1647, type=reshape];
+"1648 unsqueeze_44" [id=1648, type=unsqueeze];
+"1649 unsqueeze_45" [id=1649, type=unsqueeze];
+"1650 sub_7" [id=1650, type=sub];
+"1651 ne_7" [id=1651, type=ne];
+"1652 masked_fill_14" [id=1652, type=masked_fill];
+"1653 eq_7" [id=1653, type=eq];
+"1654 masked_fill_15" [id=1654, type=masked_fill];
+"1655 view_85" [id=1655, type=view];
+"1656 unsqueeze_46" [id=1656, type=unsqueeze];
+"1657 unsqueeze_47" [id=1657, type=unsqueeze];
+"1658 add_53" [id=1658, type=add];
+"1659 view_86" [id=1659, type=view];
+"1660 softmax_15" [id=1660, type=softmax];
+"1661 dropout_60" [id=1661, type=dropout];
+"1662 matmul_31" [id=1662, type=matmul];
+"1663 transpose_31" [id=1663, type=transpose];
+"1664 reshape_70" [id=1664, type=reshape];
+"1665 reshape_70_0_0_nncf_smooth_quant_0" [id=1665, type=call_module];
+"1666 quantize_per_tensor_default_96" [id=1666, type=quantize_per_tensor];
+"1667 dequantize_per_tensor_default_96" [id=1667, type=dequantize_per_tensor];
+"1668 features_5_11_attn_proj_bias_0_0" [id=1668, type=get_attr];
+"1669 scale_updated_constant64" [id=1669, type=get_attr];
+"1670 compressed_weight_updated_constant64" [id=1670, type=get_attr];
+"1671 mul_tensor_64" [id=1671, type=mul];
+"1672 zero_point_updated_constant64" [id=1672, type=get_attr];
+"1673 sub_tensor_64" [id=1673, type=sub];
+"1674 linear_95" [id=1674, type=linear];
+"1675 dropout_61" [id=1675, type=dropout];
+"1676 view_87" [id=1676, type=view];
+"1677 permute_72" [id=1677, type=permute];
+"1678 reshape_71" [id=1678, type=reshape];
+"1679 roll_15" [id=1679, type=roll];
+"1680 slice_241" [id=1680, type=slice];
+"1681 slice_242" [id=1681, type=slice];
+"1682 slice_243" [id=1682, type=slice];
+"1683 slice_244" [id=1683, type=slice];
+"1684 contiguous_29" [id=1684, type=contiguous];
+"1685 layer_norm_33" [id=1685, type=layer_norm];
+"1686 add_54" [id=1686, type=add];
+"1687 add_54_0_0_nncf_smooth_quant_0" [id=1687, type=call_module];
+"1688 quantize_per_tensor_default_97" [id=1688, type=quantize_per_tensor];
+"1689 dequantize_per_tensor_default_97" [id=1689, type=dequantize_per_tensor];
+"1690 features_5_11_mlp_0_bias_0_0" [id=1690, type=get_attr];
+"1691 scale_updated_constant65" [id=1691, type=get_attr];
+"1692 compressed_weight_updated_constant65" [id=1692, type=get_attr];
+"1693 mul_tensor_65" [id=1693, type=mul];
+"1694 zero_point_updated_constant65" [id=1694, type=get_attr];
+"1695 sub_tensor_65" [id=1695, type=sub];
+"1696 linear_96" [id=1696, type=linear];
+"1697 gelu_15" [id=1697, type=gelu];
+"1698 dropout_62" [id=1698, type=dropout];
+"1699 dropout_62_0_0_nncf_smooth_quant_0" [id=1699, type=call_module];
+"1700 quantize_per_tensor_default_98" [id=1700, type=quantize_per_tensor];
+"1701 dequantize_per_tensor_default_98" [id=1701, type=dequantize_per_tensor];
+"1702 features_5_11_mlp_3_bias_0_0" [id=1702, type=get_attr];
+"1703 scale_updated_constant66" [id=1703, type=get_attr];
+"1704 compressed_weight_updated_constant66" [id=1704, type=get_attr];
+"1705 mul_tensor_66" [id=1705, type=mul];
+"1706 zero_point_updated_constant66" [id=1706, type=get_attr];
+"1707 sub_tensor_66" [id=1707, type=sub];
+"1708 linear_97" [id=1708, type=linear];
+"1709 dropout_63" [id=1709, type=dropout];
+"1710 layer_norm_34" [id=1710, type=layer_norm];
+"1711 add_55" [id=1711, type=add];
+"1712 _frozen_param136" [id=1712, type=get_attr];
+"1713 pad_18" [id=1713, type=pad];
+"1714 view_90" [id=1714, type=view];
+"1715 permute_74" [id=1715, type=permute];
+"1716 reshape_72" [id=1716, type=reshape];
+"1717 reshape_72_0_0_nncf_smooth_quant_0" [id=1717, type=call_module];
+"1718 quantize_per_tensor_default_99" [id=1718, type=quantize_per_tensor];
+"1719 dequantize_per_tensor_default_99" [id=1719, type=dequantize_per_tensor];
+"1720 _frozen_param137_0_0" [id=1720, type=get_attr];
+"1721 scale_updated_constant67" [id=1721, type=get_attr];
+"1722 compressed_weight_updated_constant67" [id=1722, type=get_attr];
+"1723 mul_tensor_67" [id=1723, type=mul];
+"1724 zero_point_updated_constant67" [id=1724, type=get_attr];
+"1725 sub_tensor_67" [id=1725, type=sub];
+"1726 linear_100" [id=1726, type=linear];
+"1727 reshape_73" [id=1727, type=reshape];
+"1728 permute_75" [id=1728, type=permute];
+"1729 select_48" [id=1729, type=select];
+"1730 select_49" [id=1730, type=select];
+"1731 select_50" [id=1731, type=select];
+"1732 linalg_vector_norm_32" [id=1732, type=linalg_vector_norm];
+"1733 clamp_min_32" [id=1733, type=clamp_min];
+"1734 expand_as_32" [id=1734, type=expand_as];
+"1735 div_32" [id=1735, type=div];
+"1736 quantize_per_tensor_default_100" [id=1736, type=quantize_per_tensor];
+"1737 dequantize_per_tensor_default_100" [id=1737, type=dequantize_per_tensor];
+"1738 linalg_vector_norm_33" [id=1738, type=linalg_vector_norm];
+"1739 clamp_min_33" [id=1739, type=clamp_min];
+"1740 expand_as_33" [id=1740, type=expand_as];
+"1741 div_33" [id=1741, type=div];
+"1742 quantize_per_tensor_default_101" [id=1742, type=quantize_per_tensor];
+"1743 dequantize_per_tensor_default_101" [id=1743, type=dequantize_per_tensor];
+"1744 transpose_32" [id=1744, type=transpose];
+"1745 matmul_32" [id=1745, type=matmul];
+"1746 _frozen_param139" [id=1746, type=get_attr];
+"1747 mul_33" [id=1747, type=mul];
+"1748 add_56" [id=1748, type=add];
+"1749 softmax_16" [id=1749, type=softmax];
+"1750 dropout_64" [id=1750, type=dropout];
+"1751 matmul_33" [id=1751, type=matmul];
+"1752 transpose_33" [id=1752, type=transpose];
+"1753 reshape_74" [id=1753, type=reshape];
+"1754 reshape_74_0_0_nncf_smooth_quant_0" [id=1754, type=call_module];
+"1755 quantize_per_tensor_default_102" [id=1755, type=quantize_per_tensor];
+"1756 dequantize_per_tensor_default_102" [id=1756, type=dequantize_per_tensor];
+"1757 features_5_12_attn_proj_bias_0_0" [id=1757, type=get_attr];
+"1758 scale_updated_constant68" [id=1758, type=get_attr];
+"1759 compressed_weight_updated_constant68" [id=1759, type=get_attr];
+"1760 mul_tensor_68" [id=1760, type=mul];
+"1761 zero_point_updated_constant68" [id=1761, type=get_attr];
+"1762 sub_tensor_68" [id=1762, type=sub];
+"1763 linear_101" [id=1763, type=linear];
+"1764 dropout_65" [id=1764, type=dropout];
+"1765 view_91" [id=1765, type=view];
+"1766 permute_76" [id=1766, type=permute];
+"1767 reshape_75" [id=1767, type=reshape];
+"1768 slice_246" [id=1768, type=slice];
+"1769 slice_247" [id=1769, type=slice];
+"1770 slice_248" [id=1770, type=slice];
+"1771 slice_249" [id=1771, type=slice];
+"1772 contiguous_31" [id=1772, type=contiguous];
+"1773 layer_norm_35" [id=1773, type=layer_norm];
+"1774 add_57" [id=1774, type=add];
+"1775 add_57_0_0_nncf_smooth_quant_0" [id=1775, type=call_module];
+"1776 quantize_per_tensor_default_103" [id=1776, type=quantize_per_tensor];
+"1777 dequantize_per_tensor_default_103" [id=1777, type=dequantize_per_tensor];
+"1778 features_5_12_mlp_0_bias_0_0" [id=1778, type=get_attr];
+"1779 scale_updated_constant69" [id=1779, type=get_attr];
+"1780 compressed_weight_updated_constant69" [id=1780, type=get_attr];
+"1781 mul_tensor_69" [id=1781, type=mul];
+"1782 zero_point_updated_constant69" [id=1782, type=get_attr];
+"1783 sub_tensor_69" [id=1783, type=sub];
+"1784 linear_102" [id=1784, type=linear];
+"1785 gelu_16" [id=1785, type=gelu];
+"1786 dropout_66" [id=1786, type=dropout];
+"1787 dropout_66_0_0_nncf_smooth_quant_0" [id=1787, type=call_module];
+"1788 quantize_per_tensor_default_104" [id=1788, type=quantize_per_tensor];
+"1789 dequantize_per_tensor_default_104" [id=1789, type=dequantize_per_tensor];
+"1790 features_5_12_mlp_3_bias_0_0" [id=1790, type=get_attr];
+"1791 scale_updated_constant70" [id=1791, type=get_attr];
+"1792 compressed_weight_updated_constant70" [id=1792, type=get_attr];
+"1793 mul_tensor_70" [id=1793, type=mul];
+"1794 zero_point_updated_constant70" [id=1794, type=get_attr];
+"1795 sub_tensor_70" [id=1795, type=sub];
+"1796 linear_103" [id=1796, type=linear];
+"1797 dropout_67" [id=1797, type=dropout];
+"1798 layer_norm_36" [id=1798, type=layer_norm];
+"1799 add_58" [id=1799, type=add];
+"1800 _frozen_param140" [id=1800, type=get_attr];
+"1801 pad_19" [id=1801, type=pad];
+"1802 roll_16" [id=1802, type=roll];
+"1803 view_94" [id=1803, type=view];
+"1804 permute_78" [id=1804, type=permute];
+"1805 reshape_76" [id=1805, type=reshape];
+"1806 reshape_76_0_0_nncf_smooth_quant_0" [id=1806, type=call_module];
+"1807 quantize_per_tensor_default_105" [id=1807, type=quantize_per_tensor];
+"1808 dequantize_per_tensor_default_105" [id=1808, type=dequantize_per_tensor];
+"1809 _frozen_param141_0_0" [id=1809, type=get_attr];
+"1810 scale_updated_constant71" [id=1810, type=get_attr];
+"1811 compressed_weight_updated_constant71" [id=1811, type=get_attr];
+"1812 mul_tensor_71" [id=1812, type=mul];
+"1813 zero_point_updated_constant71" [id=1813, type=get_attr];
+"1814 sub_tensor_71" [id=1814, type=sub];
+"1815 linear_106" [id=1815, type=linear];
+"1816 reshape_77" [id=1816, type=reshape];
+"1817 permute_79" [id=1817, type=permute];
+"1818 select_51" [id=1818, type=select];
+"1819 select_52" [id=1819, type=select];
+"1820 select_53" [id=1820, type=select];
+"1821 linalg_vector_norm_34" [id=1821, type=linalg_vector_norm];
+"1822 clamp_min_34" [id=1822, type=clamp_min];
+"1823 expand_as_34" [id=1823, type=expand_as];
+"1824 div_34" [id=1824, type=div];
+"1825 quantize_per_tensor_default_106" [id=1825, type=quantize_per_tensor];
+"1826 dequantize_per_tensor_default_106" [id=1826, type=dequantize_per_tensor];
+"1827 linalg_vector_norm_35" [id=1827, type=linalg_vector_norm];
+"1828 clamp_min_35" [id=1828, type=clamp_min];
+"1829 expand_as_35" [id=1829, type=expand_as];
+"1830 div_35" [id=1830, type=div];
+"1831 quantize_per_tensor_default_107" [id=1831, type=quantize_per_tensor];
+"1832 dequantize_per_tensor_default_107" [id=1832, type=dequantize_per_tensor];
+"1833 transpose_34" [id=1833, type=transpose];
+"1834 matmul_34" [id=1834, type=matmul];
+"1835 _frozen_param143" [id=1835, type=get_attr];
+"1836 mul_35" [id=1836, type=mul];
+"1837 add_59" [id=1837, type=add];
+"1838 new_zeros_8" [id=1838, type=new_zeros];
+"1839 view_95" [id=1839, type=view];
+"1840 permute_80" [id=1840, type=permute];
+"1841 reshape_78" [id=1841, type=reshape];
+"1842 unsqueeze_50" [id=1842, type=unsqueeze];
+"1843 unsqueeze_51" [id=1843, type=unsqueeze];
+"1844 sub_8" [id=1844, type=sub];
+"1845 ne_8" [id=1845, type=ne];
+"1846 masked_fill_16" [id=1846, type=masked_fill];
+"1847 eq_8" [id=1847, type=eq];
+"1848 masked_fill_17" [id=1848, type=masked_fill];
+"1849 view_96" [id=1849, type=view];
+"1850 unsqueeze_52" [id=1850, type=unsqueeze];
+"1851 unsqueeze_53" [id=1851, type=unsqueeze];
+"1852 add_60" [id=1852, type=add];
+"1853 view_97" [id=1853, type=view];
+"1854 softmax_17" [id=1854, type=softmax];
+"1855 dropout_68" [id=1855, type=dropout];
+"1856 matmul_35" [id=1856, type=matmul];
+"1857 transpose_35" [id=1857, type=transpose];
+"1858 reshape_79" [id=1858, type=reshape];
+"1859 reshape_79_0_0_nncf_smooth_quant_0" [id=1859, type=call_module];
+"1860 quantize_per_tensor_default_108" [id=1860, type=quantize_per_tensor];
+"1861 dequantize_per_tensor_default_108" [id=1861, type=dequantize_per_tensor];
+"1862 features_5_13_attn_proj_bias_0_0" [id=1862, type=get_attr];
+"1863 scale_updated_constant72" [id=1863, type=get_attr];
+"1864 compressed_weight_updated_constant72" [id=1864, type=get_attr];
+"1865 mul_tensor_72" [id=1865, type=mul];
+"1866 zero_point_updated_constant72" [id=1866, type=get_attr];
+"1867 sub_tensor_72" [id=1867, type=sub];
+"1868 linear_107" [id=1868, type=linear];
+"1869 dropout_69" [id=1869, type=dropout];
+"1870 view_98" [id=1870, type=view];
+"1871 permute_81" [id=1871, type=permute];
+"1872 reshape_80" [id=1872, type=reshape];
+"1873 roll_17" [id=1873, type=roll];
+"1874 slice_269" [id=1874, type=slice];
+"1875 slice_270" [id=1875, type=slice];
+"1876 slice_271" [id=1876, type=slice];
+"1877 slice_272" [id=1877, type=slice];
+"1878 contiguous_33" [id=1878, type=contiguous];
+"1879 layer_norm_37" [id=1879, type=layer_norm];
+"1880 add_61" [id=1880, type=add];
+"1881 add_61_0_0_nncf_smooth_quant_0" [id=1881, type=call_module];
+"1882 quantize_per_tensor_default_109" [id=1882, type=quantize_per_tensor];
+"1883 dequantize_per_tensor_default_109" [id=1883, type=dequantize_per_tensor];
+"1884 features_5_13_mlp_0_bias_0_0" [id=1884, type=get_attr];
+"1885 scale_updated_constant73" [id=1885, type=get_attr];
+"1886 compressed_weight_updated_constant73" [id=1886, type=get_attr];
+"1887 mul_tensor_73" [id=1887, type=mul];
+"1888 zero_point_updated_constant73" [id=1888, type=get_attr];
+"1889 sub_tensor_73" [id=1889, type=sub];
+"1890 linear_108" [id=1890, type=linear];
+"1891 gelu_17" [id=1891, type=gelu];
+"1892 dropout_70" [id=1892, type=dropout];
+"1893 dropout_70_0_0_nncf_smooth_quant_0" [id=1893, type=call_module];
+"1894 quantize_per_tensor_default_110" [id=1894, type=quantize_per_tensor];
+"1895 dequantize_per_tensor_default_110" [id=1895, type=dequantize_per_tensor];
+"1896 features_5_13_mlp_3_bias_0_0" [id=1896, type=get_attr];
+"1897 scale_updated_constant74" [id=1897, type=get_attr];
+"1898 compressed_weight_updated_constant74" [id=1898, type=get_attr];
+"1899 mul_tensor_74" [id=1899, type=mul];
+"1900 zero_point_updated_constant74" [id=1900, type=get_attr];
+"1901 sub_tensor_74" [id=1901, type=sub];
+"1902 linear_109" [id=1902, type=linear];
+"1903 dropout_71" [id=1903, type=dropout];
+"1904 layer_norm_38" [id=1904, type=layer_norm];
+"1905 add_62" [id=1905, type=add];
+"1906 _frozen_param153" [id=1906, type=get_attr];
+"1907 pad_20" [id=1907, type=pad];
+"1908 view_101" [id=1908, type=view];
+"1909 permute_83" [id=1909, type=permute];
+"1910 reshape_81" [id=1910, type=reshape];
+"1911 reshape_81_0_0_nncf_smooth_quant_0" [id=1911, type=call_module];
+"1912 quantize_per_tensor_default_111" [id=1912, type=quantize_per_tensor];
+"1913 dequantize_per_tensor_default_111" [id=1913, type=dequantize_per_tensor];
+"1914 _frozen_param154_0_0" [id=1914, type=get_attr];
+"1915 scale_updated_constant75" [id=1915, type=get_attr];
+"1916 compressed_weight_updated_constant75" [id=1916, type=get_attr];
+"1917 mul_tensor_75" [id=1917, type=mul];
+"1918 zero_point_updated_constant75" [id=1918, type=get_attr];
+"1919 sub_tensor_75" [id=1919, type=sub];
+"1920 linear_112" [id=1920, type=linear];
+"1921 reshape_82" [id=1921, type=reshape];
+"1922 permute_84" [id=1922, type=permute];
+"1923 select_54" [id=1923, type=select];
+"1924 select_55" [id=1924, type=select];
+"1925 select_56" [id=1925, type=select];
+"1926 linalg_vector_norm_36" [id=1926, type=linalg_vector_norm];
+"1927 clamp_min_36" [id=1927, type=clamp_min];
+"1928 expand_as_36" [id=1928, type=expand_as];
+"1929 div_36" [id=1929, type=div];
+"1930 quantize_per_tensor_default_112" [id=1930, type=quantize_per_tensor];
+"1931 dequantize_per_tensor_default_112" [id=1931, type=dequantize_per_tensor];
+"1932 linalg_vector_norm_37" [id=1932, type=linalg_vector_norm];
+"1933 clamp_min_37" [id=1933, type=clamp_min];
+"1934 expand_as_37" [id=1934, type=expand_as];
+"1935 div_37" [id=1935, type=div];
+"1936 quantize_per_tensor_default_113" [id=1936, type=quantize_per_tensor];
+"1937 dequantize_per_tensor_default_113" [id=1937, type=dequantize_per_tensor];
+"1938 transpose_36" [id=1938, type=transpose];
+"1939 matmul_36" [id=1939, type=matmul];
+"1940 _frozen_param156" [id=1940, type=get_attr];
+"1941 mul_37" [id=1941, type=mul];
+"1942 add_63" [id=1942, type=add];
+"1943 softmax_18" [id=1943, type=softmax];
+"1944 dropout_72" [id=1944, type=dropout];
+"1945 matmul_37" [id=1945, type=matmul];
+"1946 transpose_37" [id=1946, type=transpose];
+"1947 reshape_83" [id=1947, type=reshape];
+"1948 reshape_83_0_0_nncf_smooth_quant_0" [id=1948, type=call_module];
+"1949 quantize_per_tensor_default_114" [id=1949, type=quantize_per_tensor];
+"1950 dequantize_per_tensor_default_114" [id=1950, type=dequantize_per_tensor];
+"1951 features_5_14_attn_proj_bias_0_0" [id=1951, type=get_attr];
+"1952 scale_updated_constant76" [id=1952, type=get_attr];
+"1953 compressed_weight_updated_constant76" [id=1953, type=get_attr];
+"1954 mul_tensor_76" [id=1954, type=mul];
+"1955 zero_point_updated_constant76" [id=1955, type=get_attr];
+"1956 sub_tensor_76" [id=1956, type=sub];
+"1957 linear_113" [id=1957, type=linear];
+"1958 dropout_73" [id=1958, type=dropout];
+"1959 view_102" [id=1959, type=view];
+"1960 permute_85" [id=1960, type=permute];
+"1961 reshape_84" [id=1961, type=reshape];
+"1962 slice_274" [id=1962, type=slice];
+"1963 slice_275" [id=1963, type=slice];
+"1964 slice_276" [id=1964, type=slice];
+"1965 slice_277" [id=1965, type=slice];
+"1966 contiguous_35" [id=1966, type=contiguous];
+"1967 layer_norm_39" [id=1967, type=layer_norm];
+"1968 add_64" [id=1968, type=add];
+"1969 add_64_0_0_nncf_smooth_quant_0" [id=1969, type=call_module];
+"1970 quantize_per_tensor_default_115" [id=1970, type=quantize_per_tensor];
+"1971 dequantize_per_tensor_default_115" [id=1971, type=dequantize_per_tensor];
+"1972 features_5_14_mlp_0_bias_0_0" [id=1972, type=get_attr];
+"1973 scale_updated_constant77" [id=1973, type=get_attr];
+"1974 compressed_weight_updated_constant77" [id=1974, type=get_attr];
+"1975 mul_tensor_77" [id=1975, type=mul];
+"1976 zero_point_updated_constant77" [id=1976, type=get_attr];
+"1977 sub_tensor_77" [id=1977, type=sub];
+"1978 linear_114" [id=1978, type=linear];
+"1979 gelu_18" [id=1979, type=gelu];
+"1980 dropout_74" [id=1980, type=dropout];
+"1981 dropout_74_0_0_nncf_smooth_quant_0" [id=1981, type=call_module];
+"1982 quantize_per_tensor_default_116" [id=1982, type=quantize_per_tensor];
+"1983 dequantize_per_tensor_default_116" [id=1983, type=dequantize_per_tensor];
+"1984 features_5_14_mlp_3_bias_0_0" [id=1984, type=get_attr];
+"1985 scale_updated_constant78" [id=1985, type=get_attr];
+"1986 compressed_weight_updated_constant78" [id=1986, type=get_attr];
+"1987 mul_tensor_78" [id=1987, type=mul];
+"1988 zero_point_updated_constant78" [id=1988, type=get_attr];
+"1989 sub_tensor_78" [id=1989, type=sub];
+"1990 linear_115" [id=1990, type=linear];
+"1991 dropout_75" [id=1991, type=dropout];
+"1992 layer_norm_40" [id=1992, type=layer_norm];
+"1993 add_65" [id=1993, type=add];
+"1994 _frozen_param157" [id=1994, type=get_attr];
+"1995 pad_21" [id=1995, type=pad];
+"1996 roll_18" [id=1996, type=roll];
+"1997 view_105" [id=1997, type=view];
+"1998 permute_87" [id=1998, type=permute];
+"1999 reshape_85" [id=1999, type=reshape];
+"2000 reshape_85_0_0_nncf_smooth_quant_0" [id=2000, type=call_module];
+"2001 quantize_per_tensor_default_117" [id=2001, type=quantize_per_tensor];
+"2002 dequantize_per_tensor_default_117" [id=2002, type=dequantize_per_tensor];
+"2003 _frozen_param158_0_0" [id=2003, type=get_attr];
+"2004 scale_updated_constant79" [id=2004, type=get_attr];
+"2005 compressed_weight_updated_constant79" [id=2005, type=get_attr];
+"2006 mul_tensor_79" [id=2006, type=mul];
+"2007 zero_point_updated_constant79" [id=2007, type=get_attr];
+"2008 sub_tensor_79" [id=2008, type=sub];
+"2009 linear_118" [id=2009, type=linear];
+"2010 reshape_86" [id=2010, type=reshape];
+"2011 permute_88" [id=2011, type=permute];
+"2012 select_57" [id=2012, type=select];
+"2013 select_58" [id=2013, type=select];
+"2014 select_59" [id=2014, type=select];
+"2015 linalg_vector_norm_38" [id=2015, type=linalg_vector_norm];
+"2016 clamp_min_38" [id=2016, type=clamp_min];
+"2017 expand_as_38" [id=2017, type=expand_as];
+"2018 div_38" [id=2018, type=div];
+"2019 quantize_per_tensor_default_118" [id=2019, type=quantize_per_tensor];
+"2020 dequantize_per_tensor_default_118" [id=2020, type=dequantize_per_tensor];
+"2021 linalg_vector_norm_39" [id=2021, type=linalg_vector_norm];
+"2022 clamp_min_39" [id=2022, type=clamp_min];
+"2023 expand_as_39" [id=2023, type=expand_as];
+"2024 div_39" [id=2024, type=div];
+"2025 quantize_per_tensor_default_119" [id=2025, type=quantize_per_tensor];
+"2026 dequantize_per_tensor_default_119" [id=2026, type=dequantize_per_tensor];
+"2027 transpose_38" [id=2027, type=transpose];
+"2028 matmul_38" [id=2028, type=matmul];
+"2029 _frozen_param160" [id=2029, type=get_attr];
+"2030 mul_39" [id=2030, type=mul];
+"2031 add_66" [id=2031, type=add];
+"2032 new_zeros_9" [id=2032, type=new_zeros];
+"2033 view_106" [id=2033, type=view];
+"2034 permute_89" [id=2034, type=permute];
+"2035 reshape_87" [id=2035, type=reshape];
+"2036 unsqueeze_56" [id=2036, type=unsqueeze];
+"2037 unsqueeze_57" [id=2037, type=unsqueeze];
+"2038 sub_9" [id=2038, type=sub];
+"2039 ne_9" [id=2039, type=ne];
+"2040 masked_fill_18" [id=2040, type=masked_fill];
+"2041 eq_9" [id=2041, type=eq];
+"2042 masked_fill_19" [id=2042, type=masked_fill];
+"2043 view_107" [id=2043, type=view];
+"2044 unsqueeze_58" [id=2044, type=unsqueeze];
+"2045 unsqueeze_59" [id=2045, type=unsqueeze];
+"2046 add_67" [id=2046, type=add];
+"2047 view_108" [id=2047, type=view];
+"2048 softmax_19" [id=2048, type=softmax];
+"2049 dropout_76" [id=2049, type=dropout];
+"2050 matmul_39" [id=2050, type=matmul];
+"2051 transpose_39" [id=2051, type=transpose];
+"2052 reshape_88" [id=2052, type=reshape];
+"2053 reshape_88_0_0_nncf_smooth_quant_0" [id=2053, type=call_module];
+"2054 quantize_per_tensor_default_120" [id=2054, type=quantize_per_tensor];
+"2055 dequantize_per_tensor_default_120" [id=2055, type=dequantize_per_tensor];
+"2056 features_5_15_attn_proj_bias_0_0" [id=2056, type=get_attr];
+"2057 scale_updated_constant80" [id=2057, type=get_attr];
+"2058 compressed_weight_updated_constant80" [id=2058, type=get_attr];
+"2059 mul_tensor_80" [id=2059, type=mul];
+"2060 zero_point_updated_constant80" [id=2060, type=get_attr];
+"2061 sub_tensor_80" [id=2061, type=sub];
+"2062 linear_119" [id=2062, type=linear];
+"2063 dropout_77" [id=2063, type=dropout];
+"2064 view_109" [id=2064, type=view];
+"2065 permute_90" [id=2065, type=permute];
+"2066 reshape_89" [id=2066, type=reshape];
+"2067 roll_19" [id=2067, type=roll];
+"2068 slice_297" [id=2068, type=slice];
+"2069 slice_298" [id=2069, type=slice];
+"2070 slice_299" [id=2070, type=slice];
+"2071 slice_300" [id=2071, type=slice];
+"2072 contiguous_37" [id=2072, type=contiguous];
+"2073 layer_norm_41" [id=2073, type=layer_norm];
+"2074 add_68" [id=2074, type=add];
+"2075 add_68_0_0_nncf_smooth_quant_0" [id=2075, type=call_module];
+"2076 quantize_per_tensor_default_121" [id=2076, type=quantize_per_tensor];
+"2077 dequantize_per_tensor_default_121" [id=2077, type=dequantize_per_tensor];
+"2078 features_5_15_mlp_0_bias_0_0" [id=2078, type=get_attr];
+"2079 scale_updated_constant81" [id=2079, type=get_attr];
+"2080 compressed_weight_updated_constant81" [id=2080, type=get_attr];
+"2081 mul_tensor_81" [id=2081, type=mul];
+"2082 zero_point_updated_constant81" [id=2082, type=get_attr];
+"2083 sub_tensor_81" [id=2083, type=sub];
+"2084 linear_120" [id=2084, type=linear];
+"2085 gelu_19" [id=2085, type=gelu];
+"2086 dropout_78" [id=2086, type=dropout];
+"2087 dropout_78_0_0_nncf_smooth_quant_0" [id=2087, type=call_module];
+"2088 quantize_per_tensor_default_122" [id=2088, type=quantize_per_tensor];
+"2089 dequantize_per_tensor_default_122" [id=2089, type=dequantize_per_tensor];
+"2090 features_5_15_mlp_3_bias_0_0" [id=2090, type=get_attr];
+"2091 scale_updated_constant82" [id=2091, type=get_attr];
+"2092 compressed_weight_updated_constant82" [id=2092, type=get_attr];
+"2093 mul_tensor_82" [id=2093, type=mul];
+"2094 zero_point_updated_constant82" [id=2094, type=get_attr];
+"2095 sub_tensor_82" [id=2095, type=sub];
+"2096 linear_121" [id=2096, type=linear];
+"2097 dropout_79" [id=2097, type=dropout];
+"2098 layer_norm_42" [id=2098, type=layer_norm];
+"2099 add_69" [id=2099, type=add];
+"2100 _frozen_param170" [id=2100, type=get_attr];
+"2101 pad_22" [id=2101, type=pad];
+"2102 view_112" [id=2102, type=view];
+"2103 permute_92" [id=2103, type=permute];
+"2104 reshape_90" [id=2104, type=reshape];
+"2105 reshape_90_0_0_nncf_smooth_quant_0" [id=2105, type=call_module];
+"2106 quantize_per_tensor_default_123" [id=2106, type=quantize_per_tensor];
+"2107 dequantize_per_tensor_default_123" [id=2107, type=dequantize_per_tensor];
+"2108 _frozen_param171_0_0" [id=2108, type=get_attr];
+"2109 scale_updated_constant83" [id=2109, type=get_attr];
+"2110 compressed_weight_updated_constant83" [id=2110, type=get_attr];
+"2111 mul_tensor_83" [id=2111, type=mul];
+"2112 zero_point_updated_constant83" [id=2112, type=get_attr];
+"2113 sub_tensor_83" [id=2113, type=sub];
+"2114 linear_124" [id=2114, type=linear];
+"2115 reshape_91" [id=2115, type=reshape];
+"2116 permute_93" [id=2116, type=permute];
+"2117 select_60" [id=2117, type=select];
+"2118 select_61" [id=2118, type=select];
+"2119 select_62" [id=2119, type=select];
+"2120 linalg_vector_norm_40" [id=2120, type=linalg_vector_norm];
+"2121 clamp_min_40" [id=2121, type=clamp_min];
+"2122 expand_as_40" [id=2122, type=expand_as];
+"2123 div_40" [id=2123, type=div];
+"2124 quantize_per_tensor_default_124" [id=2124, type=quantize_per_tensor];
+"2125 dequantize_per_tensor_default_124" [id=2125, type=dequantize_per_tensor];
+"2126 linalg_vector_norm_41" [id=2126, type=linalg_vector_norm];
+"2127 clamp_min_41" [id=2127, type=clamp_min];
+"2128 expand_as_41" [id=2128, type=expand_as];
+"2129 div_41" [id=2129, type=div];
+"2130 quantize_per_tensor_default_125" [id=2130, type=quantize_per_tensor];
+"2131 dequantize_per_tensor_default_125" [id=2131, type=dequantize_per_tensor];
+"2132 transpose_40" [id=2132, type=transpose];
+"2133 matmul_40" [id=2133, type=matmul];
+"2134 _frozen_param173" [id=2134, type=get_attr];
+"2135 mul_41" [id=2135, type=mul];
+"2136 add_70" [id=2136, type=add];
+"2137 softmax_20" [id=2137, type=softmax];
+"2138 dropout_80" [id=2138, type=dropout];
+"2139 matmul_41" [id=2139, type=matmul];
+"2140 transpose_41" [id=2140, type=transpose];
+"2141 reshape_92" [id=2141, type=reshape];
+"2142 reshape_92_0_0_nncf_smooth_quant_0" [id=2142, type=call_module];
+"2143 quantize_per_tensor_default_126" [id=2143, type=quantize_per_tensor];
+"2144 dequantize_per_tensor_default_126" [id=2144, type=dequantize_per_tensor];
+"2145 features_5_16_attn_proj_bias_0_0" [id=2145, type=get_attr];
+"2146 scale_updated_constant84" [id=2146, type=get_attr];
+"2147 compressed_weight_updated_constant84" [id=2147, type=get_attr];
+"2148 mul_tensor_84" [id=2148, type=mul];
+"2149 zero_point_updated_constant84" [id=2149, type=get_attr];
+"2150 sub_tensor_84" [id=2150, type=sub];
+"2151 linear_125" [id=2151, type=linear];
+"2152 dropout_81" [id=2152, type=dropout];
+"2153 view_113" [id=2153, type=view];
+"2154 permute_94" [id=2154, type=permute];
+"2155 reshape_93" [id=2155, type=reshape];
+"2156 slice_302" [id=2156, type=slice];
+"2157 slice_303" [id=2157, type=slice];
+"2158 slice_304" [id=2158, type=slice];
+"2159 slice_305" [id=2159, type=slice];
+"2160 contiguous_39" [id=2160, type=contiguous];
+"2161 layer_norm_43" [id=2161, type=layer_norm];
+"2162 add_71" [id=2162, type=add];
+"2163 add_71_0_0_nncf_smooth_quant_0" [id=2163, type=call_module];
+"2164 quantize_per_tensor_default_127" [id=2164, type=quantize_per_tensor];
+"2165 dequantize_per_tensor_default_127" [id=2165, type=dequantize_per_tensor];
+"2166 features_5_16_mlp_0_bias_0_0" [id=2166, type=get_attr];
+"2167 scale_updated_constant85" [id=2167, type=get_attr];
+"2168 compressed_weight_updated_constant85" [id=2168, type=get_attr];
+"2169 mul_tensor_85" [id=2169, type=mul];
+"2170 zero_point_updated_constant85" [id=2170, type=get_attr];
+"2171 sub_tensor_85" [id=2171, type=sub];
+"2172 linear_126" [id=2172, type=linear];
+"2173 gelu_20" [id=2173, type=gelu];
+"2174 dropout_82" [id=2174, type=dropout];
+"2175 dropout_82_0_0_nncf_smooth_quant_0" [id=2175, type=call_module];
+"2176 quantize_per_tensor_default_128" [id=2176, type=quantize_per_tensor];
+"2177 dequantize_per_tensor_default_128" [id=2177, type=dequantize_per_tensor];
+"2178 features_5_16_mlp_3_bias_0_0" [id=2178, type=get_attr];
+"2179 scale_updated_constant86" [id=2179, type=get_attr];
+"2180 compressed_weight_updated_constant86" [id=2180, type=get_attr];
+"2181 mul_tensor_86" [id=2181, type=mul];
+"2182 zero_point_updated_constant86" [id=2182, type=get_attr];
+"2183 sub_tensor_86" [id=2183, type=sub];
+"2184 linear_127" [id=2184, type=linear];
+"2185 dropout_83" [id=2185, type=dropout];
+"2186 layer_norm_44" [id=2186, type=layer_norm];
+"2187 add_72" [id=2187, type=add];
+"2188 _frozen_param174" [id=2188, type=get_attr];
+"2189 pad_23" [id=2189, type=pad];
+"2190 roll_20" [id=2190, type=roll];
+"2191 view_116" [id=2191, type=view];
+"2192 permute_96" [id=2192, type=permute];
+"2193 reshape_94" [id=2193, type=reshape];
+"2194 reshape_94_0_0_nncf_smooth_quant_0" [id=2194, type=call_module];
+"2195 quantize_per_tensor_default_129" [id=2195, type=quantize_per_tensor];
+"2196 dequantize_per_tensor_default_129" [id=2196, type=dequantize_per_tensor];
+"2197 _frozen_param175_0_0" [id=2197, type=get_attr];
+"2198 scale_updated_constant87" [id=2198, type=get_attr];
+"2199 compressed_weight_updated_constant87" [id=2199, type=get_attr];
+"2200 mul_tensor_87" [id=2200, type=mul];
+"2201 zero_point_updated_constant87" [id=2201, type=get_attr];
+"2202 sub_tensor_87" [id=2202, type=sub];
+"2203 linear_130" [id=2203, type=linear];
+"2204 reshape_95" [id=2204, type=reshape];
+"2205 permute_97" [id=2205, type=permute];
+"2206 select_63" [id=2206, type=select];
+"2207 select_64" [id=2207, type=select];
+"2208 select_65" [id=2208, type=select];
+"2209 linalg_vector_norm_42" [id=2209, type=linalg_vector_norm];
+"2210 clamp_min_42" [id=2210, type=clamp_min];
+"2211 expand_as_42" [id=2211, type=expand_as];
+"2212 div_42" [id=2212, type=div];
+"2213 quantize_per_tensor_default_130" [id=2213, type=quantize_per_tensor];
+"2214 dequantize_per_tensor_default_130" [id=2214, type=dequantize_per_tensor];
+"2215 linalg_vector_norm_43" [id=2215, type=linalg_vector_norm];
+"2216 clamp_min_43" [id=2216, type=clamp_min];
+"2217 expand_as_43" [id=2217, type=expand_as];
+"2218 div_43" [id=2218, type=div];
+"2219 quantize_per_tensor_default_131" [id=2219, type=quantize_per_tensor];
+"2220 dequantize_per_tensor_default_131" [id=2220, type=dequantize_per_tensor];
+"2221 transpose_42" [id=2221, type=transpose];
+"2222 matmul_42" [id=2222, type=matmul];
+"2223 _frozen_param177" [id=2223, type=get_attr];
+"2224 mul_43" [id=2224, type=mul];
+"2225 add_73" [id=2225, type=add];
+"2226 new_zeros_10" [id=2226, type=new_zeros];
+"2227 view_117" [id=2227, type=view];
+"2228 permute_98" [id=2228, type=permute];
+"2229 reshape_96" [id=2229, type=reshape];
+"2230 unsqueeze_62" [id=2230, type=unsqueeze];
+"2231 unsqueeze_63" [id=2231, type=unsqueeze];
+"2232 sub_10" [id=2232, type=sub];
+"2233 ne_10" [id=2233, type=ne];
+"2234 masked_fill_20" [id=2234, type=masked_fill];
+"2235 eq_10" [id=2235, type=eq];
+"2236 masked_fill_21" [id=2236, type=masked_fill];
+"2237 view_118" [id=2237, type=view];
+"2238 unsqueeze_64" [id=2238, type=unsqueeze];
+"2239 unsqueeze_65" [id=2239, type=unsqueeze];
+"2240 add_74" [id=2240, type=add];
+"2241 view_119" [id=2241, type=view];
+"2242 softmax_21" [id=2242, type=softmax];
+"2243 dropout_84" [id=2243, type=dropout];
+"2244 matmul_43" [id=2244, type=matmul];
+"2245 transpose_43" [id=2245, type=transpose];
+"2246 reshape_97" [id=2246, type=reshape];
+"2247 reshape_97_0_0_nncf_smooth_quant_0" [id=2247, type=call_module];
+"2248 quantize_per_tensor_default_132" [id=2248, type=quantize_per_tensor];
+"2249 dequantize_per_tensor_default_132" [id=2249, type=dequantize_per_tensor];
+"2250 features_5_17_attn_proj_bias_0_0" [id=2250, type=get_attr];
+"2251 scale_updated_constant88" [id=2251, type=get_attr];
+"2252 compressed_weight_updated_constant88" [id=2252, type=get_attr];
+"2253 mul_tensor_88" [id=2253, type=mul];
+"2254 zero_point_updated_constant88" [id=2254, type=get_attr];
+"2255 sub_tensor_88" [id=2255, type=sub];
+"2256 linear_131" [id=2256, type=linear];
+"2257 dropout_85" [id=2257, type=dropout];
+"2258 view_120" [id=2258, type=view];
+"2259 permute_99" [id=2259, type=permute];
+"2260 reshape_98" [id=2260, type=reshape];
+"2261 roll_21" [id=2261, type=roll];
+"2262 slice_325" [id=2262, type=slice];
+"2263 slice_326" [id=2263, type=slice];
+"2264 slice_327" [id=2264, type=slice];
+"2265 slice_328" [id=2265, type=slice];
+"2266 contiguous_41" [id=2266, type=contiguous];
+"2267 layer_norm_45" [id=2267, type=layer_norm];
+"2268 add_75" [id=2268, type=add];
+"2269 add_75_0_0_nncf_smooth_quant_0" [id=2269, type=call_module];
+"2270 quantize_per_tensor_default_133" [id=2270, type=quantize_per_tensor];
+"2271 dequantize_per_tensor_default_133" [id=2271, type=dequantize_per_tensor];
+"2272 features_5_17_mlp_0_bias_0_0" [id=2272, type=get_attr];
+"2273 scale_updated_constant89" [id=2273, type=get_attr];
+"2274 compressed_weight_updated_constant89" [id=2274, type=get_attr];
+"2275 mul_tensor_89" [id=2275, type=mul];
+"2276 zero_point_updated_constant89" [id=2276, type=get_attr];
+"2277 sub_tensor_89" [id=2277, type=sub];
+"2278 linear_132" [id=2278, type=linear];
+"2279 gelu_21" [id=2279, type=gelu];
+"2280 dropout_86" [id=2280, type=dropout];
+"2281 dropout_86_0_0_nncf_smooth_quant_0" [id=2281, type=call_module];
+"2282 quantize_per_tensor_default_134" [id=2282, type=quantize_per_tensor];
+"2283 dequantize_per_tensor_default_134" [id=2283, type=dequantize_per_tensor];
+"2284 features_5_17_mlp_3_bias_0_0" [id=2284, type=get_attr];
+"2285 scale_updated_constant90" [id=2285, type=get_attr];
+"2286 compressed_weight_updated_constant90" [id=2286, type=get_attr];
+"2287 mul_tensor_90" [id=2287, type=mul];
+"2288 zero_point_updated_constant90" [id=2288, type=get_attr];
+"2289 sub_tensor_90" [id=2289, type=sub];
+"2290 linear_133" [id=2290, type=linear];
+"2291 dropout_87" [id=2291, type=dropout];
+"2292 layer_norm_46" [id=2292, type=layer_norm];
+"2293 add_76" [id=2293, type=add];
+"2294 pad_24" [id=2294, type=pad];
+"2295 slice_329" [id=2295, type=slice];
+"2296 slice_330" [id=2296, type=slice];
+"2297 slice_331" [id=2297, type=slice];
+"2298 slice_332" [id=2298, type=slice];
+"2299 slice_333" [id=2299, type=slice];
+"2300 slice_334" [id=2300, type=slice];
+"2301 slice_335" [id=2301, type=slice];
+"2302 slice_336" [id=2302, type=slice];
+"2303 slice_337" [id=2303, type=slice];
+"2304 slice_338" [id=2304, type=slice];
+"2305 slice_339" [id=2305, type=slice];
+"2306 slice_340" [id=2306, type=slice];
+"2307 cat_2" [id=2307, type=cat];
+"2308 cat_2_0_0_nncf_smooth_quant_0" [id=2308, type=call_module];
+"2309 quantize_per_tensor_default_135" [id=2309, type=quantize_per_tensor];
+"2310 dequantize_per_tensor_default_135" [id=2310, type=dequantize_per_tensor];
+"2311 scale_updated_constant91" [id=2311, type=get_attr];
+"2312 compressed_weight_updated_constant91" [id=2312, type=get_attr];
+"2313 mul_tensor_91" [id=2313, type=mul];
+"2314 zero_point_updated_constant91" [id=2314, type=get_attr];
+"2315 sub_tensor_91" [id=2315, type=sub];
+"2316 linear_134" [id=2316, type=linear];
+"2317 layer_norm_47" [id=2317, type=layer_norm];
+"2318 _frozen_param187" [id=2318, type=get_attr];
+"2319 pad_25" [id=2319, type=pad];
+"2320 view_123" [id=2320, type=view];
+"2321 permute_101" [id=2321, type=permute];
+"2322 reshape_99" [id=2322, type=reshape];
+"2323 reshape_99_0_0_nncf_smooth_quant_0" [id=2323, type=call_module];
+"2324 quantize_per_tensor_default_136" [id=2324, type=quantize_per_tensor];
+"2325 dequantize_per_tensor_default_136" [id=2325, type=dequantize_per_tensor];
+"2326 _frozen_param188_0_0" [id=2326, type=get_attr];
+"2327 scale_updated_constant92" [id=2327, type=get_attr];
+"2328 compressed_weight_updated_constant92" [id=2328, type=get_attr];
+"2329 mul_tensor_92" [id=2329, type=mul];
+"2330 zero_point_updated_constant92" [id=2330, type=get_attr];
+"2331 sub_tensor_92" [id=2331, type=sub];
+"2332 linear_137" [id=2332, type=linear];
+"2333 reshape_100" [id=2333, type=reshape];
+"2334 permute_102" [id=2334, type=permute];
+"2335 select_66" [id=2335, type=select];
+"2336 select_67" [id=2336, type=select];
+"2337 select_68" [id=2337, type=select];
+"2338 linalg_vector_norm_44" [id=2338, type=linalg_vector_norm];
+"2339 clamp_min_44" [id=2339, type=clamp_min];
+"2340 expand_as_44" [id=2340, type=expand_as];
+"2341 div_44" [id=2341, type=div];
+"2342 quantize_per_tensor_default_137" [id=2342, type=quantize_per_tensor];
+"2343 dequantize_per_tensor_default_137" [id=2343, type=dequantize_per_tensor];
+"2344 linalg_vector_norm_45" [id=2344, type=linalg_vector_norm];
+"2345 clamp_min_45" [id=2345, type=clamp_min];
+"2346 expand_as_45" [id=2346, type=expand_as];
+"2347 div_45" [id=2347, type=div];
+"2348 quantize_per_tensor_default_138" [id=2348, type=quantize_per_tensor];
+"2349 dequantize_per_tensor_default_138" [id=2349, type=dequantize_per_tensor];
+"2350 transpose_44" [id=2350, type=transpose];
+"2351 matmul_44" [id=2351, type=matmul];
+"2352 _frozen_param190" [id=2352, type=get_attr];
+"2353 mul_45" [id=2353, type=mul];
+"2354 add_77" [id=2354, type=add];
+"2355 softmax_22" [id=2355, type=softmax];
+"2356 dropout_88" [id=2356, type=dropout];
+"2357 matmul_45" [id=2357, type=matmul];
+"2358 transpose_45" [id=2358, type=transpose];
+"2359 reshape_101" [id=2359, type=reshape];
+"2360 reshape_101_0_0_nncf_smooth_quant_0" [id=2360, type=call_module];
+"2361 quantize_per_tensor_default_139" [id=2361, type=quantize_per_tensor];
+"2362 dequantize_per_tensor_default_139" [id=2362, type=dequantize_per_tensor];
+"2363 features_7_0_attn_proj_bias_0_0" [id=2363, type=get_attr];
+"2364 scale_updated_constant93" [id=2364, type=get_attr];
+"2365 compressed_weight_updated_constant93" [id=2365, type=get_attr];
+"2366 mul_tensor_93" [id=2366, type=mul];
+"2367 zero_point_updated_constant93" [id=2367, type=get_attr];
+"2368 sub_tensor_93" [id=2368, type=sub];
+"2369 linear_138" [id=2369, type=linear];
+"2370 dropout_89" [id=2370, type=dropout];
+"2371 view_124" [id=2371, type=view];
+"2372 permute_103" [id=2372, type=permute];
+"2373 reshape_102" [id=2373, type=reshape];
+"2374 slice_342" [id=2374, type=slice];
+"2375 slice_343" [id=2375, type=slice];
+"2376 slice_344" [id=2376, type=slice];
+"2377 slice_345" [id=2377, type=slice];
+"2378 contiguous_43" [id=2378, type=contiguous];
+"2379 layer_norm_48" [id=2379, type=layer_norm];
+"2380 add_78" [id=2380, type=add];
+"2381 add_78_0_0_nncf_smooth_quant_0" [id=2381, type=call_module];
+"2382 quantize_per_tensor_default_140" [id=2382, type=quantize_per_tensor];
+"2383 dequantize_per_tensor_default_140" [id=2383, type=dequantize_per_tensor];
+"2384 features_7_0_mlp_0_bias_0_0" [id=2384, type=get_attr];
+"2385 scale_updated_constant94" [id=2385, type=get_attr];
+"2386 compressed_weight_updated_constant94" [id=2386, type=get_attr];
+"2387 mul_tensor_94" [id=2387, type=mul];
+"2388 zero_point_updated_constant94" [id=2388, type=get_attr];
+"2389 sub_tensor_94" [id=2389, type=sub];
+"2390 linear_139" [id=2390, type=linear];
+"2391 gelu_22" [id=2391, type=gelu];
+"2392 dropout_90" [id=2392, type=dropout];
+"2393 dropout_90_0_0_nncf_smooth_quant_0" [id=2393, type=call_module];
+"2394 quantize_per_tensor_default_141" [id=2394, type=quantize_per_tensor];
+"2395 dequantize_per_tensor_default_141" [id=2395, type=dequantize_per_tensor];
+"2396 features_7_0_mlp_3_bias_0_0" [id=2396, type=get_attr];
+"2397 scale_updated_constant95" [id=2397, type=get_attr];
+"2398 compressed_weight_updated_constant95" [id=2398, type=get_attr];
+"2399 mul_tensor_95" [id=2399, type=mul];
+"2400 zero_point_updated_constant95" [id=2400, type=get_attr];
+"2401 sub_tensor_95" [id=2401, type=sub];
+"2402 linear_140" [id=2402, type=linear];
+"2403 dropout_91" [id=2403, type=dropout];
+"2404 layer_norm_49" [id=2404, type=layer_norm];
+"2405 add_79" [id=2405, type=add];
+"2406 _frozen_param191" [id=2406, type=get_attr];
+"2407 pad_26" [id=2407, type=pad];
+"2408 view_127" [id=2408, type=view];
+"2409 permute_105" [id=2409, type=permute];
+"2410 reshape_103" [id=2410, type=reshape];
+"2411 reshape_103_0_0_nncf_smooth_quant_0" [id=2411, type=call_module];
+"2412 quantize_per_tensor_default_142" [id=2412, type=quantize_per_tensor];
+"2413 dequantize_per_tensor_default_142" [id=2413, type=dequantize_per_tensor];
+"2414 _frozen_param192_0_0" [id=2414, type=get_attr];
+"2415 scale_updated_constant96" [id=2415, type=get_attr];
+"2416 compressed_weight_updated_constant96" [id=2416, type=get_attr];
+"2417 mul_tensor_96" [id=2417, type=mul];
+"2418 zero_point_updated_constant96" [id=2418, type=get_attr];
+"2419 sub_tensor_96" [id=2419, type=sub];
+"2420 linear_143" [id=2420, type=linear];
+"2421 reshape_104" [id=2421, type=reshape];
+"2422 permute_106" [id=2422, type=permute];
+"2423 select_69" [id=2423, type=select];
+"2424 select_70" [id=2424, type=select];
+"2425 select_71" [id=2425, type=select];
+"2426 linalg_vector_norm_46" [id=2426, type=linalg_vector_norm];
+"2427 clamp_min_46" [id=2427, type=clamp_min];
+"2428 expand_as_46" [id=2428, type=expand_as];
+"2429 div_46" [id=2429, type=div];
+"2430 quantize_per_tensor_default_143" [id=2430, type=quantize_per_tensor];
+"2431 dequantize_per_tensor_default_143" [id=2431, type=dequantize_per_tensor];
+"2432 linalg_vector_norm_47" [id=2432, type=linalg_vector_norm];
+"2433 clamp_min_47" [id=2433, type=clamp_min];
+"2434 expand_as_47" [id=2434, type=expand_as];
+"2435 div_47" [id=2435, type=div];
+"2436 quantize_per_tensor_default_144" [id=2436, type=quantize_per_tensor];
+"2437 dequantize_per_tensor_default_144" [id=2437, type=dequantize_per_tensor];
+"2438 transpose_46" [id=2438, type=transpose];
+"2439 matmul_46" [id=2439, type=matmul];
+"2440 _frozen_param194" [id=2440, type=get_attr];
+"2441 mul_47" [id=2441, type=mul];
+"2442 add_80" [id=2442, type=add];
+"2443 softmax_23" [id=2443, type=softmax];
+"2444 dropout_92" [id=2444, type=dropout];
+"2445 matmul_47" [id=2445, type=matmul];
+"2446 transpose_47" [id=2446, type=transpose];
+"2447 reshape_105" [id=2447, type=reshape];
+"2448 reshape_105_0_0_nncf_smooth_quant_0" [id=2448, type=call_module];
+"2449 quantize_per_tensor_default_145" [id=2449, type=quantize_per_tensor];
+"2450 dequantize_per_tensor_default_145" [id=2450, type=dequantize_per_tensor];
+"2451 features_7_1_attn_proj_bias_0_0" [id=2451, type=get_attr];
+"2452 scale_updated_constant97" [id=2452, type=get_attr];
+"2453 compressed_weight_updated_constant97" [id=2453, type=get_attr];
+"2454 mul_tensor_97" [id=2454, type=mul];
+"2455 zero_point_updated_constant97" [id=2455, type=get_attr];
+"2456 sub_tensor_97" [id=2456, type=sub];
+"2457 linear_144" [id=2457, type=linear];
+"2458 dropout_93" [id=2458, type=dropout];
+"2459 view_128" [id=2459, type=view];
+"2460 permute_107" [id=2460, type=permute];
+"2461 reshape_106" [id=2461, type=reshape];
+"2462 slice_347" [id=2462, type=slice];
+"2463 slice_348" [id=2463, type=slice];
+"2464 slice_349" [id=2464, type=slice];
+"2465 slice_350" [id=2465, type=slice];
+"2466 contiguous_45" [id=2466, type=contiguous];
+"2467 layer_norm_50" [id=2467, type=layer_norm];
+"2468 add_81" [id=2468, type=add];
+"2469 add_81_0_0_nncf_smooth_quant_0" [id=2469, type=call_module];
+"2470 quantize_per_tensor_default_146" [id=2470, type=quantize_per_tensor];
+"2471 dequantize_per_tensor_default_146" [id=2471, type=dequantize_per_tensor];
+"2472 features_7_1_mlp_0_bias_0_0" [id=2472, type=get_attr];
+"2473 scale_updated_constant98" [id=2473, type=get_attr];
+"2474 compressed_weight_updated_constant98" [id=2474, type=get_attr];
+"2475 mul_tensor_98" [id=2475, type=mul];
+"2476 zero_point_updated_constant98" [id=2476, type=get_attr];
+"2477 sub_tensor_98" [id=2477, type=sub];
+"2478 linear_145" [id=2478, type=linear];
+"2479 gelu_23" [id=2479, type=gelu];
+"2480 dropout_94" [id=2480, type=dropout];
+"2481 dropout_94_0_0_nncf_smooth_quant_0" [id=2481, type=call_module];
+"2482 quantize_per_tensor_default_147" [id=2482, type=quantize_per_tensor];
+"2483 dequantize_per_tensor_default_147" [id=2483, type=dequantize_per_tensor];
+"2484 features_7_1_mlp_3_bias_0_0" [id=2484, type=get_attr];
+"2485 scale_updated_constant99" [id=2485, type=get_attr];
+"2486 compressed_weight_updated_constant99" [id=2486, type=get_attr];
+"2487 mul_tensor_99" [id=2487, type=mul];
+"2488 zero_point_updated_constant99" [id=2488, type=get_attr];
+"2489 sub_tensor_99" [id=2489, type=sub];
+"2490 linear_146" [id=2490, type=linear];
+"2491 dropout_95" [id=2491, type=dropout];
+"2492 layer_norm_51" [id=2492, type=layer_norm];
+"2493 add_82" [id=2493, type=add];
 "2494 layer_norm_52" [id=2494, type=layer_norm];
 "2495 permute_108" [id=2495, type=permute];
 "2496 adaptive_avg_pool2d" [id=2496, type=adaptive_avg_pool2d];
@@ -2500,7 +2500,7 @@ strict digraph  {
 "2498 flatten_0_0_nncf_smooth_quant_0" [id=2498, type=call_module];
 "2499 quantize_per_tensor_default_148" [id=2499, type=quantize_per_tensor];
 "2500 dequantize_per_tensor_default_148" [id=2500, type=dequantize_per_tensor];
-"2501 _param_constant400_0_0" [id=2501, type=get_attr];
+"2501 head_bias_0_0" [id=2501, type=get_attr];
 "2502 scale_updated_constant100" [id=2502, type=get_attr];
 "2503 compressed_weight_updated_constant100" [id=2503, type=get_attr];
 "2504 mul_tensor_100" [id=2504, type=mul];
@@ -2508,2745 +2508,2745 @@ strict digraph  {
 "2506 sub_tensor_100" [id=2506, type=sub];
 "2507 linear_147" [id=2507, type=linear];
 "2508 output" [id=2508, type=output];
-"0 arg0_1" -> "1 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"1 quantize_per_tensor_default" -> "2 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"2 dequantize_per_tensor_default" -> "9 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"3 _param_constant1_0_0" -> "9 conv2d"  [label="(96,)", style=solid];
-"4 scale_updated_constant0" -> "6 mul_tensor"  [label="(96, 1, 1, 1)", style=solid];
-"5 compressed_weight_updated_constant0" -> "6 mul_tensor"  [label="(96, 3, 4, 4)", style=solid];
-"6 mul_tensor" -> "8 sub_tensor"  [label="(96, 3, 4, 4)", style=solid];
-"7 zero_point_updated_constant0" -> "8 sub_tensor"  [label="(96, 1, 1, 1)", style=solid];
-"8 sub_tensor" -> "9 conv2d"  [label="(96, 3, 4, 4)", style=solid];
-"9 conv2d" -> "10 permute"  [label="(1, 96, 56, 56)", style=solid];
-"10 permute" -> "13 layer_norm"  [label="(1, 56, 56, 96)", style=solid];
-"11 _param_constant2" -> "13 layer_norm"  [label="(96,)", style=solid];
-"12 _param_constant3" -> "13 layer_norm"  [label="(96,)", style=solid];
-"13 layer_norm" -> "15 pad"  [label="(1, 56, 56, 96)", style=solid];
-"13 layer_norm" -> "75 add_1"  [label="(1, 56, 56, 96)", style=solid];
-"14 _frozen_param0" -> "50 add"  [label="(1, 3, 64, 64)", style=solid];
-"15 pad" -> "16 view_2"  [label="(1, 56, 56, 96)", style=solid];
-"16 view_2" -> "17 permute_2"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"17 permute_2" -> "18 reshape"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"18 reshape" -> "19 reshape_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
-"19 reshape_0_0_nncf_smooth_quant_0" -> "20 quantize_per_tensor_default_1"  [label="(49, 64, 96)", style=solid];
-"20 quantize_per_tensor_default_1" -> "21 dequantize_per_tensor_default_1"  [label="(49, 64, 96)", style=solid];
-"21 dequantize_per_tensor_default_1" -> "28 linear_2"  [label="(49, 64, 96)", style=solid];
-"22 _frozen_param1_0_0" -> "28 linear_2"  [label="(288,)", style=solid];
-"23 scale_updated_constant1" -> "25 mul_tensor_1"  [label="(288, 1)", style=solid];
-"24 compressed_weight_updated_constant1" -> "25 mul_tensor_1"  [label="(288, 96)", style=solid];
-"25 mul_tensor_1" -> "27 sub_tensor_1"  [label="(288, 96)", style=solid];
-"26 zero_point_updated_constant1" -> "27 sub_tensor_1"  [label="(288, 1)", style=solid];
-"27 sub_tensor_1" -> "28 linear_2"  [label="(288, 96)", style=solid];
-"28 linear_2" -> "29 reshape_1"  [label="(49, 64, 288)", style=solid];
-"29 reshape_1" -> "30 permute_3"  [label="(49, 64, 3, 3, 32)", style=solid];
-"30 permute_3" -> "31 select"  [label="(3, 49, 3, 64, 32)", style=solid];
-"30 permute_3" -> "32 select_1"  [label="(3, 49, 3, 64, 32)", style=solid];
-"30 permute_3" -> "33 select_2"  [label="(3, 49, 3, 64, 32)", style=solid];
-"31 select" -> "34 linalg_vector_norm"  [label="(49, 3, 64, 32)", style=solid];
-"31 select" -> "36 expand_as"  [label="(49, 3, 64, 32)", style=solid];
-"31 select" -> "37 div"  [label="(49, 3, 64, 32)", style=solid];
-"32 select_1" -> "40 linalg_vector_norm_1"  [label="(49, 3, 64, 32)", style=solid];
-"32 select_1" -> "42 expand_as_1"  [label="(49, 3, 64, 32)", style=solid];
-"32 select_1" -> "43 div_1"  [label="(49, 3, 64, 32)", style=solid];
-"33 select_2" -> "53 matmul_1"  [label="(49, 3, 64, 32)", style=solid];
-"34 linalg_vector_norm" -> "35 clamp_min"  [label="(49, 3, 64, 1)", style=solid];
-"35 clamp_min" -> "36 expand_as"  [label="(49, 3, 64, 1)", style=solid];
-"36 expand_as" -> "37 div"  [label="(49, 3, 64, 32)", style=solid];
-"37 div" -> "38 quantize_per_tensor_default_2"  [label="(49, 3, 64, 32)", style=solid];
-"38 quantize_per_tensor_default_2" -> "39 dequantize_per_tensor_default_2"  [label="(49, 3, 64, 32)", style=solid];
-"39 dequantize_per_tensor_default_2" -> "47 matmul"  [label="(49, 3, 64, 32)", style=solid];
-"40 linalg_vector_norm_1" -> "41 clamp_min_1"  [label="(49, 3, 64, 1)", style=solid];
-"41 clamp_min_1" -> "42 expand_as_1"  [label="(49, 3, 64, 1)", style=solid];
-"42 expand_as_1" -> "43 div_1"  [label="(49, 3, 64, 32)", style=solid];
-"43 div_1" -> "44 quantize_per_tensor_default_3"  [label="(49, 3, 64, 32)", style=solid];
-"44 quantize_per_tensor_default_3" -> "45 dequantize_per_tensor_default_3"  [label="(49, 3, 64, 32)", style=solid];
-"45 dequantize_per_tensor_default_3" -> "46 transpose"  [label="(49, 3, 64, 32)", style=solid];
-"46 transpose" -> "47 matmul"  [label="(49, 3, 32, 64)", style=solid];
-"47 matmul" -> "49 mul_1"  [label="(49, 3, 64, 64)", style=solid];
-"48 _frozen_param3" -> "49 mul_1"  [label="(3, 1, 1)", style=solid];
-"49 mul_1" -> "50 add"  [label="(49, 3, 64, 64)", style=solid];
-"50 add" -> "51 softmax"  [label="(49, 3, 64, 64)", style=solid];
-"51 softmax" -> "52 dropout"  [label="(49, 3, 64, 64)", style=solid];
-"52 dropout" -> "53 matmul_1"  [label="(49, 3, 64, 64)", style=solid];
-"53 matmul_1" -> "54 transpose_1"  [label="(49, 3, 64, 32)", style=solid];
-"54 transpose_1" -> "55 reshape_2"  [label="(49, 64, 3, 32)", style=solid];
-"55 reshape_2" -> "56 reshape_2_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
-"56 reshape_2_0_0_nncf_smooth_quant_0" -> "57 quantize_per_tensor_default_4"  [label="(49, 64, 96)", style=solid];
-"57 quantize_per_tensor_default_4" -> "58 dequantize_per_tensor_default_4"  [label="(49, 64, 96)", style=solid];
-"58 dequantize_per_tensor_default_4" -> "65 linear_3"  [label="(49, 64, 96)", style=solid];
-"59 _param_constant11_0_0" -> "65 linear_3"  [label="(96,)", style=solid];
-"60 scale_updated_constant2" -> "62 mul_tensor_2"  [label="(96, 1)", style=solid];
-"61 compressed_weight_updated_constant2" -> "62 mul_tensor_2"  [label="(96, 96)", style=solid];
-"62 mul_tensor_2" -> "64 sub_tensor_2"  [label="(96, 96)", style=solid];
-"63 zero_point_updated_constant2" -> "64 sub_tensor_2"  [label="(96, 1)", style=solid];
-"64 sub_tensor_2" -> "65 linear_3"  [label="(96, 96)", style=solid];
-"65 linear_3" -> "66 dropout_1"  [label="(49, 64, 96)", style=solid];
-"66 dropout_1" -> "67 view_3"  [label="(49, 64, 96)", style=solid];
-"67 view_3" -> "68 permute_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"68 permute_4" -> "69 reshape_3"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"69 reshape_3" -> "70 slice_2"  [label="(1, 56, 56, 96)", style=solid];
-"70 slice_2" -> "71 slice_3"  [label="(1, 56, 56, 96)", style=solid];
-"71 slice_3" -> "74 layer_norm_1"  [label="(1, 56, 56, 96)", style=solid];
-"72 _param_constant12" -> "74 layer_norm_1"  [label="(96,)", style=solid];
-"73 _param_constant13" -> "74 layer_norm_1"  [label="(96,)", style=solid];
-"74 layer_norm_1" -> "75 add_1"  [label="(1, 56, 56, 96)", style=solid];
-"75 add_1" -> "76 add_1_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 96)", style=solid];
-"75 add_1" -> "102 add_2"  [label="(1, 56, 56, 96)", style=solid];
-"76 add_1_0_0_nncf_smooth_quant_0" -> "77 quantize_per_tensor_default_5"  [label="(1, 56, 56, 96)", style=solid];
-"77 quantize_per_tensor_default_5" -> "78 dequantize_per_tensor_default_5"  [label="(1, 56, 56, 96)", style=solid];
-"78 dequantize_per_tensor_default_5" -> "85 linear_4"  [label="(1, 56, 56, 96)", style=solid];
-"79 _param_constant15_0_0" -> "85 linear_4"  [label="(384,)", style=solid];
-"80 scale_updated_constant3" -> "82 mul_tensor_3"  [label="(384, 1)", style=solid];
-"81 compressed_weight_updated_constant3" -> "82 mul_tensor_3"  [label="(384, 96)", style=solid];
-"82 mul_tensor_3" -> "84 sub_tensor_3"  [label="(384, 96)", style=solid];
-"83 zero_point_updated_constant3" -> "84 sub_tensor_3"  [label="(384, 1)", style=solid];
-"84 sub_tensor_3" -> "85 linear_4"  [label="(384, 96)", style=solid];
-"85 linear_4" -> "86 gelu"  [label="(1, 56, 56, 384)", style=solid];
-"86 gelu" -> "87 dropout_2"  [label="(1, 56, 56, 384)", style=solid];
-"87 dropout_2" -> "88 dropout_2_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 384)", style=solid];
-"88 dropout_2_0_0_nncf_smooth_quant_0" -> "89 quantize_per_tensor_default_6"  [label="(1, 56, 56, 384)", style=solid];
-"89 quantize_per_tensor_default_6" -> "90 dequantize_per_tensor_default_6"  [label="(1, 56, 56, 384)", style=solid];
-"90 dequantize_per_tensor_default_6" -> "97 linear_5"  [label="(1, 56, 56, 384)", style=solid];
-"91 _param_constant17_0_0" -> "97 linear_5"  [label="(96,)", style=solid];
-"92 scale_updated_constant4" -> "94 mul_tensor_4"  [label="(96, 1)", style=solid];
-"93 compressed_weight_updated_constant4" -> "94 mul_tensor_4"  [label="(96, 384)", style=solid];
-"94 mul_tensor_4" -> "96 sub_tensor_4"  [label="(96, 384)", style=solid];
-"95 zero_point_updated_constant4" -> "96 sub_tensor_4"  [label="(96, 1)", style=solid];
-"96 sub_tensor_4" -> "97 linear_5"  [label="(96, 384)", style=solid];
-"97 linear_5" -> "98 dropout_3"  [label="(1, 56, 56, 96)", style=solid];
-"98 dropout_3" -> "101 layer_norm_2"  [label="(1, 56, 56, 96)", style=solid];
-"99 _param_constant18" -> "101 layer_norm_2"  [label="(96,)", style=solid];
-"100 _param_constant19" -> "101 layer_norm_2"  [label="(96,)", style=solid];
-"101 layer_norm_2" -> "102 add_2"  [label="(1, 56, 56, 96)", style=solid];
-"102 add_2" -> "104 pad_1"  [label="(1, 56, 56, 96)", style=solid];
-"102 add_2" -> "182 add_5"  [label="(1, 56, 56, 96)", style=solid];
-"103 _frozen_param4" -> "140 add_3"  [label="(1, 3, 64, 64)", style=solid];
-"104 pad_1" -> "105 roll"  [label="(1, 56, 56, 96)", style=solid];
-"105 roll" -> "106 view_6"  [label="(1, 56, 56, 96)", style=solid];
-"106 view_6" -> "107 permute_6"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"107 permute_6" -> "108 reshape_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"108 reshape_4" -> "109 reshape_4_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
-"108 reshape_4" -> "141 new_zeros"  [label="(49, 64, 96)", style=solid];
-"109 reshape_4_0_0_nncf_smooth_quant_0" -> "110 quantize_per_tensor_default_7"  [label="(49, 64, 96)", style=solid];
-"110 quantize_per_tensor_default_7" -> "111 dequantize_per_tensor_default_7"  [label="(49, 64, 96)", style=solid];
-"111 dequantize_per_tensor_default_7" -> "118 linear_8"  [label="(49, 64, 96)", style=solid];
-"112 _frozen_param5_0_0" -> "118 linear_8"  [label="(288,)", style=solid];
-"113 scale_updated_constant5" -> "115 mul_tensor_5"  [label="(288, 1)", style=solid];
-"114 compressed_weight_updated_constant5" -> "115 mul_tensor_5"  [label="(288, 96)", style=solid];
-"115 mul_tensor_5" -> "117 sub_tensor_5"  [label="(288, 96)", style=solid];
-"116 zero_point_updated_constant5" -> "117 sub_tensor_5"  [label="(288, 1)", style=solid];
-"117 sub_tensor_5" -> "118 linear_8"  [label="(288, 96)", style=solid];
-"118 linear_8" -> "119 reshape_5"  [label="(49, 64, 288)", style=solid];
-"119 reshape_5" -> "120 permute_7"  [label="(49, 64, 3, 3, 32)", style=solid];
-"120 permute_7" -> "121 select_3"  [label="(3, 49, 3, 64, 32)", style=solid];
-"120 permute_7" -> "122 select_4"  [label="(3, 49, 3, 64, 32)", style=solid];
-"120 permute_7" -> "123 select_5"  [label="(3, 49, 3, 64, 32)", style=solid];
-"121 select_3" -> "124 linalg_vector_norm_2"  [label="(49, 3, 64, 32)", style=solid];
-"121 select_3" -> "126 expand_as_2"  [label="(49, 3, 64, 32)", style=solid];
-"121 select_3" -> "127 div_2"  [label="(49, 3, 64, 32)", style=solid];
-"122 select_4" -> "130 linalg_vector_norm_3"  [label="(49, 3, 64, 32)", style=solid];
-"122 select_4" -> "132 expand_as_3"  [label="(49, 3, 64, 32)", style=solid];
-"122 select_4" -> "133 div_3"  [label="(49, 3, 64, 32)", style=solid];
-"123 select_5" -> "159 matmul_3"  [label="(49, 3, 64, 32)", style=solid];
-"124 linalg_vector_norm_2" -> "125 clamp_min_2"  [label="(49, 3, 64, 1)", style=solid];
-"125 clamp_min_2" -> "126 expand_as_2"  [label="(49, 3, 64, 1)", style=solid];
-"126 expand_as_2" -> "127 div_2"  [label="(49, 3, 64, 32)", style=solid];
-"127 div_2" -> "128 quantize_per_tensor_default_8"  [label="(49, 3, 64, 32)", style=solid];
-"128 quantize_per_tensor_default_8" -> "129 dequantize_per_tensor_default_8"  [label="(49, 3, 64, 32)", style=solid];
-"129 dequantize_per_tensor_default_8" -> "137 matmul_2"  [label="(49, 3, 64, 32)", style=solid];
-"130 linalg_vector_norm_3" -> "131 clamp_min_3"  [label="(49, 3, 64, 1)", style=solid];
-"131 clamp_min_3" -> "132 expand_as_3"  [label="(49, 3, 64, 1)", style=solid];
-"132 expand_as_3" -> "133 div_3"  [label="(49, 3, 64, 32)", style=solid];
-"133 div_3" -> "134 quantize_per_tensor_default_9"  [label="(49, 3, 64, 32)", style=solid];
-"134 quantize_per_tensor_default_9" -> "135 dequantize_per_tensor_default_9"  [label="(49, 3, 64, 32)", style=solid];
-"135 dequantize_per_tensor_default_9" -> "136 transpose_2"  [label="(49, 3, 64, 32)", style=solid];
-"136 transpose_2" -> "137 matmul_2"  [label="(49, 3, 32, 64)", style=solid];
-"137 matmul_2" -> "139 mul_3"  [label="(49, 3, 64, 64)", style=solid];
-"138 _frozen_param7" -> "139 mul_3"  [label="(3, 1, 1)", style=solid];
-"139 mul_3" -> "140 add_3"  [label="(49, 3, 64, 64)", style=solid];
-"140 add_3" -> "152 view_8"  [label="(49, 3, 64, 64)", style=solid];
-"141 new_zeros" -> "142 view_7"  [label="(56, 56)", style=solid];
-"142 view_7" -> "143 permute_8"  [label="(7, 8, 7, 8)", style=solid];
-"143 permute_8" -> "144 reshape_6"  [label="(7, 7, 8, 8)", style=solid];
-"144 reshape_6" -> "145 unsqueeze_2"  [label="(49, 64)", style=solid];
-"144 reshape_6" -> "146 unsqueeze_3"  [label="(49, 64)", style=solid];
-"145 unsqueeze_2" -> "147 sub"  [label="(49, 1, 64)", style=solid];
-"146 unsqueeze_3" -> "147 sub"  [label="(49, 64, 1)", style=solid];
-"147 sub" -> "148 ne"  [label="(49, 64, 64)", style=solid];
-"147 sub" -> "149 masked_fill"  [label="(49, 64, 64)", style=solid];
-"147 sub" -> "150 eq"  [label="(49, 64, 64)", style=solid];
-"148 ne" -> "149 masked_fill"  [label="(49, 64, 64)", style=solid];
-"149 masked_fill" -> "151 masked_fill_1"  [label="(49, 64, 64)", style=solid];
-"150 eq" -> "151 masked_fill_1"  [label="(49, 64, 64)", style=solid];
-"151 masked_fill_1" -> "153 unsqueeze_4"  [label="(49, 64, 64)", style=solid];
-"152 view_8" -> "155 add_4"  [label="(1, 49, 3, 64, 64)", style=solid];
-"153 unsqueeze_4" -> "154 unsqueeze_5"  [label="(49, 1, 64, 64)", style=solid];
-"154 unsqueeze_5" -> "155 add_4"  [label="(1, 49, 1, 64, 64)", style=solid];
-"155 add_4" -> "156 view_9"  [label="(1, 49, 3, 64, 64)", style=solid];
-"156 view_9" -> "157 softmax_1"  [label="(49, 3, 64, 64)", style=solid];
-"157 softmax_1" -> "158 dropout_4"  [label="(49, 3, 64, 64)", style=solid];
-"158 dropout_4" -> "159 matmul_3"  [label="(49, 3, 64, 64)", style=solid];
-"159 matmul_3" -> "160 transpose_3"  [label="(49, 3, 64, 32)", style=solid];
-"160 transpose_3" -> "161 reshape_7"  [label="(49, 64, 3, 32)", style=solid];
-"161 reshape_7" -> "162 reshape_7_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
-"162 reshape_7_0_0_nncf_smooth_quant_0" -> "163 quantize_per_tensor_default_10"  [label="(49, 64, 96)", style=solid];
-"163 quantize_per_tensor_default_10" -> "164 dequantize_per_tensor_default_10"  [label="(49, 64, 96)", style=solid];
-"164 dequantize_per_tensor_default_10" -> "171 linear_9"  [label="(49, 64, 96)", style=solid];
-"165 _param_constant27_0_0" -> "171 linear_9"  [label="(96,)", style=solid];
-"166 scale_updated_constant6" -> "168 mul_tensor_6"  [label="(96, 1)", style=solid];
-"167 compressed_weight_updated_constant6" -> "168 mul_tensor_6"  [label="(96, 96)", style=solid];
-"168 mul_tensor_6" -> "170 sub_tensor_6"  [label="(96, 96)", style=solid];
-"169 zero_point_updated_constant6" -> "170 sub_tensor_6"  [label="(96, 1)", style=solid];
-"170 sub_tensor_6" -> "171 linear_9"  [label="(96, 96)", style=solid];
-"171 linear_9" -> "172 dropout_5"  [label="(49, 64, 96)", style=solid];
-"172 dropout_5" -> "173 view_10"  [label="(49, 64, 96)", style=solid];
-"173 view_10" -> "174 permute_9"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"174 permute_9" -> "175 reshape_8"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"175 reshape_8" -> "176 roll_1"  [label="(1, 56, 56, 96)", style=solid];
-"176 roll_1" -> "177 slice_23"  [label="(1, 56, 56, 96)", style=solid];
-"177 slice_23" -> "178 slice_24"  [label="(1, 56, 56, 96)", style=solid];
-"178 slice_24" -> "181 layer_norm_3"  [label="(1, 56, 56, 96)", style=solid];
-"179 _param_constant28" -> "181 layer_norm_3"  [label="(96,)", style=solid];
-"180 _param_constant29" -> "181 layer_norm_3"  [label="(96,)", style=solid];
-"181 layer_norm_3" -> "182 add_5"  [label="(1, 56, 56, 96)", style=solid];
-"182 add_5" -> "183 add_5_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 96)", style=solid];
-"182 add_5" -> "209 add_6"  [label="(1, 56, 56, 96)", style=solid];
-"183 add_5_0_0_nncf_smooth_quant_0" -> "184 quantize_per_tensor_default_11"  [label="(1, 56, 56, 96)", style=solid];
-"184 quantize_per_tensor_default_11" -> "185 dequantize_per_tensor_default_11"  [label="(1, 56, 56, 96)", style=solid];
-"185 dequantize_per_tensor_default_11" -> "192 linear_10"  [label="(1, 56, 56, 96)", style=solid];
-"186 _param_constant31_0_0" -> "192 linear_10"  [label="(384,)", style=solid];
-"187 scale_updated_constant7" -> "189 mul_tensor_7"  [label="(384, 1)", style=solid];
-"188 compressed_weight_updated_constant7" -> "189 mul_tensor_7"  [label="(384, 96)", style=solid];
-"189 mul_tensor_7" -> "191 sub_tensor_7"  [label="(384, 96)", style=solid];
-"190 zero_point_updated_constant7" -> "191 sub_tensor_7"  [label="(384, 1)", style=solid];
-"191 sub_tensor_7" -> "192 linear_10"  [label="(384, 96)", style=solid];
-"192 linear_10" -> "193 gelu_1"  [label="(1, 56, 56, 384)", style=solid];
-"193 gelu_1" -> "194 dropout_6"  [label="(1, 56, 56, 384)", style=solid];
-"194 dropout_6" -> "195 dropout_6_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 384)", style=solid];
-"195 dropout_6_0_0_nncf_smooth_quant_0" -> "196 quantize_per_tensor_default_12"  [label="(1, 56, 56, 384)", style=solid];
-"196 quantize_per_tensor_default_12" -> "197 dequantize_per_tensor_default_12"  [label="(1, 56, 56, 384)", style=solid];
-"197 dequantize_per_tensor_default_12" -> "204 linear_11"  [label="(1, 56, 56, 384)", style=solid];
-"198 _param_constant33_0_0" -> "204 linear_11"  [label="(96,)", style=solid];
-"199 scale_updated_constant8" -> "201 mul_tensor_8"  [label="(96, 1)", style=solid];
-"200 compressed_weight_updated_constant8" -> "201 mul_tensor_8"  [label="(96, 384)", style=solid];
-"201 mul_tensor_8" -> "203 sub_tensor_8"  [label="(96, 384)", style=solid];
-"202 zero_point_updated_constant8" -> "203 sub_tensor_8"  [label="(96, 1)", style=solid];
-"203 sub_tensor_8" -> "204 linear_11"  [label="(96, 384)", style=solid];
-"204 linear_11" -> "205 dropout_7"  [label="(1, 56, 56, 96)", style=solid];
-"205 dropout_7" -> "208 layer_norm_4"  [label="(1, 56, 56, 96)", style=solid];
-"206 _param_constant34" -> "208 layer_norm_4"  [label="(96,)", style=solid];
-"207 _param_constant35" -> "208 layer_norm_4"  [label="(96,)", style=solid];
-"208 layer_norm_4" -> "209 add_6"  [label="(1, 56, 56, 96)", style=solid];
-"209 add_6" -> "210 pad_2"  [label="(1, 56, 56, 96)", style=solid];
-"210 pad_2" -> "211 slice_25"  [label="(1, 56, 56, 96)", style=solid];
-"210 pad_2" -> "214 slice_28"  [label="(1, 56, 56, 96)", style=solid];
-"210 pad_2" -> "217 slice_31"  [label="(1, 56, 56, 96)", style=solid];
-"210 pad_2" -> "220 slice_34"  [label="(1, 56, 56, 96)", style=solid];
-"211 slice_25" -> "212 slice_26"  [label="(1, 28, 56, 96)", style=solid];
-"212 slice_26" -> "213 slice_27"  [label="(1, 28, 28, 96)", style=solid];
-"213 slice_27" -> "223 cat"  [label="(1, 28, 28, 96)", style=solid];
-"214 slice_28" -> "215 slice_29"  [label="(1, 28, 56, 96)", style=solid];
-"215 slice_29" -> "216 slice_30"  [label="(1, 28, 28, 96)", style=solid];
-"216 slice_30" -> "223 cat"  [label="(1, 28, 28, 96)", style=solid];
-"217 slice_31" -> "218 slice_32"  [label="(1, 28, 56, 96)", style=solid];
-"218 slice_32" -> "219 slice_33"  [label="(1, 28, 28, 96)", style=solid];
-"219 slice_33" -> "223 cat"  [label="(1, 28, 28, 96)", style=solid];
-"220 slice_34" -> "221 slice_35"  [label="(1, 28, 56, 96)", style=solid];
-"221 slice_35" -> "222 slice_36"  [label="(1, 28, 28, 96)", style=solid];
-"222 slice_36" -> "223 cat"  [label="(1, 28, 28, 96)", style=solid];
-"223 cat" -> "224 cat_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 384)", style=solid];
-"224 cat_0_0_nncf_smooth_quant_0" -> "225 quantize_per_tensor_default_13"  [label="(1, 28, 28, 384)", style=solid];
-"225 quantize_per_tensor_default_13" -> "226 dequantize_per_tensor_default_13"  [label="(1, 28, 28, 384)", style=solid];
-"226 dequantize_per_tensor_default_13" -> "232 linear_12"  [label="(1, 28, 28, 384)", style=solid];
-"227 scale_updated_constant9" -> "229 mul_tensor_9"  [label="(192, 1)", style=solid];
-"228 compressed_weight_updated_constant9" -> "229 mul_tensor_9"  [label="(192, 384)", style=solid];
-"229 mul_tensor_9" -> "231 sub_tensor_9"  [label="(192, 384)", style=solid];
-"230 zero_point_updated_constant9" -> "231 sub_tensor_9"  [label="(192, 1)", style=solid];
-"231 sub_tensor_9" -> "232 linear_12"  [label="(192, 384)", style=solid];
-"232 linear_12" -> "235 layer_norm_5"  [label="(1, 28, 28, 192)", style=solid];
-"233 _param_constant37" -> "235 layer_norm_5"  [label="(192,)", style=solid];
-"234 _param_constant38" -> "235 layer_norm_5"  [label="(192,)", style=solid];
-"235 layer_norm_5" -> "237 pad_3"  [label="(1, 28, 28, 192)", style=solid];
-"235 layer_norm_5" -> "300 add_8"  [label="(1, 28, 28, 192)", style=solid];
-"236 _frozen_param17" -> "272 add_7"  [label="(1, 6, 64, 64)", style=solid];
-"237 pad_3" -> "238 view_13"  [label="(1, 32, 32, 192)", style=solid];
-"238 view_13" -> "239 permute_11"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"239 permute_11" -> "240 reshape_9"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"240 reshape_9" -> "241 reshape_9_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
-"241 reshape_9_0_0_nncf_smooth_quant_0" -> "242 quantize_per_tensor_default_14"  [label="(16, 64, 192)", style=solid];
-"242 quantize_per_tensor_default_14" -> "243 dequantize_per_tensor_default_14"  [label="(16, 64, 192)", style=solid];
-"243 dequantize_per_tensor_default_14" -> "250 linear_15"  [label="(16, 64, 192)", style=solid];
-"244 _frozen_param18_0_0" -> "250 linear_15"  [label="(576,)", style=solid];
-"245 scale_updated_constant10" -> "247 mul_tensor_10"  [label="(576, 1)", style=solid];
-"246 compressed_weight_updated_constant10" -> "247 mul_tensor_10"  [label="(576, 192)", style=solid];
-"247 mul_tensor_10" -> "249 sub_tensor_10"  [label="(576, 192)", style=solid];
-"248 zero_point_updated_constant10" -> "249 sub_tensor_10"  [label="(576, 1)", style=solid];
-"249 sub_tensor_10" -> "250 linear_15"  [label="(576, 192)", style=solid];
-"250 linear_15" -> "251 reshape_10"  [label="(16, 64, 576)", style=solid];
-"251 reshape_10" -> "252 permute_12"  [label="(16, 64, 3, 6, 32)", style=solid];
-"252 permute_12" -> "253 select_6"  [label="(3, 16, 6, 64, 32)", style=solid];
-"252 permute_12" -> "254 select_7"  [label="(3, 16, 6, 64, 32)", style=solid];
-"252 permute_12" -> "255 select_8"  [label="(3, 16, 6, 64, 32)", style=solid];
-"253 select_6" -> "256 linalg_vector_norm_4"  [label="(16, 6, 64, 32)", style=solid];
-"253 select_6" -> "258 expand_as_4"  [label="(16, 6, 64, 32)", style=solid];
-"253 select_6" -> "259 div_4"  [label="(16, 6, 64, 32)", style=solid];
-"254 select_7" -> "262 linalg_vector_norm_5"  [label="(16, 6, 64, 32)", style=solid];
-"254 select_7" -> "264 expand_as_5"  [label="(16, 6, 64, 32)", style=solid];
-"254 select_7" -> "265 div_5"  [label="(16, 6, 64, 32)", style=solid];
-"255 select_8" -> "275 matmul_5"  [label="(16, 6, 64, 32)", style=solid];
-"256 linalg_vector_norm_4" -> "257 clamp_min_4"  [label="(16, 6, 64, 1)", style=solid];
-"257 clamp_min_4" -> "258 expand_as_4"  [label="(16, 6, 64, 1)", style=solid];
-"258 expand_as_4" -> "259 div_4"  [label="(16, 6, 64, 32)", style=solid];
-"259 div_4" -> "260 quantize_per_tensor_default_15"  [label="(16, 6, 64, 32)", style=solid];
-"260 quantize_per_tensor_default_15" -> "261 dequantize_per_tensor_default_15"  [label="(16, 6, 64, 32)", style=solid];
-"261 dequantize_per_tensor_default_15" -> "269 matmul_4"  [label="(16, 6, 64, 32)", style=solid];
-"262 linalg_vector_norm_5" -> "263 clamp_min_5"  [label="(16, 6, 64, 1)", style=solid];
-"263 clamp_min_5" -> "264 expand_as_5"  [label="(16, 6, 64, 1)", style=solid];
-"264 expand_as_5" -> "265 div_5"  [label="(16, 6, 64, 32)", style=solid];
-"265 div_5" -> "266 quantize_per_tensor_default_16"  [label="(16, 6, 64, 32)", style=solid];
-"266 quantize_per_tensor_default_16" -> "267 dequantize_per_tensor_default_16"  [label="(16, 6, 64, 32)", style=solid];
-"267 dequantize_per_tensor_default_16" -> "268 transpose_4"  [label="(16, 6, 64, 32)", style=solid];
-"268 transpose_4" -> "269 matmul_4"  [label="(16, 6, 32, 64)", style=solid];
-"269 matmul_4" -> "271 mul_5"  [label="(16, 6, 64, 64)", style=solid];
-"270 _frozen_param20" -> "271 mul_5"  [label="(6, 1, 1)", style=solid];
-"271 mul_5" -> "272 add_7"  [label="(16, 6, 64, 64)", style=solid];
-"272 add_7" -> "273 softmax_2"  [label="(16, 6, 64, 64)", style=solid];
-"273 softmax_2" -> "274 dropout_8"  [label="(16, 6, 64, 64)", style=solid];
-"274 dropout_8" -> "275 matmul_5"  [label="(16, 6, 64, 64)", style=solid];
-"275 matmul_5" -> "276 transpose_5"  [label="(16, 6, 64, 32)", style=solid];
-"276 transpose_5" -> "277 reshape_11"  [label="(16, 64, 6, 32)", style=solid];
-"277 reshape_11" -> "278 reshape_11_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
-"278 reshape_11_0_0_nncf_smooth_quant_0" -> "279 quantize_per_tensor_default_17"  [label="(16, 64, 192)", style=solid];
-"279 quantize_per_tensor_default_17" -> "280 dequantize_per_tensor_default_17"  [label="(16, 64, 192)", style=solid];
-"280 dequantize_per_tensor_default_17" -> "287 linear_16"  [label="(16, 64, 192)", style=solid];
-"281 _param_constant46_0_0" -> "287 linear_16"  [label="(192,)", style=solid];
-"282 scale_updated_constant11" -> "284 mul_tensor_11"  [label="(192, 1)", style=solid];
-"283 compressed_weight_updated_constant11" -> "284 mul_tensor_11"  [label="(192, 192)", style=solid];
-"284 mul_tensor_11" -> "286 sub_tensor_11"  [label="(192, 192)", style=solid];
-"285 zero_point_updated_constant11" -> "286 sub_tensor_11"  [label="(192, 1)", style=solid];
-"286 sub_tensor_11" -> "287 linear_16"  [label="(192, 192)", style=solid];
-"287 linear_16" -> "288 dropout_9"  [label="(16, 64, 192)", style=solid];
-"288 dropout_9" -> "289 view_14"  [label="(16, 64, 192)", style=solid];
-"289 view_14" -> "290 permute_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"290 permute_13" -> "291 reshape_12"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"291 reshape_12" -> "292 slice_38"  [label="(1, 32, 32, 192)", style=solid];
-"292 slice_38" -> "293 slice_39"  [label="(1, 32, 32, 192)", style=solid];
-"293 slice_39" -> "294 slice_40"  [label="(1, 28, 32, 192)", style=solid];
-"294 slice_40" -> "295 slice_41"  [label="(1, 28, 28, 192)", style=solid];
-"295 slice_41" -> "296 contiguous_3"  [label="(1, 28, 28, 192)", style=solid];
-"296 contiguous_3" -> "299 layer_norm_6"  [label="(1, 28, 28, 192)", style=solid];
-"297 _param_constant47" -> "299 layer_norm_6"  [label="(192,)", style=solid];
-"298 _param_constant48" -> "299 layer_norm_6"  [label="(192,)", style=solid];
-"299 layer_norm_6" -> "300 add_8"  [label="(1, 28, 28, 192)", style=solid];
-"300 add_8" -> "301 add_8_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 192)", style=solid];
-"300 add_8" -> "327 add_9"  [label="(1, 28, 28, 192)", style=solid];
-"301 add_8_0_0_nncf_smooth_quant_0" -> "302 quantize_per_tensor_default_18"  [label="(1, 28, 28, 192)", style=solid];
-"302 quantize_per_tensor_default_18" -> "303 dequantize_per_tensor_default_18"  [label="(1, 28, 28, 192)", style=solid];
-"303 dequantize_per_tensor_default_18" -> "310 linear_17"  [label="(1, 28, 28, 192)", style=solid];
-"304 _param_constant50_0_0" -> "310 linear_17"  [label="(768,)", style=solid];
-"305 scale_updated_constant12" -> "307 mul_tensor_12"  [label="(768, 1)", style=solid];
-"306 compressed_weight_updated_constant12" -> "307 mul_tensor_12"  [label="(768, 192)", style=solid];
-"307 mul_tensor_12" -> "309 sub_tensor_12"  [label="(768, 192)", style=solid];
-"308 zero_point_updated_constant12" -> "309 sub_tensor_12"  [label="(768, 1)", style=solid];
-"309 sub_tensor_12" -> "310 linear_17"  [label="(768, 192)", style=solid];
-"310 linear_17" -> "311 gelu_2"  [label="(1, 28, 28, 768)", style=solid];
-"311 gelu_2" -> "312 dropout_10"  [label="(1, 28, 28, 768)", style=solid];
-"312 dropout_10" -> "313 dropout_10_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 768)", style=solid];
-"313 dropout_10_0_0_nncf_smooth_quant_0" -> "314 quantize_per_tensor_default_19"  [label="(1, 28, 28, 768)", style=solid];
-"314 quantize_per_tensor_default_19" -> "315 dequantize_per_tensor_default_19"  [label="(1, 28, 28, 768)", style=solid];
-"315 dequantize_per_tensor_default_19" -> "322 linear_18"  [label="(1, 28, 28, 768)", style=solid];
-"316 _param_constant52_0_0" -> "322 linear_18"  [label="(192,)", style=solid];
-"317 scale_updated_constant13" -> "319 mul_tensor_13"  [label="(192, 1)", style=solid];
-"318 compressed_weight_updated_constant13" -> "319 mul_tensor_13"  [label="(192, 768)", style=solid];
-"319 mul_tensor_13" -> "321 sub_tensor_13"  [label="(192, 768)", style=solid];
-"320 zero_point_updated_constant13" -> "321 sub_tensor_13"  [label="(192, 1)", style=solid];
-"321 sub_tensor_13" -> "322 linear_18"  [label="(192, 768)", style=solid];
-"322 linear_18" -> "323 dropout_11"  [label="(1, 28, 28, 192)", style=solid];
-"323 dropout_11" -> "326 layer_norm_7"  [label="(1, 28, 28, 192)", style=solid];
-"324 _param_constant53" -> "326 layer_norm_7"  [label="(192,)", style=solid];
-"325 _param_constant54" -> "326 layer_norm_7"  [label="(192,)", style=solid];
-"326 layer_norm_7" -> "327 add_9"  [label="(1, 28, 28, 192)", style=solid];
-"327 add_9" -> "329 pad_4"  [label="(1, 28, 28, 192)", style=solid];
-"327 add_9" -> "410 add_12"  [label="(1, 28, 28, 192)", style=solid];
-"328 _frozen_param21" -> "365 add_10"  [label="(1, 6, 64, 64)", style=solid];
-"329 pad_4" -> "330 roll_2"  [label="(1, 32, 32, 192)", style=solid];
-"330 roll_2" -> "331 view_17"  [label="(1, 32, 32, 192)", style=solid];
-"331 view_17" -> "332 permute_15"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"332 permute_15" -> "333 reshape_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"333 reshape_13" -> "334 reshape_13_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
-"333 reshape_13" -> "366 new_zeros_1"  [label="(16, 64, 192)", style=solid];
-"334 reshape_13_0_0_nncf_smooth_quant_0" -> "335 quantize_per_tensor_default_20"  [label="(16, 64, 192)", style=solid];
-"335 quantize_per_tensor_default_20" -> "336 dequantize_per_tensor_default_20"  [label="(16, 64, 192)", style=solid];
-"336 dequantize_per_tensor_default_20" -> "343 linear_21"  [label="(16, 64, 192)", style=solid];
-"337 _frozen_param22_0_0" -> "343 linear_21"  [label="(576,)", style=solid];
-"338 scale_updated_constant14" -> "340 mul_tensor_14"  [label="(576, 1)", style=solid];
-"339 compressed_weight_updated_constant14" -> "340 mul_tensor_14"  [label="(576, 192)", style=solid];
-"340 mul_tensor_14" -> "342 sub_tensor_14"  [label="(576, 192)", style=solid];
-"341 zero_point_updated_constant14" -> "342 sub_tensor_14"  [label="(576, 1)", style=solid];
-"342 sub_tensor_14" -> "343 linear_21"  [label="(576, 192)", style=solid];
-"343 linear_21" -> "344 reshape_14"  [label="(16, 64, 576)", style=solid];
-"344 reshape_14" -> "345 permute_16"  [label="(16, 64, 3, 6, 32)", style=solid];
-"345 permute_16" -> "346 select_9"  [label="(3, 16, 6, 64, 32)", style=solid];
-"345 permute_16" -> "347 select_10"  [label="(3, 16, 6, 64, 32)", style=solid];
-"345 permute_16" -> "348 select_11"  [label="(3, 16, 6, 64, 32)", style=solid];
-"346 select_9" -> "349 linalg_vector_norm_6"  [label="(16, 6, 64, 32)", style=solid];
-"346 select_9" -> "351 expand_as_6"  [label="(16, 6, 64, 32)", style=solid];
-"346 select_9" -> "352 div_6"  [label="(16, 6, 64, 32)", style=solid];
-"347 select_10" -> "355 linalg_vector_norm_7"  [label="(16, 6, 64, 32)", style=solid];
-"347 select_10" -> "357 expand_as_7"  [label="(16, 6, 64, 32)", style=solid];
-"347 select_10" -> "358 div_7"  [label="(16, 6, 64, 32)", style=solid];
-"348 select_11" -> "384 matmul_7"  [label="(16, 6, 64, 32)", style=solid];
-"349 linalg_vector_norm_6" -> "350 clamp_min_6"  [label="(16, 6, 64, 1)", style=solid];
-"350 clamp_min_6" -> "351 expand_as_6"  [label="(16, 6, 64, 1)", style=solid];
-"351 expand_as_6" -> "352 div_6"  [label="(16, 6, 64, 32)", style=solid];
-"352 div_6" -> "353 quantize_per_tensor_default_21"  [label="(16, 6, 64, 32)", style=solid];
-"353 quantize_per_tensor_default_21" -> "354 dequantize_per_tensor_default_21"  [label="(16, 6, 64, 32)", style=solid];
-"354 dequantize_per_tensor_default_21" -> "362 matmul_6"  [label="(16, 6, 64, 32)", style=solid];
-"355 linalg_vector_norm_7" -> "356 clamp_min_7"  [label="(16, 6, 64, 1)", style=solid];
-"356 clamp_min_7" -> "357 expand_as_7"  [label="(16, 6, 64, 1)", style=solid];
-"357 expand_as_7" -> "358 div_7"  [label="(16, 6, 64, 32)", style=solid];
-"358 div_7" -> "359 quantize_per_tensor_default_22"  [label="(16, 6, 64, 32)", style=solid];
-"359 quantize_per_tensor_default_22" -> "360 dequantize_per_tensor_default_22"  [label="(16, 6, 64, 32)", style=solid];
-"360 dequantize_per_tensor_default_22" -> "361 transpose_6"  [label="(16, 6, 64, 32)", style=solid];
-"361 transpose_6" -> "362 matmul_6"  [label="(16, 6, 32, 64)", style=solid];
-"362 matmul_6" -> "364 mul_7"  [label="(16, 6, 64, 64)", style=solid];
-"363 _frozen_param24" -> "364 mul_7"  [label="(6, 1, 1)", style=solid];
-"364 mul_7" -> "365 add_10"  [label="(16, 6, 64, 64)", style=solid];
-"365 add_10" -> "377 view_19"  [label="(16, 6, 64, 64)", style=solid];
-"366 new_zeros_1" -> "367 view_18"  [label="(32, 32)", style=solid];
-"367 view_18" -> "368 permute_17"  [label="(4, 8, 4, 8)", style=solid];
-"368 permute_17" -> "369 reshape_15"  [label="(4, 4, 8, 8)", style=solid];
-"369 reshape_15" -> "370 unsqueeze_8"  [label="(16, 64)", style=solid];
-"369 reshape_15" -> "371 unsqueeze_9"  [label="(16, 64)", style=solid];
-"370 unsqueeze_8" -> "372 sub_1"  [label="(16, 1, 64)", style=solid];
-"371 unsqueeze_9" -> "372 sub_1"  [label="(16, 64, 1)", style=solid];
-"372 sub_1" -> "373 ne_1"  [label="(16, 64, 64)", style=solid];
-"372 sub_1" -> "374 masked_fill_2"  [label="(16, 64, 64)", style=solid];
-"372 sub_1" -> "375 eq_1"  [label="(16, 64, 64)", style=solid];
-"373 ne_1" -> "374 masked_fill_2"  [label="(16, 64, 64)", style=solid];
-"374 masked_fill_2" -> "376 masked_fill_3"  [label="(16, 64, 64)", style=solid];
-"375 eq_1" -> "376 masked_fill_3"  [label="(16, 64, 64)", style=solid];
-"376 masked_fill_3" -> "378 unsqueeze_10"  [label="(16, 64, 64)", style=solid];
-"377 view_19" -> "380 add_11"  [label="(1, 16, 6, 64, 64)", style=solid];
-"378 unsqueeze_10" -> "379 unsqueeze_11"  [label="(16, 1, 64, 64)", style=solid];
-"379 unsqueeze_11" -> "380 add_11"  [label="(1, 16, 1, 64, 64)", style=solid];
-"380 add_11" -> "381 view_20"  [label="(1, 16, 6, 64, 64)", style=solid];
-"381 view_20" -> "382 softmax_3"  [label="(16, 6, 64, 64)", style=solid];
-"382 softmax_3" -> "383 dropout_12"  [label="(16, 6, 64, 64)", style=solid];
-"383 dropout_12" -> "384 matmul_7"  [label="(16, 6, 64, 64)", style=solid];
-"384 matmul_7" -> "385 transpose_7"  [label="(16, 6, 64, 32)", style=solid];
-"385 transpose_7" -> "386 reshape_16"  [label="(16, 64, 6, 32)", style=solid];
-"386 reshape_16" -> "387 reshape_16_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
-"387 reshape_16_0_0_nncf_smooth_quant_0" -> "388 quantize_per_tensor_default_23"  [label="(16, 64, 192)", style=solid];
-"388 quantize_per_tensor_default_23" -> "389 dequantize_per_tensor_default_23"  [label="(16, 64, 192)", style=solid];
-"389 dequantize_per_tensor_default_23" -> "396 linear_22"  [label="(16, 64, 192)", style=solid];
-"390 _param_constant62_0_0" -> "396 linear_22"  [label="(192,)", style=solid];
-"391 scale_updated_constant15" -> "393 mul_tensor_15"  [label="(192, 1)", style=solid];
-"392 compressed_weight_updated_constant15" -> "393 mul_tensor_15"  [label="(192, 192)", style=solid];
-"393 mul_tensor_15" -> "395 sub_tensor_15"  [label="(192, 192)", style=solid];
-"394 zero_point_updated_constant15" -> "395 sub_tensor_15"  [label="(192, 1)", style=solid];
-"395 sub_tensor_15" -> "396 linear_22"  [label="(192, 192)", style=solid];
-"396 linear_22" -> "397 dropout_13"  [label="(16, 64, 192)", style=solid];
-"397 dropout_13" -> "398 view_21"  [label="(16, 64, 192)", style=solid];
-"398 view_21" -> "399 permute_18"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"399 permute_18" -> "400 reshape_17"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"400 reshape_17" -> "401 roll_3"  [label="(1, 32, 32, 192)", style=solid];
-"401 roll_3" -> "402 slice_61"  [label="(1, 32, 32, 192)", style=solid];
-"402 slice_61" -> "403 slice_62"  [label="(1, 32, 32, 192)", style=solid];
-"403 slice_62" -> "404 slice_63"  [label="(1, 28, 32, 192)", style=solid];
-"404 slice_63" -> "405 slice_64"  [label="(1, 28, 28, 192)", style=solid];
-"405 slice_64" -> "406 contiguous_5"  [label="(1, 28, 28, 192)", style=solid];
-"406 contiguous_5" -> "409 layer_norm_8"  [label="(1, 28, 28, 192)", style=solid];
-"407 _param_constant63" -> "409 layer_norm_8"  [label="(192,)", style=solid];
-"408 _param_constant64" -> "409 layer_norm_8"  [label="(192,)", style=solid];
-"409 layer_norm_8" -> "410 add_12"  [label="(1, 28, 28, 192)", style=solid];
-"410 add_12" -> "411 add_12_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 192)", style=solid];
-"410 add_12" -> "437 add_13"  [label="(1, 28, 28, 192)", style=solid];
-"411 add_12_0_0_nncf_smooth_quant_0" -> "412 quantize_per_tensor_default_24"  [label="(1, 28, 28, 192)", style=solid];
-"412 quantize_per_tensor_default_24" -> "413 dequantize_per_tensor_default_24"  [label="(1, 28, 28, 192)", style=solid];
-"413 dequantize_per_tensor_default_24" -> "420 linear_23"  [label="(1, 28, 28, 192)", style=solid];
-"414 _param_constant66_0_0" -> "420 linear_23"  [label="(768,)", style=solid];
-"415 scale_updated_constant16" -> "417 mul_tensor_16"  [label="(768, 1)", style=solid];
-"416 compressed_weight_updated_constant16" -> "417 mul_tensor_16"  [label="(768, 192)", style=solid];
-"417 mul_tensor_16" -> "419 sub_tensor_16"  [label="(768, 192)", style=solid];
-"418 zero_point_updated_constant16" -> "419 sub_tensor_16"  [label="(768, 1)", style=solid];
-"419 sub_tensor_16" -> "420 linear_23"  [label="(768, 192)", style=solid];
-"420 linear_23" -> "421 gelu_3"  [label="(1, 28, 28, 768)", style=solid];
-"421 gelu_3" -> "422 dropout_14"  [label="(1, 28, 28, 768)", style=solid];
-"422 dropout_14" -> "423 dropout_14_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 768)", style=solid];
-"423 dropout_14_0_0_nncf_smooth_quant_0" -> "424 quantize_per_tensor_default_25"  [label="(1, 28, 28, 768)", style=solid];
-"424 quantize_per_tensor_default_25" -> "425 dequantize_per_tensor_default_25"  [label="(1, 28, 28, 768)", style=solid];
-"425 dequantize_per_tensor_default_25" -> "432 linear_24"  [label="(1, 28, 28, 768)", style=solid];
-"426 _param_constant68_0_0" -> "432 linear_24"  [label="(192,)", style=solid];
-"427 scale_updated_constant17" -> "429 mul_tensor_17"  [label="(192, 1)", style=solid];
-"428 compressed_weight_updated_constant17" -> "429 mul_tensor_17"  [label="(192, 768)", style=solid];
-"429 mul_tensor_17" -> "431 sub_tensor_17"  [label="(192, 768)", style=solid];
-"430 zero_point_updated_constant17" -> "431 sub_tensor_17"  [label="(192, 1)", style=solid];
-"431 sub_tensor_17" -> "432 linear_24"  [label="(192, 768)", style=solid];
-"432 linear_24" -> "433 dropout_15"  [label="(1, 28, 28, 192)", style=solid];
-"433 dropout_15" -> "436 layer_norm_9"  [label="(1, 28, 28, 192)", style=solid];
-"434 _param_constant69" -> "436 layer_norm_9"  [label="(192,)", style=solid];
-"435 _param_constant70" -> "436 layer_norm_9"  [label="(192,)", style=solid];
-"436 layer_norm_9" -> "437 add_13"  [label="(1, 28, 28, 192)", style=solid];
-"437 add_13" -> "438 pad_5"  [label="(1, 28, 28, 192)", style=solid];
-"438 pad_5" -> "439 slice_65"  [label="(1, 28, 28, 192)", style=solid];
-"438 pad_5" -> "442 slice_68"  [label="(1, 28, 28, 192)", style=solid];
-"438 pad_5" -> "445 slice_71"  [label="(1, 28, 28, 192)", style=solid];
-"438 pad_5" -> "448 slice_74"  [label="(1, 28, 28, 192)", style=solid];
-"439 slice_65" -> "440 slice_66"  [label="(1, 14, 28, 192)", style=solid];
-"440 slice_66" -> "441 slice_67"  [label="(1, 14, 14, 192)", style=solid];
-"441 slice_67" -> "451 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"442 slice_68" -> "443 slice_69"  [label="(1, 14, 28, 192)", style=solid];
-"443 slice_69" -> "444 slice_70"  [label="(1, 14, 14, 192)", style=solid];
-"444 slice_70" -> "451 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"445 slice_71" -> "446 slice_72"  [label="(1, 14, 28, 192)", style=solid];
-"446 slice_72" -> "447 slice_73"  [label="(1, 14, 14, 192)", style=solid];
-"447 slice_73" -> "451 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"448 slice_74" -> "449 slice_75"  [label="(1, 14, 28, 192)", style=solid];
-"449 slice_75" -> "450 slice_76"  [label="(1, 14, 14, 192)", style=solid];
-"450 slice_76" -> "451 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"451 cat_1" -> "452 cat_1_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 768)", style=solid];
-"452 cat_1_0_0_nncf_smooth_quant_0" -> "453 quantize_per_tensor_default_26"  [label="(1, 14, 14, 768)", style=solid];
-"453 quantize_per_tensor_default_26" -> "454 dequantize_per_tensor_default_26"  [label="(1, 14, 14, 768)", style=solid];
-"454 dequantize_per_tensor_default_26" -> "460 linear_25"  [label="(1, 14, 14, 768)", style=solid];
-"455 scale_updated_constant18" -> "457 mul_tensor_18"  [label="(384, 1)", style=solid];
-"456 compressed_weight_updated_constant18" -> "457 mul_tensor_18"  [label="(384, 768)", style=solid];
-"457 mul_tensor_18" -> "459 sub_tensor_18"  [label="(384, 768)", style=solid];
-"458 zero_point_updated_constant18" -> "459 sub_tensor_18"  [label="(384, 1)", style=solid];
-"459 sub_tensor_18" -> "460 linear_25"  [label="(384, 768)", style=solid];
-"460 linear_25" -> "463 layer_norm_10"  [label="(1, 14, 14, 384)", style=solid];
-"461 _param_constant72" -> "463 layer_norm_10"  [label="(384,)", style=solid];
-"462 _param_constant73" -> "463 layer_norm_10"  [label="(384,)", style=solid];
-"463 layer_norm_10" -> "465 pad_6"  [label="(1, 14, 14, 384)", style=solid];
-"463 layer_norm_10" -> "528 add_15"  [label="(1, 14, 14, 384)", style=solid];
-"464 _frozen_param34" -> "500 add_14"  [label="(1, 12, 64, 64)", style=solid];
-"465 pad_6" -> "466 view_24"  [label="(1, 16, 16, 384)", style=solid];
-"466 view_24" -> "467 permute_20"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"467 permute_20" -> "468 reshape_18"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"468 reshape_18" -> "469 reshape_18_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"469 reshape_18_0_0_nncf_smooth_quant_0" -> "470 quantize_per_tensor_default_27"  [label="(4, 64, 384)", style=solid];
-"470 quantize_per_tensor_default_27" -> "471 dequantize_per_tensor_default_27"  [label="(4, 64, 384)", style=solid];
-"471 dequantize_per_tensor_default_27" -> "478 linear_28"  [label="(4, 64, 384)", style=solid];
-"472 _frozen_param35_0_0" -> "478 linear_28"  [label="(1152,)", style=solid];
-"473 scale_updated_constant19" -> "475 mul_tensor_19"  [label="(1152, 1)", style=solid];
-"474 compressed_weight_updated_constant19" -> "475 mul_tensor_19"  [label="(1152, 384)", style=solid];
-"475 mul_tensor_19" -> "477 sub_tensor_19"  [label="(1152, 384)", style=solid];
-"476 zero_point_updated_constant19" -> "477 sub_tensor_19"  [label="(1152, 1)", style=solid];
-"477 sub_tensor_19" -> "478 linear_28"  [label="(1152, 384)", style=solid];
-"478 linear_28" -> "479 reshape_19"  [label="(4, 64, 1152)", style=solid];
-"479 reshape_19" -> "480 permute_21"  [label="(4, 64, 3, 12, 32)", style=solid];
-"480 permute_21" -> "481 select_12"  [label="(3, 4, 12, 64, 32)", style=solid];
-"480 permute_21" -> "482 select_13"  [label="(3, 4, 12, 64, 32)", style=solid];
-"480 permute_21" -> "483 select_14"  [label="(3, 4, 12, 64, 32)", style=solid];
-"481 select_12" -> "484 linalg_vector_norm_8"  [label="(4, 12, 64, 32)", style=solid];
-"481 select_12" -> "486 expand_as_8"  [label="(4, 12, 64, 32)", style=solid];
-"481 select_12" -> "487 div_8"  [label="(4, 12, 64, 32)", style=solid];
-"482 select_13" -> "490 linalg_vector_norm_9"  [label="(4, 12, 64, 32)", style=solid];
-"482 select_13" -> "492 expand_as_9"  [label="(4, 12, 64, 32)", style=solid];
-"482 select_13" -> "493 div_9"  [label="(4, 12, 64, 32)", style=solid];
-"483 select_14" -> "503 matmul_9"  [label="(4, 12, 64, 32)", style=solid];
-"484 linalg_vector_norm_8" -> "485 clamp_min_8"  [label="(4, 12, 64, 1)", style=solid];
-"485 clamp_min_8" -> "486 expand_as_8"  [label="(4, 12, 64, 1)", style=solid];
-"486 expand_as_8" -> "487 div_8"  [label="(4, 12, 64, 32)", style=solid];
-"487 div_8" -> "488 quantize_per_tensor_default_28"  [label="(4, 12, 64, 32)", style=solid];
-"488 quantize_per_tensor_default_28" -> "489 dequantize_per_tensor_default_28"  [label="(4, 12, 64, 32)", style=solid];
-"489 dequantize_per_tensor_default_28" -> "497 matmul_8"  [label="(4, 12, 64, 32)", style=solid];
-"490 linalg_vector_norm_9" -> "491 clamp_min_9"  [label="(4, 12, 64, 1)", style=solid];
-"491 clamp_min_9" -> "492 expand_as_9"  [label="(4, 12, 64, 1)", style=solid];
-"492 expand_as_9" -> "493 div_9"  [label="(4, 12, 64, 32)", style=solid];
-"493 div_9" -> "494 quantize_per_tensor_default_29"  [label="(4, 12, 64, 32)", style=solid];
-"494 quantize_per_tensor_default_29" -> "495 dequantize_per_tensor_default_29"  [label="(4, 12, 64, 32)", style=solid];
-"495 dequantize_per_tensor_default_29" -> "496 transpose_8"  [label="(4, 12, 64, 32)", style=solid];
-"496 transpose_8" -> "497 matmul_8"  [label="(4, 12, 32, 64)", style=solid];
-"497 matmul_8" -> "499 mul_9"  [label="(4, 12, 64, 64)", style=solid];
-"498 _frozen_param37" -> "499 mul_9"  [label="(12, 1, 1)", style=solid];
-"499 mul_9" -> "500 add_14"  [label="(4, 12, 64, 64)", style=solid];
-"500 add_14" -> "501 softmax_4"  [label="(4, 12, 64, 64)", style=solid];
-"501 softmax_4" -> "502 dropout_16"  [label="(4, 12, 64, 64)", style=solid];
-"502 dropout_16" -> "503 matmul_9"  [label="(4, 12, 64, 64)", style=solid];
-"503 matmul_9" -> "504 transpose_9"  [label="(4, 12, 64, 32)", style=solid];
-"504 transpose_9" -> "505 reshape_20"  [label="(4, 64, 12, 32)", style=solid];
-"505 reshape_20" -> "506 reshape_20_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"506 reshape_20_0_0_nncf_smooth_quant_0" -> "507 quantize_per_tensor_default_30"  [label="(4, 64, 384)", style=solid];
-"507 quantize_per_tensor_default_30" -> "508 dequantize_per_tensor_default_30"  [label="(4, 64, 384)", style=solid];
-"508 dequantize_per_tensor_default_30" -> "515 linear_29"  [label="(4, 64, 384)", style=solid];
-"509 _param_constant81_0_0" -> "515 linear_29"  [label="(384,)", style=solid];
-"510 scale_updated_constant20" -> "512 mul_tensor_20"  [label="(384, 1)", style=solid];
-"511 compressed_weight_updated_constant20" -> "512 mul_tensor_20"  [label="(384, 384)", style=solid];
-"512 mul_tensor_20" -> "514 sub_tensor_20"  [label="(384, 384)", style=solid];
-"513 zero_point_updated_constant20" -> "514 sub_tensor_20"  [label="(384, 1)", style=solid];
-"514 sub_tensor_20" -> "515 linear_29"  [label="(384, 384)", style=solid];
-"515 linear_29" -> "516 dropout_17"  [label="(4, 64, 384)", style=solid];
-"516 dropout_17" -> "517 view_25"  [label="(4, 64, 384)", style=solid];
-"517 view_25" -> "518 permute_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"518 permute_22" -> "519 reshape_21"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"519 reshape_21" -> "520 slice_78"  [label="(1, 16, 16, 384)", style=solid];
-"520 slice_78" -> "521 slice_79"  [label="(1, 16, 16, 384)", style=solid];
-"521 slice_79" -> "522 slice_80"  [label="(1, 14, 16, 384)", style=solid];
-"522 slice_80" -> "523 slice_81"  [label="(1, 14, 14, 384)", style=solid];
-"523 slice_81" -> "524 contiguous_7"  [label="(1, 14, 14, 384)", style=solid];
-"524 contiguous_7" -> "527 layer_norm_11"  [label="(1, 14, 14, 384)", style=solid];
-"525 _param_constant82" -> "527 layer_norm_11"  [label="(384,)", style=solid];
-"526 _param_constant83" -> "527 layer_norm_11"  [label="(384,)", style=solid];
-"527 layer_norm_11" -> "528 add_15"  [label="(1, 14, 14, 384)", style=solid];
-"528 add_15" -> "529 add_15_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"528 add_15" -> "555 add_16"  [label="(1, 14, 14, 384)", style=solid];
-"529 add_15_0_0_nncf_smooth_quant_0" -> "530 quantize_per_tensor_default_31"  [label="(1, 14, 14, 384)", style=solid];
-"530 quantize_per_tensor_default_31" -> "531 dequantize_per_tensor_default_31"  [label="(1, 14, 14, 384)", style=solid];
-"531 dequantize_per_tensor_default_31" -> "538 linear_30"  [label="(1, 14, 14, 384)", style=solid];
-"532 _param_constant85_0_0" -> "538 linear_30"  [label="(1536,)", style=solid];
-"533 scale_updated_constant21" -> "535 mul_tensor_21"  [label="(1536, 1)", style=solid];
-"534 compressed_weight_updated_constant21" -> "535 mul_tensor_21"  [label="(1536, 384)", style=solid];
-"535 mul_tensor_21" -> "537 sub_tensor_21"  [label="(1536, 384)", style=solid];
-"536 zero_point_updated_constant21" -> "537 sub_tensor_21"  [label="(1536, 1)", style=solid];
-"537 sub_tensor_21" -> "538 linear_30"  [label="(1536, 384)", style=solid];
-"538 linear_30" -> "539 gelu_4"  [label="(1, 14, 14, 1536)", style=solid];
-"539 gelu_4" -> "540 dropout_18"  [label="(1, 14, 14, 1536)", style=solid];
-"540 dropout_18" -> "541 dropout_18_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"541 dropout_18_0_0_nncf_smooth_quant_0" -> "542 quantize_per_tensor_default_32"  [label="(1, 14, 14, 1536)", style=solid];
-"542 quantize_per_tensor_default_32" -> "543 dequantize_per_tensor_default_32"  [label="(1, 14, 14, 1536)", style=solid];
-"543 dequantize_per_tensor_default_32" -> "550 linear_31"  [label="(1, 14, 14, 1536)", style=solid];
-"544 _param_constant87_0_0" -> "550 linear_31"  [label="(384,)", style=solid];
-"545 scale_updated_constant22" -> "547 mul_tensor_22"  [label="(384, 1)", style=solid];
-"546 compressed_weight_updated_constant22" -> "547 mul_tensor_22"  [label="(384, 1536)", style=solid];
-"547 mul_tensor_22" -> "549 sub_tensor_22"  [label="(384, 1536)", style=solid];
-"548 zero_point_updated_constant22" -> "549 sub_tensor_22"  [label="(384, 1)", style=solid];
-"549 sub_tensor_22" -> "550 linear_31"  [label="(384, 1536)", style=solid];
-"550 linear_31" -> "551 dropout_19"  [label="(1, 14, 14, 384)", style=solid];
-"551 dropout_19" -> "554 layer_norm_12"  [label="(1, 14, 14, 384)", style=solid];
-"552 _param_constant88" -> "554 layer_norm_12"  [label="(384,)", style=solid];
-"553 _param_constant89" -> "554 layer_norm_12"  [label="(384,)", style=solid];
-"554 layer_norm_12" -> "555 add_16"  [label="(1, 14, 14, 384)", style=solid];
-"555 add_16" -> "557 pad_7"  [label="(1, 14, 14, 384)", style=solid];
-"555 add_16" -> "638 add_19"  [label="(1, 14, 14, 384)", style=solid];
-"556 _frozen_param38" -> "593 add_17"  [label="(1, 12, 64, 64)", style=solid];
-"557 pad_7" -> "558 roll_4"  [label="(1, 16, 16, 384)", style=solid];
-"558 roll_4" -> "559 view_28"  [label="(1, 16, 16, 384)", style=solid];
-"559 view_28" -> "560 permute_24"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"560 permute_24" -> "561 reshape_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"561 reshape_22" -> "562 reshape_22_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"561 reshape_22" -> "594 new_zeros_2"  [label="(4, 64, 384)", style=solid];
-"562 reshape_22_0_0_nncf_smooth_quant_0" -> "563 quantize_per_tensor_default_33"  [label="(4, 64, 384)", style=solid];
-"563 quantize_per_tensor_default_33" -> "564 dequantize_per_tensor_default_33"  [label="(4, 64, 384)", style=solid];
-"564 dequantize_per_tensor_default_33" -> "571 linear_34"  [label="(4, 64, 384)", style=solid];
-"565 _frozen_param39_0_0" -> "571 linear_34"  [label="(1152,)", style=solid];
-"566 scale_updated_constant23" -> "568 mul_tensor_23"  [label="(1152, 1)", style=solid];
-"567 compressed_weight_updated_constant23" -> "568 mul_tensor_23"  [label="(1152, 384)", style=solid];
-"568 mul_tensor_23" -> "570 sub_tensor_23"  [label="(1152, 384)", style=solid];
-"569 zero_point_updated_constant23" -> "570 sub_tensor_23"  [label="(1152, 1)", style=solid];
-"570 sub_tensor_23" -> "571 linear_34"  [label="(1152, 384)", style=solid];
-"571 linear_34" -> "572 reshape_23"  [label="(4, 64, 1152)", style=solid];
-"572 reshape_23" -> "573 permute_25"  [label="(4, 64, 3, 12, 32)", style=solid];
-"573 permute_25" -> "574 select_15"  [label="(3, 4, 12, 64, 32)", style=solid];
-"573 permute_25" -> "575 select_16"  [label="(3, 4, 12, 64, 32)", style=solid];
-"573 permute_25" -> "576 select_17"  [label="(3, 4, 12, 64, 32)", style=solid];
-"574 select_15" -> "577 linalg_vector_norm_10"  [label="(4, 12, 64, 32)", style=solid];
-"574 select_15" -> "579 expand_as_10"  [label="(4, 12, 64, 32)", style=solid];
-"574 select_15" -> "580 div_10"  [label="(4, 12, 64, 32)", style=solid];
-"575 select_16" -> "583 linalg_vector_norm_11"  [label="(4, 12, 64, 32)", style=solid];
-"575 select_16" -> "585 expand_as_11"  [label="(4, 12, 64, 32)", style=solid];
-"575 select_16" -> "586 div_11"  [label="(4, 12, 64, 32)", style=solid];
-"576 select_17" -> "612 matmul_11"  [label="(4, 12, 64, 32)", style=solid];
-"577 linalg_vector_norm_10" -> "578 clamp_min_10"  [label="(4, 12, 64, 1)", style=solid];
-"578 clamp_min_10" -> "579 expand_as_10"  [label="(4, 12, 64, 1)", style=solid];
-"579 expand_as_10" -> "580 div_10"  [label="(4, 12, 64, 32)", style=solid];
-"580 div_10" -> "581 quantize_per_tensor_default_34"  [label="(4, 12, 64, 32)", style=solid];
-"581 quantize_per_tensor_default_34" -> "582 dequantize_per_tensor_default_34"  [label="(4, 12, 64, 32)", style=solid];
-"582 dequantize_per_tensor_default_34" -> "590 matmul_10"  [label="(4, 12, 64, 32)", style=solid];
-"583 linalg_vector_norm_11" -> "584 clamp_min_11"  [label="(4, 12, 64, 1)", style=solid];
-"584 clamp_min_11" -> "585 expand_as_11"  [label="(4, 12, 64, 1)", style=solid];
-"585 expand_as_11" -> "586 div_11"  [label="(4, 12, 64, 32)", style=solid];
-"586 div_11" -> "587 quantize_per_tensor_default_35"  [label="(4, 12, 64, 32)", style=solid];
-"587 quantize_per_tensor_default_35" -> "588 dequantize_per_tensor_default_35"  [label="(4, 12, 64, 32)", style=solid];
-"588 dequantize_per_tensor_default_35" -> "589 transpose_10"  [label="(4, 12, 64, 32)", style=solid];
-"589 transpose_10" -> "590 matmul_10"  [label="(4, 12, 32, 64)", style=solid];
-"590 matmul_10" -> "592 mul_11"  [label="(4, 12, 64, 64)", style=solid];
-"591 _frozen_param41" -> "592 mul_11"  [label="(12, 1, 1)", style=solid];
-"592 mul_11" -> "593 add_17"  [label="(4, 12, 64, 64)", style=solid];
-"593 add_17" -> "605 view_30"  [label="(4, 12, 64, 64)", style=solid];
-"594 new_zeros_2" -> "595 view_29"  [label="(16, 16)", style=solid];
-"595 view_29" -> "596 permute_26"  [label="(2, 8, 2, 8)", style=solid];
-"596 permute_26" -> "597 reshape_24"  [label="(2, 2, 8, 8)", style=solid];
-"597 reshape_24" -> "598 unsqueeze_14"  [label="(4, 64)", style=solid];
-"597 reshape_24" -> "599 unsqueeze_15"  [label="(4, 64)", style=solid];
-"598 unsqueeze_14" -> "600 sub_2"  [label="(4, 1, 64)", style=solid];
-"599 unsqueeze_15" -> "600 sub_2"  [label="(4, 64, 1)", style=solid];
-"600 sub_2" -> "601 ne_2"  [label="(4, 64, 64)", style=solid];
-"600 sub_2" -> "602 masked_fill_4"  [label="(4, 64, 64)", style=solid];
-"600 sub_2" -> "603 eq_2"  [label="(4, 64, 64)", style=solid];
-"601 ne_2" -> "602 masked_fill_4"  [label="(4, 64, 64)", style=solid];
-"602 masked_fill_4" -> "604 masked_fill_5"  [label="(4, 64, 64)", style=solid];
-"603 eq_2" -> "604 masked_fill_5"  [label="(4, 64, 64)", style=solid];
-"604 masked_fill_5" -> "606 unsqueeze_16"  [label="(4, 64, 64)", style=solid];
-"605 view_30" -> "608 add_18"  [label="(1, 4, 12, 64, 64)", style=solid];
-"606 unsqueeze_16" -> "607 unsqueeze_17"  [label="(4, 1, 64, 64)", style=solid];
-"607 unsqueeze_17" -> "608 add_18"  [label="(1, 4, 1, 64, 64)", style=solid];
-"608 add_18" -> "609 view_31"  [label="(1, 4, 12, 64, 64)", style=solid];
-"609 view_31" -> "610 softmax_5"  [label="(4, 12, 64, 64)", style=solid];
-"610 softmax_5" -> "611 dropout_20"  [label="(4, 12, 64, 64)", style=solid];
-"611 dropout_20" -> "612 matmul_11"  [label="(4, 12, 64, 64)", style=solid];
-"612 matmul_11" -> "613 transpose_11"  [label="(4, 12, 64, 32)", style=solid];
-"613 transpose_11" -> "614 reshape_25"  [label="(4, 64, 12, 32)", style=solid];
-"614 reshape_25" -> "615 reshape_25_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"615 reshape_25_0_0_nncf_smooth_quant_0" -> "616 quantize_per_tensor_default_36"  [label="(4, 64, 384)", style=solid];
-"616 quantize_per_tensor_default_36" -> "617 dequantize_per_tensor_default_36"  [label="(4, 64, 384)", style=solid];
-"617 dequantize_per_tensor_default_36" -> "624 linear_35"  [label="(4, 64, 384)", style=solid];
-"618 _param_constant97_0_0" -> "624 linear_35"  [label="(384,)", style=solid];
-"619 scale_updated_constant24" -> "621 mul_tensor_24"  [label="(384, 1)", style=solid];
-"620 compressed_weight_updated_constant24" -> "621 mul_tensor_24"  [label="(384, 384)", style=solid];
-"621 mul_tensor_24" -> "623 sub_tensor_24"  [label="(384, 384)", style=solid];
-"622 zero_point_updated_constant24" -> "623 sub_tensor_24"  [label="(384, 1)", style=solid];
-"623 sub_tensor_24" -> "624 linear_35"  [label="(384, 384)", style=solid];
-"624 linear_35" -> "625 dropout_21"  [label="(4, 64, 384)", style=solid];
-"625 dropout_21" -> "626 view_32"  [label="(4, 64, 384)", style=solid];
-"626 view_32" -> "627 permute_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"627 permute_27" -> "628 reshape_26"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"628 reshape_26" -> "629 roll_5"  [label="(1, 16, 16, 384)", style=solid];
-"629 roll_5" -> "630 slice_101"  [label="(1, 16, 16, 384)", style=solid];
-"630 slice_101" -> "631 slice_102"  [label="(1, 16, 16, 384)", style=solid];
-"631 slice_102" -> "632 slice_103"  [label="(1, 14, 16, 384)", style=solid];
-"632 slice_103" -> "633 slice_104"  [label="(1, 14, 14, 384)", style=solid];
-"633 slice_104" -> "634 contiguous_9"  [label="(1, 14, 14, 384)", style=solid];
-"634 contiguous_9" -> "637 layer_norm_13"  [label="(1, 14, 14, 384)", style=solid];
-"635 _param_constant98" -> "637 layer_norm_13"  [label="(384,)", style=solid];
-"636 _param_constant99" -> "637 layer_norm_13"  [label="(384,)", style=solid];
-"637 layer_norm_13" -> "638 add_19"  [label="(1, 14, 14, 384)", style=solid];
-"638 add_19" -> "639 add_19_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"638 add_19" -> "665 add_20"  [label="(1, 14, 14, 384)", style=solid];
-"639 add_19_0_0_nncf_smooth_quant_0" -> "640 quantize_per_tensor_default_37"  [label="(1, 14, 14, 384)", style=solid];
-"640 quantize_per_tensor_default_37" -> "641 dequantize_per_tensor_default_37"  [label="(1, 14, 14, 384)", style=solid];
-"641 dequantize_per_tensor_default_37" -> "648 linear_36"  [label="(1, 14, 14, 384)", style=solid];
-"642 _param_constant101_0_0" -> "648 linear_36"  [label="(1536,)", style=solid];
-"643 scale_updated_constant25" -> "645 mul_tensor_25"  [label="(1536, 1)", style=solid];
-"644 compressed_weight_updated_constant25" -> "645 mul_tensor_25"  [label="(1536, 384)", style=solid];
-"645 mul_tensor_25" -> "647 sub_tensor_25"  [label="(1536, 384)", style=solid];
-"646 zero_point_updated_constant25" -> "647 sub_tensor_25"  [label="(1536, 1)", style=solid];
-"647 sub_tensor_25" -> "648 linear_36"  [label="(1536, 384)", style=solid];
-"648 linear_36" -> "649 gelu_5"  [label="(1, 14, 14, 1536)", style=solid];
-"649 gelu_5" -> "650 dropout_22"  [label="(1, 14, 14, 1536)", style=solid];
-"650 dropout_22" -> "651 dropout_22_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"651 dropout_22_0_0_nncf_smooth_quant_0" -> "652 quantize_per_tensor_default_38"  [label="(1, 14, 14, 1536)", style=solid];
-"652 quantize_per_tensor_default_38" -> "653 dequantize_per_tensor_default_38"  [label="(1, 14, 14, 1536)", style=solid];
-"653 dequantize_per_tensor_default_38" -> "660 linear_37"  [label="(1, 14, 14, 1536)", style=solid];
-"654 _param_constant103_0_0" -> "660 linear_37"  [label="(384,)", style=solid];
-"655 scale_updated_constant26" -> "657 mul_tensor_26"  [label="(384, 1)", style=solid];
-"656 compressed_weight_updated_constant26" -> "657 mul_tensor_26"  [label="(384, 1536)", style=solid];
-"657 mul_tensor_26" -> "659 sub_tensor_26"  [label="(384, 1536)", style=solid];
-"658 zero_point_updated_constant26" -> "659 sub_tensor_26"  [label="(384, 1)", style=solid];
-"659 sub_tensor_26" -> "660 linear_37"  [label="(384, 1536)", style=solid];
-"660 linear_37" -> "661 dropout_23"  [label="(1, 14, 14, 384)", style=solid];
-"661 dropout_23" -> "664 layer_norm_14"  [label="(1, 14, 14, 384)", style=solid];
-"662 _param_constant104" -> "664 layer_norm_14"  [label="(384,)", style=solid];
-"663 _param_constant105" -> "664 layer_norm_14"  [label="(384,)", style=solid];
-"664 layer_norm_14" -> "665 add_20"  [label="(1, 14, 14, 384)", style=solid];
-"665 add_20" -> "667 pad_8"  [label="(1, 14, 14, 384)", style=solid];
-"665 add_20" -> "730 add_22"  [label="(1, 14, 14, 384)", style=solid];
-"666 _frozen_param51" -> "702 add_21"  [label="(1, 12, 64, 64)", style=solid];
-"667 pad_8" -> "668 view_35"  [label="(1, 16, 16, 384)", style=solid];
-"668 view_35" -> "669 permute_29"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"669 permute_29" -> "670 reshape_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"670 reshape_27" -> "671 reshape_27_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"671 reshape_27_0_0_nncf_smooth_quant_0" -> "672 quantize_per_tensor_default_39"  [label="(4, 64, 384)", style=solid];
-"672 quantize_per_tensor_default_39" -> "673 dequantize_per_tensor_default_39"  [label="(4, 64, 384)", style=solid];
-"673 dequantize_per_tensor_default_39" -> "680 linear_40"  [label="(4, 64, 384)", style=solid];
-"674 _frozen_param52_0_0" -> "680 linear_40"  [label="(1152,)", style=solid];
-"675 scale_updated_constant27" -> "677 mul_tensor_27"  [label="(1152, 1)", style=solid];
-"676 compressed_weight_updated_constant27" -> "677 mul_tensor_27"  [label="(1152, 384)", style=solid];
-"677 mul_tensor_27" -> "679 sub_tensor_27"  [label="(1152, 384)", style=solid];
-"678 zero_point_updated_constant27" -> "679 sub_tensor_27"  [label="(1152, 1)", style=solid];
-"679 sub_tensor_27" -> "680 linear_40"  [label="(1152, 384)", style=solid];
-"680 linear_40" -> "681 reshape_28"  [label="(4, 64, 1152)", style=solid];
-"681 reshape_28" -> "682 permute_30"  [label="(4, 64, 3, 12, 32)", style=solid];
-"682 permute_30" -> "683 select_18"  [label="(3, 4, 12, 64, 32)", style=solid];
-"682 permute_30" -> "684 select_19"  [label="(3, 4, 12, 64, 32)", style=solid];
-"682 permute_30" -> "685 select_20"  [label="(3, 4, 12, 64, 32)", style=solid];
-"683 select_18" -> "686 linalg_vector_norm_12"  [label="(4, 12, 64, 32)", style=solid];
-"683 select_18" -> "688 expand_as_12"  [label="(4, 12, 64, 32)", style=solid];
-"683 select_18" -> "689 div_12"  [label="(4, 12, 64, 32)", style=solid];
-"684 select_19" -> "692 linalg_vector_norm_13"  [label="(4, 12, 64, 32)", style=solid];
-"684 select_19" -> "694 expand_as_13"  [label="(4, 12, 64, 32)", style=solid];
-"684 select_19" -> "695 div_13"  [label="(4, 12, 64, 32)", style=solid];
-"685 select_20" -> "705 matmul_13"  [label="(4, 12, 64, 32)", style=solid];
-"686 linalg_vector_norm_12" -> "687 clamp_min_12"  [label="(4, 12, 64, 1)", style=solid];
-"687 clamp_min_12" -> "688 expand_as_12"  [label="(4, 12, 64, 1)", style=solid];
-"688 expand_as_12" -> "689 div_12"  [label="(4, 12, 64, 32)", style=solid];
-"689 div_12" -> "690 quantize_per_tensor_default_40"  [label="(4, 12, 64, 32)", style=solid];
-"690 quantize_per_tensor_default_40" -> "691 dequantize_per_tensor_default_40"  [label="(4, 12, 64, 32)", style=solid];
-"691 dequantize_per_tensor_default_40" -> "699 matmul_12"  [label="(4, 12, 64, 32)", style=solid];
-"692 linalg_vector_norm_13" -> "693 clamp_min_13"  [label="(4, 12, 64, 1)", style=solid];
-"693 clamp_min_13" -> "694 expand_as_13"  [label="(4, 12, 64, 1)", style=solid];
-"694 expand_as_13" -> "695 div_13"  [label="(4, 12, 64, 32)", style=solid];
-"695 div_13" -> "696 quantize_per_tensor_default_41"  [label="(4, 12, 64, 32)", style=solid];
-"696 quantize_per_tensor_default_41" -> "697 dequantize_per_tensor_default_41"  [label="(4, 12, 64, 32)", style=solid];
-"697 dequantize_per_tensor_default_41" -> "698 transpose_12"  [label="(4, 12, 64, 32)", style=solid];
-"698 transpose_12" -> "699 matmul_12"  [label="(4, 12, 32, 64)", style=solid];
-"699 matmul_12" -> "701 mul_13"  [label="(4, 12, 64, 64)", style=solid];
-"700 _frozen_param54" -> "701 mul_13"  [label="(12, 1, 1)", style=solid];
-"701 mul_13" -> "702 add_21"  [label="(4, 12, 64, 64)", style=solid];
-"702 add_21" -> "703 softmax_6"  [label="(4, 12, 64, 64)", style=solid];
-"703 softmax_6" -> "704 dropout_24"  [label="(4, 12, 64, 64)", style=solid];
-"704 dropout_24" -> "705 matmul_13"  [label="(4, 12, 64, 64)", style=solid];
-"705 matmul_13" -> "706 transpose_13"  [label="(4, 12, 64, 32)", style=solid];
-"706 transpose_13" -> "707 reshape_29"  [label="(4, 64, 12, 32)", style=solid];
-"707 reshape_29" -> "708 reshape_29_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"708 reshape_29_0_0_nncf_smooth_quant_0" -> "709 quantize_per_tensor_default_42"  [label="(4, 64, 384)", style=solid];
-"709 quantize_per_tensor_default_42" -> "710 dequantize_per_tensor_default_42"  [label="(4, 64, 384)", style=solid];
-"710 dequantize_per_tensor_default_42" -> "717 linear_41"  [label="(4, 64, 384)", style=solid];
-"711 _param_constant113_0_0" -> "717 linear_41"  [label="(384,)", style=solid];
-"712 scale_updated_constant28" -> "714 mul_tensor_28"  [label="(384, 1)", style=solid];
-"713 compressed_weight_updated_constant28" -> "714 mul_tensor_28"  [label="(384, 384)", style=solid];
-"714 mul_tensor_28" -> "716 sub_tensor_28"  [label="(384, 384)", style=solid];
-"715 zero_point_updated_constant28" -> "716 sub_tensor_28"  [label="(384, 1)", style=solid];
-"716 sub_tensor_28" -> "717 linear_41"  [label="(384, 384)", style=solid];
-"717 linear_41" -> "718 dropout_25"  [label="(4, 64, 384)", style=solid];
-"718 dropout_25" -> "719 view_36"  [label="(4, 64, 384)", style=solid];
-"719 view_36" -> "720 permute_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"720 permute_31" -> "721 reshape_30"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"721 reshape_30" -> "722 slice_106"  [label="(1, 16, 16, 384)", style=solid];
-"722 slice_106" -> "723 slice_107"  [label="(1, 16, 16, 384)", style=solid];
-"723 slice_107" -> "724 slice_108"  [label="(1, 14, 16, 384)", style=solid];
-"724 slice_108" -> "725 slice_109"  [label="(1, 14, 14, 384)", style=solid];
-"725 slice_109" -> "726 contiguous_11"  [label="(1, 14, 14, 384)", style=solid];
-"726 contiguous_11" -> "729 layer_norm_15"  [label="(1, 14, 14, 384)", style=solid];
-"727 _param_constant114" -> "729 layer_norm_15"  [label="(384,)", style=solid];
-"728 _param_constant115" -> "729 layer_norm_15"  [label="(384,)", style=solid];
-"729 layer_norm_15" -> "730 add_22"  [label="(1, 14, 14, 384)", style=solid];
-"730 add_22" -> "731 add_22_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"730 add_22" -> "757 add_23"  [label="(1, 14, 14, 384)", style=solid];
-"731 add_22_0_0_nncf_smooth_quant_0" -> "732 quantize_per_tensor_default_43"  [label="(1, 14, 14, 384)", style=solid];
-"732 quantize_per_tensor_default_43" -> "733 dequantize_per_tensor_default_43"  [label="(1, 14, 14, 384)", style=solid];
-"733 dequantize_per_tensor_default_43" -> "740 linear_42"  [label="(1, 14, 14, 384)", style=solid];
-"734 _param_constant117_0_0" -> "740 linear_42"  [label="(1536,)", style=solid];
-"735 scale_updated_constant29" -> "737 mul_tensor_29"  [label="(1536, 1)", style=solid];
-"736 compressed_weight_updated_constant29" -> "737 mul_tensor_29"  [label="(1536, 384)", style=solid];
-"737 mul_tensor_29" -> "739 sub_tensor_29"  [label="(1536, 384)", style=solid];
-"738 zero_point_updated_constant29" -> "739 sub_tensor_29"  [label="(1536, 1)", style=solid];
-"739 sub_tensor_29" -> "740 linear_42"  [label="(1536, 384)", style=solid];
-"740 linear_42" -> "741 gelu_6"  [label="(1, 14, 14, 1536)", style=solid];
-"741 gelu_6" -> "742 dropout_26"  [label="(1, 14, 14, 1536)", style=solid];
-"742 dropout_26" -> "743 dropout_26_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"743 dropout_26_0_0_nncf_smooth_quant_0" -> "744 quantize_per_tensor_default_44"  [label="(1, 14, 14, 1536)", style=solid];
-"744 quantize_per_tensor_default_44" -> "745 dequantize_per_tensor_default_44"  [label="(1, 14, 14, 1536)", style=solid];
-"745 dequantize_per_tensor_default_44" -> "752 linear_43"  [label="(1, 14, 14, 1536)", style=solid];
-"746 _param_constant119_0_0" -> "752 linear_43"  [label="(384,)", style=solid];
-"747 scale_updated_constant30" -> "749 mul_tensor_30"  [label="(384, 1)", style=solid];
-"748 compressed_weight_updated_constant30" -> "749 mul_tensor_30"  [label="(384, 1536)", style=solid];
-"749 mul_tensor_30" -> "751 sub_tensor_30"  [label="(384, 1536)", style=solid];
-"750 zero_point_updated_constant30" -> "751 sub_tensor_30"  [label="(384, 1)", style=solid];
-"751 sub_tensor_30" -> "752 linear_43"  [label="(384, 1536)", style=solid];
-"752 linear_43" -> "753 dropout_27"  [label="(1, 14, 14, 384)", style=solid];
-"753 dropout_27" -> "756 layer_norm_16"  [label="(1, 14, 14, 384)", style=solid];
-"754 _param_constant120" -> "756 layer_norm_16"  [label="(384,)", style=solid];
-"755 _param_constant121" -> "756 layer_norm_16"  [label="(384,)", style=solid];
-"756 layer_norm_16" -> "757 add_23"  [label="(1, 14, 14, 384)", style=solid];
-"757 add_23" -> "759 pad_9"  [label="(1, 14, 14, 384)", style=solid];
-"757 add_23" -> "840 add_26"  [label="(1, 14, 14, 384)", style=solid];
-"758 _frozen_param55" -> "795 add_24"  [label="(1, 12, 64, 64)", style=solid];
-"759 pad_9" -> "760 roll_6"  [label="(1, 16, 16, 384)", style=solid];
-"760 roll_6" -> "761 view_39"  [label="(1, 16, 16, 384)", style=solid];
-"761 view_39" -> "762 permute_33"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"762 permute_33" -> "763 reshape_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"763 reshape_31" -> "764 reshape_31_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"763 reshape_31" -> "796 new_zeros_3"  [label="(4, 64, 384)", style=solid];
-"764 reshape_31_0_0_nncf_smooth_quant_0" -> "765 quantize_per_tensor_default_45"  [label="(4, 64, 384)", style=solid];
-"765 quantize_per_tensor_default_45" -> "766 dequantize_per_tensor_default_45"  [label="(4, 64, 384)", style=solid];
-"766 dequantize_per_tensor_default_45" -> "773 linear_46"  [label="(4, 64, 384)", style=solid];
-"767 _frozen_param56_0_0" -> "773 linear_46"  [label="(1152,)", style=solid];
-"768 scale_updated_constant31" -> "770 mul_tensor_31"  [label="(1152, 1)", style=solid];
-"769 compressed_weight_updated_constant31" -> "770 mul_tensor_31"  [label="(1152, 384)", style=solid];
-"770 mul_tensor_31" -> "772 sub_tensor_31"  [label="(1152, 384)", style=solid];
-"771 zero_point_updated_constant31" -> "772 sub_tensor_31"  [label="(1152, 1)", style=solid];
-"772 sub_tensor_31" -> "773 linear_46"  [label="(1152, 384)", style=solid];
-"773 linear_46" -> "774 reshape_32"  [label="(4, 64, 1152)", style=solid];
-"774 reshape_32" -> "775 permute_34"  [label="(4, 64, 3, 12, 32)", style=solid];
-"775 permute_34" -> "776 select_21"  [label="(3, 4, 12, 64, 32)", style=solid];
-"775 permute_34" -> "777 select_22"  [label="(3, 4, 12, 64, 32)", style=solid];
-"775 permute_34" -> "778 select_23"  [label="(3, 4, 12, 64, 32)", style=solid];
-"776 select_21" -> "779 linalg_vector_norm_14"  [label="(4, 12, 64, 32)", style=solid];
-"776 select_21" -> "781 expand_as_14"  [label="(4, 12, 64, 32)", style=solid];
-"776 select_21" -> "782 div_14"  [label="(4, 12, 64, 32)", style=solid];
-"777 select_22" -> "785 linalg_vector_norm_15"  [label="(4, 12, 64, 32)", style=solid];
-"777 select_22" -> "787 expand_as_15"  [label="(4, 12, 64, 32)", style=solid];
-"777 select_22" -> "788 div_15"  [label="(4, 12, 64, 32)", style=solid];
-"778 select_23" -> "814 matmul_15"  [label="(4, 12, 64, 32)", style=solid];
-"779 linalg_vector_norm_14" -> "780 clamp_min_14"  [label="(4, 12, 64, 1)", style=solid];
-"780 clamp_min_14" -> "781 expand_as_14"  [label="(4, 12, 64, 1)", style=solid];
-"781 expand_as_14" -> "782 div_14"  [label="(4, 12, 64, 32)", style=solid];
-"782 div_14" -> "783 quantize_per_tensor_default_46"  [label="(4, 12, 64, 32)", style=solid];
-"783 quantize_per_tensor_default_46" -> "784 dequantize_per_tensor_default_46"  [label="(4, 12, 64, 32)", style=solid];
-"784 dequantize_per_tensor_default_46" -> "792 matmul_14"  [label="(4, 12, 64, 32)", style=solid];
-"785 linalg_vector_norm_15" -> "786 clamp_min_15"  [label="(4, 12, 64, 1)", style=solid];
-"786 clamp_min_15" -> "787 expand_as_15"  [label="(4, 12, 64, 1)", style=solid];
-"787 expand_as_15" -> "788 div_15"  [label="(4, 12, 64, 32)", style=solid];
-"788 div_15" -> "789 quantize_per_tensor_default_47"  [label="(4, 12, 64, 32)", style=solid];
-"789 quantize_per_tensor_default_47" -> "790 dequantize_per_tensor_default_47"  [label="(4, 12, 64, 32)", style=solid];
-"790 dequantize_per_tensor_default_47" -> "791 transpose_14"  [label="(4, 12, 64, 32)", style=solid];
-"791 transpose_14" -> "792 matmul_14"  [label="(4, 12, 32, 64)", style=solid];
-"792 matmul_14" -> "794 mul_15"  [label="(4, 12, 64, 64)", style=solid];
-"793 _frozen_param58" -> "794 mul_15"  [label="(12, 1, 1)", style=solid];
-"794 mul_15" -> "795 add_24"  [label="(4, 12, 64, 64)", style=solid];
-"795 add_24" -> "807 view_41"  [label="(4, 12, 64, 64)", style=solid];
-"796 new_zeros_3" -> "797 view_40"  [label="(16, 16)", style=solid];
-"797 view_40" -> "798 permute_35"  [label="(2, 8, 2, 8)", style=solid];
-"798 permute_35" -> "799 reshape_33"  [label="(2, 2, 8, 8)", style=solid];
-"799 reshape_33" -> "800 unsqueeze_20"  [label="(4, 64)", style=solid];
-"799 reshape_33" -> "801 unsqueeze_21"  [label="(4, 64)", style=solid];
-"800 unsqueeze_20" -> "802 sub_3"  [label="(4, 1, 64)", style=solid];
-"801 unsqueeze_21" -> "802 sub_3"  [label="(4, 64, 1)", style=solid];
-"802 sub_3" -> "803 ne_3"  [label="(4, 64, 64)", style=solid];
-"802 sub_3" -> "804 masked_fill_6"  [label="(4, 64, 64)", style=solid];
-"802 sub_3" -> "805 eq_3"  [label="(4, 64, 64)", style=solid];
-"803 ne_3" -> "804 masked_fill_6"  [label="(4, 64, 64)", style=solid];
-"804 masked_fill_6" -> "806 masked_fill_7"  [label="(4, 64, 64)", style=solid];
-"805 eq_3" -> "806 masked_fill_7"  [label="(4, 64, 64)", style=solid];
-"806 masked_fill_7" -> "808 unsqueeze_22"  [label="(4, 64, 64)", style=solid];
-"807 view_41" -> "810 add_25"  [label="(1, 4, 12, 64, 64)", style=solid];
-"808 unsqueeze_22" -> "809 unsqueeze_23"  [label="(4, 1, 64, 64)", style=solid];
-"809 unsqueeze_23" -> "810 add_25"  [label="(1, 4, 1, 64, 64)", style=solid];
-"810 add_25" -> "811 view_42"  [label="(1, 4, 12, 64, 64)", style=solid];
-"811 view_42" -> "812 softmax_7"  [label="(4, 12, 64, 64)", style=solid];
-"812 softmax_7" -> "813 dropout_28"  [label="(4, 12, 64, 64)", style=solid];
-"813 dropout_28" -> "814 matmul_15"  [label="(4, 12, 64, 64)", style=solid];
-"814 matmul_15" -> "815 transpose_15"  [label="(4, 12, 64, 32)", style=solid];
-"815 transpose_15" -> "816 reshape_34"  [label="(4, 64, 12, 32)", style=solid];
-"816 reshape_34" -> "817 reshape_34_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"817 reshape_34_0_0_nncf_smooth_quant_0" -> "818 quantize_per_tensor_default_48"  [label="(4, 64, 384)", style=solid];
-"818 quantize_per_tensor_default_48" -> "819 dequantize_per_tensor_default_48"  [label="(4, 64, 384)", style=solid];
-"819 dequantize_per_tensor_default_48" -> "826 linear_47"  [label="(4, 64, 384)", style=solid];
-"820 _param_constant129_0_0" -> "826 linear_47"  [label="(384,)", style=solid];
-"821 scale_updated_constant32" -> "823 mul_tensor_32"  [label="(384, 1)", style=solid];
-"822 compressed_weight_updated_constant32" -> "823 mul_tensor_32"  [label="(384, 384)", style=solid];
-"823 mul_tensor_32" -> "825 sub_tensor_32"  [label="(384, 384)", style=solid];
-"824 zero_point_updated_constant32" -> "825 sub_tensor_32"  [label="(384, 1)", style=solid];
-"825 sub_tensor_32" -> "826 linear_47"  [label="(384, 384)", style=solid];
-"826 linear_47" -> "827 dropout_29"  [label="(4, 64, 384)", style=solid];
-"827 dropout_29" -> "828 view_43"  [label="(4, 64, 384)", style=solid];
-"828 view_43" -> "829 permute_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"829 permute_36" -> "830 reshape_35"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"830 reshape_35" -> "831 roll_7"  [label="(1, 16, 16, 384)", style=solid];
-"831 roll_7" -> "832 slice_129"  [label="(1, 16, 16, 384)", style=solid];
-"832 slice_129" -> "833 slice_130"  [label="(1, 16, 16, 384)", style=solid];
-"833 slice_130" -> "834 slice_131"  [label="(1, 14, 16, 384)", style=solid];
-"834 slice_131" -> "835 slice_132"  [label="(1, 14, 14, 384)", style=solid];
-"835 slice_132" -> "836 contiguous_13"  [label="(1, 14, 14, 384)", style=solid];
-"836 contiguous_13" -> "839 layer_norm_17"  [label="(1, 14, 14, 384)", style=solid];
-"837 _param_constant130" -> "839 layer_norm_17"  [label="(384,)", style=solid];
-"838 _param_constant131" -> "839 layer_norm_17"  [label="(384,)", style=solid];
-"839 layer_norm_17" -> "840 add_26"  [label="(1, 14, 14, 384)", style=solid];
-"840 add_26" -> "841 add_26_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"840 add_26" -> "867 add_27"  [label="(1, 14, 14, 384)", style=solid];
-"841 add_26_0_0_nncf_smooth_quant_0" -> "842 quantize_per_tensor_default_49"  [label="(1, 14, 14, 384)", style=solid];
-"842 quantize_per_tensor_default_49" -> "843 dequantize_per_tensor_default_49"  [label="(1, 14, 14, 384)", style=solid];
-"843 dequantize_per_tensor_default_49" -> "850 linear_48"  [label="(1, 14, 14, 384)", style=solid];
-"844 _param_constant133_0_0" -> "850 linear_48"  [label="(1536,)", style=solid];
-"845 scale_updated_constant33" -> "847 mul_tensor_33"  [label="(1536, 1)", style=solid];
-"846 compressed_weight_updated_constant33" -> "847 mul_tensor_33"  [label="(1536, 384)", style=solid];
-"847 mul_tensor_33" -> "849 sub_tensor_33"  [label="(1536, 384)", style=solid];
-"848 zero_point_updated_constant33" -> "849 sub_tensor_33"  [label="(1536, 1)", style=solid];
-"849 sub_tensor_33" -> "850 linear_48"  [label="(1536, 384)", style=solid];
-"850 linear_48" -> "851 gelu_7"  [label="(1, 14, 14, 1536)", style=solid];
-"851 gelu_7" -> "852 dropout_30"  [label="(1, 14, 14, 1536)", style=solid];
-"852 dropout_30" -> "853 dropout_30_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"853 dropout_30_0_0_nncf_smooth_quant_0" -> "854 quantize_per_tensor_default_50"  [label="(1, 14, 14, 1536)", style=solid];
-"854 quantize_per_tensor_default_50" -> "855 dequantize_per_tensor_default_50"  [label="(1, 14, 14, 1536)", style=solid];
-"855 dequantize_per_tensor_default_50" -> "862 linear_49"  [label="(1, 14, 14, 1536)", style=solid];
-"856 _param_constant135_0_0" -> "862 linear_49"  [label="(384,)", style=solid];
-"857 scale_updated_constant34" -> "859 mul_tensor_34"  [label="(384, 1)", style=solid];
-"858 compressed_weight_updated_constant34" -> "859 mul_tensor_34"  [label="(384, 1536)", style=solid];
-"859 mul_tensor_34" -> "861 sub_tensor_34"  [label="(384, 1536)", style=solid];
-"860 zero_point_updated_constant34" -> "861 sub_tensor_34"  [label="(384, 1)", style=solid];
-"861 sub_tensor_34" -> "862 linear_49"  [label="(384, 1536)", style=solid];
-"862 linear_49" -> "863 dropout_31"  [label="(1, 14, 14, 384)", style=solid];
-"863 dropout_31" -> "866 layer_norm_18"  [label="(1, 14, 14, 384)", style=solid];
-"864 _param_constant136" -> "866 layer_norm_18"  [label="(384,)", style=solid];
-"865 _param_constant137" -> "866 layer_norm_18"  [label="(384,)", style=solid];
-"866 layer_norm_18" -> "867 add_27"  [label="(1, 14, 14, 384)", style=solid];
-"867 add_27" -> "869 pad_10"  [label="(1, 14, 14, 384)", style=solid];
-"867 add_27" -> "932 add_29"  [label="(1, 14, 14, 384)", style=solid];
-"868 _frozen_param68" -> "904 add_28"  [label="(1, 12, 64, 64)", style=solid];
-"869 pad_10" -> "870 view_46"  [label="(1, 16, 16, 384)", style=solid];
-"870 view_46" -> "871 permute_38"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"871 permute_38" -> "872 reshape_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"872 reshape_36" -> "873 reshape_36_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"873 reshape_36_0_0_nncf_smooth_quant_0" -> "874 quantize_per_tensor_default_51"  [label="(4, 64, 384)", style=solid];
-"874 quantize_per_tensor_default_51" -> "875 dequantize_per_tensor_default_51"  [label="(4, 64, 384)", style=solid];
-"875 dequantize_per_tensor_default_51" -> "882 linear_52"  [label="(4, 64, 384)", style=solid];
-"876 _frozen_param69_0_0" -> "882 linear_52"  [label="(1152,)", style=solid];
-"877 scale_updated_constant35" -> "879 mul_tensor_35"  [label="(1152, 1)", style=solid];
-"878 compressed_weight_updated_constant35" -> "879 mul_tensor_35"  [label="(1152, 384)", style=solid];
-"879 mul_tensor_35" -> "881 sub_tensor_35"  [label="(1152, 384)", style=solid];
-"880 zero_point_updated_constant35" -> "881 sub_tensor_35"  [label="(1152, 1)", style=solid];
-"881 sub_tensor_35" -> "882 linear_52"  [label="(1152, 384)", style=solid];
-"882 linear_52" -> "883 reshape_37"  [label="(4, 64, 1152)", style=solid];
-"883 reshape_37" -> "884 permute_39"  [label="(4, 64, 3, 12, 32)", style=solid];
-"884 permute_39" -> "885 select_24"  [label="(3, 4, 12, 64, 32)", style=solid];
-"884 permute_39" -> "886 select_25"  [label="(3, 4, 12, 64, 32)", style=solid];
-"884 permute_39" -> "887 select_26"  [label="(3, 4, 12, 64, 32)", style=solid];
-"885 select_24" -> "888 linalg_vector_norm_16"  [label="(4, 12, 64, 32)", style=solid];
-"885 select_24" -> "890 expand_as_16"  [label="(4, 12, 64, 32)", style=solid];
-"885 select_24" -> "891 div_16"  [label="(4, 12, 64, 32)", style=solid];
-"886 select_25" -> "894 linalg_vector_norm_17"  [label="(4, 12, 64, 32)", style=solid];
-"886 select_25" -> "896 expand_as_17"  [label="(4, 12, 64, 32)", style=solid];
-"886 select_25" -> "897 div_17"  [label="(4, 12, 64, 32)", style=solid];
-"887 select_26" -> "907 matmul_17"  [label="(4, 12, 64, 32)", style=solid];
-"888 linalg_vector_norm_16" -> "889 clamp_min_16"  [label="(4, 12, 64, 1)", style=solid];
-"889 clamp_min_16" -> "890 expand_as_16"  [label="(4, 12, 64, 1)", style=solid];
-"890 expand_as_16" -> "891 div_16"  [label="(4, 12, 64, 32)", style=solid];
-"891 div_16" -> "892 quantize_per_tensor_default_52"  [label="(4, 12, 64, 32)", style=solid];
-"892 quantize_per_tensor_default_52" -> "893 dequantize_per_tensor_default_52"  [label="(4, 12, 64, 32)", style=solid];
-"893 dequantize_per_tensor_default_52" -> "901 matmul_16"  [label="(4, 12, 64, 32)", style=solid];
-"894 linalg_vector_norm_17" -> "895 clamp_min_17"  [label="(4, 12, 64, 1)", style=solid];
-"895 clamp_min_17" -> "896 expand_as_17"  [label="(4, 12, 64, 1)", style=solid];
-"896 expand_as_17" -> "897 div_17"  [label="(4, 12, 64, 32)", style=solid];
-"897 div_17" -> "898 quantize_per_tensor_default_53"  [label="(4, 12, 64, 32)", style=solid];
-"898 quantize_per_tensor_default_53" -> "899 dequantize_per_tensor_default_53"  [label="(4, 12, 64, 32)", style=solid];
-"899 dequantize_per_tensor_default_53" -> "900 transpose_16"  [label="(4, 12, 64, 32)", style=solid];
-"900 transpose_16" -> "901 matmul_16"  [label="(4, 12, 32, 64)", style=solid];
-"901 matmul_16" -> "903 mul_17"  [label="(4, 12, 64, 64)", style=solid];
-"902 _frozen_param71" -> "903 mul_17"  [label="(12, 1, 1)", style=solid];
-"903 mul_17" -> "904 add_28"  [label="(4, 12, 64, 64)", style=solid];
-"904 add_28" -> "905 softmax_8"  [label="(4, 12, 64, 64)", style=solid];
-"905 softmax_8" -> "906 dropout_32"  [label="(4, 12, 64, 64)", style=solid];
-"906 dropout_32" -> "907 matmul_17"  [label="(4, 12, 64, 64)", style=solid];
-"907 matmul_17" -> "908 transpose_17"  [label="(4, 12, 64, 32)", style=solid];
-"908 transpose_17" -> "909 reshape_38"  [label="(4, 64, 12, 32)", style=solid];
-"909 reshape_38" -> "910 reshape_38_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"910 reshape_38_0_0_nncf_smooth_quant_0" -> "911 quantize_per_tensor_default_54"  [label="(4, 64, 384)", style=solid];
-"911 quantize_per_tensor_default_54" -> "912 dequantize_per_tensor_default_54"  [label="(4, 64, 384)", style=solid];
-"912 dequantize_per_tensor_default_54" -> "919 linear_53"  [label="(4, 64, 384)", style=solid];
-"913 _param_constant145_0_0" -> "919 linear_53"  [label="(384,)", style=solid];
-"914 scale_updated_constant36" -> "916 mul_tensor_36"  [label="(384, 1)", style=solid];
-"915 compressed_weight_updated_constant36" -> "916 mul_tensor_36"  [label="(384, 384)", style=solid];
-"916 mul_tensor_36" -> "918 sub_tensor_36"  [label="(384, 384)", style=solid];
-"917 zero_point_updated_constant36" -> "918 sub_tensor_36"  [label="(384, 1)", style=solid];
-"918 sub_tensor_36" -> "919 linear_53"  [label="(384, 384)", style=solid];
-"919 linear_53" -> "920 dropout_33"  [label="(4, 64, 384)", style=solid];
-"920 dropout_33" -> "921 view_47"  [label="(4, 64, 384)", style=solid];
-"921 view_47" -> "922 permute_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"922 permute_40" -> "923 reshape_39"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"923 reshape_39" -> "924 slice_134"  [label="(1, 16, 16, 384)", style=solid];
-"924 slice_134" -> "925 slice_135"  [label="(1, 16, 16, 384)", style=solid];
-"925 slice_135" -> "926 slice_136"  [label="(1, 14, 16, 384)", style=solid];
-"926 slice_136" -> "927 slice_137"  [label="(1, 14, 14, 384)", style=solid];
-"927 slice_137" -> "928 contiguous_15"  [label="(1, 14, 14, 384)", style=solid];
-"928 contiguous_15" -> "931 layer_norm_19"  [label="(1, 14, 14, 384)", style=solid];
-"929 _param_constant146" -> "931 layer_norm_19"  [label="(384,)", style=solid];
-"930 _param_constant147" -> "931 layer_norm_19"  [label="(384,)", style=solid];
-"931 layer_norm_19" -> "932 add_29"  [label="(1, 14, 14, 384)", style=solid];
-"932 add_29" -> "933 add_29_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"932 add_29" -> "959 add_30"  [label="(1, 14, 14, 384)", style=solid];
-"933 add_29_0_0_nncf_smooth_quant_0" -> "934 quantize_per_tensor_default_55"  [label="(1, 14, 14, 384)", style=solid];
-"934 quantize_per_tensor_default_55" -> "935 dequantize_per_tensor_default_55"  [label="(1, 14, 14, 384)", style=solid];
-"935 dequantize_per_tensor_default_55" -> "942 linear_54"  [label="(1, 14, 14, 384)", style=solid];
-"936 _param_constant149_0_0" -> "942 linear_54"  [label="(1536,)", style=solid];
-"937 scale_updated_constant37" -> "939 mul_tensor_37"  [label="(1536, 1)", style=solid];
-"938 compressed_weight_updated_constant37" -> "939 mul_tensor_37"  [label="(1536, 384)", style=solid];
-"939 mul_tensor_37" -> "941 sub_tensor_37"  [label="(1536, 384)", style=solid];
-"940 zero_point_updated_constant37" -> "941 sub_tensor_37"  [label="(1536, 1)", style=solid];
-"941 sub_tensor_37" -> "942 linear_54"  [label="(1536, 384)", style=solid];
-"942 linear_54" -> "943 gelu_8"  [label="(1, 14, 14, 1536)", style=solid];
-"943 gelu_8" -> "944 dropout_34"  [label="(1, 14, 14, 1536)", style=solid];
-"944 dropout_34" -> "945 dropout_34_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"945 dropout_34_0_0_nncf_smooth_quant_0" -> "946 quantize_per_tensor_default_56"  [label="(1, 14, 14, 1536)", style=solid];
-"946 quantize_per_tensor_default_56" -> "947 dequantize_per_tensor_default_56"  [label="(1, 14, 14, 1536)", style=solid];
-"947 dequantize_per_tensor_default_56" -> "954 linear_55"  [label="(1, 14, 14, 1536)", style=solid];
-"948 _param_constant151_0_0" -> "954 linear_55"  [label="(384,)", style=solid];
-"949 scale_updated_constant38" -> "951 mul_tensor_38"  [label="(384, 1)", style=solid];
-"950 compressed_weight_updated_constant38" -> "951 mul_tensor_38"  [label="(384, 1536)", style=solid];
-"951 mul_tensor_38" -> "953 sub_tensor_38"  [label="(384, 1536)", style=solid];
-"952 zero_point_updated_constant38" -> "953 sub_tensor_38"  [label="(384, 1)", style=solid];
-"953 sub_tensor_38" -> "954 linear_55"  [label="(384, 1536)", style=solid];
-"954 linear_55" -> "955 dropout_35"  [label="(1, 14, 14, 384)", style=solid];
-"955 dropout_35" -> "958 layer_norm_20"  [label="(1, 14, 14, 384)", style=solid];
-"956 _param_constant152" -> "958 layer_norm_20"  [label="(384,)", style=solid];
-"957 _param_constant153" -> "958 layer_norm_20"  [label="(384,)", style=solid];
-"958 layer_norm_20" -> "959 add_30"  [label="(1, 14, 14, 384)", style=solid];
-"959 add_30" -> "961 pad_11"  [label="(1, 14, 14, 384)", style=solid];
-"959 add_30" -> "1042 add_33"  [label="(1, 14, 14, 384)", style=solid];
-"960 _frozen_param72" -> "997 add_31"  [label="(1, 12, 64, 64)", style=solid];
-"961 pad_11" -> "962 roll_8"  [label="(1, 16, 16, 384)", style=solid];
-"962 roll_8" -> "963 view_50"  [label="(1, 16, 16, 384)", style=solid];
-"963 view_50" -> "964 permute_42"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"964 permute_42" -> "965 reshape_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"965 reshape_40" -> "966 reshape_40_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"965 reshape_40" -> "998 new_zeros_4"  [label="(4, 64, 384)", style=solid];
-"966 reshape_40_0_0_nncf_smooth_quant_0" -> "967 quantize_per_tensor_default_57"  [label="(4, 64, 384)", style=solid];
-"967 quantize_per_tensor_default_57" -> "968 dequantize_per_tensor_default_57"  [label="(4, 64, 384)", style=solid];
-"968 dequantize_per_tensor_default_57" -> "975 linear_58"  [label="(4, 64, 384)", style=solid];
-"969 _frozen_param73_0_0" -> "975 linear_58"  [label="(1152,)", style=solid];
-"970 scale_updated_constant39" -> "972 mul_tensor_39"  [label="(1152, 1)", style=solid];
-"971 compressed_weight_updated_constant39" -> "972 mul_tensor_39"  [label="(1152, 384)", style=solid];
-"972 mul_tensor_39" -> "974 sub_tensor_39"  [label="(1152, 384)", style=solid];
-"973 zero_point_updated_constant39" -> "974 sub_tensor_39"  [label="(1152, 1)", style=solid];
-"974 sub_tensor_39" -> "975 linear_58"  [label="(1152, 384)", style=solid];
-"975 linear_58" -> "976 reshape_41"  [label="(4, 64, 1152)", style=solid];
-"976 reshape_41" -> "977 permute_43"  [label="(4, 64, 3, 12, 32)", style=solid];
-"977 permute_43" -> "978 select_27"  [label="(3, 4, 12, 64, 32)", style=solid];
-"977 permute_43" -> "979 select_28"  [label="(3, 4, 12, 64, 32)", style=solid];
-"977 permute_43" -> "980 select_29"  [label="(3, 4, 12, 64, 32)", style=solid];
-"978 select_27" -> "981 linalg_vector_norm_18"  [label="(4, 12, 64, 32)", style=solid];
-"978 select_27" -> "983 expand_as_18"  [label="(4, 12, 64, 32)", style=solid];
-"978 select_27" -> "984 div_18"  [label="(4, 12, 64, 32)", style=solid];
-"979 select_28" -> "987 linalg_vector_norm_19"  [label="(4, 12, 64, 32)", style=solid];
-"979 select_28" -> "989 expand_as_19"  [label="(4, 12, 64, 32)", style=solid];
-"979 select_28" -> "990 div_19"  [label="(4, 12, 64, 32)", style=solid];
-"980 select_29" -> "1016 matmul_19"  [label="(4, 12, 64, 32)", style=solid];
-"981 linalg_vector_norm_18" -> "982 clamp_min_18"  [label="(4, 12, 64, 1)", style=solid];
-"982 clamp_min_18" -> "983 expand_as_18"  [label="(4, 12, 64, 1)", style=solid];
-"983 expand_as_18" -> "984 div_18"  [label="(4, 12, 64, 32)", style=solid];
-"984 div_18" -> "985 quantize_per_tensor_default_58"  [label="(4, 12, 64, 32)", style=solid];
-"985 quantize_per_tensor_default_58" -> "986 dequantize_per_tensor_default_58"  [label="(4, 12, 64, 32)", style=solid];
-"986 dequantize_per_tensor_default_58" -> "994 matmul_18"  [label="(4, 12, 64, 32)", style=solid];
-"987 linalg_vector_norm_19" -> "988 clamp_min_19"  [label="(4, 12, 64, 1)", style=solid];
-"988 clamp_min_19" -> "989 expand_as_19"  [label="(4, 12, 64, 1)", style=solid];
-"989 expand_as_19" -> "990 div_19"  [label="(4, 12, 64, 32)", style=solid];
-"990 div_19" -> "991 quantize_per_tensor_default_59"  [label="(4, 12, 64, 32)", style=solid];
-"991 quantize_per_tensor_default_59" -> "992 dequantize_per_tensor_default_59"  [label="(4, 12, 64, 32)", style=solid];
-"992 dequantize_per_tensor_default_59" -> "993 transpose_18"  [label="(4, 12, 64, 32)", style=solid];
-"993 transpose_18" -> "994 matmul_18"  [label="(4, 12, 32, 64)", style=solid];
-"994 matmul_18" -> "996 mul_19"  [label="(4, 12, 64, 64)", style=solid];
-"995 _frozen_param75" -> "996 mul_19"  [label="(12, 1, 1)", style=solid];
-"996 mul_19" -> "997 add_31"  [label="(4, 12, 64, 64)", style=solid];
-"997 add_31" -> "1009 view_52"  [label="(4, 12, 64, 64)", style=solid];
-"998 new_zeros_4" -> "999 view_51"  [label="(16, 16)", style=solid];
-"999 view_51" -> "1000 permute_44"  [label="(2, 8, 2, 8)", style=solid];
-"1000 permute_44" -> "1001 reshape_42"  [label="(2, 2, 8, 8)", style=solid];
-"1001 reshape_42" -> "1002 unsqueeze_26"  [label="(4, 64)", style=solid];
-"1001 reshape_42" -> "1003 unsqueeze_27"  [label="(4, 64)", style=solid];
-"1002 unsqueeze_26" -> "1004 sub_4"  [label="(4, 1, 64)", style=solid];
-"1003 unsqueeze_27" -> "1004 sub_4"  [label="(4, 64, 1)", style=solid];
-"1004 sub_4" -> "1005 ne_4"  [label="(4, 64, 64)", style=solid];
-"1004 sub_4" -> "1006 masked_fill_8"  [label="(4, 64, 64)", style=solid];
-"1004 sub_4" -> "1007 eq_4"  [label="(4, 64, 64)", style=solid];
-"1005 ne_4" -> "1006 masked_fill_8"  [label="(4, 64, 64)", style=solid];
-"1006 masked_fill_8" -> "1008 masked_fill_9"  [label="(4, 64, 64)", style=solid];
-"1007 eq_4" -> "1008 masked_fill_9"  [label="(4, 64, 64)", style=solid];
-"1008 masked_fill_9" -> "1010 unsqueeze_28"  [label="(4, 64, 64)", style=solid];
-"1009 view_52" -> "1012 add_32"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1010 unsqueeze_28" -> "1011 unsqueeze_29"  [label="(4, 1, 64, 64)", style=solid];
-"1011 unsqueeze_29" -> "1012 add_32"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1012 add_32" -> "1013 view_53"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1013 view_53" -> "1014 softmax_9"  [label="(4, 12, 64, 64)", style=solid];
-"1014 softmax_9" -> "1015 dropout_36"  [label="(4, 12, 64, 64)", style=solid];
-"1015 dropout_36" -> "1016 matmul_19"  [label="(4, 12, 64, 64)", style=solid];
-"1016 matmul_19" -> "1017 transpose_19"  [label="(4, 12, 64, 32)", style=solid];
-"1017 transpose_19" -> "1018 reshape_43"  [label="(4, 64, 12, 32)", style=solid];
-"1018 reshape_43" -> "1019 reshape_43_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1019 reshape_43_0_0_nncf_smooth_quant_0" -> "1020 quantize_per_tensor_default_60"  [label="(4, 64, 384)", style=solid];
-"1020 quantize_per_tensor_default_60" -> "1021 dequantize_per_tensor_default_60"  [label="(4, 64, 384)", style=solid];
-"1021 dequantize_per_tensor_default_60" -> "1028 linear_59"  [label="(4, 64, 384)", style=solid];
-"1022 _param_constant161_0_0" -> "1028 linear_59"  [label="(384,)", style=solid];
-"1023 scale_updated_constant40" -> "1025 mul_tensor_40"  [label="(384, 1)", style=solid];
-"1024 compressed_weight_updated_constant40" -> "1025 mul_tensor_40"  [label="(384, 384)", style=solid];
-"1025 mul_tensor_40" -> "1027 sub_tensor_40"  [label="(384, 384)", style=solid];
-"1026 zero_point_updated_constant40" -> "1027 sub_tensor_40"  [label="(384, 1)", style=solid];
-"1027 sub_tensor_40" -> "1028 linear_59"  [label="(384, 384)", style=solid];
-"1028 linear_59" -> "1029 dropout_37"  [label="(4, 64, 384)", style=solid];
-"1029 dropout_37" -> "1030 view_54"  [label="(4, 64, 384)", style=solid];
-"1030 view_54" -> "1031 permute_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1031 permute_45" -> "1032 reshape_44"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1032 reshape_44" -> "1033 roll_9"  [label="(1, 16, 16, 384)", style=solid];
-"1033 roll_9" -> "1034 slice_157"  [label="(1, 16, 16, 384)", style=solid];
-"1034 slice_157" -> "1035 slice_158"  [label="(1, 16, 16, 384)", style=solid];
-"1035 slice_158" -> "1036 slice_159"  [label="(1, 14, 16, 384)", style=solid];
-"1036 slice_159" -> "1037 slice_160"  [label="(1, 14, 14, 384)", style=solid];
-"1037 slice_160" -> "1038 contiguous_17"  [label="(1, 14, 14, 384)", style=solid];
-"1038 contiguous_17" -> "1041 layer_norm_21"  [label="(1, 14, 14, 384)", style=solid];
-"1039 _param_constant162" -> "1041 layer_norm_21"  [label="(384,)", style=solid];
-"1040 _param_constant163" -> "1041 layer_norm_21"  [label="(384,)", style=solid];
-"1041 layer_norm_21" -> "1042 add_33"  [label="(1, 14, 14, 384)", style=solid];
-"1042 add_33" -> "1043 add_33_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1042 add_33" -> "1069 add_34"  [label="(1, 14, 14, 384)", style=solid];
-"1043 add_33_0_0_nncf_smooth_quant_0" -> "1044 quantize_per_tensor_default_61"  [label="(1, 14, 14, 384)", style=solid];
-"1044 quantize_per_tensor_default_61" -> "1045 dequantize_per_tensor_default_61"  [label="(1, 14, 14, 384)", style=solid];
-"1045 dequantize_per_tensor_default_61" -> "1052 linear_60"  [label="(1, 14, 14, 384)", style=solid];
-"1046 _param_constant165_0_0" -> "1052 linear_60"  [label="(1536,)", style=solid];
-"1047 scale_updated_constant41" -> "1049 mul_tensor_41"  [label="(1536, 1)", style=solid];
-"1048 compressed_weight_updated_constant41" -> "1049 mul_tensor_41"  [label="(1536, 384)", style=solid];
-"1049 mul_tensor_41" -> "1051 sub_tensor_41"  [label="(1536, 384)", style=solid];
-"1050 zero_point_updated_constant41" -> "1051 sub_tensor_41"  [label="(1536, 1)", style=solid];
-"1051 sub_tensor_41" -> "1052 linear_60"  [label="(1536, 384)", style=solid];
-"1052 linear_60" -> "1053 gelu_9"  [label="(1, 14, 14, 1536)", style=solid];
-"1053 gelu_9" -> "1054 dropout_38"  [label="(1, 14, 14, 1536)", style=solid];
-"1054 dropout_38" -> "1055 dropout_38_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1055 dropout_38_0_0_nncf_smooth_quant_0" -> "1056 quantize_per_tensor_default_62"  [label="(1, 14, 14, 1536)", style=solid];
-"1056 quantize_per_tensor_default_62" -> "1057 dequantize_per_tensor_default_62"  [label="(1, 14, 14, 1536)", style=solid];
-"1057 dequantize_per_tensor_default_62" -> "1064 linear_61"  [label="(1, 14, 14, 1536)", style=solid];
-"1058 _param_constant167_0_0" -> "1064 linear_61"  [label="(384,)", style=solid];
-"1059 scale_updated_constant42" -> "1061 mul_tensor_42"  [label="(384, 1)", style=solid];
-"1060 compressed_weight_updated_constant42" -> "1061 mul_tensor_42"  [label="(384, 1536)", style=solid];
-"1061 mul_tensor_42" -> "1063 sub_tensor_42"  [label="(384, 1536)", style=solid];
-"1062 zero_point_updated_constant42" -> "1063 sub_tensor_42"  [label="(384, 1)", style=solid];
-"1063 sub_tensor_42" -> "1064 linear_61"  [label="(384, 1536)", style=solid];
-"1064 linear_61" -> "1065 dropout_39"  [label="(1, 14, 14, 384)", style=solid];
-"1065 dropout_39" -> "1068 layer_norm_22"  [label="(1, 14, 14, 384)", style=solid];
-"1066 _param_constant168" -> "1068 layer_norm_22"  [label="(384,)", style=solid];
-"1067 _param_constant169" -> "1068 layer_norm_22"  [label="(384,)", style=solid];
-"1068 layer_norm_22" -> "1069 add_34"  [label="(1, 14, 14, 384)", style=solid];
-"1069 add_34" -> "1071 pad_12"  [label="(1, 14, 14, 384)", style=solid];
-"1069 add_34" -> "1134 add_36"  [label="(1, 14, 14, 384)", style=solid];
-"1070 _frozen_param85" -> "1106 add_35"  [label="(1, 12, 64, 64)", style=solid];
-"1071 pad_12" -> "1072 view_57"  [label="(1, 16, 16, 384)", style=solid];
-"1072 view_57" -> "1073 permute_47"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1073 permute_47" -> "1074 reshape_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1074 reshape_45" -> "1075 reshape_45_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1075 reshape_45_0_0_nncf_smooth_quant_0" -> "1076 quantize_per_tensor_default_63"  [label="(4, 64, 384)", style=solid];
-"1076 quantize_per_tensor_default_63" -> "1077 dequantize_per_tensor_default_63"  [label="(4, 64, 384)", style=solid];
-"1077 dequantize_per_tensor_default_63" -> "1084 linear_64"  [label="(4, 64, 384)", style=solid];
-"1078 _frozen_param86_0_0" -> "1084 linear_64"  [label="(1152,)", style=solid];
-"1079 scale_updated_constant43" -> "1081 mul_tensor_43"  [label="(1152, 1)", style=solid];
-"1080 compressed_weight_updated_constant43" -> "1081 mul_tensor_43"  [label="(1152, 384)", style=solid];
-"1081 mul_tensor_43" -> "1083 sub_tensor_43"  [label="(1152, 384)", style=solid];
-"1082 zero_point_updated_constant43" -> "1083 sub_tensor_43"  [label="(1152, 1)", style=solid];
-"1083 sub_tensor_43" -> "1084 linear_64"  [label="(1152, 384)", style=solid];
-"1084 linear_64" -> "1085 reshape_46"  [label="(4, 64, 1152)", style=solid];
-"1085 reshape_46" -> "1086 permute_48"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1086 permute_48" -> "1087 select_30"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1086 permute_48" -> "1088 select_31"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1086 permute_48" -> "1089 select_32"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1087 select_30" -> "1090 linalg_vector_norm_20"  [label="(4, 12, 64, 32)", style=solid];
-"1087 select_30" -> "1092 expand_as_20"  [label="(4, 12, 64, 32)", style=solid];
-"1087 select_30" -> "1093 div_20"  [label="(4, 12, 64, 32)", style=solid];
-"1088 select_31" -> "1096 linalg_vector_norm_21"  [label="(4, 12, 64, 32)", style=solid];
-"1088 select_31" -> "1098 expand_as_21"  [label="(4, 12, 64, 32)", style=solid];
-"1088 select_31" -> "1099 div_21"  [label="(4, 12, 64, 32)", style=solid];
-"1089 select_32" -> "1109 matmul_21"  [label="(4, 12, 64, 32)", style=solid];
-"1090 linalg_vector_norm_20" -> "1091 clamp_min_20"  [label="(4, 12, 64, 1)", style=solid];
-"1091 clamp_min_20" -> "1092 expand_as_20"  [label="(4, 12, 64, 1)", style=solid];
-"1092 expand_as_20" -> "1093 div_20"  [label="(4, 12, 64, 32)", style=solid];
-"1093 div_20" -> "1094 quantize_per_tensor_default_64"  [label="(4, 12, 64, 32)", style=solid];
-"1094 quantize_per_tensor_default_64" -> "1095 dequantize_per_tensor_default_64"  [label="(4, 12, 64, 32)", style=solid];
-"1095 dequantize_per_tensor_default_64" -> "1103 matmul_20"  [label="(4, 12, 64, 32)", style=solid];
-"1096 linalg_vector_norm_21" -> "1097 clamp_min_21"  [label="(4, 12, 64, 1)", style=solid];
-"1097 clamp_min_21" -> "1098 expand_as_21"  [label="(4, 12, 64, 1)", style=solid];
-"1098 expand_as_21" -> "1099 div_21"  [label="(4, 12, 64, 32)", style=solid];
-"1099 div_21" -> "1100 quantize_per_tensor_default_65"  [label="(4, 12, 64, 32)", style=solid];
-"1100 quantize_per_tensor_default_65" -> "1101 dequantize_per_tensor_default_65"  [label="(4, 12, 64, 32)", style=solid];
-"1101 dequantize_per_tensor_default_65" -> "1102 transpose_20"  [label="(4, 12, 64, 32)", style=solid];
-"1102 transpose_20" -> "1103 matmul_20"  [label="(4, 12, 32, 64)", style=solid];
-"1103 matmul_20" -> "1105 mul_21"  [label="(4, 12, 64, 64)", style=solid];
-"1104 _frozen_param88" -> "1105 mul_21"  [label="(12, 1, 1)", style=solid];
-"1105 mul_21" -> "1106 add_35"  [label="(4, 12, 64, 64)", style=solid];
-"1106 add_35" -> "1107 softmax_10"  [label="(4, 12, 64, 64)", style=solid];
-"1107 softmax_10" -> "1108 dropout_40"  [label="(4, 12, 64, 64)", style=solid];
-"1108 dropout_40" -> "1109 matmul_21"  [label="(4, 12, 64, 64)", style=solid];
-"1109 matmul_21" -> "1110 transpose_21"  [label="(4, 12, 64, 32)", style=solid];
-"1110 transpose_21" -> "1111 reshape_47"  [label="(4, 64, 12, 32)", style=solid];
-"1111 reshape_47" -> "1112 reshape_47_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1112 reshape_47_0_0_nncf_smooth_quant_0" -> "1113 quantize_per_tensor_default_66"  [label="(4, 64, 384)", style=solid];
-"1113 quantize_per_tensor_default_66" -> "1114 dequantize_per_tensor_default_66"  [label="(4, 64, 384)", style=solid];
-"1114 dequantize_per_tensor_default_66" -> "1121 linear_65"  [label="(4, 64, 384)", style=solid];
-"1115 _param_constant177_0_0" -> "1121 linear_65"  [label="(384,)", style=solid];
-"1116 scale_updated_constant44" -> "1118 mul_tensor_44"  [label="(384, 1)", style=solid];
-"1117 compressed_weight_updated_constant44" -> "1118 mul_tensor_44"  [label="(384, 384)", style=solid];
-"1118 mul_tensor_44" -> "1120 sub_tensor_44"  [label="(384, 384)", style=solid];
-"1119 zero_point_updated_constant44" -> "1120 sub_tensor_44"  [label="(384, 1)", style=solid];
-"1120 sub_tensor_44" -> "1121 linear_65"  [label="(384, 384)", style=solid];
-"1121 linear_65" -> "1122 dropout_41"  [label="(4, 64, 384)", style=solid];
-"1122 dropout_41" -> "1123 view_58"  [label="(4, 64, 384)", style=solid];
-"1123 view_58" -> "1124 permute_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1124 permute_49" -> "1125 reshape_48"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1125 reshape_48" -> "1126 slice_162"  [label="(1, 16, 16, 384)", style=solid];
-"1126 slice_162" -> "1127 slice_163"  [label="(1, 16, 16, 384)", style=solid];
-"1127 slice_163" -> "1128 slice_164"  [label="(1, 14, 16, 384)", style=solid];
-"1128 slice_164" -> "1129 slice_165"  [label="(1, 14, 14, 384)", style=solid];
-"1129 slice_165" -> "1130 contiguous_19"  [label="(1, 14, 14, 384)", style=solid];
-"1130 contiguous_19" -> "1133 layer_norm_23"  [label="(1, 14, 14, 384)", style=solid];
-"1131 _param_constant178" -> "1133 layer_norm_23"  [label="(384,)", style=solid];
-"1132 _param_constant179" -> "1133 layer_norm_23"  [label="(384,)", style=solid];
-"1133 layer_norm_23" -> "1134 add_36"  [label="(1, 14, 14, 384)", style=solid];
-"1134 add_36" -> "1135 add_36_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1134 add_36" -> "1161 add_37"  [label="(1, 14, 14, 384)", style=solid];
-"1135 add_36_0_0_nncf_smooth_quant_0" -> "1136 quantize_per_tensor_default_67"  [label="(1, 14, 14, 384)", style=solid];
-"1136 quantize_per_tensor_default_67" -> "1137 dequantize_per_tensor_default_67"  [label="(1, 14, 14, 384)", style=solid];
-"1137 dequantize_per_tensor_default_67" -> "1144 linear_66"  [label="(1, 14, 14, 384)", style=solid];
-"1138 _param_constant181_0_0" -> "1144 linear_66"  [label="(1536,)", style=solid];
-"1139 scale_updated_constant45" -> "1141 mul_tensor_45"  [label="(1536, 1)", style=solid];
-"1140 compressed_weight_updated_constant45" -> "1141 mul_tensor_45"  [label="(1536, 384)", style=solid];
-"1141 mul_tensor_45" -> "1143 sub_tensor_45"  [label="(1536, 384)", style=solid];
-"1142 zero_point_updated_constant45" -> "1143 sub_tensor_45"  [label="(1536, 1)", style=solid];
-"1143 sub_tensor_45" -> "1144 linear_66"  [label="(1536, 384)", style=solid];
-"1144 linear_66" -> "1145 gelu_10"  [label="(1, 14, 14, 1536)", style=solid];
-"1145 gelu_10" -> "1146 dropout_42"  [label="(1, 14, 14, 1536)", style=solid];
-"1146 dropout_42" -> "1147 dropout_42_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1147 dropout_42_0_0_nncf_smooth_quant_0" -> "1148 quantize_per_tensor_default_68"  [label="(1, 14, 14, 1536)", style=solid];
-"1148 quantize_per_tensor_default_68" -> "1149 dequantize_per_tensor_default_68"  [label="(1, 14, 14, 1536)", style=solid];
-"1149 dequantize_per_tensor_default_68" -> "1156 linear_67"  [label="(1, 14, 14, 1536)", style=solid];
-"1150 _param_constant183_0_0" -> "1156 linear_67"  [label="(384,)", style=solid];
-"1151 scale_updated_constant46" -> "1153 mul_tensor_46"  [label="(384, 1)", style=solid];
-"1152 compressed_weight_updated_constant46" -> "1153 mul_tensor_46"  [label="(384, 1536)", style=solid];
-"1153 mul_tensor_46" -> "1155 sub_tensor_46"  [label="(384, 1536)", style=solid];
-"1154 zero_point_updated_constant46" -> "1155 sub_tensor_46"  [label="(384, 1)", style=solid];
-"1155 sub_tensor_46" -> "1156 linear_67"  [label="(384, 1536)", style=solid];
-"1156 linear_67" -> "1157 dropout_43"  [label="(1, 14, 14, 384)", style=solid];
-"1157 dropout_43" -> "1160 layer_norm_24"  [label="(1, 14, 14, 384)", style=solid];
-"1158 _param_constant184" -> "1160 layer_norm_24"  [label="(384,)", style=solid];
-"1159 _param_constant185" -> "1160 layer_norm_24"  [label="(384,)", style=solid];
-"1160 layer_norm_24" -> "1161 add_37"  [label="(1, 14, 14, 384)", style=solid];
-"1161 add_37" -> "1163 pad_13"  [label="(1, 14, 14, 384)", style=solid];
-"1161 add_37" -> "1244 add_40"  [label="(1, 14, 14, 384)", style=solid];
-"1162 _frozen_param89" -> "1199 add_38"  [label="(1, 12, 64, 64)", style=solid];
-"1163 pad_13" -> "1164 roll_10"  [label="(1, 16, 16, 384)", style=solid];
-"1164 roll_10" -> "1165 view_61"  [label="(1, 16, 16, 384)", style=solid];
-"1165 view_61" -> "1166 permute_51"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1166 permute_51" -> "1167 reshape_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1167 reshape_49" -> "1168 reshape_49_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1167 reshape_49" -> "1200 new_zeros_5"  [label="(4, 64, 384)", style=solid];
-"1168 reshape_49_0_0_nncf_smooth_quant_0" -> "1169 quantize_per_tensor_default_69"  [label="(4, 64, 384)", style=solid];
-"1169 quantize_per_tensor_default_69" -> "1170 dequantize_per_tensor_default_69"  [label="(4, 64, 384)", style=solid];
-"1170 dequantize_per_tensor_default_69" -> "1177 linear_70"  [label="(4, 64, 384)", style=solid];
-"1171 _frozen_param90_0_0" -> "1177 linear_70"  [label="(1152,)", style=solid];
-"1172 scale_updated_constant47" -> "1174 mul_tensor_47"  [label="(1152, 1)", style=solid];
-"1173 compressed_weight_updated_constant47" -> "1174 mul_tensor_47"  [label="(1152, 384)", style=solid];
-"1174 mul_tensor_47" -> "1176 sub_tensor_47"  [label="(1152, 384)", style=solid];
-"1175 zero_point_updated_constant47" -> "1176 sub_tensor_47"  [label="(1152, 1)", style=solid];
-"1176 sub_tensor_47" -> "1177 linear_70"  [label="(1152, 384)", style=solid];
-"1177 linear_70" -> "1178 reshape_50"  [label="(4, 64, 1152)", style=solid];
-"1178 reshape_50" -> "1179 permute_52"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1179 permute_52" -> "1180 select_33"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1179 permute_52" -> "1181 select_34"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1179 permute_52" -> "1182 select_35"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1180 select_33" -> "1183 linalg_vector_norm_22"  [label="(4, 12, 64, 32)", style=solid];
-"1180 select_33" -> "1185 expand_as_22"  [label="(4, 12, 64, 32)", style=solid];
-"1180 select_33" -> "1186 div_22"  [label="(4, 12, 64, 32)", style=solid];
-"1181 select_34" -> "1189 linalg_vector_norm_23"  [label="(4, 12, 64, 32)", style=solid];
-"1181 select_34" -> "1191 expand_as_23"  [label="(4, 12, 64, 32)", style=solid];
-"1181 select_34" -> "1192 div_23"  [label="(4, 12, 64, 32)", style=solid];
-"1182 select_35" -> "1218 matmul_23"  [label="(4, 12, 64, 32)", style=solid];
-"1183 linalg_vector_norm_22" -> "1184 clamp_min_22"  [label="(4, 12, 64, 1)", style=solid];
-"1184 clamp_min_22" -> "1185 expand_as_22"  [label="(4, 12, 64, 1)", style=solid];
-"1185 expand_as_22" -> "1186 div_22"  [label="(4, 12, 64, 32)", style=solid];
-"1186 div_22" -> "1187 quantize_per_tensor_default_70"  [label="(4, 12, 64, 32)", style=solid];
-"1187 quantize_per_tensor_default_70" -> "1188 dequantize_per_tensor_default_70"  [label="(4, 12, 64, 32)", style=solid];
-"1188 dequantize_per_tensor_default_70" -> "1196 matmul_22"  [label="(4, 12, 64, 32)", style=solid];
-"1189 linalg_vector_norm_23" -> "1190 clamp_min_23"  [label="(4, 12, 64, 1)", style=solid];
-"1190 clamp_min_23" -> "1191 expand_as_23"  [label="(4, 12, 64, 1)", style=solid];
-"1191 expand_as_23" -> "1192 div_23"  [label="(4, 12, 64, 32)", style=solid];
-"1192 div_23" -> "1193 quantize_per_tensor_default_71"  [label="(4, 12, 64, 32)", style=solid];
-"1193 quantize_per_tensor_default_71" -> "1194 dequantize_per_tensor_default_71"  [label="(4, 12, 64, 32)", style=solid];
-"1194 dequantize_per_tensor_default_71" -> "1195 transpose_22"  [label="(4, 12, 64, 32)", style=solid];
-"1195 transpose_22" -> "1196 matmul_22"  [label="(4, 12, 32, 64)", style=solid];
-"1196 matmul_22" -> "1198 mul_23"  [label="(4, 12, 64, 64)", style=solid];
-"1197 _frozen_param92" -> "1198 mul_23"  [label="(12, 1, 1)", style=solid];
-"1198 mul_23" -> "1199 add_38"  [label="(4, 12, 64, 64)", style=solid];
-"1199 add_38" -> "1211 view_63"  [label="(4, 12, 64, 64)", style=solid];
-"1200 new_zeros_5" -> "1201 view_62"  [label="(16, 16)", style=solid];
-"1201 view_62" -> "1202 permute_53"  [label="(2, 8, 2, 8)", style=solid];
-"1202 permute_53" -> "1203 reshape_51"  [label="(2, 2, 8, 8)", style=solid];
-"1203 reshape_51" -> "1204 unsqueeze_32"  [label="(4, 64)", style=solid];
-"1203 reshape_51" -> "1205 unsqueeze_33"  [label="(4, 64)", style=solid];
-"1204 unsqueeze_32" -> "1206 sub_5"  [label="(4, 1, 64)", style=solid];
-"1205 unsqueeze_33" -> "1206 sub_5"  [label="(4, 64, 1)", style=solid];
-"1206 sub_5" -> "1207 ne_5"  [label="(4, 64, 64)", style=solid];
-"1206 sub_5" -> "1208 masked_fill_10"  [label="(4, 64, 64)", style=solid];
-"1206 sub_5" -> "1209 eq_5"  [label="(4, 64, 64)", style=solid];
-"1207 ne_5" -> "1208 masked_fill_10"  [label="(4, 64, 64)", style=solid];
-"1208 masked_fill_10" -> "1210 masked_fill_11"  [label="(4, 64, 64)", style=solid];
-"1209 eq_5" -> "1210 masked_fill_11"  [label="(4, 64, 64)", style=solid];
-"1210 masked_fill_11" -> "1212 unsqueeze_34"  [label="(4, 64, 64)", style=solid];
-"1211 view_63" -> "1214 add_39"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1212 unsqueeze_34" -> "1213 unsqueeze_35"  [label="(4, 1, 64, 64)", style=solid];
-"1213 unsqueeze_35" -> "1214 add_39"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1214 add_39" -> "1215 view_64"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1215 view_64" -> "1216 softmax_11"  [label="(4, 12, 64, 64)", style=solid];
-"1216 softmax_11" -> "1217 dropout_44"  [label="(4, 12, 64, 64)", style=solid];
-"1217 dropout_44" -> "1218 matmul_23"  [label="(4, 12, 64, 64)", style=solid];
-"1218 matmul_23" -> "1219 transpose_23"  [label="(4, 12, 64, 32)", style=solid];
-"1219 transpose_23" -> "1220 reshape_52"  [label="(4, 64, 12, 32)", style=solid];
-"1220 reshape_52" -> "1221 reshape_52_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1221 reshape_52_0_0_nncf_smooth_quant_0" -> "1222 quantize_per_tensor_default_72"  [label="(4, 64, 384)", style=solid];
-"1222 quantize_per_tensor_default_72" -> "1223 dequantize_per_tensor_default_72"  [label="(4, 64, 384)", style=solid];
-"1223 dequantize_per_tensor_default_72" -> "1230 linear_71"  [label="(4, 64, 384)", style=solid];
-"1224 _param_constant193_0_0" -> "1230 linear_71"  [label="(384,)", style=solid];
-"1225 scale_updated_constant48" -> "1227 mul_tensor_48"  [label="(384, 1)", style=solid];
-"1226 compressed_weight_updated_constant48" -> "1227 mul_tensor_48"  [label="(384, 384)", style=solid];
-"1227 mul_tensor_48" -> "1229 sub_tensor_48"  [label="(384, 384)", style=solid];
-"1228 zero_point_updated_constant48" -> "1229 sub_tensor_48"  [label="(384, 1)", style=solid];
-"1229 sub_tensor_48" -> "1230 linear_71"  [label="(384, 384)", style=solid];
-"1230 linear_71" -> "1231 dropout_45"  [label="(4, 64, 384)", style=solid];
-"1231 dropout_45" -> "1232 view_65"  [label="(4, 64, 384)", style=solid];
-"1232 view_65" -> "1233 permute_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1233 permute_54" -> "1234 reshape_53"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1234 reshape_53" -> "1235 roll_11"  [label="(1, 16, 16, 384)", style=solid];
-"1235 roll_11" -> "1236 slice_185"  [label="(1, 16, 16, 384)", style=solid];
-"1236 slice_185" -> "1237 slice_186"  [label="(1, 16, 16, 384)", style=solid];
-"1237 slice_186" -> "1238 slice_187"  [label="(1, 14, 16, 384)", style=solid];
-"1238 slice_187" -> "1239 slice_188"  [label="(1, 14, 14, 384)", style=solid];
-"1239 slice_188" -> "1240 contiguous_21"  [label="(1, 14, 14, 384)", style=solid];
-"1240 contiguous_21" -> "1243 layer_norm_25"  [label="(1, 14, 14, 384)", style=solid];
-"1241 _param_constant194" -> "1243 layer_norm_25"  [label="(384,)", style=solid];
-"1242 _param_constant195" -> "1243 layer_norm_25"  [label="(384,)", style=solid];
-"1243 layer_norm_25" -> "1244 add_40"  [label="(1, 14, 14, 384)", style=solid];
-"1244 add_40" -> "1245 add_40_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1244 add_40" -> "1271 add_41"  [label="(1, 14, 14, 384)", style=solid];
-"1245 add_40_0_0_nncf_smooth_quant_0" -> "1246 quantize_per_tensor_default_73"  [label="(1, 14, 14, 384)", style=solid];
-"1246 quantize_per_tensor_default_73" -> "1247 dequantize_per_tensor_default_73"  [label="(1, 14, 14, 384)", style=solid];
-"1247 dequantize_per_tensor_default_73" -> "1254 linear_72"  [label="(1, 14, 14, 384)", style=solid];
-"1248 _param_constant197_0_0" -> "1254 linear_72"  [label="(1536,)", style=solid];
-"1249 scale_updated_constant49" -> "1251 mul_tensor_49"  [label="(1536, 1)", style=solid];
-"1250 compressed_weight_updated_constant49" -> "1251 mul_tensor_49"  [label="(1536, 384)", style=solid];
-"1251 mul_tensor_49" -> "1253 sub_tensor_49"  [label="(1536, 384)", style=solid];
-"1252 zero_point_updated_constant49" -> "1253 sub_tensor_49"  [label="(1536, 1)", style=solid];
-"1253 sub_tensor_49" -> "1254 linear_72"  [label="(1536, 384)", style=solid];
-"1254 linear_72" -> "1255 gelu_11"  [label="(1, 14, 14, 1536)", style=solid];
-"1255 gelu_11" -> "1256 dropout_46"  [label="(1, 14, 14, 1536)", style=solid];
-"1256 dropout_46" -> "1257 dropout_46_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1257 dropout_46_0_0_nncf_smooth_quant_0" -> "1258 quantize_per_tensor_default_74"  [label="(1, 14, 14, 1536)", style=solid];
-"1258 quantize_per_tensor_default_74" -> "1259 dequantize_per_tensor_default_74"  [label="(1, 14, 14, 1536)", style=solid];
-"1259 dequantize_per_tensor_default_74" -> "1266 linear_73"  [label="(1, 14, 14, 1536)", style=solid];
-"1260 _param_constant199_0_0" -> "1266 linear_73"  [label="(384,)", style=solid];
-"1261 scale_updated_constant50" -> "1263 mul_tensor_50"  [label="(384, 1)", style=solid];
-"1262 compressed_weight_updated_constant50" -> "1263 mul_tensor_50"  [label="(384, 1536)", style=solid];
-"1263 mul_tensor_50" -> "1265 sub_tensor_50"  [label="(384, 1536)", style=solid];
-"1264 zero_point_updated_constant50" -> "1265 sub_tensor_50"  [label="(384, 1)", style=solid];
-"1265 sub_tensor_50" -> "1266 linear_73"  [label="(384, 1536)", style=solid];
-"1266 linear_73" -> "1267 dropout_47"  [label="(1, 14, 14, 384)", style=solid];
-"1267 dropout_47" -> "1270 layer_norm_26"  [label="(1, 14, 14, 384)", style=solid];
-"1268 _param_constant200" -> "1270 layer_norm_26"  [label="(384,)", style=solid];
-"1269 _param_constant201" -> "1270 layer_norm_26"  [label="(384,)", style=solid];
-"1270 layer_norm_26" -> "1271 add_41"  [label="(1, 14, 14, 384)", style=solid];
-"1271 add_41" -> "1273 pad_14"  [label="(1, 14, 14, 384)", style=solid];
-"1271 add_41" -> "1336 add_43"  [label="(1, 14, 14, 384)", style=solid];
-"1272 _frozen_param102" -> "1308 add_42"  [label="(1, 12, 64, 64)", style=solid];
-"1273 pad_14" -> "1274 view_68"  [label="(1, 16, 16, 384)", style=solid];
-"1274 view_68" -> "1275 permute_56"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1275 permute_56" -> "1276 reshape_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1276 reshape_54" -> "1277 reshape_54_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1277 reshape_54_0_0_nncf_smooth_quant_0" -> "1278 quantize_per_tensor_default_75"  [label="(4, 64, 384)", style=solid];
-"1278 quantize_per_tensor_default_75" -> "1279 dequantize_per_tensor_default_75"  [label="(4, 64, 384)", style=solid];
-"1279 dequantize_per_tensor_default_75" -> "1286 linear_76"  [label="(4, 64, 384)", style=solid];
-"1280 _frozen_param103_0_0" -> "1286 linear_76"  [label="(1152,)", style=solid];
-"1281 scale_updated_constant51" -> "1283 mul_tensor_51"  [label="(1152, 1)", style=solid];
-"1282 compressed_weight_updated_constant51" -> "1283 mul_tensor_51"  [label="(1152, 384)", style=solid];
-"1283 mul_tensor_51" -> "1285 sub_tensor_51"  [label="(1152, 384)", style=solid];
-"1284 zero_point_updated_constant51" -> "1285 sub_tensor_51"  [label="(1152, 1)", style=solid];
-"1285 sub_tensor_51" -> "1286 linear_76"  [label="(1152, 384)", style=solid];
-"1286 linear_76" -> "1287 reshape_55"  [label="(4, 64, 1152)", style=solid];
-"1287 reshape_55" -> "1288 permute_57"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1288 permute_57" -> "1289 select_36"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1288 permute_57" -> "1290 select_37"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1288 permute_57" -> "1291 select_38"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1289 select_36" -> "1292 linalg_vector_norm_24"  [label="(4, 12, 64, 32)", style=solid];
-"1289 select_36" -> "1294 expand_as_24"  [label="(4, 12, 64, 32)", style=solid];
-"1289 select_36" -> "1295 div_24"  [label="(4, 12, 64, 32)", style=solid];
-"1290 select_37" -> "1298 linalg_vector_norm_25"  [label="(4, 12, 64, 32)", style=solid];
-"1290 select_37" -> "1300 expand_as_25"  [label="(4, 12, 64, 32)", style=solid];
-"1290 select_37" -> "1301 div_25"  [label="(4, 12, 64, 32)", style=solid];
-"1291 select_38" -> "1311 matmul_25"  [label="(4, 12, 64, 32)", style=solid];
-"1292 linalg_vector_norm_24" -> "1293 clamp_min_24"  [label="(4, 12, 64, 1)", style=solid];
-"1293 clamp_min_24" -> "1294 expand_as_24"  [label="(4, 12, 64, 1)", style=solid];
-"1294 expand_as_24" -> "1295 div_24"  [label="(4, 12, 64, 32)", style=solid];
-"1295 div_24" -> "1296 quantize_per_tensor_default_76"  [label="(4, 12, 64, 32)", style=solid];
-"1296 quantize_per_tensor_default_76" -> "1297 dequantize_per_tensor_default_76"  [label="(4, 12, 64, 32)", style=solid];
-"1297 dequantize_per_tensor_default_76" -> "1305 matmul_24"  [label="(4, 12, 64, 32)", style=solid];
-"1298 linalg_vector_norm_25" -> "1299 clamp_min_25"  [label="(4, 12, 64, 1)", style=solid];
-"1299 clamp_min_25" -> "1300 expand_as_25"  [label="(4, 12, 64, 1)", style=solid];
-"1300 expand_as_25" -> "1301 div_25"  [label="(4, 12, 64, 32)", style=solid];
-"1301 div_25" -> "1302 quantize_per_tensor_default_77"  [label="(4, 12, 64, 32)", style=solid];
-"1302 quantize_per_tensor_default_77" -> "1303 dequantize_per_tensor_default_77"  [label="(4, 12, 64, 32)", style=solid];
-"1303 dequantize_per_tensor_default_77" -> "1304 transpose_24"  [label="(4, 12, 64, 32)", style=solid];
-"1304 transpose_24" -> "1305 matmul_24"  [label="(4, 12, 32, 64)", style=solid];
-"1305 matmul_24" -> "1307 mul_25"  [label="(4, 12, 64, 64)", style=solid];
-"1306 _frozen_param105" -> "1307 mul_25"  [label="(12, 1, 1)", style=solid];
-"1307 mul_25" -> "1308 add_42"  [label="(4, 12, 64, 64)", style=solid];
-"1308 add_42" -> "1309 softmax_12"  [label="(4, 12, 64, 64)", style=solid];
-"1309 softmax_12" -> "1310 dropout_48"  [label="(4, 12, 64, 64)", style=solid];
-"1310 dropout_48" -> "1311 matmul_25"  [label="(4, 12, 64, 64)", style=solid];
-"1311 matmul_25" -> "1312 transpose_25"  [label="(4, 12, 64, 32)", style=solid];
-"1312 transpose_25" -> "1313 reshape_56"  [label="(4, 64, 12, 32)", style=solid];
-"1313 reshape_56" -> "1314 reshape_56_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1314 reshape_56_0_0_nncf_smooth_quant_0" -> "1315 quantize_per_tensor_default_78"  [label="(4, 64, 384)", style=solid];
-"1315 quantize_per_tensor_default_78" -> "1316 dequantize_per_tensor_default_78"  [label="(4, 64, 384)", style=solid];
-"1316 dequantize_per_tensor_default_78" -> "1323 linear_77"  [label="(4, 64, 384)", style=solid];
-"1317 _param_constant209_0_0" -> "1323 linear_77"  [label="(384,)", style=solid];
-"1318 scale_updated_constant52" -> "1320 mul_tensor_52"  [label="(384, 1)", style=solid];
-"1319 compressed_weight_updated_constant52" -> "1320 mul_tensor_52"  [label="(384, 384)", style=solid];
-"1320 mul_tensor_52" -> "1322 sub_tensor_52"  [label="(384, 384)", style=solid];
-"1321 zero_point_updated_constant52" -> "1322 sub_tensor_52"  [label="(384, 1)", style=solid];
-"1322 sub_tensor_52" -> "1323 linear_77"  [label="(384, 384)", style=solid];
-"1323 linear_77" -> "1324 dropout_49"  [label="(4, 64, 384)", style=solid];
-"1324 dropout_49" -> "1325 view_69"  [label="(4, 64, 384)", style=solid];
-"1325 view_69" -> "1326 permute_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1326 permute_58" -> "1327 reshape_57"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1327 reshape_57" -> "1328 slice_190"  [label="(1, 16, 16, 384)", style=solid];
-"1328 slice_190" -> "1329 slice_191"  [label="(1, 16, 16, 384)", style=solid];
-"1329 slice_191" -> "1330 slice_192"  [label="(1, 14, 16, 384)", style=solid];
-"1330 slice_192" -> "1331 slice_193"  [label="(1, 14, 14, 384)", style=solid];
-"1331 slice_193" -> "1332 contiguous_23"  [label="(1, 14, 14, 384)", style=solid];
-"1332 contiguous_23" -> "1335 layer_norm_27"  [label="(1, 14, 14, 384)", style=solid];
-"1333 _param_constant210" -> "1335 layer_norm_27"  [label="(384,)", style=solid];
-"1334 _param_constant211" -> "1335 layer_norm_27"  [label="(384,)", style=solid];
-"1335 layer_norm_27" -> "1336 add_43"  [label="(1, 14, 14, 384)", style=solid];
-"1336 add_43" -> "1337 add_43_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1336 add_43" -> "1363 add_44"  [label="(1, 14, 14, 384)", style=solid];
-"1337 add_43_0_0_nncf_smooth_quant_0" -> "1338 quantize_per_tensor_default_79"  [label="(1, 14, 14, 384)", style=solid];
-"1338 quantize_per_tensor_default_79" -> "1339 dequantize_per_tensor_default_79"  [label="(1, 14, 14, 384)", style=solid];
-"1339 dequantize_per_tensor_default_79" -> "1346 linear_78"  [label="(1, 14, 14, 384)", style=solid];
-"1340 _param_constant213_0_0" -> "1346 linear_78"  [label="(1536,)", style=solid];
-"1341 scale_updated_constant53" -> "1343 mul_tensor_53"  [label="(1536, 1)", style=solid];
-"1342 compressed_weight_updated_constant53" -> "1343 mul_tensor_53"  [label="(1536, 384)", style=solid];
-"1343 mul_tensor_53" -> "1345 sub_tensor_53"  [label="(1536, 384)", style=solid];
-"1344 zero_point_updated_constant53" -> "1345 sub_tensor_53"  [label="(1536, 1)", style=solid];
-"1345 sub_tensor_53" -> "1346 linear_78"  [label="(1536, 384)", style=solid];
-"1346 linear_78" -> "1347 gelu_12"  [label="(1, 14, 14, 1536)", style=solid];
-"1347 gelu_12" -> "1348 dropout_50"  [label="(1, 14, 14, 1536)", style=solid];
-"1348 dropout_50" -> "1349 dropout_50_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1349 dropout_50_0_0_nncf_smooth_quant_0" -> "1350 quantize_per_tensor_default_80"  [label="(1, 14, 14, 1536)", style=solid];
-"1350 quantize_per_tensor_default_80" -> "1351 dequantize_per_tensor_default_80"  [label="(1, 14, 14, 1536)", style=solid];
-"1351 dequantize_per_tensor_default_80" -> "1358 linear_79"  [label="(1, 14, 14, 1536)", style=solid];
-"1352 _param_constant215_0_0" -> "1358 linear_79"  [label="(384,)", style=solid];
-"1353 scale_updated_constant54" -> "1355 mul_tensor_54"  [label="(384, 1)", style=solid];
-"1354 compressed_weight_updated_constant54" -> "1355 mul_tensor_54"  [label="(384, 1536)", style=solid];
-"1355 mul_tensor_54" -> "1357 sub_tensor_54"  [label="(384, 1536)", style=solid];
-"1356 zero_point_updated_constant54" -> "1357 sub_tensor_54"  [label="(384, 1)", style=solid];
-"1357 sub_tensor_54" -> "1358 linear_79"  [label="(384, 1536)", style=solid];
-"1358 linear_79" -> "1359 dropout_51"  [label="(1, 14, 14, 384)", style=solid];
-"1359 dropout_51" -> "1362 layer_norm_28"  [label="(1, 14, 14, 384)", style=solid];
-"1360 _param_constant216" -> "1362 layer_norm_28"  [label="(384,)", style=solid];
-"1361 _param_constant217" -> "1362 layer_norm_28"  [label="(384,)", style=solid];
-"1362 layer_norm_28" -> "1363 add_44"  [label="(1, 14, 14, 384)", style=solid];
-"1363 add_44" -> "1365 pad_15"  [label="(1, 14, 14, 384)", style=solid];
-"1363 add_44" -> "1446 add_47"  [label="(1, 14, 14, 384)", style=solid];
-"1364 _frozen_param106" -> "1401 add_45"  [label="(1, 12, 64, 64)", style=solid];
-"1365 pad_15" -> "1366 roll_12"  [label="(1, 16, 16, 384)", style=solid];
-"1366 roll_12" -> "1367 view_72"  [label="(1, 16, 16, 384)", style=solid];
-"1367 view_72" -> "1368 permute_60"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1368 permute_60" -> "1369 reshape_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1369 reshape_58" -> "1370 reshape_58_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1369 reshape_58" -> "1402 new_zeros_6"  [label="(4, 64, 384)", style=solid];
-"1370 reshape_58_0_0_nncf_smooth_quant_0" -> "1371 quantize_per_tensor_default_81"  [label="(4, 64, 384)", style=solid];
-"1371 quantize_per_tensor_default_81" -> "1372 dequantize_per_tensor_default_81"  [label="(4, 64, 384)", style=solid];
-"1372 dequantize_per_tensor_default_81" -> "1379 linear_82"  [label="(4, 64, 384)", style=solid];
-"1373 _frozen_param107_0_0" -> "1379 linear_82"  [label="(1152,)", style=solid];
-"1374 scale_updated_constant55" -> "1376 mul_tensor_55"  [label="(1152, 1)", style=solid];
-"1375 compressed_weight_updated_constant55" -> "1376 mul_tensor_55"  [label="(1152, 384)", style=solid];
-"1376 mul_tensor_55" -> "1378 sub_tensor_55"  [label="(1152, 384)", style=solid];
-"1377 zero_point_updated_constant55" -> "1378 sub_tensor_55"  [label="(1152, 1)", style=solid];
-"1378 sub_tensor_55" -> "1379 linear_82"  [label="(1152, 384)", style=solid];
-"1379 linear_82" -> "1380 reshape_59"  [label="(4, 64, 1152)", style=solid];
-"1380 reshape_59" -> "1381 permute_61"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1381 permute_61" -> "1382 select_39"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1381 permute_61" -> "1383 select_40"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1381 permute_61" -> "1384 select_41"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1382 select_39" -> "1385 linalg_vector_norm_26"  [label="(4, 12, 64, 32)", style=solid];
-"1382 select_39" -> "1387 expand_as_26"  [label="(4, 12, 64, 32)", style=solid];
-"1382 select_39" -> "1388 div_26"  [label="(4, 12, 64, 32)", style=solid];
-"1383 select_40" -> "1391 linalg_vector_norm_27"  [label="(4, 12, 64, 32)", style=solid];
-"1383 select_40" -> "1393 expand_as_27"  [label="(4, 12, 64, 32)", style=solid];
-"1383 select_40" -> "1394 div_27"  [label="(4, 12, 64, 32)", style=solid];
-"1384 select_41" -> "1420 matmul_27"  [label="(4, 12, 64, 32)", style=solid];
-"1385 linalg_vector_norm_26" -> "1386 clamp_min_26"  [label="(4, 12, 64, 1)", style=solid];
-"1386 clamp_min_26" -> "1387 expand_as_26"  [label="(4, 12, 64, 1)", style=solid];
-"1387 expand_as_26" -> "1388 div_26"  [label="(4, 12, 64, 32)", style=solid];
-"1388 div_26" -> "1389 quantize_per_tensor_default_82"  [label="(4, 12, 64, 32)", style=solid];
-"1389 quantize_per_tensor_default_82" -> "1390 dequantize_per_tensor_default_82"  [label="(4, 12, 64, 32)", style=solid];
-"1390 dequantize_per_tensor_default_82" -> "1398 matmul_26"  [label="(4, 12, 64, 32)", style=solid];
-"1391 linalg_vector_norm_27" -> "1392 clamp_min_27"  [label="(4, 12, 64, 1)", style=solid];
-"1392 clamp_min_27" -> "1393 expand_as_27"  [label="(4, 12, 64, 1)", style=solid];
-"1393 expand_as_27" -> "1394 div_27"  [label="(4, 12, 64, 32)", style=solid];
-"1394 div_27" -> "1395 quantize_per_tensor_default_83"  [label="(4, 12, 64, 32)", style=solid];
-"1395 quantize_per_tensor_default_83" -> "1396 dequantize_per_tensor_default_83"  [label="(4, 12, 64, 32)", style=solid];
-"1396 dequantize_per_tensor_default_83" -> "1397 transpose_26"  [label="(4, 12, 64, 32)", style=solid];
-"1397 transpose_26" -> "1398 matmul_26"  [label="(4, 12, 32, 64)", style=solid];
-"1398 matmul_26" -> "1400 mul_27"  [label="(4, 12, 64, 64)", style=solid];
-"1399 _frozen_param109" -> "1400 mul_27"  [label="(12, 1, 1)", style=solid];
-"1400 mul_27" -> "1401 add_45"  [label="(4, 12, 64, 64)", style=solid];
-"1401 add_45" -> "1413 view_74"  [label="(4, 12, 64, 64)", style=solid];
-"1402 new_zeros_6" -> "1403 view_73"  [label="(16, 16)", style=solid];
-"1403 view_73" -> "1404 permute_62"  [label="(2, 8, 2, 8)", style=solid];
-"1404 permute_62" -> "1405 reshape_60"  [label="(2, 2, 8, 8)", style=solid];
-"1405 reshape_60" -> "1406 unsqueeze_38"  [label="(4, 64)", style=solid];
-"1405 reshape_60" -> "1407 unsqueeze_39"  [label="(4, 64)", style=solid];
-"1406 unsqueeze_38" -> "1408 sub_6"  [label="(4, 1, 64)", style=solid];
-"1407 unsqueeze_39" -> "1408 sub_6"  [label="(4, 64, 1)", style=solid];
-"1408 sub_6" -> "1409 ne_6"  [label="(4, 64, 64)", style=solid];
-"1408 sub_6" -> "1410 masked_fill_12"  [label="(4, 64, 64)", style=solid];
-"1408 sub_6" -> "1411 eq_6"  [label="(4, 64, 64)", style=solid];
-"1409 ne_6" -> "1410 masked_fill_12"  [label="(4, 64, 64)", style=solid];
-"1410 masked_fill_12" -> "1412 masked_fill_13"  [label="(4, 64, 64)", style=solid];
-"1411 eq_6" -> "1412 masked_fill_13"  [label="(4, 64, 64)", style=solid];
-"1412 masked_fill_13" -> "1414 unsqueeze_40"  [label="(4, 64, 64)", style=solid];
-"1413 view_74" -> "1416 add_46"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1414 unsqueeze_40" -> "1415 unsqueeze_41"  [label="(4, 1, 64, 64)", style=solid];
-"1415 unsqueeze_41" -> "1416 add_46"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1416 add_46" -> "1417 view_75"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1417 view_75" -> "1418 softmax_13"  [label="(4, 12, 64, 64)", style=solid];
-"1418 softmax_13" -> "1419 dropout_52"  [label="(4, 12, 64, 64)", style=solid];
-"1419 dropout_52" -> "1420 matmul_27"  [label="(4, 12, 64, 64)", style=solid];
-"1420 matmul_27" -> "1421 transpose_27"  [label="(4, 12, 64, 32)", style=solid];
-"1421 transpose_27" -> "1422 reshape_61"  [label="(4, 64, 12, 32)", style=solid];
-"1422 reshape_61" -> "1423 reshape_61_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1423 reshape_61_0_0_nncf_smooth_quant_0" -> "1424 quantize_per_tensor_default_84"  [label="(4, 64, 384)", style=solid];
-"1424 quantize_per_tensor_default_84" -> "1425 dequantize_per_tensor_default_84"  [label="(4, 64, 384)", style=solid];
-"1425 dequantize_per_tensor_default_84" -> "1432 linear_83"  [label="(4, 64, 384)", style=solid];
-"1426 _param_constant225_0_0" -> "1432 linear_83"  [label="(384,)", style=solid];
-"1427 scale_updated_constant56" -> "1429 mul_tensor_56"  [label="(384, 1)", style=solid];
-"1428 compressed_weight_updated_constant56" -> "1429 mul_tensor_56"  [label="(384, 384)", style=solid];
-"1429 mul_tensor_56" -> "1431 sub_tensor_56"  [label="(384, 384)", style=solid];
-"1430 zero_point_updated_constant56" -> "1431 sub_tensor_56"  [label="(384, 1)", style=solid];
-"1431 sub_tensor_56" -> "1432 linear_83"  [label="(384, 384)", style=solid];
-"1432 linear_83" -> "1433 dropout_53"  [label="(4, 64, 384)", style=solid];
-"1433 dropout_53" -> "1434 view_76"  [label="(4, 64, 384)", style=solid];
-"1434 view_76" -> "1435 permute_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1435 permute_63" -> "1436 reshape_62"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1436 reshape_62" -> "1437 roll_13"  [label="(1, 16, 16, 384)", style=solid];
-"1437 roll_13" -> "1438 slice_213"  [label="(1, 16, 16, 384)", style=solid];
-"1438 slice_213" -> "1439 slice_214"  [label="(1, 16, 16, 384)", style=solid];
-"1439 slice_214" -> "1440 slice_215"  [label="(1, 14, 16, 384)", style=solid];
-"1440 slice_215" -> "1441 slice_216"  [label="(1, 14, 14, 384)", style=solid];
-"1441 slice_216" -> "1442 contiguous_25"  [label="(1, 14, 14, 384)", style=solid];
-"1442 contiguous_25" -> "1445 layer_norm_29"  [label="(1, 14, 14, 384)", style=solid];
-"1443 _param_constant226" -> "1445 layer_norm_29"  [label="(384,)", style=solid];
-"1444 _param_constant227" -> "1445 layer_norm_29"  [label="(384,)", style=solid];
-"1445 layer_norm_29" -> "1446 add_47"  [label="(1, 14, 14, 384)", style=solid];
-"1446 add_47" -> "1447 add_47_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1446 add_47" -> "1473 add_48"  [label="(1, 14, 14, 384)", style=solid];
-"1447 add_47_0_0_nncf_smooth_quant_0" -> "1448 quantize_per_tensor_default_85"  [label="(1, 14, 14, 384)", style=solid];
-"1448 quantize_per_tensor_default_85" -> "1449 dequantize_per_tensor_default_85"  [label="(1, 14, 14, 384)", style=solid];
-"1449 dequantize_per_tensor_default_85" -> "1456 linear_84"  [label="(1, 14, 14, 384)", style=solid];
-"1450 _param_constant229_0_0" -> "1456 linear_84"  [label="(1536,)", style=solid];
-"1451 scale_updated_constant57" -> "1453 mul_tensor_57"  [label="(1536, 1)", style=solid];
-"1452 compressed_weight_updated_constant57" -> "1453 mul_tensor_57"  [label="(1536, 384)", style=solid];
-"1453 mul_tensor_57" -> "1455 sub_tensor_57"  [label="(1536, 384)", style=solid];
-"1454 zero_point_updated_constant57" -> "1455 sub_tensor_57"  [label="(1536, 1)", style=solid];
-"1455 sub_tensor_57" -> "1456 linear_84"  [label="(1536, 384)", style=solid];
-"1456 linear_84" -> "1457 gelu_13"  [label="(1, 14, 14, 1536)", style=solid];
-"1457 gelu_13" -> "1458 dropout_54"  [label="(1, 14, 14, 1536)", style=solid];
-"1458 dropout_54" -> "1459 dropout_54_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1459 dropout_54_0_0_nncf_smooth_quant_0" -> "1460 quantize_per_tensor_default_86"  [label="(1, 14, 14, 1536)", style=solid];
-"1460 quantize_per_tensor_default_86" -> "1461 dequantize_per_tensor_default_86"  [label="(1, 14, 14, 1536)", style=solid];
-"1461 dequantize_per_tensor_default_86" -> "1468 linear_85"  [label="(1, 14, 14, 1536)", style=solid];
-"1462 _param_constant231_0_0" -> "1468 linear_85"  [label="(384,)", style=solid];
-"1463 scale_updated_constant58" -> "1465 mul_tensor_58"  [label="(384, 1)", style=solid];
-"1464 compressed_weight_updated_constant58" -> "1465 mul_tensor_58"  [label="(384, 1536)", style=solid];
-"1465 mul_tensor_58" -> "1467 sub_tensor_58"  [label="(384, 1536)", style=solid];
-"1466 zero_point_updated_constant58" -> "1467 sub_tensor_58"  [label="(384, 1)", style=solid];
-"1467 sub_tensor_58" -> "1468 linear_85"  [label="(384, 1536)", style=solid];
-"1468 linear_85" -> "1469 dropout_55"  [label="(1, 14, 14, 384)", style=solid];
-"1469 dropout_55" -> "1472 layer_norm_30"  [label="(1, 14, 14, 384)", style=solid];
-"1470 _param_constant232" -> "1472 layer_norm_30"  [label="(384,)", style=solid];
-"1471 _param_constant233" -> "1472 layer_norm_30"  [label="(384,)", style=solid];
-"1472 layer_norm_30" -> "1473 add_48"  [label="(1, 14, 14, 384)", style=solid];
-"1473 add_48" -> "1475 pad_16"  [label="(1, 14, 14, 384)", style=solid];
-"1473 add_48" -> "1538 add_50"  [label="(1, 14, 14, 384)", style=solid];
-"1474 _frozen_param119" -> "1510 add_49"  [label="(1, 12, 64, 64)", style=solid];
-"1475 pad_16" -> "1476 view_79"  [label="(1, 16, 16, 384)", style=solid];
-"1476 view_79" -> "1477 permute_65"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1477 permute_65" -> "1478 reshape_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1478 reshape_63" -> "1479 reshape_63_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1479 reshape_63_0_0_nncf_smooth_quant_0" -> "1480 quantize_per_tensor_default_87"  [label="(4, 64, 384)", style=solid];
-"1480 quantize_per_tensor_default_87" -> "1481 dequantize_per_tensor_default_87"  [label="(4, 64, 384)", style=solid];
-"1481 dequantize_per_tensor_default_87" -> "1488 linear_88"  [label="(4, 64, 384)", style=solid];
-"1482 _frozen_param120_0_0" -> "1488 linear_88"  [label="(1152,)", style=solid];
-"1483 scale_updated_constant59" -> "1485 mul_tensor_59"  [label="(1152, 1)", style=solid];
-"1484 compressed_weight_updated_constant59" -> "1485 mul_tensor_59"  [label="(1152, 384)", style=solid];
-"1485 mul_tensor_59" -> "1487 sub_tensor_59"  [label="(1152, 384)", style=solid];
-"1486 zero_point_updated_constant59" -> "1487 sub_tensor_59"  [label="(1152, 1)", style=solid];
-"1487 sub_tensor_59" -> "1488 linear_88"  [label="(1152, 384)", style=solid];
-"1488 linear_88" -> "1489 reshape_64"  [label="(4, 64, 1152)", style=solid];
-"1489 reshape_64" -> "1490 permute_66"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1490 permute_66" -> "1491 select_42"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1490 permute_66" -> "1492 select_43"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1490 permute_66" -> "1493 select_44"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1491 select_42" -> "1494 linalg_vector_norm_28"  [label="(4, 12, 64, 32)", style=solid];
-"1491 select_42" -> "1496 expand_as_28"  [label="(4, 12, 64, 32)", style=solid];
-"1491 select_42" -> "1497 div_28"  [label="(4, 12, 64, 32)", style=solid];
-"1492 select_43" -> "1500 linalg_vector_norm_29"  [label="(4, 12, 64, 32)", style=solid];
-"1492 select_43" -> "1502 expand_as_29"  [label="(4, 12, 64, 32)", style=solid];
-"1492 select_43" -> "1503 div_29"  [label="(4, 12, 64, 32)", style=solid];
-"1493 select_44" -> "1513 matmul_29"  [label="(4, 12, 64, 32)", style=solid];
-"1494 linalg_vector_norm_28" -> "1495 clamp_min_28"  [label="(4, 12, 64, 1)", style=solid];
-"1495 clamp_min_28" -> "1496 expand_as_28"  [label="(4, 12, 64, 1)", style=solid];
-"1496 expand_as_28" -> "1497 div_28"  [label="(4, 12, 64, 32)", style=solid];
-"1497 div_28" -> "1498 quantize_per_tensor_default_88"  [label="(4, 12, 64, 32)", style=solid];
-"1498 quantize_per_tensor_default_88" -> "1499 dequantize_per_tensor_default_88"  [label="(4, 12, 64, 32)", style=solid];
-"1499 dequantize_per_tensor_default_88" -> "1507 matmul_28"  [label="(4, 12, 64, 32)", style=solid];
-"1500 linalg_vector_norm_29" -> "1501 clamp_min_29"  [label="(4, 12, 64, 1)", style=solid];
-"1501 clamp_min_29" -> "1502 expand_as_29"  [label="(4, 12, 64, 1)", style=solid];
-"1502 expand_as_29" -> "1503 div_29"  [label="(4, 12, 64, 32)", style=solid];
-"1503 div_29" -> "1504 quantize_per_tensor_default_89"  [label="(4, 12, 64, 32)", style=solid];
-"1504 quantize_per_tensor_default_89" -> "1505 dequantize_per_tensor_default_89"  [label="(4, 12, 64, 32)", style=solid];
-"1505 dequantize_per_tensor_default_89" -> "1506 transpose_28"  [label="(4, 12, 64, 32)", style=solid];
-"1506 transpose_28" -> "1507 matmul_28"  [label="(4, 12, 32, 64)", style=solid];
-"1507 matmul_28" -> "1509 mul_29"  [label="(4, 12, 64, 64)", style=solid];
-"1508 _frozen_param122" -> "1509 mul_29"  [label="(12, 1, 1)", style=solid];
-"1509 mul_29" -> "1510 add_49"  [label="(4, 12, 64, 64)", style=solid];
-"1510 add_49" -> "1511 softmax_14"  [label="(4, 12, 64, 64)", style=solid];
-"1511 softmax_14" -> "1512 dropout_56"  [label="(4, 12, 64, 64)", style=solid];
-"1512 dropout_56" -> "1513 matmul_29"  [label="(4, 12, 64, 64)", style=solid];
-"1513 matmul_29" -> "1514 transpose_29"  [label="(4, 12, 64, 32)", style=solid];
-"1514 transpose_29" -> "1515 reshape_65"  [label="(4, 64, 12, 32)", style=solid];
-"1515 reshape_65" -> "1516 reshape_65_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1516 reshape_65_0_0_nncf_smooth_quant_0" -> "1517 quantize_per_tensor_default_90"  [label="(4, 64, 384)", style=solid];
-"1517 quantize_per_tensor_default_90" -> "1518 dequantize_per_tensor_default_90"  [label="(4, 64, 384)", style=solid];
-"1518 dequantize_per_tensor_default_90" -> "1525 linear_89"  [label="(4, 64, 384)", style=solid];
-"1519 _param_constant241_0_0" -> "1525 linear_89"  [label="(384,)", style=solid];
-"1520 scale_updated_constant60" -> "1522 mul_tensor_60"  [label="(384, 1)", style=solid];
-"1521 compressed_weight_updated_constant60" -> "1522 mul_tensor_60"  [label="(384, 384)", style=solid];
-"1522 mul_tensor_60" -> "1524 sub_tensor_60"  [label="(384, 384)", style=solid];
-"1523 zero_point_updated_constant60" -> "1524 sub_tensor_60"  [label="(384, 1)", style=solid];
-"1524 sub_tensor_60" -> "1525 linear_89"  [label="(384, 384)", style=solid];
-"1525 linear_89" -> "1526 dropout_57"  [label="(4, 64, 384)", style=solid];
-"1526 dropout_57" -> "1527 view_80"  [label="(4, 64, 384)", style=solid];
-"1527 view_80" -> "1528 permute_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1528 permute_67" -> "1529 reshape_66"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1529 reshape_66" -> "1530 slice_218"  [label="(1, 16, 16, 384)", style=solid];
-"1530 slice_218" -> "1531 slice_219"  [label="(1, 16, 16, 384)", style=solid];
-"1531 slice_219" -> "1532 slice_220"  [label="(1, 14, 16, 384)", style=solid];
-"1532 slice_220" -> "1533 slice_221"  [label="(1, 14, 14, 384)", style=solid];
-"1533 slice_221" -> "1534 contiguous_27"  [label="(1, 14, 14, 384)", style=solid];
-"1534 contiguous_27" -> "1537 layer_norm_31"  [label="(1, 14, 14, 384)", style=solid];
-"1535 _param_constant242" -> "1537 layer_norm_31"  [label="(384,)", style=solid];
-"1536 _param_constant243" -> "1537 layer_norm_31"  [label="(384,)", style=solid];
-"1537 layer_norm_31" -> "1538 add_50"  [label="(1, 14, 14, 384)", style=solid];
-"1538 add_50" -> "1539 add_50_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1538 add_50" -> "1565 add_51"  [label="(1, 14, 14, 384)", style=solid];
-"1539 add_50_0_0_nncf_smooth_quant_0" -> "1540 quantize_per_tensor_default_91"  [label="(1, 14, 14, 384)", style=solid];
-"1540 quantize_per_tensor_default_91" -> "1541 dequantize_per_tensor_default_91"  [label="(1, 14, 14, 384)", style=solid];
-"1541 dequantize_per_tensor_default_91" -> "1548 linear_90"  [label="(1, 14, 14, 384)", style=solid];
-"1542 _param_constant245_0_0" -> "1548 linear_90"  [label="(1536,)", style=solid];
-"1543 scale_updated_constant61" -> "1545 mul_tensor_61"  [label="(1536, 1)", style=solid];
-"1544 compressed_weight_updated_constant61" -> "1545 mul_tensor_61"  [label="(1536, 384)", style=solid];
-"1545 mul_tensor_61" -> "1547 sub_tensor_61"  [label="(1536, 384)", style=solid];
-"1546 zero_point_updated_constant61" -> "1547 sub_tensor_61"  [label="(1536, 1)", style=solid];
-"1547 sub_tensor_61" -> "1548 linear_90"  [label="(1536, 384)", style=solid];
-"1548 linear_90" -> "1549 gelu_14"  [label="(1, 14, 14, 1536)", style=solid];
-"1549 gelu_14" -> "1550 dropout_58"  [label="(1, 14, 14, 1536)", style=solid];
-"1550 dropout_58" -> "1551 dropout_58_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1551 dropout_58_0_0_nncf_smooth_quant_0" -> "1552 quantize_per_tensor_default_92"  [label="(1, 14, 14, 1536)", style=solid];
-"1552 quantize_per_tensor_default_92" -> "1553 dequantize_per_tensor_default_92"  [label="(1, 14, 14, 1536)", style=solid];
-"1553 dequantize_per_tensor_default_92" -> "1560 linear_91"  [label="(1, 14, 14, 1536)", style=solid];
-"1554 _param_constant247_0_0" -> "1560 linear_91"  [label="(384,)", style=solid];
-"1555 scale_updated_constant62" -> "1557 mul_tensor_62"  [label="(384, 1)", style=solid];
-"1556 compressed_weight_updated_constant62" -> "1557 mul_tensor_62"  [label="(384, 1536)", style=solid];
-"1557 mul_tensor_62" -> "1559 sub_tensor_62"  [label="(384, 1536)", style=solid];
-"1558 zero_point_updated_constant62" -> "1559 sub_tensor_62"  [label="(384, 1)", style=solid];
-"1559 sub_tensor_62" -> "1560 linear_91"  [label="(384, 1536)", style=solid];
-"1560 linear_91" -> "1561 dropout_59"  [label="(1, 14, 14, 384)", style=solid];
-"1561 dropout_59" -> "1564 layer_norm_32"  [label="(1, 14, 14, 384)", style=solid];
-"1562 _param_constant248" -> "1564 layer_norm_32"  [label="(384,)", style=solid];
-"1563 _param_constant249" -> "1564 layer_norm_32"  [label="(384,)", style=solid];
-"1564 layer_norm_32" -> "1565 add_51"  [label="(1, 14, 14, 384)", style=solid];
-"1565 add_51" -> "1567 pad_17"  [label="(1, 14, 14, 384)", style=solid];
-"1565 add_51" -> "1648 add_54"  [label="(1, 14, 14, 384)", style=solid];
-"1566 _frozen_param123" -> "1603 add_52"  [label="(1, 12, 64, 64)", style=solid];
-"1567 pad_17" -> "1568 roll_14"  [label="(1, 16, 16, 384)", style=solid];
-"1568 roll_14" -> "1569 view_83"  [label="(1, 16, 16, 384)", style=solid];
-"1569 view_83" -> "1570 permute_69"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1570 permute_69" -> "1571 reshape_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1571 reshape_67" -> "1572 reshape_67_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1571 reshape_67" -> "1604 new_zeros_7"  [label="(4, 64, 384)", style=solid];
-"1572 reshape_67_0_0_nncf_smooth_quant_0" -> "1573 quantize_per_tensor_default_93"  [label="(4, 64, 384)", style=solid];
-"1573 quantize_per_tensor_default_93" -> "1574 dequantize_per_tensor_default_93"  [label="(4, 64, 384)", style=solid];
-"1574 dequantize_per_tensor_default_93" -> "1581 linear_94"  [label="(4, 64, 384)", style=solid];
-"1575 _frozen_param124_0_0" -> "1581 linear_94"  [label="(1152,)", style=solid];
-"1576 scale_updated_constant63" -> "1578 mul_tensor_63"  [label="(1152, 1)", style=solid];
-"1577 compressed_weight_updated_constant63" -> "1578 mul_tensor_63"  [label="(1152, 384)", style=solid];
-"1578 mul_tensor_63" -> "1580 sub_tensor_63"  [label="(1152, 384)", style=solid];
-"1579 zero_point_updated_constant63" -> "1580 sub_tensor_63"  [label="(1152, 1)", style=solid];
-"1580 sub_tensor_63" -> "1581 linear_94"  [label="(1152, 384)", style=solid];
-"1581 linear_94" -> "1582 reshape_68"  [label="(4, 64, 1152)", style=solid];
-"1582 reshape_68" -> "1583 permute_70"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1583 permute_70" -> "1584 select_45"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1583 permute_70" -> "1585 select_46"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1583 permute_70" -> "1586 select_47"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1584 select_45" -> "1587 linalg_vector_norm_30"  [label="(4, 12, 64, 32)", style=solid];
-"1584 select_45" -> "1589 expand_as_30"  [label="(4, 12, 64, 32)", style=solid];
-"1584 select_45" -> "1590 div_30"  [label="(4, 12, 64, 32)", style=solid];
-"1585 select_46" -> "1593 linalg_vector_norm_31"  [label="(4, 12, 64, 32)", style=solid];
-"1585 select_46" -> "1595 expand_as_31"  [label="(4, 12, 64, 32)", style=solid];
-"1585 select_46" -> "1596 div_31"  [label="(4, 12, 64, 32)", style=solid];
-"1586 select_47" -> "1622 matmul_31"  [label="(4, 12, 64, 32)", style=solid];
-"1587 linalg_vector_norm_30" -> "1588 clamp_min_30"  [label="(4, 12, 64, 1)", style=solid];
-"1588 clamp_min_30" -> "1589 expand_as_30"  [label="(4, 12, 64, 1)", style=solid];
-"1589 expand_as_30" -> "1590 div_30"  [label="(4, 12, 64, 32)", style=solid];
-"1590 div_30" -> "1591 quantize_per_tensor_default_94"  [label="(4, 12, 64, 32)", style=solid];
-"1591 quantize_per_tensor_default_94" -> "1592 dequantize_per_tensor_default_94"  [label="(4, 12, 64, 32)", style=solid];
-"1592 dequantize_per_tensor_default_94" -> "1600 matmul_30"  [label="(4, 12, 64, 32)", style=solid];
-"1593 linalg_vector_norm_31" -> "1594 clamp_min_31"  [label="(4, 12, 64, 1)", style=solid];
-"1594 clamp_min_31" -> "1595 expand_as_31"  [label="(4, 12, 64, 1)", style=solid];
-"1595 expand_as_31" -> "1596 div_31"  [label="(4, 12, 64, 32)", style=solid];
-"1596 div_31" -> "1597 quantize_per_tensor_default_95"  [label="(4, 12, 64, 32)", style=solid];
-"1597 quantize_per_tensor_default_95" -> "1598 dequantize_per_tensor_default_95"  [label="(4, 12, 64, 32)", style=solid];
-"1598 dequantize_per_tensor_default_95" -> "1599 transpose_30"  [label="(4, 12, 64, 32)", style=solid];
-"1599 transpose_30" -> "1600 matmul_30"  [label="(4, 12, 32, 64)", style=solid];
-"1600 matmul_30" -> "1602 mul_31"  [label="(4, 12, 64, 64)", style=solid];
-"1601 _frozen_param126" -> "1602 mul_31"  [label="(12, 1, 1)", style=solid];
-"1602 mul_31" -> "1603 add_52"  [label="(4, 12, 64, 64)", style=solid];
-"1603 add_52" -> "1615 view_85"  [label="(4, 12, 64, 64)", style=solid];
-"1604 new_zeros_7" -> "1605 view_84"  [label="(16, 16)", style=solid];
-"1605 view_84" -> "1606 permute_71"  [label="(2, 8, 2, 8)", style=solid];
-"1606 permute_71" -> "1607 reshape_69"  [label="(2, 2, 8, 8)", style=solid];
-"1607 reshape_69" -> "1608 unsqueeze_44"  [label="(4, 64)", style=solid];
-"1607 reshape_69" -> "1609 unsqueeze_45"  [label="(4, 64)", style=solid];
-"1608 unsqueeze_44" -> "1610 sub_7"  [label="(4, 1, 64)", style=solid];
-"1609 unsqueeze_45" -> "1610 sub_7"  [label="(4, 64, 1)", style=solid];
-"1610 sub_7" -> "1611 ne_7"  [label="(4, 64, 64)", style=solid];
-"1610 sub_7" -> "1612 masked_fill_14"  [label="(4, 64, 64)", style=solid];
-"1610 sub_7" -> "1613 eq_7"  [label="(4, 64, 64)", style=solid];
-"1611 ne_7" -> "1612 masked_fill_14"  [label="(4, 64, 64)", style=solid];
-"1612 masked_fill_14" -> "1614 masked_fill_15"  [label="(4, 64, 64)", style=solid];
-"1613 eq_7" -> "1614 masked_fill_15"  [label="(4, 64, 64)", style=solid];
-"1614 masked_fill_15" -> "1616 unsqueeze_46"  [label="(4, 64, 64)", style=solid];
-"1615 view_85" -> "1618 add_53"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1616 unsqueeze_46" -> "1617 unsqueeze_47"  [label="(4, 1, 64, 64)", style=solid];
-"1617 unsqueeze_47" -> "1618 add_53"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1618 add_53" -> "1619 view_86"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1619 view_86" -> "1620 softmax_15"  [label="(4, 12, 64, 64)", style=solid];
-"1620 softmax_15" -> "1621 dropout_60"  [label="(4, 12, 64, 64)", style=solid];
-"1621 dropout_60" -> "1622 matmul_31"  [label="(4, 12, 64, 64)", style=solid];
-"1622 matmul_31" -> "1623 transpose_31"  [label="(4, 12, 64, 32)", style=solid];
-"1623 transpose_31" -> "1624 reshape_70"  [label="(4, 64, 12, 32)", style=solid];
-"1624 reshape_70" -> "1625 reshape_70_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1625 reshape_70_0_0_nncf_smooth_quant_0" -> "1626 quantize_per_tensor_default_96"  [label="(4, 64, 384)", style=solid];
-"1626 quantize_per_tensor_default_96" -> "1627 dequantize_per_tensor_default_96"  [label="(4, 64, 384)", style=solid];
-"1627 dequantize_per_tensor_default_96" -> "1634 linear_95"  [label="(4, 64, 384)", style=solid];
-"1628 _param_constant257_0_0" -> "1634 linear_95"  [label="(384,)", style=solid];
-"1629 scale_updated_constant64" -> "1631 mul_tensor_64"  [label="(384, 1)", style=solid];
-"1630 compressed_weight_updated_constant64" -> "1631 mul_tensor_64"  [label="(384, 384)", style=solid];
-"1631 mul_tensor_64" -> "1633 sub_tensor_64"  [label="(384, 384)", style=solid];
-"1632 zero_point_updated_constant64" -> "1633 sub_tensor_64"  [label="(384, 1)", style=solid];
-"1633 sub_tensor_64" -> "1634 linear_95"  [label="(384, 384)", style=solid];
-"1634 linear_95" -> "1635 dropout_61"  [label="(4, 64, 384)", style=solid];
-"1635 dropout_61" -> "1636 view_87"  [label="(4, 64, 384)", style=solid];
-"1636 view_87" -> "1637 permute_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1637 permute_72" -> "1638 reshape_71"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1638 reshape_71" -> "1639 roll_15"  [label="(1, 16, 16, 384)", style=solid];
-"1639 roll_15" -> "1640 slice_241"  [label="(1, 16, 16, 384)", style=solid];
-"1640 slice_241" -> "1641 slice_242"  [label="(1, 16, 16, 384)", style=solid];
-"1641 slice_242" -> "1642 slice_243"  [label="(1, 14, 16, 384)", style=solid];
-"1642 slice_243" -> "1643 slice_244"  [label="(1, 14, 14, 384)", style=solid];
-"1643 slice_244" -> "1644 contiguous_29"  [label="(1, 14, 14, 384)", style=solid];
-"1644 contiguous_29" -> "1647 layer_norm_33"  [label="(1, 14, 14, 384)", style=solid];
-"1645 _param_constant258" -> "1647 layer_norm_33"  [label="(384,)", style=solid];
-"1646 _param_constant259" -> "1647 layer_norm_33"  [label="(384,)", style=solid];
-"1647 layer_norm_33" -> "1648 add_54"  [label="(1, 14, 14, 384)", style=solid];
-"1648 add_54" -> "1649 add_54_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1648 add_54" -> "1675 add_55"  [label="(1, 14, 14, 384)", style=solid];
-"1649 add_54_0_0_nncf_smooth_quant_0" -> "1650 quantize_per_tensor_default_97"  [label="(1, 14, 14, 384)", style=solid];
-"1650 quantize_per_tensor_default_97" -> "1651 dequantize_per_tensor_default_97"  [label="(1, 14, 14, 384)", style=solid];
-"1651 dequantize_per_tensor_default_97" -> "1658 linear_96"  [label="(1, 14, 14, 384)", style=solid];
-"1652 _param_constant261_0_0" -> "1658 linear_96"  [label="(1536,)", style=solid];
-"1653 scale_updated_constant65" -> "1655 mul_tensor_65"  [label="(1536, 1)", style=solid];
-"1654 compressed_weight_updated_constant65" -> "1655 mul_tensor_65"  [label="(1536, 384)", style=solid];
-"1655 mul_tensor_65" -> "1657 sub_tensor_65"  [label="(1536, 384)", style=solid];
-"1656 zero_point_updated_constant65" -> "1657 sub_tensor_65"  [label="(1536, 1)", style=solid];
-"1657 sub_tensor_65" -> "1658 linear_96"  [label="(1536, 384)", style=solid];
-"1658 linear_96" -> "1659 gelu_15"  [label="(1, 14, 14, 1536)", style=solid];
-"1659 gelu_15" -> "1660 dropout_62"  [label="(1, 14, 14, 1536)", style=solid];
-"1660 dropout_62" -> "1661 dropout_62_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1661 dropout_62_0_0_nncf_smooth_quant_0" -> "1662 quantize_per_tensor_default_98"  [label="(1, 14, 14, 1536)", style=solid];
-"1662 quantize_per_tensor_default_98" -> "1663 dequantize_per_tensor_default_98"  [label="(1, 14, 14, 1536)", style=solid];
-"1663 dequantize_per_tensor_default_98" -> "1670 linear_97"  [label="(1, 14, 14, 1536)", style=solid];
-"1664 _param_constant263_0_0" -> "1670 linear_97"  [label="(384,)", style=solid];
-"1665 scale_updated_constant66" -> "1667 mul_tensor_66"  [label="(384, 1)", style=solid];
-"1666 compressed_weight_updated_constant66" -> "1667 mul_tensor_66"  [label="(384, 1536)", style=solid];
-"1667 mul_tensor_66" -> "1669 sub_tensor_66"  [label="(384, 1536)", style=solid];
-"1668 zero_point_updated_constant66" -> "1669 sub_tensor_66"  [label="(384, 1)", style=solid];
-"1669 sub_tensor_66" -> "1670 linear_97"  [label="(384, 1536)", style=solid];
-"1670 linear_97" -> "1671 dropout_63"  [label="(1, 14, 14, 384)", style=solid];
-"1671 dropout_63" -> "1674 layer_norm_34"  [label="(1, 14, 14, 384)", style=solid];
-"1672 _param_constant264" -> "1674 layer_norm_34"  [label="(384,)", style=solid];
-"1673 _param_constant265" -> "1674 layer_norm_34"  [label="(384,)", style=solid];
-"1674 layer_norm_34" -> "1675 add_55"  [label="(1, 14, 14, 384)", style=solid];
-"1675 add_55" -> "1677 pad_18"  [label="(1, 14, 14, 384)", style=solid];
-"1675 add_55" -> "1740 add_57"  [label="(1, 14, 14, 384)", style=solid];
-"1676 _frozen_param136" -> "1712 add_56"  [label="(1, 12, 64, 64)", style=solid];
-"1677 pad_18" -> "1678 view_90"  [label="(1, 16, 16, 384)", style=solid];
-"1678 view_90" -> "1679 permute_74"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1679 permute_74" -> "1680 reshape_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1680 reshape_72" -> "1681 reshape_72_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1681 reshape_72_0_0_nncf_smooth_quant_0" -> "1682 quantize_per_tensor_default_99"  [label="(4, 64, 384)", style=solid];
-"1682 quantize_per_tensor_default_99" -> "1683 dequantize_per_tensor_default_99"  [label="(4, 64, 384)", style=solid];
-"1683 dequantize_per_tensor_default_99" -> "1690 linear_100"  [label="(4, 64, 384)", style=solid];
-"1684 _frozen_param137_0_0" -> "1690 linear_100"  [label="(1152,)", style=solid];
-"1685 scale_updated_constant67" -> "1687 mul_tensor_67"  [label="(1152, 1)", style=solid];
-"1686 compressed_weight_updated_constant67" -> "1687 mul_tensor_67"  [label="(1152, 384)", style=solid];
-"1687 mul_tensor_67" -> "1689 sub_tensor_67"  [label="(1152, 384)", style=solid];
-"1688 zero_point_updated_constant67" -> "1689 sub_tensor_67"  [label="(1152, 1)", style=solid];
-"1689 sub_tensor_67" -> "1690 linear_100"  [label="(1152, 384)", style=solid];
-"1690 linear_100" -> "1691 reshape_73"  [label="(4, 64, 1152)", style=solid];
-"1691 reshape_73" -> "1692 permute_75"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1692 permute_75" -> "1693 select_48"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1692 permute_75" -> "1694 select_49"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1692 permute_75" -> "1695 select_50"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1693 select_48" -> "1696 linalg_vector_norm_32"  [label="(4, 12, 64, 32)", style=solid];
-"1693 select_48" -> "1698 expand_as_32"  [label="(4, 12, 64, 32)", style=solid];
-"1693 select_48" -> "1699 div_32"  [label="(4, 12, 64, 32)", style=solid];
-"1694 select_49" -> "1702 linalg_vector_norm_33"  [label="(4, 12, 64, 32)", style=solid];
-"1694 select_49" -> "1704 expand_as_33"  [label="(4, 12, 64, 32)", style=solid];
-"1694 select_49" -> "1705 div_33"  [label="(4, 12, 64, 32)", style=solid];
-"1695 select_50" -> "1715 matmul_33"  [label="(4, 12, 64, 32)", style=solid];
-"1696 linalg_vector_norm_32" -> "1697 clamp_min_32"  [label="(4, 12, 64, 1)", style=solid];
-"1697 clamp_min_32" -> "1698 expand_as_32"  [label="(4, 12, 64, 1)", style=solid];
-"1698 expand_as_32" -> "1699 div_32"  [label="(4, 12, 64, 32)", style=solid];
-"1699 div_32" -> "1700 quantize_per_tensor_default_100"  [label="(4, 12, 64, 32)", style=solid];
-"1700 quantize_per_tensor_default_100" -> "1701 dequantize_per_tensor_default_100"  [label="(4, 12, 64, 32)", style=solid];
-"1701 dequantize_per_tensor_default_100" -> "1709 matmul_32"  [label="(4, 12, 64, 32)", style=solid];
-"1702 linalg_vector_norm_33" -> "1703 clamp_min_33"  [label="(4, 12, 64, 1)", style=solid];
-"1703 clamp_min_33" -> "1704 expand_as_33"  [label="(4, 12, 64, 1)", style=solid];
-"1704 expand_as_33" -> "1705 div_33"  [label="(4, 12, 64, 32)", style=solid];
-"1705 div_33" -> "1706 quantize_per_tensor_default_101"  [label="(4, 12, 64, 32)", style=solid];
-"1706 quantize_per_tensor_default_101" -> "1707 dequantize_per_tensor_default_101"  [label="(4, 12, 64, 32)", style=solid];
-"1707 dequantize_per_tensor_default_101" -> "1708 transpose_32"  [label="(4, 12, 64, 32)", style=solid];
-"1708 transpose_32" -> "1709 matmul_32"  [label="(4, 12, 32, 64)", style=solid];
-"1709 matmul_32" -> "1711 mul_33"  [label="(4, 12, 64, 64)", style=solid];
-"1710 _frozen_param139" -> "1711 mul_33"  [label="(12, 1, 1)", style=solid];
-"1711 mul_33" -> "1712 add_56"  [label="(4, 12, 64, 64)", style=solid];
-"1712 add_56" -> "1713 softmax_16"  [label="(4, 12, 64, 64)", style=solid];
-"1713 softmax_16" -> "1714 dropout_64"  [label="(4, 12, 64, 64)", style=solid];
-"1714 dropout_64" -> "1715 matmul_33"  [label="(4, 12, 64, 64)", style=solid];
-"1715 matmul_33" -> "1716 transpose_33"  [label="(4, 12, 64, 32)", style=solid];
-"1716 transpose_33" -> "1717 reshape_74"  [label="(4, 64, 12, 32)", style=solid];
-"1717 reshape_74" -> "1718 reshape_74_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1718 reshape_74_0_0_nncf_smooth_quant_0" -> "1719 quantize_per_tensor_default_102"  [label="(4, 64, 384)", style=solid];
-"1719 quantize_per_tensor_default_102" -> "1720 dequantize_per_tensor_default_102"  [label="(4, 64, 384)", style=solid];
-"1720 dequantize_per_tensor_default_102" -> "1727 linear_101"  [label="(4, 64, 384)", style=solid];
-"1721 _param_constant273_0_0" -> "1727 linear_101"  [label="(384,)", style=solid];
-"1722 scale_updated_constant68" -> "1724 mul_tensor_68"  [label="(384, 1)", style=solid];
-"1723 compressed_weight_updated_constant68" -> "1724 mul_tensor_68"  [label="(384, 384)", style=solid];
-"1724 mul_tensor_68" -> "1726 sub_tensor_68"  [label="(384, 384)", style=solid];
-"1725 zero_point_updated_constant68" -> "1726 sub_tensor_68"  [label="(384, 1)", style=solid];
-"1726 sub_tensor_68" -> "1727 linear_101"  [label="(384, 384)", style=solid];
-"1727 linear_101" -> "1728 dropout_65"  [label="(4, 64, 384)", style=solid];
-"1728 dropout_65" -> "1729 view_91"  [label="(4, 64, 384)", style=solid];
-"1729 view_91" -> "1730 permute_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1730 permute_76" -> "1731 reshape_75"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1731 reshape_75" -> "1732 slice_246"  [label="(1, 16, 16, 384)", style=solid];
-"1732 slice_246" -> "1733 slice_247"  [label="(1, 16, 16, 384)", style=solid];
-"1733 slice_247" -> "1734 slice_248"  [label="(1, 14, 16, 384)", style=solid];
-"1734 slice_248" -> "1735 slice_249"  [label="(1, 14, 14, 384)", style=solid];
-"1735 slice_249" -> "1736 contiguous_31"  [label="(1, 14, 14, 384)", style=solid];
-"1736 contiguous_31" -> "1739 layer_norm_35"  [label="(1, 14, 14, 384)", style=solid];
-"1737 _param_constant274" -> "1739 layer_norm_35"  [label="(384,)", style=solid];
-"1738 _param_constant275" -> "1739 layer_norm_35"  [label="(384,)", style=solid];
-"1739 layer_norm_35" -> "1740 add_57"  [label="(1, 14, 14, 384)", style=solid];
-"1740 add_57" -> "1741 add_57_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1740 add_57" -> "1767 add_58"  [label="(1, 14, 14, 384)", style=solid];
-"1741 add_57_0_0_nncf_smooth_quant_0" -> "1742 quantize_per_tensor_default_103"  [label="(1, 14, 14, 384)", style=solid];
-"1742 quantize_per_tensor_default_103" -> "1743 dequantize_per_tensor_default_103"  [label="(1, 14, 14, 384)", style=solid];
-"1743 dequantize_per_tensor_default_103" -> "1750 linear_102"  [label="(1, 14, 14, 384)", style=solid];
-"1744 _param_constant277_0_0" -> "1750 linear_102"  [label="(1536,)", style=solid];
-"1745 scale_updated_constant69" -> "1747 mul_tensor_69"  [label="(1536, 1)", style=solid];
-"1746 compressed_weight_updated_constant69" -> "1747 mul_tensor_69"  [label="(1536, 384)", style=solid];
-"1747 mul_tensor_69" -> "1749 sub_tensor_69"  [label="(1536, 384)", style=solid];
-"1748 zero_point_updated_constant69" -> "1749 sub_tensor_69"  [label="(1536, 1)", style=solid];
-"1749 sub_tensor_69" -> "1750 linear_102"  [label="(1536, 384)", style=solid];
-"1750 linear_102" -> "1751 gelu_16"  [label="(1, 14, 14, 1536)", style=solid];
-"1751 gelu_16" -> "1752 dropout_66"  [label="(1, 14, 14, 1536)", style=solid];
-"1752 dropout_66" -> "1753 dropout_66_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1753 dropout_66_0_0_nncf_smooth_quant_0" -> "1754 quantize_per_tensor_default_104"  [label="(1, 14, 14, 1536)", style=solid];
-"1754 quantize_per_tensor_default_104" -> "1755 dequantize_per_tensor_default_104"  [label="(1, 14, 14, 1536)", style=solid];
-"1755 dequantize_per_tensor_default_104" -> "1762 linear_103"  [label="(1, 14, 14, 1536)", style=solid];
-"1756 _param_constant279_0_0" -> "1762 linear_103"  [label="(384,)", style=solid];
-"1757 scale_updated_constant70" -> "1759 mul_tensor_70"  [label="(384, 1)", style=solid];
-"1758 compressed_weight_updated_constant70" -> "1759 mul_tensor_70"  [label="(384, 1536)", style=solid];
-"1759 mul_tensor_70" -> "1761 sub_tensor_70"  [label="(384, 1536)", style=solid];
-"1760 zero_point_updated_constant70" -> "1761 sub_tensor_70"  [label="(384, 1)", style=solid];
-"1761 sub_tensor_70" -> "1762 linear_103"  [label="(384, 1536)", style=solid];
-"1762 linear_103" -> "1763 dropout_67"  [label="(1, 14, 14, 384)", style=solid];
-"1763 dropout_67" -> "1766 layer_norm_36"  [label="(1, 14, 14, 384)", style=solid];
-"1764 _param_constant280" -> "1766 layer_norm_36"  [label="(384,)", style=solid];
-"1765 _param_constant281" -> "1766 layer_norm_36"  [label="(384,)", style=solid];
-"1766 layer_norm_36" -> "1767 add_58"  [label="(1, 14, 14, 384)", style=solid];
-"1767 add_58" -> "1769 pad_19"  [label="(1, 14, 14, 384)", style=solid];
-"1767 add_58" -> "1850 add_61"  [label="(1, 14, 14, 384)", style=solid];
-"1768 _frozen_param140" -> "1805 add_59"  [label="(1, 12, 64, 64)", style=solid];
-"1769 pad_19" -> "1770 roll_16"  [label="(1, 16, 16, 384)", style=solid];
-"1770 roll_16" -> "1771 view_94"  [label="(1, 16, 16, 384)", style=solid];
-"1771 view_94" -> "1772 permute_78"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1772 permute_78" -> "1773 reshape_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1773 reshape_76" -> "1774 reshape_76_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1773 reshape_76" -> "1806 new_zeros_8"  [label="(4, 64, 384)", style=solid];
-"1774 reshape_76_0_0_nncf_smooth_quant_0" -> "1775 quantize_per_tensor_default_105"  [label="(4, 64, 384)", style=solid];
-"1775 quantize_per_tensor_default_105" -> "1776 dequantize_per_tensor_default_105"  [label="(4, 64, 384)", style=solid];
-"1776 dequantize_per_tensor_default_105" -> "1783 linear_106"  [label="(4, 64, 384)", style=solid];
-"1777 _frozen_param141_0_0" -> "1783 linear_106"  [label="(1152,)", style=solid];
-"1778 scale_updated_constant71" -> "1780 mul_tensor_71"  [label="(1152, 1)", style=solid];
-"1779 compressed_weight_updated_constant71" -> "1780 mul_tensor_71"  [label="(1152, 384)", style=solid];
-"1780 mul_tensor_71" -> "1782 sub_tensor_71"  [label="(1152, 384)", style=solid];
-"1781 zero_point_updated_constant71" -> "1782 sub_tensor_71"  [label="(1152, 1)", style=solid];
-"1782 sub_tensor_71" -> "1783 linear_106"  [label="(1152, 384)", style=solid];
-"1783 linear_106" -> "1784 reshape_77"  [label="(4, 64, 1152)", style=solid];
-"1784 reshape_77" -> "1785 permute_79"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1785 permute_79" -> "1786 select_51"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1785 permute_79" -> "1787 select_52"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1785 permute_79" -> "1788 select_53"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1786 select_51" -> "1789 linalg_vector_norm_34"  [label="(4, 12, 64, 32)", style=solid];
-"1786 select_51" -> "1791 expand_as_34"  [label="(4, 12, 64, 32)", style=solid];
-"1786 select_51" -> "1792 div_34"  [label="(4, 12, 64, 32)", style=solid];
-"1787 select_52" -> "1795 linalg_vector_norm_35"  [label="(4, 12, 64, 32)", style=solid];
-"1787 select_52" -> "1797 expand_as_35"  [label="(4, 12, 64, 32)", style=solid];
-"1787 select_52" -> "1798 div_35"  [label="(4, 12, 64, 32)", style=solid];
-"1788 select_53" -> "1824 matmul_35"  [label="(4, 12, 64, 32)", style=solid];
-"1789 linalg_vector_norm_34" -> "1790 clamp_min_34"  [label="(4, 12, 64, 1)", style=solid];
-"1790 clamp_min_34" -> "1791 expand_as_34"  [label="(4, 12, 64, 1)", style=solid];
-"1791 expand_as_34" -> "1792 div_34"  [label="(4, 12, 64, 32)", style=solid];
-"1792 div_34" -> "1793 quantize_per_tensor_default_106"  [label="(4, 12, 64, 32)", style=solid];
-"1793 quantize_per_tensor_default_106" -> "1794 dequantize_per_tensor_default_106"  [label="(4, 12, 64, 32)", style=solid];
-"1794 dequantize_per_tensor_default_106" -> "1802 matmul_34"  [label="(4, 12, 64, 32)", style=solid];
-"1795 linalg_vector_norm_35" -> "1796 clamp_min_35"  [label="(4, 12, 64, 1)", style=solid];
-"1796 clamp_min_35" -> "1797 expand_as_35"  [label="(4, 12, 64, 1)", style=solid];
-"1797 expand_as_35" -> "1798 div_35"  [label="(4, 12, 64, 32)", style=solid];
-"1798 div_35" -> "1799 quantize_per_tensor_default_107"  [label="(4, 12, 64, 32)", style=solid];
-"1799 quantize_per_tensor_default_107" -> "1800 dequantize_per_tensor_default_107"  [label="(4, 12, 64, 32)", style=solid];
-"1800 dequantize_per_tensor_default_107" -> "1801 transpose_34"  [label="(4, 12, 64, 32)", style=solid];
-"1801 transpose_34" -> "1802 matmul_34"  [label="(4, 12, 32, 64)", style=solid];
-"1802 matmul_34" -> "1804 mul_35"  [label="(4, 12, 64, 64)", style=solid];
-"1803 _frozen_param143" -> "1804 mul_35"  [label="(12, 1, 1)", style=solid];
-"1804 mul_35" -> "1805 add_59"  [label="(4, 12, 64, 64)", style=solid];
-"1805 add_59" -> "1817 view_96"  [label="(4, 12, 64, 64)", style=solid];
-"1806 new_zeros_8" -> "1807 view_95"  [label="(16, 16)", style=solid];
-"1807 view_95" -> "1808 permute_80"  [label="(2, 8, 2, 8)", style=solid];
-"1808 permute_80" -> "1809 reshape_78"  [label="(2, 2, 8, 8)", style=solid];
-"1809 reshape_78" -> "1810 unsqueeze_50"  [label="(4, 64)", style=solid];
-"1809 reshape_78" -> "1811 unsqueeze_51"  [label="(4, 64)", style=solid];
-"1810 unsqueeze_50" -> "1812 sub_8"  [label="(4, 1, 64)", style=solid];
-"1811 unsqueeze_51" -> "1812 sub_8"  [label="(4, 64, 1)", style=solid];
-"1812 sub_8" -> "1813 ne_8"  [label="(4, 64, 64)", style=solid];
-"1812 sub_8" -> "1814 masked_fill_16"  [label="(4, 64, 64)", style=solid];
-"1812 sub_8" -> "1815 eq_8"  [label="(4, 64, 64)", style=solid];
-"1813 ne_8" -> "1814 masked_fill_16"  [label="(4, 64, 64)", style=solid];
-"1814 masked_fill_16" -> "1816 masked_fill_17"  [label="(4, 64, 64)", style=solid];
-"1815 eq_8" -> "1816 masked_fill_17"  [label="(4, 64, 64)", style=solid];
-"1816 masked_fill_17" -> "1818 unsqueeze_52"  [label="(4, 64, 64)", style=solid];
-"1817 view_96" -> "1820 add_60"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1818 unsqueeze_52" -> "1819 unsqueeze_53"  [label="(4, 1, 64, 64)", style=solid];
-"1819 unsqueeze_53" -> "1820 add_60"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1820 add_60" -> "1821 view_97"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1821 view_97" -> "1822 softmax_17"  [label="(4, 12, 64, 64)", style=solid];
-"1822 softmax_17" -> "1823 dropout_68"  [label="(4, 12, 64, 64)", style=solid];
-"1823 dropout_68" -> "1824 matmul_35"  [label="(4, 12, 64, 64)", style=solid];
-"1824 matmul_35" -> "1825 transpose_35"  [label="(4, 12, 64, 32)", style=solid];
-"1825 transpose_35" -> "1826 reshape_79"  [label="(4, 64, 12, 32)", style=solid];
-"1826 reshape_79" -> "1827 reshape_79_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1827 reshape_79_0_0_nncf_smooth_quant_0" -> "1828 quantize_per_tensor_default_108"  [label="(4, 64, 384)", style=solid];
-"1828 quantize_per_tensor_default_108" -> "1829 dequantize_per_tensor_default_108"  [label="(4, 64, 384)", style=solid];
-"1829 dequantize_per_tensor_default_108" -> "1836 linear_107"  [label="(4, 64, 384)", style=solid];
-"1830 _param_constant289_0_0" -> "1836 linear_107"  [label="(384,)", style=solid];
-"1831 scale_updated_constant72" -> "1833 mul_tensor_72"  [label="(384, 1)", style=solid];
-"1832 compressed_weight_updated_constant72" -> "1833 mul_tensor_72"  [label="(384, 384)", style=solid];
-"1833 mul_tensor_72" -> "1835 sub_tensor_72"  [label="(384, 384)", style=solid];
-"1834 zero_point_updated_constant72" -> "1835 sub_tensor_72"  [label="(384, 1)", style=solid];
-"1835 sub_tensor_72" -> "1836 linear_107"  [label="(384, 384)", style=solid];
-"1836 linear_107" -> "1837 dropout_69"  [label="(4, 64, 384)", style=solid];
-"1837 dropout_69" -> "1838 view_98"  [label="(4, 64, 384)", style=solid];
-"1838 view_98" -> "1839 permute_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1839 permute_81" -> "1840 reshape_80"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1840 reshape_80" -> "1841 roll_17"  [label="(1, 16, 16, 384)", style=solid];
-"1841 roll_17" -> "1842 slice_269"  [label="(1, 16, 16, 384)", style=solid];
-"1842 slice_269" -> "1843 slice_270"  [label="(1, 16, 16, 384)", style=solid];
-"1843 slice_270" -> "1844 slice_271"  [label="(1, 14, 16, 384)", style=solid];
-"1844 slice_271" -> "1845 slice_272"  [label="(1, 14, 14, 384)", style=solid];
-"1845 slice_272" -> "1846 contiguous_33"  [label="(1, 14, 14, 384)", style=solid];
-"1846 contiguous_33" -> "1849 layer_norm_37"  [label="(1, 14, 14, 384)", style=solid];
-"1847 _param_constant290" -> "1849 layer_norm_37"  [label="(384,)", style=solid];
-"1848 _param_constant291" -> "1849 layer_norm_37"  [label="(384,)", style=solid];
-"1849 layer_norm_37" -> "1850 add_61"  [label="(1, 14, 14, 384)", style=solid];
-"1850 add_61" -> "1851 add_61_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1850 add_61" -> "1877 add_62"  [label="(1, 14, 14, 384)", style=solid];
-"1851 add_61_0_0_nncf_smooth_quant_0" -> "1852 quantize_per_tensor_default_109"  [label="(1, 14, 14, 384)", style=solid];
-"1852 quantize_per_tensor_default_109" -> "1853 dequantize_per_tensor_default_109"  [label="(1, 14, 14, 384)", style=solid];
-"1853 dequantize_per_tensor_default_109" -> "1860 linear_108"  [label="(1, 14, 14, 384)", style=solid];
-"1854 _param_constant293_0_0" -> "1860 linear_108"  [label="(1536,)", style=solid];
-"1855 scale_updated_constant73" -> "1857 mul_tensor_73"  [label="(1536, 1)", style=solid];
-"1856 compressed_weight_updated_constant73" -> "1857 mul_tensor_73"  [label="(1536, 384)", style=solid];
-"1857 mul_tensor_73" -> "1859 sub_tensor_73"  [label="(1536, 384)", style=solid];
-"1858 zero_point_updated_constant73" -> "1859 sub_tensor_73"  [label="(1536, 1)", style=solid];
-"1859 sub_tensor_73" -> "1860 linear_108"  [label="(1536, 384)", style=solid];
-"1860 linear_108" -> "1861 gelu_17"  [label="(1, 14, 14, 1536)", style=solid];
-"1861 gelu_17" -> "1862 dropout_70"  [label="(1, 14, 14, 1536)", style=solid];
-"1862 dropout_70" -> "1863 dropout_70_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1863 dropout_70_0_0_nncf_smooth_quant_0" -> "1864 quantize_per_tensor_default_110"  [label="(1, 14, 14, 1536)", style=solid];
-"1864 quantize_per_tensor_default_110" -> "1865 dequantize_per_tensor_default_110"  [label="(1, 14, 14, 1536)", style=solid];
-"1865 dequantize_per_tensor_default_110" -> "1872 linear_109"  [label="(1, 14, 14, 1536)", style=solid];
-"1866 _param_constant295_0_0" -> "1872 linear_109"  [label="(384,)", style=solid];
-"1867 scale_updated_constant74" -> "1869 mul_tensor_74"  [label="(384, 1)", style=solid];
-"1868 compressed_weight_updated_constant74" -> "1869 mul_tensor_74"  [label="(384, 1536)", style=solid];
-"1869 mul_tensor_74" -> "1871 sub_tensor_74"  [label="(384, 1536)", style=solid];
-"1870 zero_point_updated_constant74" -> "1871 sub_tensor_74"  [label="(384, 1)", style=solid];
-"1871 sub_tensor_74" -> "1872 linear_109"  [label="(384, 1536)", style=solid];
-"1872 linear_109" -> "1873 dropout_71"  [label="(1, 14, 14, 384)", style=solid];
-"1873 dropout_71" -> "1876 layer_norm_38"  [label="(1, 14, 14, 384)", style=solid];
-"1874 _param_constant296" -> "1876 layer_norm_38"  [label="(384,)", style=solid];
-"1875 _param_constant297" -> "1876 layer_norm_38"  [label="(384,)", style=solid];
-"1876 layer_norm_38" -> "1877 add_62"  [label="(1, 14, 14, 384)", style=solid];
-"1877 add_62" -> "1879 pad_20"  [label="(1, 14, 14, 384)", style=solid];
-"1877 add_62" -> "1942 add_64"  [label="(1, 14, 14, 384)", style=solid];
-"1878 _frozen_param153" -> "1914 add_63"  [label="(1, 12, 64, 64)", style=solid];
-"1879 pad_20" -> "1880 view_101"  [label="(1, 16, 16, 384)", style=solid];
-"1880 view_101" -> "1881 permute_83"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1881 permute_83" -> "1882 reshape_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1882 reshape_81" -> "1883 reshape_81_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1883 reshape_81_0_0_nncf_smooth_quant_0" -> "1884 quantize_per_tensor_default_111"  [label="(4, 64, 384)", style=solid];
-"1884 quantize_per_tensor_default_111" -> "1885 dequantize_per_tensor_default_111"  [label="(4, 64, 384)", style=solid];
-"1885 dequantize_per_tensor_default_111" -> "1892 linear_112"  [label="(4, 64, 384)", style=solid];
-"1886 _frozen_param154_0_0" -> "1892 linear_112"  [label="(1152,)", style=solid];
-"1887 scale_updated_constant75" -> "1889 mul_tensor_75"  [label="(1152, 1)", style=solid];
-"1888 compressed_weight_updated_constant75" -> "1889 mul_tensor_75"  [label="(1152, 384)", style=solid];
-"1889 mul_tensor_75" -> "1891 sub_tensor_75"  [label="(1152, 384)", style=solid];
-"1890 zero_point_updated_constant75" -> "1891 sub_tensor_75"  [label="(1152, 1)", style=solid];
-"1891 sub_tensor_75" -> "1892 linear_112"  [label="(1152, 384)", style=solid];
-"1892 linear_112" -> "1893 reshape_82"  [label="(4, 64, 1152)", style=solid];
-"1893 reshape_82" -> "1894 permute_84"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1894 permute_84" -> "1895 select_54"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1894 permute_84" -> "1896 select_55"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1894 permute_84" -> "1897 select_56"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1895 select_54" -> "1898 linalg_vector_norm_36"  [label="(4, 12, 64, 32)", style=solid];
-"1895 select_54" -> "1900 expand_as_36"  [label="(4, 12, 64, 32)", style=solid];
-"1895 select_54" -> "1901 div_36"  [label="(4, 12, 64, 32)", style=solid];
-"1896 select_55" -> "1904 linalg_vector_norm_37"  [label="(4, 12, 64, 32)", style=solid];
-"1896 select_55" -> "1906 expand_as_37"  [label="(4, 12, 64, 32)", style=solid];
-"1896 select_55" -> "1907 div_37"  [label="(4, 12, 64, 32)", style=solid];
-"1897 select_56" -> "1917 matmul_37"  [label="(4, 12, 64, 32)", style=solid];
-"1898 linalg_vector_norm_36" -> "1899 clamp_min_36"  [label="(4, 12, 64, 1)", style=solid];
-"1899 clamp_min_36" -> "1900 expand_as_36"  [label="(4, 12, 64, 1)", style=solid];
-"1900 expand_as_36" -> "1901 div_36"  [label="(4, 12, 64, 32)", style=solid];
-"1901 div_36" -> "1902 quantize_per_tensor_default_112"  [label="(4, 12, 64, 32)", style=solid];
-"1902 quantize_per_tensor_default_112" -> "1903 dequantize_per_tensor_default_112"  [label="(4, 12, 64, 32)", style=solid];
-"1903 dequantize_per_tensor_default_112" -> "1911 matmul_36"  [label="(4, 12, 64, 32)", style=solid];
-"1904 linalg_vector_norm_37" -> "1905 clamp_min_37"  [label="(4, 12, 64, 1)", style=solid];
-"1905 clamp_min_37" -> "1906 expand_as_37"  [label="(4, 12, 64, 1)", style=solid];
-"1906 expand_as_37" -> "1907 div_37"  [label="(4, 12, 64, 32)", style=solid];
-"1907 div_37" -> "1908 quantize_per_tensor_default_113"  [label="(4, 12, 64, 32)", style=solid];
-"1908 quantize_per_tensor_default_113" -> "1909 dequantize_per_tensor_default_113"  [label="(4, 12, 64, 32)", style=solid];
-"1909 dequantize_per_tensor_default_113" -> "1910 transpose_36"  [label="(4, 12, 64, 32)", style=solid];
-"1910 transpose_36" -> "1911 matmul_36"  [label="(4, 12, 32, 64)", style=solid];
-"1911 matmul_36" -> "1913 mul_37"  [label="(4, 12, 64, 64)", style=solid];
-"1912 _frozen_param156" -> "1913 mul_37"  [label="(12, 1, 1)", style=solid];
-"1913 mul_37" -> "1914 add_63"  [label="(4, 12, 64, 64)", style=solid];
-"1914 add_63" -> "1915 softmax_18"  [label="(4, 12, 64, 64)", style=solid];
-"1915 softmax_18" -> "1916 dropout_72"  [label="(4, 12, 64, 64)", style=solid];
-"1916 dropout_72" -> "1917 matmul_37"  [label="(4, 12, 64, 64)", style=solid];
-"1917 matmul_37" -> "1918 transpose_37"  [label="(4, 12, 64, 32)", style=solid];
-"1918 transpose_37" -> "1919 reshape_83"  [label="(4, 64, 12, 32)", style=solid];
-"1919 reshape_83" -> "1920 reshape_83_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1920 reshape_83_0_0_nncf_smooth_quant_0" -> "1921 quantize_per_tensor_default_114"  [label="(4, 64, 384)", style=solid];
-"1921 quantize_per_tensor_default_114" -> "1922 dequantize_per_tensor_default_114"  [label="(4, 64, 384)", style=solid];
-"1922 dequantize_per_tensor_default_114" -> "1929 linear_113"  [label="(4, 64, 384)", style=solid];
-"1923 _param_constant305_0_0" -> "1929 linear_113"  [label="(384,)", style=solid];
-"1924 scale_updated_constant76" -> "1926 mul_tensor_76"  [label="(384, 1)", style=solid];
-"1925 compressed_weight_updated_constant76" -> "1926 mul_tensor_76"  [label="(384, 384)", style=solid];
-"1926 mul_tensor_76" -> "1928 sub_tensor_76"  [label="(384, 384)", style=solid];
-"1927 zero_point_updated_constant76" -> "1928 sub_tensor_76"  [label="(384, 1)", style=solid];
-"1928 sub_tensor_76" -> "1929 linear_113"  [label="(384, 384)", style=solid];
-"1929 linear_113" -> "1930 dropout_73"  [label="(4, 64, 384)", style=solid];
-"1930 dropout_73" -> "1931 view_102"  [label="(4, 64, 384)", style=solid];
-"1931 view_102" -> "1932 permute_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1932 permute_85" -> "1933 reshape_84"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1933 reshape_84" -> "1934 slice_274"  [label="(1, 16, 16, 384)", style=solid];
-"1934 slice_274" -> "1935 slice_275"  [label="(1, 16, 16, 384)", style=solid];
-"1935 slice_275" -> "1936 slice_276"  [label="(1, 14, 16, 384)", style=solid];
-"1936 slice_276" -> "1937 slice_277"  [label="(1, 14, 14, 384)", style=solid];
-"1937 slice_277" -> "1938 contiguous_35"  [label="(1, 14, 14, 384)", style=solid];
-"1938 contiguous_35" -> "1941 layer_norm_39"  [label="(1, 14, 14, 384)", style=solid];
-"1939 _param_constant306" -> "1941 layer_norm_39"  [label="(384,)", style=solid];
-"1940 _param_constant307" -> "1941 layer_norm_39"  [label="(384,)", style=solid];
-"1941 layer_norm_39" -> "1942 add_64"  [label="(1, 14, 14, 384)", style=solid];
-"1942 add_64" -> "1943 add_64_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1942 add_64" -> "1969 add_65"  [label="(1, 14, 14, 384)", style=solid];
-"1943 add_64_0_0_nncf_smooth_quant_0" -> "1944 quantize_per_tensor_default_115"  [label="(1, 14, 14, 384)", style=solid];
-"1944 quantize_per_tensor_default_115" -> "1945 dequantize_per_tensor_default_115"  [label="(1, 14, 14, 384)", style=solid];
-"1945 dequantize_per_tensor_default_115" -> "1952 linear_114"  [label="(1, 14, 14, 384)", style=solid];
-"1946 _param_constant309_0_0" -> "1952 linear_114"  [label="(1536,)", style=solid];
-"1947 scale_updated_constant77" -> "1949 mul_tensor_77"  [label="(1536, 1)", style=solid];
-"1948 compressed_weight_updated_constant77" -> "1949 mul_tensor_77"  [label="(1536, 384)", style=solid];
-"1949 mul_tensor_77" -> "1951 sub_tensor_77"  [label="(1536, 384)", style=solid];
-"1950 zero_point_updated_constant77" -> "1951 sub_tensor_77"  [label="(1536, 1)", style=solid];
-"1951 sub_tensor_77" -> "1952 linear_114"  [label="(1536, 384)", style=solid];
-"1952 linear_114" -> "1953 gelu_18"  [label="(1, 14, 14, 1536)", style=solid];
-"1953 gelu_18" -> "1954 dropout_74"  [label="(1, 14, 14, 1536)", style=solid];
-"1954 dropout_74" -> "1955 dropout_74_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1955 dropout_74_0_0_nncf_smooth_quant_0" -> "1956 quantize_per_tensor_default_116"  [label="(1, 14, 14, 1536)", style=solid];
-"1956 quantize_per_tensor_default_116" -> "1957 dequantize_per_tensor_default_116"  [label="(1, 14, 14, 1536)", style=solid];
-"1957 dequantize_per_tensor_default_116" -> "1964 linear_115"  [label="(1, 14, 14, 1536)", style=solid];
-"1958 _param_constant311_0_0" -> "1964 linear_115"  [label="(384,)", style=solid];
-"1959 scale_updated_constant78" -> "1961 mul_tensor_78"  [label="(384, 1)", style=solid];
-"1960 compressed_weight_updated_constant78" -> "1961 mul_tensor_78"  [label="(384, 1536)", style=solid];
-"1961 mul_tensor_78" -> "1963 sub_tensor_78"  [label="(384, 1536)", style=solid];
-"1962 zero_point_updated_constant78" -> "1963 sub_tensor_78"  [label="(384, 1)", style=solid];
-"1963 sub_tensor_78" -> "1964 linear_115"  [label="(384, 1536)", style=solid];
-"1964 linear_115" -> "1965 dropout_75"  [label="(1, 14, 14, 384)", style=solid];
-"1965 dropout_75" -> "1968 layer_norm_40"  [label="(1, 14, 14, 384)", style=solid];
-"1966 _param_constant312" -> "1968 layer_norm_40"  [label="(384,)", style=solid];
-"1967 _param_constant313" -> "1968 layer_norm_40"  [label="(384,)", style=solid];
-"1968 layer_norm_40" -> "1969 add_65"  [label="(1, 14, 14, 384)", style=solid];
-"1969 add_65" -> "1971 pad_21"  [label="(1, 14, 14, 384)", style=solid];
-"1969 add_65" -> "2052 add_68"  [label="(1, 14, 14, 384)", style=solid];
-"1970 _frozen_param157" -> "2007 add_66"  [label="(1, 12, 64, 64)", style=solid];
-"1971 pad_21" -> "1972 roll_18"  [label="(1, 16, 16, 384)", style=solid];
-"1972 roll_18" -> "1973 view_105"  [label="(1, 16, 16, 384)", style=solid];
-"1973 view_105" -> "1974 permute_87"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1974 permute_87" -> "1975 reshape_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1975 reshape_85" -> "1976 reshape_85_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1975 reshape_85" -> "2008 new_zeros_9"  [label="(4, 64, 384)", style=solid];
-"1976 reshape_85_0_0_nncf_smooth_quant_0" -> "1977 quantize_per_tensor_default_117"  [label="(4, 64, 384)", style=solid];
-"1977 quantize_per_tensor_default_117" -> "1978 dequantize_per_tensor_default_117"  [label="(4, 64, 384)", style=solid];
-"1978 dequantize_per_tensor_default_117" -> "1985 linear_118"  [label="(4, 64, 384)", style=solid];
-"1979 _frozen_param158_0_0" -> "1985 linear_118"  [label="(1152,)", style=solid];
-"1980 scale_updated_constant79" -> "1982 mul_tensor_79"  [label="(1152, 1)", style=solid];
-"1981 compressed_weight_updated_constant79" -> "1982 mul_tensor_79"  [label="(1152, 384)", style=solid];
-"1982 mul_tensor_79" -> "1984 sub_tensor_79"  [label="(1152, 384)", style=solid];
-"1983 zero_point_updated_constant79" -> "1984 sub_tensor_79"  [label="(1152, 1)", style=solid];
-"1984 sub_tensor_79" -> "1985 linear_118"  [label="(1152, 384)", style=solid];
-"1985 linear_118" -> "1986 reshape_86"  [label="(4, 64, 1152)", style=solid];
-"1986 reshape_86" -> "1987 permute_88"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1987 permute_88" -> "1988 select_57"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1987 permute_88" -> "1989 select_58"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1987 permute_88" -> "1990 select_59"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1988 select_57" -> "1991 linalg_vector_norm_38"  [label="(4, 12, 64, 32)", style=solid];
-"1988 select_57" -> "1993 expand_as_38"  [label="(4, 12, 64, 32)", style=solid];
-"1988 select_57" -> "1994 div_38"  [label="(4, 12, 64, 32)", style=solid];
-"1989 select_58" -> "1997 linalg_vector_norm_39"  [label="(4, 12, 64, 32)", style=solid];
-"1989 select_58" -> "1999 expand_as_39"  [label="(4, 12, 64, 32)", style=solid];
-"1989 select_58" -> "2000 div_39"  [label="(4, 12, 64, 32)", style=solid];
-"1990 select_59" -> "2026 matmul_39"  [label="(4, 12, 64, 32)", style=solid];
-"1991 linalg_vector_norm_38" -> "1992 clamp_min_38"  [label="(4, 12, 64, 1)", style=solid];
-"1992 clamp_min_38" -> "1993 expand_as_38"  [label="(4, 12, 64, 1)", style=solid];
-"1993 expand_as_38" -> "1994 div_38"  [label="(4, 12, 64, 32)", style=solid];
-"1994 div_38" -> "1995 quantize_per_tensor_default_118"  [label="(4, 12, 64, 32)", style=solid];
-"1995 quantize_per_tensor_default_118" -> "1996 dequantize_per_tensor_default_118"  [label="(4, 12, 64, 32)", style=solid];
-"1996 dequantize_per_tensor_default_118" -> "2004 matmul_38"  [label="(4, 12, 64, 32)", style=solid];
-"1997 linalg_vector_norm_39" -> "1998 clamp_min_39"  [label="(4, 12, 64, 1)", style=solid];
-"1998 clamp_min_39" -> "1999 expand_as_39"  [label="(4, 12, 64, 1)", style=solid];
-"1999 expand_as_39" -> "2000 div_39"  [label="(4, 12, 64, 32)", style=solid];
-"2000 div_39" -> "2001 quantize_per_tensor_default_119"  [label="(4, 12, 64, 32)", style=solid];
-"2001 quantize_per_tensor_default_119" -> "2002 dequantize_per_tensor_default_119"  [label="(4, 12, 64, 32)", style=solid];
-"2002 dequantize_per_tensor_default_119" -> "2003 transpose_38"  [label="(4, 12, 64, 32)", style=solid];
-"2003 transpose_38" -> "2004 matmul_38"  [label="(4, 12, 32, 64)", style=solid];
-"2004 matmul_38" -> "2006 mul_39"  [label="(4, 12, 64, 64)", style=solid];
-"2005 _frozen_param160" -> "2006 mul_39"  [label="(12, 1, 1)", style=solid];
-"2006 mul_39" -> "2007 add_66"  [label="(4, 12, 64, 64)", style=solid];
-"2007 add_66" -> "2019 view_107"  [label="(4, 12, 64, 64)", style=solid];
-"2008 new_zeros_9" -> "2009 view_106"  [label="(16, 16)", style=solid];
-"2009 view_106" -> "2010 permute_89"  [label="(2, 8, 2, 8)", style=solid];
-"2010 permute_89" -> "2011 reshape_87"  [label="(2, 2, 8, 8)", style=solid];
-"2011 reshape_87" -> "2012 unsqueeze_56"  [label="(4, 64)", style=solid];
-"2011 reshape_87" -> "2013 unsqueeze_57"  [label="(4, 64)", style=solid];
-"2012 unsqueeze_56" -> "2014 sub_9"  [label="(4, 1, 64)", style=solid];
-"2013 unsqueeze_57" -> "2014 sub_9"  [label="(4, 64, 1)", style=solid];
-"2014 sub_9" -> "2015 ne_9"  [label="(4, 64, 64)", style=solid];
-"2014 sub_9" -> "2016 masked_fill_18"  [label="(4, 64, 64)", style=solid];
-"2014 sub_9" -> "2017 eq_9"  [label="(4, 64, 64)", style=solid];
-"2015 ne_9" -> "2016 masked_fill_18"  [label="(4, 64, 64)", style=solid];
-"2016 masked_fill_18" -> "2018 masked_fill_19"  [label="(4, 64, 64)", style=solid];
-"2017 eq_9" -> "2018 masked_fill_19"  [label="(4, 64, 64)", style=solid];
-"2018 masked_fill_19" -> "2020 unsqueeze_58"  [label="(4, 64, 64)", style=solid];
-"2019 view_107" -> "2022 add_67"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2020 unsqueeze_58" -> "2021 unsqueeze_59"  [label="(4, 1, 64, 64)", style=solid];
-"2021 unsqueeze_59" -> "2022 add_67"  [label="(1, 4, 1, 64, 64)", style=solid];
-"2022 add_67" -> "2023 view_108"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2023 view_108" -> "2024 softmax_19"  [label="(4, 12, 64, 64)", style=solid];
-"2024 softmax_19" -> "2025 dropout_76"  [label="(4, 12, 64, 64)", style=solid];
-"2025 dropout_76" -> "2026 matmul_39"  [label="(4, 12, 64, 64)", style=solid];
-"2026 matmul_39" -> "2027 transpose_39"  [label="(4, 12, 64, 32)", style=solid];
-"2027 transpose_39" -> "2028 reshape_88"  [label="(4, 64, 12, 32)", style=solid];
-"2028 reshape_88" -> "2029 reshape_88_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"2029 reshape_88_0_0_nncf_smooth_quant_0" -> "2030 quantize_per_tensor_default_120"  [label="(4, 64, 384)", style=solid];
-"2030 quantize_per_tensor_default_120" -> "2031 dequantize_per_tensor_default_120"  [label="(4, 64, 384)", style=solid];
-"2031 dequantize_per_tensor_default_120" -> "2038 linear_119"  [label="(4, 64, 384)", style=solid];
-"2032 _param_constant321_0_0" -> "2038 linear_119"  [label="(384,)", style=solid];
-"2033 scale_updated_constant80" -> "2035 mul_tensor_80"  [label="(384, 1)", style=solid];
-"2034 compressed_weight_updated_constant80" -> "2035 mul_tensor_80"  [label="(384, 384)", style=solid];
-"2035 mul_tensor_80" -> "2037 sub_tensor_80"  [label="(384, 384)", style=solid];
-"2036 zero_point_updated_constant80" -> "2037 sub_tensor_80"  [label="(384, 1)", style=solid];
-"2037 sub_tensor_80" -> "2038 linear_119"  [label="(384, 384)", style=solid];
-"2038 linear_119" -> "2039 dropout_77"  [label="(4, 64, 384)", style=solid];
-"2039 dropout_77" -> "2040 view_109"  [label="(4, 64, 384)", style=solid];
-"2040 view_109" -> "2041 permute_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2041 permute_90" -> "2042 reshape_89"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2042 reshape_89" -> "2043 roll_19"  [label="(1, 16, 16, 384)", style=solid];
-"2043 roll_19" -> "2044 slice_297"  [label="(1, 16, 16, 384)", style=solid];
-"2044 slice_297" -> "2045 slice_298"  [label="(1, 16, 16, 384)", style=solid];
-"2045 slice_298" -> "2046 slice_299"  [label="(1, 14, 16, 384)", style=solid];
-"2046 slice_299" -> "2047 slice_300"  [label="(1, 14, 14, 384)", style=solid];
-"2047 slice_300" -> "2048 contiguous_37"  [label="(1, 14, 14, 384)", style=solid];
-"2048 contiguous_37" -> "2051 layer_norm_41"  [label="(1, 14, 14, 384)", style=solid];
-"2049 _param_constant322" -> "2051 layer_norm_41"  [label="(384,)", style=solid];
-"2050 _param_constant323" -> "2051 layer_norm_41"  [label="(384,)", style=solid];
-"2051 layer_norm_41" -> "2052 add_68"  [label="(1, 14, 14, 384)", style=solid];
-"2052 add_68" -> "2053 add_68_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"2052 add_68" -> "2079 add_69"  [label="(1, 14, 14, 384)", style=solid];
-"2053 add_68_0_0_nncf_smooth_quant_0" -> "2054 quantize_per_tensor_default_121"  [label="(1, 14, 14, 384)", style=solid];
-"2054 quantize_per_tensor_default_121" -> "2055 dequantize_per_tensor_default_121"  [label="(1, 14, 14, 384)", style=solid];
-"2055 dequantize_per_tensor_default_121" -> "2062 linear_120"  [label="(1, 14, 14, 384)", style=solid];
-"2056 _param_constant325_0_0" -> "2062 linear_120"  [label="(1536,)", style=solid];
-"2057 scale_updated_constant81" -> "2059 mul_tensor_81"  [label="(1536, 1)", style=solid];
-"2058 compressed_weight_updated_constant81" -> "2059 mul_tensor_81"  [label="(1536, 384)", style=solid];
-"2059 mul_tensor_81" -> "2061 sub_tensor_81"  [label="(1536, 384)", style=solid];
-"2060 zero_point_updated_constant81" -> "2061 sub_tensor_81"  [label="(1536, 1)", style=solid];
-"2061 sub_tensor_81" -> "2062 linear_120"  [label="(1536, 384)", style=solid];
-"2062 linear_120" -> "2063 gelu_19"  [label="(1, 14, 14, 1536)", style=solid];
-"2063 gelu_19" -> "2064 dropout_78"  [label="(1, 14, 14, 1536)", style=solid];
-"2064 dropout_78" -> "2065 dropout_78_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"2065 dropout_78_0_0_nncf_smooth_quant_0" -> "2066 quantize_per_tensor_default_122"  [label="(1, 14, 14, 1536)", style=solid];
-"2066 quantize_per_tensor_default_122" -> "2067 dequantize_per_tensor_default_122"  [label="(1, 14, 14, 1536)", style=solid];
-"2067 dequantize_per_tensor_default_122" -> "2074 linear_121"  [label="(1, 14, 14, 1536)", style=solid];
-"2068 _param_constant327_0_0" -> "2074 linear_121"  [label="(384,)", style=solid];
-"2069 scale_updated_constant82" -> "2071 mul_tensor_82"  [label="(384, 1)", style=solid];
-"2070 compressed_weight_updated_constant82" -> "2071 mul_tensor_82"  [label="(384, 1536)", style=solid];
-"2071 mul_tensor_82" -> "2073 sub_tensor_82"  [label="(384, 1536)", style=solid];
-"2072 zero_point_updated_constant82" -> "2073 sub_tensor_82"  [label="(384, 1)", style=solid];
-"2073 sub_tensor_82" -> "2074 linear_121"  [label="(384, 1536)", style=solid];
-"2074 linear_121" -> "2075 dropout_79"  [label="(1, 14, 14, 384)", style=solid];
-"2075 dropout_79" -> "2078 layer_norm_42"  [label="(1, 14, 14, 384)", style=solid];
-"2076 _param_constant328" -> "2078 layer_norm_42"  [label="(384,)", style=solid];
-"2077 _param_constant329" -> "2078 layer_norm_42"  [label="(384,)", style=solid];
-"2078 layer_norm_42" -> "2079 add_69"  [label="(1, 14, 14, 384)", style=solid];
-"2079 add_69" -> "2081 pad_22"  [label="(1, 14, 14, 384)", style=solid];
-"2079 add_69" -> "2144 add_71"  [label="(1, 14, 14, 384)", style=solid];
-"2080 _frozen_param170" -> "2116 add_70"  [label="(1, 12, 64, 64)", style=solid];
-"2081 pad_22" -> "2082 view_112"  [label="(1, 16, 16, 384)", style=solid];
-"2082 view_112" -> "2083 permute_92"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2083 permute_92" -> "2084 reshape_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2084 reshape_90" -> "2085 reshape_90_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"2085 reshape_90_0_0_nncf_smooth_quant_0" -> "2086 quantize_per_tensor_default_123"  [label="(4, 64, 384)", style=solid];
-"2086 quantize_per_tensor_default_123" -> "2087 dequantize_per_tensor_default_123"  [label="(4, 64, 384)", style=solid];
-"2087 dequantize_per_tensor_default_123" -> "2094 linear_124"  [label="(4, 64, 384)", style=solid];
-"2088 _frozen_param171_0_0" -> "2094 linear_124"  [label="(1152,)", style=solid];
-"2089 scale_updated_constant83" -> "2091 mul_tensor_83"  [label="(1152, 1)", style=solid];
-"2090 compressed_weight_updated_constant83" -> "2091 mul_tensor_83"  [label="(1152, 384)", style=solid];
-"2091 mul_tensor_83" -> "2093 sub_tensor_83"  [label="(1152, 384)", style=solid];
-"2092 zero_point_updated_constant83" -> "2093 sub_tensor_83"  [label="(1152, 1)", style=solid];
-"2093 sub_tensor_83" -> "2094 linear_124"  [label="(1152, 384)", style=solid];
-"2094 linear_124" -> "2095 reshape_91"  [label="(4, 64, 1152)", style=solid];
-"2095 reshape_91" -> "2096 permute_93"  [label="(4, 64, 3, 12, 32)", style=solid];
-"2096 permute_93" -> "2097 select_60"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2096 permute_93" -> "2098 select_61"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2096 permute_93" -> "2099 select_62"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2097 select_60" -> "2100 linalg_vector_norm_40"  [label="(4, 12, 64, 32)", style=solid];
-"2097 select_60" -> "2102 expand_as_40"  [label="(4, 12, 64, 32)", style=solid];
-"2097 select_60" -> "2103 div_40"  [label="(4, 12, 64, 32)", style=solid];
-"2098 select_61" -> "2106 linalg_vector_norm_41"  [label="(4, 12, 64, 32)", style=solid];
-"2098 select_61" -> "2108 expand_as_41"  [label="(4, 12, 64, 32)", style=solid];
-"2098 select_61" -> "2109 div_41"  [label="(4, 12, 64, 32)", style=solid];
-"2099 select_62" -> "2119 matmul_41"  [label="(4, 12, 64, 32)", style=solid];
-"2100 linalg_vector_norm_40" -> "2101 clamp_min_40"  [label="(4, 12, 64, 1)", style=solid];
-"2101 clamp_min_40" -> "2102 expand_as_40"  [label="(4, 12, 64, 1)", style=solid];
-"2102 expand_as_40" -> "2103 div_40"  [label="(4, 12, 64, 32)", style=solid];
-"2103 div_40" -> "2104 quantize_per_tensor_default_124"  [label="(4, 12, 64, 32)", style=solid];
-"2104 quantize_per_tensor_default_124" -> "2105 dequantize_per_tensor_default_124"  [label="(4, 12, 64, 32)", style=solid];
-"2105 dequantize_per_tensor_default_124" -> "2113 matmul_40"  [label="(4, 12, 64, 32)", style=solid];
-"2106 linalg_vector_norm_41" -> "2107 clamp_min_41"  [label="(4, 12, 64, 1)", style=solid];
-"2107 clamp_min_41" -> "2108 expand_as_41"  [label="(4, 12, 64, 1)", style=solid];
-"2108 expand_as_41" -> "2109 div_41"  [label="(4, 12, 64, 32)", style=solid];
-"2109 div_41" -> "2110 quantize_per_tensor_default_125"  [label="(4, 12, 64, 32)", style=solid];
-"2110 quantize_per_tensor_default_125" -> "2111 dequantize_per_tensor_default_125"  [label="(4, 12, 64, 32)", style=solid];
-"2111 dequantize_per_tensor_default_125" -> "2112 transpose_40"  [label="(4, 12, 64, 32)", style=solid];
-"2112 transpose_40" -> "2113 matmul_40"  [label="(4, 12, 32, 64)", style=solid];
-"2113 matmul_40" -> "2115 mul_41"  [label="(4, 12, 64, 64)", style=solid];
-"2114 _frozen_param173" -> "2115 mul_41"  [label="(12, 1, 1)", style=solid];
-"2115 mul_41" -> "2116 add_70"  [label="(4, 12, 64, 64)", style=solid];
-"2116 add_70" -> "2117 softmax_20"  [label="(4, 12, 64, 64)", style=solid];
-"2117 softmax_20" -> "2118 dropout_80"  [label="(4, 12, 64, 64)", style=solid];
-"2118 dropout_80" -> "2119 matmul_41"  [label="(4, 12, 64, 64)", style=solid];
-"2119 matmul_41" -> "2120 transpose_41"  [label="(4, 12, 64, 32)", style=solid];
-"2120 transpose_41" -> "2121 reshape_92"  [label="(4, 64, 12, 32)", style=solid];
-"2121 reshape_92" -> "2122 reshape_92_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"2122 reshape_92_0_0_nncf_smooth_quant_0" -> "2123 quantize_per_tensor_default_126"  [label="(4, 64, 384)", style=solid];
-"2123 quantize_per_tensor_default_126" -> "2124 dequantize_per_tensor_default_126"  [label="(4, 64, 384)", style=solid];
-"2124 dequantize_per_tensor_default_126" -> "2131 linear_125"  [label="(4, 64, 384)", style=solid];
-"2125 _param_constant337_0_0" -> "2131 linear_125"  [label="(384,)", style=solid];
-"2126 scale_updated_constant84" -> "2128 mul_tensor_84"  [label="(384, 1)", style=solid];
-"2127 compressed_weight_updated_constant84" -> "2128 mul_tensor_84"  [label="(384, 384)", style=solid];
-"2128 mul_tensor_84" -> "2130 sub_tensor_84"  [label="(384, 384)", style=solid];
-"2129 zero_point_updated_constant84" -> "2130 sub_tensor_84"  [label="(384, 1)", style=solid];
-"2130 sub_tensor_84" -> "2131 linear_125"  [label="(384, 384)", style=solid];
-"2131 linear_125" -> "2132 dropout_81"  [label="(4, 64, 384)", style=solid];
-"2132 dropout_81" -> "2133 view_113"  [label="(4, 64, 384)", style=solid];
-"2133 view_113" -> "2134 permute_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2134 permute_94" -> "2135 reshape_93"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2135 reshape_93" -> "2136 slice_302"  [label="(1, 16, 16, 384)", style=solid];
-"2136 slice_302" -> "2137 slice_303"  [label="(1, 16, 16, 384)", style=solid];
-"2137 slice_303" -> "2138 slice_304"  [label="(1, 14, 16, 384)", style=solid];
-"2138 slice_304" -> "2139 slice_305"  [label="(1, 14, 14, 384)", style=solid];
-"2139 slice_305" -> "2140 contiguous_39"  [label="(1, 14, 14, 384)", style=solid];
-"2140 contiguous_39" -> "2143 layer_norm_43"  [label="(1, 14, 14, 384)", style=solid];
-"2141 _param_constant338" -> "2143 layer_norm_43"  [label="(384,)", style=solid];
-"2142 _param_constant339" -> "2143 layer_norm_43"  [label="(384,)", style=solid];
-"2143 layer_norm_43" -> "2144 add_71"  [label="(1, 14, 14, 384)", style=solid];
-"2144 add_71" -> "2145 add_71_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"2144 add_71" -> "2171 add_72"  [label="(1, 14, 14, 384)", style=solid];
-"2145 add_71_0_0_nncf_smooth_quant_0" -> "2146 quantize_per_tensor_default_127"  [label="(1, 14, 14, 384)", style=solid];
-"2146 quantize_per_tensor_default_127" -> "2147 dequantize_per_tensor_default_127"  [label="(1, 14, 14, 384)", style=solid];
-"2147 dequantize_per_tensor_default_127" -> "2154 linear_126"  [label="(1, 14, 14, 384)", style=solid];
-"2148 _param_constant341_0_0" -> "2154 linear_126"  [label="(1536,)", style=solid];
-"2149 scale_updated_constant85" -> "2151 mul_tensor_85"  [label="(1536, 1)", style=solid];
-"2150 compressed_weight_updated_constant85" -> "2151 mul_tensor_85"  [label="(1536, 384)", style=solid];
-"2151 mul_tensor_85" -> "2153 sub_tensor_85"  [label="(1536, 384)", style=solid];
-"2152 zero_point_updated_constant85" -> "2153 sub_tensor_85"  [label="(1536, 1)", style=solid];
-"2153 sub_tensor_85" -> "2154 linear_126"  [label="(1536, 384)", style=solid];
-"2154 linear_126" -> "2155 gelu_20"  [label="(1, 14, 14, 1536)", style=solid];
-"2155 gelu_20" -> "2156 dropout_82"  [label="(1, 14, 14, 1536)", style=solid];
-"2156 dropout_82" -> "2157 dropout_82_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"2157 dropout_82_0_0_nncf_smooth_quant_0" -> "2158 quantize_per_tensor_default_128"  [label="(1, 14, 14, 1536)", style=solid];
-"2158 quantize_per_tensor_default_128" -> "2159 dequantize_per_tensor_default_128"  [label="(1, 14, 14, 1536)", style=solid];
-"2159 dequantize_per_tensor_default_128" -> "2166 linear_127"  [label="(1, 14, 14, 1536)", style=solid];
-"2160 _param_constant343_0_0" -> "2166 linear_127"  [label="(384,)", style=solid];
-"2161 scale_updated_constant86" -> "2163 mul_tensor_86"  [label="(384, 1)", style=solid];
-"2162 compressed_weight_updated_constant86" -> "2163 mul_tensor_86"  [label="(384, 1536)", style=solid];
-"2163 mul_tensor_86" -> "2165 sub_tensor_86"  [label="(384, 1536)", style=solid];
-"2164 zero_point_updated_constant86" -> "2165 sub_tensor_86"  [label="(384, 1)", style=solid];
-"2165 sub_tensor_86" -> "2166 linear_127"  [label="(384, 1536)", style=solid];
-"2166 linear_127" -> "2167 dropout_83"  [label="(1, 14, 14, 384)", style=solid];
-"2167 dropout_83" -> "2170 layer_norm_44"  [label="(1, 14, 14, 384)", style=solid];
-"2168 _param_constant344" -> "2170 layer_norm_44"  [label="(384,)", style=solid];
-"2169 _param_constant345" -> "2170 layer_norm_44"  [label="(384,)", style=solid];
-"2170 layer_norm_44" -> "2171 add_72"  [label="(1, 14, 14, 384)", style=solid];
-"2171 add_72" -> "2173 pad_23"  [label="(1, 14, 14, 384)", style=solid];
-"2171 add_72" -> "2254 add_75"  [label="(1, 14, 14, 384)", style=solid];
-"2172 _frozen_param174" -> "2209 add_73"  [label="(1, 12, 64, 64)", style=solid];
-"2173 pad_23" -> "2174 roll_20"  [label="(1, 16, 16, 384)", style=solid];
-"2174 roll_20" -> "2175 view_116"  [label="(1, 16, 16, 384)", style=solid];
-"2175 view_116" -> "2176 permute_96"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2176 permute_96" -> "2177 reshape_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2177 reshape_94" -> "2178 reshape_94_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"2177 reshape_94" -> "2210 new_zeros_10"  [label="(4, 64, 384)", style=solid];
-"2178 reshape_94_0_0_nncf_smooth_quant_0" -> "2179 quantize_per_tensor_default_129"  [label="(4, 64, 384)", style=solid];
-"2179 quantize_per_tensor_default_129" -> "2180 dequantize_per_tensor_default_129"  [label="(4, 64, 384)", style=solid];
-"2180 dequantize_per_tensor_default_129" -> "2187 linear_130"  [label="(4, 64, 384)", style=solid];
-"2181 _frozen_param175_0_0" -> "2187 linear_130"  [label="(1152,)", style=solid];
-"2182 scale_updated_constant87" -> "2184 mul_tensor_87"  [label="(1152, 1)", style=solid];
-"2183 compressed_weight_updated_constant87" -> "2184 mul_tensor_87"  [label="(1152, 384)", style=solid];
-"2184 mul_tensor_87" -> "2186 sub_tensor_87"  [label="(1152, 384)", style=solid];
-"2185 zero_point_updated_constant87" -> "2186 sub_tensor_87"  [label="(1152, 1)", style=solid];
-"2186 sub_tensor_87" -> "2187 linear_130"  [label="(1152, 384)", style=solid];
-"2187 linear_130" -> "2188 reshape_95"  [label="(4, 64, 1152)", style=solid];
-"2188 reshape_95" -> "2189 permute_97"  [label="(4, 64, 3, 12, 32)", style=solid];
-"2189 permute_97" -> "2190 select_63"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2189 permute_97" -> "2191 select_64"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2189 permute_97" -> "2192 select_65"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2190 select_63" -> "2193 linalg_vector_norm_42"  [label="(4, 12, 64, 32)", style=solid];
-"2190 select_63" -> "2195 expand_as_42"  [label="(4, 12, 64, 32)", style=solid];
-"2190 select_63" -> "2196 div_42"  [label="(4, 12, 64, 32)", style=solid];
-"2191 select_64" -> "2199 linalg_vector_norm_43"  [label="(4, 12, 64, 32)", style=solid];
-"2191 select_64" -> "2201 expand_as_43"  [label="(4, 12, 64, 32)", style=solid];
-"2191 select_64" -> "2202 div_43"  [label="(4, 12, 64, 32)", style=solid];
-"2192 select_65" -> "2228 matmul_43"  [label="(4, 12, 64, 32)", style=solid];
-"2193 linalg_vector_norm_42" -> "2194 clamp_min_42"  [label="(4, 12, 64, 1)", style=solid];
-"2194 clamp_min_42" -> "2195 expand_as_42"  [label="(4, 12, 64, 1)", style=solid];
-"2195 expand_as_42" -> "2196 div_42"  [label="(4, 12, 64, 32)", style=solid];
-"2196 div_42" -> "2197 quantize_per_tensor_default_130"  [label="(4, 12, 64, 32)", style=solid];
-"2197 quantize_per_tensor_default_130" -> "2198 dequantize_per_tensor_default_130"  [label="(4, 12, 64, 32)", style=solid];
-"2198 dequantize_per_tensor_default_130" -> "2206 matmul_42"  [label="(4, 12, 64, 32)", style=solid];
-"2199 linalg_vector_norm_43" -> "2200 clamp_min_43"  [label="(4, 12, 64, 1)", style=solid];
-"2200 clamp_min_43" -> "2201 expand_as_43"  [label="(4, 12, 64, 1)", style=solid];
-"2201 expand_as_43" -> "2202 div_43"  [label="(4, 12, 64, 32)", style=solid];
-"2202 div_43" -> "2203 quantize_per_tensor_default_131"  [label="(4, 12, 64, 32)", style=solid];
-"2203 quantize_per_tensor_default_131" -> "2204 dequantize_per_tensor_default_131"  [label="(4, 12, 64, 32)", style=solid];
-"2204 dequantize_per_tensor_default_131" -> "2205 transpose_42"  [label="(4, 12, 64, 32)", style=solid];
-"2205 transpose_42" -> "2206 matmul_42"  [label="(4, 12, 32, 64)", style=solid];
-"2206 matmul_42" -> "2208 mul_43"  [label="(4, 12, 64, 64)", style=solid];
-"2207 _frozen_param177" -> "2208 mul_43"  [label="(12, 1, 1)", style=solid];
-"2208 mul_43" -> "2209 add_73"  [label="(4, 12, 64, 64)", style=solid];
-"2209 add_73" -> "2221 view_118"  [label="(4, 12, 64, 64)", style=solid];
-"2210 new_zeros_10" -> "2211 view_117"  [label="(16, 16)", style=solid];
-"2211 view_117" -> "2212 permute_98"  [label="(2, 8, 2, 8)", style=solid];
-"2212 permute_98" -> "2213 reshape_96"  [label="(2, 2, 8, 8)", style=solid];
-"2213 reshape_96" -> "2214 unsqueeze_62"  [label="(4, 64)", style=solid];
-"2213 reshape_96" -> "2215 unsqueeze_63"  [label="(4, 64)", style=solid];
-"2214 unsqueeze_62" -> "2216 sub_10"  [label="(4, 1, 64)", style=solid];
-"2215 unsqueeze_63" -> "2216 sub_10"  [label="(4, 64, 1)", style=solid];
-"2216 sub_10" -> "2217 ne_10"  [label="(4, 64, 64)", style=solid];
-"2216 sub_10" -> "2218 masked_fill_20"  [label="(4, 64, 64)", style=solid];
-"2216 sub_10" -> "2219 eq_10"  [label="(4, 64, 64)", style=solid];
-"2217 ne_10" -> "2218 masked_fill_20"  [label="(4, 64, 64)", style=solid];
-"2218 masked_fill_20" -> "2220 masked_fill_21"  [label="(4, 64, 64)", style=solid];
-"2219 eq_10" -> "2220 masked_fill_21"  [label="(4, 64, 64)", style=solid];
-"2220 masked_fill_21" -> "2222 unsqueeze_64"  [label="(4, 64, 64)", style=solid];
-"2221 view_118" -> "2224 add_74"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2222 unsqueeze_64" -> "2223 unsqueeze_65"  [label="(4, 1, 64, 64)", style=solid];
-"2223 unsqueeze_65" -> "2224 add_74"  [label="(1, 4, 1, 64, 64)", style=solid];
-"2224 add_74" -> "2225 view_119"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2225 view_119" -> "2226 softmax_21"  [label="(4, 12, 64, 64)", style=solid];
-"2226 softmax_21" -> "2227 dropout_84"  [label="(4, 12, 64, 64)", style=solid];
-"2227 dropout_84" -> "2228 matmul_43"  [label="(4, 12, 64, 64)", style=solid];
-"2228 matmul_43" -> "2229 transpose_43"  [label="(4, 12, 64, 32)", style=solid];
-"2229 transpose_43" -> "2230 reshape_97"  [label="(4, 64, 12, 32)", style=solid];
-"2230 reshape_97" -> "2231 reshape_97_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"2231 reshape_97_0_0_nncf_smooth_quant_0" -> "2232 quantize_per_tensor_default_132"  [label="(4, 64, 384)", style=solid];
-"2232 quantize_per_tensor_default_132" -> "2233 dequantize_per_tensor_default_132"  [label="(4, 64, 384)", style=solid];
-"2233 dequantize_per_tensor_default_132" -> "2240 linear_131"  [label="(4, 64, 384)", style=solid];
-"2234 _param_constant353_0_0" -> "2240 linear_131"  [label="(384,)", style=solid];
-"2235 scale_updated_constant88" -> "2237 mul_tensor_88"  [label="(384, 1)", style=solid];
-"2236 compressed_weight_updated_constant88" -> "2237 mul_tensor_88"  [label="(384, 384)", style=solid];
-"2237 mul_tensor_88" -> "2239 sub_tensor_88"  [label="(384, 384)", style=solid];
-"2238 zero_point_updated_constant88" -> "2239 sub_tensor_88"  [label="(384, 1)", style=solid];
-"2239 sub_tensor_88" -> "2240 linear_131"  [label="(384, 384)", style=solid];
-"2240 linear_131" -> "2241 dropout_85"  [label="(4, 64, 384)", style=solid];
-"2241 dropout_85" -> "2242 view_120"  [label="(4, 64, 384)", style=solid];
-"2242 view_120" -> "2243 permute_99"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2243 permute_99" -> "2244 reshape_98"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2244 reshape_98" -> "2245 roll_21"  [label="(1, 16, 16, 384)", style=solid];
-"2245 roll_21" -> "2246 slice_325"  [label="(1, 16, 16, 384)", style=solid];
-"2246 slice_325" -> "2247 slice_326"  [label="(1, 16, 16, 384)", style=solid];
-"2247 slice_326" -> "2248 slice_327"  [label="(1, 14, 16, 384)", style=solid];
-"2248 slice_327" -> "2249 slice_328"  [label="(1, 14, 14, 384)", style=solid];
-"2249 slice_328" -> "2250 contiguous_41"  [label="(1, 14, 14, 384)", style=solid];
-"2250 contiguous_41" -> "2253 layer_norm_45"  [label="(1, 14, 14, 384)", style=solid];
-"2251 _param_constant354" -> "2253 layer_norm_45"  [label="(384,)", style=solid];
-"2252 _param_constant355" -> "2253 layer_norm_45"  [label="(384,)", style=solid];
-"2253 layer_norm_45" -> "2254 add_75"  [label="(1, 14, 14, 384)", style=solid];
-"2254 add_75" -> "2255 add_75_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"2254 add_75" -> "2281 add_76"  [label="(1, 14, 14, 384)", style=solid];
-"2255 add_75_0_0_nncf_smooth_quant_0" -> "2256 quantize_per_tensor_default_133"  [label="(1, 14, 14, 384)", style=solid];
-"2256 quantize_per_tensor_default_133" -> "2257 dequantize_per_tensor_default_133"  [label="(1, 14, 14, 384)", style=solid];
-"2257 dequantize_per_tensor_default_133" -> "2264 linear_132"  [label="(1, 14, 14, 384)", style=solid];
-"2258 _param_constant357_0_0" -> "2264 linear_132"  [label="(1536,)", style=solid];
-"2259 scale_updated_constant89" -> "2261 mul_tensor_89"  [label="(1536, 1)", style=solid];
-"2260 compressed_weight_updated_constant89" -> "2261 mul_tensor_89"  [label="(1536, 384)", style=solid];
-"2261 mul_tensor_89" -> "2263 sub_tensor_89"  [label="(1536, 384)", style=solid];
-"2262 zero_point_updated_constant89" -> "2263 sub_tensor_89"  [label="(1536, 1)", style=solid];
-"2263 sub_tensor_89" -> "2264 linear_132"  [label="(1536, 384)", style=solid];
-"2264 linear_132" -> "2265 gelu_21"  [label="(1, 14, 14, 1536)", style=solid];
-"2265 gelu_21" -> "2266 dropout_86"  [label="(1, 14, 14, 1536)", style=solid];
-"2266 dropout_86" -> "2267 dropout_86_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"2267 dropout_86_0_0_nncf_smooth_quant_0" -> "2268 quantize_per_tensor_default_134"  [label="(1, 14, 14, 1536)", style=solid];
-"2268 quantize_per_tensor_default_134" -> "2269 dequantize_per_tensor_default_134"  [label="(1, 14, 14, 1536)", style=solid];
-"2269 dequantize_per_tensor_default_134" -> "2276 linear_133"  [label="(1, 14, 14, 1536)", style=solid];
-"2270 _param_constant359_0_0" -> "2276 linear_133"  [label="(384,)", style=solid];
-"2271 scale_updated_constant90" -> "2273 mul_tensor_90"  [label="(384, 1)", style=solid];
-"2272 compressed_weight_updated_constant90" -> "2273 mul_tensor_90"  [label="(384, 1536)", style=solid];
-"2273 mul_tensor_90" -> "2275 sub_tensor_90"  [label="(384, 1536)", style=solid];
-"2274 zero_point_updated_constant90" -> "2275 sub_tensor_90"  [label="(384, 1)", style=solid];
-"2275 sub_tensor_90" -> "2276 linear_133"  [label="(384, 1536)", style=solid];
-"2276 linear_133" -> "2277 dropout_87"  [label="(1, 14, 14, 384)", style=solid];
-"2277 dropout_87" -> "2280 layer_norm_46"  [label="(1, 14, 14, 384)", style=solid];
-"2278 _param_constant360" -> "2280 layer_norm_46"  [label="(384,)", style=solid];
-"2279 _param_constant361" -> "2280 layer_norm_46"  [label="(384,)", style=solid];
-"2280 layer_norm_46" -> "2281 add_76"  [label="(1, 14, 14, 384)", style=solid];
-"2281 add_76" -> "2282 pad_24"  [label="(1, 14, 14, 384)", style=solid];
-"2282 pad_24" -> "2283 slice_329"  [label="(1, 14, 14, 384)", style=solid];
-"2282 pad_24" -> "2286 slice_332"  [label="(1, 14, 14, 384)", style=solid];
-"2282 pad_24" -> "2289 slice_335"  [label="(1, 14, 14, 384)", style=solid];
-"2282 pad_24" -> "2292 slice_338"  [label="(1, 14, 14, 384)", style=solid];
-"2283 slice_329" -> "2284 slice_330"  [label="(1, 7, 14, 384)", style=solid];
-"2284 slice_330" -> "2285 slice_331"  [label="(1, 7, 7, 384)", style=solid];
-"2285 slice_331" -> "2295 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2286 slice_332" -> "2287 slice_333"  [label="(1, 7, 14, 384)", style=solid];
-"2287 slice_333" -> "2288 slice_334"  [label="(1, 7, 7, 384)", style=solid];
-"2288 slice_334" -> "2295 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2289 slice_335" -> "2290 slice_336"  [label="(1, 7, 14, 384)", style=solid];
-"2290 slice_336" -> "2291 slice_337"  [label="(1, 7, 7, 384)", style=solid];
-"2291 slice_337" -> "2295 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2292 slice_338" -> "2293 slice_339"  [label="(1, 7, 14, 384)", style=solid];
-"2293 slice_339" -> "2294 slice_340"  [label="(1, 7, 7, 384)", style=solid];
-"2294 slice_340" -> "2295 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2295 cat_2" -> "2296 cat_2_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 1536)", style=solid];
-"2296 cat_2_0_0_nncf_smooth_quant_0" -> "2297 quantize_per_tensor_default_135"  [label="(1, 7, 7, 1536)", style=solid];
-"2297 quantize_per_tensor_default_135" -> "2298 dequantize_per_tensor_default_135"  [label="(1, 7, 7, 1536)", style=solid];
-"2298 dequantize_per_tensor_default_135" -> "2304 linear_134"  [label="(1, 7, 7, 1536)", style=solid];
-"2299 scale_updated_constant91" -> "2301 mul_tensor_91"  [label="(768, 1)", style=solid];
-"2300 compressed_weight_updated_constant91" -> "2301 mul_tensor_91"  [label="(768, 1536)", style=solid];
-"2301 mul_tensor_91" -> "2303 sub_tensor_91"  [label="(768, 1536)", style=solid];
-"2302 zero_point_updated_constant91" -> "2303 sub_tensor_91"  [label="(768, 1)", style=solid];
-"2303 sub_tensor_91" -> "2304 linear_134"  [label="(768, 1536)", style=solid];
-"2304 linear_134" -> "2307 layer_norm_47"  [label="(1, 7, 7, 768)", style=solid];
-"2305 _param_constant363" -> "2307 layer_norm_47"  [label="(768,)", style=solid];
-"2306 _param_constant364" -> "2307 layer_norm_47"  [label="(768,)", style=solid];
-"2307 layer_norm_47" -> "2309 pad_25"  [label="(1, 7, 7, 768)", style=solid];
-"2307 layer_norm_47" -> "2372 add_78"  [label="(1, 7, 7, 768)", style=solid];
-"2308 _frozen_param187" -> "2344 add_77"  [label="(1, 24, 64, 64)", style=solid];
-"2309 pad_25" -> "2310 view_123"  [label="(1, 8, 8, 768)", style=solid];
-"2310 view_123" -> "2311 permute_101"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2311 permute_101" -> "2312 reshape_99"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2312 reshape_99" -> "2313 reshape_99_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
-"2313 reshape_99_0_0_nncf_smooth_quant_0" -> "2314 quantize_per_tensor_default_136"  [label="(1, 64, 768)", style=solid];
-"2314 quantize_per_tensor_default_136" -> "2315 dequantize_per_tensor_default_136"  [label="(1, 64, 768)", style=solid];
-"2315 dequantize_per_tensor_default_136" -> "2322 linear_137"  [label="(1, 64, 768)", style=solid];
-"2316 _frozen_param188_0_0" -> "2322 linear_137"  [label="(2304,)", style=solid];
-"2317 scale_updated_constant92" -> "2319 mul_tensor_92"  [label="(2304, 1)", style=solid];
-"2318 compressed_weight_updated_constant92" -> "2319 mul_tensor_92"  [label="(2304, 768)", style=solid];
-"2319 mul_tensor_92" -> "2321 sub_tensor_92"  [label="(2304, 768)", style=solid];
-"2320 zero_point_updated_constant92" -> "2321 sub_tensor_92"  [label="(2304, 1)", style=solid];
-"2321 sub_tensor_92" -> "2322 linear_137"  [label="(2304, 768)", style=solid];
-"2322 linear_137" -> "2323 reshape_100"  [label="(1, 64, 2304)", style=solid];
-"2323 reshape_100" -> "2324 permute_102"  [label="(1, 64, 3, 24, 32)", style=solid];
-"2324 permute_102" -> "2325 select_66"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2324 permute_102" -> "2326 select_67"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2324 permute_102" -> "2327 select_68"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2325 select_66" -> "2328 linalg_vector_norm_44"  [label="(1, 24, 64, 32)", style=solid];
-"2325 select_66" -> "2330 expand_as_44"  [label="(1, 24, 64, 32)", style=solid];
-"2325 select_66" -> "2331 div_44"  [label="(1, 24, 64, 32)", style=solid];
-"2326 select_67" -> "2334 linalg_vector_norm_45"  [label="(1, 24, 64, 32)", style=solid];
-"2326 select_67" -> "2336 expand_as_45"  [label="(1, 24, 64, 32)", style=solid];
-"2326 select_67" -> "2337 div_45"  [label="(1, 24, 64, 32)", style=solid];
-"2327 select_68" -> "2347 matmul_45"  [label="(1, 24, 64, 32)", style=solid];
-"2328 linalg_vector_norm_44" -> "2329 clamp_min_44"  [label="(1, 24, 64, 1)", style=solid];
-"2329 clamp_min_44" -> "2330 expand_as_44"  [label="(1, 24, 64, 1)", style=solid];
-"2330 expand_as_44" -> "2331 div_44"  [label="(1, 24, 64, 32)", style=solid];
-"2331 div_44" -> "2332 quantize_per_tensor_default_137"  [label="(1, 24, 64, 32)", style=solid];
-"2332 quantize_per_tensor_default_137" -> "2333 dequantize_per_tensor_default_137"  [label="(1, 24, 64, 32)", style=solid];
-"2333 dequantize_per_tensor_default_137" -> "2341 matmul_44"  [label="(1, 24, 64, 32)", style=solid];
-"2334 linalg_vector_norm_45" -> "2335 clamp_min_45"  [label="(1, 24, 64, 1)", style=solid];
-"2335 clamp_min_45" -> "2336 expand_as_45"  [label="(1, 24, 64, 1)", style=solid];
-"2336 expand_as_45" -> "2337 div_45"  [label="(1, 24, 64, 32)", style=solid];
-"2337 div_45" -> "2338 quantize_per_tensor_default_138"  [label="(1, 24, 64, 32)", style=solid];
-"2338 quantize_per_tensor_default_138" -> "2339 dequantize_per_tensor_default_138"  [label="(1, 24, 64, 32)", style=solid];
-"2339 dequantize_per_tensor_default_138" -> "2340 transpose_44"  [label="(1, 24, 64, 32)", style=solid];
-"2340 transpose_44" -> "2341 matmul_44"  [label="(1, 24, 32, 64)", style=solid];
-"2341 matmul_44" -> "2343 mul_45"  [label="(1, 24, 64, 64)", style=solid];
-"2342 _frozen_param190" -> "2343 mul_45"  [label="(24, 1, 1)", style=solid];
-"2343 mul_45" -> "2344 add_77"  [label="(1, 24, 64, 64)", style=solid];
-"2344 add_77" -> "2345 softmax_22"  [label="(1, 24, 64, 64)", style=solid];
-"2345 softmax_22" -> "2346 dropout_88"  [label="(1, 24, 64, 64)", style=solid];
-"2346 dropout_88" -> "2347 matmul_45"  [label="(1, 24, 64, 64)", style=solid];
-"2347 matmul_45" -> "2348 transpose_45"  [label="(1, 24, 64, 32)", style=solid];
-"2348 transpose_45" -> "2349 reshape_101"  [label="(1, 64, 24, 32)", style=solid];
-"2349 reshape_101" -> "2350 reshape_101_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
-"2350 reshape_101_0_0_nncf_smooth_quant_0" -> "2351 quantize_per_tensor_default_139"  [label="(1, 64, 768)", style=solid];
-"2351 quantize_per_tensor_default_139" -> "2352 dequantize_per_tensor_default_139"  [label="(1, 64, 768)", style=solid];
-"2352 dequantize_per_tensor_default_139" -> "2359 linear_138"  [label="(1, 64, 768)", style=solid];
-"2353 _param_constant372_0_0" -> "2359 linear_138"  [label="(768,)", style=solid];
-"2354 scale_updated_constant93" -> "2356 mul_tensor_93"  [label="(768, 1)", style=solid];
-"2355 compressed_weight_updated_constant93" -> "2356 mul_tensor_93"  [label="(768, 768)", style=solid];
-"2356 mul_tensor_93" -> "2358 sub_tensor_93"  [label="(768, 768)", style=solid];
-"2357 zero_point_updated_constant93" -> "2358 sub_tensor_93"  [label="(768, 1)", style=solid];
-"2358 sub_tensor_93" -> "2359 linear_138"  [label="(768, 768)", style=solid];
-"2359 linear_138" -> "2360 dropout_89"  [label="(1, 64, 768)", style=solid];
-"2360 dropout_89" -> "2361 view_124"  [label="(1, 64, 768)", style=solid];
-"2361 view_124" -> "2362 permute_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2362 permute_103" -> "2363 reshape_102"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2363 reshape_102" -> "2364 slice_342"  [label="(1, 8, 8, 768)", style=solid];
-"2364 slice_342" -> "2365 slice_343"  [label="(1, 8, 8, 768)", style=solid];
-"2365 slice_343" -> "2366 slice_344"  [label="(1, 7, 8, 768)", style=solid];
-"2366 slice_344" -> "2367 slice_345"  [label="(1, 7, 7, 768)", style=solid];
-"2367 slice_345" -> "2368 contiguous_43"  [label="(1, 7, 7, 768)", style=solid];
-"2368 contiguous_43" -> "2371 layer_norm_48"  [label="(1, 7, 7, 768)", style=solid];
-"2369 _param_constant373" -> "2371 layer_norm_48"  [label="(768,)", style=solid];
-"2370 _param_constant374" -> "2371 layer_norm_48"  [label="(768,)", style=solid];
-"2371 layer_norm_48" -> "2372 add_78"  [label="(1, 7, 7, 768)", style=solid];
-"2372 add_78" -> "2373 add_78_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 768)", style=solid];
-"2372 add_78" -> "2399 add_79"  [label="(1, 7, 7, 768)", style=solid];
-"2373 add_78_0_0_nncf_smooth_quant_0" -> "2374 quantize_per_tensor_default_140"  [label="(1, 7, 7, 768)", style=solid];
-"2374 quantize_per_tensor_default_140" -> "2375 dequantize_per_tensor_default_140"  [label="(1, 7, 7, 768)", style=solid];
-"2375 dequantize_per_tensor_default_140" -> "2382 linear_139"  [label="(1, 7, 7, 768)", style=solid];
-"2376 _param_constant376_0_0" -> "2382 linear_139"  [label="(3072,)", style=solid];
-"2377 scale_updated_constant94" -> "2379 mul_tensor_94"  [label="(3072, 1)", style=solid];
-"2378 compressed_weight_updated_constant94" -> "2379 mul_tensor_94"  [label="(3072, 768)", style=solid];
-"2379 mul_tensor_94" -> "2381 sub_tensor_94"  [label="(3072, 768)", style=solid];
-"2380 zero_point_updated_constant94" -> "2381 sub_tensor_94"  [label="(3072, 1)", style=solid];
-"2381 sub_tensor_94" -> "2382 linear_139"  [label="(3072, 768)", style=solid];
-"2382 linear_139" -> "2383 gelu_22"  [label="(1, 7, 7, 3072)", style=solid];
-"2383 gelu_22" -> "2384 dropout_90"  [label="(1, 7, 7, 3072)", style=solid];
-"2384 dropout_90" -> "2385 dropout_90_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 3072)", style=solid];
-"2385 dropout_90_0_0_nncf_smooth_quant_0" -> "2386 quantize_per_tensor_default_141"  [label="(1, 7, 7, 3072)", style=solid];
-"2386 quantize_per_tensor_default_141" -> "2387 dequantize_per_tensor_default_141"  [label="(1, 7, 7, 3072)", style=solid];
-"2387 dequantize_per_tensor_default_141" -> "2394 linear_140"  [label="(1, 7, 7, 3072)", style=solid];
-"2388 _param_constant378_0_0" -> "2394 linear_140"  [label="(768,)", style=solid];
-"2389 scale_updated_constant95" -> "2391 mul_tensor_95"  [label="(768, 1)", style=solid];
-"2390 compressed_weight_updated_constant95" -> "2391 mul_tensor_95"  [label="(768, 3072)", style=solid];
-"2391 mul_tensor_95" -> "2393 sub_tensor_95"  [label="(768, 3072)", style=solid];
-"2392 zero_point_updated_constant95" -> "2393 sub_tensor_95"  [label="(768, 1)", style=solid];
-"2393 sub_tensor_95" -> "2394 linear_140"  [label="(768, 3072)", style=solid];
-"2394 linear_140" -> "2395 dropout_91"  [label="(1, 7, 7, 768)", style=solid];
-"2395 dropout_91" -> "2398 layer_norm_49"  [label="(1, 7, 7, 768)", style=solid];
-"2396 _param_constant379" -> "2398 layer_norm_49"  [label="(768,)", style=solid];
-"2397 _param_constant380" -> "2398 layer_norm_49"  [label="(768,)", style=solid];
-"2398 layer_norm_49" -> "2399 add_79"  [label="(1, 7, 7, 768)", style=solid];
-"2399 add_79" -> "2401 pad_26"  [label="(1, 7, 7, 768)", style=solid];
-"2399 add_79" -> "2464 add_81"  [label="(1, 7, 7, 768)", style=solid];
-"2400 _frozen_param191" -> "2436 add_80"  [label="(1, 24, 64, 64)", style=solid];
-"2401 pad_26" -> "2402 view_127"  [label="(1, 8, 8, 768)", style=solid];
-"2402 view_127" -> "2403 permute_105"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2403 permute_105" -> "2404 reshape_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2404 reshape_103" -> "2405 reshape_103_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
-"2405 reshape_103_0_0_nncf_smooth_quant_0" -> "2406 quantize_per_tensor_default_142"  [label="(1, 64, 768)", style=solid];
-"2406 quantize_per_tensor_default_142" -> "2407 dequantize_per_tensor_default_142"  [label="(1, 64, 768)", style=solid];
-"2407 dequantize_per_tensor_default_142" -> "2414 linear_143"  [label="(1, 64, 768)", style=solid];
-"2408 _frozen_param192_0_0" -> "2414 linear_143"  [label="(2304,)", style=solid];
-"2409 scale_updated_constant96" -> "2411 mul_tensor_96"  [label="(2304, 1)", style=solid];
-"2410 compressed_weight_updated_constant96" -> "2411 mul_tensor_96"  [label="(2304, 768)", style=solid];
-"2411 mul_tensor_96" -> "2413 sub_tensor_96"  [label="(2304, 768)", style=solid];
-"2412 zero_point_updated_constant96" -> "2413 sub_tensor_96"  [label="(2304, 1)", style=solid];
-"2413 sub_tensor_96" -> "2414 linear_143"  [label="(2304, 768)", style=solid];
-"2414 linear_143" -> "2415 reshape_104"  [label="(1, 64, 2304)", style=solid];
-"2415 reshape_104" -> "2416 permute_106"  [label="(1, 64, 3, 24, 32)", style=solid];
-"2416 permute_106" -> "2417 select_69"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2416 permute_106" -> "2418 select_70"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2416 permute_106" -> "2419 select_71"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2417 select_69" -> "2420 linalg_vector_norm_46"  [label="(1, 24, 64, 32)", style=solid];
-"2417 select_69" -> "2422 expand_as_46"  [label="(1, 24, 64, 32)", style=solid];
-"2417 select_69" -> "2423 div_46"  [label="(1, 24, 64, 32)", style=solid];
-"2418 select_70" -> "2426 linalg_vector_norm_47"  [label="(1, 24, 64, 32)", style=solid];
-"2418 select_70" -> "2428 expand_as_47"  [label="(1, 24, 64, 32)", style=solid];
-"2418 select_70" -> "2429 div_47"  [label="(1, 24, 64, 32)", style=solid];
-"2419 select_71" -> "2439 matmul_47"  [label="(1, 24, 64, 32)", style=solid];
-"2420 linalg_vector_norm_46" -> "2421 clamp_min_46"  [label="(1, 24, 64, 1)", style=solid];
-"2421 clamp_min_46" -> "2422 expand_as_46"  [label="(1, 24, 64, 1)", style=solid];
-"2422 expand_as_46" -> "2423 div_46"  [label="(1, 24, 64, 32)", style=solid];
-"2423 div_46" -> "2424 quantize_per_tensor_default_143"  [label="(1, 24, 64, 32)", style=solid];
-"2424 quantize_per_tensor_default_143" -> "2425 dequantize_per_tensor_default_143"  [label="(1, 24, 64, 32)", style=solid];
-"2425 dequantize_per_tensor_default_143" -> "2433 matmul_46"  [label="(1, 24, 64, 32)", style=solid];
-"2426 linalg_vector_norm_47" -> "2427 clamp_min_47"  [label="(1, 24, 64, 1)", style=solid];
-"2427 clamp_min_47" -> "2428 expand_as_47"  [label="(1, 24, 64, 1)", style=solid];
-"2428 expand_as_47" -> "2429 div_47"  [label="(1, 24, 64, 32)", style=solid];
-"2429 div_47" -> "2430 quantize_per_tensor_default_144"  [label="(1, 24, 64, 32)", style=solid];
-"2430 quantize_per_tensor_default_144" -> "2431 dequantize_per_tensor_default_144"  [label="(1, 24, 64, 32)", style=solid];
-"2431 dequantize_per_tensor_default_144" -> "2432 transpose_46"  [label="(1, 24, 64, 32)", style=solid];
-"2432 transpose_46" -> "2433 matmul_46"  [label="(1, 24, 32, 64)", style=solid];
-"2433 matmul_46" -> "2435 mul_47"  [label="(1, 24, 64, 64)", style=solid];
-"2434 _frozen_param194" -> "2435 mul_47"  [label="(24, 1, 1)", style=solid];
-"2435 mul_47" -> "2436 add_80"  [label="(1, 24, 64, 64)", style=solid];
-"2436 add_80" -> "2437 softmax_23"  [label="(1, 24, 64, 64)", style=solid];
-"2437 softmax_23" -> "2438 dropout_92"  [label="(1, 24, 64, 64)", style=solid];
-"2438 dropout_92" -> "2439 matmul_47"  [label="(1, 24, 64, 64)", style=solid];
-"2439 matmul_47" -> "2440 transpose_47"  [label="(1, 24, 64, 32)", style=solid];
-"2440 transpose_47" -> "2441 reshape_105"  [label="(1, 64, 24, 32)", style=solid];
-"2441 reshape_105" -> "2442 reshape_105_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
-"2442 reshape_105_0_0_nncf_smooth_quant_0" -> "2443 quantize_per_tensor_default_145"  [label="(1, 64, 768)", style=solid];
-"2443 quantize_per_tensor_default_145" -> "2444 dequantize_per_tensor_default_145"  [label="(1, 64, 768)", style=solid];
-"2444 dequantize_per_tensor_default_145" -> "2451 linear_144"  [label="(1, 64, 768)", style=solid];
-"2445 _param_constant388_0_0" -> "2451 linear_144"  [label="(768,)", style=solid];
-"2446 scale_updated_constant97" -> "2448 mul_tensor_97"  [label="(768, 1)", style=solid];
-"2447 compressed_weight_updated_constant97" -> "2448 mul_tensor_97"  [label="(768, 768)", style=solid];
-"2448 mul_tensor_97" -> "2450 sub_tensor_97"  [label="(768, 768)", style=solid];
-"2449 zero_point_updated_constant97" -> "2450 sub_tensor_97"  [label="(768, 1)", style=solid];
-"2450 sub_tensor_97" -> "2451 linear_144"  [label="(768, 768)", style=solid];
-"2451 linear_144" -> "2452 dropout_93"  [label="(1, 64, 768)", style=solid];
-"2452 dropout_93" -> "2453 view_128"  [label="(1, 64, 768)", style=solid];
-"2453 view_128" -> "2454 permute_107"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2454 permute_107" -> "2455 reshape_106"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2455 reshape_106" -> "2456 slice_347"  [label="(1, 8, 8, 768)", style=solid];
-"2456 slice_347" -> "2457 slice_348"  [label="(1, 8, 8, 768)", style=solid];
-"2457 slice_348" -> "2458 slice_349"  [label="(1, 7, 8, 768)", style=solid];
-"2458 slice_349" -> "2459 slice_350"  [label="(1, 7, 7, 768)", style=solid];
-"2459 slice_350" -> "2460 contiguous_45"  [label="(1, 7, 7, 768)", style=solid];
-"2460 contiguous_45" -> "2463 layer_norm_50"  [label="(1, 7, 7, 768)", style=solid];
-"2461 _param_constant389" -> "2463 layer_norm_50"  [label="(768,)", style=solid];
-"2462 _param_constant390" -> "2463 layer_norm_50"  [label="(768,)", style=solid];
-"2463 layer_norm_50" -> "2464 add_81"  [label="(1, 7, 7, 768)", style=solid];
-"2464 add_81" -> "2465 add_81_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 768)", style=solid];
-"2464 add_81" -> "2491 add_82"  [label="(1, 7, 7, 768)", style=solid];
-"2465 add_81_0_0_nncf_smooth_quant_0" -> "2466 quantize_per_tensor_default_146"  [label="(1, 7, 7, 768)", style=solid];
-"2466 quantize_per_tensor_default_146" -> "2467 dequantize_per_tensor_default_146"  [label="(1, 7, 7, 768)", style=solid];
-"2467 dequantize_per_tensor_default_146" -> "2474 linear_145"  [label="(1, 7, 7, 768)", style=solid];
-"2468 _param_constant392_0_0" -> "2474 linear_145"  [label="(3072,)", style=solid];
-"2469 scale_updated_constant98" -> "2471 mul_tensor_98"  [label="(3072, 1)", style=solid];
-"2470 compressed_weight_updated_constant98" -> "2471 mul_tensor_98"  [label="(3072, 768)", style=solid];
-"2471 mul_tensor_98" -> "2473 sub_tensor_98"  [label="(3072, 768)", style=solid];
-"2472 zero_point_updated_constant98" -> "2473 sub_tensor_98"  [label="(3072, 1)", style=solid];
-"2473 sub_tensor_98" -> "2474 linear_145"  [label="(3072, 768)", style=solid];
-"2474 linear_145" -> "2475 gelu_23"  [label="(1, 7, 7, 3072)", style=solid];
-"2475 gelu_23" -> "2476 dropout_94"  [label="(1, 7, 7, 3072)", style=solid];
-"2476 dropout_94" -> "2477 dropout_94_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 3072)", style=solid];
-"2477 dropout_94_0_0_nncf_smooth_quant_0" -> "2478 quantize_per_tensor_default_147"  [label="(1, 7, 7, 3072)", style=solid];
-"2478 quantize_per_tensor_default_147" -> "2479 dequantize_per_tensor_default_147"  [label="(1, 7, 7, 3072)", style=solid];
-"2479 dequantize_per_tensor_default_147" -> "2486 linear_146"  [label="(1, 7, 7, 3072)", style=solid];
-"2480 _param_constant394_0_0" -> "2486 linear_146"  [label="(768,)", style=solid];
-"2481 scale_updated_constant99" -> "2483 mul_tensor_99"  [label="(768, 1)", style=solid];
-"2482 compressed_weight_updated_constant99" -> "2483 mul_tensor_99"  [label="(768, 3072)", style=solid];
-"2483 mul_tensor_99" -> "2485 sub_tensor_99"  [label="(768, 3072)", style=solid];
-"2484 zero_point_updated_constant99" -> "2485 sub_tensor_99"  [label="(768, 1)", style=solid];
-"2485 sub_tensor_99" -> "2486 linear_146"  [label="(768, 3072)", style=solid];
-"2486 linear_146" -> "2487 dropout_95"  [label="(1, 7, 7, 768)", style=solid];
-"2487 dropout_95" -> "2490 layer_norm_51"  [label="(1, 7, 7, 768)", style=solid];
-"2488 _param_constant395" -> "2490 layer_norm_51"  [label="(768,)", style=solid];
-"2489 _param_constant396" -> "2490 layer_norm_51"  [label="(768,)", style=solid];
-"2490 layer_norm_51" -> "2491 add_82"  [label="(1, 7, 7, 768)", style=solid];
-"2491 add_82" -> "2494 layer_norm_52"  [label="(1, 7, 7, 768)", style=solid];
-"2492 _param_constant397" -> "2494 layer_norm_52"  [label="(768,)", style=solid];
-"2493 _param_constant398" -> "2494 layer_norm_52"  [label="(768,)", style=solid];
+"0 features_0_2_weight" -> "117 layer_norm"  [label="(96,)", style=solid];
+"1 features_0_2_bias" -> "117 layer_norm"  [label="(96,)", style=solid];
+"2 features_1_0_norm1_weight" -> "176 layer_norm_1"  [label="(96,)", style=solid];
+"3 features_1_0_norm1_bias" -> "176 layer_norm_1"  [label="(96,)", style=solid];
+"4 features_1_0_norm2_weight" -> "201 layer_norm_2"  [label="(96,)", style=solid];
+"5 features_1_0_norm2_bias" -> "201 layer_norm_2"  [label="(96,)", style=solid];
+"6 features_1_1_norm1_weight" -> "279 layer_norm_3"  [label="(96,)", style=solid];
+"7 features_1_1_norm1_bias" -> "279 layer_norm_3"  [label="(96,)", style=solid];
+"8 features_1_1_norm2_weight" -> "304 layer_norm_4"  [label="(96,)", style=solid];
+"9 features_1_1_norm2_bias" -> "304 layer_norm_4"  [label="(96,)", style=solid];
+"10 features_2_norm_weight" -> "329 layer_norm_5"  [label="(192,)", style=solid];
+"11 features_2_norm_bias" -> "329 layer_norm_5"  [label="(192,)", style=solid];
+"12 features_3_0_norm1_weight" -> "391 layer_norm_6"  [label="(192,)", style=solid];
+"13 features_3_0_norm1_bias" -> "391 layer_norm_6"  [label="(192,)", style=solid];
+"14 features_3_0_norm2_weight" -> "416 layer_norm_7"  [label="(192,)", style=solid];
+"15 features_3_0_norm2_bias" -> "416 layer_norm_7"  [label="(192,)", style=solid];
+"16 features_3_1_norm1_weight" -> "497 layer_norm_8"  [label="(192,)", style=solid];
+"17 features_3_1_norm1_bias" -> "497 layer_norm_8"  [label="(192,)", style=solid];
+"18 features_3_1_norm2_weight" -> "522 layer_norm_9"  [label="(192,)", style=solid];
+"19 features_3_1_norm2_bias" -> "522 layer_norm_9"  [label="(192,)", style=solid];
+"20 features_4_norm_weight" -> "547 layer_norm_10"  [label="(384,)", style=solid];
+"21 features_4_norm_bias" -> "547 layer_norm_10"  [label="(384,)", style=solid];
+"22 features_5_0_norm1_weight" -> "609 layer_norm_11"  [label="(384,)", style=solid];
+"23 features_5_0_norm1_bias" -> "609 layer_norm_11"  [label="(384,)", style=solid];
+"24 features_5_0_norm2_weight" -> "634 layer_norm_12"  [label="(384,)", style=solid];
+"25 features_5_0_norm2_bias" -> "634 layer_norm_12"  [label="(384,)", style=solid];
+"26 features_5_1_norm1_weight" -> "715 layer_norm_13"  [label="(384,)", style=solid];
+"27 features_5_1_norm1_bias" -> "715 layer_norm_13"  [label="(384,)", style=solid];
+"28 features_5_1_norm2_weight" -> "740 layer_norm_14"  [label="(384,)", style=solid];
+"29 features_5_1_norm2_bias" -> "740 layer_norm_14"  [label="(384,)", style=solid];
+"30 features_5_2_norm1_weight" -> "803 layer_norm_15"  [label="(384,)", style=solid];
+"31 features_5_2_norm1_bias" -> "803 layer_norm_15"  [label="(384,)", style=solid];
+"32 features_5_2_norm2_weight" -> "828 layer_norm_16"  [label="(384,)", style=solid];
+"33 features_5_2_norm2_bias" -> "828 layer_norm_16"  [label="(384,)", style=solid];
+"34 features_5_3_norm1_weight" -> "909 layer_norm_17"  [label="(384,)", style=solid];
+"35 features_5_3_norm1_bias" -> "909 layer_norm_17"  [label="(384,)", style=solid];
+"36 features_5_3_norm2_weight" -> "934 layer_norm_18"  [label="(384,)", style=solid];
+"37 features_5_3_norm2_bias" -> "934 layer_norm_18"  [label="(384,)", style=solid];
+"38 features_5_4_norm1_weight" -> "997 layer_norm_19"  [label="(384,)", style=solid];
+"39 features_5_4_norm1_bias" -> "997 layer_norm_19"  [label="(384,)", style=solid];
+"40 features_5_4_norm2_weight" -> "1022 layer_norm_20"  [label="(384,)", style=solid];
+"41 features_5_4_norm2_bias" -> "1022 layer_norm_20"  [label="(384,)", style=solid];
+"42 features_5_5_norm1_weight" -> "1103 layer_norm_21"  [label="(384,)", style=solid];
+"43 features_5_5_norm1_bias" -> "1103 layer_norm_21"  [label="(384,)", style=solid];
+"44 features_5_5_norm2_weight" -> "1128 layer_norm_22"  [label="(384,)", style=solid];
+"45 features_5_5_norm2_bias" -> "1128 layer_norm_22"  [label="(384,)", style=solid];
+"46 features_5_6_norm1_weight" -> "1191 layer_norm_23"  [label="(384,)", style=solid];
+"47 features_5_6_norm1_bias" -> "1191 layer_norm_23"  [label="(384,)", style=solid];
+"48 features_5_6_norm2_weight" -> "1216 layer_norm_24"  [label="(384,)", style=solid];
+"49 features_5_6_norm2_bias" -> "1216 layer_norm_24"  [label="(384,)", style=solid];
+"50 features_5_7_norm1_weight" -> "1297 layer_norm_25"  [label="(384,)", style=solid];
+"51 features_5_7_norm1_bias" -> "1297 layer_norm_25"  [label="(384,)", style=solid];
+"52 features_5_7_norm2_weight" -> "1322 layer_norm_26"  [label="(384,)", style=solid];
+"53 features_5_7_norm2_bias" -> "1322 layer_norm_26"  [label="(384,)", style=solid];
+"54 features_5_8_norm1_weight" -> "1385 layer_norm_27"  [label="(384,)", style=solid];
+"55 features_5_8_norm1_bias" -> "1385 layer_norm_27"  [label="(384,)", style=solid];
+"56 features_5_8_norm2_weight" -> "1410 layer_norm_28"  [label="(384,)", style=solid];
+"57 features_5_8_norm2_bias" -> "1410 layer_norm_28"  [label="(384,)", style=solid];
+"58 features_5_9_norm1_weight" -> "1491 layer_norm_29"  [label="(384,)", style=solid];
+"59 features_5_9_norm1_bias" -> "1491 layer_norm_29"  [label="(384,)", style=solid];
+"60 features_5_9_norm2_weight" -> "1516 layer_norm_30"  [label="(384,)", style=solid];
+"61 features_5_9_norm2_bias" -> "1516 layer_norm_30"  [label="(384,)", style=solid];
+"62 features_5_10_norm1_weight" -> "1579 layer_norm_31"  [label="(384,)", style=solid];
+"63 features_5_10_norm1_bias" -> "1579 layer_norm_31"  [label="(384,)", style=solid];
+"64 features_5_10_norm2_weight" -> "1604 layer_norm_32"  [label="(384,)", style=solid];
+"65 features_5_10_norm2_bias" -> "1604 layer_norm_32"  [label="(384,)", style=solid];
+"66 features_5_11_norm1_weight" -> "1685 layer_norm_33"  [label="(384,)", style=solid];
+"67 features_5_11_norm1_bias" -> "1685 layer_norm_33"  [label="(384,)", style=solid];
+"68 features_5_11_norm2_weight" -> "1710 layer_norm_34"  [label="(384,)", style=solid];
+"69 features_5_11_norm2_bias" -> "1710 layer_norm_34"  [label="(384,)", style=solid];
+"70 features_5_12_norm1_weight" -> "1773 layer_norm_35"  [label="(384,)", style=solid];
+"71 features_5_12_norm1_bias" -> "1773 layer_norm_35"  [label="(384,)", style=solid];
+"72 features_5_12_norm2_weight" -> "1798 layer_norm_36"  [label="(384,)", style=solid];
+"73 features_5_12_norm2_bias" -> "1798 layer_norm_36"  [label="(384,)", style=solid];
+"74 features_5_13_norm1_weight" -> "1879 layer_norm_37"  [label="(384,)", style=solid];
+"75 features_5_13_norm1_bias" -> "1879 layer_norm_37"  [label="(384,)", style=solid];
+"76 features_5_13_norm2_weight" -> "1904 layer_norm_38"  [label="(384,)", style=solid];
+"77 features_5_13_norm2_bias" -> "1904 layer_norm_38"  [label="(384,)", style=solid];
+"78 features_5_14_norm1_weight" -> "1967 layer_norm_39"  [label="(384,)", style=solid];
+"79 features_5_14_norm1_bias" -> "1967 layer_norm_39"  [label="(384,)", style=solid];
+"80 features_5_14_norm2_weight" -> "1992 layer_norm_40"  [label="(384,)", style=solid];
+"81 features_5_14_norm2_bias" -> "1992 layer_norm_40"  [label="(384,)", style=solid];
+"82 features_5_15_norm1_weight" -> "2073 layer_norm_41"  [label="(384,)", style=solid];
+"83 features_5_15_norm1_bias" -> "2073 layer_norm_41"  [label="(384,)", style=solid];
+"84 features_5_15_norm2_weight" -> "2098 layer_norm_42"  [label="(384,)", style=solid];
+"85 features_5_15_norm2_bias" -> "2098 layer_norm_42"  [label="(384,)", style=solid];
+"86 features_5_16_norm1_weight" -> "2161 layer_norm_43"  [label="(384,)", style=solid];
+"87 features_5_16_norm1_bias" -> "2161 layer_norm_43"  [label="(384,)", style=solid];
+"88 features_5_16_norm2_weight" -> "2186 layer_norm_44"  [label="(384,)", style=solid];
+"89 features_5_16_norm2_bias" -> "2186 layer_norm_44"  [label="(384,)", style=solid];
+"90 features_5_17_norm1_weight" -> "2267 layer_norm_45"  [label="(384,)", style=solid];
+"91 features_5_17_norm1_bias" -> "2267 layer_norm_45"  [label="(384,)", style=solid];
+"92 features_5_17_norm2_weight" -> "2292 layer_norm_46"  [label="(384,)", style=solid];
+"93 features_5_17_norm2_bias" -> "2292 layer_norm_46"  [label="(384,)", style=solid];
+"94 features_6_norm_weight" -> "2317 layer_norm_47"  [label="(768,)", style=solid];
+"95 features_6_norm_bias" -> "2317 layer_norm_47"  [label="(768,)", style=solid];
+"96 features_7_0_norm1_weight" -> "2379 layer_norm_48"  [label="(768,)", style=solid];
+"97 features_7_0_norm1_bias" -> "2379 layer_norm_48"  [label="(768,)", style=solid];
+"98 features_7_0_norm2_weight" -> "2404 layer_norm_49"  [label="(768,)", style=solid];
+"99 features_7_0_norm2_bias" -> "2404 layer_norm_49"  [label="(768,)", style=solid];
+"100 features_7_1_norm1_weight" -> "2467 layer_norm_50"  [label="(768,)", style=solid];
+"101 features_7_1_norm1_bias" -> "2467 layer_norm_50"  [label="(768,)", style=solid];
+"102 features_7_1_norm2_weight" -> "2492 layer_norm_51"  [label="(768,)", style=solid];
+"103 features_7_1_norm2_bias" -> "2492 layer_norm_51"  [label="(768,)", style=solid];
+"104 norm_weight" -> "2494 layer_norm_52"  [label="(768,)", style=solid];
+"105 norm_bias" -> "2494 layer_norm_52"  [label="(768,)", style=solid];
+"106 x" -> "107 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"107 quantize_per_tensor_default" -> "108 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"108 dequantize_per_tensor_default" -> "115 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"109 features_0_0_bias_0_0" -> "115 conv2d"  [label="(96,)", style=solid];
+"110 scale_updated_constant0" -> "112 mul_tensor"  [label="(96, 1, 1, 1)", style=solid];
+"111 compressed_weight_updated_constant0" -> "112 mul_tensor"  [label="(96, 3, 4, 4)", style=solid];
+"112 mul_tensor" -> "114 sub_tensor"  [label="(96, 3, 4, 4)", style=solid];
+"113 zero_point_updated_constant0" -> "114 sub_tensor"  [label="(96, 1, 1, 1)", style=solid];
+"114 sub_tensor" -> "115 conv2d"  [label="(96, 3, 4, 4)", style=solid];
+"115 conv2d" -> "116 permute"  [label="(1, 96, 56, 56)", style=solid];
+"116 permute" -> "117 layer_norm"  [label="(1, 56, 56, 96)", style=solid];
+"117 layer_norm" -> "119 pad"  [label="(1, 56, 56, 96)", style=solid];
+"117 layer_norm" -> "177 add_1"  [label="(1, 56, 56, 96)", style=solid];
+"118 _frozen_param0" -> "154 add"  [label="(1, 3, 64, 64)", style=solid];
+"119 pad" -> "120 view_2"  [label="(1, 56, 56, 96)", style=solid];
+"120 view_2" -> "121 permute_2"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"121 permute_2" -> "122 reshape"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"122 reshape" -> "123 reshape_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
+"123 reshape_0_0_nncf_smooth_quant_0" -> "124 quantize_per_tensor_default_1"  [label="(49, 64, 96)", style=solid];
+"124 quantize_per_tensor_default_1" -> "125 dequantize_per_tensor_default_1"  [label="(49, 64, 96)", style=solid];
+"125 dequantize_per_tensor_default_1" -> "132 linear_2"  [label="(49, 64, 96)", style=solid];
+"126 _frozen_param1_0_0" -> "132 linear_2"  [label="(288,)", style=solid];
+"127 scale_updated_constant1" -> "129 mul_tensor_1"  [label="(288, 1)", style=solid];
+"128 compressed_weight_updated_constant1" -> "129 mul_tensor_1"  [label="(288, 96)", style=solid];
+"129 mul_tensor_1" -> "131 sub_tensor_1"  [label="(288, 96)", style=solid];
+"130 zero_point_updated_constant1" -> "131 sub_tensor_1"  [label="(288, 1)", style=solid];
+"131 sub_tensor_1" -> "132 linear_2"  [label="(288, 96)", style=solid];
+"132 linear_2" -> "133 reshape_1"  [label="(49, 64, 288)", style=solid];
+"133 reshape_1" -> "134 permute_3"  [label="(49, 64, 3, 3, 32)", style=solid];
+"134 permute_3" -> "135 select"  [label="(3, 49, 3, 64, 32)", style=solid];
+"134 permute_3" -> "136 select_1"  [label="(3, 49, 3, 64, 32)", style=solid];
+"134 permute_3" -> "137 select_2"  [label="(3, 49, 3, 64, 32)", style=solid];
+"135 select" -> "138 linalg_vector_norm"  [label="(49, 3, 64, 32)", style=solid];
+"135 select" -> "140 expand_as"  [label="(49, 3, 64, 32)", style=solid];
+"135 select" -> "141 div"  [label="(49, 3, 64, 32)", style=solid];
+"136 select_1" -> "144 linalg_vector_norm_1"  [label="(49, 3, 64, 32)", style=solid];
+"136 select_1" -> "146 expand_as_1"  [label="(49, 3, 64, 32)", style=solid];
+"136 select_1" -> "147 div_1"  [label="(49, 3, 64, 32)", style=solid];
+"137 select_2" -> "157 matmul_1"  [label="(49, 3, 64, 32)", style=solid];
+"138 linalg_vector_norm" -> "139 clamp_min"  [label="(49, 3, 64, 1)", style=solid];
+"139 clamp_min" -> "140 expand_as"  [label="(49, 3, 64, 1)", style=solid];
+"140 expand_as" -> "141 div"  [label="(49, 3, 64, 32)", style=solid];
+"141 div" -> "142 quantize_per_tensor_default_2"  [label="(49, 3, 64, 32)", style=solid];
+"142 quantize_per_tensor_default_2" -> "143 dequantize_per_tensor_default_2"  [label="(49, 3, 64, 32)", style=solid];
+"143 dequantize_per_tensor_default_2" -> "151 matmul"  [label="(49, 3, 64, 32)", style=solid];
+"144 linalg_vector_norm_1" -> "145 clamp_min_1"  [label="(49, 3, 64, 1)", style=solid];
+"145 clamp_min_1" -> "146 expand_as_1"  [label="(49, 3, 64, 1)", style=solid];
+"146 expand_as_1" -> "147 div_1"  [label="(49, 3, 64, 32)", style=solid];
+"147 div_1" -> "148 quantize_per_tensor_default_3"  [label="(49, 3, 64, 32)", style=solid];
+"148 quantize_per_tensor_default_3" -> "149 dequantize_per_tensor_default_3"  [label="(49, 3, 64, 32)", style=solid];
+"149 dequantize_per_tensor_default_3" -> "150 transpose"  [label="(49, 3, 64, 32)", style=solid];
+"150 transpose" -> "151 matmul"  [label="(49, 3, 32, 64)", style=solid];
+"151 matmul" -> "153 mul_1"  [label="(49, 3, 64, 64)", style=solid];
+"152 _frozen_param3" -> "153 mul_1"  [label="(3, 1, 1)", style=solid];
+"153 mul_1" -> "154 add"  [label="(49, 3, 64, 64)", style=solid];
+"154 add" -> "155 softmax"  [label="(49, 3, 64, 64)", style=solid];
+"155 softmax" -> "156 dropout"  [label="(49, 3, 64, 64)", style=solid];
+"156 dropout" -> "157 matmul_1"  [label="(49, 3, 64, 64)", style=solid];
+"157 matmul_1" -> "158 transpose_1"  [label="(49, 3, 64, 32)", style=solid];
+"158 transpose_1" -> "159 reshape_2"  [label="(49, 64, 3, 32)", style=solid];
+"159 reshape_2" -> "160 reshape_2_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
+"160 reshape_2_0_0_nncf_smooth_quant_0" -> "161 quantize_per_tensor_default_4"  [label="(49, 64, 96)", style=solid];
+"161 quantize_per_tensor_default_4" -> "162 dequantize_per_tensor_default_4"  [label="(49, 64, 96)", style=solid];
+"162 dequantize_per_tensor_default_4" -> "169 linear_3"  [label="(49, 64, 96)", style=solid];
+"163 features_1_0_attn_proj_bias_0_0" -> "169 linear_3"  [label="(96,)", style=solid];
+"164 scale_updated_constant2" -> "166 mul_tensor_2"  [label="(96, 1)", style=solid];
+"165 compressed_weight_updated_constant2" -> "166 mul_tensor_2"  [label="(96, 96)", style=solid];
+"166 mul_tensor_2" -> "168 sub_tensor_2"  [label="(96, 96)", style=solid];
+"167 zero_point_updated_constant2" -> "168 sub_tensor_2"  [label="(96, 1)", style=solid];
+"168 sub_tensor_2" -> "169 linear_3"  [label="(96, 96)", style=solid];
+"169 linear_3" -> "170 dropout_1"  [label="(49, 64, 96)", style=solid];
+"170 dropout_1" -> "171 view_3"  [label="(49, 64, 96)", style=solid];
+"171 view_3" -> "172 permute_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"172 permute_4" -> "173 reshape_3"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"173 reshape_3" -> "174 slice_2"  [label="(1, 56, 56, 96)", style=solid];
+"174 slice_2" -> "175 slice_3"  [label="(1, 56, 56, 96)", style=solid];
+"175 slice_3" -> "176 layer_norm_1"  [label="(1, 56, 56, 96)", style=solid];
+"176 layer_norm_1" -> "177 add_1"  [label="(1, 56, 56, 96)", style=solid];
+"177 add_1" -> "178 add_1_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 96)", style=solid];
+"177 add_1" -> "202 add_2"  [label="(1, 56, 56, 96)", style=solid];
+"178 add_1_0_0_nncf_smooth_quant_0" -> "179 quantize_per_tensor_default_5"  [label="(1, 56, 56, 96)", style=solid];
+"179 quantize_per_tensor_default_5" -> "180 dequantize_per_tensor_default_5"  [label="(1, 56, 56, 96)", style=solid];
+"180 dequantize_per_tensor_default_5" -> "187 linear_4"  [label="(1, 56, 56, 96)", style=solid];
+"181 features_1_0_mlp_0_bias_0_0" -> "187 linear_4"  [label="(384,)", style=solid];
+"182 scale_updated_constant3" -> "184 mul_tensor_3"  [label="(384, 1)", style=solid];
+"183 compressed_weight_updated_constant3" -> "184 mul_tensor_3"  [label="(384, 96)", style=solid];
+"184 mul_tensor_3" -> "186 sub_tensor_3"  [label="(384, 96)", style=solid];
+"185 zero_point_updated_constant3" -> "186 sub_tensor_3"  [label="(384, 1)", style=solid];
+"186 sub_tensor_3" -> "187 linear_4"  [label="(384, 96)", style=solid];
+"187 linear_4" -> "188 gelu"  [label="(1, 56, 56, 384)", style=solid];
+"188 gelu" -> "189 dropout_2"  [label="(1, 56, 56, 384)", style=solid];
+"189 dropout_2" -> "190 dropout_2_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 384)", style=solid];
+"190 dropout_2_0_0_nncf_smooth_quant_0" -> "191 quantize_per_tensor_default_6"  [label="(1, 56, 56, 384)", style=solid];
+"191 quantize_per_tensor_default_6" -> "192 dequantize_per_tensor_default_6"  [label="(1, 56, 56, 384)", style=solid];
+"192 dequantize_per_tensor_default_6" -> "199 linear_5"  [label="(1, 56, 56, 384)", style=solid];
+"193 features_1_0_mlp_3_bias_0_0" -> "199 linear_5"  [label="(96,)", style=solid];
+"194 scale_updated_constant4" -> "196 mul_tensor_4"  [label="(96, 1)", style=solid];
+"195 compressed_weight_updated_constant4" -> "196 mul_tensor_4"  [label="(96, 384)", style=solid];
+"196 mul_tensor_4" -> "198 sub_tensor_4"  [label="(96, 384)", style=solid];
+"197 zero_point_updated_constant4" -> "198 sub_tensor_4"  [label="(96, 1)", style=solid];
+"198 sub_tensor_4" -> "199 linear_5"  [label="(96, 384)", style=solid];
+"199 linear_5" -> "200 dropout_3"  [label="(1, 56, 56, 96)", style=solid];
+"200 dropout_3" -> "201 layer_norm_2"  [label="(1, 56, 56, 96)", style=solid];
+"201 layer_norm_2" -> "202 add_2"  [label="(1, 56, 56, 96)", style=solid];
+"202 add_2" -> "204 pad_1"  [label="(1, 56, 56, 96)", style=solid];
+"202 add_2" -> "280 add_5"  [label="(1, 56, 56, 96)", style=solid];
+"203 _frozen_param4" -> "240 add_3"  [label="(1, 3, 64, 64)", style=solid];
+"204 pad_1" -> "205 roll"  [label="(1, 56, 56, 96)", style=solid];
+"205 roll" -> "206 view_6"  [label="(1, 56, 56, 96)", style=solid];
+"206 view_6" -> "207 permute_6"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"207 permute_6" -> "208 reshape_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"208 reshape_4" -> "209 reshape_4_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
+"208 reshape_4" -> "241 new_zeros"  [label="(49, 64, 96)", style=solid];
+"209 reshape_4_0_0_nncf_smooth_quant_0" -> "210 quantize_per_tensor_default_7"  [label="(49, 64, 96)", style=solid];
+"210 quantize_per_tensor_default_7" -> "211 dequantize_per_tensor_default_7"  [label="(49, 64, 96)", style=solid];
+"211 dequantize_per_tensor_default_7" -> "218 linear_8"  [label="(49, 64, 96)", style=solid];
+"212 _frozen_param5_0_0" -> "218 linear_8"  [label="(288,)", style=solid];
+"213 scale_updated_constant5" -> "215 mul_tensor_5"  [label="(288, 1)", style=solid];
+"214 compressed_weight_updated_constant5" -> "215 mul_tensor_5"  [label="(288, 96)", style=solid];
+"215 mul_tensor_5" -> "217 sub_tensor_5"  [label="(288, 96)", style=solid];
+"216 zero_point_updated_constant5" -> "217 sub_tensor_5"  [label="(288, 1)", style=solid];
+"217 sub_tensor_5" -> "218 linear_8"  [label="(288, 96)", style=solid];
+"218 linear_8" -> "219 reshape_5"  [label="(49, 64, 288)", style=solid];
+"219 reshape_5" -> "220 permute_7"  [label="(49, 64, 3, 3, 32)", style=solid];
+"220 permute_7" -> "221 select_3"  [label="(3, 49, 3, 64, 32)", style=solid];
+"220 permute_7" -> "222 select_4"  [label="(3, 49, 3, 64, 32)", style=solid];
+"220 permute_7" -> "223 select_5"  [label="(3, 49, 3, 64, 32)", style=solid];
+"221 select_3" -> "224 linalg_vector_norm_2"  [label="(49, 3, 64, 32)", style=solid];
+"221 select_3" -> "226 expand_as_2"  [label="(49, 3, 64, 32)", style=solid];
+"221 select_3" -> "227 div_2"  [label="(49, 3, 64, 32)", style=solid];
+"222 select_4" -> "230 linalg_vector_norm_3"  [label="(49, 3, 64, 32)", style=solid];
+"222 select_4" -> "232 expand_as_3"  [label="(49, 3, 64, 32)", style=solid];
+"222 select_4" -> "233 div_3"  [label="(49, 3, 64, 32)", style=solid];
+"223 select_5" -> "259 matmul_3"  [label="(49, 3, 64, 32)", style=solid];
+"224 linalg_vector_norm_2" -> "225 clamp_min_2"  [label="(49, 3, 64, 1)", style=solid];
+"225 clamp_min_2" -> "226 expand_as_2"  [label="(49, 3, 64, 1)", style=solid];
+"226 expand_as_2" -> "227 div_2"  [label="(49, 3, 64, 32)", style=solid];
+"227 div_2" -> "228 quantize_per_tensor_default_8"  [label="(49, 3, 64, 32)", style=solid];
+"228 quantize_per_tensor_default_8" -> "229 dequantize_per_tensor_default_8"  [label="(49, 3, 64, 32)", style=solid];
+"229 dequantize_per_tensor_default_8" -> "237 matmul_2"  [label="(49, 3, 64, 32)", style=solid];
+"230 linalg_vector_norm_3" -> "231 clamp_min_3"  [label="(49, 3, 64, 1)", style=solid];
+"231 clamp_min_3" -> "232 expand_as_3"  [label="(49, 3, 64, 1)", style=solid];
+"232 expand_as_3" -> "233 div_3"  [label="(49, 3, 64, 32)", style=solid];
+"233 div_3" -> "234 quantize_per_tensor_default_9"  [label="(49, 3, 64, 32)", style=solid];
+"234 quantize_per_tensor_default_9" -> "235 dequantize_per_tensor_default_9"  [label="(49, 3, 64, 32)", style=solid];
+"235 dequantize_per_tensor_default_9" -> "236 transpose_2"  [label="(49, 3, 64, 32)", style=solid];
+"236 transpose_2" -> "237 matmul_2"  [label="(49, 3, 32, 64)", style=solid];
+"237 matmul_2" -> "239 mul_3"  [label="(49, 3, 64, 64)", style=solid];
+"238 _frozen_param7" -> "239 mul_3"  [label="(3, 1, 1)", style=solid];
+"239 mul_3" -> "240 add_3"  [label="(49, 3, 64, 64)", style=solid];
+"240 add_3" -> "252 view_8"  [label="(49, 3, 64, 64)", style=solid];
+"241 new_zeros" -> "242 view_7"  [label="(56, 56)", style=solid];
+"242 view_7" -> "243 permute_8"  [label="(7, 8, 7, 8)", style=solid];
+"243 permute_8" -> "244 reshape_6"  [label="(7, 7, 8, 8)", style=solid];
+"244 reshape_6" -> "245 unsqueeze_2"  [label="(49, 64)", style=solid];
+"244 reshape_6" -> "246 unsqueeze_3"  [label="(49, 64)", style=solid];
+"245 unsqueeze_2" -> "247 sub"  [label="(49, 1, 64)", style=solid];
+"246 unsqueeze_3" -> "247 sub"  [label="(49, 64, 1)", style=solid];
+"247 sub" -> "248 ne"  [label="(49, 64, 64)", style=solid];
+"247 sub" -> "249 masked_fill"  [label="(49, 64, 64)", style=solid];
+"247 sub" -> "250 eq"  [label="(49, 64, 64)", style=solid];
+"248 ne" -> "249 masked_fill"  [label="(49, 64, 64)", style=solid];
+"249 masked_fill" -> "251 masked_fill_1"  [label="(49, 64, 64)", style=solid];
+"250 eq" -> "251 masked_fill_1"  [label="(49, 64, 64)", style=solid];
+"251 masked_fill_1" -> "253 unsqueeze_4"  [label="(49, 64, 64)", style=solid];
+"252 view_8" -> "255 add_4"  [label="(1, 49, 3, 64, 64)", style=solid];
+"253 unsqueeze_4" -> "254 unsqueeze_5"  [label="(49, 1, 64, 64)", style=solid];
+"254 unsqueeze_5" -> "255 add_4"  [label="(1, 49, 1, 64, 64)", style=solid];
+"255 add_4" -> "256 view_9"  [label="(1, 49, 3, 64, 64)", style=solid];
+"256 view_9" -> "257 softmax_1"  [label="(49, 3, 64, 64)", style=solid];
+"257 softmax_1" -> "258 dropout_4"  [label="(49, 3, 64, 64)", style=solid];
+"258 dropout_4" -> "259 matmul_3"  [label="(49, 3, 64, 64)", style=solid];
+"259 matmul_3" -> "260 transpose_3"  [label="(49, 3, 64, 32)", style=solid];
+"260 transpose_3" -> "261 reshape_7"  [label="(49, 64, 3, 32)", style=solid];
+"261 reshape_7" -> "262 reshape_7_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
+"262 reshape_7_0_0_nncf_smooth_quant_0" -> "263 quantize_per_tensor_default_10"  [label="(49, 64, 96)", style=solid];
+"263 quantize_per_tensor_default_10" -> "264 dequantize_per_tensor_default_10"  [label="(49, 64, 96)", style=solid];
+"264 dequantize_per_tensor_default_10" -> "271 linear_9"  [label="(49, 64, 96)", style=solid];
+"265 features_1_1_attn_proj_bias_0_0" -> "271 linear_9"  [label="(96,)", style=solid];
+"266 scale_updated_constant6" -> "268 mul_tensor_6"  [label="(96, 1)", style=solid];
+"267 compressed_weight_updated_constant6" -> "268 mul_tensor_6"  [label="(96, 96)", style=solid];
+"268 mul_tensor_6" -> "270 sub_tensor_6"  [label="(96, 96)", style=solid];
+"269 zero_point_updated_constant6" -> "270 sub_tensor_6"  [label="(96, 1)", style=solid];
+"270 sub_tensor_6" -> "271 linear_9"  [label="(96, 96)", style=solid];
+"271 linear_9" -> "272 dropout_5"  [label="(49, 64, 96)", style=solid];
+"272 dropout_5" -> "273 view_10"  [label="(49, 64, 96)", style=solid];
+"273 view_10" -> "274 permute_9"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"274 permute_9" -> "275 reshape_8"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"275 reshape_8" -> "276 roll_1"  [label="(1, 56, 56, 96)", style=solid];
+"276 roll_1" -> "277 slice_23"  [label="(1, 56, 56, 96)", style=solid];
+"277 slice_23" -> "278 slice_24"  [label="(1, 56, 56, 96)", style=solid];
+"278 slice_24" -> "279 layer_norm_3"  [label="(1, 56, 56, 96)", style=solid];
+"279 layer_norm_3" -> "280 add_5"  [label="(1, 56, 56, 96)", style=solid];
+"280 add_5" -> "281 add_5_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 96)", style=solid];
+"280 add_5" -> "305 add_6"  [label="(1, 56, 56, 96)", style=solid];
+"281 add_5_0_0_nncf_smooth_quant_0" -> "282 quantize_per_tensor_default_11"  [label="(1, 56, 56, 96)", style=solid];
+"282 quantize_per_tensor_default_11" -> "283 dequantize_per_tensor_default_11"  [label="(1, 56, 56, 96)", style=solid];
+"283 dequantize_per_tensor_default_11" -> "290 linear_10"  [label="(1, 56, 56, 96)", style=solid];
+"284 features_1_1_mlp_0_bias_0_0" -> "290 linear_10"  [label="(384,)", style=solid];
+"285 scale_updated_constant7" -> "287 mul_tensor_7"  [label="(384, 1)", style=solid];
+"286 compressed_weight_updated_constant7" -> "287 mul_tensor_7"  [label="(384, 96)", style=solid];
+"287 mul_tensor_7" -> "289 sub_tensor_7"  [label="(384, 96)", style=solid];
+"288 zero_point_updated_constant7" -> "289 sub_tensor_7"  [label="(384, 1)", style=solid];
+"289 sub_tensor_7" -> "290 linear_10"  [label="(384, 96)", style=solid];
+"290 linear_10" -> "291 gelu_1"  [label="(1, 56, 56, 384)", style=solid];
+"291 gelu_1" -> "292 dropout_6"  [label="(1, 56, 56, 384)", style=solid];
+"292 dropout_6" -> "293 dropout_6_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 384)", style=solid];
+"293 dropout_6_0_0_nncf_smooth_quant_0" -> "294 quantize_per_tensor_default_12"  [label="(1, 56, 56, 384)", style=solid];
+"294 quantize_per_tensor_default_12" -> "295 dequantize_per_tensor_default_12"  [label="(1, 56, 56, 384)", style=solid];
+"295 dequantize_per_tensor_default_12" -> "302 linear_11"  [label="(1, 56, 56, 384)", style=solid];
+"296 features_1_1_mlp_3_bias_0_0" -> "302 linear_11"  [label="(96,)", style=solid];
+"297 scale_updated_constant8" -> "299 mul_tensor_8"  [label="(96, 1)", style=solid];
+"298 compressed_weight_updated_constant8" -> "299 mul_tensor_8"  [label="(96, 384)", style=solid];
+"299 mul_tensor_8" -> "301 sub_tensor_8"  [label="(96, 384)", style=solid];
+"300 zero_point_updated_constant8" -> "301 sub_tensor_8"  [label="(96, 1)", style=solid];
+"301 sub_tensor_8" -> "302 linear_11"  [label="(96, 384)", style=solid];
+"302 linear_11" -> "303 dropout_7"  [label="(1, 56, 56, 96)", style=solid];
+"303 dropout_7" -> "304 layer_norm_4"  [label="(1, 56, 56, 96)", style=solid];
+"304 layer_norm_4" -> "305 add_6"  [label="(1, 56, 56, 96)", style=solid];
+"305 add_6" -> "306 pad_2"  [label="(1, 56, 56, 96)", style=solid];
+"306 pad_2" -> "307 slice_25"  [label="(1, 56, 56, 96)", style=solid];
+"306 pad_2" -> "310 slice_28"  [label="(1, 56, 56, 96)", style=solid];
+"306 pad_2" -> "313 slice_31"  [label="(1, 56, 56, 96)", style=solid];
+"306 pad_2" -> "316 slice_34"  [label="(1, 56, 56, 96)", style=solid];
+"307 slice_25" -> "308 slice_26"  [label="(1, 28, 56, 96)", style=solid];
+"308 slice_26" -> "309 slice_27"  [label="(1, 28, 28, 96)", style=solid];
+"309 slice_27" -> "319 cat"  [label="(1, 28, 28, 96)", style=solid];
+"310 slice_28" -> "311 slice_29"  [label="(1, 28, 56, 96)", style=solid];
+"311 slice_29" -> "312 slice_30"  [label="(1, 28, 28, 96)", style=solid];
+"312 slice_30" -> "319 cat"  [label="(1, 28, 28, 96)", style=solid];
+"313 slice_31" -> "314 slice_32"  [label="(1, 28, 56, 96)", style=solid];
+"314 slice_32" -> "315 slice_33"  [label="(1, 28, 28, 96)", style=solid];
+"315 slice_33" -> "319 cat"  [label="(1, 28, 28, 96)", style=solid];
+"316 slice_34" -> "317 slice_35"  [label="(1, 28, 56, 96)", style=solid];
+"317 slice_35" -> "318 slice_36"  [label="(1, 28, 28, 96)", style=solid];
+"318 slice_36" -> "319 cat"  [label="(1, 28, 28, 96)", style=solid];
+"319 cat" -> "320 cat_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 384)", style=solid];
+"320 cat_0_0_nncf_smooth_quant_0" -> "321 quantize_per_tensor_default_13"  [label="(1, 28, 28, 384)", style=solid];
+"321 quantize_per_tensor_default_13" -> "322 dequantize_per_tensor_default_13"  [label="(1, 28, 28, 384)", style=solid];
+"322 dequantize_per_tensor_default_13" -> "328 linear_12"  [label="(1, 28, 28, 384)", style=solid];
+"323 scale_updated_constant9" -> "325 mul_tensor_9"  [label="(192, 1)", style=solid];
+"324 compressed_weight_updated_constant9" -> "325 mul_tensor_9"  [label="(192, 384)", style=solid];
+"325 mul_tensor_9" -> "327 sub_tensor_9"  [label="(192, 384)", style=solid];
+"326 zero_point_updated_constant9" -> "327 sub_tensor_9"  [label="(192, 1)", style=solid];
+"327 sub_tensor_9" -> "328 linear_12"  [label="(192, 384)", style=solid];
+"328 linear_12" -> "329 layer_norm_5"  [label="(1, 28, 28, 192)", style=solid];
+"329 layer_norm_5" -> "331 pad_3"  [label="(1, 28, 28, 192)", style=solid];
+"329 layer_norm_5" -> "392 add_8"  [label="(1, 28, 28, 192)", style=solid];
+"330 _frozen_param17" -> "366 add_7"  [label="(1, 6, 64, 64)", style=solid];
+"331 pad_3" -> "332 view_13"  [label="(1, 32, 32, 192)", style=solid];
+"332 view_13" -> "333 permute_11"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"333 permute_11" -> "334 reshape_9"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"334 reshape_9" -> "335 reshape_9_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
+"335 reshape_9_0_0_nncf_smooth_quant_0" -> "336 quantize_per_tensor_default_14"  [label="(16, 64, 192)", style=solid];
+"336 quantize_per_tensor_default_14" -> "337 dequantize_per_tensor_default_14"  [label="(16, 64, 192)", style=solid];
+"337 dequantize_per_tensor_default_14" -> "344 linear_15"  [label="(16, 64, 192)", style=solid];
+"338 _frozen_param18_0_0" -> "344 linear_15"  [label="(576,)", style=solid];
+"339 scale_updated_constant10" -> "341 mul_tensor_10"  [label="(576, 1)", style=solid];
+"340 compressed_weight_updated_constant10" -> "341 mul_tensor_10"  [label="(576, 192)", style=solid];
+"341 mul_tensor_10" -> "343 sub_tensor_10"  [label="(576, 192)", style=solid];
+"342 zero_point_updated_constant10" -> "343 sub_tensor_10"  [label="(576, 1)", style=solid];
+"343 sub_tensor_10" -> "344 linear_15"  [label="(576, 192)", style=solid];
+"344 linear_15" -> "345 reshape_10"  [label="(16, 64, 576)", style=solid];
+"345 reshape_10" -> "346 permute_12"  [label="(16, 64, 3, 6, 32)", style=solid];
+"346 permute_12" -> "347 select_6"  [label="(3, 16, 6, 64, 32)", style=solid];
+"346 permute_12" -> "348 select_7"  [label="(3, 16, 6, 64, 32)", style=solid];
+"346 permute_12" -> "349 select_8"  [label="(3, 16, 6, 64, 32)", style=solid];
+"347 select_6" -> "350 linalg_vector_norm_4"  [label="(16, 6, 64, 32)", style=solid];
+"347 select_6" -> "352 expand_as_4"  [label="(16, 6, 64, 32)", style=solid];
+"347 select_6" -> "353 div_4"  [label="(16, 6, 64, 32)", style=solid];
+"348 select_7" -> "356 linalg_vector_norm_5"  [label="(16, 6, 64, 32)", style=solid];
+"348 select_7" -> "358 expand_as_5"  [label="(16, 6, 64, 32)", style=solid];
+"348 select_7" -> "359 div_5"  [label="(16, 6, 64, 32)", style=solid];
+"349 select_8" -> "369 matmul_5"  [label="(16, 6, 64, 32)", style=solid];
+"350 linalg_vector_norm_4" -> "351 clamp_min_4"  [label="(16, 6, 64, 1)", style=solid];
+"351 clamp_min_4" -> "352 expand_as_4"  [label="(16, 6, 64, 1)", style=solid];
+"352 expand_as_4" -> "353 div_4"  [label="(16, 6, 64, 32)", style=solid];
+"353 div_4" -> "354 quantize_per_tensor_default_15"  [label="(16, 6, 64, 32)", style=solid];
+"354 quantize_per_tensor_default_15" -> "355 dequantize_per_tensor_default_15"  [label="(16, 6, 64, 32)", style=solid];
+"355 dequantize_per_tensor_default_15" -> "363 matmul_4"  [label="(16, 6, 64, 32)", style=solid];
+"356 linalg_vector_norm_5" -> "357 clamp_min_5"  [label="(16, 6, 64, 1)", style=solid];
+"357 clamp_min_5" -> "358 expand_as_5"  [label="(16, 6, 64, 1)", style=solid];
+"358 expand_as_5" -> "359 div_5"  [label="(16, 6, 64, 32)", style=solid];
+"359 div_5" -> "360 quantize_per_tensor_default_16"  [label="(16, 6, 64, 32)", style=solid];
+"360 quantize_per_tensor_default_16" -> "361 dequantize_per_tensor_default_16"  [label="(16, 6, 64, 32)", style=solid];
+"361 dequantize_per_tensor_default_16" -> "362 transpose_4"  [label="(16, 6, 64, 32)", style=solid];
+"362 transpose_4" -> "363 matmul_4"  [label="(16, 6, 32, 64)", style=solid];
+"363 matmul_4" -> "365 mul_5"  [label="(16, 6, 64, 64)", style=solid];
+"364 _frozen_param20" -> "365 mul_5"  [label="(6, 1, 1)", style=solid];
+"365 mul_5" -> "366 add_7"  [label="(16, 6, 64, 64)", style=solid];
+"366 add_7" -> "367 softmax_2"  [label="(16, 6, 64, 64)", style=solid];
+"367 softmax_2" -> "368 dropout_8"  [label="(16, 6, 64, 64)", style=solid];
+"368 dropout_8" -> "369 matmul_5"  [label="(16, 6, 64, 64)", style=solid];
+"369 matmul_5" -> "370 transpose_5"  [label="(16, 6, 64, 32)", style=solid];
+"370 transpose_5" -> "371 reshape_11"  [label="(16, 64, 6, 32)", style=solid];
+"371 reshape_11" -> "372 reshape_11_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
+"372 reshape_11_0_0_nncf_smooth_quant_0" -> "373 quantize_per_tensor_default_17"  [label="(16, 64, 192)", style=solid];
+"373 quantize_per_tensor_default_17" -> "374 dequantize_per_tensor_default_17"  [label="(16, 64, 192)", style=solid];
+"374 dequantize_per_tensor_default_17" -> "381 linear_16"  [label="(16, 64, 192)", style=solid];
+"375 features_3_0_attn_proj_bias_0_0" -> "381 linear_16"  [label="(192,)", style=solid];
+"376 scale_updated_constant11" -> "378 mul_tensor_11"  [label="(192, 1)", style=solid];
+"377 compressed_weight_updated_constant11" -> "378 mul_tensor_11"  [label="(192, 192)", style=solid];
+"378 mul_tensor_11" -> "380 sub_tensor_11"  [label="(192, 192)", style=solid];
+"379 zero_point_updated_constant11" -> "380 sub_tensor_11"  [label="(192, 1)", style=solid];
+"380 sub_tensor_11" -> "381 linear_16"  [label="(192, 192)", style=solid];
+"381 linear_16" -> "382 dropout_9"  [label="(16, 64, 192)", style=solid];
+"382 dropout_9" -> "383 view_14"  [label="(16, 64, 192)", style=solid];
+"383 view_14" -> "384 permute_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"384 permute_13" -> "385 reshape_12"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"385 reshape_12" -> "386 slice_38"  [label="(1, 32, 32, 192)", style=solid];
+"386 slice_38" -> "387 slice_39"  [label="(1, 32, 32, 192)", style=solid];
+"387 slice_39" -> "388 slice_40"  [label="(1, 28, 32, 192)", style=solid];
+"388 slice_40" -> "389 slice_41"  [label="(1, 28, 28, 192)", style=solid];
+"389 slice_41" -> "390 contiguous_3"  [label="(1, 28, 28, 192)", style=solid];
+"390 contiguous_3" -> "391 layer_norm_6"  [label="(1, 28, 28, 192)", style=solid];
+"391 layer_norm_6" -> "392 add_8"  [label="(1, 28, 28, 192)", style=solid];
+"392 add_8" -> "393 add_8_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 192)", style=solid];
+"392 add_8" -> "417 add_9"  [label="(1, 28, 28, 192)", style=solid];
+"393 add_8_0_0_nncf_smooth_quant_0" -> "394 quantize_per_tensor_default_18"  [label="(1, 28, 28, 192)", style=solid];
+"394 quantize_per_tensor_default_18" -> "395 dequantize_per_tensor_default_18"  [label="(1, 28, 28, 192)", style=solid];
+"395 dequantize_per_tensor_default_18" -> "402 linear_17"  [label="(1, 28, 28, 192)", style=solid];
+"396 features_3_0_mlp_0_bias_0_0" -> "402 linear_17"  [label="(768,)", style=solid];
+"397 scale_updated_constant12" -> "399 mul_tensor_12"  [label="(768, 1)", style=solid];
+"398 compressed_weight_updated_constant12" -> "399 mul_tensor_12"  [label="(768, 192)", style=solid];
+"399 mul_tensor_12" -> "401 sub_tensor_12"  [label="(768, 192)", style=solid];
+"400 zero_point_updated_constant12" -> "401 sub_tensor_12"  [label="(768, 1)", style=solid];
+"401 sub_tensor_12" -> "402 linear_17"  [label="(768, 192)", style=solid];
+"402 linear_17" -> "403 gelu_2"  [label="(1, 28, 28, 768)", style=solid];
+"403 gelu_2" -> "404 dropout_10"  [label="(1, 28, 28, 768)", style=solid];
+"404 dropout_10" -> "405 dropout_10_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 768)", style=solid];
+"405 dropout_10_0_0_nncf_smooth_quant_0" -> "406 quantize_per_tensor_default_19"  [label="(1, 28, 28, 768)", style=solid];
+"406 quantize_per_tensor_default_19" -> "407 dequantize_per_tensor_default_19"  [label="(1, 28, 28, 768)", style=solid];
+"407 dequantize_per_tensor_default_19" -> "414 linear_18"  [label="(1, 28, 28, 768)", style=solid];
+"408 features_3_0_mlp_3_bias_0_0" -> "414 linear_18"  [label="(192,)", style=solid];
+"409 scale_updated_constant13" -> "411 mul_tensor_13"  [label="(192, 1)", style=solid];
+"410 compressed_weight_updated_constant13" -> "411 mul_tensor_13"  [label="(192, 768)", style=solid];
+"411 mul_tensor_13" -> "413 sub_tensor_13"  [label="(192, 768)", style=solid];
+"412 zero_point_updated_constant13" -> "413 sub_tensor_13"  [label="(192, 1)", style=solid];
+"413 sub_tensor_13" -> "414 linear_18"  [label="(192, 768)", style=solid];
+"414 linear_18" -> "415 dropout_11"  [label="(1, 28, 28, 192)", style=solid];
+"415 dropout_11" -> "416 layer_norm_7"  [label="(1, 28, 28, 192)", style=solid];
+"416 layer_norm_7" -> "417 add_9"  [label="(1, 28, 28, 192)", style=solid];
+"417 add_9" -> "419 pad_4"  [label="(1, 28, 28, 192)", style=solid];
+"417 add_9" -> "498 add_12"  [label="(1, 28, 28, 192)", style=solid];
+"418 _frozen_param21" -> "455 add_10"  [label="(1, 6, 64, 64)", style=solid];
+"419 pad_4" -> "420 roll_2"  [label="(1, 32, 32, 192)", style=solid];
+"420 roll_2" -> "421 view_17"  [label="(1, 32, 32, 192)", style=solid];
+"421 view_17" -> "422 permute_15"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"422 permute_15" -> "423 reshape_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"423 reshape_13" -> "424 reshape_13_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
+"423 reshape_13" -> "456 new_zeros_1"  [label="(16, 64, 192)", style=solid];
+"424 reshape_13_0_0_nncf_smooth_quant_0" -> "425 quantize_per_tensor_default_20"  [label="(16, 64, 192)", style=solid];
+"425 quantize_per_tensor_default_20" -> "426 dequantize_per_tensor_default_20"  [label="(16, 64, 192)", style=solid];
+"426 dequantize_per_tensor_default_20" -> "433 linear_21"  [label="(16, 64, 192)", style=solid];
+"427 _frozen_param22_0_0" -> "433 linear_21"  [label="(576,)", style=solid];
+"428 scale_updated_constant14" -> "430 mul_tensor_14"  [label="(576, 1)", style=solid];
+"429 compressed_weight_updated_constant14" -> "430 mul_tensor_14"  [label="(576, 192)", style=solid];
+"430 mul_tensor_14" -> "432 sub_tensor_14"  [label="(576, 192)", style=solid];
+"431 zero_point_updated_constant14" -> "432 sub_tensor_14"  [label="(576, 1)", style=solid];
+"432 sub_tensor_14" -> "433 linear_21"  [label="(576, 192)", style=solid];
+"433 linear_21" -> "434 reshape_14"  [label="(16, 64, 576)", style=solid];
+"434 reshape_14" -> "435 permute_16"  [label="(16, 64, 3, 6, 32)", style=solid];
+"435 permute_16" -> "436 select_9"  [label="(3, 16, 6, 64, 32)", style=solid];
+"435 permute_16" -> "437 select_10"  [label="(3, 16, 6, 64, 32)", style=solid];
+"435 permute_16" -> "438 select_11"  [label="(3, 16, 6, 64, 32)", style=solid];
+"436 select_9" -> "439 linalg_vector_norm_6"  [label="(16, 6, 64, 32)", style=solid];
+"436 select_9" -> "441 expand_as_6"  [label="(16, 6, 64, 32)", style=solid];
+"436 select_9" -> "442 div_6"  [label="(16, 6, 64, 32)", style=solid];
+"437 select_10" -> "445 linalg_vector_norm_7"  [label="(16, 6, 64, 32)", style=solid];
+"437 select_10" -> "447 expand_as_7"  [label="(16, 6, 64, 32)", style=solid];
+"437 select_10" -> "448 div_7"  [label="(16, 6, 64, 32)", style=solid];
+"438 select_11" -> "474 matmul_7"  [label="(16, 6, 64, 32)", style=solid];
+"439 linalg_vector_norm_6" -> "440 clamp_min_6"  [label="(16, 6, 64, 1)", style=solid];
+"440 clamp_min_6" -> "441 expand_as_6"  [label="(16, 6, 64, 1)", style=solid];
+"441 expand_as_6" -> "442 div_6"  [label="(16, 6, 64, 32)", style=solid];
+"442 div_6" -> "443 quantize_per_tensor_default_21"  [label="(16, 6, 64, 32)", style=solid];
+"443 quantize_per_tensor_default_21" -> "444 dequantize_per_tensor_default_21"  [label="(16, 6, 64, 32)", style=solid];
+"444 dequantize_per_tensor_default_21" -> "452 matmul_6"  [label="(16, 6, 64, 32)", style=solid];
+"445 linalg_vector_norm_7" -> "446 clamp_min_7"  [label="(16, 6, 64, 1)", style=solid];
+"446 clamp_min_7" -> "447 expand_as_7"  [label="(16, 6, 64, 1)", style=solid];
+"447 expand_as_7" -> "448 div_7"  [label="(16, 6, 64, 32)", style=solid];
+"448 div_7" -> "449 quantize_per_tensor_default_22"  [label="(16, 6, 64, 32)", style=solid];
+"449 quantize_per_tensor_default_22" -> "450 dequantize_per_tensor_default_22"  [label="(16, 6, 64, 32)", style=solid];
+"450 dequantize_per_tensor_default_22" -> "451 transpose_6"  [label="(16, 6, 64, 32)", style=solid];
+"451 transpose_6" -> "452 matmul_6"  [label="(16, 6, 32, 64)", style=solid];
+"452 matmul_6" -> "454 mul_7"  [label="(16, 6, 64, 64)", style=solid];
+"453 _frozen_param24" -> "454 mul_7"  [label="(6, 1, 1)", style=solid];
+"454 mul_7" -> "455 add_10"  [label="(16, 6, 64, 64)", style=solid];
+"455 add_10" -> "467 view_19"  [label="(16, 6, 64, 64)", style=solid];
+"456 new_zeros_1" -> "457 view_18"  [label="(32, 32)", style=solid];
+"457 view_18" -> "458 permute_17"  [label="(4, 8, 4, 8)", style=solid];
+"458 permute_17" -> "459 reshape_15"  [label="(4, 4, 8, 8)", style=solid];
+"459 reshape_15" -> "460 unsqueeze_8"  [label="(16, 64)", style=solid];
+"459 reshape_15" -> "461 unsqueeze_9"  [label="(16, 64)", style=solid];
+"460 unsqueeze_8" -> "462 sub_1"  [label="(16, 1, 64)", style=solid];
+"461 unsqueeze_9" -> "462 sub_1"  [label="(16, 64, 1)", style=solid];
+"462 sub_1" -> "463 ne_1"  [label="(16, 64, 64)", style=solid];
+"462 sub_1" -> "464 masked_fill_2"  [label="(16, 64, 64)", style=solid];
+"462 sub_1" -> "465 eq_1"  [label="(16, 64, 64)", style=solid];
+"463 ne_1" -> "464 masked_fill_2"  [label="(16, 64, 64)", style=solid];
+"464 masked_fill_2" -> "466 masked_fill_3"  [label="(16, 64, 64)", style=solid];
+"465 eq_1" -> "466 masked_fill_3"  [label="(16, 64, 64)", style=solid];
+"466 masked_fill_3" -> "468 unsqueeze_10"  [label="(16, 64, 64)", style=solid];
+"467 view_19" -> "470 add_11"  [label="(1, 16, 6, 64, 64)", style=solid];
+"468 unsqueeze_10" -> "469 unsqueeze_11"  [label="(16, 1, 64, 64)", style=solid];
+"469 unsqueeze_11" -> "470 add_11"  [label="(1, 16, 1, 64, 64)", style=solid];
+"470 add_11" -> "471 view_20"  [label="(1, 16, 6, 64, 64)", style=solid];
+"471 view_20" -> "472 softmax_3"  [label="(16, 6, 64, 64)", style=solid];
+"472 softmax_3" -> "473 dropout_12"  [label="(16, 6, 64, 64)", style=solid];
+"473 dropout_12" -> "474 matmul_7"  [label="(16, 6, 64, 64)", style=solid];
+"474 matmul_7" -> "475 transpose_7"  [label="(16, 6, 64, 32)", style=solid];
+"475 transpose_7" -> "476 reshape_16"  [label="(16, 64, 6, 32)", style=solid];
+"476 reshape_16" -> "477 reshape_16_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
+"477 reshape_16_0_0_nncf_smooth_quant_0" -> "478 quantize_per_tensor_default_23"  [label="(16, 64, 192)", style=solid];
+"478 quantize_per_tensor_default_23" -> "479 dequantize_per_tensor_default_23"  [label="(16, 64, 192)", style=solid];
+"479 dequantize_per_tensor_default_23" -> "486 linear_22"  [label="(16, 64, 192)", style=solid];
+"480 features_3_1_attn_proj_bias_0_0" -> "486 linear_22"  [label="(192,)", style=solid];
+"481 scale_updated_constant15" -> "483 mul_tensor_15"  [label="(192, 1)", style=solid];
+"482 compressed_weight_updated_constant15" -> "483 mul_tensor_15"  [label="(192, 192)", style=solid];
+"483 mul_tensor_15" -> "485 sub_tensor_15"  [label="(192, 192)", style=solid];
+"484 zero_point_updated_constant15" -> "485 sub_tensor_15"  [label="(192, 1)", style=solid];
+"485 sub_tensor_15" -> "486 linear_22"  [label="(192, 192)", style=solid];
+"486 linear_22" -> "487 dropout_13"  [label="(16, 64, 192)", style=solid];
+"487 dropout_13" -> "488 view_21"  [label="(16, 64, 192)", style=solid];
+"488 view_21" -> "489 permute_18"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"489 permute_18" -> "490 reshape_17"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"490 reshape_17" -> "491 roll_3"  [label="(1, 32, 32, 192)", style=solid];
+"491 roll_3" -> "492 slice_61"  [label="(1, 32, 32, 192)", style=solid];
+"492 slice_61" -> "493 slice_62"  [label="(1, 32, 32, 192)", style=solid];
+"493 slice_62" -> "494 slice_63"  [label="(1, 28, 32, 192)", style=solid];
+"494 slice_63" -> "495 slice_64"  [label="(1, 28, 28, 192)", style=solid];
+"495 slice_64" -> "496 contiguous_5"  [label="(1, 28, 28, 192)", style=solid];
+"496 contiguous_5" -> "497 layer_norm_8"  [label="(1, 28, 28, 192)", style=solid];
+"497 layer_norm_8" -> "498 add_12"  [label="(1, 28, 28, 192)", style=solid];
+"498 add_12" -> "499 add_12_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 192)", style=solid];
+"498 add_12" -> "523 add_13"  [label="(1, 28, 28, 192)", style=solid];
+"499 add_12_0_0_nncf_smooth_quant_0" -> "500 quantize_per_tensor_default_24"  [label="(1, 28, 28, 192)", style=solid];
+"500 quantize_per_tensor_default_24" -> "501 dequantize_per_tensor_default_24"  [label="(1, 28, 28, 192)", style=solid];
+"501 dequantize_per_tensor_default_24" -> "508 linear_23"  [label="(1, 28, 28, 192)", style=solid];
+"502 features_3_1_mlp_0_bias_0_0" -> "508 linear_23"  [label="(768,)", style=solid];
+"503 scale_updated_constant16" -> "505 mul_tensor_16"  [label="(768, 1)", style=solid];
+"504 compressed_weight_updated_constant16" -> "505 mul_tensor_16"  [label="(768, 192)", style=solid];
+"505 mul_tensor_16" -> "507 sub_tensor_16"  [label="(768, 192)", style=solid];
+"506 zero_point_updated_constant16" -> "507 sub_tensor_16"  [label="(768, 1)", style=solid];
+"507 sub_tensor_16" -> "508 linear_23"  [label="(768, 192)", style=solid];
+"508 linear_23" -> "509 gelu_3"  [label="(1, 28, 28, 768)", style=solid];
+"509 gelu_3" -> "510 dropout_14"  [label="(1, 28, 28, 768)", style=solid];
+"510 dropout_14" -> "511 dropout_14_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 768)", style=solid];
+"511 dropout_14_0_0_nncf_smooth_quant_0" -> "512 quantize_per_tensor_default_25"  [label="(1, 28, 28, 768)", style=solid];
+"512 quantize_per_tensor_default_25" -> "513 dequantize_per_tensor_default_25"  [label="(1, 28, 28, 768)", style=solid];
+"513 dequantize_per_tensor_default_25" -> "520 linear_24"  [label="(1, 28, 28, 768)", style=solid];
+"514 features_3_1_mlp_3_bias_0_0" -> "520 linear_24"  [label="(192,)", style=solid];
+"515 scale_updated_constant17" -> "517 mul_tensor_17"  [label="(192, 1)", style=solid];
+"516 compressed_weight_updated_constant17" -> "517 mul_tensor_17"  [label="(192, 768)", style=solid];
+"517 mul_tensor_17" -> "519 sub_tensor_17"  [label="(192, 768)", style=solid];
+"518 zero_point_updated_constant17" -> "519 sub_tensor_17"  [label="(192, 1)", style=solid];
+"519 sub_tensor_17" -> "520 linear_24"  [label="(192, 768)", style=solid];
+"520 linear_24" -> "521 dropout_15"  [label="(1, 28, 28, 192)", style=solid];
+"521 dropout_15" -> "522 layer_norm_9"  [label="(1, 28, 28, 192)", style=solid];
+"522 layer_norm_9" -> "523 add_13"  [label="(1, 28, 28, 192)", style=solid];
+"523 add_13" -> "524 pad_5"  [label="(1, 28, 28, 192)", style=solid];
+"524 pad_5" -> "525 slice_65"  [label="(1, 28, 28, 192)", style=solid];
+"524 pad_5" -> "528 slice_68"  [label="(1, 28, 28, 192)", style=solid];
+"524 pad_5" -> "531 slice_71"  [label="(1, 28, 28, 192)", style=solid];
+"524 pad_5" -> "534 slice_74"  [label="(1, 28, 28, 192)", style=solid];
+"525 slice_65" -> "526 slice_66"  [label="(1, 14, 28, 192)", style=solid];
+"526 slice_66" -> "527 slice_67"  [label="(1, 14, 14, 192)", style=solid];
+"527 slice_67" -> "537 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"528 slice_68" -> "529 slice_69"  [label="(1, 14, 28, 192)", style=solid];
+"529 slice_69" -> "530 slice_70"  [label="(1, 14, 14, 192)", style=solid];
+"530 slice_70" -> "537 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"531 slice_71" -> "532 slice_72"  [label="(1, 14, 28, 192)", style=solid];
+"532 slice_72" -> "533 slice_73"  [label="(1, 14, 14, 192)", style=solid];
+"533 slice_73" -> "537 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"534 slice_74" -> "535 slice_75"  [label="(1, 14, 28, 192)", style=solid];
+"535 slice_75" -> "536 slice_76"  [label="(1, 14, 14, 192)", style=solid];
+"536 slice_76" -> "537 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"537 cat_1" -> "538 cat_1_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 768)", style=solid];
+"538 cat_1_0_0_nncf_smooth_quant_0" -> "539 quantize_per_tensor_default_26"  [label="(1, 14, 14, 768)", style=solid];
+"539 quantize_per_tensor_default_26" -> "540 dequantize_per_tensor_default_26"  [label="(1, 14, 14, 768)", style=solid];
+"540 dequantize_per_tensor_default_26" -> "546 linear_25"  [label="(1, 14, 14, 768)", style=solid];
+"541 scale_updated_constant18" -> "543 mul_tensor_18"  [label="(384, 1)", style=solid];
+"542 compressed_weight_updated_constant18" -> "543 mul_tensor_18"  [label="(384, 768)", style=solid];
+"543 mul_tensor_18" -> "545 sub_tensor_18"  [label="(384, 768)", style=solid];
+"544 zero_point_updated_constant18" -> "545 sub_tensor_18"  [label="(384, 1)", style=solid];
+"545 sub_tensor_18" -> "546 linear_25"  [label="(384, 768)", style=solid];
+"546 linear_25" -> "547 layer_norm_10"  [label="(1, 14, 14, 384)", style=solid];
+"547 layer_norm_10" -> "549 pad_6"  [label="(1, 14, 14, 384)", style=solid];
+"547 layer_norm_10" -> "610 add_15"  [label="(1, 14, 14, 384)", style=solid];
+"548 _frozen_param34" -> "584 add_14"  [label="(1, 12, 64, 64)", style=solid];
+"549 pad_6" -> "550 view_24"  [label="(1, 16, 16, 384)", style=solid];
+"550 view_24" -> "551 permute_20"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"551 permute_20" -> "552 reshape_18"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"552 reshape_18" -> "553 reshape_18_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"553 reshape_18_0_0_nncf_smooth_quant_0" -> "554 quantize_per_tensor_default_27"  [label="(4, 64, 384)", style=solid];
+"554 quantize_per_tensor_default_27" -> "555 dequantize_per_tensor_default_27"  [label="(4, 64, 384)", style=solid];
+"555 dequantize_per_tensor_default_27" -> "562 linear_28"  [label="(4, 64, 384)", style=solid];
+"556 _frozen_param35_0_0" -> "562 linear_28"  [label="(1152,)", style=solid];
+"557 scale_updated_constant19" -> "559 mul_tensor_19"  [label="(1152, 1)", style=solid];
+"558 compressed_weight_updated_constant19" -> "559 mul_tensor_19"  [label="(1152, 384)", style=solid];
+"559 mul_tensor_19" -> "561 sub_tensor_19"  [label="(1152, 384)", style=solid];
+"560 zero_point_updated_constant19" -> "561 sub_tensor_19"  [label="(1152, 1)", style=solid];
+"561 sub_tensor_19" -> "562 linear_28"  [label="(1152, 384)", style=solid];
+"562 linear_28" -> "563 reshape_19"  [label="(4, 64, 1152)", style=solid];
+"563 reshape_19" -> "564 permute_21"  [label="(4, 64, 3, 12, 32)", style=solid];
+"564 permute_21" -> "565 select_12"  [label="(3, 4, 12, 64, 32)", style=solid];
+"564 permute_21" -> "566 select_13"  [label="(3, 4, 12, 64, 32)", style=solid];
+"564 permute_21" -> "567 select_14"  [label="(3, 4, 12, 64, 32)", style=solid];
+"565 select_12" -> "568 linalg_vector_norm_8"  [label="(4, 12, 64, 32)", style=solid];
+"565 select_12" -> "570 expand_as_8"  [label="(4, 12, 64, 32)", style=solid];
+"565 select_12" -> "571 div_8"  [label="(4, 12, 64, 32)", style=solid];
+"566 select_13" -> "574 linalg_vector_norm_9"  [label="(4, 12, 64, 32)", style=solid];
+"566 select_13" -> "576 expand_as_9"  [label="(4, 12, 64, 32)", style=solid];
+"566 select_13" -> "577 div_9"  [label="(4, 12, 64, 32)", style=solid];
+"567 select_14" -> "587 matmul_9"  [label="(4, 12, 64, 32)", style=solid];
+"568 linalg_vector_norm_8" -> "569 clamp_min_8"  [label="(4, 12, 64, 1)", style=solid];
+"569 clamp_min_8" -> "570 expand_as_8"  [label="(4, 12, 64, 1)", style=solid];
+"570 expand_as_8" -> "571 div_8"  [label="(4, 12, 64, 32)", style=solid];
+"571 div_8" -> "572 quantize_per_tensor_default_28"  [label="(4, 12, 64, 32)", style=solid];
+"572 quantize_per_tensor_default_28" -> "573 dequantize_per_tensor_default_28"  [label="(4, 12, 64, 32)", style=solid];
+"573 dequantize_per_tensor_default_28" -> "581 matmul_8"  [label="(4, 12, 64, 32)", style=solid];
+"574 linalg_vector_norm_9" -> "575 clamp_min_9"  [label="(4, 12, 64, 1)", style=solid];
+"575 clamp_min_9" -> "576 expand_as_9"  [label="(4, 12, 64, 1)", style=solid];
+"576 expand_as_9" -> "577 div_9"  [label="(4, 12, 64, 32)", style=solid];
+"577 div_9" -> "578 quantize_per_tensor_default_29"  [label="(4, 12, 64, 32)", style=solid];
+"578 quantize_per_tensor_default_29" -> "579 dequantize_per_tensor_default_29"  [label="(4, 12, 64, 32)", style=solid];
+"579 dequantize_per_tensor_default_29" -> "580 transpose_8"  [label="(4, 12, 64, 32)", style=solid];
+"580 transpose_8" -> "581 matmul_8"  [label="(4, 12, 32, 64)", style=solid];
+"581 matmul_8" -> "583 mul_9"  [label="(4, 12, 64, 64)", style=solid];
+"582 _frozen_param37" -> "583 mul_9"  [label="(12, 1, 1)", style=solid];
+"583 mul_9" -> "584 add_14"  [label="(4, 12, 64, 64)", style=solid];
+"584 add_14" -> "585 softmax_4"  [label="(4, 12, 64, 64)", style=solid];
+"585 softmax_4" -> "586 dropout_16"  [label="(4, 12, 64, 64)", style=solid];
+"586 dropout_16" -> "587 matmul_9"  [label="(4, 12, 64, 64)", style=solid];
+"587 matmul_9" -> "588 transpose_9"  [label="(4, 12, 64, 32)", style=solid];
+"588 transpose_9" -> "589 reshape_20"  [label="(4, 64, 12, 32)", style=solid];
+"589 reshape_20" -> "590 reshape_20_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"590 reshape_20_0_0_nncf_smooth_quant_0" -> "591 quantize_per_tensor_default_30"  [label="(4, 64, 384)", style=solid];
+"591 quantize_per_tensor_default_30" -> "592 dequantize_per_tensor_default_30"  [label="(4, 64, 384)", style=solid];
+"592 dequantize_per_tensor_default_30" -> "599 linear_29"  [label="(4, 64, 384)", style=solid];
+"593 features_5_0_attn_proj_bias_0_0" -> "599 linear_29"  [label="(384,)", style=solid];
+"594 scale_updated_constant20" -> "596 mul_tensor_20"  [label="(384, 1)", style=solid];
+"595 compressed_weight_updated_constant20" -> "596 mul_tensor_20"  [label="(384, 384)", style=solid];
+"596 mul_tensor_20" -> "598 sub_tensor_20"  [label="(384, 384)", style=solid];
+"597 zero_point_updated_constant20" -> "598 sub_tensor_20"  [label="(384, 1)", style=solid];
+"598 sub_tensor_20" -> "599 linear_29"  [label="(384, 384)", style=solid];
+"599 linear_29" -> "600 dropout_17"  [label="(4, 64, 384)", style=solid];
+"600 dropout_17" -> "601 view_25"  [label="(4, 64, 384)", style=solid];
+"601 view_25" -> "602 permute_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"602 permute_22" -> "603 reshape_21"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"603 reshape_21" -> "604 slice_78"  [label="(1, 16, 16, 384)", style=solid];
+"604 slice_78" -> "605 slice_79"  [label="(1, 16, 16, 384)", style=solid];
+"605 slice_79" -> "606 slice_80"  [label="(1, 14, 16, 384)", style=solid];
+"606 slice_80" -> "607 slice_81"  [label="(1, 14, 14, 384)", style=solid];
+"607 slice_81" -> "608 contiguous_7"  [label="(1, 14, 14, 384)", style=solid];
+"608 contiguous_7" -> "609 layer_norm_11"  [label="(1, 14, 14, 384)", style=solid];
+"609 layer_norm_11" -> "610 add_15"  [label="(1, 14, 14, 384)", style=solid];
+"610 add_15" -> "611 add_15_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"610 add_15" -> "635 add_16"  [label="(1, 14, 14, 384)", style=solid];
+"611 add_15_0_0_nncf_smooth_quant_0" -> "612 quantize_per_tensor_default_31"  [label="(1, 14, 14, 384)", style=solid];
+"612 quantize_per_tensor_default_31" -> "613 dequantize_per_tensor_default_31"  [label="(1, 14, 14, 384)", style=solid];
+"613 dequantize_per_tensor_default_31" -> "620 linear_30"  [label="(1, 14, 14, 384)", style=solid];
+"614 features_5_0_mlp_0_bias_0_0" -> "620 linear_30"  [label="(1536,)", style=solid];
+"615 scale_updated_constant21" -> "617 mul_tensor_21"  [label="(1536, 1)", style=solid];
+"616 compressed_weight_updated_constant21" -> "617 mul_tensor_21"  [label="(1536, 384)", style=solid];
+"617 mul_tensor_21" -> "619 sub_tensor_21"  [label="(1536, 384)", style=solid];
+"618 zero_point_updated_constant21" -> "619 sub_tensor_21"  [label="(1536, 1)", style=solid];
+"619 sub_tensor_21" -> "620 linear_30"  [label="(1536, 384)", style=solid];
+"620 linear_30" -> "621 gelu_4"  [label="(1, 14, 14, 1536)", style=solid];
+"621 gelu_4" -> "622 dropout_18"  [label="(1, 14, 14, 1536)", style=solid];
+"622 dropout_18" -> "623 dropout_18_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"623 dropout_18_0_0_nncf_smooth_quant_0" -> "624 quantize_per_tensor_default_32"  [label="(1, 14, 14, 1536)", style=solid];
+"624 quantize_per_tensor_default_32" -> "625 dequantize_per_tensor_default_32"  [label="(1, 14, 14, 1536)", style=solid];
+"625 dequantize_per_tensor_default_32" -> "632 linear_31"  [label="(1, 14, 14, 1536)", style=solid];
+"626 features_5_0_mlp_3_bias_0_0" -> "632 linear_31"  [label="(384,)", style=solid];
+"627 scale_updated_constant22" -> "629 mul_tensor_22"  [label="(384, 1)", style=solid];
+"628 compressed_weight_updated_constant22" -> "629 mul_tensor_22"  [label="(384, 1536)", style=solid];
+"629 mul_tensor_22" -> "631 sub_tensor_22"  [label="(384, 1536)", style=solid];
+"630 zero_point_updated_constant22" -> "631 sub_tensor_22"  [label="(384, 1)", style=solid];
+"631 sub_tensor_22" -> "632 linear_31"  [label="(384, 1536)", style=solid];
+"632 linear_31" -> "633 dropout_19"  [label="(1, 14, 14, 384)", style=solid];
+"633 dropout_19" -> "634 layer_norm_12"  [label="(1, 14, 14, 384)", style=solid];
+"634 layer_norm_12" -> "635 add_16"  [label="(1, 14, 14, 384)", style=solid];
+"635 add_16" -> "637 pad_7"  [label="(1, 14, 14, 384)", style=solid];
+"635 add_16" -> "716 add_19"  [label="(1, 14, 14, 384)", style=solid];
+"636 _frozen_param38" -> "673 add_17"  [label="(1, 12, 64, 64)", style=solid];
+"637 pad_7" -> "638 roll_4"  [label="(1, 16, 16, 384)", style=solid];
+"638 roll_4" -> "639 view_28"  [label="(1, 16, 16, 384)", style=solid];
+"639 view_28" -> "640 permute_24"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"640 permute_24" -> "641 reshape_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"641 reshape_22" -> "642 reshape_22_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"641 reshape_22" -> "674 new_zeros_2"  [label="(4, 64, 384)", style=solid];
+"642 reshape_22_0_0_nncf_smooth_quant_0" -> "643 quantize_per_tensor_default_33"  [label="(4, 64, 384)", style=solid];
+"643 quantize_per_tensor_default_33" -> "644 dequantize_per_tensor_default_33"  [label="(4, 64, 384)", style=solid];
+"644 dequantize_per_tensor_default_33" -> "651 linear_34"  [label="(4, 64, 384)", style=solid];
+"645 _frozen_param39_0_0" -> "651 linear_34"  [label="(1152,)", style=solid];
+"646 scale_updated_constant23" -> "648 mul_tensor_23"  [label="(1152, 1)", style=solid];
+"647 compressed_weight_updated_constant23" -> "648 mul_tensor_23"  [label="(1152, 384)", style=solid];
+"648 mul_tensor_23" -> "650 sub_tensor_23"  [label="(1152, 384)", style=solid];
+"649 zero_point_updated_constant23" -> "650 sub_tensor_23"  [label="(1152, 1)", style=solid];
+"650 sub_tensor_23" -> "651 linear_34"  [label="(1152, 384)", style=solid];
+"651 linear_34" -> "652 reshape_23"  [label="(4, 64, 1152)", style=solid];
+"652 reshape_23" -> "653 permute_25"  [label="(4, 64, 3, 12, 32)", style=solid];
+"653 permute_25" -> "654 select_15"  [label="(3, 4, 12, 64, 32)", style=solid];
+"653 permute_25" -> "655 select_16"  [label="(3, 4, 12, 64, 32)", style=solid];
+"653 permute_25" -> "656 select_17"  [label="(3, 4, 12, 64, 32)", style=solid];
+"654 select_15" -> "657 linalg_vector_norm_10"  [label="(4, 12, 64, 32)", style=solid];
+"654 select_15" -> "659 expand_as_10"  [label="(4, 12, 64, 32)", style=solid];
+"654 select_15" -> "660 div_10"  [label="(4, 12, 64, 32)", style=solid];
+"655 select_16" -> "663 linalg_vector_norm_11"  [label="(4, 12, 64, 32)", style=solid];
+"655 select_16" -> "665 expand_as_11"  [label="(4, 12, 64, 32)", style=solid];
+"655 select_16" -> "666 div_11"  [label="(4, 12, 64, 32)", style=solid];
+"656 select_17" -> "692 matmul_11"  [label="(4, 12, 64, 32)", style=solid];
+"657 linalg_vector_norm_10" -> "658 clamp_min_10"  [label="(4, 12, 64, 1)", style=solid];
+"658 clamp_min_10" -> "659 expand_as_10"  [label="(4, 12, 64, 1)", style=solid];
+"659 expand_as_10" -> "660 div_10"  [label="(4, 12, 64, 32)", style=solid];
+"660 div_10" -> "661 quantize_per_tensor_default_34"  [label="(4, 12, 64, 32)", style=solid];
+"661 quantize_per_tensor_default_34" -> "662 dequantize_per_tensor_default_34"  [label="(4, 12, 64, 32)", style=solid];
+"662 dequantize_per_tensor_default_34" -> "670 matmul_10"  [label="(4, 12, 64, 32)", style=solid];
+"663 linalg_vector_norm_11" -> "664 clamp_min_11"  [label="(4, 12, 64, 1)", style=solid];
+"664 clamp_min_11" -> "665 expand_as_11"  [label="(4, 12, 64, 1)", style=solid];
+"665 expand_as_11" -> "666 div_11"  [label="(4, 12, 64, 32)", style=solid];
+"666 div_11" -> "667 quantize_per_tensor_default_35"  [label="(4, 12, 64, 32)", style=solid];
+"667 quantize_per_tensor_default_35" -> "668 dequantize_per_tensor_default_35"  [label="(4, 12, 64, 32)", style=solid];
+"668 dequantize_per_tensor_default_35" -> "669 transpose_10"  [label="(4, 12, 64, 32)", style=solid];
+"669 transpose_10" -> "670 matmul_10"  [label="(4, 12, 32, 64)", style=solid];
+"670 matmul_10" -> "672 mul_11"  [label="(4, 12, 64, 64)", style=solid];
+"671 _frozen_param41" -> "672 mul_11"  [label="(12, 1, 1)", style=solid];
+"672 mul_11" -> "673 add_17"  [label="(4, 12, 64, 64)", style=solid];
+"673 add_17" -> "685 view_30"  [label="(4, 12, 64, 64)", style=solid];
+"674 new_zeros_2" -> "675 view_29"  [label="(16, 16)", style=solid];
+"675 view_29" -> "676 permute_26"  [label="(2, 8, 2, 8)", style=solid];
+"676 permute_26" -> "677 reshape_24"  [label="(2, 2, 8, 8)", style=solid];
+"677 reshape_24" -> "678 unsqueeze_14"  [label="(4, 64)", style=solid];
+"677 reshape_24" -> "679 unsqueeze_15"  [label="(4, 64)", style=solid];
+"678 unsqueeze_14" -> "680 sub_2"  [label="(4, 1, 64)", style=solid];
+"679 unsqueeze_15" -> "680 sub_2"  [label="(4, 64, 1)", style=solid];
+"680 sub_2" -> "681 ne_2"  [label="(4, 64, 64)", style=solid];
+"680 sub_2" -> "682 masked_fill_4"  [label="(4, 64, 64)", style=solid];
+"680 sub_2" -> "683 eq_2"  [label="(4, 64, 64)", style=solid];
+"681 ne_2" -> "682 masked_fill_4"  [label="(4, 64, 64)", style=solid];
+"682 masked_fill_4" -> "684 masked_fill_5"  [label="(4, 64, 64)", style=solid];
+"683 eq_2" -> "684 masked_fill_5"  [label="(4, 64, 64)", style=solid];
+"684 masked_fill_5" -> "686 unsqueeze_16"  [label="(4, 64, 64)", style=solid];
+"685 view_30" -> "688 add_18"  [label="(1, 4, 12, 64, 64)", style=solid];
+"686 unsqueeze_16" -> "687 unsqueeze_17"  [label="(4, 1, 64, 64)", style=solid];
+"687 unsqueeze_17" -> "688 add_18"  [label="(1, 4, 1, 64, 64)", style=solid];
+"688 add_18" -> "689 view_31"  [label="(1, 4, 12, 64, 64)", style=solid];
+"689 view_31" -> "690 softmax_5"  [label="(4, 12, 64, 64)", style=solid];
+"690 softmax_5" -> "691 dropout_20"  [label="(4, 12, 64, 64)", style=solid];
+"691 dropout_20" -> "692 matmul_11"  [label="(4, 12, 64, 64)", style=solid];
+"692 matmul_11" -> "693 transpose_11"  [label="(4, 12, 64, 32)", style=solid];
+"693 transpose_11" -> "694 reshape_25"  [label="(4, 64, 12, 32)", style=solid];
+"694 reshape_25" -> "695 reshape_25_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"695 reshape_25_0_0_nncf_smooth_quant_0" -> "696 quantize_per_tensor_default_36"  [label="(4, 64, 384)", style=solid];
+"696 quantize_per_tensor_default_36" -> "697 dequantize_per_tensor_default_36"  [label="(4, 64, 384)", style=solid];
+"697 dequantize_per_tensor_default_36" -> "704 linear_35"  [label="(4, 64, 384)", style=solid];
+"698 features_5_1_attn_proj_bias_0_0" -> "704 linear_35"  [label="(384,)", style=solid];
+"699 scale_updated_constant24" -> "701 mul_tensor_24"  [label="(384, 1)", style=solid];
+"700 compressed_weight_updated_constant24" -> "701 mul_tensor_24"  [label="(384, 384)", style=solid];
+"701 mul_tensor_24" -> "703 sub_tensor_24"  [label="(384, 384)", style=solid];
+"702 zero_point_updated_constant24" -> "703 sub_tensor_24"  [label="(384, 1)", style=solid];
+"703 sub_tensor_24" -> "704 linear_35"  [label="(384, 384)", style=solid];
+"704 linear_35" -> "705 dropout_21"  [label="(4, 64, 384)", style=solid];
+"705 dropout_21" -> "706 view_32"  [label="(4, 64, 384)", style=solid];
+"706 view_32" -> "707 permute_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"707 permute_27" -> "708 reshape_26"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"708 reshape_26" -> "709 roll_5"  [label="(1, 16, 16, 384)", style=solid];
+"709 roll_5" -> "710 slice_101"  [label="(1, 16, 16, 384)", style=solid];
+"710 slice_101" -> "711 slice_102"  [label="(1, 16, 16, 384)", style=solid];
+"711 slice_102" -> "712 slice_103"  [label="(1, 14, 16, 384)", style=solid];
+"712 slice_103" -> "713 slice_104"  [label="(1, 14, 14, 384)", style=solid];
+"713 slice_104" -> "714 contiguous_9"  [label="(1, 14, 14, 384)", style=solid];
+"714 contiguous_9" -> "715 layer_norm_13"  [label="(1, 14, 14, 384)", style=solid];
+"715 layer_norm_13" -> "716 add_19"  [label="(1, 14, 14, 384)", style=solid];
+"716 add_19" -> "717 add_19_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"716 add_19" -> "741 add_20"  [label="(1, 14, 14, 384)", style=solid];
+"717 add_19_0_0_nncf_smooth_quant_0" -> "718 quantize_per_tensor_default_37"  [label="(1, 14, 14, 384)", style=solid];
+"718 quantize_per_tensor_default_37" -> "719 dequantize_per_tensor_default_37"  [label="(1, 14, 14, 384)", style=solid];
+"719 dequantize_per_tensor_default_37" -> "726 linear_36"  [label="(1, 14, 14, 384)", style=solid];
+"720 features_5_1_mlp_0_bias_0_0" -> "726 linear_36"  [label="(1536,)", style=solid];
+"721 scale_updated_constant25" -> "723 mul_tensor_25"  [label="(1536, 1)", style=solid];
+"722 compressed_weight_updated_constant25" -> "723 mul_tensor_25"  [label="(1536, 384)", style=solid];
+"723 mul_tensor_25" -> "725 sub_tensor_25"  [label="(1536, 384)", style=solid];
+"724 zero_point_updated_constant25" -> "725 sub_tensor_25"  [label="(1536, 1)", style=solid];
+"725 sub_tensor_25" -> "726 linear_36"  [label="(1536, 384)", style=solid];
+"726 linear_36" -> "727 gelu_5"  [label="(1, 14, 14, 1536)", style=solid];
+"727 gelu_5" -> "728 dropout_22"  [label="(1, 14, 14, 1536)", style=solid];
+"728 dropout_22" -> "729 dropout_22_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"729 dropout_22_0_0_nncf_smooth_quant_0" -> "730 quantize_per_tensor_default_38"  [label="(1, 14, 14, 1536)", style=solid];
+"730 quantize_per_tensor_default_38" -> "731 dequantize_per_tensor_default_38"  [label="(1, 14, 14, 1536)", style=solid];
+"731 dequantize_per_tensor_default_38" -> "738 linear_37"  [label="(1, 14, 14, 1536)", style=solid];
+"732 features_5_1_mlp_3_bias_0_0" -> "738 linear_37"  [label="(384,)", style=solid];
+"733 scale_updated_constant26" -> "735 mul_tensor_26"  [label="(384, 1)", style=solid];
+"734 compressed_weight_updated_constant26" -> "735 mul_tensor_26"  [label="(384, 1536)", style=solid];
+"735 mul_tensor_26" -> "737 sub_tensor_26"  [label="(384, 1536)", style=solid];
+"736 zero_point_updated_constant26" -> "737 sub_tensor_26"  [label="(384, 1)", style=solid];
+"737 sub_tensor_26" -> "738 linear_37"  [label="(384, 1536)", style=solid];
+"738 linear_37" -> "739 dropout_23"  [label="(1, 14, 14, 384)", style=solid];
+"739 dropout_23" -> "740 layer_norm_14"  [label="(1, 14, 14, 384)", style=solid];
+"740 layer_norm_14" -> "741 add_20"  [label="(1, 14, 14, 384)", style=solid];
+"741 add_20" -> "743 pad_8"  [label="(1, 14, 14, 384)", style=solid];
+"741 add_20" -> "804 add_22"  [label="(1, 14, 14, 384)", style=solid];
+"742 _frozen_param51" -> "778 add_21"  [label="(1, 12, 64, 64)", style=solid];
+"743 pad_8" -> "744 view_35"  [label="(1, 16, 16, 384)", style=solid];
+"744 view_35" -> "745 permute_29"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"745 permute_29" -> "746 reshape_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"746 reshape_27" -> "747 reshape_27_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"747 reshape_27_0_0_nncf_smooth_quant_0" -> "748 quantize_per_tensor_default_39"  [label="(4, 64, 384)", style=solid];
+"748 quantize_per_tensor_default_39" -> "749 dequantize_per_tensor_default_39"  [label="(4, 64, 384)", style=solid];
+"749 dequantize_per_tensor_default_39" -> "756 linear_40"  [label="(4, 64, 384)", style=solid];
+"750 _frozen_param52_0_0" -> "756 linear_40"  [label="(1152,)", style=solid];
+"751 scale_updated_constant27" -> "753 mul_tensor_27"  [label="(1152, 1)", style=solid];
+"752 compressed_weight_updated_constant27" -> "753 mul_tensor_27"  [label="(1152, 384)", style=solid];
+"753 mul_tensor_27" -> "755 sub_tensor_27"  [label="(1152, 384)", style=solid];
+"754 zero_point_updated_constant27" -> "755 sub_tensor_27"  [label="(1152, 1)", style=solid];
+"755 sub_tensor_27" -> "756 linear_40"  [label="(1152, 384)", style=solid];
+"756 linear_40" -> "757 reshape_28"  [label="(4, 64, 1152)", style=solid];
+"757 reshape_28" -> "758 permute_30"  [label="(4, 64, 3, 12, 32)", style=solid];
+"758 permute_30" -> "759 select_18"  [label="(3, 4, 12, 64, 32)", style=solid];
+"758 permute_30" -> "760 select_19"  [label="(3, 4, 12, 64, 32)", style=solid];
+"758 permute_30" -> "761 select_20"  [label="(3, 4, 12, 64, 32)", style=solid];
+"759 select_18" -> "762 linalg_vector_norm_12"  [label="(4, 12, 64, 32)", style=solid];
+"759 select_18" -> "764 expand_as_12"  [label="(4, 12, 64, 32)", style=solid];
+"759 select_18" -> "765 div_12"  [label="(4, 12, 64, 32)", style=solid];
+"760 select_19" -> "768 linalg_vector_norm_13"  [label="(4, 12, 64, 32)", style=solid];
+"760 select_19" -> "770 expand_as_13"  [label="(4, 12, 64, 32)", style=solid];
+"760 select_19" -> "771 div_13"  [label="(4, 12, 64, 32)", style=solid];
+"761 select_20" -> "781 matmul_13"  [label="(4, 12, 64, 32)", style=solid];
+"762 linalg_vector_norm_12" -> "763 clamp_min_12"  [label="(4, 12, 64, 1)", style=solid];
+"763 clamp_min_12" -> "764 expand_as_12"  [label="(4, 12, 64, 1)", style=solid];
+"764 expand_as_12" -> "765 div_12"  [label="(4, 12, 64, 32)", style=solid];
+"765 div_12" -> "766 quantize_per_tensor_default_40"  [label="(4, 12, 64, 32)", style=solid];
+"766 quantize_per_tensor_default_40" -> "767 dequantize_per_tensor_default_40"  [label="(4, 12, 64, 32)", style=solid];
+"767 dequantize_per_tensor_default_40" -> "775 matmul_12"  [label="(4, 12, 64, 32)", style=solid];
+"768 linalg_vector_norm_13" -> "769 clamp_min_13"  [label="(4, 12, 64, 1)", style=solid];
+"769 clamp_min_13" -> "770 expand_as_13"  [label="(4, 12, 64, 1)", style=solid];
+"770 expand_as_13" -> "771 div_13"  [label="(4, 12, 64, 32)", style=solid];
+"771 div_13" -> "772 quantize_per_tensor_default_41"  [label="(4, 12, 64, 32)", style=solid];
+"772 quantize_per_tensor_default_41" -> "773 dequantize_per_tensor_default_41"  [label="(4, 12, 64, 32)", style=solid];
+"773 dequantize_per_tensor_default_41" -> "774 transpose_12"  [label="(4, 12, 64, 32)", style=solid];
+"774 transpose_12" -> "775 matmul_12"  [label="(4, 12, 32, 64)", style=solid];
+"775 matmul_12" -> "777 mul_13"  [label="(4, 12, 64, 64)", style=solid];
+"776 _frozen_param54" -> "777 mul_13"  [label="(12, 1, 1)", style=solid];
+"777 mul_13" -> "778 add_21"  [label="(4, 12, 64, 64)", style=solid];
+"778 add_21" -> "779 softmax_6"  [label="(4, 12, 64, 64)", style=solid];
+"779 softmax_6" -> "780 dropout_24"  [label="(4, 12, 64, 64)", style=solid];
+"780 dropout_24" -> "781 matmul_13"  [label="(4, 12, 64, 64)", style=solid];
+"781 matmul_13" -> "782 transpose_13"  [label="(4, 12, 64, 32)", style=solid];
+"782 transpose_13" -> "783 reshape_29"  [label="(4, 64, 12, 32)", style=solid];
+"783 reshape_29" -> "784 reshape_29_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"784 reshape_29_0_0_nncf_smooth_quant_0" -> "785 quantize_per_tensor_default_42"  [label="(4, 64, 384)", style=solid];
+"785 quantize_per_tensor_default_42" -> "786 dequantize_per_tensor_default_42"  [label="(4, 64, 384)", style=solid];
+"786 dequantize_per_tensor_default_42" -> "793 linear_41"  [label="(4, 64, 384)", style=solid];
+"787 features_5_2_attn_proj_bias_0_0" -> "793 linear_41"  [label="(384,)", style=solid];
+"788 scale_updated_constant28" -> "790 mul_tensor_28"  [label="(384, 1)", style=solid];
+"789 compressed_weight_updated_constant28" -> "790 mul_tensor_28"  [label="(384, 384)", style=solid];
+"790 mul_tensor_28" -> "792 sub_tensor_28"  [label="(384, 384)", style=solid];
+"791 zero_point_updated_constant28" -> "792 sub_tensor_28"  [label="(384, 1)", style=solid];
+"792 sub_tensor_28" -> "793 linear_41"  [label="(384, 384)", style=solid];
+"793 linear_41" -> "794 dropout_25"  [label="(4, 64, 384)", style=solid];
+"794 dropout_25" -> "795 view_36"  [label="(4, 64, 384)", style=solid];
+"795 view_36" -> "796 permute_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"796 permute_31" -> "797 reshape_30"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"797 reshape_30" -> "798 slice_106"  [label="(1, 16, 16, 384)", style=solid];
+"798 slice_106" -> "799 slice_107"  [label="(1, 16, 16, 384)", style=solid];
+"799 slice_107" -> "800 slice_108"  [label="(1, 14, 16, 384)", style=solid];
+"800 slice_108" -> "801 slice_109"  [label="(1, 14, 14, 384)", style=solid];
+"801 slice_109" -> "802 contiguous_11"  [label="(1, 14, 14, 384)", style=solid];
+"802 contiguous_11" -> "803 layer_norm_15"  [label="(1, 14, 14, 384)", style=solid];
+"803 layer_norm_15" -> "804 add_22"  [label="(1, 14, 14, 384)", style=solid];
+"804 add_22" -> "805 add_22_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"804 add_22" -> "829 add_23"  [label="(1, 14, 14, 384)", style=solid];
+"805 add_22_0_0_nncf_smooth_quant_0" -> "806 quantize_per_tensor_default_43"  [label="(1, 14, 14, 384)", style=solid];
+"806 quantize_per_tensor_default_43" -> "807 dequantize_per_tensor_default_43"  [label="(1, 14, 14, 384)", style=solid];
+"807 dequantize_per_tensor_default_43" -> "814 linear_42"  [label="(1, 14, 14, 384)", style=solid];
+"808 features_5_2_mlp_0_bias_0_0" -> "814 linear_42"  [label="(1536,)", style=solid];
+"809 scale_updated_constant29" -> "811 mul_tensor_29"  [label="(1536, 1)", style=solid];
+"810 compressed_weight_updated_constant29" -> "811 mul_tensor_29"  [label="(1536, 384)", style=solid];
+"811 mul_tensor_29" -> "813 sub_tensor_29"  [label="(1536, 384)", style=solid];
+"812 zero_point_updated_constant29" -> "813 sub_tensor_29"  [label="(1536, 1)", style=solid];
+"813 sub_tensor_29" -> "814 linear_42"  [label="(1536, 384)", style=solid];
+"814 linear_42" -> "815 gelu_6"  [label="(1, 14, 14, 1536)", style=solid];
+"815 gelu_6" -> "816 dropout_26"  [label="(1, 14, 14, 1536)", style=solid];
+"816 dropout_26" -> "817 dropout_26_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"817 dropout_26_0_0_nncf_smooth_quant_0" -> "818 quantize_per_tensor_default_44"  [label="(1, 14, 14, 1536)", style=solid];
+"818 quantize_per_tensor_default_44" -> "819 dequantize_per_tensor_default_44"  [label="(1, 14, 14, 1536)", style=solid];
+"819 dequantize_per_tensor_default_44" -> "826 linear_43"  [label="(1, 14, 14, 1536)", style=solid];
+"820 features_5_2_mlp_3_bias_0_0" -> "826 linear_43"  [label="(384,)", style=solid];
+"821 scale_updated_constant30" -> "823 mul_tensor_30"  [label="(384, 1)", style=solid];
+"822 compressed_weight_updated_constant30" -> "823 mul_tensor_30"  [label="(384, 1536)", style=solid];
+"823 mul_tensor_30" -> "825 sub_tensor_30"  [label="(384, 1536)", style=solid];
+"824 zero_point_updated_constant30" -> "825 sub_tensor_30"  [label="(384, 1)", style=solid];
+"825 sub_tensor_30" -> "826 linear_43"  [label="(384, 1536)", style=solid];
+"826 linear_43" -> "827 dropout_27"  [label="(1, 14, 14, 384)", style=solid];
+"827 dropout_27" -> "828 layer_norm_16"  [label="(1, 14, 14, 384)", style=solid];
+"828 layer_norm_16" -> "829 add_23"  [label="(1, 14, 14, 384)", style=solid];
+"829 add_23" -> "831 pad_9"  [label="(1, 14, 14, 384)", style=solid];
+"829 add_23" -> "910 add_26"  [label="(1, 14, 14, 384)", style=solid];
+"830 _frozen_param55" -> "867 add_24"  [label="(1, 12, 64, 64)", style=solid];
+"831 pad_9" -> "832 roll_6"  [label="(1, 16, 16, 384)", style=solid];
+"832 roll_6" -> "833 view_39"  [label="(1, 16, 16, 384)", style=solid];
+"833 view_39" -> "834 permute_33"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"834 permute_33" -> "835 reshape_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"835 reshape_31" -> "836 reshape_31_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"835 reshape_31" -> "868 new_zeros_3"  [label="(4, 64, 384)", style=solid];
+"836 reshape_31_0_0_nncf_smooth_quant_0" -> "837 quantize_per_tensor_default_45"  [label="(4, 64, 384)", style=solid];
+"837 quantize_per_tensor_default_45" -> "838 dequantize_per_tensor_default_45"  [label="(4, 64, 384)", style=solid];
+"838 dequantize_per_tensor_default_45" -> "845 linear_46"  [label="(4, 64, 384)", style=solid];
+"839 _frozen_param56_0_0" -> "845 linear_46"  [label="(1152,)", style=solid];
+"840 scale_updated_constant31" -> "842 mul_tensor_31"  [label="(1152, 1)", style=solid];
+"841 compressed_weight_updated_constant31" -> "842 mul_tensor_31"  [label="(1152, 384)", style=solid];
+"842 mul_tensor_31" -> "844 sub_tensor_31"  [label="(1152, 384)", style=solid];
+"843 zero_point_updated_constant31" -> "844 sub_tensor_31"  [label="(1152, 1)", style=solid];
+"844 sub_tensor_31" -> "845 linear_46"  [label="(1152, 384)", style=solid];
+"845 linear_46" -> "846 reshape_32"  [label="(4, 64, 1152)", style=solid];
+"846 reshape_32" -> "847 permute_34"  [label="(4, 64, 3, 12, 32)", style=solid];
+"847 permute_34" -> "848 select_21"  [label="(3, 4, 12, 64, 32)", style=solid];
+"847 permute_34" -> "849 select_22"  [label="(3, 4, 12, 64, 32)", style=solid];
+"847 permute_34" -> "850 select_23"  [label="(3, 4, 12, 64, 32)", style=solid];
+"848 select_21" -> "851 linalg_vector_norm_14"  [label="(4, 12, 64, 32)", style=solid];
+"848 select_21" -> "853 expand_as_14"  [label="(4, 12, 64, 32)", style=solid];
+"848 select_21" -> "854 div_14"  [label="(4, 12, 64, 32)", style=solid];
+"849 select_22" -> "857 linalg_vector_norm_15"  [label="(4, 12, 64, 32)", style=solid];
+"849 select_22" -> "859 expand_as_15"  [label="(4, 12, 64, 32)", style=solid];
+"849 select_22" -> "860 div_15"  [label="(4, 12, 64, 32)", style=solid];
+"850 select_23" -> "886 matmul_15"  [label="(4, 12, 64, 32)", style=solid];
+"851 linalg_vector_norm_14" -> "852 clamp_min_14"  [label="(4, 12, 64, 1)", style=solid];
+"852 clamp_min_14" -> "853 expand_as_14"  [label="(4, 12, 64, 1)", style=solid];
+"853 expand_as_14" -> "854 div_14"  [label="(4, 12, 64, 32)", style=solid];
+"854 div_14" -> "855 quantize_per_tensor_default_46"  [label="(4, 12, 64, 32)", style=solid];
+"855 quantize_per_tensor_default_46" -> "856 dequantize_per_tensor_default_46"  [label="(4, 12, 64, 32)", style=solid];
+"856 dequantize_per_tensor_default_46" -> "864 matmul_14"  [label="(4, 12, 64, 32)", style=solid];
+"857 linalg_vector_norm_15" -> "858 clamp_min_15"  [label="(4, 12, 64, 1)", style=solid];
+"858 clamp_min_15" -> "859 expand_as_15"  [label="(4, 12, 64, 1)", style=solid];
+"859 expand_as_15" -> "860 div_15"  [label="(4, 12, 64, 32)", style=solid];
+"860 div_15" -> "861 quantize_per_tensor_default_47"  [label="(4, 12, 64, 32)", style=solid];
+"861 quantize_per_tensor_default_47" -> "862 dequantize_per_tensor_default_47"  [label="(4, 12, 64, 32)", style=solid];
+"862 dequantize_per_tensor_default_47" -> "863 transpose_14"  [label="(4, 12, 64, 32)", style=solid];
+"863 transpose_14" -> "864 matmul_14"  [label="(4, 12, 32, 64)", style=solid];
+"864 matmul_14" -> "866 mul_15"  [label="(4, 12, 64, 64)", style=solid];
+"865 _frozen_param58" -> "866 mul_15"  [label="(12, 1, 1)", style=solid];
+"866 mul_15" -> "867 add_24"  [label="(4, 12, 64, 64)", style=solid];
+"867 add_24" -> "879 view_41"  [label="(4, 12, 64, 64)", style=solid];
+"868 new_zeros_3" -> "869 view_40"  [label="(16, 16)", style=solid];
+"869 view_40" -> "870 permute_35"  [label="(2, 8, 2, 8)", style=solid];
+"870 permute_35" -> "871 reshape_33"  [label="(2, 2, 8, 8)", style=solid];
+"871 reshape_33" -> "872 unsqueeze_20"  [label="(4, 64)", style=solid];
+"871 reshape_33" -> "873 unsqueeze_21"  [label="(4, 64)", style=solid];
+"872 unsqueeze_20" -> "874 sub_3"  [label="(4, 1, 64)", style=solid];
+"873 unsqueeze_21" -> "874 sub_3"  [label="(4, 64, 1)", style=solid];
+"874 sub_3" -> "875 ne_3"  [label="(4, 64, 64)", style=solid];
+"874 sub_3" -> "876 masked_fill_6"  [label="(4, 64, 64)", style=solid];
+"874 sub_3" -> "877 eq_3"  [label="(4, 64, 64)", style=solid];
+"875 ne_3" -> "876 masked_fill_6"  [label="(4, 64, 64)", style=solid];
+"876 masked_fill_6" -> "878 masked_fill_7"  [label="(4, 64, 64)", style=solid];
+"877 eq_3" -> "878 masked_fill_7"  [label="(4, 64, 64)", style=solid];
+"878 masked_fill_7" -> "880 unsqueeze_22"  [label="(4, 64, 64)", style=solid];
+"879 view_41" -> "882 add_25"  [label="(1, 4, 12, 64, 64)", style=solid];
+"880 unsqueeze_22" -> "881 unsqueeze_23"  [label="(4, 1, 64, 64)", style=solid];
+"881 unsqueeze_23" -> "882 add_25"  [label="(1, 4, 1, 64, 64)", style=solid];
+"882 add_25" -> "883 view_42"  [label="(1, 4, 12, 64, 64)", style=solid];
+"883 view_42" -> "884 softmax_7"  [label="(4, 12, 64, 64)", style=solid];
+"884 softmax_7" -> "885 dropout_28"  [label="(4, 12, 64, 64)", style=solid];
+"885 dropout_28" -> "886 matmul_15"  [label="(4, 12, 64, 64)", style=solid];
+"886 matmul_15" -> "887 transpose_15"  [label="(4, 12, 64, 32)", style=solid];
+"887 transpose_15" -> "888 reshape_34"  [label="(4, 64, 12, 32)", style=solid];
+"888 reshape_34" -> "889 reshape_34_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"889 reshape_34_0_0_nncf_smooth_quant_0" -> "890 quantize_per_tensor_default_48"  [label="(4, 64, 384)", style=solid];
+"890 quantize_per_tensor_default_48" -> "891 dequantize_per_tensor_default_48"  [label="(4, 64, 384)", style=solid];
+"891 dequantize_per_tensor_default_48" -> "898 linear_47"  [label="(4, 64, 384)", style=solid];
+"892 features_5_3_attn_proj_bias_0_0" -> "898 linear_47"  [label="(384,)", style=solid];
+"893 scale_updated_constant32" -> "895 mul_tensor_32"  [label="(384, 1)", style=solid];
+"894 compressed_weight_updated_constant32" -> "895 mul_tensor_32"  [label="(384, 384)", style=solid];
+"895 mul_tensor_32" -> "897 sub_tensor_32"  [label="(384, 384)", style=solid];
+"896 zero_point_updated_constant32" -> "897 sub_tensor_32"  [label="(384, 1)", style=solid];
+"897 sub_tensor_32" -> "898 linear_47"  [label="(384, 384)", style=solid];
+"898 linear_47" -> "899 dropout_29"  [label="(4, 64, 384)", style=solid];
+"899 dropout_29" -> "900 view_43"  [label="(4, 64, 384)", style=solid];
+"900 view_43" -> "901 permute_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"901 permute_36" -> "902 reshape_35"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"902 reshape_35" -> "903 roll_7"  [label="(1, 16, 16, 384)", style=solid];
+"903 roll_7" -> "904 slice_129"  [label="(1, 16, 16, 384)", style=solid];
+"904 slice_129" -> "905 slice_130"  [label="(1, 16, 16, 384)", style=solid];
+"905 slice_130" -> "906 slice_131"  [label="(1, 14, 16, 384)", style=solid];
+"906 slice_131" -> "907 slice_132"  [label="(1, 14, 14, 384)", style=solid];
+"907 slice_132" -> "908 contiguous_13"  [label="(1, 14, 14, 384)", style=solid];
+"908 contiguous_13" -> "909 layer_norm_17"  [label="(1, 14, 14, 384)", style=solid];
+"909 layer_norm_17" -> "910 add_26"  [label="(1, 14, 14, 384)", style=solid];
+"910 add_26" -> "911 add_26_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"910 add_26" -> "935 add_27"  [label="(1, 14, 14, 384)", style=solid];
+"911 add_26_0_0_nncf_smooth_quant_0" -> "912 quantize_per_tensor_default_49"  [label="(1, 14, 14, 384)", style=solid];
+"912 quantize_per_tensor_default_49" -> "913 dequantize_per_tensor_default_49"  [label="(1, 14, 14, 384)", style=solid];
+"913 dequantize_per_tensor_default_49" -> "920 linear_48"  [label="(1, 14, 14, 384)", style=solid];
+"914 features_5_3_mlp_0_bias_0_0" -> "920 linear_48"  [label="(1536,)", style=solid];
+"915 scale_updated_constant33" -> "917 mul_tensor_33"  [label="(1536, 1)", style=solid];
+"916 compressed_weight_updated_constant33" -> "917 mul_tensor_33"  [label="(1536, 384)", style=solid];
+"917 mul_tensor_33" -> "919 sub_tensor_33"  [label="(1536, 384)", style=solid];
+"918 zero_point_updated_constant33" -> "919 sub_tensor_33"  [label="(1536, 1)", style=solid];
+"919 sub_tensor_33" -> "920 linear_48"  [label="(1536, 384)", style=solid];
+"920 linear_48" -> "921 gelu_7"  [label="(1, 14, 14, 1536)", style=solid];
+"921 gelu_7" -> "922 dropout_30"  [label="(1, 14, 14, 1536)", style=solid];
+"922 dropout_30" -> "923 dropout_30_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"923 dropout_30_0_0_nncf_smooth_quant_0" -> "924 quantize_per_tensor_default_50"  [label="(1, 14, 14, 1536)", style=solid];
+"924 quantize_per_tensor_default_50" -> "925 dequantize_per_tensor_default_50"  [label="(1, 14, 14, 1536)", style=solid];
+"925 dequantize_per_tensor_default_50" -> "932 linear_49"  [label="(1, 14, 14, 1536)", style=solid];
+"926 features_5_3_mlp_3_bias_0_0" -> "932 linear_49"  [label="(384,)", style=solid];
+"927 scale_updated_constant34" -> "929 mul_tensor_34"  [label="(384, 1)", style=solid];
+"928 compressed_weight_updated_constant34" -> "929 mul_tensor_34"  [label="(384, 1536)", style=solid];
+"929 mul_tensor_34" -> "931 sub_tensor_34"  [label="(384, 1536)", style=solid];
+"930 zero_point_updated_constant34" -> "931 sub_tensor_34"  [label="(384, 1)", style=solid];
+"931 sub_tensor_34" -> "932 linear_49"  [label="(384, 1536)", style=solid];
+"932 linear_49" -> "933 dropout_31"  [label="(1, 14, 14, 384)", style=solid];
+"933 dropout_31" -> "934 layer_norm_18"  [label="(1, 14, 14, 384)", style=solid];
+"934 layer_norm_18" -> "935 add_27"  [label="(1, 14, 14, 384)", style=solid];
+"935 add_27" -> "937 pad_10"  [label="(1, 14, 14, 384)", style=solid];
+"935 add_27" -> "998 add_29"  [label="(1, 14, 14, 384)", style=solid];
+"936 _frozen_param68" -> "972 add_28"  [label="(1, 12, 64, 64)", style=solid];
+"937 pad_10" -> "938 view_46"  [label="(1, 16, 16, 384)", style=solid];
+"938 view_46" -> "939 permute_38"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"939 permute_38" -> "940 reshape_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"940 reshape_36" -> "941 reshape_36_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"941 reshape_36_0_0_nncf_smooth_quant_0" -> "942 quantize_per_tensor_default_51"  [label="(4, 64, 384)", style=solid];
+"942 quantize_per_tensor_default_51" -> "943 dequantize_per_tensor_default_51"  [label="(4, 64, 384)", style=solid];
+"943 dequantize_per_tensor_default_51" -> "950 linear_52"  [label="(4, 64, 384)", style=solid];
+"944 _frozen_param69_0_0" -> "950 linear_52"  [label="(1152,)", style=solid];
+"945 scale_updated_constant35" -> "947 mul_tensor_35"  [label="(1152, 1)", style=solid];
+"946 compressed_weight_updated_constant35" -> "947 mul_tensor_35"  [label="(1152, 384)", style=solid];
+"947 mul_tensor_35" -> "949 sub_tensor_35"  [label="(1152, 384)", style=solid];
+"948 zero_point_updated_constant35" -> "949 sub_tensor_35"  [label="(1152, 1)", style=solid];
+"949 sub_tensor_35" -> "950 linear_52"  [label="(1152, 384)", style=solid];
+"950 linear_52" -> "951 reshape_37"  [label="(4, 64, 1152)", style=solid];
+"951 reshape_37" -> "952 permute_39"  [label="(4, 64, 3, 12, 32)", style=solid];
+"952 permute_39" -> "953 select_24"  [label="(3, 4, 12, 64, 32)", style=solid];
+"952 permute_39" -> "954 select_25"  [label="(3, 4, 12, 64, 32)", style=solid];
+"952 permute_39" -> "955 select_26"  [label="(3, 4, 12, 64, 32)", style=solid];
+"953 select_24" -> "956 linalg_vector_norm_16"  [label="(4, 12, 64, 32)", style=solid];
+"953 select_24" -> "958 expand_as_16"  [label="(4, 12, 64, 32)", style=solid];
+"953 select_24" -> "959 div_16"  [label="(4, 12, 64, 32)", style=solid];
+"954 select_25" -> "962 linalg_vector_norm_17"  [label="(4, 12, 64, 32)", style=solid];
+"954 select_25" -> "964 expand_as_17"  [label="(4, 12, 64, 32)", style=solid];
+"954 select_25" -> "965 div_17"  [label="(4, 12, 64, 32)", style=solid];
+"955 select_26" -> "975 matmul_17"  [label="(4, 12, 64, 32)", style=solid];
+"956 linalg_vector_norm_16" -> "957 clamp_min_16"  [label="(4, 12, 64, 1)", style=solid];
+"957 clamp_min_16" -> "958 expand_as_16"  [label="(4, 12, 64, 1)", style=solid];
+"958 expand_as_16" -> "959 div_16"  [label="(4, 12, 64, 32)", style=solid];
+"959 div_16" -> "960 quantize_per_tensor_default_52"  [label="(4, 12, 64, 32)", style=solid];
+"960 quantize_per_tensor_default_52" -> "961 dequantize_per_tensor_default_52"  [label="(4, 12, 64, 32)", style=solid];
+"961 dequantize_per_tensor_default_52" -> "969 matmul_16"  [label="(4, 12, 64, 32)", style=solid];
+"962 linalg_vector_norm_17" -> "963 clamp_min_17"  [label="(4, 12, 64, 1)", style=solid];
+"963 clamp_min_17" -> "964 expand_as_17"  [label="(4, 12, 64, 1)", style=solid];
+"964 expand_as_17" -> "965 div_17"  [label="(4, 12, 64, 32)", style=solid];
+"965 div_17" -> "966 quantize_per_tensor_default_53"  [label="(4, 12, 64, 32)", style=solid];
+"966 quantize_per_tensor_default_53" -> "967 dequantize_per_tensor_default_53"  [label="(4, 12, 64, 32)", style=solid];
+"967 dequantize_per_tensor_default_53" -> "968 transpose_16"  [label="(4, 12, 64, 32)", style=solid];
+"968 transpose_16" -> "969 matmul_16"  [label="(4, 12, 32, 64)", style=solid];
+"969 matmul_16" -> "971 mul_17"  [label="(4, 12, 64, 64)", style=solid];
+"970 _frozen_param71" -> "971 mul_17"  [label="(12, 1, 1)", style=solid];
+"971 mul_17" -> "972 add_28"  [label="(4, 12, 64, 64)", style=solid];
+"972 add_28" -> "973 softmax_8"  [label="(4, 12, 64, 64)", style=solid];
+"973 softmax_8" -> "974 dropout_32"  [label="(4, 12, 64, 64)", style=solid];
+"974 dropout_32" -> "975 matmul_17"  [label="(4, 12, 64, 64)", style=solid];
+"975 matmul_17" -> "976 transpose_17"  [label="(4, 12, 64, 32)", style=solid];
+"976 transpose_17" -> "977 reshape_38"  [label="(4, 64, 12, 32)", style=solid];
+"977 reshape_38" -> "978 reshape_38_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"978 reshape_38_0_0_nncf_smooth_quant_0" -> "979 quantize_per_tensor_default_54"  [label="(4, 64, 384)", style=solid];
+"979 quantize_per_tensor_default_54" -> "980 dequantize_per_tensor_default_54"  [label="(4, 64, 384)", style=solid];
+"980 dequantize_per_tensor_default_54" -> "987 linear_53"  [label="(4, 64, 384)", style=solid];
+"981 features_5_4_attn_proj_bias_0_0" -> "987 linear_53"  [label="(384,)", style=solid];
+"982 scale_updated_constant36" -> "984 mul_tensor_36"  [label="(384, 1)", style=solid];
+"983 compressed_weight_updated_constant36" -> "984 mul_tensor_36"  [label="(384, 384)", style=solid];
+"984 mul_tensor_36" -> "986 sub_tensor_36"  [label="(384, 384)", style=solid];
+"985 zero_point_updated_constant36" -> "986 sub_tensor_36"  [label="(384, 1)", style=solid];
+"986 sub_tensor_36" -> "987 linear_53"  [label="(384, 384)", style=solid];
+"987 linear_53" -> "988 dropout_33"  [label="(4, 64, 384)", style=solid];
+"988 dropout_33" -> "989 view_47"  [label="(4, 64, 384)", style=solid];
+"989 view_47" -> "990 permute_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"990 permute_40" -> "991 reshape_39"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"991 reshape_39" -> "992 slice_134"  [label="(1, 16, 16, 384)", style=solid];
+"992 slice_134" -> "993 slice_135"  [label="(1, 16, 16, 384)", style=solid];
+"993 slice_135" -> "994 slice_136"  [label="(1, 14, 16, 384)", style=solid];
+"994 slice_136" -> "995 slice_137"  [label="(1, 14, 14, 384)", style=solid];
+"995 slice_137" -> "996 contiguous_15"  [label="(1, 14, 14, 384)", style=solid];
+"996 contiguous_15" -> "997 layer_norm_19"  [label="(1, 14, 14, 384)", style=solid];
+"997 layer_norm_19" -> "998 add_29"  [label="(1, 14, 14, 384)", style=solid];
+"998 add_29" -> "999 add_29_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"998 add_29" -> "1023 add_30"  [label="(1, 14, 14, 384)", style=solid];
+"999 add_29_0_0_nncf_smooth_quant_0" -> "1000 quantize_per_tensor_default_55"  [label="(1, 14, 14, 384)", style=solid];
+"1000 quantize_per_tensor_default_55" -> "1001 dequantize_per_tensor_default_55"  [label="(1, 14, 14, 384)", style=solid];
+"1001 dequantize_per_tensor_default_55" -> "1008 linear_54"  [label="(1, 14, 14, 384)", style=solid];
+"1002 features_5_4_mlp_0_bias_0_0" -> "1008 linear_54"  [label="(1536,)", style=solid];
+"1003 scale_updated_constant37" -> "1005 mul_tensor_37"  [label="(1536, 1)", style=solid];
+"1004 compressed_weight_updated_constant37" -> "1005 mul_tensor_37"  [label="(1536, 384)", style=solid];
+"1005 mul_tensor_37" -> "1007 sub_tensor_37"  [label="(1536, 384)", style=solid];
+"1006 zero_point_updated_constant37" -> "1007 sub_tensor_37"  [label="(1536, 1)", style=solid];
+"1007 sub_tensor_37" -> "1008 linear_54"  [label="(1536, 384)", style=solid];
+"1008 linear_54" -> "1009 gelu_8"  [label="(1, 14, 14, 1536)", style=solid];
+"1009 gelu_8" -> "1010 dropout_34"  [label="(1, 14, 14, 1536)", style=solid];
+"1010 dropout_34" -> "1011 dropout_34_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1011 dropout_34_0_0_nncf_smooth_quant_0" -> "1012 quantize_per_tensor_default_56"  [label="(1, 14, 14, 1536)", style=solid];
+"1012 quantize_per_tensor_default_56" -> "1013 dequantize_per_tensor_default_56"  [label="(1, 14, 14, 1536)", style=solid];
+"1013 dequantize_per_tensor_default_56" -> "1020 linear_55"  [label="(1, 14, 14, 1536)", style=solid];
+"1014 features_5_4_mlp_3_bias_0_0" -> "1020 linear_55"  [label="(384,)", style=solid];
+"1015 scale_updated_constant38" -> "1017 mul_tensor_38"  [label="(384, 1)", style=solid];
+"1016 compressed_weight_updated_constant38" -> "1017 mul_tensor_38"  [label="(384, 1536)", style=solid];
+"1017 mul_tensor_38" -> "1019 sub_tensor_38"  [label="(384, 1536)", style=solid];
+"1018 zero_point_updated_constant38" -> "1019 sub_tensor_38"  [label="(384, 1)", style=solid];
+"1019 sub_tensor_38" -> "1020 linear_55"  [label="(384, 1536)", style=solid];
+"1020 linear_55" -> "1021 dropout_35"  [label="(1, 14, 14, 384)", style=solid];
+"1021 dropout_35" -> "1022 layer_norm_20"  [label="(1, 14, 14, 384)", style=solid];
+"1022 layer_norm_20" -> "1023 add_30"  [label="(1, 14, 14, 384)", style=solid];
+"1023 add_30" -> "1025 pad_11"  [label="(1, 14, 14, 384)", style=solid];
+"1023 add_30" -> "1104 add_33"  [label="(1, 14, 14, 384)", style=solid];
+"1024 _frozen_param72" -> "1061 add_31"  [label="(1, 12, 64, 64)", style=solid];
+"1025 pad_11" -> "1026 roll_8"  [label="(1, 16, 16, 384)", style=solid];
+"1026 roll_8" -> "1027 view_50"  [label="(1, 16, 16, 384)", style=solid];
+"1027 view_50" -> "1028 permute_42"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1028 permute_42" -> "1029 reshape_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1029 reshape_40" -> "1030 reshape_40_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1029 reshape_40" -> "1062 new_zeros_4"  [label="(4, 64, 384)", style=solid];
+"1030 reshape_40_0_0_nncf_smooth_quant_0" -> "1031 quantize_per_tensor_default_57"  [label="(4, 64, 384)", style=solid];
+"1031 quantize_per_tensor_default_57" -> "1032 dequantize_per_tensor_default_57"  [label="(4, 64, 384)", style=solid];
+"1032 dequantize_per_tensor_default_57" -> "1039 linear_58"  [label="(4, 64, 384)", style=solid];
+"1033 _frozen_param73_0_0" -> "1039 linear_58"  [label="(1152,)", style=solid];
+"1034 scale_updated_constant39" -> "1036 mul_tensor_39"  [label="(1152, 1)", style=solid];
+"1035 compressed_weight_updated_constant39" -> "1036 mul_tensor_39"  [label="(1152, 384)", style=solid];
+"1036 mul_tensor_39" -> "1038 sub_tensor_39"  [label="(1152, 384)", style=solid];
+"1037 zero_point_updated_constant39" -> "1038 sub_tensor_39"  [label="(1152, 1)", style=solid];
+"1038 sub_tensor_39" -> "1039 linear_58"  [label="(1152, 384)", style=solid];
+"1039 linear_58" -> "1040 reshape_41"  [label="(4, 64, 1152)", style=solid];
+"1040 reshape_41" -> "1041 permute_43"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1041 permute_43" -> "1042 select_27"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1041 permute_43" -> "1043 select_28"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1041 permute_43" -> "1044 select_29"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1042 select_27" -> "1045 linalg_vector_norm_18"  [label="(4, 12, 64, 32)", style=solid];
+"1042 select_27" -> "1047 expand_as_18"  [label="(4, 12, 64, 32)", style=solid];
+"1042 select_27" -> "1048 div_18"  [label="(4, 12, 64, 32)", style=solid];
+"1043 select_28" -> "1051 linalg_vector_norm_19"  [label="(4, 12, 64, 32)", style=solid];
+"1043 select_28" -> "1053 expand_as_19"  [label="(4, 12, 64, 32)", style=solid];
+"1043 select_28" -> "1054 div_19"  [label="(4, 12, 64, 32)", style=solid];
+"1044 select_29" -> "1080 matmul_19"  [label="(4, 12, 64, 32)", style=solid];
+"1045 linalg_vector_norm_18" -> "1046 clamp_min_18"  [label="(4, 12, 64, 1)", style=solid];
+"1046 clamp_min_18" -> "1047 expand_as_18"  [label="(4, 12, 64, 1)", style=solid];
+"1047 expand_as_18" -> "1048 div_18"  [label="(4, 12, 64, 32)", style=solid];
+"1048 div_18" -> "1049 quantize_per_tensor_default_58"  [label="(4, 12, 64, 32)", style=solid];
+"1049 quantize_per_tensor_default_58" -> "1050 dequantize_per_tensor_default_58"  [label="(4, 12, 64, 32)", style=solid];
+"1050 dequantize_per_tensor_default_58" -> "1058 matmul_18"  [label="(4, 12, 64, 32)", style=solid];
+"1051 linalg_vector_norm_19" -> "1052 clamp_min_19"  [label="(4, 12, 64, 1)", style=solid];
+"1052 clamp_min_19" -> "1053 expand_as_19"  [label="(4, 12, 64, 1)", style=solid];
+"1053 expand_as_19" -> "1054 div_19"  [label="(4, 12, 64, 32)", style=solid];
+"1054 div_19" -> "1055 quantize_per_tensor_default_59"  [label="(4, 12, 64, 32)", style=solid];
+"1055 quantize_per_tensor_default_59" -> "1056 dequantize_per_tensor_default_59"  [label="(4, 12, 64, 32)", style=solid];
+"1056 dequantize_per_tensor_default_59" -> "1057 transpose_18"  [label="(4, 12, 64, 32)", style=solid];
+"1057 transpose_18" -> "1058 matmul_18"  [label="(4, 12, 32, 64)", style=solid];
+"1058 matmul_18" -> "1060 mul_19"  [label="(4, 12, 64, 64)", style=solid];
+"1059 _frozen_param75" -> "1060 mul_19"  [label="(12, 1, 1)", style=solid];
+"1060 mul_19" -> "1061 add_31"  [label="(4, 12, 64, 64)", style=solid];
+"1061 add_31" -> "1073 view_52"  [label="(4, 12, 64, 64)", style=solid];
+"1062 new_zeros_4" -> "1063 view_51"  [label="(16, 16)", style=solid];
+"1063 view_51" -> "1064 permute_44"  [label="(2, 8, 2, 8)", style=solid];
+"1064 permute_44" -> "1065 reshape_42"  [label="(2, 2, 8, 8)", style=solid];
+"1065 reshape_42" -> "1066 unsqueeze_26"  [label="(4, 64)", style=solid];
+"1065 reshape_42" -> "1067 unsqueeze_27"  [label="(4, 64)", style=solid];
+"1066 unsqueeze_26" -> "1068 sub_4"  [label="(4, 1, 64)", style=solid];
+"1067 unsqueeze_27" -> "1068 sub_4"  [label="(4, 64, 1)", style=solid];
+"1068 sub_4" -> "1069 ne_4"  [label="(4, 64, 64)", style=solid];
+"1068 sub_4" -> "1070 masked_fill_8"  [label="(4, 64, 64)", style=solid];
+"1068 sub_4" -> "1071 eq_4"  [label="(4, 64, 64)", style=solid];
+"1069 ne_4" -> "1070 masked_fill_8"  [label="(4, 64, 64)", style=solid];
+"1070 masked_fill_8" -> "1072 masked_fill_9"  [label="(4, 64, 64)", style=solid];
+"1071 eq_4" -> "1072 masked_fill_9"  [label="(4, 64, 64)", style=solid];
+"1072 masked_fill_9" -> "1074 unsqueeze_28"  [label="(4, 64, 64)", style=solid];
+"1073 view_52" -> "1076 add_32"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1074 unsqueeze_28" -> "1075 unsqueeze_29"  [label="(4, 1, 64, 64)", style=solid];
+"1075 unsqueeze_29" -> "1076 add_32"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1076 add_32" -> "1077 view_53"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1077 view_53" -> "1078 softmax_9"  [label="(4, 12, 64, 64)", style=solid];
+"1078 softmax_9" -> "1079 dropout_36"  [label="(4, 12, 64, 64)", style=solid];
+"1079 dropout_36" -> "1080 matmul_19"  [label="(4, 12, 64, 64)", style=solid];
+"1080 matmul_19" -> "1081 transpose_19"  [label="(4, 12, 64, 32)", style=solid];
+"1081 transpose_19" -> "1082 reshape_43"  [label="(4, 64, 12, 32)", style=solid];
+"1082 reshape_43" -> "1083 reshape_43_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1083 reshape_43_0_0_nncf_smooth_quant_0" -> "1084 quantize_per_tensor_default_60"  [label="(4, 64, 384)", style=solid];
+"1084 quantize_per_tensor_default_60" -> "1085 dequantize_per_tensor_default_60"  [label="(4, 64, 384)", style=solid];
+"1085 dequantize_per_tensor_default_60" -> "1092 linear_59"  [label="(4, 64, 384)", style=solid];
+"1086 features_5_5_attn_proj_bias_0_0" -> "1092 linear_59"  [label="(384,)", style=solid];
+"1087 scale_updated_constant40" -> "1089 mul_tensor_40"  [label="(384, 1)", style=solid];
+"1088 compressed_weight_updated_constant40" -> "1089 mul_tensor_40"  [label="(384, 384)", style=solid];
+"1089 mul_tensor_40" -> "1091 sub_tensor_40"  [label="(384, 384)", style=solid];
+"1090 zero_point_updated_constant40" -> "1091 sub_tensor_40"  [label="(384, 1)", style=solid];
+"1091 sub_tensor_40" -> "1092 linear_59"  [label="(384, 384)", style=solid];
+"1092 linear_59" -> "1093 dropout_37"  [label="(4, 64, 384)", style=solid];
+"1093 dropout_37" -> "1094 view_54"  [label="(4, 64, 384)", style=solid];
+"1094 view_54" -> "1095 permute_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1095 permute_45" -> "1096 reshape_44"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1096 reshape_44" -> "1097 roll_9"  [label="(1, 16, 16, 384)", style=solid];
+"1097 roll_9" -> "1098 slice_157"  [label="(1, 16, 16, 384)", style=solid];
+"1098 slice_157" -> "1099 slice_158"  [label="(1, 16, 16, 384)", style=solid];
+"1099 slice_158" -> "1100 slice_159"  [label="(1, 14, 16, 384)", style=solid];
+"1100 slice_159" -> "1101 slice_160"  [label="(1, 14, 14, 384)", style=solid];
+"1101 slice_160" -> "1102 contiguous_17"  [label="(1, 14, 14, 384)", style=solid];
+"1102 contiguous_17" -> "1103 layer_norm_21"  [label="(1, 14, 14, 384)", style=solid];
+"1103 layer_norm_21" -> "1104 add_33"  [label="(1, 14, 14, 384)", style=solid];
+"1104 add_33" -> "1105 add_33_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1104 add_33" -> "1129 add_34"  [label="(1, 14, 14, 384)", style=solid];
+"1105 add_33_0_0_nncf_smooth_quant_0" -> "1106 quantize_per_tensor_default_61"  [label="(1, 14, 14, 384)", style=solid];
+"1106 quantize_per_tensor_default_61" -> "1107 dequantize_per_tensor_default_61"  [label="(1, 14, 14, 384)", style=solid];
+"1107 dequantize_per_tensor_default_61" -> "1114 linear_60"  [label="(1, 14, 14, 384)", style=solid];
+"1108 features_5_5_mlp_0_bias_0_0" -> "1114 linear_60"  [label="(1536,)", style=solid];
+"1109 scale_updated_constant41" -> "1111 mul_tensor_41"  [label="(1536, 1)", style=solid];
+"1110 compressed_weight_updated_constant41" -> "1111 mul_tensor_41"  [label="(1536, 384)", style=solid];
+"1111 mul_tensor_41" -> "1113 sub_tensor_41"  [label="(1536, 384)", style=solid];
+"1112 zero_point_updated_constant41" -> "1113 sub_tensor_41"  [label="(1536, 1)", style=solid];
+"1113 sub_tensor_41" -> "1114 linear_60"  [label="(1536, 384)", style=solid];
+"1114 linear_60" -> "1115 gelu_9"  [label="(1, 14, 14, 1536)", style=solid];
+"1115 gelu_9" -> "1116 dropout_38"  [label="(1, 14, 14, 1536)", style=solid];
+"1116 dropout_38" -> "1117 dropout_38_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1117 dropout_38_0_0_nncf_smooth_quant_0" -> "1118 quantize_per_tensor_default_62"  [label="(1, 14, 14, 1536)", style=solid];
+"1118 quantize_per_tensor_default_62" -> "1119 dequantize_per_tensor_default_62"  [label="(1, 14, 14, 1536)", style=solid];
+"1119 dequantize_per_tensor_default_62" -> "1126 linear_61"  [label="(1, 14, 14, 1536)", style=solid];
+"1120 features_5_5_mlp_3_bias_0_0" -> "1126 linear_61"  [label="(384,)", style=solid];
+"1121 scale_updated_constant42" -> "1123 mul_tensor_42"  [label="(384, 1)", style=solid];
+"1122 compressed_weight_updated_constant42" -> "1123 mul_tensor_42"  [label="(384, 1536)", style=solid];
+"1123 mul_tensor_42" -> "1125 sub_tensor_42"  [label="(384, 1536)", style=solid];
+"1124 zero_point_updated_constant42" -> "1125 sub_tensor_42"  [label="(384, 1)", style=solid];
+"1125 sub_tensor_42" -> "1126 linear_61"  [label="(384, 1536)", style=solid];
+"1126 linear_61" -> "1127 dropout_39"  [label="(1, 14, 14, 384)", style=solid];
+"1127 dropout_39" -> "1128 layer_norm_22"  [label="(1, 14, 14, 384)", style=solid];
+"1128 layer_norm_22" -> "1129 add_34"  [label="(1, 14, 14, 384)", style=solid];
+"1129 add_34" -> "1131 pad_12"  [label="(1, 14, 14, 384)", style=solid];
+"1129 add_34" -> "1192 add_36"  [label="(1, 14, 14, 384)", style=solid];
+"1130 _frozen_param85" -> "1166 add_35"  [label="(1, 12, 64, 64)", style=solid];
+"1131 pad_12" -> "1132 view_57"  [label="(1, 16, 16, 384)", style=solid];
+"1132 view_57" -> "1133 permute_47"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1133 permute_47" -> "1134 reshape_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1134 reshape_45" -> "1135 reshape_45_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1135 reshape_45_0_0_nncf_smooth_quant_0" -> "1136 quantize_per_tensor_default_63"  [label="(4, 64, 384)", style=solid];
+"1136 quantize_per_tensor_default_63" -> "1137 dequantize_per_tensor_default_63"  [label="(4, 64, 384)", style=solid];
+"1137 dequantize_per_tensor_default_63" -> "1144 linear_64"  [label="(4, 64, 384)", style=solid];
+"1138 _frozen_param86_0_0" -> "1144 linear_64"  [label="(1152,)", style=solid];
+"1139 scale_updated_constant43" -> "1141 mul_tensor_43"  [label="(1152, 1)", style=solid];
+"1140 compressed_weight_updated_constant43" -> "1141 mul_tensor_43"  [label="(1152, 384)", style=solid];
+"1141 mul_tensor_43" -> "1143 sub_tensor_43"  [label="(1152, 384)", style=solid];
+"1142 zero_point_updated_constant43" -> "1143 sub_tensor_43"  [label="(1152, 1)", style=solid];
+"1143 sub_tensor_43" -> "1144 linear_64"  [label="(1152, 384)", style=solid];
+"1144 linear_64" -> "1145 reshape_46"  [label="(4, 64, 1152)", style=solid];
+"1145 reshape_46" -> "1146 permute_48"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1146 permute_48" -> "1147 select_30"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1146 permute_48" -> "1148 select_31"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1146 permute_48" -> "1149 select_32"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1147 select_30" -> "1150 linalg_vector_norm_20"  [label="(4, 12, 64, 32)", style=solid];
+"1147 select_30" -> "1152 expand_as_20"  [label="(4, 12, 64, 32)", style=solid];
+"1147 select_30" -> "1153 div_20"  [label="(4, 12, 64, 32)", style=solid];
+"1148 select_31" -> "1156 linalg_vector_norm_21"  [label="(4, 12, 64, 32)", style=solid];
+"1148 select_31" -> "1158 expand_as_21"  [label="(4, 12, 64, 32)", style=solid];
+"1148 select_31" -> "1159 div_21"  [label="(4, 12, 64, 32)", style=solid];
+"1149 select_32" -> "1169 matmul_21"  [label="(4, 12, 64, 32)", style=solid];
+"1150 linalg_vector_norm_20" -> "1151 clamp_min_20"  [label="(4, 12, 64, 1)", style=solid];
+"1151 clamp_min_20" -> "1152 expand_as_20"  [label="(4, 12, 64, 1)", style=solid];
+"1152 expand_as_20" -> "1153 div_20"  [label="(4, 12, 64, 32)", style=solid];
+"1153 div_20" -> "1154 quantize_per_tensor_default_64"  [label="(4, 12, 64, 32)", style=solid];
+"1154 quantize_per_tensor_default_64" -> "1155 dequantize_per_tensor_default_64"  [label="(4, 12, 64, 32)", style=solid];
+"1155 dequantize_per_tensor_default_64" -> "1163 matmul_20"  [label="(4, 12, 64, 32)", style=solid];
+"1156 linalg_vector_norm_21" -> "1157 clamp_min_21"  [label="(4, 12, 64, 1)", style=solid];
+"1157 clamp_min_21" -> "1158 expand_as_21"  [label="(4, 12, 64, 1)", style=solid];
+"1158 expand_as_21" -> "1159 div_21"  [label="(4, 12, 64, 32)", style=solid];
+"1159 div_21" -> "1160 quantize_per_tensor_default_65"  [label="(4, 12, 64, 32)", style=solid];
+"1160 quantize_per_tensor_default_65" -> "1161 dequantize_per_tensor_default_65"  [label="(4, 12, 64, 32)", style=solid];
+"1161 dequantize_per_tensor_default_65" -> "1162 transpose_20"  [label="(4, 12, 64, 32)", style=solid];
+"1162 transpose_20" -> "1163 matmul_20"  [label="(4, 12, 32, 64)", style=solid];
+"1163 matmul_20" -> "1165 mul_21"  [label="(4, 12, 64, 64)", style=solid];
+"1164 _frozen_param88" -> "1165 mul_21"  [label="(12, 1, 1)", style=solid];
+"1165 mul_21" -> "1166 add_35"  [label="(4, 12, 64, 64)", style=solid];
+"1166 add_35" -> "1167 softmax_10"  [label="(4, 12, 64, 64)", style=solid];
+"1167 softmax_10" -> "1168 dropout_40"  [label="(4, 12, 64, 64)", style=solid];
+"1168 dropout_40" -> "1169 matmul_21"  [label="(4, 12, 64, 64)", style=solid];
+"1169 matmul_21" -> "1170 transpose_21"  [label="(4, 12, 64, 32)", style=solid];
+"1170 transpose_21" -> "1171 reshape_47"  [label="(4, 64, 12, 32)", style=solid];
+"1171 reshape_47" -> "1172 reshape_47_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1172 reshape_47_0_0_nncf_smooth_quant_0" -> "1173 quantize_per_tensor_default_66"  [label="(4, 64, 384)", style=solid];
+"1173 quantize_per_tensor_default_66" -> "1174 dequantize_per_tensor_default_66"  [label="(4, 64, 384)", style=solid];
+"1174 dequantize_per_tensor_default_66" -> "1181 linear_65"  [label="(4, 64, 384)", style=solid];
+"1175 features_5_6_attn_proj_bias_0_0" -> "1181 linear_65"  [label="(384,)", style=solid];
+"1176 scale_updated_constant44" -> "1178 mul_tensor_44"  [label="(384, 1)", style=solid];
+"1177 compressed_weight_updated_constant44" -> "1178 mul_tensor_44"  [label="(384, 384)", style=solid];
+"1178 mul_tensor_44" -> "1180 sub_tensor_44"  [label="(384, 384)", style=solid];
+"1179 zero_point_updated_constant44" -> "1180 sub_tensor_44"  [label="(384, 1)", style=solid];
+"1180 sub_tensor_44" -> "1181 linear_65"  [label="(384, 384)", style=solid];
+"1181 linear_65" -> "1182 dropout_41"  [label="(4, 64, 384)", style=solid];
+"1182 dropout_41" -> "1183 view_58"  [label="(4, 64, 384)", style=solid];
+"1183 view_58" -> "1184 permute_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1184 permute_49" -> "1185 reshape_48"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1185 reshape_48" -> "1186 slice_162"  [label="(1, 16, 16, 384)", style=solid];
+"1186 slice_162" -> "1187 slice_163"  [label="(1, 16, 16, 384)", style=solid];
+"1187 slice_163" -> "1188 slice_164"  [label="(1, 14, 16, 384)", style=solid];
+"1188 slice_164" -> "1189 slice_165"  [label="(1, 14, 14, 384)", style=solid];
+"1189 slice_165" -> "1190 contiguous_19"  [label="(1, 14, 14, 384)", style=solid];
+"1190 contiguous_19" -> "1191 layer_norm_23"  [label="(1, 14, 14, 384)", style=solid];
+"1191 layer_norm_23" -> "1192 add_36"  [label="(1, 14, 14, 384)", style=solid];
+"1192 add_36" -> "1193 add_36_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1192 add_36" -> "1217 add_37"  [label="(1, 14, 14, 384)", style=solid];
+"1193 add_36_0_0_nncf_smooth_quant_0" -> "1194 quantize_per_tensor_default_67"  [label="(1, 14, 14, 384)", style=solid];
+"1194 quantize_per_tensor_default_67" -> "1195 dequantize_per_tensor_default_67"  [label="(1, 14, 14, 384)", style=solid];
+"1195 dequantize_per_tensor_default_67" -> "1202 linear_66"  [label="(1, 14, 14, 384)", style=solid];
+"1196 features_5_6_mlp_0_bias_0_0" -> "1202 linear_66"  [label="(1536,)", style=solid];
+"1197 scale_updated_constant45" -> "1199 mul_tensor_45"  [label="(1536, 1)", style=solid];
+"1198 compressed_weight_updated_constant45" -> "1199 mul_tensor_45"  [label="(1536, 384)", style=solid];
+"1199 mul_tensor_45" -> "1201 sub_tensor_45"  [label="(1536, 384)", style=solid];
+"1200 zero_point_updated_constant45" -> "1201 sub_tensor_45"  [label="(1536, 1)", style=solid];
+"1201 sub_tensor_45" -> "1202 linear_66"  [label="(1536, 384)", style=solid];
+"1202 linear_66" -> "1203 gelu_10"  [label="(1, 14, 14, 1536)", style=solid];
+"1203 gelu_10" -> "1204 dropout_42"  [label="(1, 14, 14, 1536)", style=solid];
+"1204 dropout_42" -> "1205 dropout_42_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1205 dropout_42_0_0_nncf_smooth_quant_0" -> "1206 quantize_per_tensor_default_68"  [label="(1, 14, 14, 1536)", style=solid];
+"1206 quantize_per_tensor_default_68" -> "1207 dequantize_per_tensor_default_68"  [label="(1, 14, 14, 1536)", style=solid];
+"1207 dequantize_per_tensor_default_68" -> "1214 linear_67"  [label="(1, 14, 14, 1536)", style=solid];
+"1208 features_5_6_mlp_3_bias_0_0" -> "1214 linear_67"  [label="(384,)", style=solid];
+"1209 scale_updated_constant46" -> "1211 mul_tensor_46"  [label="(384, 1)", style=solid];
+"1210 compressed_weight_updated_constant46" -> "1211 mul_tensor_46"  [label="(384, 1536)", style=solid];
+"1211 mul_tensor_46" -> "1213 sub_tensor_46"  [label="(384, 1536)", style=solid];
+"1212 zero_point_updated_constant46" -> "1213 sub_tensor_46"  [label="(384, 1)", style=solid];
+"1213 sub_tensor_46" -> "1214 linear_67"  [label="(384, 1536)", style=solid];
+"1214 linear_67" -> "1215 dropout_43"  [label="(1, 14, 14, 384)", style=solid];
+"1215 dropout_43" -> "1216 layer_norm_24"  [label="(1, 14, 14, 384)", style=solid];
+"1216 layer_norm_24" -> "1217 add_37"  [label="(1, 14, 14, 384)", style=solid];
+"1217 add_37" -> "1219 pad_13"  [label="(1, 14, 14, 384)", style=solid];
+"1217 add_37" -> "1298 add_40"  [label="(1, 14, 14, 384)", style=solid];
+"1218 _frozen_param89" -> "1255 add_38"  [label="(1, 12, 64, 64)", style=solid];
+"1219 pad_13" -> "1220 roll_10"  [label="(1, 16, 16, 384)", style=solid];
+"1220 roll_10" -> "1221 view_61"  [label="(1, 16, 16, 384)", style=solid];
+"1221 view_61" -> "1222 permute_51"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1222 permute_51" -> "1223 reshape_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1223 reshape_49" -> "1224 reshape_49_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1223 reshape_49" -> "1256 new_zeros_5"  [label="(4, 64, 384)", style=solid];
+"1224 reshape_49_0_0_nncf_smooth_quant_0" -> "1225 quantize_per_tensor_default_69"  [label="(4, 64, 384)", style=solid];
+"1225 quantize_per_tensor_default_69" -> "1226 dequantize_per_tensor_default_69"  [label="(4, 64, 384)", style=solid];
+"1226 dequantize_per_tensor_default_69" -> "1233 linear_70"  [label="(4, 64, 384)", style=solid];
+"1227 _frozen_param90_0_0" -> "1233 linear_70"  [label="(1152,)", style=solid];
+"1228 scale_updated_constant47" -> "1230 mul_tensor_47"  [label="(1152, 1)", style=solid];
+"1229 compressed_weight_updated_constant47" -> "1230 mul_tensor_47"  [label="(1152, 384)", style=solid];
+"1230 mul_tensor_47" -> "1232 sub_tensor_47"  [label="(1152, 384)", style=solid];
+"1231 zero_point_updated_constant47" -> "1232 sub_tensor_47"  [label="(1152, 1)", style=solid];
+"1232 sub_tensor_47" -> "1233 linear_70"  [label="(1152, 384)", style=solid];
+"1233 linear_70" -> "1234 reshape_50"  [label="(4, 64, 1152)", style=solid];
+"1234 reshape_50" -> "1235 permute_52"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1235 permute_52" -> "1236 select_33"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1235 permute_52" -> "1237 select_34"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1235 permute_52" -> "1238 select_35"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1236 select_33" -> "1239 linalg_vector_norm_22"  [label="(4, 12, 64, 32)", style=solid];
+"1236 select_33" -> "1241 expand_as_22"  [label="(4, 12, 64, 32)", style=solid];
+"1236 select_33" -> "1242 div_22"  [label="(4, 12, 64, 32)", style=solid];
+"1237 select_34" -> "1245 linalg_vector_norm_23"  [label="(4, 12, 64, 32)", style=solid];
+"1237 select_34" -> "1247 expand_as_23"  [label="(4, 12, 64, 32)", style=solid];
+"1237 select_34" -> "1248 div_23"  [label="(4, 12, 64, 32)", style=solid];
+"1238 select_35" -> "1274 matmul_23"  [label="(4, 12, 64, 32)", style=solid];
+"1239 linalg_vector_norm_22" -> "1240 clamp_min_22"  [label="(4, 12, 64, 1)", style=solid];
+"1240 clamp_min_22" -> "1241 expand_as_22"  [label="(4, 12, 64, 1)", style=solid];
+"1241 expand_as_22" -> "1242 div_22"  [label="(4, 12, 64, 32)", style=solid];
+"1242 div_22" -> "1243 quantize_per_tensor_default_70"  [label="(4, 12, 64, 32)", style=solid];
+"1243 quantize_per_tensor_default_70" -> "1244 dequantize_per_tensor_default_70"  [label="(4, 12, 64, 32)", style=solid];
+"1244 dequantize_per_tensor_default_70" -> "1252 matmul_22"  [label="(4, 12, 64, 32)", style=solid];
+"1245 linalg_vector_norm_23" -> "1246 clamp_min_23"  [label="(4, 12, 64, 1)", style=solid];
+"1246 clamp_min_23" -> "1247 expand_as_23"  [label="(4, 12, 64, 1)", style=solid];
+"1247 expand_as_23" -> "1248 div_23"  [label="(4, 12, 64, 32)", style=solid];
+"1248 div_23" -> "1249 quantize_per_tensor_default_71"  [label="(4, 12, 64, 32)", style=solid];
+"1249 quantize_per_tensor_default_71" -> "1250 dequantize_per_tensor_default_71"  [label="(4, 12, 64, 32)", style=solid];
+"1250 dequantize_per_tensor_default_71" -> "1251 transpose_22"  [label="(4, 12, 64, 32)", style=solid];
+"1251 transpose_22" -> "1252 matmul_22"  [label="(4, 12, 32, 64)", style=solid];
+"1252 matmul_22" -> "1254 mul_23"  [label="(4, 12, 64, 64)", style=solid];
+"1253 _frozen_param92" -> "1254 mul_23"  [label="(12, 1, 1)", style=solid];
+"1254 mul_23" -> "1255 add_38"  [label="(4, 12, 64, 64)", style=solid];
+"1255 add_38" -> "1267 view_63"  [label="(4, 12, 64, 64)", style=solid];
+"1256 new_zeros_5" -> "1257 view_62"  [label="(16, 16)", style=solid];
+"1257 view_62" -> "1258 permute_53"  [label="(2, 8, 2, 8)", style=solid];
+"1258 permute_53" -> "1259 reshape_51"  [label="(2, 2, 8, 8)", style=solid];
+"1259 reshape_51" -> "1260 unsqueeze_32"  [label="(4, 64)", style=solid];
+"1259 reshape_51" -> "1261 unsqueeze_33"  [label="(4, 64)", style=solid];
+"1260 unsqueeze_32" -> "1262 sub_5"  [label="(4, 1, 64)", style=solid];
+"1261 unsqueeze_33" -> "1262 sub_5"  [label="(4, 64, 1)", style=solid];
+"1262 sub_5" -> "1263 ne_5"  [label="(4, 64, 64)", style=solid];
+"1262 sub_5" -> "1264 masked_fill_10"  [label="(4, 64, 64)", style=solid];
+"1262 sub_5" -> "1265 eq_5"  [label="(4, 64, 64)", style=solid];
+"1263 ne_5" -> "1264 masked_fill_10"  [label="(4, 64, 64)", style=solid];
+"1264 masked_fill_10" -> "1266 masked_fill_11"  [label="(4, 64, 64)", style=solid];
+"1265 eq_5" -> "1266 masked_fill_11"  [label="(4, 64, 64)", style=solid];
+"1266 masked_fill_11" -> "1268 unsqueeze_34"  [label="(4, 64, 64)", style=solid];
+"1267 view_63" -> "1270 add_39"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1268 unsqueeze_34" -> "1269 unsqueeze_35"  [label="(4, 1, 64, 64)", style=solid];
+"1269 unsqueeze_35" -> "1270 add_39"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1270 add_39" -> "1271 view_64"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1271 view_64" -> "1272 softmax_11"  [label="(4, 12, 64, 64)", style=solid];
+"1272 softmax_11" -> "1273 dropout_44"  [label="(4, 12, 64, 64)", style=solid];
+"1273 dropout_44" -> "1274 matmul_23"  [label="(4, 12, 64, 64)", style=solid];
+"1274 matmul_23" -> "1275 transpose_23"  [label="(4, 12, 64, 32)", style=solid];
+"1275 transpose_23" -> "1276 reshape_52"  [label="(4, 64, 12, 32)", style=solid];
+"1276 reshape_52" -> "1277 reshape_52_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1277 reshape_52_0_0_nncf_smooth_quant_0" -> "1278 quantize_per_tensor_default_72"  [label="(4, 64, 384)", style=solid];
+"1278 quantize_per_tensor_default_72" -> "1279 dequantize_per_tensor_default_72"  [label="(4, 64, 384)", style=solid];
+"1279 dequantize_per_tensor_default_72" -> "1286 linear_71"  [label="(4, 64, 384)", style=solid];
+"1280 features_5_7_attn_proj_bias_0_0" -> "1286 linear_71"  [label="(384,)", style=solid];
+"1281 scale_updated_constant48" -> "1283 mul_tensor_48"  [label="(384, 1)", style=solid];
+"1282 compressed_weight_updated_constant48" -> "1283 mul_tensor_48"  [label="(384, 384)", style=solid];
+"1283 mul_tensor_48" -> "1285 sub_tensor_48"  [label="(384, 384)", style=solid];
+"1284 zero_point_updated_constant48" -> "1285 sub_tensor_48"  [label="(384, 1)", style=solid];
+"1285 sub_tensor_48" -> "1286 linear_71"  [label="(384, 384)", style=solid];
+"1286 linear_71" -> "1287 dropout_45"  [label="(4, 64, 384)", style=solid];
+"1287 dropout_45" -> "1288 view_65"  [label="(4, 64, 384)", style=solid];
+"1288 view_65" -> "1289 permute_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1289 permute_54" -> "1290 reshape_53"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1290 reshape_53" -> "1291 roll_11"  [label="(1, 16, 16, 384)", style=solid];
+"1291 roll_11" -> "1292 slice_185"  [label="(1, 16, 16, 384)", style=solid];
+"1292 slice_185" -> "1293 slice_186"  [label="(1, 16, 16, 384)", style=solid];
+"1293 slice_186" -> "1294 slice_187"  [label="(1, 14, 16, 384)", style=solid];
+"1294 slice_187" -> "1295 slice_188"  [label="(1, 14, 14, 384)", style=solid];
+"1295 slice_188" -> "1296 contiguous_21"  [label="(1, 14, 14, 384)", style=solid];
+"1296 contiguous_21" -> "1297 layer_norm_25"  [label="(1, 14, 14, 384)", style=solid];
+"1297 layer_norm_25" -> "1298 add_40"  [label="(1, 14, 14, 384)", style=solid];
+"1298 add_40" -> "1299 add_40_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1298 add_40" -> "1323 add_41"  [label="(1, 14, 14, 384)", style=solid];
+"1299 add_40_0_0_nncf_smooth_quant_0" -> "1300 quantize_per_tensor_default_73"  [label="(1, 14, 14, 384)", style=solid];
+"1300 quantize_per_tensor_default_73" -> "1301 dequantize_per_tensor_default_73"  [label="(1, 14, 14, 384)", style=solid];
+"1301 dequantize_per_tensor_default_73" -> "1308 linear_72"  [label="(1, 14, 14, 384)", style=solid];
+"1302 features_5_7_mlp_0_bias_0_0" -> "1308 linear_72"  [label="(1536,)", style=solid];
+"1303 scale_updated_constant49" -> "1305 mul_tensor_49"  [label="(1536, 1)", style=solid];
+"1304 compressed_weight_updated_constant49" -> "1305 mul_tensor_49"  [label="(1536, 384)", style=solid];
+"1305 mul_tensor_49" -> "1307 sub_tensor_49"  [label="(1536, 384)", style=solid];
+"1306 zero_point_updated_constant49" -> "1307 sub_tensor_49"  [label="(1536, 1)", style=solid];
+"1307 sub_tensor_49" -> "1308 linear_72"  [label="(1536, 384)", style=solid];
+"1308 linear_72" -> "1309 gelu_11"  [label="(1, 14, 14, 1536)", style=solid];
+"1309 gelu_11" -> "1310 dropout_46"  [label="(1, 14, 14, 1536)", style=solid];
+"1310 dropout_46" -> "1311 dropout_46_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1311 dropout_46_0_0_nncf_smooth_quant_0" -> "1312 quantize_per_tensor_default_74"  [label="(1, 14, 14, 1536)", style=solid];
+"1312 quantize_per_tensor_default_74" -> "1313 dequantize_per_tensor_default_74"  [label="(1, 14, 14, 1536)", style=solid];
+"1313 dequantize_per_tensor_default_74" -> "1320 linear_73"  [label="(1, 14, 14, 1536)", style=solid];
+"1314 features_5_7_mlp_3_bias_0_0" -> "1320 linear_73"  [label="(384,)", style=solid];
+"1315 scale_updated_constant50" -> "1317 mul_tensor_50"  [label="(384, 1)", style=solid];
+"1316 compressed_weight_updated_constant50" -> "1317 mul_tensor_50"  [label="(384, 1536)", style=solid];
+"1317 mul_tensor_50" -> "1319 sub_tensor_50"  [label="(384, 1536)", style=solid];
+"1318 zero_point_updated_constant50" -> "1319 sub_tensor_50"  [label="(384, 1)", style=solid];
+"1319 sub_tensor_50" -> "1320 linear_73"  [label="(384, 1536)", style=solid];
+"1320 linear_73" -> "1321 dropout_47"  [label="(1, 14, 14, 384)", style=solid];
+"1321 dropout_47" -> "1322 layer_norm_26"  [label="(1, 14, 14, 384)", style=solid];
+"1322 layer_norm_26" -> "1323 add_41"  [label="(1, 14, 14, 384)", style=solid];
+"1323 add_41" -> "1325 pad_14"  [label="(1, 14, 14, 384)", style=solid];
+"1323 add_41" -> "1386 add_43"  [label="(1, 14, 14, 384)", style=solid];
+"1324 _frozen_param102" -> "1360 add_42"  [label="(1, 12, 64, 64)", style=solid];
+"1325 pad_14" -> "1326 view_68"  [label="(1, 16, 16, 384)", style=solid];
+"1326 view_68" -> "1327 permute_56"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1327 permute_56" -> "1328 reshape_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1328 reshape_54" -> "1329 reshape_54_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1329 reshape_54_0_0_nncf_smooth_quant_0" -> "1330 quantize_per_tensor_default_75"  [label="(4, 64, 384)", style=solid];
+"1330 quantize_per_tensor_default_75" -> "1331 dequantize_per_tensor_default_75"  [label="(4, 64, 384)", style=solid];
+"1331 dequantize_per_tensor_default_75" -> "1338 linear_76"  [label="(4, 64, 384)", style=solid];
+"1332 _frozen_param103_0_0" -> "1338 linear_76"  [label="(1152,)", style=solid];
+"1333 scale_updated_constant51" -> "1335 mul_tensor_51"  [label="(1152, 1)", style=solid];
+"1334 compressed_weight_updated_constant51" -> "1335 mul_tensor_51"  [label="(1152, 384)", style=solid];
+"1335 mul_tensor_51" -> "1337 sub_tensor_51"  [label="(1152, 384)", style=solid];
+"1336 zero_point_updated_constant51" -> "1337 sub_tensor_51"  [label="(1152, 1)", style=solid];
+"1337 sub_tensor_51" -> "1338 linear_76"  [label="(1152, 384)", style=solid];
+"1338 linear_76" -> "1339 reshape_55"  [label="(4, 64, 1152)", style=solid];
+"1339 reshape_55" -> "1340 permute_57"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1340 permute_57" -> "1341 select_36"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1340 permute_57" -> "1342 select_37"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1340 permute_57" -> "1343 select_38"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1341 select_36" -> "1344 linalg_vector_norm_24"  [label="(4, 12, 64, 32)", style=solid];
+"1341 select_36" -> "1346 expand_as_24"  [label="(4, 12, 64, 32)", style=solid];
+"1341 select_36" -> "1347 div_24"  [label="(4, 12, 64, 32)", style=solid];
+"1342 select_37" -> "1350 linalg_vector_norm_25"  [label="(4, 12, 64, 32)", style=solid];
+"1342 select_37" -> "1352 expand_as_25"  [label="(4, 12, 64, 32)", style=solid];
+"1342 select_37" -> "1353 div_25"  [label="(4, 12, 64, 32)", style=solid];
+"1343 select_38" -> "1363 matmul_25"  [label="(4, 12, 64, 32)", style=solid];
+"1344 linalg_vector_norm_24" -> "1345 clamp_min_24"  [label="(4, 12, 64, 1)", style=solid];
+"1345 clamp_min_24" -> "1346 expand_as_24"  [label="(4, 12, 64, 1)", style=solid];
+"1346 expand_as_24" -> "1347 div_24"  [label="(4, 12, 64, 32)", style=solid];
+"1347 div_24" -> "1348 quantize_per_tensor_default_76"  [label="(4, 12, 64, 32)", style=solid];
+"1348 quantize_per_tensor_default_76" -> "1349 dequantize_per_tensor_default_76"  [label="(4, 12, 64, 32)", style=solid];
+"1349 dequantize_per_tensor_default_76" -> "1357 matmul_24"  [label="(4, 12, 64, 32)", style=solid];
+"1350 linalg_vector_norm_25" -> "1351 clamp_min_25"  [label="(4, 12, 64, 1)", style=solid];
+"1351 clamp_min_25" -> "1352 expand_as_25"  [label="(4, 12, 64, 1)", style=solid];
+"1352 expand_as_25" -> "1353 div_25"  [label="(4, 12, 64, 32)", style=solid];
+"1353 div_25" -> "1354 quantize_per_tensor_default_77"  [label="(4, 12, 64, 32)", style=solid];
+"1354 quantize_per_tensor_default_77" -> "1355 dequantize_per_tensor_default_77"  [label="(4, 12, 64, 32)", style=solid];
+"1355 dequantize_per_tensor_default_77" -> "1356 transpose_24"  [label="(4, 12, 64, 32)", style=solid];
+"1356 transpose_24" -> "1357 matmul_24"  [label="(4, 12, 32, 64)", style=solid];
+"1357 matmul_24" -> "1359 mul_25"  [label="(4, 12, 64, 64)", style=solid];
+"1358 _frozen_param105" -> "1359 mul_25"  [label="(12, 1, 1)", style=solid];
+"1359 mul_25" -> "1360 add_42"  [label="(4, 12, 64, 64)", style=solid];
+"1360 add_42" -> "1361 softmax_12"  [label="(4, 12, 64, 64)", style=solid];
+"1361 softmax_12" -> "1362 dropout_48"  [label="(4, 12, 64, 64)", style=solid];
+"1362 dropout_48" -> "1363 matmul_25"  [label="(4, 12, 64, 64)", style=solid];
+"1363 matmul_25" -> "1364 transpose_25"  [label="(4, 12, 64, 32)", style=solid];
+"1364 transpose_25" -> "1365 reshape_56"  [label="(4, 64, 12, 32)", style=solid];
+"1365 reshape_56" -> "1366 reshape_56_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1366 reshape_56_0_0_nncf_smooth_quant_0" -> "1367 quantize_per_tensor_default_78"  [label="(4, 64, 384)", style=solid];
+"1367 quantize_per_tensor_default_78" -> "1368 dequantize_per_tensor_default_78"  [label="(4, 64, 384)", style=solid];
+"1368 dequantize_per_tensor_default_78" -> "1375 linear_77"  [label="(4, 64, 384)", style=solid];
+"1369 features_5_8_attn_proj_bias_0_0" -> "1375 linear_77"  [label="(384,)", style=solid];
+"1370 scale_updated_constant52" -> "1372 mul_tensor_52"  [label="(384, 1)", style=solid];
+"1371 compressed_weight_updated_constant52" -> "1372 mul_tensor_52"  [label="(384, 384)", style=solid];
+"1372 mul_tensor_52" -> "1374 sub_tensor_52"  [label="(384, 384)", style=solid];
+"1373 zero_point_updated_constant52" -> "1374 sub_tensor_52"  [label="(384, 1)", style=solid];
+"1374 sub_tensor_52" -> "1375 linear_77"  [label="(384, 384)", style=solid];
+"1375 linear_77" -> "1376 dropout_49"  [label="(4, 64, 384)", style=solid];
+"1376 dropout_49" -> "1377 view_69"  [label="(4, 64, 384)", style=solid];
+"1377 view_69" -> "1378 permute_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1378 permute_58" -> "1379 reshape_57"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1379 reshape_57" -> "1380 slice_190"  [label="(1, 16, 16, 384)", style=solid];
+"1380 slice_190" -> "1381 slice_191"  [label="(1, 16, 16, 384)", style=solid];
+"1381 slice_191" -> "1382 slice_192"  [label="(1, 14, 16, 384)", style=solid];
+"1382 slice_192" -> "1383 slice_193"  [label="(1, 14, 14, 384)", style=solid];
+"1383 slice_193" -> "1384 contiguous_23"  [label="(1, 14, 14, 384)", style=solid];
+"1384 contiguous_23" -> "1385 layer_norm_27"  [label="(1, 14, 14, 384)", style=solid];
+"1385 layer_norm_27" -> "1386 add_43"  [label="(1, 14, 14, 384)", style=solid];
+"1386 add_43" -> "1387 add_43_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1386 add_43" -> "1411 add_44"  [label="(1, 14, 14, 384)", style=solid];
+"1387 add_43_0_0_nncf_smooth_quant_0" -> "1388 quantize_per_tensor_default_79"  [label="(1, 14, 14, 384)", style=solid];
+"1388 quantize_per_tensor_default_79" -> "1389 dequantize_per_tensor_default_79"  [label="(1, 14, 14, 384)", style=solid];
+"1389 dequantize_per_tensor_default_79" -> "1396 linear_78"  [label="(1, 14, 14, 384)", style=solid];
+"1390 features_5_8_mlp_0_bias_0_0" -> "1396 linear_78"  [label="(1536,)", style=solid];
+"1391 scale_updated_constant53" -> "1393 mul_tensor_53"  [label="(1536, 1)", style=solid];
+"1392 compressed_weight_updated_constant53" -> "1393 mul_tensor_53"  [label="(1536, 384)", style=solid];
+"1393 mul_tensor_53" -> "1395 sub_tensor_53"  [label="(1536, 384)", style=solid];
+"1394 zero_point_updated_constant53" -> "1395 sub_tensor_53"  [label="(1536, 1)", style=solid];
+"1395 sub_tensor_53" -> "1396 linear_78"  [label="(1536, 384)", style=solid];
+"1396 linear_78" -> "1397 gelu_12"  [label="(1, 14, 14, 1536)", style=solid];
+"1397 gelu_12" -> "1398 dropout_50"  [label="(1, 14, 14, 1536)", style=solid];
+"1398 dropout_50" -> "1399 dropout_50_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1399 dropout_50_0_0_nncf_smooth_quant_0" -> "1400 quantize_per_tensor_default_80"  [label="(1, 14, 14, 1536)", style=solid];
+"1400 quantize_per_tensor_default_80" -> "1401 dequantize_per_tensor_default_80"  [label="(1, 14, 14, 1536)", style=solid];
+"1401 dequantize_per_tensor_default_80" -> "1408 linear_79"  [label="(1, 14, 14, 1536)", style=solid];
+"1402 features_5_8_mlp_3_bias_0_0" -> "1408 linear_79"  [label="(384,)", style=solid];
+"1403 scale_updated_constant54" -> "1405 mul_tensor_54"  [label="(384, 1)", style=solid];
+"1404 compressed_weight_updated_constant54" -> "1405 mul_tensor_54"  [label="(384, 1536)", style=solid];
+"1405 mul_tensor_54" -> "1407 sub_tensor_54"  [label="(384, 1536)", style=solid];
+"1406 zero_point_updated_constant54" -> "1407 sub_tensor_54"  [label="(384, 1)", style=solid];
+"1407 sub_tensor_54" -> "1408 linear_79"  [label="(384, 1536)", style=solid];
+"1408 linear_79" -> "1409 dropout_51"  [label="(1, 14, 14, 384)", style=solid];
+"1409 dropout_51" -> "1410 layer_norm_28"  [label="(1, 14, 14, 384)", style=solid];
+"1410 layer_norm_28" -> "1411 add_44"  [label="(1, 14, 14, 384)", style=solid];
+"1411 add_44" -> "1413 pad_15"  [label="(1, 14, 14, 384)", style=solid];
+"1411 add_44" -> "1492 add_47"  [label="(1, 14, 14, 384)", style=solid];
+"1412 _frozen_param106" -> "1449 add_45"  [label="(1, 12, 64, 64)", style=solid];
+"1413 pad_15" -> "1414 roll_12"  [label="(1, 16, 16, 384)", style=solid];
+"1414 roll_12" -> "1415 view_72"  [label="(1, 16, 16, 384)", style=solid];
+"1415 view_72" -> "1416 permute_60"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1416 permute_60" -> "1417 reshape_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1417 reshape_58" -> "1418 reshape_58_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1417 reshape_58" -> "1450 new_zeros_6"  [label="(4, 64, 384)", style=solid];
+"1418 reshape_58_0_0_nncf_smooth_quant_0" -> "1419 quantize_per_tensor_default_81"  [label="(4, 64, 384)", style=solid];
+"1419 quantize_per_tensor_default_81" -> "1420 dequantize_per_tensor_default_81"  [label="(4, 64, 384)", style=solid];
+"1420 dequantize_per_tensor_default_81" -> "1427 linear_82"  [label="(4, 64, 384)", style=solid];
+"1421 _frozen_param107_0_0" -> "1427 linear_82"  [label="(1152,)", style=solid];
+"1422 scale_updated_constant55" -> "1424 mul_tensor_55"  [label="(1152, 1)", style=solid];
+"1423 compressed_weight_updated_constant55" -> "1424 mul_tensor_55"  [label="(1152, 384)", style=solid];
+"1424 mul_tensor_55" -> "1426 sub_tensor_55"  [label="(1152, 384)", style=solid];
+"1425 zero_point_updated_constant55" -> "1426 sub_tensor_55"  [label="(1152, 1)", style=solid];
+"1426 sub_tensor_55" -> "1427 linear_82"  [label="(1152, 384)", style=solid];
+"1427 linear_82" -> "1428 reshape_59"  [label="(4, 64, 1152)", style=solid];
+"1428 reshape_59" -> "1429 permute_61"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1429 permute_61" -> "1430 select_39"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1429 permute_61" -> "1431 select_40"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1429 permute_61" -> "1432 select_41"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1430 select_39" -> "1433 linalg_vector_norm_26"  [label="(4, 12, 64, 32)", style=solid];
+"1430 select_39" -> "1435 expand_as_26"  [label="(4, 12, 64, 32)", style=solid];
+"1430 select_39" -> "1436 div_26"  [label="(4, 12, 64, 32)", style=solid];
+"1431 select_40" -> "1439 linalg_vector_norm_27"  [label="(4, 12, 64, 32)", style=solid];
+"1431 select_40" -> "1441 expand_as_27"  [label="(4, 12, 64, 32)", style=solid];
+"1431 select_40" -> "1442 div_27"  [label="(4, 12, 64, 32)", style=solid];
+"1432 select_41" -> "1468 matmul_27"  [label="(4, 12, 64, 32)", style=solid];
+"1433 linalg_vector_norm_26" -> "1434 clamp_min_26"  [label="(4, 12, 64, 1)", style=solid];
+"1434 clamp_min_26" -> "1435 expand_as_26"  [label="(4, 12, 64, 1)", style=solid];
+"1435 expand_as_26" -> "1436 div_26"  [label="(4, 12, 64, 32)", style=solid];
+"1436 div_26" -> "1437 quantize_per_tensor_default_82"  [label="(4, 12, 64, 32)", style=solid];
+"1437 quantize_per_tensor_default_82" -> "1438 dequantize_per_tensor_default_82"  [label="(4, 12, 64, 32)", style=solid];
+"1438 dequantize_per_tensor_default_82" -> "1446 matmul_26"  [label="(4, 12, 64, 32)", style=solid];
+"1439 linalg_vector_norm_27" -> "1440 clamp_min_27"  [label="(4, 12, 64, 1)", style=solid];
+"1440 clamp_min_27" -> "1441 expand_as_27"  [label="(4, 12, 64, 1)", style=solid];
+"1441 expand_as_27" -> "1442 div_27"  [label="(4, 12, 64, 32)", style=solid];
+"1442 div_27" -> "1443 quantize_per_tensor_default_83"  [label="(4, 12, 64, 32)", style=solid];
+"1443 quantize_per_tensor_default_83" -> "1444 dequantize_per_tensor_default_83"  [label="(4, 12, 64, 32)", style=solid];
+"1444 dequantize_per_tensor_default_83" -> "1445 transpose_26"  [label="(4, 12, 64, 32)", style=solid];
+"1445 transpose_26" -> "1446 matmul_26"  [label="(4, 12, 32, 64)", style=solid];
+"1446 matmul_26" -> "1448 mul_27"  [label="(4, 12, 64, 64)", style=solid];
+"1447 _frozen_param109" -> "1448 mul_27"  [label="(12, 1, 1)", style=solid];
+"1448 mul_27" -> "1449 add_45"  [label="(4, 12, 64, 64)", style=solid];
+"1449 add_45" -> "1461 view_74"  [label="(4, 12, 64, 64)", style=solid];
+"1450 new_zeros_6" -> "1451 view_73"  [label="(16, 16)", style=solid];
+"1451 view_73" -> "1452 permute_62"  [label="(2, 8, 2, 8)", style=solid];
+"1452 permute_62" -> "1453 reshape_60"  [label="(2, 2, 8, 8)", style=solid];
+"1453 reshape_60" -> "1454 unsqueeze_38"  [label="(4, 64)", style=solid];
+"1453 reshape_60" -> "1455 unsqueeze_39"  [label="(4, 64)", style=solid];
+"1454 unsqueeze_38" -> "1456 sub_6"  [label="(4, 1, 64)", style=solid];
+"1455 unsqueeze_39" -> "1456 sub_6"  [label="(4, 64, 1)", style=solid];
+"1456 sub_6" -> "1457 ne_6"  [label="(4, 64, 64)", style=solid];
+"1456 sub_6" -> "1458 masked_fill_12"  [label="(4, 64, 64)", style=solid];
+"1456 sub_6" -> "1459 eq_6"  [label="(4, 64, 64)", style=solid];
+"1457 ne_6" -> "1458 masked_fill_12"  [label="(4, 64, 64)", style=solid];
+"1458 masked_fill_12" -> "1460 masked_fill_13"  [label="(4, 64, 64)", style=solid];
+"1459 eq_6" -> "1460 masked_fill_13"  [label="(4, 64, 64)", style=solid];
+"1460 masked_fill_13" -> "1462 unsqueeze_40"  [label="(4, 64, 64)", style=solid];
+"1461 view_74" -> "1464 add_46"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1462 unsqueeze_40" -> "1463 unsqueeze_41"  [label="(4, 1, 64, 64)", style=solid];
+"1463 unsqueeze_41" -> "1464 add_46"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1464 add_46" -> "1465 view_75"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1465 view_75" -> "1466 softmax_13"  [label="(4, 12, 64, 64)", style=solid];
+"1466 softmax_13" -> "1467 dropout_52"  [label="(4, 12, 64, 64)", style=solid];
+"1467 dropout_52" -> "1468 matmul_27"  [label="(4, 12, 64, 64)", style=solid];
+"1468 matmul_27" -> "1469 transpose_27"  [label="(4, 12, 64, 32)", style=solid];
+"1469 transpose_27" -> "1470 reshape_61"  [label="(4, 64, 12, 32)", style=solid];
+"1470 reshape_61" -> "1471 reshape_61_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1471 reshape_61_0_0_nncf_smooth_quant_0" -> "1472 quantize_per_tensor_default_84"  [label="(4, 64, 384)", style=solid];
+"1472 quantize_per_tensor_default_84" -> "1473 dequantize_per_tensor_default_84"  [label="(4, 64, 384)", style=solid];
+"1473 dequantize_per_tensor_default_84" -> "1480 linear_83"  [label="(4, 64, 384)", style=solid];
+"1474 features_5_9_attn_proj_bias_0_0" -> "1480 linear_83"  [label="(384,)", style=solid];
+"1475 scale_updated_constant56" -> "1477 mul_tensor_56"  [label="(384, 1)", style=solid];
+"1476 compressed_weight_updated_constant56" -> "1477 mul_tensor_56"  [label="(384, 384)", style=solid];
+"1477 mul_tensor_56" -> "1479 sub_tensor_56"  [label="(384, 384)", style=solid];
+"1478 zero_point_updated_constant56" -> "1479 sub_tensor_56"  [label="(384, 1)", style=solid];
+"1479 sub_tensor_56" -> "1480 linear_83"  [label="(384, 384)", style=solid];
+"1480 linear_83" -> "1481 dropout_53"  [label="(4, 64, 384)", style=solid];
+"1481 dropout_53" -> "1482 view_76"  [label="(4, 64, 384)", style=solid];
+"1482 view_76" -> "1483 permute_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1483 permute_63" -> "1484 reshape_62"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1484 reshape_62" -> "1485 roll_13"  [label="(1, 16, 16, 384)", style=solid];
+"1485 roll_13" -> "1486 slice_213"  [label="(1, 16, 16, 384)", style=solid];
+"1486 slice_213" -> "1487 slice_214"  [label="(1, 16, 16, 384)", style=solid];
+"1487 slice_214" -> "1488 slice_215"  [label="(1, 14, 16, 384)", style=solid];
+"1488 slice_215" -> "1489 slice_216"  [label="(1, 14, 14, 384)", style=solid];
+"1489 slice_216" -> "1490 contiguous_25"  [label="(1, 14, 14, 384)", style=solid];
+"1490 contiguous_25" -> "1491 layer_norm_29"  [label="(1, 14, 14, 384)", style=solid];
+"1491 layer_norm_29" -> "1492 add_47"  [label="(1, 14, 14, 384)", style=solid];
+"1492 add_47" -> "1493 add_47_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1492 add_47" -> "1517 add_48"  [label="(1, 14, 14, 384)", style=solid];
+"1493 add_47_0_0_nncf_smooth_quant_0" -> "1494 quantize_per_tensor_default_85"  [label="(1, 14, 14, 384)", style=solid];
+"1494 quantize_per_tensor_default_85" -> "1495 dequantize_per_tensor_default_85"  [label="(1, 14, 14, 384)", style=solid];
+"1495 dequantize_per_tensor_default_85" -> "1502 linear_84"  [label="(1, 14, 14, 384)", style=solid];
+"1496 features_5_9_mlp_0_bias_0_0" -> "1502 linear_84"  [label="(1536,)", style=solid];
+"1497 scale_updated_constant57" -> "1499 mul_tensor_57"  [label="(1536, 1)", style=solid];
+"1498 compressed_weight_updated_constant57" -> "1499 mul_tensor_57"  [label="(1536, 384)", style=solid];
+"1499 mul_tensor_57" -> "1501 sub_tensor_57"  [label="(1536, 384)", style=solid];
+"1500 zero_point_updated_constant57" -> "1501 sub_tensor_57"  [label="(1536, 1)", style=solid];
+"1501 sub_tensor_57" -> "1502 linear_84"  [label="(1536, 384)", style=solid];
+"1502 linear_84" -> "1503 gelu_13"  [label="(1, 14, 14, 1536)", style=solid];
+"1503 gelu_13" -> "1504 dropout_54"  [label="(1, 14, 14, 1536)", style=solid];
+"1504 dropout_54" -> "1505 dropout_54_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1505 dropout_54_0_0_nncf_smooth_quant_0" -> "1506 quantize_per_tensor_default_86"  [label="(1, 14, 14, 1536)", style=solid];
+"1506 quantize_per_tensor_default_86" -> "1507 dequantize_per_tensor_default_86"  [label="(1, 14, 14, 1536)", style=solid];
+"1507 dequantize_per_tensor_default_86" -> "1514 linear_85"  [label="(1, 14, 14, 1536)", style=solid];
+"1508 features_5_9_mlp_3_bias_0_0" -> "1514 linear_85"  [label="(384,)", style=solid];
+"1509 scale_updated_constant58" -> "1511 mul_tensor_58"  [label="(384, 1)", style=solid];
+"1510 compressed_weight_updated_constant58" -> "1511 mul_tensor_58"  [label="(384, 1536)", style=solid];
+"1511 mul_tensor_58" -> "1513 sub_tensor_58"  [label="(384, 1536)", style=solid];
+"1512 zero_point_updated_constant58" -> "1513 sub_tensor_58"  [label="(384, 1)", style=solid];
+"1513 sub_tensor_58" -> "1514 linear_85"  [label="(384, 1536)", style=solid];
+"1514 linear_85" -> "1515 dropout_55"  [label="(1, 14, 14, 384)", style=solid];
+"1515 dropout_55" -> "1516 layer_norm_30"  [label="(1, 14, 14, 384)", style=solid];
+"1516 layer_norm_30" -> "1517 add_48"  [label="(1, 14, 14, 384)", style=solid];
+"1517 add_48" -> "1519 pad_16"  [label="(1, 14, 14, 384)", style=solid];
+"1517 add_48" -> "1580 add_50"  [label="(1, 14, 14, 384)", style=solid];
+"1518 _frozen_param119" -> "1554 add_49"  [label="(1, 12, 64, 64)", style=solid];
+"1519 pad_16" -> "1520 view_79"  [label="(1, 16, 16, 384)", style=solid];
+"1520 view_79" -> "1521 permute_65"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1521 permute_65" -> "1522 reshape_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1522 reshape_63" -> "1523 reshape_63_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1523 reshape_63_0_0_nncf_smooth_quant_0" -> "1524 quantize_per_tensor_default_87"  [label="(4, 64, 384)", style=solid];
+"1524 quantize_per_tensor_default_87" -> "1525 dequantize_per_tensor_default_87"  [label="(4, 64, 384)", style=solid];
+"1525 dequantize_per_tensor_default_87" -> "1532 linear_88"  [label="(4, 64, 384)", style=solid];
+"1526 _frozen_param120_0_0" -> "1532 linear_88"  [label="(1152,)", style=solid];
+"1527 scale_updated_constant59" -> "1529 mul_tensor_59"  [label="(1152, 1)", style=solid];
+"1528 compressed_weight_updated_constant59" -> "1529 mul_tensor_59"  [label="(1152, 384)", style=solid];
+"1529 mul_tensor_59" -> "1531 sub_tensor_59"  [label="(1152, 384)", style=solid];
+"1530 zero_point_updated_constant59" -> "1531 sub_tensor_59"  [label="(1152, 1)", style=solid];
+"1531 sub_tensor_59" -> "1532 linear_88"  [label="(1152, 384)", style=solid];
+"1532 linear_88" -> "1533 reshape_64"  [label="(4, 64, 1152)", style=solid];
+"1533 reshape_64" -> "1534 permute_66"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1534 permute_66" -> "1535 select_42"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1534 permute_66" -> "1536 select_43"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1534 permute_66" -> "1537 select_44"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1535 select_42" -> "1538 linalg_vector_norm_28"  [label="(4, 12, 64, 32)", style=solid];
+"1535 select_42" -> "1540 expand_as_28"  [label="(4, 12, 64, 32)", style=solid];
+"1535 select_42" -> "1541 div_28"  [label="(4, 12, 64, 32)", style=solid];
+"1536 select_43" -> "1544 linalg_vector_norm_29"  [label="(4, 12, 64, 32)", style=solid];
+"1536 select_43" -> "1546 expand_as_29"  [label="(4, 12, 64, 32)", style=solid];
+"1536 select_43" -> "1547 div_29"  [label="(4, 12, 64, 32)", style=solid];
+"1537 select_44" -> "1557 matmul_29"  [label="(4, 12, 64, 32)", style=solid];
+"1538 linalg_vector_norm_28" -> "1539 clamp_min_28"  [label="(4, 12, 64, 1)", style=solid];
+"1539 clamp_min_28" -> "1540 expand_as_28"  [label="(4, 12, 64, 1)", style=solid];
+"1540 expand_as_28" -> "1541 div_28"  [label="(4, 12, 64, 32)", style=solid];
+"1541 div_28" -> "1542 quantize_per_tensor_default_88"  [label="(4, 12, 64, 32)", style=solid];
+"1542 quantize_per_tensor_default_88" -> "1543 dequantize_per_tensor_default_88"  [label="(4, 12, 64, 32)", style=solid];
+"1543 dequantize_per_tensor_default_88" -> "1551 matmul_28"  [label="(4, 12, 64, 32)", style=solid];
+"1544 linalg_vector_norm_29" -> "1545 clamp_min_29"  [label="(4, 12, 64, 1)", style=solid];
+"1545 clamp_min_29" -> "1546 expand_as_29"  [label="(4, 12, 64, 1)", style=solid];
+"1546 expand_as_29" -> "1547 div_29"  [label="(4, 12, 64, 32)", style=solid];
+"1547 div_29" -> "1548 quantize_per_tensor_default_89"  [label="(4, 12, 64, 32)", style=solid];
+"1548 quantize_per_tensor_default_89" -> "1549 dequantize_per_tensor_default_89"  [label="(4, 12, 64, 32)", style=solid];
+"1549 dequantize_per_tensor_default_89" -> "1550 transpose_28"  [label="(4, 12, 64, 32)", style=solid];
+"1550 transpose_28" -> "1551 matmul_28"  [label="(4, 12, 32, 64)", style=solid];
+"1551 matmul_28" -> "1553 mul_29"  [label="(4, 12, 64, 64)", style=solid];
+"1552 _frozen_param122" -> "1553 mul_29"  [label="(12, 1, 1)", style=solid];
+"1553 mul_29" -> "1554 add_49"  [label="(4, 12, 64, 64)", style=solid];
+"1554 add_49" -> "1555 softmax_14"  [label="(4, 12, 64, 64)", style=solid];
+"1555 softmax_14" -> "1556 dropout_56"  [label="(4, 12, 64, 64)", style=solid];
+"1556 dropout_56" -> "1557 matmul_29"  [label="(4, 12, 64, 64)", style=solid];
+"1557 matmul_29" -> "1558 transpose_29"  [label="(4, 12, 64, 32)", style=solid];
+"1558 transpose_29" -> "1559 reshape_65"  [label="(4, 64, 12, 32)", style=solid];
+"1559 reshape_65" -> "1560 reshape_65_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1560 reshape_65_0_0_nncf_smooth_quant_0" -> "1561 quantize_per_tensor_default_90"  [label="(4, 64, 384)", style=solid];
+"1561 quantize_per_tensor_default_90" -> "1562 dequantize_per_tensor_default_90"  [label="(4, 64, 384)", style=solid];
+"1562 dequantize_per_tensor_default_90" -> "1569 linear_89"  [label="(4, 64, 384)", style=solid];
+"1563 features_5_10_attn_proj_bias_0_0" -> "1569 linear_89"  [label="(384,)", style=solid];
+"1564 scale_updated_constant60" -> "1566 mul_tensor_60"  [label="(384, 1)", style=solid];
+"1565 compressed_weight_updated_constant60" -> "1566 mul_tensor_60"  [label="(384, 384)", style=solid];
+"1566 mul_tensor_60" -> "1568 sub_tensor_60"  [label="(384, 384)", style=solid];
+"1567 zero_point_updated_constant60" -> "1568 sub_tensor_60"  [label="(384, 1)", style=solid];
+"1568 sub_tensor_60" -> "1569 linear_89"  [label="(384, 384)", style=solid];
+"1569 linear_89" -> "1570 dropout_57"  [label="(4, 64, 384)", style=solid];
+"1570 dropout_57" -> "1571 view_80"  [label="(4, 64, 384)", style=solid];
+"1571 view_80" -> "1572 permute_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1572 permute_67" -> "1573 reshape_66"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1573 reshape_66" -> "1574 slice_218"  [label="(1, 16, 16, 384)", style=solid];
+"1574 slice_218" -> "1575 slice_219"  [label="(1, 16, 16, 384)", style=solid];
+"1575 slice_219" -> "1576 slice_220"  [label="(1, 14, 16, 384)", style=solid];
+"1576 slice_220" -> "1577 slice_221"  [label="(1, 14, 14, 384)", style=solid];
+"1577 slice_221" -> "1578 contiguous_27"  [label="(1, 14, 14, 384)", style=solid];
+"1578 contiguous_27" -> "1579 layer_norm_31"  [label="(1, 14, 14, 384)", style=solid];
+"1579 layer_norm_31" -> "1580 add_50"  [label="(1, 14, 14, 384)", style=solid];
+"1580 add_50" -> "1581 add_50_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1580 add_50" -> "1605 add_51"  [label="(1, 14, 14, 384)", style=solid];
+"1581 add_50_0_0_nncf_smooth_quant_0" -> "1582 quantize_per_tensor_default_91"  [label="(1, 14, 14, 384)", style=solid];
+"1582 quantize_per_tensor_default_91" -> "1583 dequantize_per_tensor_default_91"  [label="(1, 14, 14, 384)", style=solid];
+"1583 dequantize_per_tensor_default_91" -> "1590 linear_90"  [label="(1, 14, 14, 384)", style=solid];
+"1584 features_5_10_mlp_0_bias_0_0" -> "1590 linear_90"  [label="(1536,)", style=solid];
+"1585 scale_updated_constant61" -> "1587 mul_tensor_61"  [label="(1536, 1)", style=solid];
+"1586 compressed_weight_updated_constant61" -> "1587 mul_tensor_61"  [label="(1536, 384)", style=solid];
+"1587 mul_tensor_61" -> "1589 sub_tensor_61"  [label="(1536, 384)", style=solid];
+"1588 zero_point_updated_constant61" -> "1589 sub_tensor_61"  [label="(1536, 1)", style=solid];
+"1589 sub_tensor_61" -> "1590 linear_90"  [label="(1536, 384)", style=solid];
+"1590 linear_90" -> "1591 gelu_14"  [label="(1, 14, 14, 1536)", style=solid];
+"1591 gelu_14" -> "1592 dropout_58"  [label="(1, 14, 14, 1536)", style=solid];
+"1592 dropout_58" -> "1593 dropout_58_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1593 dropout_58_0_0_nncf_smooth_quant_0" -> "1594 quantize_per_tensor_default_92"  [label="(1, 14, 14, 1536)", style=solid];
+"1594 quantize_per_tensor_default_92" -> "1595 dequantize_per_tensor_default_92"  [label="(1, 14, 14, 1536)", style=solid];
+"1595 dequantize_per_tensor_default_92" -> "1602 linear_91"  [label="(1, 14, 14, 1536)", style=solid];
+"1596 features_5_10_mlp_3_bias_0_0" -> "1602 linear_91"  [label="(384,)", style=solid];
+"1597 scale_updated_constant62" -> "1599 mul_tensor_62"  [label="(384, 1)", style=solid];
+"1598 compressed_weight_updated_constant62" -> "1599 mul_tensor_62"  [label="(384, 1536)", style=solid];
+"1599 mul_tensor_62" -> "1601 sub_tensor_62"  [label="(384, 1536)", style=solid];
+"1600 zero_point_updated_constant62" -> "1601 sub_tensor_62"  [label="(384, 1)", style=solid];
+"1601 sub_tensor_62" -> "1602 linear_91"  [label="(384, 1536)", style=solid];
+"1602 linear_91" -> "1603 dropout_59"  [label="(1, 14, 14, 384)", style=solid];
+"1603 dropout_59" -> "1604 layer_norm_32"  [label="(1, 14, 14, 384)", style=solid];
+"1604 layer_norm_32" -> "1605 add_51"  [label="(1, 14, 14, 384)", style=solid];
+"1605 add_51" -> "1607 pad_17"  [label="(1, 14, 14, 384)", style=solid];
+"1605 add_51" -> "1686 add_54"  [label="(1, 14, 14, 384)", style=solid];
+"1606 _frozen_param123" -> "1643 add_52"  [label="(1, 12, 64, 64)", style=solid];
+"1607 pad_17" -> "1608 roll_14"  [label="(1, 16, 16, 384)", style=solid];
+"1608 roll_14" -> "1609 view_83"  [label="(1, 16, 16, 384)", style=solid];
+"1609 view_83" -> "1610 permute_69"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1610 permute_69" -> "1611 reshape_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1611 reshape_67" -> "1612 reshape_67_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1611 reshape_67" -> "1644 new_zeros_7"  [label="(4, 64, 384)", style=solid];
+"1612 reshape_67_0_0_nncf_smooth_quant_0" -> "1613 quantize_per_tensor_default_93"  [label="(4, 64, 384)", style=solid];
+"1613 quantize_per_tensor_default_93" -> "1614 dequantize_per_tensor_default_93"  [label="(4, 64, 384)", style=solid];
+"1614 dequantize_per_tensor_default_93" -> "1621 linear_94"  [label="(4, 64, 384)", style=solid];
+"1615 _frozen_param124_0_0" -> "1621 linear_94"  [label="(1152,)", style=solid];
+"1616 scale_updated_constant63" -> "1618 mul_tensor_63"  [label="(1152, 1)", style=solid];
+"1617 compressed_weight_updated_constant63" -> "1618 mul_tensor_63"  [label="(1152, 384)", style=solid];
+"1618 mul_tensor_63" -> "1620 sub_tensor_63"  [label="(1152, 384)", style=solid];
+"1619 zero_point_updated_constant63" -> "1620 sub_tensor_63"  [label="(1152, 1)", style=solid];
+"1620 sub_tensor_63" -> "1621 linear_94"  [label="(1152, 384)", style=solid];
+"1621 linear_94" -> "1622 reshape_68"  [label="(4, 64, 1152)", style=solid];
+"1622 reshape_68" -> "1623 permute_70"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1623 permute_70" -> "1624 select_45"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1623 permute_70" -> "1625 select_46"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1623 permute_70" -> "1626 select_47"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1624 select_45" -> "1627 linalg_vector_norm_30"  [label="(4, 12, 64, 32)", style=solid];
+"1624 select_45" -> "1629 expand_as_30"  [label="(4, 12, 64, 32)", style=solid];
+"1624 select_45" -> "1630 div_30"  [label="(4, 12, 64, 32)", style=solid];
+"1625 select_46" -> "1633 linalg_vector_norm_31"  [label="(4, 12, 64, 32)", style=solid];
+"1625 select_46" -> "1635 expand_as_31"  [label="(4, 12, 64, 32)", style=solid];
+"1625 select_46" -> "1636 div_31"  [label="(4, 12, 64, 32)", style=solid];
+"1626 select_47" -> "1662 matmul_31"  [label="(4, 12, 64, 32)", style=solid];
+"1627 linalg_vector_norm_30" -> "1628 clamp_min_30"  [label="(4, 12, 64, 1)", style=solid];
+"1628 clamp_min_30" -> "1629 expand_as_30"  [label="(4, 12, 64, 1)", style=solid];
+"1629 expand_as_30" -> "1630 div_30"  [label="(4, 12, 64, 32)", style=solid];
+"1630 div_30" -> "1631 quantize_per_tensor_default_94"  [label="(4, 12, 64, 32)", style=solid];
+"1631 quantize_per_tensor_default_94" -> "1632 dequantize_per_tensor_default_94"  [label="(4, 12, 64, 32)", style=solid];
+"1632 dequantize_per_tensor_default_94" -> "1640 matmul_30"  [label="(4, 12, 64, 32)", style=solid];
+"1633 linalg_vector_norm_31" -> "1634 clamp_min_31"  [label="(4, 12, 64, 1)", style=solid];
+"1634 clamp_min_31" -> "1635 expand_as_31"  [label="(4, 12, 64, 1)", style=solid];
+"1635 expand_as_31" -> "1636 div_31"  [label="(4, 12, 64, 32)", style=solid];
+"1636 div_31" -> "1637 quantize_per_tensor_default_95"  [label="(4, 12, 64, 32)", style=solid];
+"1637 quantize_per_tensor_default_95" -> "1638 dequantize_per_tensor_default_95"  [label="(4, 12, 64, 32)", style=solid];
+"1638 dequantize_per_tensor_default_95" -> "1639 transpose_30"  [label="(4, 12, 64, 32)", style=solid];
+"1639 transpose_30" -> "1640 matmul_30"  [label="(4, 12, 32, 64)", style=solid];
+"1640 matmul_30" -> "1642 mul_31"  [label="(4, 12, 64, 64)", style=solid];
+"1641 _frozen_param126" -> "1642 mul_31"  [label="(12, 1, 1)", style=solid];
+"1642 mul_31" -> "1643 add_52"  [label="(4, 12, 64, 64)", style=solid];
+"1643 add_52" -> "1655 view_85"  [label="(4, 12, 64, 64)", style=solid];
+"1644 new_zeros_7" -> "1645 view_84"  [label="(16, 16)", style=solid];
+"1645 view_84" -> "1646 permute_71"  [label="(2, 8, 2, 8)", style=solid];
+"1646 permute_71" -> "1647 reshape_69"  [label="(2, 2, 8, 8)", style=solid];
+"1647 reshape_69" -> "1648 unsqueeze_44"  [label="(4, 64)", style=solid];
+"1647 reshape_69" -> "1649 unsqueeze_45"  [label="(4, 64)", style=solid];
+"1648 unsqueeze_44" -> "1650 sub_7"  [label="(4, 1, 64)", style=solid];
+"1649 unsqueeze_45" -> "1650 sub_7"  [label="(4, 64, 1)", style=solid];
+"1650 sub_7" -> "1651 ne_7"  [label="(4, 64, 64)", style=solid];
+"1650 sub_7" -> "1652 masked_fill_14"  [label="(4, 64, 64)", style=solid];
+"1650 sub_7" -> "1653 eq_7"  [label="(4, 64, 64)", style=solid];
+"1651 ne_7" -> "1652 masked_fill_14"  [label="(4, 64, 64)", style=solid];
+"1652 masked_fill_14" -> "1654 masked_fill_15"  [label="(4, 64, 64)", style=solid];
+"1653 eq_7" -> "1654 masked_fill_15"  [label="(4, 64, 64)", style=solid];
+"1654 masked_fill_15" -> "1656 unsqueeze_46"  [label="(4, 64, 64)", style=solid];
+"1655 view_85" -> "1658 add_53"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1656 unsqueeze_46" -> "1657 unsqueeze_47"  [label="(4, 1, 64, 64)", style=solid];
+"1657 unsqueeze_47" -> "1658 add_53"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1658 add_53" -> "1659 view_86"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1659 view_86" -> "1660 softmax_15"  [label="(4, 12, 64, 64)", style=solid];
+"1660 softmax_15" -> "1661 dropout_60"  [label="(4, 12, 64, 64)", style=solid];
+"1661 dropout_60" -> "1662 matmul_31"  [label="(4, 12, 64, 64)", style=solid];
+"1662 matmul_31" -> "1663 transpose_31"  [label="(4, 12, 64, 32)", style=solid];
+"1663 transpose_31" -> "1664 reshape_70"  [label="(4, 64, 12, 32)", style=solid];
+"1664 reshape_70" -> "1665 reshape_70_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1665 reshape_70_0_0_nncf_smooth_quant_0" -> "1666 quantize_per_tensor_default_96"  [label="(4, 64, 384)", style=solid];
+"1666 quantize_per_tensor_default_96" -> "1667 dequantize_per_tensor_default_96"  [label="(4, 64, 384)", style=solid];
+"1667 dequantize_per_tensor_default_96" -> "1674 linear_95"  [label="(4, 64, 384)", style=solid];
+"1668 features_5_11_attn_proj_bias_0_0" -> "1674 linear_95"  [label="(384,)", style=solid];
+"1669 scale_updated_constant64" -> "1671 mul_tensor_64"  [label="(384, 1)", style=solid];
+"1670 compressed_weight_updated_constant64" -> "1671 mul_tensor_64"  [label="(384, 384)", style=solid];
+"1671 mul_tensor_64" -> "1673 sub_tensor_64"  [label="(384, 384)", style=solid];
+"1672 zero_point_updated_constant64" -> "1673 sub_tensor_64"  [label="(384, 1)", style=solid];
+"1673 sub_tensor_64" -> "1674 linear_95"  [label="(384, 384)", style=solid];
+"1674 linear_95" -> "1675 dropout_61"  [label="(4, 64, 384)", style=solid];
+"1675 dropout_61" -> "1676 view_87"  [label="(4, 64, 384)", style=solid];
+"1676 view_87" -> "1677 permute_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1677 permute_72" -> "1678 reshape_71"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1678 reshape_71" -> "1679 roll_15"  [label="(1, 16, 16, 384)", style=solid];
+"1679 roll_15" -> "1680 slice_241"  [label="(1, 16, 16, 384)", style=solid];
+"1680 slice_241" -> "1681 slice_242"  [label="(1, 16, 16, 384)", style=solid];
+"1681 slice_242" -> "1682 slice_243"  [label="(1, 14, 16, 384)", style=solid];
+"1682 slice_243" -> "1683 slice_244"  [label="(1, 14, 14, 384)", style=solid];
+"1683 slice_244" -> "1684 contiguous_29"  [label="(1, 14, 14, 384)", style=solid];
+"1684 contiguous_29" -> "1685 layer_norm_33"  [label="(1, 14, 14, 384)", style=solid];
+"1685 layer_norm_33" -> "1686 add_54"  [label="(1, 14, 14, 384)", style=solid];
+"1686 add_54" -> "1687 add_54_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1686 add_54" -> "1711 add_55"  [label="(1, 14, 14, 384)", style=solid];
+"1687 add_54_0_0_nncf_smooth_quant_0" -> "1688 quantize_per_tensor_default_97"  [label="(1, 14, 14, 384)", style=solid];
+"1688 quantize_per_tensor_default_97" -> "1689 dequantize_per_tensor_default_97"  [label="(1, 14, 14, 384)", style=solid];
+"1689 dequantize_per_tensor_default_97" -> "1696 linear_96"  [label="(1, 14, 14, 384)", style=solid];
+"1690 features_5_11_mlp_0_bias_0_0" -> "1696 linear_96"  [label="(1536,)", style=solid];
+"1691 scale_updated_constant65" -> "1693 mul_tensor_65"  [label="(1536, 1)", style=solid];
+"1692 compressed_weight_updated_constant65" -> "1693 mul_tensor_65"  [label="(1536, 384)", style=solid];
+"1693 mul_tensor_65" -> "1695 sub_tensor_65"  [label="(1536, 384)", style=solid];
+"1694 zero_point_updated_constant65" -> "1695 sub_tensor_65"  [label="(1536, 1)", style=solid];
+"1695 sub_tensor_65" -> "1696 linear_96"  [label="(1536, 384)", style=solid];
+"1696 linear_96" -> "1697 gelu_15"  [label="(1, 14, 14, 1536)", style=solid];
+"1697 gelu_15" -> "1698 dropout_62"  [label="(1, 14, 14, 1536)", style=solid];
+"1698 dropout_62" -> "1699 dropout_62_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1699 dropout_62_0_0_nncf_smooth_quant_0" -> "1700 quantize_per_tensor_default_98"  [label="(1, 14, 14, 1536)", style=solid];
+"1700 quantize_per_tensor_default_98" -> "1701 dequantize_per_tensor_default_98"  [label="(1, 14, 14, 1536)", style=solid];
+"1701 dequantize_per_tensor_default_98" -> "1708 linear_97"  [label="(1, 14, 14, 1536)", style=solid];
+"1702 features_5_11_mlp_3_bias_0_0" -> "1708 linear_97"  [label="(384,)", style=solid];
+"1703 scale_updated_constant66" -> "1705 mul_tensor_66"  [label="(384, 1)", style=solid];
+"1704 compressed_weight_updated_constant66" -> "1705 mul_tensor_66"  [label="(384, 1536)", style=solid];
+"1705 mul_tensor_66" -> "1707 sub_tensor_66"  [label="(384, 1536)", style=solid];
+"1706 zero_point_updated_constant66" -> "1707 sub_tensor_66"  [label="(384, 1)", style=solid];
+"1707 sub_tensor_66" -> "1708 linear_97"  [label="(384, 1536)", style=solid];
+"1708 linear_97" -> "1709 dropout_63"  [label="(1, 14, 14, 384)", style=solid];
+"1709 dropout_63" -> "1710 layer_norm_34"  [label="(1, 14, 14, 384)", style=solid];
+"1710 layer_norm_34" -> "1711 add_55"  [label="(1, 14, 14, 384)", style=solid];
+"1711 add_55" -> "1713 pad_18"  [label="(1, 14, 14, 384)", style=solid];
+"1711 add_55" -> "1774 add_57"  [label="(1, 14, 14, 384)", style=solid];
+"1712 _frozen_param136" -> "1748 add_56"  [label="(1, 12, 64, 64)", style=solid];
+"1713 pad_18" -> "1714 view_90"  [label="(1, 16, 16, 384)", style=solid];
+"1714 view_90" -> "1715 permute_74"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1715 permute_74" -> "1716 reshape_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1716 reshape_72" -> "1717 reshape_72_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1717 reshape_72_0_0_nncf_smooth_quant_0" -> "1718 quantize_per_tensor_default_99"  [label="(4, 64, 384)", style=solid];
+"1718 quantize_per_tensor_default_99" -> "1719 dequantize_per_tensor_default_99"  [label="(4, 64, 384)", style=solid];
+"1719 dequantize_per_tensor_default_99" -> "1726 linear_100"  [label="(4, 64, 384)", style=solid];
+"1720 _frozen_param137_0_0" -> "1726 linear_100"  [label="(1152,)", style=solid];
+"1721 scale_updated_constant67" -> "1723 mul_tensor_67"  [label="(1152, 1)", style=solid];
+"1722 compressed_weight_updated_constant67" -> "1723 mul_tensor_67"  [label="(1152, 384)", style=solid];
+"1723 mul_tensor_67" -> "1725 sub_tensor_67"  [label="(1152, 384)", style=solid];
+"1724 zero_point_updated_constant67" -> "1725 sub_tensor_67"  [label="(1152, 1)", style=solid];
+"1725 sub_tensor_67" -> "1726 linear_100"  [label="(1152, 384)", style=solid];
+"1726 linear_100" -> "1727 reshape_73"  [label="(4, 64, 1152)", style=solid];
+"1727 reshape_73" -> "1728 permute_75"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1728 permute_75" -> "1729 select_48"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1728 permute_75" -> "1730 select_49"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1728 permute_75" -> "1731 select_50"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1729 select_48" -> "1732 linalg_vector_norm_32"  [label="(4, 12, 64, 32)", style=solid];
+"1729 select_48" -> "1734 expand_as_32"  [label="(4, 12, 64, 32)", style=solid];
+"1729 select_48" -> "1735 div_32"  [label="(4, 12, 64, 32)", style=solid];
+"1730 select_49" -> "1738 linalg_vector_norm_33"  [label="(4, 12, 64, 32)", style=solid];
+"1730 select_49" -> "1740 expand_as_33"  [label="(4, 12, 64, 32)", style=solid];
+"1730 select_49" -> "1741 div_33"  [label="(4, 12, 64, 32)", style=solid];
+"1731 select_50" -> "1751 matmul_33"  [label="(4, 12, 64, 32)", style=solid];
+"1732 linalg_vector_norm_32" -> "1733 clamp_min_32"  [label="(4, 12, 64, 1)", style=solid];
+"1733 clamp_min_32" -> "1734 expand_as_32"  [label="(4, 12, 64, 1)", style=solid];
+"1734 expand_as_32" -> "1735 div_32"  [label="(4, 12, 64, 32)", style=solid];
+"1735 div_32" -> "1736 quantize_per_tensor_default_100"  [label="(4, 12, 64, 32)", style=solid];
+"1736 quantize_per_tensor_default_100" -> "1737 dequantize_per_tensor_default_100"  [label="(4, 12, 64, 32)", style=solid];
+"1737 dequantize_per_tensor_default_100" -> "1745 matmul_32"  [label="(4, 12, 64, 32)", style=solid];
+"1738 linalg_vector_norm_33" -> "1739 clamp_min_33"  [label="(4, 12, 64, 1)", style=solid];
+"1739 clamp_min_33" -> "1740 expand_as_33"  [label="(4, 12, 64, 1)", style=solid];
+"1740 expand_as_33" -> "1741 div_33"  [label="(4, 12, 64, 32)", style=solid];
+"1741 div_33" -> "1742 quantize_per_tensor_default_101"  [label="(4, 12, 64, 32)", style=solid];
+"1742 quantize_per_tensor_default_101" -> "1743 dequantize_per_tensor_default_101"  [label="(4, 12, 64, 32)", style=solid];
+"1743 dequantize_per_tensor_default_101" -> "1744 transpose_32"  [label="(4, 12, 64, 32)", style=solid];
+"1744 transpose_32" -> "1745 matmul_32"  [label="(4, 12, 32, 64)", style=solid];
+"1745 matmul_32" -> "1747 mul_33"  [label="(4, 12, 64, 64)", style=solid];
+"1746 _frozen_param139" -> "1747 mul_33"  [label="(12, 1, 1)", style=solid];
+"1747 mul_33" -> "1748 add_56"  [label="(4, 12, 64, 64)", style=solid];
+"1748 add_56" -> "1749 softmax_16"  [label="(4, 12, 64, 64)", style=solid];
+"1749 softmax_16" -> "1750 dropout_64"  [label="(4, 12, 64, 64)", style=solid];
+"1750 dropout_64" -> "1751 matmul_33"  [label="(4, 12, 64, 64)", style=solid];
+"1751 matmul_33" -> "1752 transpose_33"  [label="(4, 12, 64, 32)", style=solid];
+"1752 transpose_33" -> "1753 reshape_74"  [label="(4, 64, 12, 32)", style=solid];
+"1753 reshape_74" -> "1754 reshape_74_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1754 reshape_74_0_0_nncf_smooth_quant_0" -> "1755 quantize_per_tensor_default_102"  [label="(4, 64, 384)", style=solid];
+"1755 quantize_per_tensor_default_102" -> "1756 dequantize_per_tensor_default_102"  [label="(4, 64, 384)", style=solid];
+"1756 dequantize_per_tensor_default_102" -> "1763 linear_101"  [label="(4, 64, 384)", style=solid];
+"1757 features_5_12_attn_proj_bias_0_0" -> "1763 linear_101"  [label="(384,)", style=solid];
+"1758 scale_updated_constant68" -> "1760 mul_tensor_68"  [label="(384, 1)", style=solid];
+"1759 compressed_weight_updated_constant68" -> "1760 mul_tensor_68"  [label="(384, 384)", style=solid];
+"1760 mul_tensor_68" -> "1762 sub_tensor_68"  [label="(384, 384)", style=solid];
+"1761 zero_point_updated_constant68" -> "1762 sub_tensor_68"  [label="(384, 1)", style=solid];
+"1762 sub_tensor_68" -> "1763 linear_101"  [label="(384, 384)", style=solid];
+"1763 linear_101" -> "1764 dropout_65"  [label="(4, 64, 384)", style=solid];
+"1764 dropout_65" -> "1765 view_91"  [label="(4, 64, 384)", style=solid];
+"1765 view_91" -> "1766 permute_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1766 permute_76" -> "1767 reshape_75"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1767 reshape_75" -> "1768 slice_246"  [label="(1, 16, 16, 384)", style=solid];
+"1768 slice_246" -> "1769 slice_247"  [label="(1, 16, 16, 384)", style=solid];
+"1769 slice_247" -> "1770 slice_248"  [label="(1, 14, 16, 384)", style=solid];
+"1770 slice_248" -> "1771 slice_249"  [label="(1, 14, 14, 384)", style=solid];
+"1771 slice_249" -> "1772 contiguous_31"  [label="(1, 14, 14, 384)", style=solid];
+"1772 contiguous_31" -> "1773 layer_norm_35"  [label="(1, 14, 14, 384)", style=solid];
+"1773 layer_norm_35" -> "1774 add_57"  [label="(1, 14, 14, 384)", style=solid];
+"1774 add_57" -> "1775 add_57_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1774 add_57" -> "1799 add_58"  [label="(1, 14, 14, 384)", style=solid];
+"1775 add_57_0_0_nncf_smooth_quant_0" -> "1776 quantize_per_tensor_default_103"  [label="(1, 14, 14, 384)", style=solid];
+"1776 quantize_per_tensor_default_103" -> "1777 dequantize_per_tensor_default_103"  [label="(1, 14, 14, 384)", style=solid];
+"1777 dequantize_per_tensor_default_103" -> "1784 linear_102"  [label="(1, 14, 14, 384)", style=solid];
+"1778 features_5_12_mlp_0_bias_0_0" -> "1784 linear_102"  [label="(1536,)", style=solid];
+"1779 scale_updated_constant69" -> "1781 mul_tensor_69"  [label="(1536, 1)", style=solid];
+"1780 compressed_weight_updated_constant69" -> "1781 mul_tensor_69"  [label="(1536, 384)", style=solid];
+"1781 mul_tensor_69" -> "1783 sub_tensor_69"  [label="(1536, 384)", style=solid];
+"1782 zero_point_updated_constant69" -> "1783 sub_tensor_69"  [label="(1536, 1)", style=solid];
+"1783 sub_tensor_69" -> "1784 linear_102"  [label="(1536, 384)", style=solid];
+"1784 linear_102" -> "1785 gelu_16"  [label="(1, 14, 14, 1536)", style=solid];
+"1785 gelu_16" -> "1786 dropout_66"  [label="(1, 14, 14, 1536)", style=solid];
+"1786 dropout_66" -> "1787 dropout_66_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1787 dropout_66_0_0_nncf_smooth_quant_0" -> "1788 quantize_per_tensor_default_104"  [label="(1, 14, 14, 1536)", style=solid];
+"1788 quantize_per_tensor_default_104" -> "1789 dequantize_per_tensor_default_104"  [label="(1, 14, 14, 1536)", style=solid];
+"1789 dequantize_per_tensor_default_104" -> "1796 linear_103"  [label="(1, 14, 14, 1536)", style=solid];
+"1790 features_5_12_mlp_3_bias_0_0" -> "1796 linear_103"  [label="(384,)", style=solid];
+"1791 scale_updated_constant70" -> "1793 mul_tensor_70"  [label="(384, 1)", style=solid];
+"1792 compressed_weight_updated_constant70" -> "1793 mul_tensor_70"  [label="(384, 1536)", style=solid];
+"1793 mul_tensor_70" -> "1795 sub_tensor_70"  [label="(384, 1536)", style=solid];
+"1794 zero_point_updated_constant70" -> "1795 sub_tensor_70"  [label="(384, 1)", style=solid];
+"1795 sub_tensor_70" -> "1796 linear_103"  [label="(384, 1536)", style=solid];
+"1796 linear_103" -> "1797 dropout_67"  [label="(1, 14, 14, 384)", style=solid];
+"1797 dropout_67" -> "1798 layer_norm_36"  [label="(1, 14, 14, 384)", style=solid];
+"1798 layer_norm_36" -> "1799 add_58"  [label="(1, 14, 14, 384)", style=solid];
+"1799 add_58" -> "1801 pad_19"  [label="(1, 14, 14, 384)", style=solid];
+"1799 add_58" -> "1880 add_61"  [label="(1, 14, 14, 384)", style=solid];
+"1800 _frozen_param140" -> "1837 add_59"  [label="(1, 12, 64, 64)", style=solid];
+"1801 pad_19" -> "1802 roll_16"  [label="(1, 16, 16, 384)", style=solid];
+"1802 roll_16" -> "1803 view_94"  [label="(1, 16, 16, 384)", style=solid];
+"1803 view_94" -> "1804 permute_78"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1804 permute_78" -> "1805 reshape_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1805 reshape_76" -> "1806 reshape_76_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1805 reshape_76" -> "1838 new_zeros_8"  [label="(4, 64, 384)", style=solid];
+"1806 reshape_76_0_0_nncf_smooth_quant_0" -> "1807 quantize_per_tensor_default_105"  [label="(4, 64, 384)", style=solid];
+"1807 quantize_per_tensor_default_105" -> "1808 dequantize_per_tensor_default_105"  [label="(4, 64, 384)", style=solid];
+"1808 dequantize_per_tensor_default_105" -> "1815 linear_106"  [label="(4, 64, 384)", style=solid];
+"1809 _frozen_param141_0_0" -> "1815 linear_106"  [label="(1152,)", style=solid];
+"1810 scale_updated_constant71" -> "1812 mul_tensor_71"  [label="(1152, 1)", style=solid];
+"1811 compressed_weight_updated_constant71" -> "1812 mul_tensor_71"  [label="(1152, 384)", style=solid];
+"1812 mul_tensor_71" -> "1814 sub_tensor_71"  [label="(1152, 384)", style=solid];
+"1813 zero_point_updated_constant71" -> "1814 sub_tensor_71"  [label="(1152, 1)", style=solid];
+"1814 sub_tensor_71" -> "1815 linear_106"  [label="(1152, 384)", style=solid];
+"1815 linear_106" -> "1816 reshape_77"  [label="(4, 64, 1152)", style=solid];
+"1816 reshape_77" -> "1817 permute_79"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1817 permute_79" -> "1818 select_51"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1817 permute_79" -> "1819 select_52"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1817 permute_79" -> "1820 select_53"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1818 select_51" -> "1821 linalg_vector_norm_34"  [label="(4, 12, 64, 32)", style=solid];
+"1818 select_51" -> "1823 expand_as_34"  [label="(4, 12, 64, 32)", style=solid];
+"1818 select_51" -> "1824 div_34"  [label="(4, 12, 64, 32)", style=solid];
+"1819 select_52" -> "1827 linalg_vector_norm_35"  [label="(4, 12, 64, 32)", style=solid];
+"1819 select_52" -> "1829 expand_as_35"  [label="(4, 12, 64, 32)", style=solid];
+"1819 select_52" -> "1830 div_35"  [label="(4, 12, 64, 32)", style=solid];
+"1820 select_53" -> "1856 matmul_35"  [label="(4, 12, 64, 32)", style=solid];
+"1821 linalg_vector_norm_34" -> "1822 clamp_min_34"  [label="(4, 12, 64, 1)", style=solid];
+"1822 clamp_min_34" -> "1823 expand_as_34"  [label="(4, 12, 64, 1)", style=solid];
+"1823 expand_as_34" -> "1824 div_34"  [label="(4, 12, 64, 32)", style=solid];
+"1824 div_34" -> "1825 quantize_per_tensor_default_106"  [label="(4, 12, 64, 32)", style=solid];
+"1825 quantize_per_tensor_default_106" -> "1826 dequantize_per_tensor_default_106"  [label="(4, 12, 64, 32)", style=solid];
+"1826 dequantize_per_tensor_default_106" -> "1834 matmul_34"  [label="(4, 12, 64, 32)", style=solid];
+"1827 linalg_vector_norm_35" -> "1828 clamp_min_35"  [label="(4, 12, 64, 1)", style=solid];
+"1828 clamp_min_35" -> "1829 expand_as_35"  [label="(4, 12, 64, 1)", style=solid];
+"1829 expand_as_35" -> "1830 div_35"  [label="(4, 12, 64, 32)", style=solid];
+"1830 div_35" -> "1831 quantize_per_tensor_default_107"  [label="(4, 12, 64, 32)", style=solid];
+"1831 quantize_per_tensor_default_107" -> "1832 dequantize_per_tensor_default_107"  [label="(4, 12, 64, 32)", style=solid];
+"1832 dequantize_per_tensor_default_107" -> "1833 transpose_34"  [label="(4, 12, 64, 32)", style=solid];
+"1833 transpose_34" -> "1834 matmul_34"  [label="(4, 12, 32, 64)", style=solid];
+"1834 matmul_34" -> "1836 mul_35"  [label="(4, 12, 64, 64)", style=solid];
+"1835 _frozen_param143" -> "1836 mul_35"  [label="(12, 1, 1)", style=solid];
+"1836 mul_35" -> "1837 add_59"  [label="(4, 12, 64, 64)", style=solid];
+"1837 add_59" -> "1849 view_96"  [label="(4, 12, 64, 64)", style=solid];
+"1838 new_zeros_8" -> "1839 view_95"  [label="(16, 16)", style=solid];
+"1839 view_95" -> "1840 permute_80"  [label="(2, 8, 2, 8)", style=solid];
+"1840 permute_80" -> "1841 reshape_78"  [label="(2, 2, 8, 8)", style=solid];
+"1841 reshape_78" -> "1842 unsqueeze_50"  [label="(4, 64)", style=solid];
+"1841 reshape_78" -> "1843 unsqueeze_51"  [label="(4, 64)", style=solid];
+"1842 unsqueeze_50" -> "1844 sub_8"  [label="(4, 1, 64)", style=solid];
+"1843 unsqueeze_51" -> "1844 sub_8"  [label="(4, 64, 1)", style=solid];
+"1844 sub_8" -> "1845 ne_8"  [label="(4, 64, 64)", style=solid];
+"1844 sub_8" -> "1846 masked_fill_16"  [label="(4, 64, 64)", style=solid];
+"1844 sub_8" -> "1847 eq_8"  [label="(4, 64, 64)", style=solid];
+"1845 ne_8" -> "1846 masked_fill_16"  [label="(4, 64, 64)", style=solid];
+"1846 masked_fill_16" -> "1848 masked_fill_17"  [label="(4, 64, 64)", style=solid];
+"1847 eq_8" -> "1848 masked_fill_17"  [label="(4, 64, 64)", style=solid];
+"1848 masked_fill_17" -> "1850 unsqueeze_52"  [label="(4, 64, 64)", style=solid];
+"1849 view_96" -> "1852 add_60"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1850 unsqueeze_52" -> "1851 unsqueeze_53"  [label="(4, 1, 64, 64)", style=solid];
+"1851 unsqueeze_53" -> "1852 add_60"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1852 add_60" -> "1853 view_97"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1853 view_97" -> "1854 softmax_17"  [label="(4, 12, 64, 64)", style=solid];
+"1854 softmax_17" -> "1855 dropout_68"  [label="(4, 12, 64, 64)", style=solid];
+"1855 dropout_68" -> "1856 matmul_35"  [label="(4, 12, 64, 64)", style=solid];
+"1856 matmul_35" -> "1857 transpose_35"  [label="(4, 12, 64, 32)", style=solid];
+"1857 transpose_35" -> "1858 reshape_79"  [label="(4, 64, 12, 32)", style=solid];
+"1858 reshape_79" -> "1859 reshape_79_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1859 reshape_79_0_0_nncf_smooth_quant_0" -> "1860 quantize_per_tensor_default_108"  [label="(4, 64, 384)", style=solid];
+"1860 quantize_per_tensor_default_108" -> "1861 dequantize_per_tensor_default_108"  [label="(4, 64, 384)", style=solid];
+"1861 dequantize_per_tensor_default_108" -> "1868 linear_107"  [label="(4, 64, 384)", style=solid];
+"1862 features_5_13_attn_proj_bias_0_0" -> "1868 linear_107"  [label="(384,)", style=solid];
+"1863 scale_updated_constant72" -> "1865 mul_tensor_72"  [label="(384, 1)", style=solid];
+"1864 compressed_weight_updated_constant72" -> "1865 mul_tensor_72"  [label="(384, 384)", style=solid];
+"1865 mul_tensor_72" -> "1867 sub_tensor_72"  [label="(384, 384)", style=solid];
+"1866 zero_point_updated_constant72" -> "1867 sub_tensor_72"  [label="(384, 1)", style=solid];
+"1867 sub_tensor_72" -> "1868 linear_107"  [label="(384, 384)", style=solid];
+"1868 linear_107" -> "1869 dropout_69"  [label="(4, 64, 384)", style=solid];
+"1869 dropout_69" -> "1870 view_98"  [label="(4, 64, 384)", style=solid];
+"1870 view_98" -> "1871 permute_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1871 permute_81" -> "1872 reshape_80"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1872 reshape_80" -> "1873 roll_17"  [label="(1, 16, 16, 384)", style=solid];
+"1873 roll_17" -> "1874 slice_269"  [label="(1, 16, 16, 384)", style=solid];
+"1874 slice_269" -> "1875 slice_270"  [label="(1, 16, 16, 384)", style=solid];
+"1875 slice_270" -> "1876 slice_271"  [label="(1, 14, 16, 384)", style=solid];
+"1876 slice_271" -> "1877 slice_272"  [label="(1, 14, 14, 384)", style=solid];
+"1877 slice_272" -> "1878 contiguous_33"  [label="(1, 14, 14, 384)", style=solid];
+"1878 contiguous_33" -> "1879 layer_norm_37"  [label="(1, 14, 14, 384)", style=solid];
+"1879 layer_norm_37" -> "1880 add_61"  [label="(1, 14, 14, 384)", style=solid];
+"1880 add_61" -> "1881 add_61_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1880 add_61" -> "1905 add_62"  [label="(1, 14, 14, 384)", style=solid];
+"1881 add_61_0_0_nncf_smooth_quant_0" -> "1882 quantize_per_tensor_default_109"  [label="(1, 14, 14, 384)", style=solid];
+"1882 quantize_per_tensor_default_109" -> "1883 dequantize_per_tensor_default_109"  [label="(1, 14, 14, 384)", style=solid];
+"1883 dequantize_per_tensor_default_109" -> "1890 linear_108"  [label="(1, 14, 14, 384)", style=solid];
+"1884 features_5_13_mlp_0_bias_0_0" -> "1890 linear_108"  [label="(1536,)", style=solid];
+"1885 scale_updated_constant73" -> "1887 mul_tensor_73"  [label="(1536, 1)", style=solid];
+"1886 compressed_weight_updated_constant73" -> "1887 mul_tensor_73"  [label="(1536, 384)", style=solid];
+"1887 mul_tensor_73" -> "1889 sub_tensor_73"  [label="(1536, 384)", style=solid];
+"1888 zero_point_updated_constant73" -> "1889 sub_tensor_73"  [label="(1536, 1)", style=solid];
+"1889 sub_tensor_73" -> "1890 linear_108"  [label="(1536, 384)", style=solid];
+"1890 linear_108" -> "1891 gelu_17"  [label="(1, 14, 14, 1536)", style=solid];
+"1891 gelu_17" -> "1892 dropout_70"  [label="(1, 14, 14, 1536)", style=solid];
+"1892 dropout_70" -> "1893 dropout_70_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1893 dropout_70_0_0_nncf_smooth_quant_0" -> "1894 quantize_per_tensor_default_110"  [label="(1, 14, 14, 1536)", style=solid];
+"1894 quantize_per_tensor_default_110" -> "1895 dequantize_per_tensor_default_110"  [label="(1, 14, 14, 1536)", style=solid];
+"1895 dequantize_per_tensor_default_110" -> "1902 linear_109"  [label="(1, 14, 14, 1536)", style=solid];
+"1896 features_5_13_mlp_3_bias_0_0" -> "1902 linear_109"  [label="(384,)", style=solid];
+"1897 scale_updated_constant74" -> "1899 mul_tensor_74"  [label="(384, 1)", style=solid];
+"1898 compressed_weight_updated_constant74" -> "1899 mul_tensor_74"  [label="(384, 1536)", style=solid];
+"1899 mul_tensor_74" -> "1901 sub_tensor_74"  [label="(384, 1536)", style=solid];
+"1900 zero_point_updated_constant74" -> "1901 sub_tensor_74"  [label="(384, 1)", style=solid];
+"1901 sub_tensor_74" -> "1902 linear_109"  [label="(384, 1536)", style=solid];
+"1902 linear_109" -> "1903 dropout_71"  [label="(1, 14, 14, 384)", style=solid];
+"1903 dropout_71" -> "1904 layer_norm_38"  [label="(1, 14, 14, 384)", style=solid];
+"1904 layer_norm_38" -> "1905 add_62"  [label="(1, 14, 14, 384)", style=solid];
+"1905 add_62" -> "1907 pad_20"  [label="(1, 14, 14, 384)", style=solid];
+"1905 add_62" -> "1968 add_64"  [label="(1, 14, 14, 384)", style=solid];
+"1906 _frozen_param153" -> "1942 add_63"  [label="(1, 12, 64, 64)", style=solid];
+"1907 pad_20" -> "1908 view_101"  [label="(1, 16, 16, 384)", style=solid];
+"1908 view_101" -> "1909 permute_83"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1909 permute_83" -> "1910 reshape_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1910 reshape_81" -> "1911 reshape_81_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1911 reshape_81_0_0_nncf_smooth_quant_0" -> "1912 quantize_per_tensor_default_111"  [label="(4, 64, 384)", style=solid];
+"1912 quantize_per_tensor_default_111" -> "1913 dequantize_per_tensor_default_111"  [label="(4, 64, 384)", style=solid];
+"1913 dequantize_per_tensor_default_111" -> "1920 linear_112"  [label="(4, 64, 384)", style=solid];
+"1914 _frozen_param154_0_0" -> "1920 linear_112"  [label="(1152,)", style=solid];
+"1915 scale_updated_constant75" -> "1917 mul_tensor_75"  [label="(1152, 1)", style=solid];
+"1916 compressed_weight_updated_constant75" -> "1917 mul_tensor_75"  [label="(1152, 384)", style=solid];
+"1917 mul_tensor_75" -> "1919 sub_tensor_75"  [label="(1152, 384)", style=solid];
+"1918 zero_point_updated_constant75" -> "1919 sub_tensor_75"  [label="(1152, 1)", style=solid];
+"1919 sub_tensor_75" -> "1920 linear_112"  [label="(1152, 384)", style=solid];
+"1920 linear_112" -> "1921 reshape_82"  [label="(4, 64, 1152)", style=solid];
+"1921 reshape_82" -> "1922 permute_84"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1922 permute_84" -> "1923 select_54"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1922 permute_84" -> "1924 select_55"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1922 permute_84" -> "1925 select_56"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1923 select_54" -> "1926 linalg_vector_norm_36"  [label="(4, 12, 64, 32)", style=solid];
+"1923 select_54" -> "1928 expand_as_36"  [label="(4, 12, 64, 32)", style=solid];
+"1923 select_54" -> "1929 div_36"  [label="(4, 12, 64, 32)", style=solid];
+"1924 select_55" -> "1932 linalg_vector_norm_37"  [label="(4, 12, 64, 32)", style=solid];
+"1924 select_55" -> "1934 expand_as_37"  [label="(4, 12, 64, 32)", style=solid];
+"1924 select_55" -> "1935 div_37"  [label="(4, 12, 64, 32)", style=solid];
+"1925 select_56" -> "1945 matmul_37"  [label="(4, 12, 64, 32)", style=solid];
+"1926 linalg_vector_norm_36" -> "1927 clamp_min_36"  [label="(4, 12, 64, 1)", style=solid];
+"1927 clamp_min_36" -> "1928 expand_as_36"  [label="(4, 12, 64, 1)", style=solid];
+"1928 expand_as_36" -> "1929 div_36"  [label="(4, 12, 64, 32)", style=solid];
+"1929 div_36" -> "1930 quantize_per_tensor_default_112"  [label="(4, 12, 64, 32)", style=solid];
+"1930 quantize_per_tensor_default_112" -> "1931 dequantize_per_tensor_default_112"  [label="(4, 12, 64, 32)", style=solid];
+"1931 dequantize_per_tensor_default_112" -> "1939 matmul_36"  [label="(4, 12, 64, 32)", style=solid];
+"1932 linalg_vector_norm_37" -> "1933 clamp_min_37"  [label="(4, 12, 64, 1)", style=solid];
+"1933 clamp_min_37" -> "1934 expand_as_37"  [label="(4, 12, 64, 1)", style=solid];
+"1934 expand_as_37" -> "1935 div_37"  [label="(4, 12, 64, 32)", style=solid];
+"1935 div_37" -> "1936 quantize_per_tensor_default_113"  [label="(4, 12, 64, 32)", style=solid];
+"1936 quantize_per_tensor_default_113" -> "1937 dequantize_per_tensor_default_113"  [label="(4, 12, 64, 32)", style=solid];
+"1937 dequantize_per_tensor_default_113" -> "1938 transpose_36"  [label="(4, 12, 64, 32)", style=solid];
+"1938 transpose_36" -> "1939 matmul_36"  [label="(4, 12, 32, 64)", style=solid];
+"1939 matmul_36" -> "1941 mul_37"  [label="(4, 12, 64, 64)", style=solid];
+"1940 _frozen_param156" -> "1941 mul_37"  [label="(12, 1, 1)", style=solid];
+"1941 mul_37" -> "1942 add_63"  [label="(4, 12, 64, 64)", style=solid];
+"1942 add_63" -> "1943 softmax_18"  [label="(4, 12, 64, 64)", style=solid];
+"1943 softmax_18" -> "1944 dropout_72"  [label="(4, 12, 64, 64)", style=solid];
+"1944 dropout_72" -> "1945 matmul_37"  [label="(4, 12, 64, 64)", style=solid];
+"1945 matmul_37" -> "1946 transpose_37"  [label="(4, 12, 64, 32)", style=solid];
+"1946 transpose_37" -> "1947 reshape_83"  [label="(4, 64, 12, 32)", style=solid];
+"1947 reshape_83" -> "1948 reshape_83_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1948 reshape_83_0_0_nncf_smooth_quant_0" -> "1949 quantize_per_tensor_default_114"  [label="(4, 64, 384)", style=solid];
+"1949 quantize_per_tensor_default_114" -> "1950 dequantize_per_tensor_default_114"  [label="(4, 64, 384)", style=solid];
+"1950 dequantize_per_tensor_default_114" -> "1957 linear_113"  [label="(4, 64, 384)", style=solid];
+"1951 features_5_14_attn_proj_bias_0_0" -> "1957 linear_113"  [label="(384,)", style=solid];
+"1952 scale_updated_constant76" -> "1954 mul_tensor_76"  [label="(384, 1)", style=solid];
+"1953 compressed_weight_updated_constant76" -> "1954 mul_tensor_76"  [label="(384, 384)", style=solid];
+"1954 mul_tensor_76" -> "1956 sub_tensor_76"  [label="(384, 384)", style=solid];
+"1955 zero_point_updated_constant76" -> "1956 sub_tensor_76"  [label="(384, 1)", style=solid];
+"1956 sub_tensor_76" -> "1957 linear_113"  [label="(384, 384)", style=solid];
+"1957 linear_113" -> "1958 dropout_73"  [label="(4, 64, 384)", style=solid];
+"1958 dropout_73" -> "1959 view_102"  [label="(4, 64, 384)", style=solid];
+"1959 view_102" -> "1960 permute_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1960 permute_85" -> "1961 reshape_84"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1961 reshape_84" -> "1962 slice_274"  [label="(1, 16, 16, 384)", style=solid];
+"1962 slice_274" -> "1963 slice_275"  [label="(1, 16, 16, 384)", style=solid];
+"1963 slice_275" -> "1964 slice_276"  [label="(1, 14, 16, 384)", style=solid];
+"1964 slice_276" -> "1965 slice_277"  [label="(1, 14, 14, 384)", style=solid];
+"1965 slice_277" -> "1966 contiguous_35"  [label="(1, 14, 14, 384)", style=solid];
+"1966 contiguous_35" -> "1967 layer_norm_39"  [label="(1, 14, 14, 384)", style=solid];
+"1967 layer_norm_39" -> "1968 add_64"  [label="(1, 14, 14, 384)", style=solid];
+"1968 add_64" -> "1969 add_64_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1968 add_64" -> "1993 add_65"  [label="(1, 14, 14, 384)", style=solid];
+"1969 add_64_0_0_nncf_smooth_quant_0" -> "1970 quantize_per_tensor_default_115"  [label="(1, 14, 14, 384)", style=solid];
+"1970 quantize_per_tensor_default_115" -> "1971 dequantize_per_tensor_default_115"  [label="(1, 14, 14, 384)", style=solid];
+"1971 dequantize_per_tensor_default_115" -> "1978 linear_114"  [label="(1, 14, 14, 384)", style=solid];
+"1972 features_5_14_mlp_0_bias_0_0" -> "1978 linear_114"  [label="(1536,)", style=solid];
+"1973 scale_updated_constant77" -> "1975 mul_tensor_77"  [label="(1536, 1)", style=solid];
+"1974 compressed_weight_updated_constant77" -> "1975 mul_tensor_77"  [label="(1536, 384)", style=solid];
+"1975 mul_tensor_77" -> "1977 sub_tensor_77"  [label="(1536, 384)", style=solid];
+"1976 zero_point_updated_constant77" -> "1977 sub_tensor_77"  [label="(1536, 1)", style=solid];
+"1977 sub_tensor_77" -> "1978 linear_114"  [label="(1536, 384)", style=solid];
+"1978 linear_114" -> "1979 gelu_18"  [label="(1, 14, 14, 1536)", style=solid];
+"1979 gelu_18" -> "1980 dropout_74"  [label="(1, 14, 14, 1536)", style=solid];
+"1980 dropout_74" -> "1981 dropout_74_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1981 dropout_74_0_0_nncf_smooth_quant_0" -> "1982 quantize_per_tensor_default_116"  [label="(1, 14, 14, 1536)", style=solid];
+"1982 quantize_per_tensor_default_116" -> "1983 dequantize_per_tensor_default_116"  [label="(1, 14, 14, 1536)", style=solid];
+"1983 dequantize_per_tensor_default_116" -> "1990 linear_115"  [label="(1, 14, 14, 1536)", style=solid];
+"1984 features_5_14_mlp_3_bias_0_0" -> "1990 linear_115"  [label="(384,)", style=solid];
+"1985 scale_updated_constant78" -> "1987 mul_tensor_78"  [label="(384, 1)", style=solid];
+"1986 compressed_weight_updated_constant78" -> "1987 mul_tensor_78"  [label="(384, 1536)", style=solid];
+"1987 mul_tensor_78" -> "1989 sub_tensor_78"  [label="(384, 1536)", style=solid];
+"1988 zero_point_updated_constant78" -> "1989 sub_tensor_78"  [label="(384, 1)", style=solid];
+"1989 sub_tensor_78" -> "1990 linear_115"  [label="(384, 1536)", style=solid];
+"1990 linear_115" -> "1991 dropout_75"  [label="(1, 14, 14, 384)", style=solid];
+"1991 dropout_75" -> "1992 layer_norm_40"  [label="(1, 14, 14, 384)", style=solid];
+"1992 layer_norm_40" -> "1993 add_65"  [label="(1, 14, 14, 384)", style=solid];
+"1993 add_65" -> "1995 pad_21"  [label="(1, 14, 14, 384)", style=solid];
+"1993 add_65" -> "2074 add_68"  [label="(1, 14, 14, 384)", style=solid];
+"1994 _frozen_param157" -> "2031 add_66"  [label="(1, 12, 64, 64)", style=solid];
+"1995 pad_21" -> "1996 roll_18"  [label="(1, 16, 16, 384)", style=solid];
+"1996 roll_18" -> "1997 view_105"  [label="(1, 16, 16, 384)", style=solid];
+"1997 view_105" -> "1998 permute_87"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1998 permute_87" -> "1999 reshape_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1999 reshape_85" -> "2000 reshape_85_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1999 reshape_85" -> "2032 new_zeros_9"  [label="(4, 64, 384)", style=solid];
+"2000 reshape_85_0_0_nncf_smooth_quant_0" -> "2001 quantize_per_tensor_default_117"  [label="(4, 64, 384)", style=solid];
+"2001 quantize_per_tensor_default_117" -> "2002 dequantize_per_tensor_default_117"  [label="(4, 64, 384)", style=solid];
+"2002 dequantize_per_tensor_default_117" -> "2009 linear_118"  [label="(4, 64, 384)", style=solid];
+"2003 _frozen_param158_0_0" -> "2009 linear_118"  [label="(1152,)", style=solid];
+"2004 scale_updated_constant79" -> "2006 mul_tensor_79"  [label="(1152, 1)", style=solid];
+"2005 compressed_weight_updated_constant79" -> "2006 mul_tensor_79"  [label="(1152, 384)", style=solid];
+"2006 mul_tensor_79" -> "2008 sub_tensor_79"  [label="(1152, 384)", style=solid];
+"2007 zero_point_updated_constant79" -> "2008 sub_tensor_79"  [label="(1152, 1)", style=solid];
+"2008 sub_tensor_79" -> "2009 linear_118"  [label="(1152, 384)", style=solid];
+"2009 linear_118" -> "2010 reshape_86"  [label="(4, 64, 1152)", style=solid];
+"2010 reshape_86" -> "2011 permute_88"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2011 permute_88" -> "2012 select_57"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2011 permute_88" -> "2013 select_58"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2011 permute_88" -> "2014 select_59"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2012 select_57" -> "2015 linalg_vector_norm_38"  [label="(4, 12, 64, 32)", style=solid];
+"2012 select_57" -> "2017 expand_as_38"  [label="(4, 12, 64, 32)", style=solid];
+"2012 select_57" -> "2018 div_38"  [label="(4, 12, 64, 32)", style=solid];
+"2013 select_58" -> "2021 linalg_vector_norm_39"  [label="(4, 12, 64, 32)", style=solid];
+"2013 select_58" -> "2023 expand_as_39"  [label="(4, 12, 64, 32)", style=solid];
+"2013 select_58" -> "2024 div_39"  [label="(4, 12, 64, 32)", style=solid];
+"2014 select_59" -> "2050 matmul_39"  [label="(4, 12, 64, 32)", style=solid];
+"2015 linalg_vector_norm_38" -> "2016 clamp_min_38"  [label="(4, 12, 64, 1)", style=solid];
+"2016 clamp_min_38" -> "2017 expand_as_38"  [label="(4, 12, 64, 1)", style=solid];
+"2017 expand_as_38" -> "2018 div_38"  [label="(4, 12, 64, 32)", style=solid];
+"2018 div_38" -> "2019 quantize_per_tensor_default_118"  [label="(4, 12, 64, 32)", style=solid];
+"2019 quantize_per_tensor_default_118" -> "2020 dequantize_per_tensor_default_118"  [label="(4, 12, 64, 32)", style=solid];
+"2020 dequantize_per_tensor_default_118" -> "2028 matmul_38"  [label="(4, 12, 64, 32)", style=solid];
+"2021 linalg_vector_norm_39" -> "2022 clamp_min_39"  [label="(4, 12, 64, 1)", style=solid];
+"2022 clamp_min_39" -> "2023 expand_as_39"  [label="(4, 12, 64, 1)", style=solid];
+"2023 expand_as_39" -> "2024 div_39"  [label="(4, 12, 64, 32)", style=solid];
+"2024 div_39" -> "2025 quantize_per_tensor_default_119"  [label="(4, 12, 64, 32)", style=solid];
+"2025 quantize_per_tensor_default_119" -> "2026 dequantize_per_tensor_default_119"  [label="(4, 12, 64, 32)", style=solid];
+"2026 dequantize_per_tensor_default_119" -> "2027 transpose_38"  [label="(4, 12, 64, 32)", style=solid];
+"2027 transpose_38" -> "2028 matmul_38"  [label="(4, 12, 32, 64)", style=solid];
+"2028 matmul_38" -> "2030 mul_39"  [label="(4, 12, 64, 64)", style=solid];
+"2029 _frozen_param160" -> "2030 mul_39"  [label="(12, 1, 1)", style=solid];
+"2030 mul_39" -> "2031 add_66"  [label="(4, 12, 64, 64)", style=solid];
+"2031 add_66" -> "2043 view_107"  [label="(4, 12, 64, 64)", style=solid];
+"2032 new_zeros_9" -> "2033 view_106"  [label="(16, 16)", style=solid];
+"2033 view_106" -> "2034 permute_89"  [label="(2, 8, 2, 8)", style=solid];
+"2034 permute_89" -> "2035 reshape_87"  [label="(2, 2, 8, 8)", style=solid];
+"2035 reshape_87" -> "2036 unsqueeze_56"  [label="(4, 64)", style=solid];
+"2035 reshape_87" -> "2037 unsqueeze_57"  [label="(4, 64)", style=solid];
+"2036 unsqueeze_56" -> "2038 sub_9"  [label="(4, 1, 64)", style=solid];
+"2037 unsqueeze_57" -> "2038 sub_9"  [label="(4, 64, 1)", style=solid];
+"2038 sub_9" -> "2039 ne_9"  [label="(4, 64, 64)", style=solid];
+"2038 sub_9" -> "2040 masked_fill_18"  [label="(4, 64, 64)", style=solid];
+"2038 sub_9" -> "2041 eq_9"  [label="(4, 64, 64)", style=solid];
+"2039 ne_9" -> "2040 masked_fill_18"  [label="(4, 64, 64)", style=solid];
+"2040 masked_fill_18" -> "2042 masked_fill_19"  [label="(4, 64, 64)", style=solid];
+"2041 eq_9" -> "2042 masked_fill_19"  [label="(4, 64, 64)", style=solid];
+"2042 masked_fill_19" -> "2044 unsqueeze_58"  [label="(4, 64, 64)", style=solid];
+"2043 view_107" -> "2046 add_67"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2044 unsqueeze_58" -> "2045 unsqueeze_59"  [label="(4, 1, 64, 64)", style=solid];
+"2045 unsqueeze_59" -> "2046 add_67"  [label="(1, 4, 1, 64, 64)", style=solid];
+"2046 add_67" -> "2047 view_108"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2047 view_108" -> "2048 softmax_19"  [label="(4, 12, 64, 64)", style=solid];
+"2048 softmax_19" -> "2049 dropout_76"  [label="(4, 12, 64, 64)", style=solid];
+"2049 dropout_76" -> "2050 matmul_39"  [label="(4, 12, 64, 64)", style=solid];
+"2050 matmul_39" -> "2051 transpose_39"  [label="(4, 12, 64, 32)", style=solid];
+"2051 transpose_39" -> "2052 reshape_88"  [label="(4, 64, 12, 32)", style=solid];
+"2052 reshape_88" -> "2053 reshape_88_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"2053 reshape_88_0_0_nncf_smooth_quant_0" -> "2054 quantize_per_tensor_default_120"  [label="(4, 64, 384)", style=solid];
+"2054 quantize_per_tensor_default_120" -> "2055 dequantize_per_tensor_default_120"  [label="(4, 64, 384)", style=solid];
+"2055 dequantize_per_tensor_default_120" -> "2062 linear_119"  [label="(4, 64, 384)", style=solid];
+"2056 features_5_15_attn_proj_bias_0_0" -> "2062 linear_119"  [label="(384,)", style=solid];
+"2057 scale_updated_constant80" -> "2059 mul_tensor_80"  [label="(384, 1)", style=solid];
+"2058 compressed_weight_updated_constant80" -> "2059 mul_tensor_80"  [label="(384, 384)", style=solid];
+"2059 mul_tensor_80" -> "2061 sub_tensor_80"  [label="(384, 384)", style=solid];
+"2060 zero_point_updated_constant80" -> "2061 sub_tensor_80"  [label="(384, 1)", style=solid];
+"2061 sub_tensor_80" -> "2062 linear_119"  [label="(384, 384)", style=solid];
+"2062 linear_119" -> "2063 dropout_77"  [label="(4, 64, 384)", style=solid];
+"2063 dropout_77" -> "2064 view_109"  [label="(4, 64, 384)", style=solid];
+"2064 view_109" -> "2065 permute_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2065 permute_90" -> "2066 reshape_89"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2066 reshape_89" -> "2067 roll_19"  [label="(1, 16, 16, 384)", style=solid];
+"2067 roll_19" -> "2068 slice_297"  [label="(1, 16, 16, 384)", style=solid];
+"2068 slice_297" -> "2069 slice_298"  [label="(1, 16, 16, 384)", style=solid];
+"2069 slice_298" -> "2070 slice_299"  [label="(1, 14, 16, 384)", style=solid];
+"2070 slice_299" -> "2071 slice_300"  [label="(1, 14, 14, 384)", style=solid];
+"2071 slice_300" -> "2072 contiguous_37"  [label="(1, 14, 14, 384)", style=solid];
+"2072 contiguous_37" -> "2073 layer_norm_41"  [label="(1, 14, 14, 384)", style=solid];
+"2073 layer_norm_41" -> "2074 add_68"  [label="(1, 14, 14, 384)", style=solid];
+"2074 add_68" -> "2075 add_68_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"2074 add_68" -> "2099 add_69"  [label="(1, 14, 14, 384)", style=solid];
+"2075 add_68_0_0_nncf_smooth_quant_0" -> "2076 quantize_per_tensor_default_121"  [label="(1, 14, 14, 384)", style=solid];
+"2076 quantize_per_tensor_default_121" -> "2077 dequantize_per_tensor_default_121"  [label="(1, 14, 14, 384)", style=solid];
+"2077 dequantize_per_tensor_default_121" -> "2084 linear_120"  [label="(1, 14, 14, 384)", style=solid];
+"2078 features_5_15_mlp_0_bias_0_0" -> "2084 linear_120"  [label="(1536,)", style=solid];
+"2079 scale_updated_constant81" -> "2081 mul_tensor_81"  [label="(1536, 1)", style=solid];
+"2080 compressed_weight_updated_constant81" -> "2081 mul_tensor_81"  [label="(1536, 384)", style=solid];
+"2081 mul_tensor_81" -> "2083 sub_tensor_81"  [label="(1536, 384)", style=solid];
+"2082 zero_point_updated_constant81" -> "2083 sub_tensor_81"  [label="(1536, 1)", style=solid];
+"2083 sub_tensor_81" -> "2084 linear_120"  [label="(1536, 384)", style=solid];
+"2084 linear_120" -> "2085 gelu_19"  [label="(1, 14, 14, 1536)", style=solid];
+"2085 gelu_19" -> "2086 dropout_78"  [label="(1, 14, 14, 1536)", style=solid];
+"2086 dropout_78" -> "2087 dropout_78_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"2087 dropout_78_0_0_nncf_smooth_quant_0" -> "2088 quantize_per_tensor_default_122"  [label="(1, 14, 14, 1536)", style=solid];
+"2088 quantize_per_tensor_default_122" -> "2089 dequantize_per_tensor_default_122"  [label="(1, 14, 14, 1536)", style=solid];
+"2089 dequantize_per_tensor_default_122" -> "2096 linear_121"  [label="(1, 14, 14, 1536)", style=solid];
+"2090 features_5_15_mlp_3_bias_0_0" -> "2096 linear_121"  [label="(384,)", style=solid];
+"2091 scale_updated_constant82" -> "2093 mul_tensor_82"  [label="(384, 1)", style=solid];
+"2092 compressed_weight_updated_constant82" -> "2093 mul_tensor_82"  [label="(384, 1536)", style=solid];
+"2093 mul_tensor_82" -> "2095 sub_tensor_82"  [label="(384, 1536)", style=solid];
+"2094 zero_point_updated_constant82" -> "2095 sub_tensor_82"  [label="(384, 1)", style=solid];
+"2095 sub_tensor_82" -> "2096 linear_121"  [label="(384, 1536)", style=solid];
+"2096 linear_121" -> "2097 dropout_79"  [label="(1, 14, 14, 384)", style=solid];
+"2097 dropout_79" -> "2098 layer_norm_42"  [label="(1, 14, 14, 384)", style=solid];
+"2098 layer_norm_42" -> "2099 add_69"  [label="(1, 14, 14, 384)", style=solid];
+"2099 add_69" -> "2101 pad_22"  [label="(1, 14, 14, 384)", style=solid];
+"2099 add_69" -> "2162 add_71"  [label="(1, 14, 14, 384)", style=solid];
+"2100 _frozen_param170" -> "2136 add_70"  [label="(1, 12, 64, 64)", style=solid];
+"2101 pad_22" -> "2102 view_112"  [label="(1, 16, 16, 384)", style=solid];
+"2102 view_112" -> "2103 permute_92"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2103 permute_92" -> "2104 reshape_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2104 reshape_90" -> "2105 reshape_90_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"2105 reshape_90_0_0_nncf_smooth_quant_0" -> "2106 quantize_per_tensor_default_123"  [label="(4, 64, 384)", style=solid];
+"2106 quantize_per_tensor_default_123" -> "2107 dequantize_per_tensor_default_123"  [label="(4, 64, 384)", style=solid];
+"2107 dequantize_per_tensor_default_123" -> "2114 linear_124"  [label="(4, 64, 384)", style=solid];
+"2108 _frozen_param171_0_0" -> "2114 linear_124"  [label="(1152,)", style=solid];
+"2109 scale_updated_constant83" -> "2111 mul_tensor_83"  [label="(1152, 1)", style=solid];
+"2110 compressed_weight_updated_constant83" -> "2111 mul_tensor_83"  [label="(1152, 384)", style=solid];
+"2111 mul_tensor_83" -> "2113 sub_tensor_83"  [label="(1152, 384)", style=solid];
+"2112 zero_point_updated_constant83" -> "2113 sub_tensor_83"  [label="(1152, 1)", style=solid];
+"2113 sub_tensor_83" -> "2114 linear_124"  [label="(1152, 384)", style=solid];
+"2114 linear_124" -> "2115 reshape_91"  [label="(4, 64, 1152)", style=solid];
+"2115 reshape_91" -> "2116 permute_93"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2116 permute_93" -> "2117 select_60"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2116 permute_93" -> "2118 select_61"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2116 permute_93" -> "2119 select_62"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2117 select_60" -> "2120 linalg_vector_norm_40"  [label="(4, 12, 64, 32)", style=solid];
+"2117 select_60" -> "2122 expand_as_40"  [label="(4, 12, 64, 32)", style=solid];
+"2117 select_60" -> "2123 div_40"  [label="(4, 12, 64, 32)", style=solid];
+"2118 select_61" -> "2126 linalg_vector_norm_41"  [label="(4, 12, 64, 32)", style=solid];
+"2118 select_61" -> "2128 expand_as_41"  [label="(4, 12, 64, 32)", style=solid];
+"2118 select_61" -> "2129 div_41"  [label="(4, 12, 64, 32)", style=solid];
+"2119 select_62" -> "2139 matmul_41"  [label="(4, 12, 64, 32)", style=solid];
+"2120 linalg_vector_norm_40" -> "2121 clamp_min_40"  [label="(4, 12, 64, 1)", style=solid];
+"2121 clamp_min_40" -> "2122 expand_as_40"  [label="(4, 12, 64, 1)", style=solid];
+"2122 expand_as_40" -> "2123 div_40"  [label="(4, 12, 64, 32)", style=solid];
+"2123 div_40" -> "2124 quantize_per_tensor_default_124"  [label="(4, 12, 64, 32)", style=solid];
+"2124 quantize_per_tensor_default_124" -> "2125 dequantize_per_tensor_default_124"  [label="(4, 12, 64, 32)", style=solid];
+"2125 dequantize_per_tensor_default_124" -> "2133 matmul_40"  [label="(4, 12, 64, 32)", style=solid];
+"2126 linalg_vector_norm_41" -> "2127 clamp_min_41"  [label="(4, 12, 64, 1)", style=solid];
+"2127 clamp_min_41" -> "2128 expand_as_41"  [label="(4, 12, 64, 1)", style=solid];
+"2128 expand_as_41" -> "2129 div_41"  [label="(4, 12, 64, 32)", style=solid];
+"2129 div_41" -> "2130 quantize_per_tensor_default_125"  [label="(4, 12, 64, 32)", style=solid];
+"2130 quantize_per_tensor_default_125" -> "2131 dequantize_per_tensor_default_125"  [label="(4, 12, 64, 32)", style=solid];
+"2131 dequantize_per_tensor_default_125" -> "2132 transpose_40"  [label="(4, 12, 64, 32)", style=solid];
+"2132 transpose_40" -> "2133 matmul_40"  [label="(4, 12, 32, 64)", style=solid];
+"2133 matmul_40" -> "2135 mul_41"  [label="(4, 12, 64, 64)", style=solid];
+"2134 _frozen_param173" -> "2135 mul_41"  [label="(12, 1, 1)", style=solid];
+"2135 mul_41" -> "2136 add_70"  [label="(4, 12, 64, 64)", style=solid];
+"2136 add_70" -> "2137 softmax_20"  [label="(4, 12, 64, 64)", style=solid];
+"2137 softmax_20" -> "2138 dropout_80"  [label="(4, 12, 64, 64)", style=solid];
+"2138 dropout_80" -> "2139 matmul_41"  [label="(4, 12, 64, 64)", style=solid];
+"2139 matmul_41" -> "2140 transpose_41"  [label="(4, 12, 64, 32)", style=solid];
+"2140 transpose_41" -> "2141 reshape_92"  [label="(4, 64, 12, 32)", style=solid];
+"2141 reshape_92" -> "2142 reshape_92_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"2142 reshape_92_0_0_nncf_smooth_quant_0" -> "2143 quantize_per_tensor_default_126"  [label="(4, 64, 384)", style=solid];
+"2143 quantize_per_tensor_default_126" -> "2144 dequantize_per_tensor_default_126"  [label="(4, 64, 384)", style=solid];
+"2144 dequantize_per_tensor_default_126" -> "2151 linear_125"  [label="(4, 64, 384)", style=solid];
+"2145 features_5_16_attn_proj_bias_0_0" -> "2151 linear_125"  [label="(384,)", style=solid];
+"2146 scale_updated_constant84" -> "2148 mul_tensor_84"  [label="(384, 1)", style=solid];
+"2147 compressed_weight_updated_constant84" -> "2148 mul_tensor_84"  [label="(384, 384)", style=solid];
+"2148 mul_tensor_84" -> "2150 sub_tensor_84"  [label="(384, 384)", style=solid];
+"2149 zero_point_updated_constant84" -> "2150 sub_tensor_84"  [label="(384, 1)", style=solid];
+"2150 sub_tensor_84" -> "2151 linear_125"  [label="(384, 384)", style=solid];
+"2151 linear_125" -> "2152 dropout_81"  [label="(4, 64, 384)", style=solid];
+"2152 dropout_81" -> "2153 view_113"  [label="(4, 64, 384)", style=solid];
+"2153 view_113" -> "2154 permute_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2154 permute_94" -> "2155 reshape_93"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2155 reshape_93" -> "2156 slice_302"  [label="(1, 16, 16, 384)", style=solid];
+"2156 slice_302" -> "2157 slice_303"  [label="(1, 16, 16, 384)", style=solid];
+"2157 slice_303" -> "2158 slice_304"  [label="(1, 14, 16, 384)", style=solid];
+"2158 slice_304" -> "2159 slice_305"  [label="(1, 14, 14, 384)", style=solid];
+"2159 slice_305" -> "2160 contiguous_39"  [label="(1, 14, 14, 384)", style=solid];
+"2160 contiguous_39" -> "2161 layer_norm_43"  [label="(1, 14, 14, 384)", style=solid];
+"2161 layer_norm_43" -> "2162 add_71"  [label="(1, 14, 14, 384)", style=solid];
+"2162 add_71" -> "2163 add_71_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"2162 add_71" -> "2187 add_72"  [label="(1, 14, 14, 384)", style=solid];
+"2163 add_71_0_0_nncf_smooth_quant_0" -> "2164 quantize_per_tensor_default_127"  [label="(1, 14, 14, 384)", style=solid];
+"2164 quantize_per_tensor_default_127" -> "2165 dequantize_per_tensor_default_127"  [label="(1, 14, 14, 384)", style=solid];
+"2165 dequantize_per_tensor_default_127" -> "2172 linear_126"  [label="(1, 14, 14, 384)", style=solid];
+"2166 features_5_16_mlp_0_bias_0_0" -> "2172 linear_126"  [label="(1536,)", style=solid];
+"2167 scale_updated_constant85" -> "2169 mul_tensor_85"  [label="(1536, 1)", style=solid];
+"2168 compressed_weight_updated_constant85" -> "2169 mul_tensor_85"  [label="(1536, 384)", style=solid];
+"2169 mul_tensor_85" -> "2171 sub_tensor_85"  [label="(1536, 384)", style=solid];
+"2170 zero_point_updated_constant85" -> "2171 sub_tensor_85"  [label="(1536, 1)", style=solid];
+"2171 sub_tensor_85" -> "2172 linear_126"  [label="(1536, 384)", style=solid];
+"2172 linear_126" -> "2173 gelu_20"  [label="(1, 14, 14, 1536)", style=solid];
+"2173 gelu_20" -> "2174 dropout_82"  [label="(1, 14, 14, 1536)", style=solid];
+"2174 dropout_82" -> "2175 dropout_82_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"2175 dropout_82_0_0_nncf_smooth_quant_0" -> "2176 quantize_per_tensor_default_128"  [label="(1, 14, 14, 1536)", style=solid];
+"2176 quantize_per_tensor_default_128" -> "2177 dequantize_per_tensor_default_128"  [label="(1, 14, 14, 1536)", style=solid];
+"2177 dequantize_per_tensor_default_128" -> "2184 linear_127"  [label="(1, 14, 14, 1536)", style=solid];
+"2178 features_5_16_mlp_3_bias_0_0" -> "2184 linear_127"  [label="(384,)", style=solid];
+"2179 scale_updated_constant86" -> "2181 mul_tensor_86"  [label="(384, 1)", style=solid];
+"2180 compressed_weight_updated_constant86" -> "2181 mul_tensor_86"  [label="(384, 1536)", style=solid];
+"2181 mul_tensor_86" -> "2183 sub_tensor_86"  [label="(384, 1536)", style=solid];
+"2182 zero_point_updated_constant86" -> "2183 sub_tensor_86"  [label="(384, 1)", style=solid];
+"2183 sub_tensor_86" -> "2184 linear_127"  [label="(384, 1536)", style=solid];
+"2184 linear_127" -> "2185 dropout_83"  [label="(1, 14, 14, 384)", style=solid];
+"2185 dropout_83" -> "2186 layer_norm_44"  [label="(1, 14, 14, 384)", style=solid];
+"2186 layer_norm_44" -> "2187 add_72"  [label="(1, 14, 14, 384)", style=solid];
+"2187 add_72" -> "2189 pad_23"  [label="(1, 14, 14, 384)", style=solid];
+"2187 add_72" -> "2268 add_75"  [label="(1, 14, 14, 384)", style=solid];
+"2188 _frozen_param174" -> "2225 add_73"  [label="(1, 12, 64, 64)", style=solid];
+"2189 pad_23" -> "2190 roll_20"  [label="(1, 16, 16, 384)", style=solid];
+"2190 roll_20" -> "2191 view_116"  [label="(1, 16, 16, 384)", style=solid];
+"2191 view_116" -> "2192 permute_96"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2192 permute_96" -> "2193 reshape_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2193 reshape_94" -> "2194 reshape_94_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"2193 reshape_94" -> "2226 new_zeros_10"  [label="(4, 64, 384)", style=solid];
+"2194 reshape_94_0_0_nncf_smooth_quant_0" -> "2195 quantize_per_tensor_default_129"  [label="(4, 64, 384)", style=solid];
+"2195 quantize_per_tensor_default_129" -> "2196 dequantize_per_tensor_default_129"  [label="(4, 64, 384)", style=solid];
+"2196 dequantize_per_tensor_default_129" -> "2203 linear_130"  [label="(4, 64, 384)", style=solid];
+"2197 _frozen_param175_0_0" -> "2203 linear_130"  [label="(1152,)", style=solid];
+"2198 scale_updated_constant87" -> "2200 mul_tensor_87"  [label="(1152, 1)", style=solid];
+"2199 compressed_weight_updated_constant87" -> "2200 mul_tensor_87"  [label="(1152, 384)", style=solid];
+"2200 mul_tensor_87" -> "2202 sub_tensor_87"  [label="(1152, 384)", style=solid];
+"2201 zero_point_updated_constant87" -> "2202 sub_tensor_87"  [label="(1152, 1)", style=solid];
+"2202 sub_tensor_87" -> "2203 linear_130"  [label="(1152, 384)", style=solid];
+"2203 linear_130" -> "2204 reshape_95"  [label="(4, 64, 1152)", style=solid];
+"2204 reshape_95" -> "2205 permute_97"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2205 permute_97" -> "2206 select_63"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2205 permute_97" -> "2207 select_64"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2205 permute_97" -> "2208 select_65"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2206 select_63" -> "2209 linalg_vector_norm_42"  [label="(4, 12, 64, 32)", style=solid];
+"2206 select_63" -> "2211 expand_as_42"  [label="(4, 12, 64, 32)", style=solid];
+"2206 select_63" -> "2212 div_42"  [label="(4, 12, 64, 32)", style=solid];
+"2207 select_64" -> "2215 linalg_vector_norm_43"  [label="(4, 12, 64, 32)", style=solid];
+"2207 select_64" -> "2217 expand_as_43"  [label="(4, 12, 64, 32)", style=solid];
+"2207 select_64" -> "2218 div_43"  [label="(4, 12, 64, 32)", style=solid];
+"2208 select_65" -> "2244 matmul_43"  [label="(4, 12, 64, 32)", style=solid];
+"2209 linalg_vector_norm_42" -> "2210 clamp_min_42"  [label="(4, 12, 64, 1)", style=solid];
+"2210 clamp_min_42" -> "2211 expand_as_42"  [label="(4, 12, 64, 1)", style=solid];
+"2211 expand_as_42" -> "2212 div_42"  [label="(4, 12, 64, 32)", style=solid];
+"2212 div_42" -> "2213 quantize_per_tensor_default_130"  [label="(4, 12, 64, 32)", style=solid];
+"2213 quantize_per_tensor_default_130" -> "2214 dequantize_per_tensor_default_130"  [label="(4, 12, 64, 32)", style=solid];
+"2214 dequantize_per_tensor_default_130" -> "2222 matmul_42"  [label="(4, 12, 64, 32)", style=solid];
+"2215 linalg_vector_norm_43" -> "2216 clamp_min_43"  [label="(4, 12, 64, 1)", style=solid];
+"2216 clamp_min_43" -> "2217 expand_as_43"  [label="(4, 12, 64, 1)", style=solid];
+"2217 expand_as_43" -> "2218 div_43"  [label="(4, 12, 64, 32)", style=solid];
+"2218 div_43" -> "2219 quantize_per_tensor_default_131"  [label="(4, 12, 64, 32)", style=solid];
+"2219 quantize_per_tensor_default_131" -> "2220 dequantize_per_tensor_default_131"  [label="(4, 12, 64, 32)", style=solid];
+"2220 dequantize_per_tensor_default_131" -> "2221 transpose_42"  [label="(4, 12, 64, 32)", style=solid];
+"2221 transpose_42" -> "2222 matmul_42"  [label="(4, 12, 32, 64)", style=solid];
+"2222 matmul_42" -> "2224 mul_43"  [label="(4, 12, 64, 64)", style=solid];
+"2223 _frozen_param177" -> "2224 mul_43"  [label="(12, 1, 1)", style=solid];
+"2224 mul_43" -> "2225 add_73"  [label="(4, 12, 64, 64)", style=solid];
+"2225 add_73" -> "2237 view_118"  [label="(4, 12, 64, 64)", style=solid];
+"2226 new_zeros_10" -> "2227 view_117"  [label="(16, 16)", style=solid];
+"2227 view_117" -> "2228 permute_98"  [label="(2, 8, 2, 8)", style=solid];
+"2228 permute_98" -> "2229 reshape_96"  [label="(2, 2, 8, 8)", style=solid];
+"2229 reshape_96" -> "2230 unsqueeze_62"  [label="(4, 64)", style=solid];
+"2229 reshape_96" -> "2231 unsqueeze_63"  [label="(4, 64)", style=solid];
+"2230 unsqueeze_62" -> "2232 sub_10"  [label="(4, 1, 64)", style=solid];
+"2231 unsqueeze_63" -> "2232 sub_10"  [label="(4, 64, 1)", style=solid];
+"2232 sub_10" -> "2233 ne_10"  [label="(4, 64, 64)", style=solid];
+"2232 sub_10" -> "2234 masked_fill_20"  [label="(4, 64, 64)", style=solid];
+"2232 sub_10" -> "2235 eq_10"  [label="(4, 64, 64)", style=solid];
+"2233 ne_10" -> "2234 masked_fill_20"  [label="(4, 64, 64)", style=solid];
+"2234 masked_fill_20" -> "2236 masked_fill_21"  [label="(4, 64, 64)", style=solid];
+"2235 eq_10" -> "2236 masked_fill_21"  [label="(4, 64, 64)", style=solid];
+"2236 masked_fill_21" -> "2238 unsqueeze_64"  [label="(4, 64, 64)", style=solid];
+"2237 view_118" -> "2240 add_74"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2238 unsqueeze_64" -> "2239 unsqueeze_65"  [label="(4, 1, 64, 64)", style=solid];
+"2239 unsqueeze_65" -> "2240 add_74"  [label="(1, 4, 1, 64, 64)", style=solid];
+"2240 add_74" -> "2241 view_119"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2241 view_119" -> "2242 softmax_21"  [label="(4, 12, 64, 64)", style=solid];
+"2242 softmax_21" -> "2243 dropout_84"  [label="(4, 12, 64, 64)", style=solid];
+"2243 dropout_84" -> "2244 matmul_43"  [label="(4, 12, 64, 64)", style=solid];
+"2244 matmul_43" -> "2245 transpose_43"  [label="(4, 12, 64, 32)", style=solid];
+"2245 transpose_43" -> "2246 reshape_97"  [label="(4, 64, 12, 32)", style=solid];
+"2246 reshape_97" -> "2247 reshape_97_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"2247 reshape_97_0_0_nncf_smooth_quant_0" -> "2248 quantize_per_tensor_default_132"  [label="(4, 64, 384)", style=solid];
+"2248 quantize_per_tensor_default_132" -> "2249 dequantize_per_tensor_default_132"  [label="(4, 64, 384)", style=solid];
+"2249 dequantize_per_tensor_default_132" -> "2256 linear_131"  [label="(4, 64, 384)", style=solid];
+"2250 features_5_17_attn_proj_bias_0_0" -> "2256 linear_131"  [label="(384,)", style=solid];
+"2251 scale_updated_constant88" -> "2253 mul_tensor_88"  [label="(384, 1)", style=solid];
+"2252 compressed_weight_updated_constant88" -> "2253 mul_tensor_88"  [label="(384, 384)", style=solid];
+"2253 mul_tensor_88" -> "2255 sub_tensor_88"  [label="(384, 384)", style=solid];
+"2254 zero_point_updated_constant88" -> "2255 sub_tensor_88"  [label="(384, 1)", style=solid];
+"2255 sub_tensor_88" -> "2256 linear_131"  [label="(384, 384)", style=solid];
+"2256 linear_131" -> "2257 dropout_85"  [label="(4, 64, 384)", style=solid];
+"2257 dropout_85" -> "2258 view_120"  [label="(4, 64, 384)", style=solid];
+"2258 view_120" -> "2259 permute_99"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2259 permute_99" -> "2260 reshape_98"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2260 reshape_98" -> "2261 roll_21"  [label="(1, 16, 16, 384)", style=solid];
+"2261 roll_21" -> "2262 slice_325"  [label="(1, 16, 16, 384)", style=solid];
+"2262 slice_325" -> "2263 slice_326"  [label="(1, 16, 16, 384)", style=solid];
+"2263 slice_326" -> "2264 slice_327"  [label="(1, 14, 16, 384)", style=solid];
+"2264 slice_327" -> "2265 slice_328"  [label="(1, 14, 14, 384)", style=solid];
+"2265 slice_328" -> "2266 contiguous_41"  [label="(1, 14, 14, 384)", style=solid];
+"2266 contiguous_41" -> "2267 layer_norm_45"  [label="(1, 14, 14, 384)", style=solid];
+"2267 layer_norm_45" -> "2268 add_75"  [label="(1, 14, 14, 384)", style=solid];
+"2268 add_75" -> "2269 add_75_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"2268 add_75" -> "2293 add_76"  [label="(1, 14, 14, 384)", style=solid];
+"2269 add_75_0_0_nncf_smooth_quant_0" -> "2270 quantize_per_tensor_default_133"  [label="(1, 14, 14, 384)", style=solid];
+"2270 quantize_per_tensor_default_133" -> "2271 dequantize_per_tensor_default_133"  [label="(1, 14, 14, 384)", style=solid];
+"2271 dequantize_per_tensor_default_133" -> "2278 linear_132"  [label="(1, 14, 14, 384)", style=solid];
+"2272 features_5_17_mlp_0_bias_0_0" -> "2278 linear_132"  [label="(1536,)", style=solid];
+"2273 scale_updated_constant89" -> "2275 mul_tensor_89"  [label="(1536, 1)", style=solid];
+"2274 compressed_weight_updated_constant89" -> "2275 mul_tensor_89"  [label="(1536, 384)", style=solid];
+"2275 mul_tensor_89" -> "2277 sub_tensor_89"  [label="(1536, 384)", style=solid];
+"2276 zero_point_updated_constant89" -> "2277 sub_tensor_89"  [label="(1536, 1)", style=solid];
+"2277 sub_tensor_89" -> "2278 linear_132"  [label="(1536, 384)", style=solid];
+"2278 linear_132" -> "2279 gelu_21"  [label="(1, 14, 14, 1536)", style=solid];
+"2279 gelu_21" -> "2280 dropout_86"  [label="(1, 14, 14, 1536)", style=solid];
+"2280 dropout_86" -> "2281 dropout_86_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"2281 dropout_86_0_0_nncf_smooth_quant_0" -> "2282 quantize_per_tensor_default_134"  [label="(1, 14, 14, 1536)", style=solid];
+"2282 quantize_per_tensor_default_134" -> "2283 dequantize_per_tensor_default_134"  [label="(1, 14, 14, 1536)", style=solid];
+"2283 dequantize_per_tensor_default_134" -> "2290 linear_133"  [label="(1, 14, 14, 1536)", style=solid];
+"2284 features_5_17_mlp_3_bias_0_0" -> "2290 linear_133"  [label="(384,)", style=solid];
+"2285 scale_updated_constant90" -> "2287 mul_tensor_90"  [label="(384, 1)", style=solid];
+"2286 compressed_weight_updated_constant90" -> "2287 mul_tensor_90"  [label="(384, 1536)", style=solid];
+"2287 mul_tensor_90" -> "2289 sub_tensor_90"  [label="(384, 1536)", style=solid];
+"2288 zero_point_updated_constant90" -> "2289 sub_tensor_90"  [label="(384, 1)", style=solid];
+"2289 sub_tensor_90" -> "2290 linear_133"  [label="(384, 1536)", style=solid];
+"2290 linear_133" -> "2291 dropout_87"  [label="(1, 14, 14, 384)", style=solid];
+"2291 dropout_87" -> "2292 layer_norm_46"  [label="(1, 14, 14, 384)", style=solid];
+"2292 layer_norm_46" -> "2293 add_76"  [label="(1, 14, 14, 384)", style=solid];
+"2293 add_76" -> "2294 pad_24"  [label="(1, 14, 14, 384)", style=solid];
+"2294 pad_24" -> "2295 slice_329"  [label="(1, 14, 14, 384)", style=solid];
+"2294 pad_24" -> "2298 slice_332"  [label="(1, 14, 14, 384)", style=solid];
+"2294 pad_24" -> "2301 slice_335"  [label="(1, 14, 14, 384)", style=solid];
+"2294 pad_24" -> "2304 slice_338"  [label="(1, 14, 14, 384)", style=solid];
+"2295 slice_329" -> "2296 slice_330"  [label="(1, 7, 14, 384)", style=solid];
+"2296 slice_330" -> "2297 slice_331"  [label="(1, 7, 7, 384)", style=solid];
+"2297 slice_331" -> "2307 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2298 slice_332" -> "2299 slice_333"  [label="(1, 7, 14, 384)", style=solid];
+"2299 slice_333" -> "2300 slice_334"  [label="(1, 7, 7, 384)", style=solid];
+"2300 slice_334" -> "2307 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2301 slice_335" -> "2302 slice_336"  [label="(1, 7, 14, 384)", style=solid];
+"2302 slice_336" -> "2303 slice_337"  [label="(1, 7, 7, 384)", style=solid];
+"2303 slice_337" -> "2307 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2304 slice_338" -> "2305 slice_339"  [label="(1, 7, 14, 384)", style=solid];
+"2305 slice_339" -> "2306 slice_340"  [label="(1, 7, 7, 384)", style=solid];
+"2306 slice_340" -> "2307 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2307 cat_2" -> "2308 cat_2_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 1536)", style=solid];
+"2308 cat_2_0_0_nncf_smooth_quant_0" -> "2309 quantize_per_tensor_default_135"  [label="(1, 7, 7, 1536)", style=solid];
+"2309 quantize_per_tensor_default_135" -> "2310 dequantize_per_tensor_default_135"  [label="(1, 7, 7, 1536)", style=solid];
+"2310 dequantize_per_tensor_default_135" -> "2316 linear_134"  [label="(1, 7, 7, 1536)", style=solid];
+"2311 scale_updated_constant91" -> "2313 mul_tensor_91"  [label="(768, 1)", style=solid];
+"2312 compressed_weight_updated_constant91" -> "2313 mul_tensor_91"  [label="(768, 1536)", style=solid];
+"2313 mul_tensor_91" -> "2315 sub_tensor_91"  [label="(768, 1536)", style=solid];
+"2314 zero_point_updated_constant91" -> "2315 sub_tensor_91"  [label="(768, 1)", style=solid];
+"2315 sub_tensor_91" -> "2316 linear_134"  [label="(768, 1536)", style=solid];
+"2316 linear_134" -> "2317 layer_norm_47"  [label="(1, 7, 7, 768)", style=solid];
+"2317 layer_norm_47" -> "2319 pad_25"  [label="(1, 7, 7, 768)", style=solid];
+"2317 layer_norm_47" -> "2380 add_78"  [label="(1, 7, 7, 768)", style=solid];
+"2318 _frozen_param187" -> "2354 add_77"  [label="(1, 24, 64, 64)", style=solid];
+"2319 pad_25" -> "2320 view_123"  [label="(1, 8, 8, 768)", style=solid];
+"2320 view_123" -> "2321 permute_101"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2321 permute_101" -> "2322 reshape_99"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2322 reshape_99" -> "2323 reshape_99_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
+"2323 reshape_99_0_0_nncf_smooth_quant_0" -> "2324 quantize_per_tensor_default_136"  [label="(1, 64, 768)", style=solid];
+"2324 quantize_per_tensor_default_136" -> "2325 dequantize_per_tensor_default_136"  [label="(1, 64, 768)", style=solid];
+"2325 dequantize_per_tensor_default_136" -> "2332 linear_137"  [label="(1, 64, 768)", style=solid];
+"2326 _frozen_param188_0_0" -> "2332 linear_137"  [label="(2304,)", style=solid];
+"2327 scale_updated_constant92" -> "2329 mul_tensor_92"  [label="(2304, 1)", style=solid];
+"2328 compressed_weight_updated_constant92" -> "2329 mul_tensor_92"  [label="(2304, 768)", style=solid];
+"2329 mul_tensor_92" -> "2331 sub_tensor_92"  [label="(2304, 768)", style=solid];
+"2330 zero_point_updated_constant92" -> "2331 sub_tensor_92"  [label="(2304, 1)", style=solid];
+"2331 sub_tensor_92" -> "2332 linear_137"  [label="(2304, 768)", style=solid];
+"2332 linear_137" -> "2333 reshape_100"  [label="(1, 64, 2304)", style=solid];
+"2333 reshape_100" -> "2334 permute_102"  [label="(1, 64, 3, 24, 32)", style=solid];
+"2334 permute_102" -> "2335 select_66"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2334 permute_102" -> "2336 select_67"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2334 permute_102" -> "2337 select_68"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2335 select_66" -> "2338 linalg_vector_norm_44"  [label="(1, 24, 64, 32)", style=solid];
+"2335 select_66" -> "2340 expand_as_44"  [label="(1, 24, 64, 32)", style=solid];
+"2335 select_66" -> "2341 div_44"  [label="(1, 24, 64, 32)", style=solid];
+"2336 select_67" -> "2344 linalg_vector_norm_45"  [label="(1, 24, 64, 32)", style=solid];
+"2336 select_67" -> "2346 expand_as_45"  [label="(1, 24, 64, 32)", style=solid];
+"2336 select_67" -> "2347 div_45"  [label="(1, 24, 64, 32)", style=solid];
+"2337 select_68" -> "2357 matmul_45"  [label="(1, 24, 64, 32)", style=solid];
+"2338 linalg_vector_norm_44" -> "2339 clamp_min_44"  [label="(1, 24, 64, 1)", style=solid];
+"2339 clamp_min_44" -> "2340 expand_as_44"  [label="(1, 24, 64, 1)", style=solid];
+"2340 expand_as_44" -> "2341 div_44"  [label="(1, 24, 64, 32)", style=solid];
+"2341 div_44" -> "2342 quantize_per_tensor_default_137"  [label="(1, 24, 64, 32)", style=solid];
+"2342 quantize_per_tensor_default_137" -> "2343 dequantize_per_tensor_default_137"  [label="(1, 24, 64, 32)", style=solid];
+"2343 dequantize_per_tensor_default_137" -> "2351 matmul_44"  [label="(1, 24, 64, 32)", style=solid];
+"2344 linalg_vector_norm_45" -> "2345 clamp_min_45"  [label="(1, 24, 64, 1)", style=solid];
+"2345 clamp_min_45" -> "2346 expand_as_45"  [label="(1, 24, 64, 1)", style=solid];
+"2346 expand_as_45" -> "2347 div_45"  [label="(1, 24, 64, 32)", style=solid];
+"2347 div_45" -> "2348 quantize_per_tensor_default_138"  [label="(1, 24, 64, 32)", style=solid];
+"2348 quantize_per_tensor_default_138" -> "2349 dequantize_per_tensor_default_138"  [label="(1, 24, 64, 32)", style=solid];
+"2349 dequantize_per_tensor_default_138" -> "2350 transpose_44"  [label="(1, 24, 64, 32)", style=solid];
+"2350 transpose_44" -> "2351 matmul_44"  [label="(1, 24, 32, 64)", style=solid];
+"2351 matmul_44" -> "2353 mul_45"  [label="(1, 24, 64, 64)", style=solid];
+"2352 _frozen_param190" -> "2353 mul_45"  [label="(24, 1, 1)", style=solid];
+"2353 mul_45" -> "2354 add_77"  [label="(1, 24, 64, 64)", style=solid];
+"2354 add_77" -> "2355 softmax_22"  [label="(1, 24, 64, 64)", style=solid];
+"2355 softmax_22" -> "2356 dropout_88"  [label="(1, 24, 64, 64)", style=solid];
+"2356 dropout_88" -> "2357 matmul_45"  [label="(1, 24, 64, 64)", style=solid];
+"2357 matmul_45" -> "2358 transpose_45"  [label="(1, 24, 64, 32)", style=solid];
+"2358 transpose_45" -> "2359 reshape_101"  [label="(1, 64, 24, 32)", style=solid];
+"2359 reshape_101" -> "2360 reshape_101_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
+"2360 reshape_101_0_0_nncf_smooth_quant_0" -> "2361 quantize_per_tensor_default_139"  [label="(1, 64, 768)", style=solid];
+"2361 quantize_per_tensor_default_139" -> "2362 dequantize_per_tensor_default_139"  [label="(1, 64, 768)", style=solid];
+"2362 dequantize_per_tensor_default_139" -> "2369 linear_138"  [label="(1, 64, 768)", style=solid];
+"2363 features_7_0_attn_proj_bias_0_0" -> "2369 linear_138"  [label="(768,)", style=solid];
+"2364 scale_updated_constant93" -> "2366 mul_tensor_93"  [label="(768, 1)", style=solid];
+"2365 compressed_weight_updated_constant93" -> "2366 mul_tensor_93"  [label="(768, 768)", style=solid];
+"2366 mul_tensor_93" -> "2368 sub_tensor_93"  [label="(768, 768)", style=solid];
+"2367 zero_point_updated_constant93" -> "2368 sub_tensor_93"  [label="(768, 1)", style=solid];
+"2368 sub_tensor_93" -> "2369 linear_138"  [label="(768, 768)", style=solid];
+"2369 linear_138" -> "2370 dropout_89"  [label="(1, 64, 768)", style=solid];
+"2370 dropout_89" -> "2371 view_124"  [label="(1, 64, 768)", style=solid];
+"2371 view_124" -> "2372 permute_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2372 permute_103" -> "2373 reshape_102"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2373 reshape_102" -> "2374 slice_342"  [label="(1, 8, 8, 768)", style=solid];
+"2374 slice_342" -> "2375 slice_343"  [label="(1, 8, 8, 768)", style=solid];
+"2375 slice_343" -> "2376 slice_344"  [label="(1, 7, 8, 768)", style=solid];
+"2376 slice_344" -> "2377 slice_345"  [label="(1, 7, 7, 768)", style=solid];
+"2377 slice_345" -> "2378 contiguous_43"  [label="(1, 7, 7, 768)", style=solid];
+"2378 contiguous_43" -> "2379 layer_norm_48"  [label="(1, 7, 7, 768)", style=solid];
+"2379 layer_norm_48" -> "2380 add_78"  [label="(1, 7, 7, 768)", style=solid];
+"2380 add_78" -> "2381 add_78_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 768)", style=solid];
+"2380 add_78" -> "2405 add_79"  [label="(1, 7, 7, 768)", style=solid];
+"2381 add_78_0_0_nncf_smooth_quant_0" -> "2382 quantize_per_tensor_default_140"  [label="(1, 7, 7, 768)", style=solid];
+"2382 quantize_per_tensor_default_140" -> "2383 dequantize_per_tensor_default_140"  [label="(1, 7, 7, 768)", style=solid];
+"2383 dequantize_per_tensor_default_140" -> "2390 linear_139"  [label="(1, 7, 7, 768)", style=solid];
+"2384 features_7_0_mlp_0_bias_0_0" -> "2390 linear_139"  [label="(3072,)", style=solid];
+"2385 scale_updated_constant94" -> "2387 mul_tensor_94"  [label="(3072, 1)", style=solid];
+"2386 compressed_weight_updated_constant94" -> "2387 mul_tensor_94"  [label="(3072, 768)", style=solid];
+"2387 mul_tensor_94" -> "2389 sub_tensor_94"  [label="(3072, 768)", style=solid];
+"2388 zero_point_updated_constant94" -> "2389 sub_tensor_94"  [label="(3072, 1)", style=solid];
+"2389 sub_tensor_94" -> "2390 linear_139"  [label="(3072, 768)", style=solid];
+"2390 linear_139" -> "2391 gelu_22"  [label="(1, 7, 7, 3072)", style=solid];
+"2391 gelu_22" -> "2392 dropout_90"  [label="(1, 7, 7, 3072)", style=solid];
+"2392 dropout_90" -> "2393 dropout_90_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 3072)", style=solid];
+"2393 dropout_90_0_0_nncf_smooth_quant_0" -> "2394 quantize_per_tensor_default_141"  [label="(1, 7, 7, 3072)", style=solid];
+"2394 quantize_per_tensor_default_141" -> "2395 dequantize_per_tensor_default_141"  [label="(1, 7, 7, 3072)", style=solid];
+"2395 dequantize_per_tensor_default_141" -> "2402 linear_140"  [label="(1, 7, 7, 3072)", style=solid];
+"2396 features_7_0_mlp_3_bias_0_0" -> "2402 linear_140"  [label="(768,)", style=solid];
+"2397 scale_updated_constant95" -> "2399 mul_tensor_95"  [label="(768, 1)", style=solid];
+"2398 compressed_weight_updated_constant95" -> "2399 mul_tensor_95"  [label="(768, 3072)", style=solid];
+"2399 mul_tensor_95" -> "2401 sub_tensor_95"  [label="(768, 3072)", style=solid];
+"2400 zero_point_updated_constant95" -> "2401 sub_tensor_95"  [label="(768, 1)", style=solid];
+"2401 sub_tensor_95" -> "2402 linear_140"  [label="(768, 3072)", style=solid];
+"2402 linear_140" -> "2403 dropout_91"  [label="(1, 7, 7, 768)", style=solid];
+"2403 dropout_91" -> "2404 layer_norm_49"  [label="(1, 7, 7, 768)", style=solid];
+"2404 layer_norm_49" -> "2405 add_79"  [label="(1, 7, 7, 768)", style=solid];
+"2405 add_79" -> "2407 pad_26"  [label="(1, 7, 7, 768)", style=solid];
+"2405 add_79" -> "2468 add_81"  [label="(1, 7, 7, 768)", style=solid];
+"2406 _frozen_param191" -> "2442 add_80"  [label="(1, 24, 64, 64)", style=solid];
+"2407 pad_26" -> "2408 view_127"  [label="(1, 8, 8, 768)", style=solid];
+"2408 view_127" -> "2409 permute_105"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2409 permute_105" -> "2410 reshape_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2410 reshape_103" -> "2411 reshape_103_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
+"2411 reshape_103_0_0_nncf_smooth_quant_0" -> "2412 quantize_per_tensor_default_142"  [label="(1, 64, 768)", style=solid];
+"2412 quantize_per_tensor_default_142" -> "2413 dequantize_per_tensor_default_142"  [label="(1, 64, 768)", style=solid];
+"2413 dequantize_per_tensor_default_142" -> "2420 linear_143"  [label="(1, 64, 768)", style=solid];
+"2414 _frozen_param192_0_0" -> "2420 linear_143"  [label="(2304,)", style=solid];
+"2415 scale_updated_constant96" -> "2417 mul_tensor_96"  [label="(2304, 1)", style=solid];
+"2416 compressed_weight_updated_constant96" -> "2417 mul_tensor_96"  [label="(2304, 768)", style=solid];
+"2417 mul_tensor_96" -> "2419 sub_tensor_96"  [label="(2304, 768)", style=solid];
+"2418 zero_point_updated_constant96" -> "2419 sub_tensor_96"  [label="(2304, 1)", style=solid];
+"2419 sub_tensor_96" -> "2420 linear_143"  [label="(2304, 768)", style=solid];
+"2420 linear_143" -> "2421 reshape_104"  [label="(1, 64, 2304)", style=solid];
+"2421 reshape_104" -> "2422 permute_106"  [label="(1, 64, 3, 24, 32)", style=solid];
+"2422 permute_106" -> "2423 select_69"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2422 permute_106" -> "2424 select_70"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2422 permute_106" -> "2425 select_71"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2423 select_69" -> "2426 linalg_vector_norm_46"  [label="(1, 24, 64, 32)", style=solid];
+"2423 select_69" -> "2428 expand_as_46"  [label="(1, 24, 64, 32)", style=solid];
+"2423 select_69" -> "2429 div_46"  [label="(1, 24, 64, 32)", style=solid];
+"2424 select_70" -> "2432 linalg_vector_norm_47"  [label="(1, 24, 64, 32)", style=solid];
+"2424 select_70" -> "2434 expand_as_47"  [label="(1, 24, 64, 32)", style=solid];
+"2424 select_70" -> "2435 div_47"  [label="(1, 24, 64, 32)", style=solid];
+"2425 select_71" -> "2445 matmul_47"  [label="(1, 24, 64, 32)", style=solid];
+"2426 linalg_vector_norm_46" -> "2427 clamp_min_46"  [label="(1, 24, 64, 1)", style=solid];
+"2427 clamp_min_46" -> "2428 expand_as_46"  [label="(1, 24, 64, 1)", style=solid];
+"2428 expand_as_46" -> "2429 div_46"  [label="(1, 24, 64, 32)", style=solid];
+"2429 div_46" -> "2430 quantize_per_tensor_default_143"  [label="(1, 24, 64, 32)", style=solid];
+"2430 quantize_per_tensor_default_143" -> "2431 dequantize_per_tensor_default_143"  [label="(1, 24, 64, 32)", style=solid];
+"2431 dequantize_per_tensor_default_143" -> "2439 matmul_46"  [label="(1, 24, 64, 32)", style=solid];
+"2432 linalg_vector_norm_47" -> "2433 clamp_min_47"  [label="(1, 24, 64, 1)", style=solid];
+"2433 clamp_min_47" -> "2434 expand_as_47"  [label="(1, 24, 64, 1)", style=solid];
+"2434 expand_as_47" -> "2435 div_47"  [label="(1, 24, 64, 32)", style=solid];
+"2435 div_47" -> "2436 quantize_per_tensor_default_144"  [label="(1, 24, 64, 32)", style=solid];
+"2436 quantize_per_tensor_default_144" -> "2437 dequantize_per_tensor_default_144"  [label="(1, 24, 64, 32)", style=solid];
+"2437 dequantize_per_tensor_default_144" -> "2438 transpose_46"  [label="(1, 24, 64, 32)", style=solid];
+"2438 transpose_46" -> "2439 matmul_46"  [label="(1, 24, 32, 64)", style=solid];
+"2439 matmul_46" -> "2441 mul_47"  [label="(1, 24, 64, 64)", style=solid];
+"2440 _frozen_param194" -> "2441 mul_47"  [label="(24, 1, 1)", style=solid];
+"2441 mul_47" -> "2442 add_80"  [label="(1, 24, 64, 64)", style=solid];
+"2442 add_80" -> "2443 softmax_23"  [label="(1, 24, 64, 64)", style=solid];
+"2443 softmax_23" -> "2444 dropout_92"  [label="(1, 24, 64, 64)", style=solid];
+"2444 dropout_92" -> "2445 matmul_47"  [label="(1, 24, 64, 64)", style=solid];
+"2445 matmul_47" -> "2446 transpose_47"  [label="(1, 24, 64, 32)", style=solid];
+"2446 transpose_47" -> "2447 reshape_105"  [label="(1, 64, 24, 32)", style=solid];
+"2447 reshape_105" -> "2448 reshape_105_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
+"2448 reshape_105_0_0_nncf_smooth_quant_0" -> "2449 quantize_per_tensor_default_145"  [label="(1, 64, 768)", style=solid];
+"2449 quantize_per_tensor_default_145" -> "2450 dequantize_per_tensor_default_145"  [label="(1, 64, 768)", style=solid];
+"2450 dequantize_per_tensor_default_145" -> "2457 linear_144"  [label="(1, 64, 768)", style=solid];
+"2451 features_7_1_attn_proj_bias_0_0" -> "2457 linear_144"  [label="(768,)", style=solid];
+"2452 scale_updated_constant97" -> "2454 mul_tensor_97"  [label="(768, 1)", style=solid];
+"2453 compressed_weight_updated_constant97" -> "2454 mul_tensor_97"  [label="(768, 768)", style=solid];
+"2454 mul_tensor_97" -> "2456 sub_tensor_97"  [label="(768, 768)", style=solid];
+"2455 zero_point_updated_constant97" -> "2456 sub_tensor_97"  [label="(768, 1)", style=solid];
+"2456 sub_tensor_97" -> "2457 linear_144"  [label="(768, 768)", style=solid];
+"2457 linear_144" -> "2458 dropout_93"  [label="(1, 64, 768)", style=solid];
+"2458 dropout_93" -> "2459 view_128"  [label="(1, 64, 768)", style=solid];
+"2459 view_128" -> "2460 permute_107"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2460 permute_107" -> "2461 reshape_106"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2461 reshape_106" -> "2462 slice_347"  [label="(1, 8, 8, 768)", style=solid];
+"2462 slice_347" -> "2463 slice_348"  [label="(1, 8, 8, 768)", style=solid];
+"2463 slice_348" -> "2464 slice_349"  [label="(1, 7, 8, 768)", style=solid];
+"2464 slice_349" -> "2465 slice_350"  [label="(1, 7, 7, 768)", style=solid];
+"2465 slice_350" -> "2466 contiguous_45"  [label="(1, 7, 7, 768)", style=solid];
+"2466 contiguous_45" -> "2467 layer_norm_50"  [label="(1, 7, 7, 768)", style=solid];
+"2467 layer_norm_50" -> "2468 add_81"  [label="(1, 7, 7, 768)", style=solid];
+"2468 add_81" -> "2469 add_81_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 768)", style=solid];
+"2468 add_81" -> "2493 add_82"  [label="(1, 7, 7, 768)", style=solid];
+"2469 add_81_0_0_nncf_smooth_quant_0" -> "2470 quantize_per_tensor_default_146"  [label="(1, 7, 7, 768)", style=solid];
+"2470 quantize_per_tensor_default_146" -> "2471 dequantize_per_tensor_default_146"  [label="(1, 7, 7, 768)", style=solid];
+"2471 dequantize_per_tensor_default_146" -> "2478 linear_145"  [label="(1, 7, 7, 768)", style=solid];
+"2472 features_7_1_mlp_0_bias_0_0" -> "2478 linear_145"  [label="(3072,)", style=solid];
+"2473 scale_updated_constant98" -> "2475 mul_tensor_98"  [label="(3072, 1)", style=solid];
+"2474 compressed_weight_updated_constant98" -> "2475 mul_tensor_98"  [label="(3072, 768)", style=solid];
+"2475 mul_tensor_98" -> "2477 sub_tensor_98"  [label="(3072, 768)", style=solid];
+"2476 zero_point_updated_constant98" -> "2477 sub_tensor_98"  [label="(3072, 1)", style=solid];
+"2477 sub_tensor_98" -> "2478 linear_145"  [label="(3072, 768)", style=solid];
+"2478 linear_145" -> "2479 gelu_23"  [label="(1, 7, 7, 3072)", style=solid];
+"2479 gelu_23" -> "2480 dropout_94"  [label="(1, 7, 7, 3072)", style=solid];
+"2480 dropout_94" -> "2481 dropout_94_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 3072)", style=solid];
+"2481 dropout_94_0_0_nncf_smooth_quant_0" -> "2482 quantize_per_tensor_default_147"  [label="(1, 7, 7, 3072)", style=solid];
+"2482 quantize_per_tensor_default_147" -> "2483 dequantize_per_tensor_default_147"  [label="(1, 7, 7, 3072)", style=solid];
+"2483 dequantize_per_tensor_default_147" -> "2490 linear_146"  [label="(1, 7, 7, 3072)", style=solid];
+"2484 features_7_1_mlp_3_bias_0_0" -> "2490 linear_146"  [label="(768,)", style=solid];
+"2485 scale_updated_constant99" -> "2487 mul_tensor_99"  [label="(768, 1)", style=solid];
+"2486 compressed_weight_updated_constant99" -> "2487 mul_tensor_99"  [label="(768, 3072)", style=solid];
+"2487 mul_tensor_99" -> "2489 sub_tensor_99"  [label="(768, 3072)", style=solid];
+"2488 zero_point_updated_constant99" -> "2489 sub_tensor_99"  [label="(768, 1)", style=solid];
+"2489 sub_tensor_99" -> "2490 linear_146"  [label="(768, 3072)", style=solid];
+"2490 linear_146" -> "2491 dropout_95"  [label="(1, 7, 7, 768)", style=solid];
+"2491 dropout_95" -> "2492 layer_norm_51"  [label="(1, 7, 7, 768)", style=solid];
+"2492 layer_norm_51" -> "2493 add_82"  [label="(1, 7, 7, 768)", style=solid];
+"2493 add_82" -> "2494 layer_norm_52"  [label="(1, 7, 7, 768)", style=solid];
 "2494 layer_norm_52" -> "2495 permute_108"  [label="(1, 7, 7, 768)", style=solid];
 "2495 permute_108" -> "2496 adaptive_avg_pool2d"  [label="(1, 768, 7, 7)", style=solid];
 "2496 adaptive_avg_pool2d" -> "2497 flatten"  [label="(1, 768, 1, 1)", style=solid];
@@ -5254,7 +5254,7 @@ strict digraph  {
 "2498 flatten_0_0_nncf_smooth_quant_0" -> "2499 quantize_per_tensor_default_148"  [label="(1, 768)", style=solid];
 "2499 quantize_per_tensor_default_148" -> "2500 dequantize_per_tensor_default_148"  [label="(1, 768)", style=solid];
 "2500 dequantize_per_tensor_default_148" -> "2507 linear_147"  [label="(1, 768)", style=solid];
-"2501 _param_constant400_0_0" -> "2507 linear_147"  [label="(1000,)", style=solid];
+"2501 head_bias_0_0" -> "2507 linear_147"  [label="(1000,)", style=solid];
 "2502 scale_updated_constant100" -> "2504 mul_tensor_100"  [label="(1000, 1)", style=solid];
 "2503 compressed_weight_updated_constant100" -> "2504 mul_tensor_100"  [label="(1000, 768)", style=solid];
 "2504 mul_tensor_100" -> "2506 sub_tensor_100"  [label="(1000, 768)", style=solid];
diff --git a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/synthetic_transformer.dot b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/synthetic_transformer.dot
index c7cfb557bc6..44026591fb1 100644
--- a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/synthetic_transformer.dot
+++ b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/synthetic_transformer.dot
@@ -1,11 +1,11 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
+"0 wte_weight" [id=0, type=get_attr];
+"1 input_ids" [id=1, type=input];
 "2 embedding" [id=2, type=embedding];
 "3 embedding_0_0_nncf_smooth_quant_0" [id=3, type=call_module];
 "4 quantize_per_tensor_default" [id=4, type=quantize_per_tensor];
 "5 dequantize_per_tensor_default" [id=5, type=dequantize_per_tensor];
-"6 _param_constant2_0_0" [id=6, type=get_attr];
+"6 linear_bias_0_0" [id=6, type=get_attr];
 "7 scale_updated_constant0" [id=7, type=get_attr];
 "8 compressed_weight_updated_constant0" [id=8, type=get_attr];
 "9 mul_tensor" [id=9, type=mul];
@@ -15,7 +15,7 @@ strict digraph  {
 "13 add_tensor_0_0_nncf_smooth_quant_0" [id=13, type=call_module];
 "14 quantize_per_tensor_default_1" [id=14, type=quantize_per_tensor];
 "15 dequantize_per_tensor_default_1" [id=15, type=dequantize_per_tensor];
-"16 _param_constant4_0_0" [id=16, type=get_attr];
+"16 lm_head_bias_0_0" [id=16, type=get_attr];
 "17 scale_updated_constant1" [id=17, type=get_attr];
 "18 compressed_weight_updated_constant1" [id=18, type=get_attr];
 "19 mul_tensor_1" [id=19, type=mul];
@@ -23,13 +23,13 @@ strict digraph  {
 "21 sub_tensor_1" [id=21, type=sub];
 "22 linear_1" [id=22, type=linear];
 "23 output" [id=23, type=output];
-"0 arg0_1" -> "2 embedding"  [label="(5,)", style=solid];
-"1 _param_constant0" -> "2 embedding"  [label="(10, 5)", style=solid];
+"0 wte_weight" -> "2 embedding"  [label="(10, 5)", style=solid];
+"1 input_ids" -> "2 embedding"  [label="(5,)", style=solid];
 "2 embedding" -> "3 embedding_0_0_nncf_smooth_quant_0"  [label="(5, 5)", style=solid];
 "3 embedding_0_0_nncf_smooth_quant_0" -> "4 quantize_per_tensor_default"  [label="(5, 5)", style=solid];
 "4 quantize_per_tensor_default" -> "5 dequantize_per_tensor_default"  [label="(5, 5)", style=solid];
 "5 dequantize_per_tensor_default" -> "12 linear"  [label="(5, 5)", style=solid];
-"6 _param_constant2_0_0" -> "12 linear"  [label="(5,)", style=solid];
+"6 linear_bias_0_0" -> "12 linear"  [label="(5,)", style=solid];
 "7 scale_updated_constant0" -> "9 mul_tensor"  [label="(5, 1)", style=solid];
 "8 compressed_weight_updated_constant0" -> "9 mul_tensor"  [label="(5, 5)", style=solid];
 "9 mul_tensor" -> "11 sub_tensor"  [label="(5, 5)", style=solid];
@@ -39,7 +39,7 @@ strict digraph  {
 "13 add_tensor_0_0_nncf_smooth_quant_0" -> "14 quantize_per_tensor_default_1"  [label="(5, 5)", style=solid];
 "14 quantize_per_tensor_default_1" -> "15 dequantize_per_tensor_default_1"  [label="(5, 5)", style=solid];
 "15 dequantize_per_tensor_default_1" -> "22 linear_1"  [label="(5, 5)", style=solid];
-"16 _param_constant4_0_0" -> "22 linear_1"  [label="(10,)", style=solid];
+"16 lm_head_bias_0_0" -> "22 linear_1"  [label="(10,)", style=solid];
 "17 scale_updated_constant1" -> "19 mul_tensor_1"  [label="(10, 1)", style=solid];
 "18 compressed_weight_updated_constant1" -> "19 mul_tensor_1"  [label="(10, 5)", style=solid];
 "19 mul_tensor_1" -> "21 sub_tensor_1"  [label="(10, 5)", style=solid];
diff --git a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/unet.dot b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/unet.dot
index 32bfa0b2e6f..a183dc3d499 100644
--- a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/unet.dot
+++ b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/unet.dot
@@ -1,8 +1,8 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
+"0 x" [id=0, type=input];
 "1 quantize_per_tensor_default_8" [id=1, type=quantize_per_tensor];
 "2 dequantize_per_tensor_default_12" [id=2, type=dequantize_per_tensor];
-"3 _param_constant1_0_0" [id=3, type=get_attr];
+"3 down_path_0_block_0_bias_0_0" [id=3, type=get_attr];
 "4 scale_updated_constant0" [id=4, type=get_attr];
 "5 compressed_weight_updated_constant0" [id=5, type=get_attr];
 "6 mul_tensor" [id=6, type=mul];
@@ -12,7 +12,7 @@ strict digraph  {
 "10 relu" [id=10, type=relu];
 "11 quantize_per_tensor_default_9" [id=11, type=quantize_per_tensor];
 "12 dequantize_per_tensor_default_13" [id=12, type=dequantize_per_tensor];
-"13 _param_constant5_0_0" [id=13, type=get_attr];
+"13 down_path_0_block_3_bias_0_0" [id=13, type=get_attr];
 "14 scale_updated_constant1" [id=14, type=get_attr];
 "15 compressed_weight_updated_constant1" [id=15, type=get_attr];
 "16 mul_tensor_1" [id=16, type=mul];
@@ -24,7 +24,7 @@ strict digraph  {
 "22 dequantize_per_tensor_default_1" [id=22, type=dequantize_per_tensor];
 "23 dequantize_per_tensor_default" [id=23, type=dequantize_per_tensor];
 "24 max_pool2d" [id=24, type=max_pool2d];
-"25 _param_constant9_0_0" [id=25, type=get_attr];
+"25 down_path_1_block_0_bias_0_0" [id=25, type=get_attr];
 "26 scale_updated_constant2" [id=26, type=get_attr];
 "27 compressed_weight_updated_constant2" [id=27, type=get_attr];
 "28 mul_tensor_2" [id=28, type=mul];
@@ -34,7 +34,7 @@ strict digraph  {
 "32 relu_2" [id=32, type=relu];
 "33 quantize_per_tensor_default_10" [id=33, type=quantize_per_tensor];
 "34 dequantize_per_tensor_default_14" [id=34, type=dequantize_per_tensor];
-"35 _param_constant13_0_0" [id=35, type=get_attr];
+"35 down_path_1_block_3_bias_0_0" [id=35, type=get_attr];
 "36 scale_updated_constant3" [id=36, type=get_attr];
 "37 compressed_weight_updated_constant3" [id=37, type=get_attr];
 "38 mul_tensor_3" [id=38, type=mul];
@@ -46,7 +46,7 @@ strict digraph  {
 "44 dequantize_per_tensor_default_5" [id=44, type=dequantize_per_tensor];
 "45 dequantize_per_tensor_default_4" [id=45, type=dequantize_per_tensor];
 "46 max_pool2d_1" [id=46, type=max_pool2d];
-"47 _param_constant17_0_0" [id=47, type=get_attr];
+"47 down_path_2_block_0_bias_0_0" [id=47, type=get_attr];
 "48 scale_updated_constant4" [id=48, type=get_attr];
 "49 compressed_weight_updated_constant4" [id=49, type=get_attr];
 "50 mul_tensor_4" [id=50, type=mul];
@@ -56,7 +56,7 @@ strict digraph  {
 "54 relu_4" [id=54, type=relu];
 "55 quantize_per_tensor_default_11" [id=55, type=quantize_per_tensor];
 "56 dequantize_per_tensor_default_15" [id=56, type=dequantize_per_tensor];
-"57 _param_constant21_0_0" [id=57, type=get_attr];
+"57 down_path_2_block_3_bias_0_0" [id=57, type=get_attr];
 "58 scale_updated_constant5" [id=58, type=get_attr];
 "59 compressed_weight_updated_constant5" [id=59, type=get_attr];
 "60 mul_tensor_5" [id=60, type=mul];
@@ -68,7 +68,7 @@ strict digraph  {
 "66 dequantize_per_tensor_default_7" [id=66, type=dequantize_per_tensor];
 "67 dequantize_per_tensor_default_6" [id=67, type=dequantize_per_tensor];
 "68 max_pool2d_2" [id=68, type=max_pool2d];
-"69 _param_constant25_0_0" [id=69, type=get_attr];
+"69 down_path_3_block_0_bias_0_0" [id=69, type=get_attr];
 "70 scale_updated_constant6" [id=70, type=get_attr];
 "71 compressed_weight_updated_constant6" [id=71, type=get_attr];
 "72 mul_tensor_6" [id=72, type=mul];
@@ -78,7 +78,7 @@ strict digraph  {
 "76 relu_6" [id=76, type=relu];
 "77 quantize_per_tensor_default_12" [id=77, type=quantize_per_tensor];
 "78 dequantize_per_tensor_default_16" [id=78, type=dequantize_per_tensor];
-"79 _param_constant29_0_0" [id=79, type=get_attr];
+"79 down_path_3_block_3_bias_0_0" [id=79, type=get_attr];
 "80 scale_updated_constant7" [id=80, type=get_attr];
 "81 compressed_weight_updated_constant7" [id=81, type=get_attr];
 "82 mul_tensor_7" [id=82, type=mul];
@@ -90,7 +90,7 @@ strict digraph  {
 "88 dequantize_per_tensor_default_11" [id=88, type=dequantize_per_tensor];
 "89 dequantize_per_tensor_default_10" [id=89, type=dequantize_per_tensor];
 "90 max_pool2d_3" [id=90, type=max_pool2d];
-"91 _param_constant33_0_0" [id=91, type=get_attr];
+"91 down_path_4_block_0_bias_0_0" [id=91, type=get_attr];
 "92 scale_updated_constant8" [id=92, type=get_attr];
 "93 compressed_weight_updated_constant8" [id=93, type=get_attr];
 "94 mul_tensor_8" [id=94, type=mul];
@@ -100,7 +100,7 @@ strict digraph  {
 "98 relu_8" [id=98, type=relu];
 "99 quantize_per_tensor_default_13" [id=99, type=quantize_per_tensor];
 "100 dequantize_per_tensor_default_17" [id=100, type=dequantize_per_tensor];
-"101 _param_constant37_0_0" [id=101, type=get_attr];
+"101 down_path_4_block_3_bias_0_0" [id=101, type=get_attr];
 "102 scale_updated_constant9" [id=102, type=get_attr];
 "103 compressed_weight_updated_constant9" [id=103, type=get_attr];
 "104 mul_tensor_9" [id=104, type=mul];
@@ -110,7 +110,7 @@ strict digraph  {
 "108 relu_9" [id=108, type=relu];
 "109 quantize_per_tensor_default_14" [id=109, type=quantize_per_tensor];
 "110 dequantize_per_tensor_default_18" [id=110, type=dequantize_per_tensor];
-"111 _param_constant41_0_0" [id=111, type=get_attr];
+"111 up_path_0_up_bias_0_0" [id=111, type=get_attr];
 "112 scale_updated_constant10" [id=112, type=get_attr];
 "113 compressed_weight_updated_constant10" [id=113, type=get_attr];
 "114 mul_tensor_10" [id=114, type=mul];
@@ -124,7 +124,7 @@ strict digraph  {
 "122 slice_3" [id=122, type=slice];
 "123 slice_4" [id=123, type=slice];
 "124 cat" [id=124, type=cat];
-"125 _param_constant43_0_0" [id=125, type=get_attr];
+"125 up_path_0_conv_block_block_0_bias_0_0" [id=125, type=get_attr];
 "126 scale_updated_constant11" [id=126, type=get_attr];
 "127 compressed_weight_updated_constant11" [id=127, type=get_attr];
 "128 mul_tensor_11" [id=128, type=mul];
@@ -134,7 +134,7 @@ strict digraph  {
 "132 relu_10" [id=132, type=relu];
 "133 quantize_per_tensor_default_15" [id=133, type=quantize_per_tensor];
 "134 dequantize_per_tensor_default_19" [id=134, type=dequantize_per_tensor];
-"135 _param_constant47_0_0" [id=135, type=get_attr];
+"135 up_path_0_conv_block_block_3_bias_0_0" [id=135, type=get_attr];
 "136 scale_updated_constant12" [id=136, type=get_attr];
 "137 compressed_weight_updated_constant12" [id=137, type=get_attr];
 "138 mul_tensor_12" [id=138, type=mul];
@@ -144,7 +144,7 @@ strict digraph  {
 "142 relu_11" [id=142, type=relu];
 "143 quantize_per_tensor_default_16" [id=143, type=quantize_per_tensor];
 "144 dequantize_per_tensor_default_20" [id=144, type=dequantize_per_tensor];
-"145 _param_constant51_0_0" [id=145, type=get_attr];
+"145 up_path_1_up_bias_0_0" [id=145, type=get_attr];
 "146 scale_updated_constant13" [id=146, type=get_attr];
 "147 compressed_weight_updated_constant13" [id=147, type=get_attr];
 "148 mul_tensor_13" [id=148, type=mul];
@@ -158,7 +158,7 @@ strict digraph  {
 "156 slice_7" [id=156, type=slice];
 "157 slice_8" [id=157, type=slice];
 "158 cat_1" [id=158, type=cat];
-"159 _param_constant53_0_0" [id=159, type=get_attr];
+"159 up_path_1_conv_block_block_0_bias_0_0" [id=159, type=get_attr];
 "160 scale_updated_constant14" [id=160, type=get_attr];
 "161 compressed_weight_updated_constant14" [id=161, type=get_attr];
 "162 mul_tensor_14" [id=162, type=mul];
@@ -168,7 +168,7 @@ strict digraph  {
 "166 relu_12" [id=166, type=relu];
 "167 quantize_per_tensor_default_17" [id=167, type=quantize_per_tensor];
 "168 dequantize_per_tensor_default_21" [id=168, type=dequantize_per_tensor];
-"169 _param_constant57_0_0" [id=169, type=get_attr];
+"169 up_path_1_conv_block_block_3_bias_0_0" [id=169, type=get_attr];
 "170 scale_updated_constant15" [id=170, type=get_attr];
 "171 compressed_weight_updated_constant15" [id=171, type=get_attr];
 "172 mul_tensor_15" [id=172, type=mul];
@@ -178,7 +178,7 @@ strict digraph  {
 "176 relu_13" [id=176, type=relu];
 "177 quantize_per_tensor_default_18" [id=177, type=quantize_per_tensor];
 "178 dequantize_per_tensor_default_22" [id=178, type=dequantize_per_tensor];
-"179 _param_constant61_0_0" [id=179, type=get_attr];
+"179 up_path_2_up_bias_0_0" [id=179, type=get_attr];
 "180 scale_updated_constant16" [id=180, type=get_attr];
 "181 compressed_weight_updated_constant16" [id=181, type=get_attr];
 "182 mul_tensor_16" [id=182, type=mul];
@@ -192,7 +192,7 @@ strict digraph  {
 "190 slice_11" [id=190, type=slice];
 "191 slice_12" [id=191, type=slice];
 "192 cat_2" [id=192, type=cat];
-"193 _param_constant63_0_0" [id=193, type=get_attr];
+"193 up_path_2_conv_block_block_0_bias_0_0" [id=193, type=get_attr];
 "194 scale_updated_constant17" [id=194, type=get_attr];
 "195 compressed_weight_updated_constant17" [id=195, type=get_attr];
 "196 mul_tensor_17" [id=196, type=mul];
@@ -202,7 +202,7 @@ strict digraph  {
 "200 relu_14" [id=200, type=relu];
 "201 quantize_per_tensor_default_19" [id=201, type=quantize_per_tensor];
 "202 dequantize_per_tensor_default_23" [id=202, type=dequantize_per_tensor];
-"203 _param_constant67_0_0" [id=203, type=get_attr];
+"203 up_path_2_conv_block_block_3_bias_0_0" [id=203, type=get_attr];
 "204 scale_updated_constant18" [id=204, type=get_attr];
 "205 compressed_weight_updated_constant18" [id=205, type=get_attr];
 "206 mul_tensor_18" [id=206, type=mul];
@@ -212,7 +212,7 @@ strict digraph  {
 "210 relu_15" [id=210, type=relu];
 "211 quantize_per_tensor_default_20" [id=211, type=quantize_per_tensor];
 "212 dequantize_per_tensor_default_24" [id=212, type=dequantize_per_tensor];
-"213 _param_constant71_0_0" [id=213, type=get_attr];
+"213 up_path_3_up_bias_0_0" [id=213, type=get_attr];
 "214 scale_updated_constant19" [id=214, type=get_attr];
 "215 compressed_weight_updated_constant19" [id=215, type=get_attr];
 "216 mul_tensor_19" [id=216, type=mul];
@@ -226,7 +226,7 @@ strict digraph  {
 "224 slice_15" [id=224, type=slice];
 "225 slice_16" [id=225, type=slice];
 "226 cat_3" [id=226, type=cat];
-"227 _param_constant73_0_0" [id=227, type=get_attr];
+"227 up_path_3_conv_block_block_0_bias_0_0" [id=227, type=get_attr];
 "228 scale_updated_constant20" [id=228, type=get_attr];
 "229 compressed_weight_updated_constant20" [id=229, type=get_attr];
 "230 mul_tensor_20" [id=230, type=mul];
@@ -236,7 +236,7 @@ strict digraph  {
 "234 relu_16" [id=234, type=relu];
 "235 quantize_per_tensor_default_21" [id=235, type=quantize_per_tensor];
 "236 dequantize_per_tensor_default_25" [id=236, type=dequantize_per_tensor];
-"237 _param_constant77_0_0" [id=237, type=get_attr];
+"237 up_path_3_conv_block_block_3_bias_0_0" [id=237, type=get_attr];
 "238 scale_updated_constant21" [id=238, type=get_attr];
 "239 compressed_weight_updated_constant21" [id=239, type=get_attr];
 "240 mul_tensor_21" [id=240, type=mul];
@@ -246,7 +246,7 @@ strict digraph  {
 "244 relu_17" [id=244, type=relu];
 "245 quantize_per_tensor_default_22" [id=245, type=quantize_per_tensor];
 "246 dequantize_per_tensor_default_26" [id=246, type=dequantize_per_tensor];
-"247 _param_constant81_0_0" [id=247, type=get_attr];
+"247 last_bias_0_0" [id=247, type=get_attr];
 "248 scale_updated_constant22" [id=248, type=get_attr];
 "249 compressed_weight_updated_constant22" [id=249, type=get_attr];
 "250 mul_tensor_22" [id=250, type=mul];
@@ -254,10 +254,10 @@ strict digraph  {
 "252 sub_tensor_22" [id=252, type=sub];
 "253 conv2d_18" [id=253, type=conv2d];
 "254 output" [id=254, type=output];
-"0 arg0_1" -> "1 quantize_per_tensor_default_8"  [label="(1, 3, 224, 224)", style=solid];
+"0 x" -> "1 quantize_per_tensor_default_8"  [label="(1, 3, 224, 224)", style=solid];
 "1 quantize_per_tensor_default_8" -> "2 dequantize_per_tensor_default_12"  [label="(1, 3, 224, 224)", style=solid];
 "2 dequantize_per_tensor_default_12" -> "9 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"3 _param_constant1_0_0" -> "9 conv2d"  [label="(64,)", style=solid];
+"3 down_path_0_block_0_bias_0_0" -> "9 conv2d"  [label="(64,)", style=solid];
 "4 scale_updated_constant0" -> "6 mul_tensor"  [label="(64, 1, 1, 1)", style=solid];
 "5 compressed_weight_updated_constant0" -> "6 mul_tensor"  [label="(64, 3, 3, 3)", style=solid];
 "6 mul_tensor" -> "8 sub_tensor"  [label="(64, 3, 3, 3)", style=solid];
@@ -267,7 +267,7 @@ strict digraph  {
 "10 relu" -> "11 quantize_per_tensor_default_9"  [label="(1, 64, 222, 222)", style=solid];
 "11 quantize_per_tensor_default_9" -> "12 dequantize_per_tensor_default_13"  [label="(1, 64, 222, 222)", style=solid];
 "12 dequantize_per_tensor_default_13" -> "19 conv2d_1"  [label="(1, 64, 222, 222)", style=solid];
-"13 _param_constant5_0_0" -> "19 conv2d_1"  [label="(64,)", style=solid];
+"13 down_path_0_block_3_bias_0_0" -> "19 conv2d_1"  [label="(64,)", style=solid];
 "14 scale_updated_constant1" -> "16 mul_tensor_1"  [label="(64, 1, 1, 1)", style=solid];
 "15 compressed_weight_updated_constant1" -> "16 mul_tensor_1"  [label="(64, 64, 3, 3)", style=solid];
 "16 mul_tensor_1" -> "18 sub_tensor_1"  [label="(64, 64, 3, 3)", style=solid];
@@ -280,7 +280,7 @@ strict digraph  {
 "22 dequantize_per_tensor_default_1" -> "222 slice_13"  [label="(1, 64, 220, 220)", style=solid];
 "23 dequantize_per_tensor_default" -> "24 max_pool2d"  [label="(1, 64, 220, 220)", style=solid];
 "24 max_pool2d" -> "31 conv2d_2"  [label="(1, 64, 110, 110)", style=solid];
-"25 _param_constant9_0_0" -> "31 conv2d_2"  [label="(128,)", style=solid];
+"25 down_path_1_block_0_bias_0_0" -> "31 conv2d_2"  [label="(128,)", style=solid];
 "26 scale_updated_constant2" -> "28 mul_tensor_2"  [label="(128, 1, 1, 1)", style=solid];
 "27 compressed_weight_updated_constant2" -> "28 mul_tensor_2"  [label="(128, 64, 3, 3)", style=solid];
 "28 mul_tensor_2" -> "30 sub_tensor_2"  [label="(128, 64, 3, 3)", style=solid];
@@ -290,7 +290,7 @@ strict digraph  {
 "32 relu_2" -> "33 quantize_per_tensor_default_10"  [label="(1, 128, 108, 108)", style=solid];
 "33 quantize_per_tensor_default_10" -> "34 dequantize_per_tensor_default_14"  [label="(1, 128, 108, 108)", style=solid];
 "34 dequantize_per_tensor_default_14" -> "41 conv2d_3"  [label="(1, 128, 108, 108)", style=solid];
-"35 _param_constant13_0_0" -> "41 conv2d_3"  [label="(128,)", style=solid];
+"35 down_path_1_block_3_bias_0_0" -> "41 conv2d_3"  [label="(128,)", style=solid];
 "36 scale_updated_constant3" -> "38 mul_tensor_3"  [label="(128, 1, 1, 1)", style=solid];
 "37 compressed_weight_updated_constant3" -> "38 mul_tensor_3"  [label="(128, 128, 3, 3)", style=solid];
 "38 mul_tensor_3" -> "40 sub_tensor_3"  [label="(128, 128, 3, 3)", style=solid];
@@ -303,7 +303,7 @@ strict digraph  {
 "44 dequantize_per_tensor_default_5" -> "188 slice_9"  [label="(1, 128, 106, 106)", style=solid];
 "45 dequantize_per_tensor_default_4" -> "46 max_pool2d_1"  [label="(1, 128, 106, 106)", style=solid];
 "46 max_pool2d_1" -> "53 conv2d_4"  [label="(1, 128, 53, 53)", style=solid];
-"47 _param_constant17_0_0" -> "53 conv2d_4"  [label="(256,)", style=solid];
+"47 down_path_2_block_0_bias_0_0" -> "53 conv2d_4"  [label="(256,)", style=solid];
 "48 scale_updated_constant4" -> "50 mul_tensor_4"  [label="(256, 1, 1, 1)", style=solid];
 "49 compressed_weight_updated_constant4" -> "50 mul_tensor_4"  [label="(256, 128, 3, 3)", style=solid];
 "50 mul_tensor_4" -> "52 sub_tensor_4"  [label="(256, 128, 3, 3)", style=solid];
@@ -313,7 +313,7 @@ strict digraph  {
 "54 relu_4" -> "55 quantize_per_tensor_default_11"  [label="(1, 256, 51, 51)", style=solid];
 "55 quantize_per_tensor_default_11" -> "56 dequantize_per_tensor_default_15"  [label="(1, 256, 51, 51)", style=solid];
 "56 dequantize_per_tensor_default_15" -> "63 conv2d_5"  [label="(1, 256, 51, 51)", style=solid];
-"57 _param_constant21_0_0" -> "63 conv2d_5"  [label="(256,)", style=solid];
+"57 down_path_2_block_3_bias_0_0" -> "63 conv2d_5"  [label="(256,)", style=solid];
 "58 scale_updated_constant5" -> "60 mul_tensor_5"  [label="(256, 1, 1, 1)", style=solid];
 "59 compressed_weight_updated_constant5" -> "60 mul_tensor_5"  [label="(256, 256, 3, 3)", style=solid];
 "60 mul_tensor_5" -> "62 sub_tensor_5"  [label="(256, 256, 3, 3)", style=solid];
@@ -326,7 +326,7 @@ strict digraph  {
 "66 dequantize_per_tensor_default_7" -> "154 slice_5"  [label="(1, 256, 49, 49)", style=solid];
 "67 dequantize_per_tensor_default_6" -> "68 max_pool2d_2"  [label="(1, 256, 49, 49)", style=solid];
 "68 max_pool2d_2" -> "75 conv2d_6"  [label="(1, 256, 24, 24)", style=solid];
-"69 _param_constant25_0_0" -> "75 conv2d_6"  [label="(512,)", style=solid];
+"69 down_path_3_block_0_bias_0_0" -> "75 conv2d_6"  [label="(512,)", style=solid];
 "70 scale_updated_constant6" -> "72 mul_tensor_6"  [label="(512, 1, 1, 1)", style=solid];
 "71 compressed_weight_updated_constant6" -> "72 mul_tensor_6"  [label="(512, 256, 3, 3)", style=solid];
 "72 mul_tensor_6" -> "74 sub_tensor_6"  [label="(512, 256, 3, 3)", style=solid];
@@ -336,7 +336,7 @@ strict digraph  {
 "76 relu_6" -> "77 quantize_per_tensor_default_12"  [label="(1, 512, 22, 22)", style=solid];
 "77 quantize_per_tensor_default_12" -> "78 dequantize_per_tensor_default_16"  [label="(1, 512, 22, 22)", style=solid];
 "78 dequantize_per_tensor_default_16" -> "85 conv2d_7"  [label="(1, 512, 22, 22)", style=solid];
-"79 _param_constant29_0_0" -> "85 conv2d_7"  [label="(512,)", style=solid];
+"79 down_path_3_block_3_bias_0_0" -> "85 conv2d_7"  [label="(512,)", style=solid];
 "80 scale_updated_constant7" -> "82 mul_tensor_7"  [label="(512, 1, 1, 1)", style=solid];
 "81 compressed_weight_updated_constant7" -> "82 mul_tensor_7"  [label="(512, 512, 3, 3)", style=solid];
 "82 mul_tensor_7" -> "84 sub_tensor_7"  [label="(512, 512, 3, 3)", style=solid];
@@ -349,7 +349,7 @@ strict digraph  {
 "88 dequantize_per_tensor_default_11" -> "120 slice_1"  [label="(1, 512, 20, 20)", style=solid];
 "89 dequantize_per_tensor_default_10" -> "90 max_pool2d_3"  [label="(1, 512, 20, 20)", style=solid];
 "90 max_pool2d_3" -> "97 conv2d_8"  [label="(1, 512, 10, 10)", style=solid];
-"91 _param_constant33_0_0" -> "97 conv2d_8"  [label="(1024,)", style=solid];
+"91 down_path_4_block_0_bias_0_0" -> "97 conv2d_8"  [label="(1024,)", style=solid];
 "92 scale_updated_constant8" -> "94 mul_tensor_8"  [label="(1024, 1, 1, 1)", style=solid];
 "93 compressed_weight_updated_constant8" -> "94 mul_tensor_8"  [label="(1024, 512, 3, 3)", style=solid];
 "94 mul_tensor_8" -> "96 sub_tensor_8"  [label="(1024, 512, 3, 3)", style=solid];
@@ -359,7 +359,7 @@ strict digraph  {
 "98 relu_8" -> "99 quantize_per_tensor_default_13"  [label="(1, 1024, 8, 8)", style=solid];
 "99 quantize_per_tensor_default_13" -> "100 dequantize_per_tensor_default_17"  [label="(1, 1024, 8, 8)", style=solid];
 "100 dequantize_per_tensor_default_17" -> "107 conv2d_9"  [label="(1, 1024, 8, 8)", style=solid];
-"101 _param_constant37_0_0" -> "107 conv2d_9"  [label="(1024,)", style=solid];
+"101 down_path_4_block_3_bias_0_0" -> "107 conv2d_9"  [label="(1024,)", style=solid];
 "102 scale_updated_constant9" -> "104 mul_tensor_9"  [label="(1024, 1, 1, 1)", style=solid];
 "103 compressed_weight_updated_constant9" -> "104 mul_tensor_9"  [label="(1024, 1024, 3, 3)", style=solid];
 "104 mul_tensor_9" -> "106 sub_tensor_9"  [label="(1024, 1024, 3, 3)", style=solid];
@@ -369,7 +369,7 @@ strict digraph  {
 "108 relu_9" -> "109 quantize_per_tensor_default_14"  [label="(1, 1024, 6, 6)", style=solid];
 "109 quantize_per_tensor_default_14" -> "110 dequantize_per_tensor_default_18"  [label="(1, 1024, 6, 6)", style=solid];
 "110 dequantize_per_tensor_default_18" -> "117 conv_transpose2d"  [label="(1, 1024, 6, 6)", style=solid];
-"111 _param_constant41_0_0" -> "117 conv_transpose2d"  [label="(512,)", style=solid];
+"111 up_path_0_up_bias_0_0" -> "117 conv_transpose2d"  [label="(512,)", style=solid];
 "112 scale_updated_constant10" -> "114 mul_tensor_10"  [label="(1024, 1, 1, 1)", style=solid];
 "113 compressed_weight_updated_constant10" -> "114 mul_tensor_10"  [label="(1024, 512, 2, 2)", style=solid];
 "114 mul_tensor_10" -> "116 sub_tensor_10"  [label="(1024, 512, 2, 2)", style=solid];
@@ -383,7 +383,7 @@ strict digraph  {
 "122 slice_3" -> "123 slice_4"  [label="(1, 512, 12, 20)", style=solid];
 "123 slice_4" -> "124 cat"  [label="(1, 512, 12, 12)", style=solid];
 "124 cat" -> "131 conv2d_10"  [label="(1, 1024, 12, 12)", style=solid];
-"125 _param_constant43_0_0" -> "131 conv2d_10"  [label="(512,)", style=solid];
+"125 up_path_0_conv_block_block_0_bias_0_0" -> "131 conv2d_10"  [label="(512,)", style=solid];
 "126 scale_updated_constant11" -> "128 mul_tensor_11"  [label="(512, 1, 1, 1)", style=solid];
 "127 compressed_weight_updated_constant11" -> "128 mul_tensor_11"  [label="(512, 1024, 3, 3)", style=solid];
 "128 mul_tensor_11" -> "130 sub_tensor_11"  [label="(512, 1024, 3, 3)", style=solid];
@@ -393,7 +393,7 @@ strict digraph  {
 "132 relu_10" -> "133 quantize_per_tensor_default_15"  [label="(1, 512, 10, 10)", style=solid];
 "133 quantize_per_tensor_default_15" -> "134 dequantize_per_tensor_default_19"  [label="(1, 512, 10, 10)", style=solid];
 "134 dequantize_per_tensor_default_19" -> "141 conv2d_11"  [label="(1, 512, 10, 10)", style=solid];
-"135 _param_constant47_0_0" -> "141 conv2d_11"  [label="(512,)", style=solid];
+"135 up_path_0_conv_block_block_3_bias_0_0" -> "141 conv2d_11"  [label="(512,)", style=solid];
 "136 scale_updated_constant12" -> "138 mul_tensor_12"  [label="(512, 1, 1, 1)", style=solid];
 "137 compressed_weight_updated_constant12" -> "138 mul_tensor_12"  [label="(512, 512, 3, 3)", style=solid];
 "138 mul_tensor_12" -> "140 sub_tensor_12"  [label="(512, 512, 3, 3)", style=solid];
@@ -403,7 +403,7 @@ strict digraph  {
 "142 relu_11" -> "143 quantize_per_tensor_default_16"  [label="(1, 512, 8, 8)", style=solid];
 "143 quantize_per_tensor_default_16" -> "144 dequantize_per_tensor_default_20"  [label="(1, 512, 8, 8)", style=solid];
 "144 dequantize_per_tensor_default_20" -> "151 conv_transpose2d_1"  [label="(1, 512, 8, 8)", style=solid];
-"145 _param_constant51_0_0" -> "151 conv_transpose2d_1"  [label="(256,)", style=solid];
+"145 up_path_1_up_bias_0_0" -> "151 conv_transpose2d_1"  [label="(256,)", style=solid];
 "146 scale_updated_constant13" -> "148 mul_tensor_13"  [label="(512, 1, 1, 1)", style=solid];
 "147 compressed_weight_updated_constant13" -> "148 mul_tensor_13"  [label="(512, 256, 2, 2)", style=solid];
 "148 mul_tensor_13" -> "150 sub_tensor_13"  [label="(512, 256, 2, 2)", style=solid];
@@ -417,7 +417,7 @@ strict digraph  {
 "156 slice_7" -> "157 slice_8"  [label="(1, 256, 16, 49)", style=solid];
 "157 slice_8" -> "158 cat_1"  [label="(1, 256, 16, 16)", style=solid];
 "158 cat_1" -> "165 conv2d_12"  [label="(1, 512, 16, 16)", style=solid];
-"159 _param_constant53_0_0" -> "165 conv2d_12"  [label="(256,)", style=solid];
+"159 up_path_1_conv_block_block_0_bias_0_0" -> "165 conv2d_12"  [label="(256,)", style=solid];
 "160 scale_updated_constant14" -> "162 mul_tensor_14"  [label="(256, 1, 1, 1)", style=solid];
 "161 compressed_weight_updated_constant14" -> "162 mul_tensor_14"  [label="(256, 512, 3, 3)", style=solid];
 "162 mul_tensor_14" -> "164 sub_tensor_14"  [label="(256, 512, 3, 3)", style=solid];
@@ -427,7 +427,7 @@ strict digraph  {
 "166 relu_12" -> "167 quantize_per_tensor_default_17"  [label="(1, 256, 14, 14)", style=solid];
 "167 quantize_per_tensor_default_17" -> "168 dequantize_per_tensor_default_21"  [label="(1, 256, 14, 14)", style=solid];
 "168 dequantize_per_tensor_default_21" -> "175 conv2d_13"  [label="(1, 256, 14, 14)", style=solid];
-"169 _param_constant57_0_0" -> "175 conv2d_13"  [label="(256,)", style=solid];
+"169 up_path_1_conv_block_block_3_bias_0_0" -> "175 conv2d_13"  [label="(256,)", style=solid];
 "170 scale_updated_constant15" -> "172 mul_tensor_15"  [label="(256, 1, 1, 1)", style=solid];
 "171 compressed_weight_updated_constant15" -> "172 mul_tensor_15"  [label="(256, 256, 3, 3)", style=solid];
 "172 mul_tensor_15" -> "174 sub_tensor_15"  [label="(256, 256, 3, 3)", style=solid];
@@ -437,7 +437,7 @@ strict digraph  {
 "176 relu_13" -> "177 quantize_per_tensor_default_18"  [label="(1, 256, 12, 12)", style=solid];
 "177 quantize_per_tensor_default_18" -> "178 dequantize_per_tensor_default_22"  [label="(1, 256, 12, 12)", style=solid];
 "178 dequantize_per_tensor_default_22" -> "185 conv_transpose2d_2"  [label="(1, 256, 12, 12)", style=solid];
-"179 _param_constant61_0_0" -> "185 conv_transpose2d_2"  [label="(128,)", style=solid];
+"179 up_path_2_up_bias_0_0" -> "185 conv_transpose2d_2"  [label="(128,)", style=solid];
 "180 scale_updated_constant16" -> "182 mul_tensor_16"  [label="(256, 1, 1, 1)", style=solid];
 "181 compressed_weight_updated_constant16" -> "182 mul_tensor_16"  [label="(256, 128, 2, 2)", style=solid];
 "182 mul_tensor_16" -> "184 sub_tensor_16"  [label="(256, 128, 2, 2)", style=solid];
@@ -451,7 +451,7 @@ strict digraph  {
 "190 slice_11" -> "191 slice_12"  [label="(1, 128, 24, 106)", style=solid];
 "191 slice_12" -> "192 cat_2"  [label="(1, 128, 24, 24)", style=solid];
 "192 cat_2" -> "199 conv2d_14"  [label="(1, 256, 24, 24)", style=solid];
-"193 _param_constant63_0_0" -> "199 conv2d_14"  [label="(128,)", style=solid];
+"193 up_path_2_conv_block_block_0_bias_0_0" -> "199 conv2d_14"  [label="(128,)", style=solid];
 "194 scale_updated_constant17" -> "196 mul_tensor_17"  [label="(128, 1, 1, 1)", style=solid];
 "195 compressed_weight_updated_constant17" -> "196 mul_tensor_17"  [label="(128, 256, 3, 3)", style=solid];
 "196 mul_tensor_17" -> "198 sub_tensor_17"  [label="(128, 256, 3, 3)", style=solid];
@@ -461,7 +461,7 @@ strict digraph  {
 "200 relu_14" -> "201 quantize_per_tensor_default_19"  [label="(1, 128, 22, 22)", style=solid];
 "201 quantize_per_tensor_default_19" -> "202 dequantize_per_tensor_default_23"  [label="(1, 128, 22, 22)", style=solid];
 "202 dequantize_per_tensor_default_23" -> "209 conv2d_15"  [label="(1, 128, 22, 22)", style=solid];
-"203 _param_constant67_0_0" -> "209 conv2d_15"  [label="(128,)", style=solid];
+"203 up_path_2_conv_block_block_3_bias_0_0" -> "209 conv2d_15"  [label="(128,)", style=solid];
 "204 scale_updated_constant18" -> "206 mul_tensor_18"  [label="(128, 1, 1, 1)", style=solid];
 "205 compressed_weight_updated_constant18" -> "206 mul_tensor_18"  [label="(128, 128, 3, 3)", style=solid];
 "206 mul_tensor_18" -> "208 sub_tensor_18"  [label="(128, 128, 3, 3)", style=solid];
@@ -471,7 +471,7 @@ strict digraph  {
 "210 relu_15" -> "211 quantize_per_tensor_default_20"  [label="(1, 128, 20, 20)", style=solid];
 "211 quantize_per_tensor_default_20" -> "212 dequantize_per_tensor_default_24"  [label="(1, 128, 20, 20)", style=solid];
 "212 dequantize_per_tensor_default_24" -> "219 conv_transpose2d_3"  [label="(1, 128, 20, 20)", style=solid];
-"213 _param_constant71_0_0" -> "219 conv_transpose2d_3"  [label="(64,)", style=solid];
+"213 up_path_3_up_bias_0_0" -> "219 conv_transpose2d_3"  [label="(64,)", style=solid];
 "214 scale_updated_constant19" -> "216 mul_tensor_19"  [label="(128, 1, 1, 1)", style=solid];
 "215 compressed_weight_updated_constant19" -> "216 mul_tensor_19"  [label="(128, 64, 2, 2)", style=solid];
 "216 mul_tensor_19" -> "218 sub_tensor_19"  [label="(128, 64, 2, 2)", style=solid];
@@ -485,7 +485,7 @@ strict digraph  {
 "224 slice_15" -> "225 slice_16"  [label="(1, 64, 40, 220)", style=solid];
 "225 slice_16" -> "226 cat_3"  [label="(1, 64, 40, 40)", style=solid];
 "226 cat_3" -> "233 conv2d_16"  [label="(1, 128, 40, 40)", style=solid];
-"227 _param_constant73_0_0" -> "233 conv2d_16"  [label="(64,)", style=solid];
+"227 up_path_3_conv_block_block_0_bias_0_0" -> "233 conv2d_16"  [label="(64,)", style=solid];
 "228 scale_updated_constant20" -> "230 mul_tensor_20"  [label="(64, 1, 1, 1)", style=solid];
 "229 compressed_weight_updated_constant20" -> "230 mul_tensor_20"  [label="(64, 128, 3, 3)", style=solid];
 "230 mul_tensor_20" -> "232 sub_tensor_20"  [label="(64, 128, 3, 3)", style=solid];
@@ -495,7 +495,7 @@ strict digraph  {
 "234 relu_16" -> "235 quantize_per_tensor_default_21"  [label="(1, 64, 38, 38)", style=solid];
 "235 quantize_per_tensor_default_21" -> "236 dequantize_per_tensor_default_25"  [label="(1, 64, 38, 38)", style=solid];
 "236 dequantize_per_tensor_default_25" -> "243 conv2d_17"  [label="(1, 64, 38, 38)", style=solid];
-"237 _param_constant77_0_0" -> "243 conv2d_17"  [label="(64,)", style=solid];
+"237 up_path_3_conv_block_block_3_bias_0_0" -> "243 conv2d_17"  [label="(64,)", style=solid];
 "238 scale_updated_constant21" -> "240 mul_tensor_21"  [label="(64, 1, 1, 1)", style=solid];
 "239 compressed_weight_updated_constant21" -> "240 mul_tensor_21"  [label="(64, 64, 3, 3)", style=solid];
 "240 mul_tensor_21" -> "242 sub_tensor_21"  [label="(64, 64, 3, 3)", style=solid];
@@ -505,7 +505,7 @@ strict digraph  {
 "244 relu_17" -> "245 quantize_per_tensor_default_22"  [label="(1, 64, 36, 36)", style=solid];
 "245 quantize_per_tensor_default_22" -> "246 dequantize_per_tensor_default_26"  [label="(1, 64, 36, 36)", style=solid];
 "246 dequantize_per_tensor_default_26" -> "253 conv2d_18"  [label="(1, 64, 36, 36)", style=solid];
-"247 _param_constant81_0_0" -> "253 conv2d_18"  [label="(12,)", style=solid];
+"247 last_bias_0_0" -> "253 conv2d_18"  [label="(12,)", style=solid];
 "248 scale_updated_constant22" -> "250 mul_tensor_22"  [label="(12, 1, 1, 1)", style=solid];
 "249 compressed_weight_updated_constant22" -> "250 mul_tensor_22"  [label="(12, 64, 1, 1)", style=solid];
 "250 mul_tensor_22" -> "252 sub_tensor_22"  [label="(12, 64, 1, 1)", style=solid];
diff --git a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/vit_b_16.dot b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/vit_b_16.dot
index d031abe21c5..88e4a8c30d1 100644
--- a/tests/torch/data/reference_graphs/fx/post_quantization_compressed/vit_b_16.dot
+++ b/tests/torch/data/reference_graphs/fx/post_quantization_compressed/vit_b_16.dot
@@ -1,978 +1,978 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 quantize_per_tensor_default" [id=1, type=quantize_per_tensor];
-"2 dequantize_per_tensor_default" [id=2, type=dequantize_per_tensor];
-"3 _param_constant1_0_0" [id=3, type=get_attr];
-"4 scale_updated_constant0" [id=4, type=get_attr];
-"5 compressed_weight_updated_constant0" [id=5, type=get_attr];
-"6 mul_tensor" [id=6, type=mul];
-"7 zero_point_updated_constant0" [id=7, type=get_attr];
-"8 sub_tensor" [id=8, type=sub];
-"9 conv2d" [id=9, type=conv2d];
-"10 reshape" [id=10, type=reshape];
-"11 permute" [id=11, type=permute];
-"12 _frozen_param0" [id=12, type=get_attr];
-"13 cat" [id=13, type=cat];
-"14 _param_constant3" [id=14, type=get_attr];
-"15 add" [id=15, type=add];
-"16 dropout" [id=16, type=dropout];
-"17 _param_constant4" [id=17, type=get_attr];
-"18 _param_constant5" [id=18, type=get_attr];
-"19 layer_norm" [id=19, type=layer_norm];
-"20 transpose" [id=20, type=transpose];
-"21 transpose_0_0_nncf_smooth_quant_0" [id=21, type=call_module];
-"22 quantize_per_tensor_default_1" [id=22, type=quantize_per_tensor];
-"23 dequantize_per_tensor_default_1" [id=23, type=dequantize_per_tensor];
-"24 _param_constant7_0_0" [id=24, type=get_attr];
-"25 scale_updated_constant1" [id=25, type=get_attr];
-"26 compressed_weight_updated_constant1" [id=26, type=get_attr];
-"27 mul_tensor_1" [id=27, type=mul];
-"28 zero_point_updated_constant1" [id=28, type=get_attr];
-"29 sub_tensor_1" [id=29, type=sub];
-"30 linear" [id=30, type=linear];
-"31 unflatten" [id=31, type=unflatten];
-"32 unsqueeze" [id=32, type=unsqueeze];
-"33 transpose_1" [id=33, type=transpose];
-"34 squeeze" [id=34, type=squeeze];
-"35 contiguous" [id=35, type=contiguous];
-"36 quantize_per_tensor_default_2" [id=36, type=quantize_per_tensor];
-"37 dequantize_per_tensor_default_2" [id=37, type=dequantize_per_tensor];
-"38 select" [id=38, type=select];
-"39 quantize_per_tensor_default_3" [id=39, type=quantize_per_tensor];
-"40 dequantize_per_tensor_default_3" [id=40, type=dequantize_per_tensor];
-"41 select_1" [id=41, type=select];
-"42 select_2" [id=42, type=select];
-"43 view" [id=43, type=view];
-"44 transpose_2" [id=44, type=transpose];
-"45 view_1" [id=45, type=view];
-"46 transpose_3" [id=46, type=transpose];
-"47 view_2" [id=47, type=view];
-"48 transpose_4" [id=48, type=transpose];
-"49 view_3" [id=49, type=view];
-"50 view_4" [id=50, type=view];
-"51 view_5" [id=51, type=view];
-"52 scaled_dot_product_attention" [id=52, type=scaled_dot_product_attention];
-"53 permute_1" [id=53, type=permute];
-"54 view_6" [id=54, type=view];
-"55 view_6_0_0_nncf_smooth_quant_0" [id=55, type=call_module];
-"56 quantize_per_tensor_default_4" [id=56, type=quantize_per_tensor];
-"57 dequantize_per_tensor_default_4" [id=57, type=dequantize_per_tensor];
-"58 _param_constant9_0_0" [id=58, type=get_attr];
-"59 scale_updated_constant2" [id=59, type=get_attr];
-"60 compressed_weight_updated_constant2" [id=60, type=get_attr];
-"61 mul_tensor_2" [id=61, type=mul];
-"62 zero_point_updated_constant2" [id=62, type=get_attr];
-"63 sub_tensor_2" [id=63, type=sub];
-"64 linear_1" [id=64, type=linear];
-"65 view_7" [id=65, type=view];
-"66 transpose_5" [id=66, type=transpose];
-"67 dropout_1" [id=67, type=dropout];
-"68 add_1" [id=68, type=add];
-"69 _param_constant10" [id=69, type=get_attr];
-"70 _param_constant11" [id=70, type=get_attr];
-"71 layer_norm_1" [id=71, type=layer_norm];
-"72 layer_norm_1_0_0_nncf_smooth_quant_0" [id=72, type=call_module];
-"73 quantize_per_tensor_default_5" [id=73, type=quantize_per_tensor];
-"74 dequantize_per_tensor_default_5" [id=74, type=dequantize_per_tensor];
-"75 _param_constant13_0_0" [id=75, type=get_attr];
-"76 scale_updated_constant3" [id=76, type=get_attr];
-"77 compressed_weight_updated_constant3" [id=77, type=get_attr];
-"78 mul_tensor_3" [id=78, type=mul];
-"79 zero_point_updated_constant3" [id=79, type=get_attr];
-"80 sub_tensor_3" [id=80, type=sub];
-"81 linear_2" [id=81, type=linear];
-"82 gelu" [id=82, type=gelu];
-"83 dropout_2" [id=83, type=dropout];
-"84 dropout_2_0_0_nncf_smooth_quant_0" [id=84, type=call_module];
-"85 quantize_per_tensor_default_6" [id=85, type=quantize_per_tensor];
-"86 dequantize_per_tensor_default_6" [id=86, type=dequantize_per_tensor];
-"87 _param_constant15_0_0" [id=87, type=get_attr];
-"88 scale_updated_constant4" [id=88, type=get_attr];
-"89 compressed_weight_updated_constant4" [id=89, type=get_attr];
-"90 mul_tensor_4" [id=90, type=mul];
-"91 zero_point_updated_constant4" [id=91, type=get_attr];
-"92 sub_tensor_4" [id=92, type=sub];
-"93 linear_3" [id=93, type=linear];
-"94 dropout_3" [id=94, type=dropout];
-"95 add_2" [id=95, type=add];
-"96 _param_constant16" [id=96, type=get_attr];
-"97 _param_constant17" [id=97, type=get_attr];
-"98 layer_norm_2" [id=98, type=layer_norm];
-"99 transpose_6" [id=99, type=transpose];
-"100 transpose_6_0_0_nncf_smooth_quant_0" [id=100, type=call_module];
-"101 quantize_per_tensor_default_7" [id=101, type=quantize_per_tensor];
-"102 dequantize_per_tensor_default_7" [id=102, type=dequantize_per_tensor];
-"103 _param_constant19_0_0" [id=103, type=get_attr];
-"104 scale_updated_constant5" [id=104, type=get_attr];
-"105 compressed_weight_updated_constant5" [id=105, type=get_attr];
-"106 mul_tensor_5" [id=106, type=mul];
-"107 zero_point_updated_constant5" [id=107, type=get_attr];
-"108 sub_tensor_5" [id=108, type=sub];
-"109 linear_4" [id=109, type=linear];
-"110 unflatten_1" [id=110, type=unflatten];
-"111 unsqueeze_1" [id=111, type=unsqueeze];
-"112 transpose_7" [id=112, type=transpose];
-"113 squeeze_1" [id=113, type=squeeze];
-"114 contiguous_1" [id=114, type=contiguous];
-"115 quantize_per_tensor_default_8" [id=115, type=quantize_per_tensor];
-"116 dequantize_per_tensor_default_8" [id=116, type=dequantize_per_tensor];
-"117 select_3" [id=117, type=select];
-"118 quantize_per_tensor_default_9" [id=118, type=quantize_per_tensor];
-"119 dequantize_per_tensor_default_9" [id=119, type=dequantize_per_tensor];
-"120 select_4" [id=120, type=select];
-"121 select_5" [id=121, type=select];
-"122 view_8" [id=122, type=view];
-"123 transpose_8" [id=123, type=transpose];
-"124 view_9" [id=124, type=view];
-"125 transpose_9" [id=125, type=transpose];
-"126 view_10" [id=126, type=view];
-"127 transpose_10" [id=127, type=transpose];
-"128 view_11" [id=128, type=view];
-"129 view_12" [id=129, type=view];
-"130 view_13" [id=130, type=view];
-"131 scaled_dot_product_attention_1" [id=131, type=scaled_dot_product_attention];
-"132 permute_2" [id=132, type=permute];
-"133 view_14" [id=133, type=view];
-"134 view_14_0_0_nncf_smooth_quant_0" [id=134, type=call_module];
-"135 quantize_per_tensor_default_10" [id=135, type=quantize_per_tensor];
-"136 dequantize_per_tensor_default_10" [id=136, type=dequantize_per_tensor];
-"137 _param_constant21_0_0" [id=137, type=get_attr];
-"138 scale_updated_constant6" [id=138, type=get_attr];
-"139 compressed_weight_updated_constant6" [id=139, type=get_attr];
-"140 mul_tensor_6" [id=140, type=mul];
-"141 zero_point_updated_constant6" [id=141, type=get_attr];
-"142 sub_tensor_6" [id=142, type=sub];
-"143 linear_5" [id=143, type=linear];
-"144 view_15" [id=144, type=view];
-"145 transpose_11" [id=145, type=transpose];
-"146 dropout_4" [id=146, type=dropout];
-"147 add_3" [id=147, type=add];
-"148 _param_constant22" [id=148, type=get_attr];
-"149 _param_constant23" [id=149, type=get_attr];
-"150 layer_norm_3" [id=150, type=layer_norm];
-"151 layer_norm_3_0_0_nncf_smooth_quant_0" [id=151, type=call_module];
-"152 quantize_per_tensor_default_11" [id=152, type=quantize_per_tensor];
-"153 dequantize_per_tensor_default_11" [id=153, type=dequantize_per_tensor];
-"154 _param_constant25_0_0" [id=154, type=get_attr];
-"155 scale_updated_constant7" [id=155, type=get_attr];
-"156 compressed_weight_updated_constant7" [id=156, type=get_attr];
-"157 mul_tensor_7" [id=157, type=mul];
-"158 zero_point_updated_constant7" [id=158, type=get_attr];
-"159 sub_tensor_7" [id=159, type=sub];
-"160 linear_6" [id=160, type=linear];
-"161 gelu_1" [id=161, type=gelu];
-"162 dropout_5" [id=162, type=dropout];
-"163 dropout_5_0_0_nncf_smooth_quant_0" [id=163, type=call_module];
-"164 quantize_per_tensor_default_12" [id=164, type=quantize_per_tensor];
-"165 dequantize_per_tensor_default_12" [id=165, type=dequantize_per_tensor];
-"166 _param_constant27_0_0" [id=166, type=get_attr];
-"167 scale_updated_constant8" [id=167, type=get_attr];
-"168 compressed_weight_updated_constant8" [id=168, type=get_attr];
-"169 mul_tensor_8" [id=169, type=mul];
-"170 zero_point_updated_constant8" [id=170, type=get_attr];
-"171 sub_tensor_8" [id=171, type=sub];
-"172 linear_7" [id=172, type=linear];
-"173 dropout_6" [id=173, type=dropout];
-"174 add_4" [id=174, type=add];
-"175 _param_constant28" [id=175, type=get_attr];
-"176 _param_constant29" [id=176, type=get_attr];
-"177 layer_norm_4" [id=177, type=layer_norm];
-"178 transpose_12" [id=178, type=transpose];
-"179 transpose_12_0_0_nncf_smooth_quant_0" [id=179, type=call_module];
-"180 quantize_per_tensor_default_13" [id=180, type=quantize_per_tensor];
-"181 dequantize_per_tensor_default_13" [id=181, type=dequantize_per_tensor];
-"182 _param_constant31_0_0" [id=182, type=get_attr];
-"183 scale_updated_constant9" [id=183, type=get_attr];
-"184 compressed_weight_updated_constant9" [id=184, type=get_attr];
-"185 mul_tensor_9" [id=185, type=mul];
-"186 zero_point_updated_constant9" [id=186, type=get_attr];
-"187 sub_tensor_9" [id=187, type=sub];
-"188 linear_8" [id=188, type=linear];
-"189 unflatten_2" [id=189, type=unflatten];
-"190 unsqueeze_2" [id=190, type=unsqueeze];
-"191 transpose_13" [id=191, type=transpose];
-"192 squeeze_2" [id=192, type=squeeze];
-"193 contiguous_2" [id=193, type=contiguous];
-"194 quantize_per_tensor_default_14" [id=194, type=quantize_per_tensor];
-"195 dequantize_per_tensor_default_14" [id=195, type=dequantize_per_tensor];
-"196 select_6" [id=196, type=select];
-"197 quantize_per_tensor_default_15" [id=197, type=quantize_per_tensor];
-"198 dequantize_per_tensor_default_15" [id=198, type=dequantize_per_tensor];
-"199 select_7" [id=199, type=select];
-"200 select_8" [id=200, type=select];
-"201 view_16" [id=201, type=view];
-"202 transpose_14" [id=202, type=transpose];
-"203 view_17" [id=203, type=view];
-"204 transpose_15" [id=204, type=transpose];
-"205 view_18" [id=205, type=view];
-"206 transpose_16" [id=206, type=transpose];
-"207 view_19" [id=207, type=view];
-"208 view_20" [id=208, type=view];
-"209 view_21" [id=209, type=view];
-"210 scaled_dot_product_attention_2" [id=210, type=scaled_dot_product_attention];
-"211 permute_3" [id=211, type=permute];
-"212 view_22" [id=212, type=view];
-"213 view_22_0_0_nncf_smooth_quant_0" [id=213, type=call_module];
-"214 quantize_per_tensor_default_16" [id=214, type=quantize_per_tensor];
-"215 dequantize_per_tensor_default_16" [id=215, type=dequantize_per_tensor];
-"216 _param_constant33_0_0" [id=216, type=get_attr];
-"217 scale_updated_constant10" [id=217, type=get_attr];
-"218 compressed_weight_updated_constant10" [id=218, type=get_attr];
-"219 mul_tensor_10" [id=219, type=mul];
-"220 zero_point_updated_constant10" [id=220, type=get_attr];
-"221 sub_tensor_10" [id=221, type=sub];
-"222 linear_9" [id=222, type=linear];
-"223 view_23" [id=223, type=view];
-"224 transpose_17" [id=224, type=transpose];
-"225 dropout_7" [id=225, type=dropout];
-"226 add_5" [id=226, type=add];
-"227 _param_constant34" [id=227, type=get_attr];
-"228 _param_constant35" [id=228, type=get_attr];
-"229 layer_norm_5" [id=229, type=layer_norm];
-"230 layer_norm_5_0_0_nncf_smooth_quant_0" [id=230, type=call_module];
-"231 quantize_per_tensor_default_17" [id=231, type=quantize_per_tensor];
-"232 dequantize_per_tensor_default_17" [id=232, type=dequantize_per_tensor];
-"233 _param_constant37_0_0" [id=233, type=get_attr];
-"234 scale_updated_constant11" [id=234, type=get_attr];
-"235 compressed_weight_updated_constant11" [id=235, type=get_attr];
-"236 mul_tensor_11" [id=236, type=mul];
-"237 zero_point_updated_constant11" [id=237, type=get_attr];
-"238 sub_tensor_11" [id=238, type=sub];
-"239 linear_10" [id=239, type=linear];
-"240 gelu_2" [id=240, type=gelu];
-"241 dropout_8" [id=241, type=dropout];
-"242 dropout_8_0_0_nncf_smooth_quant_0" [id=242, type=call_module];
-"243 quantize_per_tensor_default_18" [id=243, type=quantize_per_tensor];
-"244 dequantize_per_tensor_default_18" [id=244, type=dequantize_per_tensor];
-"245 _param_constant39_0_0" [id=245, type=get_attr];
-"246 scale_updated_constant12" [id=246, type=get_attr];
-"247 compressed_weight_updated_constant12" [id=247, type=get_attr];
-"248 mul_tensor_12" [id=248, type=mul];
-"249 zero_point_updated_constant12" [id=249, type=get_attr];
-"250 sub_tensor_12" [id=250, type=sub];
-"251 linear_11" [id=251, type=linear];
-"252 dropout_9" [id=252, type=dropout];
-"253 add_6" [id=253, type=add];
-"254 _param_constant40" [id=254, type=get_attr];
-"255 _param_constant41" [id=255, type=get_attr];
-"256 layer_norm_6" [id=256, type=layer_norm];
-"257 transpose_18" [id=257, type=transpose];
-"258 transpose_18_0_0_nncf_smooth_quant_0" [id=258, type=call_module];
-"259 quantize_per_tensor_default_19" [id=259, type=quantize_per_tensor];
-"260 dequantize_per_tensor_default_19" [id=260, type=dequantize_per_tensor];
-"261 _param_constant43_0_0" [id=261, type=get_attr];
-"262 scale_updated_constant13" [id=262, type=get_attr];
-"263 compressed_weight_updated_constant13" [id=263, type=get_attr];
-"264 mul_tensor_13" [id=264, type=mul];
-"265 zero_point_updated_constant13" [id=265, type=get_attr];
-"266 sub_tensor_13" [id=266, type=sub];
-"267 linear_12" [id=267, type=linear];
-"268 unflatten_3" [id=268, type=unflatten];
-"269 unsqueeze_3" [id=269, type=unsqueeze];
-"270 transpose_19" [id=270, type=transpose];
-"271 squeeze_3" [id=271, type=squeeze];
-"272 contiguous_3" [id=272, type=contiguous];
-"273 quantize_per_tensor_default_20" [id=273, type=quantize_per_tensor];
-"274 dequantize_per_tensor_default_20" [id=274, type=dequantize_per_tensor];
-"275 select_9" [id=275, type=select];
-"276 quantize_per_tensor_default_21" [id=276, type=quantize_per_tensor];
-"277 dequantize_per_tensor_default_21" [id=277, type=dequantize_per_tensor];
-"278 select_10" [id=278, type=select];
-"279 select_11" [id=279, type=select];
-"280 view_24" [id=280, type=view];
-"281 transpose_20" [id=281, type=transpose];
-"282 view_25" [id=282, type=view];
-"283 transpose_21" [id=283, type=transpose];
-"284 view_26" [id=284, type=view];
-"285 transpose_22" [id=285, type=transpose];
-"286 view_27" [id=286, type=view];
-"287 view_28" [id=287, type=view];
-"288 view_29" [id=288, type=view];
-"289 scaled_dot_product_attention_3" [id=289, type=scaled_dot_product_attention];
-"290 permute_4" [id=290, type=permute];
-"291 view_30" [id=291, type=view];
-"292 view_30_0_0_nncf_smooth_quant_0" [id=292, type=call_module];
-"293 quantize_per_tensor_default_22" [id=293, type=quantize_per_tensor];
-"294 dequantize_per_tensor_default_22" [id=294, type=dequantize_per_tensor];
-"295 _param_constant45_0_0" [id=295, type=get_attr];
-"296 scale_updated_constant14" [id=296, type=get_attr];
-"297 compressed_weight_updated_constant14" [id=297, type=get_attr];
-"298 mul_tensor_14" [id=298, type=mul];
-"299 zero_point_updated_constant14" [id=299, type=get_attr];
-"300 sub_tensor_14" [id=300, type=sub];
-"301 linear_13" [id=301, type=linear];
-"302 view_31" [id=302, type=view];
-"303 transpose_23" [id=303, type=transpose];
-"304 dropout_10" [id=304, type=dropout];
-"305 add_7" [id=305, type=add];
-"306 _param_constant46" [id=306, type=get_attr];
-"307 _param_constant47" [id=307, type=get_attr];
-"308 layer_norm_7" [id=308, type=layer_norm];
-"309 layer_norm_7_0_0_nncf_smooth_quant_0" [id=309, type=call_module];
-"310 quantize_per_tensor_default_23" [id=310, type=quantize_per_tensor];
-"311 dequantize_per_tensor_default_23" [id=311, type=dequantize_per_tensor];
-"312 _param_constant49_0_0" [id=312, type=get_attr];
-"313 scale_updated_constant15" [id=313, type=get_attr];
-"314 compressed_weight_updated_constant15" [id=314, type=get_attr];
-"315 mul_tensor_15" [id=315, type=mul];
-"316 zero_point_updated_constant15" [id=316, type=get_attr];
-"317 sub_tensor_15" [id=317, type=sub];
-"318 linear_14" [id=318, type=linear];
-"319 gelu_3" [id=319, type=gelu];
-"320 dropout_11" [id=320, type=dropout];
-"321 dropout_11_0_0_nncf_smooth_quant_0" [id=321, type=call_module];
-"322 quantize_per_tensor_default_24" [id=322, type=quantize_per_tensor];
-"323 dequantize_per_tensor_default_24" [id=323, type=dequantize_per_tensor];
-"324 _param_constant51_0_0" [id=324, type=get_attr];
-"325 scale_updated_constant16" [id=325, type=get_attr];
-"326 compressed_weight_updated_constant16" [id=326, type=get_attr];
-"327 mul_tensor_16" [id=327, type=mul];
-"328 zero_point_updated_constant16" [id=328, type=get_attr];
-"329 sub_tensor_16" [id=329, type=sub];
-"330 linear_15" [id=330, type=linear];
-"331 dropout_12" [id=331, type=dropout];
-"332 add_8" [id=332, type=add];
-"333 _param_constant52" [id=333, type=get_attr];
-"334 _param_constant53" [id=334, type=get_attr];
-"335 layer_norm_8" [id=335, type=layer_norm];
-"336 transpose_24" [id=336, type=transpose];
-"337 transpose_24_0_0_nncf_smooth_quant_0" [id=337, type=call_module];
-"338 quantize_per_tensor_default_25" [id=338, type=quantize_per_tensor];
-"339 dequantize_per_tensor_default_25" [id=339, type=dequantize_per_tensor];
-"340 _param_constant55_0_0" [id=340, type=get_attr];
-"341 scale_updated_constant17" [id=341, type=get_attr];
-"342 compressed_weight_updated_constant17" [id=342, type=get_attr];
-"343 mul_tensor_17" [id=343, type=mul];
-"344 zero_point_updated_constant17" [id=344, type=get_attr];
-"345 sub_tensor_17" [id=345, type=sub];
-"346 linear_16" [id=346, type=linear];
-"347 unflatten_4" [id=347, type=unflatten];
-"348 unsqueeze_4" [id=348, type=unsqueeze];
-"349 transpose_25" [id=349, type=transpose];
-"350 squeeze_4" [id=350, type=squeeze];
-"351 contiguous_4" [id=351, type=contiguous];
-"352 quantize_per_tensor_default_26" [id=352, type=quantize_per_tensor];
-"353 dequantize_per_tensor_default_26" [id=353, type=dequantize_per_tensor];
-"354 select_12" [id=354, type=select];
-"355 quantize_per_tensor_default_27" [id=355, type=quantize_per_tensor];
-"356 dequantize_per_tensor_default_27" [id=356, type=dequantize_per_tensor];
-"357 select_13" [id=357, type=select];
-"358 select_14" [id=358, type=select];
-"359 view_32" [id=359, type=view];
-"360 transpose_26" [id=360, type=transpose];
-"361 view_33" [id=361, type=view];
-"362 transpose_27" [id=362, type=transpose];
-"363 view_34" [id=363, type=view];
-"364 transpose_28" [id=364, type=transpose];
-"365 view_35" [id=365, type=view];
-"366 view_36" [id=366, type=view];
-"367 view_37" [id=367, type=view];
-"368 scaled_dot_product_attention_4" [id=368, type=scaled_dot_product_attention];
-"369 permute_5" [id=369, type=permute];
-"370 view_38" [id=370, type=view];
-"371 view_38_0_0_nncf_smooth_quant_0" [id=371, type=call_module];
-"372 quantize_per_tensor_default_28" [id=372, type=quantize_per_tensor];
-"373 dequantize_per_tensor_default_28" [id=373, type=dequantize_per_tensor];
-"374 _param_constant57_0_0" [id=374, type=get_attr];
-"375 scale_updated_constant18" [id=375, type=get_attr];
-"376 compressed_weight_updated_constant18" [id=376, type=get_attr];
-"377 mul_tensor_18" [id=377, type=mul];
-"378 zero_point_updated_constant18" [id=378, type=get_attr];
-"379 sub_tensor_18" [id=379, type=sub];
-"380 linear_17" [id=380, type=linear];
-"381 view_39" [id=381, type=view];
-"382 transpose_29" [id=382, type=transpose];
-"383 dropout_13" [id=383, type=dropout];
-"384 add_9" [id=384, type=add];
-"385 _param_constant58" [id=385, type=get_attr];
-"386 _param_constant59" [id=386, type=get_attr];
-"387 layer_norm_9" [id=387, type=layer_norm];
-"388 layer_norm_9_0_0_nncf_smooth_quant_0" [id=388, type=call_module];
-"389 quantize_per_tensor_default_29" [id=389, type=quantize_per_tensor];
-"390 dequantize_per_tensor_default_29" [id=390, type=dequantize_per_tensor];
-"391 _param_constant61_0_0" [id=391, type=get_attr];
-"392 scale_updated_constant19" [id=392, type=get_attr];
-"393 compressed_weight_updated_constant19" [id=393, type=get_attr];
-"394 mul_tensor_19" [id=394, type=mul];
-"395 zero_point_updated_constant19" [id=395, type=get_attr];
-"396 sub_tensor_19" [id=396, type=sub];
-"397 linear_18" [id=397, type=linear];
-"398 gelu_4" [id=398, type=gelu];
-"399 dropout_14" [id=399, type=dropout];
-"400 dropout_14_0_0_nncf_smooth_quant_0" [id=400, type=call_module];
-"401 quantize_per_tensor_default_30" [id=401, type=quantize_per_tensor];
-"402 dequantize_per_tensor_default_30" [id=402, type=dequantize_per_tensor];
-"403 _param_constant63_0_0" [id=403, type=get_attr];
-"404 scale_updated_constant20" [id=404, type=get_attr];
-"405 compressed_weight_updated_constant20" [id=405, type=get_attr];
-"406 mul_tensor_20" [id=406, type=mul];
-"407 zero_point_updated_constant20" [id=407, type=get_attr];
-"408 sub_tensor_20" [id=408, type=sub];
-"409 linear_19" [id=409, type=linear];
-"410 dropout_15" [id=410, type=dropout];
-"411 add_10" [id=411, type=add];
-"412 _param_constant64" [id=412, type=get_attr];
-"413 _param_constant65" [id=413, type=get_attr];
-"414 layer_norm_10" [id=414, type=layer_norm];
-"415 transpose_30" [id=415, type=transpose];
-"416 transpose_30_0_0_nncf_smooth_quant_0" [id=416, type=call_module];
-"417 quantize_per_tensor_default_31" [id=417, type=quantize_per_tensor];
-"418 dequantize_per_tensor_default_31" [id=418, type=dequantize_per_tensor];
-"419 _param_constant67_0_0" [id=419, type=get_attr];
-"420 scale_updated_constant21" [id=420, type=get_attr];
-"421 compressed_weight_updated_constant21" [id=421, type=get_attr];
-"422 mul_tensor_21" [id=422, type=mul];
-"423 zero_point_updated_constant21" [id=423, type=get_attr];
-"424 sub_tensor_21" [id=424, type=sub];
-"425 linear_20" [id=425, type=linear];
-"426 unflatten_5" [id=426, type=unflatten];
-"427 unsqueeze_5" [id=427, type=unsqueeze];
-"428 transpose_31" [id=428, type=transpose];
-"429 squeeze_5" [id=429, type=squeeze];
-"430 contiguous_5" [id=430, type=contiguous];
-"431 quantize_per_tensor_default_32" [id=431, type=quantize_per_tensor];
-"432 dequantize_per_tensor_default_32" [id=432, type=dequantize_per_tensor];
-"433 select_15" [id=433, type=select];
-"434 quantize_per_tensor_default_33" [id=434, type=quantize_per_tensor];
-"435 dequantize_per_tensor_default_33" [id=435, type=dequantize_per_tensor];
-"436 select_16" [id=436, type=select];
-"437 select_17" [id=437, type=select];
-"438 view_40" [id=438, type=view];
-"439 transpose_32" [id=439, type=transpose];
-"440 view_41" [id=440, type=view];
-"441 transpose_33" [id=441, type=transpose];
-"442 view_42" [id=442, type=view];
-"443 transpose_34" [id=443, type=transpose];
-"444 view_43" [id=444, type=view];
-"445 view_44" [id=445, type=view];
-"446 view_45" [id=446, type=view];
-"447 scaled_dot_product_attention_5" [id=447, type=scaled_dot_product_attention];
-"448 permute_6" [id=448, type=permute];
-"449 view_46" [id=449, type=view];
-"450 view_46_0_0_nncf_smooth_quant_0" [id=450, type=call_module];
-"451 quantize_per_tensor_default_34" [id=451, type=quantize_per_tensor];
-"452 dequantize_per_tensor_default_34" [id=452, type=dequantize_per_tensor];
-"453 _param_constant69_0_0" [id=453, type=get_attr];
-"454 scale_updated_constant22" [id=454, type=get_attr];
-"455 compressed_weight_updated_constant22" [id=455, type=get_attr];
-"456 mul_tensor_22" [id=456, type=mul];
-"457 zero_point_updated_constant22" [id=457, type=get_attr];
-"458 sub_tensor_22" [id=458, type=sub];
-"459 linear_21" [id=459, type=linear];
-"460 view_47" [id=460, type=view];
-"461 transpose_35" [id=461, type=transpose];
-"462 dropout_16" [id=462, type=dropout];
-"463 add_11" [id=463, type=add];
-"464 _param_constant70" [id=464, type=get_attr];
-"465 _param_constant71" [id=465, type=get_attr];
-"466 layer_norm_11" [id=466, type=layer_norm];
-"467 layer_norm_11_0_0_nncf_smooth_quant_0" [id=467, type=call_module];
-"468 quantize_per_tensor_default_35" [id=468, type=quantize_per_tensor];
-"469 dequantize_per_tensor_default_35" [id=469, type=dequantize_per_tensor];
-"470 _param_constant73_0_0" [id=470, type=get_attr];
-"471 scale_updated_constant23" [id=471, type=get_attr];
-"472 compressed_weight_updated_constant23" [id=472, type=get_attr];
-"473 mul_tensor_23" [id=473, type=mul];
-"474 zero_point_updated_constant23" [id=474, type=get_attr];
-"475 sub_tensor_23" [id=475, type=sub];
-"476 linear_22" [id=476, type=linear];
-"477 gelu_5" [id=477, type=gelu];
-"478 dropout_17" [id=478, type=dropout];
-"479 dropout_17_0_0_nncf_smooth_quant_0" [id=479, type=call_module];
-"480 quantize_per_tensor_default_36" [id=480, type=quantize_per_tensor];
-"481 dequantize_per_tensor_default_36" [id=481, type=dequantize_per_tensor];
-"482 _param_constant75_0_0" [id=482, type=get_attr];
-"483 scale_updated_constant24" [id=483, type=get_attr];
-"484 compressed_weight_updated_constant24" [id=484, type=get_attr];
-"485 mul_tensor_24" [id=485, type=mul];
-"486 zero_point_updated_constant24" [id=486, type=get_attr];
-"487 sub_tensor_24" [id=487, type=sub];
-"488 linear_23" [id=488, type=linear];
-"489 dropout_18" [id=489, type=dropout];
-"490 add_12" [id=490, type=add];
-"491 _param_constant76" [id=491, type=get_attr];
-"492 _param_constant77" [id=492, type=get_attr];
-"493 layer_norm_12" [id=493, type=layer_norm];
-"494 transpose_36" [id=494, type=transpose];
-"495 transpose_36_0_0_nncf_smooth_quant_0" [id=495, type=call_module];
-"496 quantize_per_tensor_default_37" [id=496, type=quantize_per_tensor];
-"497 dequantize_per_tensor_default_37" [id=497, type=dequantize_per_tensor];
-"498 _param_constant79_0_0" [id=498, type=get_attr];
-"499 scale_updated_constant25" [id=499, type=get_attr];
-"500 compressed_weight_updated_constant25" [id=500, type=get_attr];
-"501 mul_tensor_25" [id=501, type=mul];
-"502 zero_point_updated_constant25" [id=502, type=get_attr];
-"503 sub_tensor_25" [id=503, type=sub];
-"504 linear_24" [id=504, type=linear];
-"505 unflatten_6" [id=505, type=unflatten];
-"506 unsqueeze_6" [id=506, type=unsqueeze];
-"507 transpose_37" [id=507, type=transpose];
-"508 squeeze_6" [id=508, type=squeeze];
-"509 contiguous_6" [id=509, type=contiguous];
-"510 quantize_per_tensor_default_38" [id=510, type=quantize_per_tensor];
-"511 dequantize_per_tensor_default_38" [id=511, type=dequantize_per_tensor];
-"512 select_18" [id=512, type=select];
-"513 quantize_per_tensor_default_39" [id=513, type=quantize_per_tensor];
-"514 dequantize_per_tensor_default_39" [id=514, type=dequantize_per_tensor];
-"515 select_19" [id=515, type=select];
-"516 select_20" [id=516, type=select];
-"517 view_48" [id=517, type=view];
-"518 transpose_38" [id=518, type=transpose];
-"519 view_49" [id=519, type=view];
-"520 transpose_39" [id=520, type=transpose];
-"521 view_50" [id=521, type=view];
-"522 transpose_40" [id=522, type=transpose];
-"523 view_51" [id=523, type=view];
-"524 view_52" [id=524, type=view];
-"525 view_53" [id=525, type=view];
-"526 scaled_dot_product_attention_6" [id=526, type=scaled_dot_product_attention];
-"527 permute_7" [id=527, type=permute];
-"528 view_54" [id=528, type=view];
-"529 view_54_0_0_nncf_smooth_quant_0" [id=529, type=call_module];
-"530 quantize_per_tensor_default_40" [id=530, type=quantize_per_tensor];
-"531 dequantize_per_tensor_default_40" [id=531, type=dequantize_per_tensor];
-"532 _param_constant81_0_0" [id=532, type=get_attr];
-"533 scale_updated_constant26" [id=533, type=get_attr];
-"534 compressed_weight_updated_constant26" [id=534, type=get_attr];
-"535 mul_tensor_26" [id=535, type=mul];
-"536 zero_point_updated_constant26" [id=536, type=get_attr];
-"537 sub_tensor_26" [id=537, type=sub];
-"538 linear_25" [id=538, type=linear];
-"539 view_55" [id=539, type=view];
-"540 transpose_41" [id=540, type=transpose];
-"541 dropout_19" [id=541, type=dropout];
-"542 add_13" [id=542, type=add];
-"543 _param_constant82" [id=543, type=get_attr];
-"544 _param_constant83" [id=544, type=get_attr];
-"545 layer_norm_13" [id=545, type=layer_norm];
-"546 layer_norm_13_0_0_nncf_smooth_quant_0" [id=546, type=call_module];
-"547 quantize_per_tensor_default_41" [id=547, type=quantize_per_tensor];
-"548 dequantize_per_tensor_default_41" [id=548, type=dequantize_per_tensor];
-"549 _param_constant85_0_0" [id=549, type=get_attr];
-"550 scale_updated_constant27" [id=550, type=get_attr];
-"551 compressed_weight_updated_constant27" [id=551, type=get_attr];
-"552 mul_tensor_27" [id=552, type=mul];
-"553 zero_point_updated_constant27" [id=553, type=get_attr];
-"554 sub_tensor_27" [id=554, type=sub];
-"555 linear_26" [id=555, type=linear];
-"556 gelu_6" [id=556, type=gelu];
-"557 dropout_20" [id=557, type=dropout];
-"558 dropout_20_0_0_nncf_smooth_quant_0" [id=558, type=call_module];
-"559 quantize_per_tensor_default_42" [id=559, type=quantize_per_tensor];
-"560 dequantize_per_tensor_default_42" [id=560, type=dequantize_per_tensor];
-"561 _param_constant87_0_0" [id=561, type=get_attr];
-"562 scale_updated_constant28" [id=562, type=get_attr];
-"563 compressed_weight_updated_constant28" [id=563, type=get_attr];
-"564 mul_tensor_28" [id=564, type=mul];
-"565 zero_point_updated_constant28" [id=565, type=get_attr];
-"566 sub_tensor_28" [id=566, type=sub];
-"567 linear_27" [id=567, type=linear];
-"568 dropout_21" [id=568, type=dropout];
-"569 add_14" [id=569, type=add];
-"570 _param_constant88" [id=570, type=get_attr];
-"571 _param_constant89" [id=571, type=get_attr];
-"572 layer_norm_14" [id=572, type=layer_norm];
-"573 transpose_42" [id=573, type=transpose];
-"574 transpose_42_0_0_nncf_smooth_quant_0" [id=574, type=call_module];
-"575 quantize_per_tensor_default_43" [id=575, type=quantize_per_tensor];
-"576 dequantize_per_tensor_default_43" [id=576, type=dequantize_per_tensor];
-"577 _param_constant91_0_0" [id=577, type=get_attr];
-"578 scale_updated_constant29" [id=578, type=get_attr];
-"579 compressed_weight_updated_constant29" [id=579, type=get_attr];
-"580 mul_tensor_29" [id=580, type=mul];
-"581 zero_point_updated_constant29" [id=581, type=get_attr];
-"582 sub_tensor_29" [id=582, type=sub];
-"583 linear_28" [id=583, type=linear];
-"584 unflatten_7" [id=584, type=unflatten];
-"585 unsqueeze_7" [id=585, type=unsqueeze];
-"586 transpose_43" [id=586, type=transpose];
-"587 squeeze_7" [id=587, type=squeeze];
-"588 contiguous_7" [id=588, type=contiguous];
-"589 quantize_per_tensor_default_44" [id=589, type=quantize_per_tensor];
-"590 dequantize_per_tensor_default_44" [id=590, type=dequantize_per_tensor];
-"591 select_21" [id=591, type=select];
-"592 quantize_per_tensor_default_45" [id=592, type=quantize_per_tensor];
-"593 dequantize_per_tensor_default_45" [id=593, type=dequantize_per_tensor];
-"594 select_22" [id=594, type=select];
-"595 select_23" [id=595, type=select];
-"596 view_56" [id=596, type=view];
-"597 transpose_44" [id=597, type=transpose];
-"598 view_57" [id=598, type=view];
-"599 transpose_45" [id=599, type=transpose];
-"600 view_58" [id=600, type=view];
-"601 transpose_46" [id=601, type=transpose];
-"602 view_59" [id=602, type=view];
-"603 view_60" [id=603, type=view];
-"604 view_61" [id=604, type=view];
-"605 scaled_dot_product_attention_7" [id=605, type=scaled_dot_product_attention];
-"606 permute_8" [id=606, type=permute];
-"607 view_62" [id=607, type=view];
-"608 view_62_0_0_nncf_smooth_quant_0" [id=608, type=call_module];
-"609 quantize_per_tensor_default_46" [id=609, type=quantize_per_tensor];
-"610 dequantize_per_tensor_default_46" [id=610, type=dequantize_per_tensor];
-"611 _param_constant93_0_0" [id=611, type=get_attr];
-"612 scale_updated_constant30" [id=612, type=get_attr];
-"613 compressed_weight_updated_constant30" [id=613, type=get_attr];
-"614 mul_tensor_30" [id=614, type=mul];
-"615 zero_point_updated_constant30" [id=615, type=get_attr];
-"616 sub_tensor_30" [id=616, type=sub];
-"617 linear_29" [id=617, type=linear];
-"618 view_63" [id=618, type=view];
-"619 transpose_47" [id=619, type=transpose];
-"620 dropout_22" [id=620, type=dropout];
-"621 add_15" [id=621, type=add];
-"622 _param_constant94" [id=622, type=get_attr];
-"623 _param_constant95" [id=623, type=get_attr];
-"624 layer_norm_15" [id=624, type=layer_norm];
-"625 layer_norm_15_0_0_nncf_smooth_quant_0" [id=625, type=call_module];
-"626 quantize_per_tensor_default_47" [id=626, type=quantize_per_tensor];
-"627 dequantize_per_tensor_default_47" [id=627, type=dequantize_per_tensor];
-"628 _param_constant97_0_0" [id=628, type=get_attr];
-"629 scale_updated_constant31" [id=629, type=get_attr];
-"630 compressed_weight_updated_constant31" [id=630, type=get_attr];
-"631 mul_tensor_31" [id=631, type=mul];
-"632 zero_point_updated_constant31" [id=632, type=get_attr];
-"633 sub_tensor_31" [id=633, type=sub];
-"634 linear_30" [id=634, type=linear];
-"635 gelu_7" [id=635, type=gelu];
-"636 dropout_23" [id=636, type=dropout];
-"637 dropout_23_0_0_nncf_smooth_quant_0" [id=637, type=call_module];
-"638 quantize_per_tensor_default_48" [id=638, type=quantize_per_tensor];
-"639 dequantize_per_tensor_default_48" [id=639, type=dequantize_per_tensor];
-"640 _param_constant99_0_0" [id=640, type=get_attr];
-"641 scale_updated_constant32" [id=641, type=get_attr];
-"642 compressed_weight_updated_constant32" [id=642, type=get_attr];
-"643 mul_tensor_32" [id=643, type=mul];
-"644 zero_point_updated_constant32" [id=644, type=get_attr];
-"645 sub_tensor_32" [id=645, type=sub];
-"646 linear_31" [id=646, type=linear];
-"647 dropout_24" [id=647, type=dropout];
-"648 add_16" [id=648, type=add];
-"649 _param_constant100" [id=649, type=get_attr];
-"650 _param_constant101" [id=650, type=get_attr];
-"651 layer_norm_16" [id=651, type=layer_norm];
-"652 transpose_48" [id=652, type=transpose];
-"653 transpose_48_0_0_nncf_smooth_quant_0" [id=653, type=call_module];
-"654 quantize_per_tensor_default_49" [id=654, type=quantize_per_tensor];
-"655 dequantize_per_tensor_default_49" [id=655, type=dequantize_per_tensor];
-"656 _param_constant103_0_0" [id=656, type=get_attr];
-"657 scale_updated_constant33" [id=657, type=get_attr];
-"658 compressed_weight_updated_constant33" [id=658, type=get_attr];
-"659 mul_tensor_33" [id=659, type=mul];
-"660 zero_point_updated_constant33" [id=660, type=get_attr];
-"661 sub_tensor_33" [id=661, type=sub];
-"662 linear_32" [id=662, type=linear];
-"663 unflatten_8" [id=663, type=unflatten];
-"664 unsqueeze_8" [id=664, type=unsqueeze];
-"665 transpose_49" [id=665, type=transpose];
-"666 squeeze_8" [id=666, type=squeeze];
-"667 contiguous_8" [id=667, type=contiguous];
-"668 quantize_per_tensor_default_50" [id=668, type=quantize_per_tensor];
-"669 dequantize_per_tensor_default_50" [id=669, type=dequantize_per_tensor];
-"670 select_24" [id=670, type=select];
-"671 quantize_per_tensor_default_51" [id=671, type=quantize_per_tensor];
-"672 dequantize_per_tensor_default_51" [id=672, type=dequantize_per_tensor];
-"673 select_25" [id=673, type=select];
-"674 select_26" [id=674, type=select];
-"675 view_64" [id=675, type=view];
-"676 transpose_50" [id=676, type=transpose];
-"677 view_65" [id=677, type=view];
-"678 transpose_51" [id=678, type=transpose];
-"679 view_66" [id=679, type=view];
-"680 transpose_52" [id=680, type=transpose];
-"681 view_67" [id=681, type=view];
-"682 view_68" [id=682, type=view];
-"683 view_69" [id=683, type=view];
-"684 scaled_dot_product_attention_8" [id=684, type=scaled_dot_product_attention];
-"685 permute_9" [id=685, type=permute];
-"686 view_70" [id=686, type=view];
-"687 view_70_0_0_nncf_smooth_quant_0" [id=687, type=call_module];
-"688 quantize_per_tensor_default_52" [id=688, type=quantize_per_tensor];
-"689 dequantize_per_tensor_default_52" [id=689, type=dequantize_per_tensor];
-"690 _param_constant105_0_0" [id=690, type=get_attr];
-"691 scale_updated_constant34" [id=691, type=get_attr];
-"692 compressed_weight_updated_constant34" [id=692, type=get_attr];
-"693 mul_tensor_34" [id=693, type=mul];
-"694 zero_point_updated_constant34" [id=694, type=get_attr];
-"695 sub_tensor_34" [id=695, type=sub];
-"696 linear_33" [id=696, type=linear];
-"697 view_71" [id=697, type=view];
-"698 transpose_53" [id=698, type=transpose];
-"699 dropout_25" [id=699, type=dropout];
-"700 add_17" [id=700, type=add];
-"701 _param_constant106" [id=701, type=get_attr];
-"702 _param_constant107" [id=702, type=get_attr];
-"703 layer_norm_17" [id=703, type=layer_norm];
-"704 layer_norm_17_0_0_nncf_smooth_quant_0" [id=704, type=call_module];
-"705 quantize_per_tensor_default_53" [id=705, type=quantize_per_tensor];
-"706 dequantize_per_tensor_default_53" [id=706, type=dequantize_per_tensor];
-"707 _param_constant109_0_0" [id=707, type=get_attr];
-"708 scale_updated_constant35" [id=708, type=get_attr];
-"709 compressed_weight_updated_constant35" [id=709, type=get_attr];
-"710 mul_tensor_35" [id=710, type=mul];
-"711 zero_point_updated_constant35" [id=711, type=get_attr];
-"712 sub_tensor_35" [id=712, type=sub];
-"713 linear_34" [id=713, type=linear];
-"714 gelu_8" [id=714, type=gelu];
-"715 dropout_26" [id=715, type=dropout];
-"716 dropout_26_0_0_nncf_smooth_quant_0" [id=716, type=call_module];
-"717 quantize_per_tensor_default_54" [id=717, type=quantize_per_tensor];
-"718 dequantize_per_tensor_default_54" [id=718, type=dequantize_per_tensor];
-"719 _param_constant111_0_0" [id=719, type=get_attr];
-"720 scale_updated_constant36" [id=720, type=get_attr];
-"721 compressed_weight_updated_constant36" [id=721, type=get_attr];
-"722 mul_tensor_36" [id=722, type=mul];
-"723 zero_point_updated_constant36" [id=723, type=get_attr];
-"724 sub_tensor_36" [id=724, type=sub];
-"725 linear_35" [id=725, type=linear];
-"726 dropout_27" [id=726, type=dropout];
-"727 add_18" [id=727, type=add];
-"728 _param_constant112" [id=728, type=get_attr];
-"729 _param_constant113" [id=729, type=get_attr];
-"730 layer_norm_18" [id=730, type=layer_norm];
-"731 transpose_54" [id=731, type=transpose];
-"732 transpose_54_0_0_nncf_smooth_quant_0" [id=732, type=call_module];
-"733 quantize_per_tensor_default_55" [id=733, type=quantize_per_tensor];
-"734 dequantize_per_tensor_default_55" [id=734, type=dequantize_per_tensor];
-"735 _param_constant115_0_0" [id=735, type=get_attr];
-"736 scale_updated_constant37" [id=736, type=get_attr];
-"737 compressed_weight_updated_constant37" [id=737, type=get_attr];
-"738 mul_tensor_37" [id=738, type=mul];
-"739 zero_point_updated_constant37" [id=739, type=get_attr];
-"740 sub_tensor_37" [id=740, type=sub];
-"741 linear_36" [id=741, type=linear];
-"742 unflatten_9" [id=742, type=unflatten];
-"743 unsqueeze_9" [id=743, type=unsqueeze];
-"744 transpose_55" [id=744, type=transpose];
-"745 squeeze_9" [id=745, type=squeeze];
-"746 contiguous_9" [id=746, type=contiguous];
-"747 quantize_per_tensor_default_56" [id=747, type=quantize_per_tensor];
-"748 dequantize_per_tensor_default_56" [id=748, type=dequantize_per_tensor];
-"749 select_27" [id=749, type=select];
-"750 quantize_per_tensor_default_57" [id=750, type=quantize_per_tensor];
-"751 dequantize_per_tensor_default_57" [id=751, type=dequantize_per_tensor];
-"752 select_28" [id=752, type=select];
-"753 select_29" [id=753, type=select];
-"754 view_72" [id=754, type=view];
-"755 transpose_56" [id=755, type=transpose];
-"756 view_73" [id=756, type=view];
-"757 transpose_57" [id=757, type=transpose];
-"758 view_74" [id=758, type=view];
-"759 transpose_58" [id=759, type=transpose];
-"760 view_75" [id=760, type=view];
-"761 view_76" [id=761, type=view];
-"762 view_77" [id=762, type=view];
-"763 scaled_dot_product_attention_9" [id=763, type=scaled_dot_product_attention];
-"764 permute_10" [id=764, type=permute];
-"765 view_78" [id=765, type=view];
-"766 view_78_0_0_nncf_smooth_quant_0" [id=766, type=call_module];
-"767 quantize_per_tensor_default_58" [id=767, type=quantize_per_tensor];
-"768 dequantize_per_tensor_default_58" [id=768, type=dequantize_per_tensor];
-"769 _param_constant117_0_0" [id=769, type=get_attr];
-"770 scale_updated_constant38" [id=770, type=get_attr];
-"771 compressed_weight_updated_constant38" [id=771, type=get_attr];
-"772 mul_tensor_38" [id=772, type=mul];
-"773 zero_point_updated_constant38" [id=773, type=get_attr];
-"774 sub_tensor_38" [id=774, type=sub];
-"775 linear_37" [id=775, type=linear];
-"776 view_79" [id=776, type=view];
-"777 transpose_59" [id=777, type=transpose];
-"778 dropout_28" [id=778, type=dropout];
-"779 add_19" [id=779, type=add];
-"780 _param_constant118" [id=780, type=get_attr];
-"781 _param_constant119" [id=781, type=get_attr];
-"782 layer_norm_19" [id=782, type=layer_norm];
-"783 layer_norm_19_0_0_nncf_smooth_quant_0" [id=783, type=call_module];
-"784 quantize_per_tensor_default_59" [id=784, type=quantize_per_tensor];
-"785 dequantize_per_tensor_default_59" [id=785, type=dequantize_per_tensor];
-"786 _param_constant121_0_0" [id=786, type=get_attr];
-"787 scale_updated_constant39" [id=787, type=get_attr];
-"788 compressed_weight_updated_constant39" [id=788, type=get_attr];
-"789 mul_tensor_39" [id=789, type=mul];
-"790 zero_point_updated_constant39" [id=790, type=get_attr];
-"791 sub_tensor_39" [id=791, type=sub];
-"792 linear_38" [id=792, type=linear];
-"793 gelu_9" [id=793, type=gelu];
-"794 dropout_29" [id=794, type=dropout];
-"795 dropout_29_0_0_nncf_smooth_quant_0" [id=795, type=call_module];
-"796 quantize_per_tensor_default_60" [id=796, type=quantize_per_tensor];
-"797 dequantize_per_tensor_default_60" [id=797, type=dequantize_per_tensor];
-"798 _param_constant123_0_0" [id=798, type=get_attr];
-"799 scale_updated_constant40" [id=799, type=get_attr];
-"800 compressed_weight_updated_constant40" [id=800, type=get_attr];
-"801 mul_tensor_40" [id=801, type=mul];
-"802 zero_point_updated_constant40" [id=802, type=get_attr];
-"803 sub_tensor_40" [id=803, type=sub];
-"804 linear_39" [id=804, type=linear];
-"805 dropout_30" [id=805, type=dropout];
-"806 add_20" [id=806, type=add];
-"807 _param_constant124" [id=807, type=get_attr];
-"808 _param_constant125" [id=808, type=get_attr];
-"809 layer_norm_20" [id=809, type=layer_norm];
-"810 transpose_60" [id=810, type=transpose];
-"811 transpose_60_0_0_nncf_smooth_quant_0" [id=811, type=call_module];
-"812 quantize_per_tensor_default_61" [id=812, type=quantize_per_tensor];
-"813 dequantize_per_tensor_default_61" [id=813, type=dequantize_per_tensor];
-"814 _param_constant127_0_0" [id=814, type=get_attr];
-"815 scale_updated_constant41" [id=815, type=get_attr];
-"816 compressed_weight_updated_constant41" [id=816, type=get_attr];
-"817 mul_tensor_41" [id=817, type=mul];
-"818 zero_point_updated_constant41" [id=818, type=get_attr];
-"819 sub_tensor_41" [id=819, type=sub];
-"820 linear_40" [id=820, type=linear];
-"821 unflatten_10" [id=821, type=unflatten];
-"822 unsqueeze_10" [id=822, type=unsqueeze];
-"823 transpose_61" [id=823, type=transpose];
-"824 squeeze_10" [id=824, type=squeeze];
-"825 contiguous_10" [id=825, type=contiguous];
-"826 quantize_per_tensor_default_62" [id=826, type=quantize_per_tensor];
-"827 dequantize_per_tensor_default_62" [id=827, type=dequantize_per_tensor];
-"828 select_30" [id=828, type=select];
-"829 quantize_per_tensor_default_63" [id=829, type=quantize_per_tensor];
-"830 dequantize_per_tensor_default_63" [id=830, type=dequantize_per_tensor];
-"831 select_31" [id=831, type=select];
-"832 select_32" [id=832, type=select];
-"833 view_80" [id=833, type=view];
-"834 transpose_62" [id=834, type=transpose];
-"835 view_81" [id=835, type=view];
-"836 transpose_63" [id=836, type=transpose];
-"837 view_82" [id=837, type=view];
-"838 transpose_64" [id=838, type=transpose];
-"839 view_83" [id=839, type=view];
-"840 view_84" [id=840, type=view];
-"841 view_85" [id=841, type=view];
-"842 scaled_dot_product_attention_10" [id=842, type=scaled_dot_product_attention];
-"843 permute_11" [id=843, type=permute];
-"844 view_86" [id=844, type=view];
-"845 view_86_0_0_nncf_smooth_quant_0" [id=845, type=call_module];
-"846 quantize_per_tensor_default_64" [id=846, type=quantize_per_tensor];
-"847 dequantize_per_tensor_default_64" [id=847, type=dequantize_per_tensor];
-"848 _param_constant129_0_0" [id=848, type=get_attr];
-"849 scale_updated_constant42" [id=849, type=get_attr];
-"850 compressed_weight_updated_constant42" [id=850, type=get_attr];
-"851 mul_tensor_42" [id=851, type=mul];
-"852 zero_point_updated_constant42" [id=852, type=get_attr];
-"853 sub_tensor_42" [id=853, type=sub];
-"854 linear_41" [id=854, type=linear];
-"855 view_87" [id=855, type=view];
-"856 transpose_65" [id=856, type=transpose];
-"857 dropout_31" [id=857, type=dropout];
-"858 add_21" [id=858, type=add];
-"859 _param_constant130" [id=859, type=get_attr];
-"860 _param_constant131" [id=860, type=get_attr];
-"861 layer_norm_21" [id=861, type=layer_norm];
-"862 layer_norm_21_0_0_nncf_smooth_quant_0" [id=862, type=call_module];
-"863 quantize_per_tensor_default_65" [id=863, type=quantize_per_tensor];
-"864 dequantize_per_tensor_default_65" [id=864, type=dequantize_per_tensor];
-"865 _param_constant133_0_0" [id=865, type=get_attr];
-"866 scale_updated_constant43" [id=866, type=get_attr];
-"867 compressed_weight_updated_constant43" [id=867, type=get_attr];
-"868 mul_tensor_43" [id=868, type=mul];
-"869 zero_point_updated_constant43" [id=869, type=get_attr];
-"870 sub_tensor_43" [id=870, type=sub];
-"871 linear_42" [id=871, type=linear];
-"872 gelu_10" [id=872, type=gelu];
-"873 dropout_32" [id=873, type=dropout];
-"874 dropout_32_0_0_nncf_smooth_quant_0" [id=874, type=call_module];
-"875 quantize_per_tensor_default_66" [id=875, type=quantize_per_tensor];
-"876 dequantize_per_tensor_default_66" [id=876, type=dequantize_per_tensor];
-"877 _param_constant135_0_0" [id=877, type=get_attr];
-"878 scale_updated_constant44" [id=878, type=get_attr];
-"879 compressed_weight_updated_constant44" [id=879, type=get_attr];
-"880 mul_tensor_44" [id=880, type=mul];
-"881 zero_point_updated_constant44" [id=881, type=get_attr];
-"882 sub_tensor_44" [id=882, type=sub];
-"883 linear_43" [id=883, type=linear];
-"884 dropout_33" [id=884, type=dropout];
-"885 add_22" [id=885, type=add];
-"886 _param_constant136" [id=886, type=get_attr];
-"887 _param_constant137" [id=887, type=get_attr];
-"888 layer_norm_22" [id=888, type=layer_norm];
-"889 transpose_66" [id=889, type=transpose];
-"890 transpose_66_0_0_nncf_smooth_quant_0" [id=890, type=call_module];
-"891 quantize_per_tensor_default_67" [id=891, type=quantize_per_tensor];
-"892 dequantize_per_tensor_default_67" [id=892, type=dequantize_per_tensor];
-"893 _param_constant139_0_0" [id=893, type=get_attr];
-"894 scale_updated_constant45" [id=894, type=get_attr];
-"895 compressed_weight_updated_constant45" [id=895, type=get_attr];
-"896 mul_tensor_45" [id=896, type=mul];
-"897 zero_point_updated_constant45" [id=897, type=get_attr];
-"898 sub_tensor_45" [id=898, type=sub];
-"899 linear_44" [id=899, type=linear];
-"900 unflatten_11" [id=900, type=unflatten];
-"901 unsqueeze_11" [id=901, type=unsqueeze];
-"902 transpose_67" [id=902, type=transpose];
-"903 squeeze_11" [id=903, type=squeeze];
-"904 contiguous_11" [id=904, type=contiguous];
-"905 quantize_per_tensor_default_68" [id=905, type=quantize_per_tensor];
-"906 dequantize_per_tensor_default_68" [id=906, type=dequantize_per_tensor];
-"907 select_33" [id=907, type=select];
-"908 quantize_per_tensor_default_69" [id=908, type=quantize_per_tensor];
-"909 dequantize_per_tensor_default_69" [id=909, type=dequantize_per_tensor];
-"910 select_34" [id=910, type=select];
-"911 select_35" [id=911, type=select];
-"912 view_88" [id=912, type=view];
-"913 transpose_68" [id=913, type=transpose];
-"914 view_89" [id=914, type=view];
-"915 transpose_69" [id=915, type=transpose];
-"916 view_90" [id=916, type=view];
-"917 transpose_70" [id=917, type=transpose];
-"918 view_91" [id=918, type=view];
-"919 view_92" [id=919, type=view];
-"920 view_93" [id=920, type=view];
-"921 scaled_dot_product_attention_11" [id=921, type=scaled_dot_product_attention];
-"922 permute_12" [id=922, type=permute];
-"923 view_94" [id=923, type=view];
-"924 view_94_0_0_nncf_smooth_quant_0" [id=924, type=call_module];
-"925 quantize_per_tensor_default_70" [id=925, type=quantize_per_tensor];
-"926 dequantize_per_tensor_default_70" [id=926, type=dequantize_per_tensor];
-"927 _param_constant141_0_0" [id=927, type=get_attr];
-"928 scale_updated_constant46" [id=928, type=get_attr];
-"929 compressed_weight_updated_constant46" [id=929, type=get_attr];
-"930 mul_tensor_46" [id=930, type=mul];
-"931 zero_point_updated_constant46" [id=931, type=get_attr];
-"932 sub_tensor_46" [id=932, type=sub];
-"933 linear_45" [id=933, type=linear];
-"934 view_95" [id=934, type=view];
-"935 transpose_71" [id=935, type=transpose];
-"936 dropout_34" [id=936, type=dropout];
-"937 add_23" [id=937, type=add];
-"938 _param_constant142" [id=938, type=get_attr];
-"939 _param_constant143" [id=939, type=get_attr];
-"940 layer_norm_23" [id=940, type=layer_norm];
-"941 layer_norm_23_0_0_nncf_smooth_quant_0" [id=941, type=call_module];
-"942 quantize_per_tensor_default_71" [id=942, type=quantize_per_tensor];
-"943 dequantize_per_tensor_default_71" [id=943, type=dequantize_per_tensor];
-"944 _param_constant145_0_0" [id=944, type=get_attr];
-"945 scale_updated_constant47" [id=945, type=get_attr];
-"946 compressed_weight_updated_constant47" [id=946, type=get_attr];
-"947 mul_tensor_47" [id=947, type=mul];
-"948 zero_point_updated_constant47" [id=948, type=get_attr];
-"949 sub_tensor_47" [id=949, type=sub];
-"950 linear_46" [id=950, type=linear];
-"951 gelu_11" [id=951, type=gelu];
-"952 dropout_35" [id=952, type=dropout];
-"953 dropout_35_0_0_nncf_smooth_quant_0" [id=953, type=call_module];
-"954 quantize_per_tensor_default_72" [id=954, type=quantize_per_tensor];
-"955 dequantize_per_tensor_default_72" [id=955, type=dequantize_per_tensor];
-"956 _param_constant147_0_0" [id=956, type=get_attr];
-"957 scale_updated_constant48" [id=957, type=get_attr];
-"958 compressed_weight_updated_constant48" [id=958, type=get_attr];
-"959 mul_tensor_48" [id=959, type=mul];
-"960 zero_point_updated_constant48" [id=960, type=get_attr];
-"961 sub_tensor_48" [id=961, type=sub];
-"962 linear_47" [id=962, type=linear];
-"963 dropout_36" [id=963, type=dropout];
-"964 add_24" [id=964, type=add];
-"965 _param_constant148" [id=965, type=get_attr];
-"966 _param_constant149" [id=966, type=get_attr];
+"0 encoder_pos_embedding" [id=0, type=get_attr];
+"1 encoder_layers_encoder_layer_0_ln_1_weight" [id=1, type=get_attr];
+"2 encoder_layers_encoder_layer_0_ln_1_bias" [id=2, type=get_attr];
+"3 encoder_layers_encoder_layer_0_ln_2_weight" [id=3, type=get_attr];
+"4 encoder_layers_encoder_layer_0_ln_2_bias" [id=4, type=get_attr];
+"5 encoder_layers_encoder_layer_1_ln_1_weight" [id=5, type=get_attr];
+"6 encoder_layers_encoder_layer_1_ln_1_bias" [id=6, type=get_attr];
+"7 encoder_layers_encoder_layer_1_ln_2_weight" [id=7, type=get_attr];
+"8 encoder_layers_encoder_layer_1_ln_2_bias" [id=8, type=get_attr];
+"9 encoder_layers_encoder_layer_2_ln_1_weight" [id=9, type=get_attr];
+"10 encoder_layers_encoder_layer_2_ln_1_bias" [id=10, type=get_attr];
+"11 encoder_layers_encoder_layer_2_ln_2_weight" [id=11, type=get_attr];
+"12 encoder_layers_encoder_layer_2_ln_2_bias" [id=12, type=get_attr];
+"13 encoder_layers_encoder_layer_3_ln_1_weight" [id=13, type=get_attr];
+"14 encoder_layers_encoder_layer_3_ln_1_bias" [id=14, type=get_attr];
+"15 encoder_layers_encoder_layer_3_ln_2_weight" [id=15, type=get_attr];
+"16 encoder_layers_encoder_layer_3_ln_2_bias" [id=16, type=get_attr];
+"17 encoder_layers_encoder_layer_4_ln_1_weight" [id=17, type=get_attr];
+"18 encoder_layers_encoder_layer_4_ln_1_bias" [id=18, type=get_attr];
+"19 encoder_layers_encoder_layer_4_ln_2_weight" [id=19, type=get_attr];
+"20 encoder_layers_encoder_layer_4_ln_2_bias" [id=20, type=get_attr];
+"21 encoder_layers_encoder_layer_5_ln_1_weight" [id=21, type=get_attr];
+"22 encoder_layers_encoder_layer_5_ln_1_bias" [id=22, type=get_attr];
+"23 encoder_layers_encoder_layer_5_ln_2_weight" [id=23, type=get_attr];
+"24 encoder_layers_encoder_layer_5_ln_2_bias" [id=24, type=get_attr];
+"25 encoder_layers_encoder_layer_6_ln_1_weight" [id=25, type=get_attr];
+"26 encoder_layers_encoder_layer_6_ln_1_bias" [id=26, type=get_attr];
+"27 encoder_layers_encoder_layer_6_ln_2_weight" [id=27, type=get_attr];
+"28 encoder_layers_encoder_layer_6_ln_2_bias" [id=28, type=get_attr];
+"29 encoder_layers_encoder_layer_7_ln_1_weight" [id=29, type=get_attr];
+"30 encoder_layers_encoder_layer_7_ln_1_bias" [id=30, type=get_attr];
+"31 encoder_layers_encoder_layer_7_ln_2_weight" [id=31, type=get_attr];
+"32 encoder_layers_encoder_layer_7_ln_2_bias" [id=32, type=get_attr];
+"33 encoder_layers_encoder_layer_8_ln_1_weight" [id=33, type=get_attr];
+"34 encoder_layers_encoder_layer_8_ln_1_bias" [id=34, type=get_attr];
+"35 encoder_layers_encoder_layer_8_ln_2_weight" [id=35, type=get_attr];
+"36 encoder_layers_encoder_layer_8_ln_2_bias" [id=36, type=get_attr];
+"37 encoder_layers_encoder_layer_9_ln_1_weight" [id=37, type=get_attr];
+"38 encoder_layers_encoder_layer_9_ln_1_bias" [id=38, type=get_attr];
+"39 encoder_layers_encoder_layer_9_ln_2_weight" [id=39, type=get_attr];
+"40 encoder_layers_encoder_layer_9_ln_2_bias" [id=40, type=get_attr];
+"41 encoder_layers_encoder_layer_10_ln_1_weight" [id=41, type=get_attr];
+"42 encoder_layers_encoder_layer_10_ln_1_bias" [id=42, type=get_attr];
+"43 encoder_layers_encoder_layer_10_ln_2_weight" [id=43, type=get_attr];
+"44 encoder_layers_encoder_layer_10_ln_2_bias" [id=44, type=get_attr];
+"45 encoder_layers_encoder_layer_11_ln_1_weight" [id=45, type=get_attr];
+"46 encoder_layers_encoder_layer_11_ln_1_bias" [id=46, type=get_attr];
+"47 encoder_layers_encoder_layer_11_ln_2_weight" [id=47, type=get_attr];
+"48 encoder_layers_encoder_layer_11_ln_2_bias" [id=48, type=get_attr];
+"49 encoder_ln_weight" [id=49, type=get_attr];
+"50 encoder_ln_bias" [id=50, type=get_attr];
+"51 x" [id=51, type=input];
+"52 quantize_per_tensor_default" [id=52, type=quantize_per_tensor];
+"53 dequantize_per_tensor_default" [id=53, type=dequantize_per_tensor];
+"54 conv_proj_bias_0_0" [id=54, type=get_attr];
+"55 scale_updated_constant0" [id=55, type=get_attr];
+"56 compressed_weight_updated_constant0" [id=56, type=get_attr];
+"57 mul_tensor" [id=57, type=mul];
+"58 zero_point_updated_constant0" [id=58, type=get_attr];
+"59 sub_tensor" [id=59, type=sub];
+"60 conv2d" [id=60, type=conv2d];
+"61 reshape" [id=61, type=reshape];
+"62 permute" [id=62, type=permute];
+"63 _frozen_param0" [id=63, type=get_attr];
+"64 cat" [id=64, type=cat];
+"65 add" [id=65, type=add];
+"66 dropout" [id=66, type=dropout];
+"67 layer_norm" [id=67, type=layer_norm];
+"68 transpose" [id=68, type=transpose];
+"69 transpose_0_0_nncf_smooth_quant_0" [id=69, type=call_module];
+"70 quantize_per_tensor_default_1" [id=70, type=quantize_per_tensor];
+"71 dequantize_per_tensor_default_1" [id=71, type=dequantize_per_tensor];
+"72 encoder_layers_encoder_layer_0_self_attention_in_proj_bias_0_0" [id=72, type=get_attr];
+"73 scale_updated_constant1" [id=73, type=get_attr];
+"74 compressed_weight_updated_constant1" [id=74, type=get_attr];
+"75 mul_tensor_1" [id=75, type=mul];
+"76 zero_point_updated_constant1" [id=76, type=get_attr];
+"77 sub_tensor_1" [id=77, type=sub];
+"78 linear" [id=78, type=linear];
+"79 unflatten" [id=79, type=unflatten];
+"80 unsqueeze" [id=80, type=unsqueeze];
+"81 transpose_1" [id=81, type=transpose];
+"82 squeeze" [id=82, type=squeeze];
+"83 contiguous" [id=83, type=contiguous];
+"84 quantize_per_tensor_default_2" [id=84, type=quantize_per_tensor];
+"85 dequantize_per_tensor_default_2" [id=85, type=dequantize_per_tensor];
+"86 select" [id=86, type=select];
+"87 quantize_per_tensor_default_3" [id=87, type=quantize_per_tensor];
+"88 dequantize_per_tensor_default_3" [id=88, type=dequantize_per_tensor];
+"89 select_1" [id=89, type=select];
+"90 select_2" [id=90, type=select];
+"91 view" [id=91, type=view];
+"92 transpose_2" [id=92, type=transpose];
+"93 view_1" [id=93, type=view];
+"94 transpose_3" [id=94, type=transpose];
+"95 view_2" [id=95, type=view];
+"96 transpose_4" [id=96, type=transpose];
+"97 view_3" [id=97, type=view];
+"98 view_4" [id=98, type=view];
+"99 view_5" [id=99, type=view];
+"100 scaled_dot_product_attention" [id=100, type=scaled_dot_product_attention];
+"101 permute_1" [id=101, type=permute];
+"102 view_6" [id=102, type=view];
+"103 view_6_0_0_nncf_smooth_quant_0" [id=103, type=call_module];
+"104 quantize_per_tensor_default_4" [id=104, type=quantize_per_tensor];
+"105 dequantize_per_tensor_default_4" [id=105, type=dequantize_per_tensor];
+"106 encoder_layers_encoder_layer_0_self_attention_out_proj_bias_0_0" [id=106, type=get_attr];
+"107 scale_updated_constant2" [id=107, type=get_attr];
+"108 compressed_weight_updated_constant2" [id=108, type=get_attr];
+"109 mul_tensor_2" [id=109, type=mul];
+"110 zero_point_updated_constant2" [id=110, type=get_attr];
+"111 sub_tensor_2" [id=111, type=sub];
+"112 linear_1" [id=112, type=linear];
+"113 view_7" [id=113, type=view];
+"114 transpose_5" [id=114, type=transpose];
+"115 dropout_1" [id=115, type=dropout];
+"116 add_1" [id=116, type=add];
+"117 layer_norm_1" [id=117, type=layer_norm];
+"118 layer_norm_1_0_0_nncf_smooth_quant_0" [id=118, type=call_module];
+"119 quantize_per_tensor_default_5" [id=119, type=quantize_per_tensor];
+"120 dequantize_per_tensor_default_5" [id=120, type=dequantize_per_tensor];
+"121 encoder_layers_encoder_layer_0_mlp_0_bias_0_0" [id=121, type=get_attr];
+"122 scale_updated_constant3" [id=122, type=get_attr];
+"123 compressed_weight_updated_constant3" [id=123, type=get_attr];
+"124 mul_tensor_3" [id=124, type=mul];
+"125 zero_point_updated_constant3" [id=125, type=get_attr];
+"126 sub_tensor_3" [id=126, type=sub];
+"127 linear_2" [id=127, type=linear];
+"128 gelu" [id=128, type=gelu];
+"129 dropout_2" [id=129, type=dropout];
+"130 dropout_2_0_0_nncf_smooth_quant_0" [id=130, type=call_module];
+"131 quantize_per_tensor_default_6" [id=131, type=quantize_per_tensor];
+"132 dequantize_per_tensor_default_6" [id=132, type=dequantize_per_tensor];
+"133 encoder_layers_encoder_layer_0_mlp_3_bias_0_0" [id=133, type=get_attr];
+"134 scale_updated_constant4" [id=134, type=get_attr];
+"135 compressed_weight_updated_constant4" [id=135, type=get_attr];
+"136 mul_tensor_4" [id=136, type=mul];
+"137 zero_point_updated_constant4" [id=137, type=get_attr];
+"138 sub_tensor_4" [id=138, type=sub];
+"139 linear_3" [id=139, type=linear];
+"140 dropout_3" [id=140, type=dropout];
+"141 add_2" [id=141, type=add];
+"142 layer_norm_2" [id=142, type=layer_norm];
+"143 transpose_6" [id=143, type=transpose];
+"144 transpose_6_0_0_nncf_smooth_quant_0" [id=144, type=call_module];
+"145 quantize_per_tensor_default_7" [id=145, type=quantize_per_tensor];
+"146 dequantize_per_tensor_default_7" [id=146, type=dequantize_per_tensor];
+"147 encoder_layers_encoder_layer_1_self_attention_in_proj_bias_0_0" [id=147, type=get_attr];
+"148 scale_updated_constant5" [id=148, type=get_attr];
+"149 compressed_weight_updated_constant5" [id=149, type=get_attr];
+"150 mul_tensor_5" [id=150, type=mul];
+"151 zero_point_updated_constant5" [id=151, type=get_attr];
+"152 sub_tensor_5" [id=152, type=sub];
+"153 linear_4" [id=153, type=linear];
+"154 unflatten_1" [id=154, type=unflatten];
+"155 unsqueeze_1" [id=155, type=unsqueeze];
+"156 transpose_7" [id=156, type=transpose];
+"157 squeeze_1" [id=157, type=squeeze];
+"158 contiguous_1" [id=158, type=contiguous];
+"159 quantize_per_tensor_default_8" [id=159, type=quantize_per_tensor];
+"160 dequantize_per_tensor_default_8" [id=160, type=dequantize_per_tensor];
+"161 select_3" [id=161, type=select];
+"162 quantize_per_tensor_default_9" [id=162, type=quantize_per_tensor];
+"163 dequantize_per_tensor_default_9" [id=163, type=dequantize_per_tensor];
+"164 select_4" [id=164, type=select];
+"165 select_5" [id=165, type=select];
+"166 view_8" [id=166, type=view];
+"167 transpose_8" [id=167, type=transpose];
+"168 view_9" [id=168, type=view];
+"169 transpose_9" [id=169, type=transpose];
+"170 view_10" [id=170, type=view];
+"171 transpose_10" [id=171, type=transpose];
+"172 view_11" [id=172, type=view];
+"173 view_12" [id=173, type=view];
+"174 view_13" [id=174, type=view];
+"175 scaled_dot_product_attention_1" [id=175, type=scaled_dot_product_attention];
+"176 permute_2" [id=176, type=permute];
+"177 view_14" [id=177, type=view];
+"178 view_14_0_0_nncf_smooth_quant_0" [id=178, type=call_module];
+"179 quantize_per_tensor_default_10" [id=179, type=quantize_per_tensor];
+"180 dequantize_per_tensor_default_10" [id=180, type=dequantize_per_tensor];
+"181 encoder_layers_encoder_layer_1_self_attention_out_proj_bias_0_0" [id=181, type=get_attr];
+"182 scale_updated_constant6" [id=182, type=get_attr];
+"183 compressed_weight_updated_constant6" [id=183, type=get_attr];
+"184 mul_tensor_6" [id=184, type=mul];
+"185 zero_point_updated_constant6" [id=185, type=get_attr];
+"186 sub_tensor_6" [id=186, type=sub];
+"187 linear_5" [id=187, type=linear];
+"188 view_15" [id=188, type=view];
+"189 transpose_11" [id=189, type=transpose];
+"190 dropout_4" [id=190, type=dropout];
+"191 add_3" [id=191, type=add];
+"192 layer_norm_3" [id=192, type=layer_norm];
+"193 layer_norm_3_0_0_nncf_smooth_quant_0" [id=193, type=call_module];
+"194 quantize_per_tensor_default_11" [id=194, type=quantize_per_tensor];
+"195 dequantize_per_tensor_default_11" [id=195, type=dequantize_per_tensor];
+"196 encoder_layers_encoder_layer_1_mlp_0_bias_0_0" [id=196, type=get_attr];
+"197 scale_updated_constant7" [id=197, type=get_attr];
+"198 compressed_weight_updated_constant7" [id=198, type=get_attr];
+"199 mul_tensor_7" [id=199, type=mul];
+"200 zero_point_updated_constant7" [id=200, type=get_attr];
+"201 sub_tensor_7" [id=201, type=sub];
+"202 linear_6" [id=202, type=linear];
+"203 gelu_1" [id=203, type=gelu];
+"204 dropout_5" [id=204, type=dropout];
+"205 dropout_5_0_0_nncf_smooth_quant_0" [id=205, type=call_module];
+"206 quantize_per_tensor_default_12" [id=206, type=quantize_per_tensor];
+"207 dequantize_per_tensor_default_12" [id=207, type=dequantize_per_tensor];
+"208 encoder_layers_encoder_layer_1_mlp_3_bias_0_0" [id=208, type=get_attr];
+"209 scale_updated_constant8" [id=209, type=get_attr];
+"210 compressed_weight_updated_constant8" [id=210, type=get_attr];
+"211 mul_tensor_8" [id=211, type=mul];
+"212 zero_point_updated_constant8" [id=212, type=get_attr];
+"213 sub_tensor_8" [id=213, type=sub];
+"214 linear_7" [id=214, type=linear];
+"215 dropout_6" [id=215, type=dropout];
+"216 add_4" [id=216, type=add];
+"217 layer_norm_4" [id=217, type=layer_norm];
+"218 transpose_12" [id=218, type=transpose];
+"219 transpose_12_0_0_nncf_smooth_quant_0" [id=219, type=call_module];
+"220 quantize_per_tensor_default_13" [id=220, type=quantize_per_tensor];
+"221 dequantize_per_tensor_default_13" [id=221, type=dequantize_per_tensor];
+"222 encoder_layers_encoder_layer_2_self_attention_in_proj_bias_0_0" [id=222, type=get_attr];
+"223 scale_updated_constant9" [id=223, type=get_attr];
+"224 compressed_weight_updated_constant9" [id=224, type=get_attr];
+"225 mul_tensor_9" [id=225, type=mul];
+"226 zero_point_updated_constant9" [id=226, type=get_attr];
+"227 sub_tensor_9" [id=227, type=sub];
+"228 linear_8" [id=228, type=linear];
+"229 unflatten_2" [id=229, type=unflatten];
+"230 unsqueeze_2" [id=230, type=unsqueeze];
+"231 transpose_13" [id=231, type=transpose];
+"232 squeeze_2" [id=232, type=squeeze];
+"233 contiguous_2" [id=233, type=contiguous];
+"234 quantize_per_tensor_default_14" [id=234, type=quantize_per_tensor];
+"235 dequantize_per_tensor_default_14" [id=235, type=dequantize_per_tensor];
+"236 select_6" [id=236, type=select];
+"237 quantize_per_tensor_default_15" [id=237, type=quantize_per_tensor];
+"238 dequantize_per_tensor_default_15" [id=238, type=dequantize_per_tensor];
+"239 select_7" [id=239, type=select];
+"240 select_8" [id=240, type=select];
+"241 view_16" [id=241, type=view];
+"242 transpose_14" [id=242, type=transpose];
+"243 view_17" [id=243, type=view];
+"244 transpose_15" [id=244, type=transpose];
+"245 view_18" [id=245, type=view];
+"246 transpose_16" [id=246, type=transpose];
+"247 view_19" [id=247, type=view];
+"248 view_20" [id=248, type=view];
+"249 view_21" [id=249, type=view];
+"250 scaled_dot_product_attention_2" [id=250, type=scaled_dot_product_attention];
+"251 permute_3" [id=251, type=permute];
+"252 view_22" [id=252, type=view];
+"253 view_22_0_0_nncf_smooth_quant_0" [id=253, type=call_module];
+"254 quantize_per_tensor_default_16" [id=254, type=quantize_per_tensor];
+"255 dequantize_per_tensor_default_16" [id=255, type=dequantize_per_tensor];
+"256 encoder_layers_encoder_layer_2_self_attention_out_proj_bias_0_0" [id=256, type=get_attr];
+"257 scale_updated_constant10" [id=257, type=get_attr];
+"258 compressed_weight_updated_constant10" [id=258, type=get_attr];
+"259 mul_tensor_10" [id=259, type=mul];
+"260 zero_point_updated_constant10" [id=260, type=get_attr];
+"261 sub_tensor_10" [id=261, type=sub];
+"262 linear_9" [id=262, type=linear];
+"263 view_23" [id=263, type=view];
+"264 transpose_17" [id=264, type=transpose];
+"265 dropout_7" [id=265, type=dropout];
+"266 add_5" [id=266, type=add];
+"267 layer_norm_5" [id=267, type=layer_norm];
+"268 layer_norm_5_0_0_nncf_smooth_quant_0" [id=268, type=call_module];
+"269 quantize_per_tensor_default_17" [id=269, type=quantize_per_tensor];
+"270 dequantize_per_tensor_default_17" [id=270, type=dequantize_per_tensor];
+"271 encoder_layers_encoder_layer_2_mlp_0_bias_0_0" [id=271, type=get_attr];
+"272 scale_updated_constant11" [id=272, type=get_attr];
+"273 compressed_weight_updated_constant11" [id=273, type=get_attr];
+"274 mul_tensor_11" [id=274, type=mul];
+"275 zero_point_updated_constant11" [id=275, type=get_attr];
+"276 sub_tensor_11" [id=276, type=sub];
+"277 linear_10" [id=277, type=linear];
+"278 gelu_2" [id=278, type=gelu];
+"279 dropout_8" [id=279, type=dropout];
+"280 dropout_8_0_0_nncf_smooth_quant_0" [id=280, type=call_module];
+"281 quantize_per_tensor_default_18" [id=281, type=quantize_per_tensor];
+"282 dequantize_per_tensor_default_18" [id=282, type=dequantize_per_tensor];
+"283 encoder_layers_encoder_layer_2_mlp_3_bias_0_0" [id=283, type=get_attr];
+"284 scale_updated_constant12" [id=284, type=get_attr];
+"285 compressed_weight_updated_constant12" [id=285, type=get_attr];
+"286 mul_tensor_12" [id=286, type=mul];
+"287 zero_point_updated_constant12" [id=287, type=get_attr];
+"288 sub_tensor_12" [id=288, type=sub];
+"289 linear_11" [id=289, type=linear];
+"290 dropout_9" [id=290, type=dropout];
+"291 add_6" [id=291, type=add];
+"292 layer_norm_6" [id=292, type=layer_norm];
+"293 transpose_18" [id=293, type=transpose];
+"294 transpose_18_0_0_nncf_smooth_quant_0" [id=294, type=call_module];
+"295 quantize_per_tensor_default_19" [id=295, type=quantize_per_tensor];
+"296 dequantize_per_tensor_default_19" [id=296, type=dequantize_per_tensor];
+"297 encoder_layers_encoder_layer_3_self_attention_in_proj_bias_0_0" [id=297, type=get_attr];
+"298 scale_updated_constant13" [id=298, type=get_attr];
+"299 compressed_weight_updated_constant13" [id=299, type=get_attr];
+"300 mul_tensor_13" [id=300, type=mul];
+"301 zero_point_updated_constant13" [id=301, type=get_attr];
+"302 sub_tensor_13" [id=302, type=sub];
+"303 linear_12" [id=303, type=linear];
+"304 unflatten_3" [id=304, type=unflatten];
+"305 unsqueeze_3" [id=305, type=unsqueeze];
+"306 transpose_19" [id=306, type=transpose];
+"307 squeeze_3" [id=307, type=squeeze];
+"308 contiguous_3" [id=308, type=contiguous];
+"309 quantize_per_tensor_default_20" [id=309, type=quantize_per_tensor];
+"310 dequantize_per_tensor_default_20" [id=310, type=dequantize_per_tensor];
+"311 select_9" [id=311, type=select];
+"312 quantize_per_tensor_default_21" [id=312, type=quantize_per_tensor];
+"313 dequantize_per_tensor_default_21" [id=313, type=dequantize_per_tensor];
+"314 select_10" [id=314, type=select];
+"315 select_11" [id=315, type=select];
+"316 view_24" [id=316, type=view];
+"317 transpose_20" [id=317, type=transpose];
+"318 view_25" [id=318, type=view];
+"319 transpose_21" [id=319, type=transpose];
+"320 view_26" [id=320, type=view];
+"321 transpose_22" [id=321, type=transpose];
+"322 view_27" [id=322, type=view];
+"323 view_28" [id=323, type=view];
+"324 view_29" [id=324, type=view];
+"325 scaled_dot_product_attention_3" [id=325, type=scaled_dot_product_attention];
+"326 permute_4" [id=326, type=permute];
+"327 view_30" [id=327, type=view];
+"328 view_30_0_0_nncf_smooth_quant_0" [id=328, type=call_module];
+"329 quantize_per_tensor_default_22" [id=329, type=quantize_per_tensor];
+"330 dequantize_per_tensor_default_22" [id=330, type=dequantize_per_tensor];
+"331 encoder_layers_encoder_layer_3_self_attention_out_proj_bias_0_0" [id=331, type=get_attr];
+"332 scale_updated_constant14" [id=332, type=get_attr];
+"333 compressed_weight_updated_constant14" [id=333, type=get_attr];
+"334 mul_tensor_14" [id=334, type=mul];
+"335 zero_point_updated_constant14" [id=335, type=get_attr];
+"336 sub_tensor_14" [id=336, type=sub];
+"337 linear_13" [id=337, type=linear];
+"338 view_31" [id=338, type=view];
+"339 transpose_23" [id=339, type=transpose];
+"340 dropout_10" [id=340, type=dropout];
+"341 add_7" [id=341, type=add];
+"342 layer_norm_7" [id=342, type=layer_norm];
+"343 layer_norm_7_0_0_nncf_smooth_quant_0" [id=343, type=call_module];
+"344 quantize_per_tensor_default_23" [id=344, type=quantize_per_tensor];
+"345 dequantize_per_tensor_default_23" [id=345, type=dequantize_per_tensor];
+"346 encoder_layers_encoder_layer_3_mlp_0_bias_0_0" [id=346, type=get_attr];
+"347 scale_updated_constant15" [id=347, type=get_attr];
+"348 compressed_weight_updated_constant15" [id=348, type=get_attr];
+"349 mul_tensor_15" [id=349, type=mul];
+"350 zero_point_updated_constant15" [id=350, type=get_attr];
+"351 sub_tensor_15" [id=351, type=sub];
+"352 linear_14" [id=352, type=linear];
+"353 gelu_3" [id=353, type=gelu];
+"354 dropout_11" [id=354, type=dropout];
+"355 dropout_11_0_0_nncf_smooth_quant_0" [id=355, type=call_module];
+"356 quantize_per_tensor_default_24" [id=356, type=quantize_per_tensor];
+"357 dequantize_per_tensor_default_24" [id=357, type=dequantize_per_tensor];
+"358 encoder_layers_encoder_layer_3_mlp_3_bias_0_0" [id=358, type=get_attr];
+"359 scale_updated_constant16" [id=359, type=get_attr];
+"360 compressed_weight_updated_constant16" [id=360, type=get_attr];
+"361 mul_tensor_16" [id=361, type=mul];
+"362 zero_point_updated_constant16" [id=362, type=get_attr];
+"363 sub_tensor_16" [id=363, type=sub];
+"364 linear_15" [id=364, type=linear];
+"365 dropout_12" [id=365, type=dropout];
+"366 add_8" [id=366, type=add];
+"367 layer_norm_8" [id=367, type=layer_norm];
+"368 transpose_24" [id=368, type=transpose];
+"369 transpose_24_0_0_nncf_smooth_quant_0" [id=369, type=call_module];
+"370 quantize_per_tensor_default_25" [id=370, type=quantize_per_tensor];
+"371 dequantize_per_tensor_default_25" [id=371, type=dequantize_per_tensor];
+"372 encoder_layers_encoder_layer_4_self_attention_in_proj_bias_0_0" [id=372, type=get_attr];
+"373 scale_updated_constant17" [id=373, type=get_attr];
+"374 compressed_weight_updated_constant17" [id=374, type=get_attr];
+"375 mul_tensor_17" [id=375, type=mul];
+"376 zero_point_updated_constant17" [id=376, type=get_attr];
+"377 sub_tensor_17" [id=377, type=sub];
+"378 linear_16" [id=378, type=linear];
+"379 unflatten_4" [id=379, type=unflatten];
+"380 unsqueeze_4" [id=380, type=unsqueeze];
+"381 transpose_25" [id=381, type=transpose];
+"382 squeeze_4" [id=382, type=squeeze];
+"383 contiguous_4" [id=383, type=contiguous];
+"384 quantize_per_tensor_default_26" [id=384, type=quantize_per_tensor];
+"385 dequantize_per_tensor_default_26" [id=385, type=dequantize_per_tensor];
+"386 select_12" [id=386, type=select];
+"387 quantize_per_tensor_default_27" [id=387, type=quantize_per_tensor];
+"388 dequantize_per_tensor_default_27" [id=388, type=dequantize_per_tensor];
+"389 select_13" [id=389, type=select];
+"390 select_14" [id=390, type=select];
+"391 view_32" [id=391, type=view];
+"392 transpose_26" [id=392, type=transpose];
+"393 view_33" [id=393, type=view];
+"394 transpose_27" [id=394, type=transpose];
+"395 view_34" [id=395, type=view];
+"396 transpose_28" [id=396, type=transpose];
+"397 view_35" [id=397, type=view];
+"398 view_36" [id=398, type=view];
+"399 view_37" [id=399, type=view];
+"400 scaled_dot_product_attention_4" [id=400, type=scaled_dot_product_attention];
+"401 permute_5" [id=401, type=permute];
+"402 view_38" [id=402, type=view];
+"403 view_38_0_0_nncf_smooth_quant_0" [id=403, type=call_module];
+"404 quantize_per_tensor_default_28" [id=404, type=quantize_per_tensor];
+"405 dequantize_per_tensor_default_28" [id=405, type=dequantize_per_tensor];
+"406 encoder_layers_encoder_layer_4_self_attention_out_proj_bias_0_0" [id=406, type=get_attr];
+"407 scale_updated_constant18" [id=407, type=get_attr];
+"408 compressed_weight_updated_constant18" [id=408, type=get_attr];
+"409 mul_tensor_18" [id=409, type=mul];
+"410 zero_point_updated_constant18" [id=410, type=get_attr];
+"411 sub_tensor_18" [id=411, type=sub];
+"412 linear_17" [id=412, type=linear];
+"413 view_39" [id=413, type=view];
+"414 transpose_29" [id=414, type=transpose];
+"415 dropout_13" [id=415, type=dropout];
+"416 add_9" [id=416, type=add];
+"417 layer_norm_9" [id=417, type=layer_norm];
+"418 layer_norm_9_0_0_nncf_smooth_quant_0" [id=418, type=call_module];
+"419 quantize_per_tensor_default_29" [id=419, type=quantize_per_tensor];
+"420 dequantize_per_tensor_default_29" [id=420, type=dequantize_per_tensor];
+"421 encoder_layers_encoder_layer_4_mlp_0_bias_0_0" [id=421, type=get_attr];
+"422 scale_updated_constant19" [id=422, type=get_attr];
+"423 compressed_weight_updated_constant19" [id=423, type=get_attr];
+"424 mul_tensor_19" [id=424, type=mul];
+"425 zero_point_updated_constant19" [id=425, type=get_attr];
+"426 sub_tensor_19" [id=426, type=sub];
+"427 linear_18" [id=427, type=linear];
+"428 gelu_4" [id=428, type=gelu];
+"429 dropout_14" [id=429, type=dropout];
+"430 dropout_14_0_0_nncf_smooth_quant_0" [id=430, type=call_module];
+"431 quantize_per_tensor_default_30" [id=431, type=quantize_per_tensor];
+"432 dequantize_per_tensor_default_30" [id=432, type=dequantize_per_tensor];
+"433 encoder_layers_encoder_layer_4_mlp_3_bias_0_0" [id=433, type=get_attr];
+"434 scale_updated_constant20" [id=434, type=get_attr];
+"435 compressed_weight_updated_constant20" [id=435, type=get_attr];
+"436 mul_tensor_20" [id=436, type=mul];
+"437 zero_point_updated_constant20" [id=437, type=get_attr];
+"438 sub_tensor_20" [id=438, type=sub];
+"439 linear_19" [id=439, type=linear];
+"440 dropout_15" [id=440, type=dropout];
+"441 add_10" [id=441, type=add];
+"442 layer_norm_10" [id=442, type=layer_norm];
+"443 transpose_30" [id=443, type=transpose];
+"444 transpose_30_0_0_nncf_smooth_quant_0" [id=444, type=call_module];
+"445 quantize_per_tensor_default_31" [id=445, type=quantize_per_tensor];
+"446 dequantize_per_tensor_default_31" [id=446, type=dequantize_per_tensor];
+"447 encoder_layers_encoder_layer_5_self_attention_in_proj_bias_0_0" [id=447, type=get_attr];
+"448 scale_updated_constant21" [id=448, type=get_attr];
+"449 compressed_weight_updated_constant21" [id=449, type=get_attr];
+"450 mul_tensor_21" [id=450, type=mul];
+"451 zero_point_updated_constant21" [id=451, type=get_attr];
+"452 sub_tensor_21" [id=452, type=sub];
+"453 linear_20" [id=453, type=linear];
+"454 unflatten_5" [id=454, type=unflatten];
+"455 unsqueeze_5" [id=455, type=unsqueeze];
+"456 transpose_31" [id=456, type=transpose];
+"457 squeeze_5" [id=457, type=squeeze];
+"458 contiguous_5" [id=458, type=contiguous];
+"459 quantize_per_tensor_default_32" [id=459, type=quantize_per_tensor];
+"460 dequantize_per_tensor_default_32" [id=460, type=dequantize_per_tensor];
+"461 select_15" [id=461, type=select];
+"462 quantize_per_tensor_default_33" [id=462, type=quantize_per_tensor];
+"463 dequantize_per_tensor_default_33" [id=463, type=dequantize_per_tensor];
+"464 select_16" [id=464, type=select];
+"465 select_17" [id=465, type=select];
+"466 view_40" [id=466, type=view];
+"467 transpose_32" [id=467, type=transpose];
+"468 view_41" [id=468, type=view];
+"469 transpose_33" [id=469, type=transpose];
+"470 view_42" [id=470, type=view];
+"471 transpose_34" [id=471, type=transpose];
+"472 view_43" [id=472, type=view];
+"473 view_44" [id=473, type=view];
+"474 view_45" [id=474, type=view];
+"475 scaled_dot_product_attention_5" [id=475, type=scaled_dot_product_attention];
+"476 permute_6" [id=476, type=permute];
+"477 view_46" [id=477, type=view];
+"478 view_46_0_0_nncf_smooth_quant_0" [id=478, type=call_module];
+"479 quantize_per_tensor_default_34" [id=479, type=quantize_per_tensor];
+"480 dequantize_per_tensor_default_34" [id=480, type=dequantize_per_tensor];
+"481 encoder_layers_encoder_layer_5_self_attention_out_proj_bias_0_0" [id=481, type=get_attr];
+"482 scale_updated_constant22" [id=482, type=get_attr];
+"483 compressed_weight_updated_constant22" [id=483, type=get_attr];
+"484 mul_tensor_22" [id=484, type=mul];
+"485 zero_point_updated_constant22" [id=485, type=get_attr];
+"486 sub_tensor_22" [id=486, type=sub];
+"487 linear_21" [id=487, type=linear];
+"488 view_47" [id=488, type=view];
+"489 transpose_35" [id=489, type=transpose];
+"490 dropout_16" [id=490, type=dropout];
+"491 add_11" [id=491, type=add];
+"492 layer_norm_11" [id=492, type=layer_norm];
+"493 layer_norm_11_0_0_nncf_smooth_quant_0" [id=493, type=call_module];
+"494 quantize_per_tensor_default_35" [id=494, type=quantize_per_tensor];
+"495 dequantize_per_tensor_default_35" [id=495, type=dequantize_per_tensor];
+"496 encoder_layers_encoder_layer_5_mlp_0_bias_0_0" [id=496, type=get_attr];
+"497 scale_updated_constant23" [id=497, type=get_attr];
+"498 compressed_weight_updated_constant23" [id=498, type=get_attr];
+"499 mul_tensor_23" [id=499, type=mul];
+"500 zero_point_updated_constant23" [id=500, type=get_attr];
+"501 sub_tensor_23" [id=501, type=sub];
+"502 linear_22" [id=502, type=linear];
+"503 gelu_5" [id=503, type=gelu];
+"504 dropout_17" [id=504, type=dropout];
+"505 dropout_17_0_0_nncf_smooth_quant_0" [id=505, type=call_module];
+"506 quantize_per_tensor_default_36" [id=506, type=quantize_per_tensor];
+"507 dequantize_per_tensor_default_36" [id=507, type=dequantize_per_tensor];
+"508 encoder_layers_encoder_layer_5_mlp_3_bias_0_0" [id=508, type=get_attr];
+"509 scale_updated_constant24" [id=509, type=get_attr];
+"510 compressed_weight_updated_constant24" [id=510, type=get_attr];
+"511 mul_tensor_24" [id=511, type=mul];
+"512 zero_point_updated_constant24" [id=512, type=get_attr];
+"513 sub_tensor_24" [id=513, type=sub];
+"514 linear_23" [id=514, type=linear];
+"515 dropout_18" [id=515, type=dropout];
+"516 add_12" [id=516, type=add];
+"517 layer_norm_12" [id=517, type=layer_norm];
+"518 transpose_36" [id=518, type=transpose];
+"519 transpose_36_0_0_nncf_smooth_quant_0" [id=519, type=call_module];
+"520 quantize_per_tensor_default_37" [id=520, type=quantize_per_tensor];
+"521 dequantize_per_tensor_default_37" [id=521, type=dequantize_per_tensor];
+"522 encoder_layers_encoder_layer_6_self_attention_in_proj_bias_0_0" [id=522, type=get_attr];
+"523 scale_updated_constant25" [id=523, type=get_attr];
+"524 compressed_weight_updated_constant25" [id=524, type=get_attr];
+"525 mul_tensor_25" [id=525, type=mul];
+"526 zero_point_updated_constant25" [id=526, type=get_attr];
+"527 sub_tensor_25" [id=527, type=sub];
+"528 linear_24" [id=528, type=linear];
+"529 unflatten_6" [id=529, type=unflatten];
+"530 unsqueeze_6" [id=530, type=unsqueeze];
+"531 transpose_37" [id=531, type=transpose];
+"532 squeeze_6" [id=532, type=squeeze];
+"533 contiguous_6" [id=533, type=contiguous];
+"534 quantize_per_tensor_default_38" [id=534, type=quantize_per_tensor];
+"535 dequantize_per_tensor_default_38" [id=535, type=dequantize_per_tensor];
+"536 select_18" [id=536, type=select];
+"537 quantize_per_tensor_default_39" [id=537, type=quantize_per_tensor];
+"538 dequantize_per_tensor_default_39" [id=538, type=dequantize_per_tensor];
+"539 select_19" [id=539, type=select];
+"540 select_20" [id=540, type=select];
+"541 view_48" [id=541, type=view];
+"542 transpose_38" [id=542, type=transpose];
+"543 view_49" [id=543, type=view];
+"544 transpose_39" [id=544, type=transpose];
+"545 view_50" [id=545, type=view];
+"546 transpose_40" [id=546, type=transpose];
+"547 view_51" [id=547, type=view];
+"548 view_52" [id=548, type=view];
+"549 view_53" [id=549, type=view];
+"550 scaled_dot_product_attention_6" [id=550, type=scaled_dot_product_attention];
+"551 permute_7" [id=551, type=permute];
+"552 view_54" [id=552, type=view];
+"553 view_54_0_0_nncf_smooth_quant_0" [id=553, type=call_module];
+"554 quantize_per_tensor_default_40" [id=554, type=quantize_per_tensor];
+"555 dequantize_per_tensor_default_40" [id=555, type=dequantize_per_tensor];
+"556 encoder_layers_encoder_layer_6_self_attention_out_proj_bias_0_0" [id=556, type=get_attr];
+"557 scale_updated_constant26" [id=557, type=get_attr];
+"558 compressed_weight_updated_constant26" [id=558, type=get_attr];
+"559 mul_tensor_26" [id=559, type=mul];
+"560 zero_point_updated_constant26" [id=560, type=get_attr];
+"561 sub_tensor_26" [id=561, type=sub];
+"562 linear_25" [id=562, type=linear];
+"563 view_55" [id=563, type=view];
+"564 transpose_41" [id=564, type=transpose];
+"565 dropout_19" [id=565, type=dropout];
+"566 add_13" [id=566, type=add];
+"567 layer_norm_13" [id=567, type=layer_norm];
+"568 layer_norm_13_0_0_nncf_smooth_quant_0" [id=568, type=call_module];
+"569 quantize_per_tensor_default_41" [id=569, type=quantize_per_tensor];
+"570 dequantize_per_tensor_default_41" [id=570, type=dequantize_per_tensor];
+"571 encoder_layers_encoder_layer_6_mlp_0_bias_0_0" [id=571, type=get_attr];
+"572 scale_updated_constant27" [id=572, type=get_attr];
+"573 compressed_weight_updated_constant27" [id=573, type=get_attr];
+"574 mul_tensor_27" [id=574, type=mul];
+"575 zero_point_updated_constant27" [id=575, type=get_attr];
+"576 sub_tensor_27" [id=576, type=sub];
+"577 linear_26" [id=577, type=linear];
+"578 gelu_6" [id=578, type=gelu];
+"579 dropout_20" [id=579, type=dropout];
+"580 dropout_20_0_0_nncf_smooth_quant_0" [id=580, type=call_module];
+"581 quantize_per_tensor_default_42" [id=581, type=quantize_per_tensor];
+"582 dequantize_per_tensor_default_42" [id=582, type=dequantize_per_tensor];
+"583 encoder_layers_encoder_layer_6_mlp_3_bias_0_0" [id=583, type=get_attr];
+"584 scale_updated_constant28" [id=584, type=get_attr];
+"585 compressed_weight_updated_constant28" [id=585, type=get_attr];
+"586 mul_tensor_28" [id=586, type=mul];
+"587 zero_point_updated_constant28" [id=587, type=get_attr];
+"588 sub_tensor_28" [id=588, type=sub];
+"589 linear_27" [id=589, type=linear];
+"590 dropout_21" [id=590, type=dropout];
+"591 add_14" [id=591, type=add];
+"592 layer_norm_14" [id=592, type=layer_norm];
+"593 transpose_42" [id=593, type=transpose];
+"594 transpose_42_0_0_nncf_smooth_quant_0" [id=594, type=call_module];
+"595 quantize_per_tensor_default_43" [id=595, type=quantize_per_tensor];
+"596 dequantize_per_tensor_default_43" [id=596, type=dequantize_per_tensor];
+"597 encoder_layers_encoder_layer_7_self_attention_in_proj_bias_0_0" [id=597, type=get_attr];
+"598 scale_updated_constant29" [id=598, type=get_attr];
+"599 compressed_weight_updated_constant29" [id=599, type=get_attr];
+"600 mul_tensor_29" [id=600, type=mul];
+"601 zero_point_updated_constant29" [id=601, type=get_attr];
+"602 sub_tensor_29" [id=602, type=sub];
+"603 linear_28" [id=603, type=linear];
+"604 unflatten_7" [id=604, type=unflatten];
+"605 unsqueeze_7" [id=605, type=unsqueeze];
+"606 transpose_43" [id=606, type=transpose];
+"607 squeeze_7" [id=607, type=squeeze];
+"608 contiguous_7" [id=608, type=contiguous];
+"609 quantize_per_tensor_default_44" [id=609, type=quantize_per_tensor];
+"610 dequantize_per_tensor_default_44" [id=610, type=dequantize_per_tensor];
+"611 select_21" [id=611, type=select];
+"612 quantize_per_tensor_default_45" [id=612, type=quantize_per_tensor];
+"613 dequantize_per_tensor_default_45" [id=613, type=dequantize_per_tensor];
+"614 select_22" [id=614, type=select];
+"615 select_23" [id=615, type=select];
+"616 view_56" [id=616, type=view];
+"617 transpose_44" [id=617, type=transpose];
+"618 view_57" [id=618, type=view];
+"619 transpose_45" [id=619, type=transpose];
+"620 view_58" [id=620, type=view];
+"621 transpose_46" [id=621, type=transpose];
+"622 view_59" [id=622, type=view];
+"623 view_60" [id=623, type=view];
+"624 view_61" [id=624, type=view];
+"625 scaled_dot_product_attention_7" [id=625, type=scaled_dot_product_attention];
+"626 permute_8" [id=626, type=permute];
+"627 view_62" [id=627, type=view];
+"628 view_62_0_0_nncf_smooth_quant_0" [id=628, type=call_module];
+"629 quantize_per_tensor_default_46" [id=629, type=quantize_per_tensor];
+"630 dequantize_per_tensor_default_46" [id=630, type=dequantize_per_tensor];
+"631 encoder_layers_encoder_layer_7_self_attention_out_proj_bias_0_0" [id=631, type=get_attr];
+"632 scale_updated_constant30" [id=632, type=get_attr];
+"633 compressed_weight_updated_constant30" [id=633, type=get_attr];
+"634 mul_tensor_30" [id=634, type=mul];
+"635 zero_point_updated_constant30" [id=635, type=get_attr];
+"636 sub_tensor_30" [id=636, type=sub];
+"637 linear_29" [id=637, type=linear];
+"638 view_63" [id=638, type=view];
+"639 transpose_47" [id=639, type=transpose];
+"640 dropout_22" [id=640, type=dropout];
+"641 add_15" [id=641, type=add];
+"642 layer_norm_15" [id=642, type=layer_norm];
+"643 layer_norm_15_0_0_nncf_smooth_quant_0" [id=643, type=call_module];
+"644 quantize_per_tensor_default_47" [id=644, type=quantize_per_tensor];
+"645 dequantize_per_tensor_default_47" [id=645, type=dequantize_per_tensor];
+"646 encoder_layers_encoder_layer_7_mlp_0_bias_0_0" [id=646, type=get_attr];
+"647 scale_updated_constant31" [id=647, type=get_attr];
+"648 compressed_weight_updated_constant31" [id=648, type=get_attr];
+"649 mul_tensor_31" [id=649, type=mul];
+"650 zero_point_updated_constant31" [id=650, type=get_attr];
+"651 sub_tensor_31" [id=651, type=sub];
+"652 linear_30" [id=652, type=linear];
+"653 gelu_7" [id=653, type=gelu];
+"654 dropout_23" [id=654, type=dropout];
+"655 dropout_23_0_0_nncf_smooth_quant_0" [id=655, type=call_module];
+"656 quantize_per_tensor_default_48" [id=656, type=quantize_per_tensor];
+"657 dequantize_per_tensor_default_48" [id=657, type=dequantize_per_tensor];
+"658 encoder_layers_encoder_layer_7_mlp_3_bias_0_0" [id=658, type=get_attr];
+"659 scale_updated_constant32" [id=659, type=get_attr];
+"660 compressed_weight_updated_constant32" [id=660, type=get_attr];
+"661 mul_tensor_32" [id=661, type=mul];
+"662 zero_point_updated_constant32" [id=662, type=get_attr];
+"663 sub_tensor_32" [id=663, type=sub];
+"664 linear_31" [id=664, type=linear];
+"665 dropout_24" [id=665, type=dropout];
+"666 add_16" [id=666, type=add];
+"667 layer_norm_16" [id=667, type=layer_norm];
+"668 transpose_48" [id=668, type=transpose];
+"669 transpose_48_0_0_nncf_smooth_quant_0" [id=669, type=call_module];
+"670 quantize_per_tensor_default_49" [id=670, type=quantize_per_tensor];
+"671 dequantize_per_tensor_default_49" [id=671, type=dequantize_per_tensor];
+"672 encoder_layers_encoder_layer_8_self_attention_in_proj_bias_0_0" [id=672, type=get_attr];
+"673 scale_updated_constant33" [id=673, type=get_attr];
+"674 compressed_weight_updated_constant33" [id=674, type=get_attr];
+"675 mul_tensor_33" [id=675, type=mul];
+"676 zero_point_updated_constant33" [id=676, type=get_attr];
+"677 sub_tensor_33" [id=677, type=sub];
+"678 linear_32" [id=678, type=linear];
+"679 unflatten_8" [id=679, type=unflatten];
+"680 unsqueeze_8" [id=680, type=unsqueeze];
+"681 transpose_49" [id=681, type=transpose];
+"682 squeeze_8" [id=682, type=squeeze];
+"683 contiguous_8" [id=683, type=contiguous];
+"684 quantize_per_tensor_default_50" [id=684, type=quantize_per_tensor];
+"685 dequantize_per_tensor_default_50" [id=685, type=dequantize_per_tensor];
+"686 select_24" [id=686, type=select];
+"687 quantize_per_tensor_default_51" [id=687, type=quantize_per_tensor];
+"688 dequantize_per_tensor_default_51" [id=688, type=dequantize_per_tensor];
+"689 select_25" [id=689, type=select];
+"690 select_26" [id=690, type=select];
+"691 view_64" [id=691, type=view];
+"692 transpose_50" [id=692, type=transpose];
+"693 view_65" [id=693, type=view];
+"694 transpose_51" [id=694, type=transpose];
+"695 view_66" [id=695, type=view];
+"696 transpose_52" [id=696, type=transpose];
+"697 view_67" [id=697, type=view];
+"698 view_68" [id=698, type=view];
+"699 view_69" [id=699, type=view];
+"700 scaled_dot_product_attention_8" [id=700, type=scaled_dot_product_attention];
+"701 permute_9" [id=701, type=permute];
+"702 view_70" [id=702, type=view];
+"703 view_70_0_0_nncf_smooth_quant_0" [id=703, type=call_module];
+"704 quantize_per_tensor_default_52" [id=704, type=quantize_per_tensor];
+"705 dequantize_per_tensor_default_52" [id=705, type=dequantize_per_tensor];
+"706 encoder_layers_encoder_layer_8_self_attention_out_proj_bias_0_0" [id=706, type=get_attr];
+"707 scale_updated_constant34" [id=707, type=get_attr];
+"708 compressed_weight_updated_constant34" [id=708, type=get_attr];
+"709 mul_tensor_34" [id=709, type=mul];
+"710 zero_point_updated_constant34" [id=710, type=get_attr];
+"711 sub_tensor_34" [id=711, type=sub];
+"712 linear_33" [id=712, type=linear];
+"713 view_71" [id=713, type=view];
+"714 transpose_53" [id=714, type=transpose];
+"715 dropout_25" [id=715, type=dropout];
+"716 add_17" [id=716, type=add];
+"717 layer_norm_17" [id=717, type=layer_norm];
+"718 layer_norm_17_0_0_nncf_smooth_quant_0" [id=718, type=call_module];
+"719 quantize_per_tensor_default_53" [id=719, type=quantize_per_tensor];
+"720 dequantize_per_tensor_default_53" [id=720, type=dequantize_per_tensor];
+"721 encoder_layers_encoder_layer_8_mlp_0_bias_0_0" [id=721, type=get_attr];
+"722 scale_updated_constant35" [id=722, type=get_attr];
+"723 compressed_weight_updated_constant35" [id=723, type=get_attr];
+"724 mul_tensor_35" [id=724, type=mul];
+"725 zero_point_updated_constant35" [id=725, type=get_attr];
+"726 sub_tensor_35" [id=726, type=sub];
+"727 linear_34" [id=727, type=linear];
+"728 gelu_8" [id=728, type=gelu];
+"729 dropout_26" [id=729, type=dropout];
+"730 dropout_26_0_0_nncf_smooth_quant_0" [id=730, type=call_module];
+"731 quantize_per_tensor_default_54" [id=731, type=quantize_per_tensor];
+"732 dequantize_per_tensor_default_54" [id=732, type=dequantize_per_tensor];
+"733 encoder_layers_encoder_layer_8_mlp_3_bias_0_0" [id=733, type=get_attr];
+"734 scale_updated_constant36" [id=734, type=get_attr];
+"735 compressed_weight_updated_constant36" [id=735, type=get_attr];
+"736 mul_tensor_36" [id=736, type=mul];
+"737 zero_point_updated_constant36" [id=737, type=get_attr];
+"738 sub_tensor_36" [id=738, type=sub];
+"739 linear_35" [id=739, type=linear];
+"740 dropout_27" [id=740, type=dropout];
+"741 add_18" [id=741, type=add];
+"742 layer_norm_18" [id=742, type=layer_norm];
+"743 transpose_54" [id=743, type=transpose];
+"744 transpose_54_0_0_nncf_smooth_quant_0" [id=744, type=call_module];
+"745 quantize_per_tensor_default_55" [id=745, type=quantize_per_tensor];
+"746 dequantize_per_tensor_default_55" [id=746, type=dequantize_per_tensor];
+"747 encoder_layers_encoder_layer_9_self_attention_in_proj_bias_0_0" [id=747, type=get_attr];
+"748 scale_updated_constant37" [id=748, type=get_attr];
+"749 compressed_weight_updated_constant37" [id=749, type=get_attr];
+"750 mul_tensor_37" [id=750, type=mul];
+"751 zero_point_updated_constant37" [id=751, type=get_attr];
+"752 sub_tensor_37" [id=752, type=sub];
+"753 linear_36" [id=753, type=linear];
+"754 unflatten_9" [id=754, type=unflatten];
+"755 unsqueeze_9" [id=755, type=unsqueeze];
+"756 transpose_55" [id=756, type=transpose];
+"757 squeeze_9" [id=757, type=squeeze];
+"758 contiguous_9" [id=758, type=contiguous];
+"759 quantize_per_tensor_default_56" [id=759, type=quantize_per_tensor];
+"760 dequantize_per_tensor_default_56" [id=760, type=dequantize_per_tensor];
+"761 select_27" [id=761, type=select];
+"762 quantize_per_tensor_default_57" [id=762, type=quantize_per_tensor];
+"763 dequantize_per_tensor_default_57" [id=763, type=dequantize_per_tensor];
+"764 select_28" [id=764, type=select];
+"765 select_29" [id=765, type=select];
+"766 view_72" [id=766, type=view];
+"767 transpose_56" [id=767, type=transpose];
+"768 view_73" [id=768, type=view];
+"769 transpose_57" [id=769, type=transpose];
+"770 view_74" [id=770, type=view];
+"771 transpose_58" [id=771, type=transpose];
+"772 view_75" [id=772, type=view];
+"773 view_76" [id=773, type=view];
+"774 view_77" [id=774, type=view];
+"775 scaled_dot_product_attention_9" [id=775, type=scaled_dot_product_attention];
+"776 permute_10" [id=776, type=permute];
+"777 view_78" [id=777, type=view];
+"778 view_78_0_0_nncf_smooth_quant_0" [id=778, type=call_module];
+"779 quantize_per_tensor_default_58" [id=779, type=quantize_per_tensor];
+"780 dequantize_per_tensor_default_58" [id=780, type=dequantize_per_tensor];
+"781 encoder_layers_encoder_layer_9_self_attention_out_proj_bias_0_0" [id=781, type=get_attr];
+"782 scale_updated_constant38" [id=782, type=get_attr];
+"783 compressed_weight_updated_constant38" [id=783, type=get_attr];
+"784 mul_tensor_38" [id=784, type=mul];
+"785 zero_point_updated_constant38" [id=785, type=get_attr];
+"786 sub_tensor_38" [id=786, type=sub];
+"787 linear_37" [id=787, type=linear];
+"788 view_79" [id=788, type=view];
+"789 transpose_59" [id=789, type=transpose];
+"790 dropout_28" [id=790, type=dropout];
+"791 add_19" [id=791, type=add];
+"792 layer_norm_19" [id=792, type=layer_norm];
+"793 layer_norm_19_0_0_nncf_smooth_quant_0" [id=793, type=call_module];
+"794 quantize_per_tensor_default_59" [id=794, type=quantize_per_tensor];
+"795 dequantize_per_tensor_default_59" [id=795, type=dequantize_per_tensor];
+"796 encoder_layers_encoder_layer_9_mlp_0_bias_0_0" [id=796, type=get_attr];
+"797 scale_updated_constant39" [id=797, type=get_attr];
+"798 compressed_weight_updated_constant39" [id=798, type=get_attr];
+"799 mul_tensor_39" [id=799, type=mul];
+"800 zero_point_updated_constant39" [id=800, type=get_attr];
+"801 sub_tensor_39" [id=801, type=sub];
+"802 linear_38" [id=802, type=linear];
+"803 gelu_9" [id=803, type=gelu];
+"804 dropout_29" [id=804, type=dropout];
+"805 dropout_29_0_0_nncf_smooth_quant_0" [id=805, type=call_module];
+"806 quantize_per_tensor_default_60" [id=806, type=quantize_per_tensor];
+"807 dequantize_per_tensor_default_60" [id=807, type=dequantize_per_tensor];
+"808 encoder_layers_encoder_layer_9_mlp_3_bias_0_0" [id=808, type=get_attr];
+"809 scale_updated_constant40" [id=809, type=get_attr];
+"810 compressed_weight_updated_constant40" [id=810, type=get_attr];
+"811 mul_tensor_40" [id=811, type=mul];
+"812 zero_point_updated_constant40" [id=812, type=get_attr];
+"813 sub_tensor_40" [id=813, type=sub];
+"814 linear_39" [id=814, type=linear];
+"815 dropout_30" [id=815, type=dropout];
+"816 add_20" [id=816, type=add];
+"817 layer_norm_20" [id=817, type=layer_norm];
+"818 transpose_60" [id=818, type=transpose];
+"819 transpose_60_0_0_nncf_smooth_quant_0" [id=819, type=call_module];
+"820 quantize_per_tensor_default_61" [id=820, type=quantize_per_tensor];
+"821 dequantize_per_tensor_default_61" [id=821, type=dequantize_per_tensor];
+"822 encoder_layers_encoder_layer_10_self_attention_in_proj_bias_0_0" [id=822, type=get_attr];
+"823 scale_updated_constant41" [id=823, type=get_attr];
+"824 compressed_weight_updated_constant41" [id=824, type=get_attr];
+"825 mul_tensor_41" [id=825, type=mul];
+"826 zero_point_updated_constant41" [id=826, type=get_attr];
+"827 sub_tensor_41" [id=827, type=sub];
+"828 linear_40" [id=828, type=linear];
+"829 unflatten_10" [id=829, type=unflatten];
+"830 unsqueeze_10" [id=830, type=unsqueeze];
+"831 transpose_61" [id=831, type=transpose];
+"832 squeeze_10" [id=832, type=squeeze];
+"833 contiguous_10" [id=833, type=contiguous];
+"834 quantize_per_tensor_default_62" [id=834, type=quantize_per_tensor];
+"835 dequantize_per_tensor_default_62" [id=835, type=dequantize_per_tensor];
+"836 select_30" [id=836, type=select];
+"837 quantize_per_tensor_default_63" [id=837, type=quantize_per_tensor];
+"838 dequantize_per_tensor_default_63" [id=838, type=dequantize_per_tensor];
+"839 select_31" [id=839, type=select];
+"840 select_32" [id=840, type=select];
+"841 view_80" [id=841, type=view];
+"842 transpose_62" [id=842, type=transpose];
+"843 view_81" [id=843, type=view];
+"844 transpose_63" [id=844, type=transpose];
+"845 view_82" [id=845, type=view];
+"846 transpose_64" [id=846, type=transpose];
+"847 view_83" [id=847, type=view];
+"848 view_84" [id=848, type=view];
+"849 view_85" [id=849, type=view];
+"850 scaled_dot_product_attention_10" [id=850, type=scaled_dot_product_attention];
+"851 permute_11" [id=851, type=permute];
+"852 view_86" [id=852, type=view];
+"853 view_86_0_0_nncf_smooth_quant_0" [id=853, type=call_module];
+"854 quantize_per_tensor_default_64" [id=854, type=quantize_per_tensor];
+"855 dequantize_per_tensor_default_64" [id=855, type=dequantize_per_tensor];
+"856 encoder_layers_encoder_layer_10_self_attention_out_proj_bias_0_0" [id=856, type=get_attr];
+"857 scale_updated_constant42" [id=857, type=get_attr];
+"858 compressed_weight_updated_constant42" [id=858, type=get_attr];
+"859 mul_tensor_42" [id=859, type=mul];
+"860 zero_point_updated_constant42" [id=860, type=get_attr];
+"861 sub_tensor_42" [id=861, type=sub];
+"862 linear_41" [id=862, type=linear];
+"863 view_87" [id=863, type=view];
+"864 transpose_65" [id=864, type=transpose];
+"865 dropout_31" [id=865, type=dropout];
+"866 add_21" [id=866, type=add];
+"867 layer_norm_21" [id=867, type=layer_norm];
+"868 layer_norm_21_0_0_nncf_smooth_quant_0" [id=868, type=call_module];
+"869 quantize_per_tensor_default_65" [id=869, type=quantize_per_tensor];
+"870 dequantize_per_tensor_default_65" [id=870, type=dequantize_per_tensor];
+"871 encoder_layers_encoder_layer_10_mlp_0_bias_0_0" [id=871, type=get_attr];
+"872 scale_updated_constant43" [id=872, type=get_attr];
+"873 compressed_weight_updated_constant43" [id=873, type=get_attr];
+"874 mul_tensor_43" [id=874, type=mul];
+"875 zero_point_updated_constant43" [id=875, type=get_attr];
+"876 sub_tensor_43" [id=876, type=sub];
+"877 linear_42" [id=877, type=linear];
+"878 gelu_10" [id=878, type=gelu];
+"879 dropout_32" [id=879, type=dropout];
+"880 dropout_32_0_0_nncf_smooth_quant_0" [id=880, type=call_module];
+"881 quantize_per_tensor_default_66" [id=881, type=quantize_per_tensor];
+"882 dequantize_per_tensor_default_66" [id=882, type=dequantize_per_tensor];
+"883 encoder_layers_encoder_layer_10_mlp_3_bias_0_0" [id=883, type=get_attr];
+"884 scale_updated_constant44" [id=884, type=get_attr];
+"885 compressed_weight_updated_constant44" [id=885, type=get_attr];
+"886 mul_tensor_44" [id=886, type=mul];
+"887 zero_point_updated_constant44" [id=887, type=get_attr];
+"888 sub_tensor_44" [id=888, type=sub];
+"889 linear_43" [id=889, type=linear];
+"890 dropout_33" [id=890, type=dropout];
+"891 add_22" [id=891, type=add];
+"892 layer_norm_22" [id=892, type=layer_norm];
+"893 transpose_66" [id=893, type=transpose];
+"894 transpose_66_0_0_nncf_smooth_quant_0" [id=894, type=call_module];
+"895 quantize_per_tensor_default_67" [id=895, type=quantize_per_tensor];
+"896 dequantize_per_tensor_default_67" [id=896, type=dequantize_per_tensor];
+"897 encoder_layers_encoder_layer_11_self_attention_in_proj_bias_0_0" [id=897, type=get_attr];
+"898 scale_updated_constant45" [id=898, type=get_attr];
+"899 compressed_weight_updated_constant45" [id=899, type=get_attr];
+"900 mul_tensor_45" [id=900, type=mul];
+"901 zero_point_updated_constant45" [id=901, type=get_attr];
+"902 sub_tensor_45" [id=902, type=sub];
+"903 linear_44" [id=903, type=linear];
+"904 unflatten_11" [id=904, type=unflatten];
+"905 unsqueeze_11" [id=905, type=unsqueeze];
+"906 transpose_67" [id=906, type=transpose];
+"907 squeeze_11" [id=907, type=squeeze];
+"908 contiguous_11" [id=908, type=contiguous];
+"909 quantize_per_tensor_default_68" [id=909, type=quantize_per_tensor];
+"910 dequantize_per_tensor_default_68" [id=910, type=dequantize_per_tensor];
+"911 select_33" [id=911, type=select];
+"912 quantize_per_tensor_default_69" [id=912, type=quantize_per_tensor];
+"913 dequantize_per_tensor_default_69" [id=913, type=dequantize_per_tensor];
+"914 select_34" [id=914, type=select];
+"915 select_35" [id=915, type=select];
+"916 view_88" [id=916, type=view];
+"917 transpose_68" [id=917, type=transpose];
+"918 view_89" [id=918, type=view];
+"919 transpose_69" [id=919, type=transpose];
+"920 view_90" [id=920, type=view];
+"921 transpose_70" [id=921, type=transpose];
+"922 view_91" [id=922, type=view];
+"923 view_92" [id=923, type=view];
+"924 view_93" [id=924, type=view];
+"925 scaled_dot_product_attention_11" [id=925, type=scaled_dot_product_attention];
+"926 permute_12" [id=926, type=permute];
+"927 view_94" [id=927, type=view];
+"928 view_94_0_0_nncf_smooth_quant_0" [id=928, type=call_module];
+"929 quantize_per_tensor_default_70" [id=929, type=quantize_per_tensor];
+"930 dequantize_per_tensor_default_70" [id=930, type=dequantize_per_tensor];
+"931 encoder_layers_encoder_layer_11_self_attention_out_proj_bias_0_0" [id=931, type=get_attr];
+"932 scale_updated_constant46" [id=932, type=get_attr];
+"933 compressed_weight_updated_constant46" [id=933, type=get_attr];
+"934 mul_tensor_46" [id=934, type=mul];
+"935 zero_point_updated_constant46" [id=935, type=get_attr];
+"936 sub_tensor_46" [id=936, type=sub];
+"937 linear_45" [id=937, type=linear];
+"938 view_95" [id=938, type=view];
+"939 transpose_71" [id=939, type=transpose];
+"940 dropout_34" [id=940, type=dropout];
+"941 add_23" [id=941, type=add];
+"942 layer_norm_23" [id=942, type=layer_norm];
+"943 layer_norm_23_0_0_nncf_smooth_quant_0" [id=943, type=call_module];
+"944 quantize_per_tensor_default_71" [id=944, type=quantize_per_tensor];
+"945 dequantize_per_tensor_default_71" [id=945, type=dequantize_per_tensor];
+"946 encoder_layers_encoder_layer_11_mlp_0_bias_0_0" [id=946, type=get_attr];
+"947 scale_updated_constant47" [id=947, type=get_attr];
+"948 compressed_weight_updated_constant47" [id=948, type=get_attr];
+"949 mul_tensor_47" [id=949, type=mul];
+"950 zero_point_updated_constant47" [id=950, type=get_attr];
+"951 sub_tensor_47" [id=951, type=sub];
+"952 linear_46" [id=952, type=linear];
+"953 gelu_11" [id=953, type=gelu];
+"954 dropout_35" [id=954, type=dropout];
+"955 dropout_35_0_0_nncf_smooth_quant_0" [id=955, type=call_module];
+"956 quantize_per_tensor_default_72" [id=956, type=quantize_per_tensor];
+"957 dequantize_per_tensor_default_72" [id=957, type=dequantize_per_tensor];
+"958 encoder_layers_encoder_layer_11_mlp_3_bias_0_0" [id=958, type=get_attr];
+"959 scale_updated_constant48" [id=959, type=get_attr];
+"960 compressed_weight_updated_constant48" [id=960, type=get_attr];
+"961 mul_tensor_48" [id=961, type=mul];
+"962 zero_point_updated_constant48" [id=962, type=get_attr];
+"963 sub_tensor_48" [id=963, type=sub];
+"964 linear_47" [id=964, type=linear];
+"965 dropout_36" [id=965, type=dropout];
+"966 add_24" [id=966, type=add];
 "967 layer_norm_24" [id=967, type=layer_norm];
 "968 slice_1" [id=968, type=slice];
 "969 select_36" [id=969, type=select];
 "970 select_36_0_0_nncf_smooth_quant_0" [id=970, type=call_module];
 "971 quantize_per_tensor_default_73" [id=971, type=quantize_per_tensor];
 "972 dequantize_per_tensor_default_73" [id=972, type=dequantize_per_tensor];
-"973 _param_constant151_0_0" [id=973, type=get_attr];
+"973 heads_head_bias_0_0" [id=973, type=get_attr];
 "974 scale_updated_constant49" [id=974, type=get_attr];
 "975 compressed_weight_updated_constant49" [id=975, type=get_attr];
 "976 mul_tensor_49" [id=976, type=mul];
@@ -980,1028 +980,1028 @@ strict digraph  {
 "978 sub_tensor_49" [id=978, type=sub];
 "979 linear_48" [id=979, type=linear];
 "980 output" [id=980, type=output];
-"0 arg0_1" -> "1 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"1 quantize_per_tensor_default" -> "2 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"2 dequantize_per_tensor_default" -> "9 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"3 _param_constant1_0_0" -> "9 conv2d"  [label="(768,)", style=solid];
-"4 scale_updated_constant0" -> "6 mul_tensor"  [label="(768, 1, 1, 1)", style=solid];
-"5 compressed_weight_updated_constant0" -> "6 mul_tensor"  [label="(768, 3, 16, 16)", style=solid];
-"6 mul_tensor" -> "8 sub_tensor"  [label="(768, 3, 16, 16)", style=solid];
-"7 zero_point_updated_constant0" -> "8 sub_tensor"  [label="(768, 1, 1, 1)", style=solid];
-"8 sub_tensor" -> "9 conv2d"  [label="(768, 3, 16, 16)", style=solid];
-"9 conv2d" -> "10 reshape"  [label="(1, 768, 14, 14)", style=solid];
-"10 reshape" -> "11 permute"  [label="(1, 768, 196)", style=solid];
-"11 permute" -> "13 cat"  [label="(1, 196, 768)", style=solid];
-"12 _frozen_param0" -> "13 cat"  [label="(1, 1, 768)", style=solid];
-"13 cat" -> "15 add"  [label="(1, 197, 768)", style=solid];
-"14 _param_constant3" -> "15 add"  [label="(1, 197, 768)", style=solid];
-"15 add" -> "16 dropout"  [label="(1, 197, 768)", style=solid];
-"16 dropout" -> "19 layer_norm"  [label="(1, 197, 768)", style=solid];
-"16 dropout" -> "68 add_1"  [label="(1, 197, 768)", style=solid];
-"17 _param_constant4" -> "19 layer_norm"  [label="(768,)", style=solid];
-"18 _param_constant5" -> "19 layer_norm"  [label="(768,)", style=solid];
-"19 layer_norm" -> "20 transpose"  [label="(1, 197, 768)", style=solid];
-"20 transpose" -> "21 transpose_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"21 transpose_0_0_nncf_smooth_quant_0" -> "22 quantize_per_tensor_default_1"  [label="(197, 1, 768)", style=solid];
-"22 quantize_per_tensor_default_1" -> "23 dequantize_per_tensor_default_1"  [label="(197, 1, 768)", style=solid];
-"23 dequantize_per_tensor_default_1" -> "30 linear"  [label="(197, 1, 768)", style=solid];
-"24 _param_constant7_0_0" -> "30 linear"  [label="(2304,)", style=solid];
-"25 scale_updated_constant1" -> "27 mul_tensor_1"  [label="(2304, 1)", style=solid];
-"26 compressed_weight_updated_constant1" -> "27 mul_tensor_1"  [label="(2304, 768)", style=solid];
-"27 mul_tensor_1" -> "29 sub_tensor_1"  [label="(2304, 768)", style=solid];
-"28 zero_point_updated_constant1" -> "29 sub_tensor_1"  [label="(2304, 1)", style=solid];
-"29 sub_tensor_1" -> "30 linear"  [label="(2304, 768)", style=solid];
-"30 linear" -> "31 unflatten"  [label="(197, 1, 2304)", style=solid];
-"31 unflatten" -> "32 unsqueeze"  [label="(197, 1, 3, 768)", style=solid];
-"32 unsqueeze" -> "33 transpose_1"  [label="(1, 197, 1, 3, 768)", style=solid];
-"33 transpose_1" -> "34 squeeze"  [label="(3, 197, 1, 1, 768)", style=solid];
-"34 squeeze" -> "35 contiguous"  [label="(3, 197, 1, 768)", style=solid];
-"35 contiguous" -> "36 quantize_per_tensor_default_2"  [label="(3, 197, 1, 768)", style=solid];
-"35 contiguous" -> "39 quantize_per_tensor_default_3"  [label="(3, 197, 1, 768)", style=solid];
-"35 contiguous" -> "42 select_2"  [label="(3, 197, 1, 768)", style=solid];
-"36 quantize_per_tensor_default_2" -> "37 dequantize_per_tensor_default_2"  [label="(3, 197, 1, 768)", style=solid];
-"37 dequantize_per_tensor_default_2" -> "38 select"  [label="(3, 197, 1, 768)", style=solid];
-"38 select" -> "43 view"  [label="(197, 1, 768)", style=solid];
-"39 quantize_per_tensor_default_3" -> "40 dequantize_per_tensor_default_3"  [label="(3, 197, 1, 768)", style=solid];
-"40 dequantize_per_tensor_default_3" -> "41 select_1"  [label="(3, 197, 1, 768)", style=solid];
-"41 select_1" -> "45 view_1"  [label="(197, 1, 768)", style=solid];
-"42 select_2" -> "47 view_2"  [label="(197, 1, 768)", style=solid];
-"43 view" -> "44 transpose_2"  [label="(197, 12, 64)", style=solid];
-"44 transpose_2" -> "49 view_3"  [label="(12, 197, 64)", style=solid];
-"45 view_1" -> "46 transpose_3"  [label="(197, 12, 64)", style=solid];
-"46 transpose_3" -> "50 view_4"  [label="(12, 197, 64)", style=solid];
-"47 view_2" -> "48 transpose_4"  [label="(197, 12, 64)", style=solid];
-"48 transpose_4" -> "51 view_5"  [label="(12, 197, 64)", style=solid];
-"49 view_3" -> "52 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
-"50 view_4" -> "52 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
-"51 view_5" -> "52 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
-"52 scaled_dot_product_attention" -> "53 permute_1"  [label="(1, 12, 197, 64)", style=solid];
-"53 permute_1" -> "54 view_6"  [label="(197, 1, 12, 64)", style=solid];
-"54 view_6" -> "55 view_6_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"55 view_6_0_0_nncf_smooth_quant_0" -> "56 quantize_per_tensor_default_4"  [label="(197, 768)", style=solid];
-"56 quantize_per_tensor_default_4" -> "57 dequantize_per_tensor_default_4"  [label="(197, 768)", style=solid];
-"57 dequantize_per_tensor_default_4" -> "64 linear_1"  [label="(197, 768)", style=solid];
-"58 _param_constant9_0_0" -> "64 linear_1"  [label="(768,)", style=solid];
-"59 scale_updated_constant2" -> "61 mul_tensor_2"  [label="(768, 1)", style=solid];
-"60 compressed_weight_updated_constant2" -> "61 mul_tensor_2"  [label="(768, 768)", style=solid];
-"61 mul_tensor_2" -> "63 sub_tensor_2"  [label="(768, 768)", style=solid];
-"62 zero_point_updated_constant2" -> "63 sub_tensor_2"  [label="(768, 1)", style=solid];
-"63 sub_tensor_2" -> "64 linear_1"  [label="(768, 768)", style=solid];
-"64 linear_1" -> "65 view_7"  [label="(197, 768)", style=solid];
-"65 view_7" -> "66 transpose_5"  [label="(197, 1, 768)", style=solid];
-"66 transpose_5" -> "67 dropout_1"  [label="(1, 197, 768)", style=solid];
-"67 dropout_1" -> "68 add_1"  [label="(1, 197, 768)", style=solid];
-"68 add_1" -> "71 layer_norm_1"  [label="(1, 197, 768)", style=solid];
-"68 add_1" -> "95 add_2"  [label="(1, 197, 768)", style=solid];
-"69 _param_constant10" -> "71 layer_norm_1"  [label="(768,)", style=solid];
-"70 _param_constant11" -> "71 layer_norm_1"  [label="(768,)", style=solid];
-"71 layer_norm_1" -> "72 layer_norm_1_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"72 layer_norm_1_0_0_nncf_smooth_quant_0" -> "73 quantize_per_tensor_default_5"  [label="(1, 197, 768)", style=solid];
-"73 quantize_per_tensor_default_5" -> "74 dequantize_per_tensor_default_5"  [label="(1, 197, 768)", style=solid];
-"74 dequantize_per_tensor_default_5" -> "81 linear_2"  [label="(1, 197, 768)", style=solid];
-"75 _param_constant13_0_0" -> "81 linear_2"  [label="(3072,)", style=solid];
-"76 scale_updated_constant3" -> "78 mul_tensor_3"  [label="(3072, 1)", style=solid];
-"77 compressed_weight_updated_constant3" -> "78 mul_tensor_3"  [label="(3072, 768)", style=solid];
-"78 mul_tensor_3" -> "80 sub_tensor_3"  [label="(3072, 768)", style=solid];
-"79 zero_point_updated_constant3" -> "80 sub_tensor_3"  [label="(3072, 1)", style=solid];
-"80 sub_tensor_3" -> "81 linear_2"  [label="(3072, 768)", style=solid];
-"81 linear_2" -> "82 gelu"  [label="(1, 197, 3072)", style=solid];
-"82 gelu" -> "83 dropout_2"  [label="(1, 197, 3072)", style=solid];
-"83 dropout_2" -> "84 dropout_2_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"84 dropout_2_0_0_nncf_smooth_quant_0" -> "85 quantize_per_tensor_default_6"  [label="(1, 197, 3072)", style=solid];
-"85 quantize_per_tensor_default_6" -> "86 dequantize_per_tensor_default_6"  [label="(1, 197, 3072)", style=solid];
-"86 dequantize_per_tensor_default_6" -> "93 linear_3"  [label="(1, 197, 3072)", style=solid];
-"87 _param_constant15_0_0" -> "93 linear_3"  [label="(768,)", style=solid];
-"88 scale_updated_constant4" -> "90 mul_tensor_4"  [label="(768, 1)", style=solid];
-"89 compressed_weight_updated_constant4" -> "90 mul_tensor_4"  [label="(768, 3072)", style=solid];
-"90 mul_tensor_4" -> "92 sub_tensor_4"  [label="(768, 3072)", style=solid];
-"91 zero_point_updated_constant4" -> "92 sub_tensor_4"  [label="(768, 1)", style=solid];
-"92 sub_tensor_4" -> "93 linear_3"  [label="(768, 3072)", style=solid];
-"93 linear_3" -> "94 dropout_3"  [label="(1, 197, 768)", style=solid];
-"94 dropout_3" -> "95 add_2"  [label="(1, 197, 768)", style=solid];
-"95 add_2" -> "98 layer_norm_2"  [label="(1, 197, 768)", style=solid];
-"95 add_2" -> "147 add_3"  [label="(1, 197, 768)", style=solid];
-"96 _param_constant16" -> "98 layer_norm_2"  [label="(768,)", style=solid];
-"97 _param_constant17" -> "98 layer_norm_2"  [label="(768,)", style=solid];
-"98 layer_norm_2" -> "99 transpose_6"  [label="(1, 197, 768)", style=solid];
-"99 transpose_6" -> "100 transpose_6_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"100 transpose_6_0_0_nncf_smooth_quant_0" -> "101 quantize_per_tensor_default_7"  [label="(197, 1, 768)", style=solid];
-"101 quantize_per_tensor_default_7" -> "102 dequantize_per_tensor_default_7"  [label="(197, 1, 768)", style=solid];
-"102 dequantize_per_tensor_default_7" -> "109 linear_4"  [label="(197, 1, 768)", style=solid];
-"103 _param_constant19_0_0" -> "109 linear_4"  [label="(2304,)", style=solid];
-"104 scale_updated_constant5" -> "106 mul_tensor_5"  [label="(2304, 1)", style=solid];
-"105 compressed_weight_updated_constant5" -> "106 mul_tensor_5"  [label="(2304, 768)", style=solid];
-"106 mul_tensor_5" -> "108 sub_tensor_5"  [label="(2304, 768)", style=solid];
-"107 zero_point_updated_constant5" -> "108 sub_tensor_5"  [label="(2304, 1)", style=solid];
-"108 sub_tensor_5" -> "109 linear_4"  [label="(2304, 768)", style=solid];
-"109 linear_4" -> "110 unflatten_1"  [label="(197, 1, 2304)", style=solid];
-"110 unflatten_1" -> "111 unsqueeze_1"  [label="(197, 1, 3, 768)", style=solid];
-"111 unsqueeze_1" -> "112 transpose_7"  [label="(1, 197, 1, 3, 768)", style=solid];
-"112 transpose_7" -> "113 squeeze_1"  [label="(3, 197, 1, 1, 768)", style=solid];
-"113 squeeze_1" -> "114 contiguous_1"  [label="(3, 197, 1, 768)", style=solid];
-"114 contiguous_1" -> "115 quantize_per_tensor_default_8"  [label="(3, 197, 1, 768)", style=solid];
-"114 contiguous_1" -> "118 quantize_per_tensor_default_9"  [label="(3, 197, 1, 768)", style=solid];
-"114 contiguous_1" -> "121 select_5"  [label="(3, 197, 1, 768)", style=solid];
-"115 quantize_per_tensor_default_8" -> "116 dequantize_per_tensor_default_8"  [label="(3, 197, 1, 768)", style=solid];
-"116 dequantize_per_tensor_default_8" -> "117 select_3"  [label="(3, 197, 1, 768)", style=solid];
-"117 select_3" -> "122 view_8"  [label="(197, 1, 768)", style=solid];
-"118 quantize_per_tensor_default_9" -> "119 dequantize_per_tensor_default_9"  [label="(3, 197, 1, 768)", style=solid];
-"119 dequantize_per_tensor_default_9" -> "120 select_4"  [label="(3, 197, 1, 768)", style=solid];
-"120 select_4" -> "124 view_9"  [label="(197, 1, 768)", style=solid];
-"121 select_5" -> "126 view_10"  [label="(197, 1, 768)", style=solid];
-"122 view_8" -> "123 transpose_8"  [label="(197, 12, 64)", style=solid];
-"123 transpose_8" -> "128 view_11"  [label="(12, 197, 64)", style=solid];
-"124 view_9" -> "125 transpose_9"  [label="(197, 12, 64)", style=solid];
-"125 transpose_9" -> "129 view_12"  [label="(12, 197, 64)", style=solid];
-"126 view_10" -> "127 transpose_10"  [label="(197, 12, 64)", style=solid];
-"127 transpose_10" -> "130 view_13"  [label="(12, 197, 64)", style=solid];
-"128 view_11" -> "131 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
-"129 view_12" -> "131 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
-"130 view_13" -> "131 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
-"131 scaled_dot_product_attention_1" -> "132 permute_2"  [label="(1, 12, 197, 64)", style=solid];
-"132 permute_2" -> "133 view_14"  [label="(197, 1, 12, 64)", style=solid];
-"133 view_14" -> "134 view_14_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"134 view_14_0_0_nncf_smooth_quant_0" -> "135 quantize_per_tensor_default_10"  [label="(197, 768)", style=solid];
-"135 quantize_per_tensor_default_10" -> "136 dequantize_per_tensor_default_10"  [label="(197, 768)", style=solid];
-"136 dequantize_per_tensor_default_10" -> "143 linear_5"  [label="(197, 768)", style=solid];
-"137 _param_constant21_0_0" -> "143 linear_5"  [label="(768,)", style=solid];
-"138 scale_updated_constant6" -> "140 mul_tensor_6"  [label="(768, 1)", style=solid];
-"139 compressed_weight_updated_constant6" -> "140 mul_tensor_6"  [label="(768, 768)", style=solid];
-"140 mul_tensor_6" -> "142 sub_tensor_6"  [label="(768, 768)", style=solid];
-"141 zero_point_updated_constant6" -> "142 sub_tensor_6"  [label="(768, 1)", style=solid];
-"142 sub_tensor_6" -> "143 linear_5"  [label="(768, 768)", style=solid];
-"143 linear_5" -> "144 view_15"  [label="(197, 768)", style=solid];
-"144 view_15" -> "145 transpose_11"  [label="(197, 1, 768)", style=solid];
-"145 transpose_11" -> "146 dropout_4"  [label="(1, 197, 768)", style=solid];
-"146 dropout_4" -> "147 add_3"  [label="(1, 197, 768)", style=solid];
-"147 add_3" -> "150 layer_norm_3"  [label="(1, 197, 768)", style=solid];
-"147 add_3" -> "174 add_4"  [label="(1, 197, 768)", style=solid];
-"148 _param_constant22" -> "150 layer_norm_3"  [label="(768,)", style=solid];
-"149 _param_constant23" -> "150 layer_norm_3"  [label="(768,)", style=solid];
-"150 layer_norm_3" -> "151 layer_norm_3_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"151 layer_norm_3_0_0_nncf_smooth_quant_0" -> "152 quantize_per_tensor_default_11"  [label="(1, 197, 768)", style=solid];
-"152 quantize_per_tensor_default_11" -> "153 dequantize_per_tensor_default_11"  [label="(1, 197, 768)", style=solid];
-"153 dequantize_per_tensor_default_11" -> "160 linear_6"  [label="(1, 197, 768)", style=solid];
-"154 _param_constant25_0_0" -> "160 linear_6"  [label="(3072,)", style=solid];
-"155 scale_updated_constant7" -> "157 mul_tensor_7"  [label="(3072, 1)", style=solid];
-"156 compressed_weight_updated_constant7" -> "157 mul_tensor_7"  [label="(3072, 768)", style=solid];
-"157 mul_tensor_7" -> "159 sub_tensor_7"  [label="(3072, 768)", style=solid];
-"158 zero_point_updated_constant7" -> "159 sub_tensor_7"  [label="(3072, 1)", style=solid];
-"159 sub_tensor_7" -> "160 linear_6"  [label="(3072, 768)", style=solid];
-"160 linear_6" -> "161 gelu_1"  [label="(1, 197, 3072)", style=solid];
-"161 gelu_1" -> "162 dropout_5"  [label="(1, 197, 3072)", style=solid];
-"162 dropout_5" -> "163 dropout_5_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"163 dropout_5_0_0_nncf_smooth_quant_0" -> "164 quantize_per_tensor_default_12"  [label="(1, 197, 3072)", style=solid];
-"164 quantize_per_tensor_default_12" -> "165 dequantize_per_tensor_default_12"  [label="(1, 197, 3072)", style=solid];
-"165 dequantize_per_tensor_default_12" -> "172 linear_7"  [label="(1, 197, 3072)", style=solid];
-"166 _param_constant27_0_0" -> "172 linear_7"  [label="(768,)", style=solid];
-"167 scale_updated_constant8" -> "169 mul_tensor_8"  [label="(768, 1)", style=solid];
-"168 compressed_weight_updated_constant8" -> "169 mul_tensor_8"  [label="(768, 3072)", style=solid];
-"169 mul_tensor_8" -> "171 sub_tensor_8"  [label="(768, 3072)", style=solid];
-"170 zero_point_updated_constant8" -> "171 sub_tensor_8"  [label="(768, 1)", style=solid];
-"171 sub_tensor_8" -> "172 linear_7"  [label="(768, 3072)", style=solid];
-"172 linear_7" -> "173 dropout_6"  [label="(1, 197, 768)", style=solid];
-"173 dropout_6" -> "174 add_4"  [label="(1, 197, 768)", style=solid];
-"174 add_4" -> "177 layer_norm_4"  [label="(1, 197, 768)", style=solid];
-"174 add_4" -> "226 add_5"  [label="(1, 197, 768)", style=solid];
-"175 _param_constant28" -> "177 layer_norm_4"  [label="(768,)", style=solid];
-"176 _param_constant29" -> "177 layer_norm_4"  [label="(768,)", style=solid];
-"177 layer_norm_4" -> "178 transpose_12"  [label="(1, 197, 768)", style=solid];
-"178 transpose_12" -> "179 transpose_12_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"179 transpose_12_0_0_nncf_smooth_quant_0" -> "180 quantize_per_tensor_default_13"  [label="(197, 1, 768)", style=solid];
-"180 quantize_per_tensor_default_13" -> "181 dequantize_per_tensor_default_13"  [label="(197, 1, 768)", style=solid];
-"181 dequantize_per_tensor_default_13" -> "188 linear_8"  [label="(197, 1, 768)", style=solid];
-"182 _param_constant31_0_0" -> "188 linear_8"  [label="(2304,)", style=solid];
-"183 scale_updated_constant9" -> "185 mul_tensor_9"  [label="(2304, 1)", style=solid];
-"184 compressed_weight_updated_constant9" -> "185 mul_tensor_9"  [label="(2304, 768)", style=solid];
-"185 mul_tensor_9" -> "187 sub_tensor_9"  [label="(2304, 768)", style=solid];
-"186 zero_point_updated_constant9" -> "187 sub_tensor_9"  [label="(2304, 1)", style=solid];
-"187 sub_tensor_9" -> "188 linear_8"  [label="(2304, 768)", style=solid];
-"188 linear_8" -> "189 unflatten_2"  [label="(197, 1, 2304)", style=solid];
-"189 unflatten_2" -> "190 unsqueeze_2"  [label="(197, 1, 3, 768)", style=solid];
-"190 unsqueeze_2" -> "191 transpose_13"  [label="(1, 197, 1, 3, 768)", style=solid];
-"191 transpose_13" -> "192 squeeze_2"  [label="(3, 197, 1, 1, 768)", style=solid];
-"192 squeeze_2" -> "193 contiguous_2"  [label="(3, 197, 1, 768)", style=solid];
-"193 contiguous_2" -> "194 quantize_per_tensor_default_14"  [label="(3, 197, 1, 768)", style=solid];
-"193 contiguous_2" -> "197 quantize_per_tensor_default_15"  [label="(3, 197, 1, 768)", style=solid];
-"193 contiguous_2" -> "200 select_8"  [label="(3, 197, 1, 768)", style=solid];
-"194 quantize_per_tensor_default_14" -> "195 dequantize_per_tensor_default_14"  [label="(3, 197, 1, 768)", style=solid];
-"195 dequantize_per_tensor_default_14" -> "196 select_6"  [label="(3, 197, 1, 768)", style=solid];
-"196 select_6" -> "201 view_16"  [label="(197, 1, 768)", style=solid];
-"197 quantize_per_tensor_default_15" -> "198 dequantize_per_tensor_default_15"  [label="(3, 197, 1, 768)", style=solid];
-"198 dequantize_per_tensor_default_15" -> "199 select_7"  [label="(3, 197, 1, 768)", style=solid];
-"199 select_7" -> "203 view_17"  [label="(197, 1, 768)", style=solid];
-"200 select_8" -> "205 view_18"  [label="(197, 1, 768)", style=solid];
-"201 view_16" -> "202 transpose_14"  [label="(197, 12, 64)", style=solid];
-"202 transpose_14" -> "207 view_19"  [label="(12, 197, 64)", style=solid];
-"203 view_17" -> "204 transpose_15"  [label="(197, 12, 64)", style=solid];
-"204 transpose_15" -> "208 view_20"  [label="(12, 197, 64)", style=solid];
-"205 view_18" -> "206 transpose_16"  [label="(197, 12, 64)", style=solid];
-"206 transpose_16" -> "209 view_21"  [label="(12, 197, 64)", style=solid];
-"207 view_19" -> "210 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
-"208 view_20" -> "210 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
-"209 view_21" -> "210 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
-"210 scaled_dot_product_attention_2" -> "211 permute_3"  [label="(1, 12, 197, 64)", style=solid];
-"211 permute_3" -> "212 view_22"  [label="(197, 1, 12, 64)", style=solid];
-"212 view_22" -> "213 view_22_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"213 view_22_0_0_nncf_smooth_quant_0" -> "214 quantize_per_tensor_default_16"  [label="(197, 768)", style=solid];
-"214 quantize_per_tensor_default_16" -> "215 dequantize_per_tensor_default_16"  [label="(197, 768)", style=solid];
-"215 dequantize_per_tensor_default_16" -> "222 linear_9"  [label="(197, 768)", style=solid];
-"216 _param_constant33_0_0" -> "222 linear_9"  [label="(768,)", style=solid];
-"217 scale_updated_constant10" -> "219 mul_tensor_10"  [label="(768, 1)", style=solid];
-"218 compressed_weight_updated_constant10" -> "219 mul_tensor_10"  [label="(768, 768)", style=solid];
-"219 mul_tensor_10" -> "221 sub_tensor_10"  [label="(768, 768)", style=solid];
-"220 zero_point_updated_constant10" -> "221 sub_tensor_10"  [label="(768, 1)", style=solid];
-"221 sub_tensor_10" -> "222 linear_9"  [label="(768, 768)", style=solid];
-"222 linear_9" -> "223 view_23"  [label="(197, 768)", style=solid];
-"223 view_23" -> "224 transpose_17"  [label="(197, 1, 768)", style=solid];
-"224 transpose_17" -> "225 dropout_7"  [label="(1, 197, 768)", style=solid];
-"225 dropout_7" -> "226 add_5"  [label="(1, 197, 768)", style=solid];
-"226 add_5" -> "229 layer_norm_5"  [label="(1, 197, 768)", style=solid];
-"226 add_5" -> "253 add_6"  [label="(1, 197, 768)", style=solid];
-"227 _param_constant34" -> "229 layer_norm_5"  [label="(768,)", style=solid];
-"228 _param_constant35" -> "229 layer_norm_5"  [label="(768,)", style=solid];
-"229 layer_norm_5" -> "230 layer_norm_5_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"230 layer_norm_5_0_0_nncf_smooth_quant_0" -> "231 quantize_per_tensor_default_17"  [label="(1, 197, 768)", style=solid];
-"231 quantize_per_tensor_default_17" -> "232 dequantize_per_tensor_default_17"  [label="(1, 197, 768)", style=solid];
-"232 dequantize_per_tensor_default_17" -> "239 linear_10"  [label="(1, 197, 768)", style=solid];
-"233 _param_constant37_0_0" -> "239 linear_10"  [label="(3072,)", style=solid];
-"234 scale_updated_constant11" -> "236 mul_tensor_11"  [label="(3072, 1)", style=solid];
-"235 compressed_weight_updated_constant11" -> "236 mul_tensor_11"  [label="(3072, 768)", style=solid];
-"236 mul_tensor_11" -> "238 sub_tensor_11"  [label="(3072, 768)", style=solid];
-"237 zero_point_updated_constant11" -> "238 sub_tensor_11"  [label="(3072, 1)", style=solid];
-"238 sub_tensor_11" -> "239 linear_10"  [label="(3072, 768)", style=solid];
-"239 linear_10" -> "240 gelu_2"  [label="(1, 197, 3072)", style=solid];
-"240 gelu_2" -> "241 dropout_8"  [label="(1, 197, 3072)", style=solid];
-"241 dropout_8" -> "242 dropout_8_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"242 dropout_8_0_0_nncf_smooth_quant_0" -> "243 quantize_per_tensor_default_18"  [label="(1, 197, 3072)", style=solid];
-"243 quantize_per_tensor_default_18" -> "244 dequantize_per_tensor_default_18"  [label="(1, 197, 3072)", style=solid];
-"244 dequantize_per_tensor_default_18" -> "251 linear_11"  [label="(1, 197, 3072)", style=solid];
-"245 _param_constant39_0_0" -> "251 linear_11"  [label="(768,)", style=solid];
-"246 scale_updated_constant12" -> "248 mul_tensor_12"  [label="(768, 1)", style=solid];
-"247 compressed_weight_updated_constant12" -> "248 mul_tensor_12"  [label="(768, 3072)", style=solid];
-"248 mul_tensor_12" -> "250 sub_tensor_12"  [label="(768, 3072)", style=solid];
-"249 zero_point_updated_constant12" -> "250 sub_tensor_12"  [label="(768, 1)", style=solid];
-"250 sub_tensor_12" -> "251 linear_11"  [label="(768, 3072)", style=solid];
-"251 linear_11" -> "252 dropout_9"  [label="(1, 197, 768)", style=solid];
-"252 dropout_9" -> "253 add_6"  [label="(1, 197, 768)", style=solid];
-"253 add_6" -> "256 layer_norm_6"  [label="(1, 197, 768)", style=solid];
-"253 add_6" -> "305 add_7"  [label="(1, 197, 768)", style=solid];
-"254 _param_constant40" -> "256 layer_norm_6"  [label="(768,)", style=solid];
-"255 _param_constant41" -> "256 layer_norm_6"  [label="(768,)", style=solid];
-"256 layer_norm_6" -> "257 transpose_18"  [label="(1, 197, 768)", style=solid];
-"257 transpose_18" -> "258 transpose_18_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"258 transpose_18_0_0_nncf_smooth_quant_0" -> "259 quantize_per_tensor_default_19"  [label="(197, 1, 768)", style=solid];
-"259 quantize_per_tensor_default_19" -> "260 dequantize_per_tensor_default_19"  [label="(197, 1, 768)", style=solid];
-"260 dequantize_per_tensor_default_19" -> "267 linear_12"  [label="(197, 1, 768)", style=solid];
-"261 _param_constant43_0_0" -> "267 linear_12"  [label="(2304,)", style=solid];
-"262 scale_updated_constant13" -> "264 mul_tensor_13"  [label="(2304, 1)", style=solid];
-"263 compressed_weight_updated_constant13" -> "264 mul_tensor_13"  [label="(2304, 768)", style=solid];
-"264 mul_tensor_13" -> "266 sub_tensor_13"  [label="(2304, 768)", style=solid];
-"265 zero_point_updated_constant13" -> "266 sub_tensor_13"  [label="(2304, 1)", style=solid];
-"266 sub_tensor_13" -> "267 linear_12"  [label="(2304, 768)", style=solid];
-"267 linear_12" -> "268 unflatten_3"  [label="(197, 1, 2304)", style=solid];
-"268 unflatten_3" -> "269 unsqueeze_3"  [label="(197, 1, 3, 768)", style=solid];
-"269 unsqueeze_3" -> "270 transpose_19"  [label="(1, 197, 1, 3, 768)", style=solid];
-"270 transpose_19" -> "271 squeeze_3"  [label="(3, 197, 1, 1, 768)", style=solid];
-"271 squeeze_3" -> "272 contiguous_3"  [label="(3, 197, 1, 768)", style=solid];
-"272 contiguous_3" -> "273 quantize_per_tensor_default_20"  [label="(3, 197, 1, 768)", style=solid];
-"272 contiguous_3" -> "276 quantize_per_tensor_default_21"  [label="(3, 197, 1, 768)", style=solid];
-"272 contiguous_3" -> "279 select_11"  [label="(3, 197, 1, 768)", style=solid];
-"273 quantize_per_tensor_default_20" -> "274 dequantize_per_tensor_default_20"  [label="(3, 197, 1, 768)", style=solid];
-"274 dequantize_per_tensor_default_20" -> "275 select_9"  [label="(3, 197, 1, 768)", style=solid];
-"275 select_9" -> "280 view_24"  [label="(197, 1, 768)", style=solid];
-"276 quantize_per_tensor_default_21" -> "277 dequantize_per_tensor_default_21"  [label="(3, 197, 1, 768)", style=solid];
-"277 dequantize_per_tensor_default_21" -> "278 select_10"  [label="(3, 197, 1, 768)", style=solid];
-"278 select_10" -> "282 view_25"  [label="(197, 1, 768)", style=solid];
-"279 select_11" -> "284 view_26"  [label="(197, 1, 768)", style=solid];
-"280 view_24" -> "281 transpose_20"  [label="(197, 12, 64)", style=solid];
-"281 transpose_20" -> "286 view_27"  [label="(12, 197, 64)", style=solid];
-"282 view_25" -> "283 transpose_21"  [label="(197, 12, 64)", style=solid];
-"283 transpose_21" -> "287 view_28"  [label="(12, 197, 64)", style=solid];
-"284 view_26" -> "285 transpose_22"  [label="(197, 12, 64)", style=solid];
-"285 transpose_22" -> "288 view_29"  [label="(12, 197, 64)", style=solid];
-"286 view_27" -> "289 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
-"287 view_28" -> "289 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
-"288 view_29" -> "289 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
-"289 scaled_dot_product_attention_3" -> "290 permute_4"  [label="(1, 12, 197, 64)", style=solid];
-"290 permute_4" -> "291 view_30"  [label="(197, 1, 12, 64)", style=solid];
-"291 view_30" -> "292 view_30_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"292 view_30_0_0_nncf_smooth_quant_0" -> "293 quantize_per_tensor_default_22"  [label="(197, 768)", style=solid];
-"293 quantize_per_tensor_default_22" -> "294 dequantize_per_tensor_default_22"  [label="(197, 768)", style=solid];
-"294 dequantize_per_tensor_default_22" -> "301 linear_13"  [label="(197, 768)", style=solid];
-"295 _param_constant45_0_0" -> "301 linear_13"  [label="(768,)", style=solid];
-"296 scale_updated_constant14" -> "298 mul_tensor_14"  [label="(768, 1)", style=solid];
-"297 compressed_weight_updated_constant14" -> "298 mul_tensor_14"  [label="(768, 768)", style=solid];
-"298 mul_tensor_14" -> "300 sub_tensor_14"  [label="(768, 768)", style=solid];
-"299 zero_point_updated_constant14" -> "300 sub_tensor_14"  [label="(768, 1)", style=solid];
-"300 sub_tensor_14" -> "301 linear_13"  [label="(768, 768)", style=solid];
-"301 linear_13" -> "302 view_31"  [label="(197, 768)", style=solid];
-"302 view_31" -> "303 transpose_23"  [label="(197, 1, 768)", style=solid];
-"303 transpose_23" -> "304 dropout_10"  [label="(1, 197, 768)", style=solid];
-"304 dropout_10" -> "305 add_7"  [label="(1, 197, 768)", style=solid];
-"305 add_7" -> "308 layer_norm_7"  [label="(1, 197, 768)", style=solid];
-"305 add_7" -> "332 add_8"  [label="(1, 197, 768)", style=solid];
-"306 _param_constant46" -> "308 layer_norm_7"  [label="(768,)", style=solid];
-"307 _param_constant47" -> "308 layer_norm_7"  [label="(768,)", style=solid];
-"308 layer_norm_7" -> "309 layer_norm_7_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"309 layer_norm_7_0_0_nncf_smooth_quant_0" -> "310 quantize_per_tensor_default_23"  [label="(1, 197, 768)", style=solid];
-"310 quantize_per_tensor_default_23" -> "311 dequantize_per_tensor_default_23"  [label="(1, 197, 768)", style=solid];
-"311 dequantize_per_tensor_default_23" -> "318 linear_14"  [label="(1, 197, 768)", style=solid];
-"312 _param_constant49_0_0" -> "318 linear_14"  [label="(3072,)", style=solid];
-"313 scale_updated_constant15" -> "315 mul_tensor_15"  [label="(3072, 1)", style=solid];
-"314 compressed_weight_updated_constant15" -> "315 mul_tensor_15"  [label="(3072, 768)", style=solid];
-"315 mul_tensor_15" -> "317 sub_tensor_15"  [label="(3072, 768)", style=solid];
-"316 zero_point_updated_constant15" -> "317 sub_tensor_15"  [label="(3072, 1)", style=solid];
-"317 sub_tensor_15" -> "318 linear_14"  [label="(3072, 768)", style=solid];
-"318 linear_14" -> "319 gelu_3"  [label="(1, 197, 3072)", style=solid];
-"319 gelu_3" -> "320 dropout_11"  [label="(1, 197, 3072)", style=solid];
-"320 dropout_11" -> "321 dropout_11_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"321 dropout_11_0_0_nncf_smooth_quant_0" -> "322 quantize_per_tensor_default_24"  [label="(1, 197, 3072)", style=solid];
-"322 quantize_per_tensor_default_24" -> "323 dequantize_per_tensor_default_24"  [label="(1, 197, 3072)", style=solid];
-"323 dequantize_per_tensor_default_24" -> "330 linear_15"  [label="(1, 197, 3072)", style=solid];
-"324 _param_constant51_0_0" -> "330 linear_15"  [label="(768,)", style=solid];
-"325 scale_updated_constant16" -> "327 mul_tensor_16"  [label="(768, 1)", style=solid];
-"326 compressed_weight_updated_constant16" -> "327 mul_tensor_16"  [label="(768, 3072)", style=solid];
-"327 mul_tensor_16" -> "329 sub_tensor_16"  [label="(768, 3072)", style=solid];
-"328 zero_point_updated_constant16" -> "329 sub_tensor_16"  [label="(768, 1)", style=solid];
-"329 sub_tensor_16" -> "330 linear_15"  [label="(768, 3072)", style=solid];
-"330 linear_15" -> "331 dropout_12"  [label="(1, 197, 768)", style=solid];
-"331 dropout_12" -> "332 add_8"  [label="(1, 197, 768)", style=solid];
-"332 add_8" -> "335 layer_norm_8"  [label="(1, 197, 768)", style=solid];
-"332 add_8" -> "384 add_9"  [label="(1, 197, 768)", style=solid];
-"333 _param_constant52" -> "335 layer_norm_8"  [label="(768,)", style=solid];
-"334 _param_constant53" -> "335 layer_norm_8"  [label="(768,)", style=solid];
-"335 layer_norm_8" -> "336 transpose_24"  [label="(1, 197, 768)", style=solid];
-"336 transpose_24" -> "337 transpose_24_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"337 transpose_24_0_0_nncf_smooth_quant_0" -> "338 quantize_per_tensor_default_25"  [label="(197, 1, 768)", style=solid];
-"338 quantize_per_tensor_default_25" -> "339 dequantize_per_tensor_default_25"  [label="(197, 1, 768)", style=solid];
-"339 dequantize_per_tensor_default_25" -> "346 linear_16"  [label="(197, 1, 768)", style=solid];
-"340 _param_constant55_0_0" -> "346 linear_16"  [label="(2304,)", style=solid];
-"341 scale_updated_constant17" -> "343 mul_tensor_17"  [label="(2304, 1)", style=solid];
-"342 compressed_weight_updated_constant17" -> "343 mul_tensor_17"  [label="(2304, 768)", style=solid];
-"343 mul_tensor_17" -> "345 sub_tensor_17"  [label="(2304, 768)", style=solid];
-"344 zero_point_updated_constant17" -> "345 sub_tensor_17"  [label="(2304, 1)", style=solid];
-"345 sub_tensor_17" -> "346 linear_16"  [label="(2304, 768)", style=solid];
-"346 linear_16" -> "347 unflatten_4"  [label="(197, 1, 2304)", style=solid];
-"347 unflatten_4" -> "348 unsqueeze_4"  [label="(197, 1, 3, 768)", style=solid];
-"348 unsqueeze_4" -> "349 transpose_25"  [label="(1, 197, 1, 3, 768)", style=solid];
-"349 transpose_25" -> "350 squeeze_4"  [label="(3, 197, 1, 1, 768)", style=solid];
-"350 squeeze_4" -> "351 contiguous_4"  [label="(3, 197, 1, 768)", style=solid];
-"351 contiguous_4" -> "352 quantize_per_tensor_default_26"  [label="(3, 197, 1, 768)", style=solid];
-"351 contiguous_4" -> "355 quantize_per_tensor_default_27"  [label="(3, 197, 1, 768)", style=solid];
-"351 contiguous_4" -> "358 select_14"  [label="(3, 197, 1, 768)", style=solid];
-"352 quantize_per_tensor_default_26" -> "353 dequantize_per_tensor_default_26"  [label="(3, 197, 1, 768)", style=solid];
-"353 dequantize_per_tensor_default_26" -> "354 select_12"  [label="(3, 197, 1, 768)", style=solid];
-"354 select_12" -> "359 view_32"  [label="(197, 1, 768)", style=solid];
-"355 quantize_per_tensor_default_27" -> "356 dequantize_per_tensor_default_27"  [label="(3, 197, 1, 768)", style=solid];
-"356 dequantize_per_tensor_default_27" -> "357 select_13"  [label="(3, 197, 1, 768)", style=solid];
-"357 select_13" -> "361 view_33"  [label="(197, 1, 768)", style=solid];
-"358 select_14" -> "363 view_34"  [label="(197, 1, 768)", style=solid];
-"359 view_32" -> "360 transpose_26"  [label="(197, 12, 64)", style=solid];
-"360 transpose_26" -> "365 view_35"  [label="(12, 197, 64)", style=solid];
-"361 view_33" -> "362 transpose_27"  [label="(197, 12, 64)", style=solid];
-"362 transpose_27" -> "366 view_36"  [label="(12, 197, 64)", style=solid];
-"363 view_34" -> "364 transpose_28"  [label="(197, 12, 64)", style=solid];
-"364 transpose_28" -> "367 view_37"  [label="(12, 197, 64)", style=solid];
-"365 view_35" -> "368 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
-"366 view_36" -> "368 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
-"367 view_37" -> "368 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
-"368 scaled_dot_product_attention_4" -> "369 permute_5"  [label="(1, 12, 197, 64)", style=solid];
-"369 permute_5" -> "370 view_38"  [label="(197, 1, 12, 64)", style=solid];
-"370 view_38" -> "371 view_38_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"371 view_38_0_0_nncf_smooth_quant_0" -> "372 quantize_per_tensor_default_28"  [label="(197, 768)", style=solid];
-"372 quantize_per_tensor_default_28" -> "373 dequantize_per_tensor_default_28"  [label="(197, 768)", style=solid];
-"373 dequantize_per_tensor_default_28" -> "380 linear_17"  [label="(197, 768)", style=solid];
-"374 _param_constant57_0_0" -> "380 linear_17"  [label="(768,)", style=solid];
-"375 scale_updated_constant18" -> "377 mul_tensor_18"  [label="(768, 1)", style=solid];
-"376 compressed_weight_updated_constant18" -> "377 mul_tensor_18"  [label="(768, 768)", style=solid];
-"377 mul_tensor_18" -> "379 sub_tensor_18"  [label="(768, 768)", style=solid];
-"378 zero_point_updated_constant18" -> "379 sub_tensor_18"  [label="(768, 1)", style=solid];
-"379 sub_tensor_18" -> "380 linear_17"  [label="(768, 768)", style=solid];
-"380 linear_17" -> "381 view_39"  [label="(197, 768)", style=solid];
-"381 view_39" -> "382 transpose_29"  [label="(197, 1, 768)", style=solid];
-"382 transpose_29" -> "383 dropout_13"  [label="(1, 197, 768)", style=solid];
-"383 dropout_13" -> "384 add_9"  [label="(1, 197, 768)", style=solid];
-"384 add_9" -> "387 layer_norm_9"  [label="(1, 197, 768)", style=solid];
-"384 add_9" -> "411 add_10"  [label="(1, 197, 768)", style=solid];
-"385 _param_constant58" -> "387 layer_norm_9"  [label="(768,)", style=solid];
-"386 _param_constant59" -> "387 layer_norm_9"  [label="(768,)", style=solid];
-"387 layer_norm_9" -> "388 layer_norm_9_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"388 layer_norm_9_0_0_nncf_smooth_quant_0" -> "389 quantize_per_tensor_default_29"  [label="(1, 197, 768)", style=solid];
-"389 quantize_per_tensor_default_29" -> "390 dequantize_per_tensor_default_29"  [label="(1, 197, 768)", style=solid];
-"390 dequantize_per_tensor_default_29" -> "397 linear_18"  [label="(1, 197, 768)", style=solid];
-"391 _param_constant61_0_0" -> "397 linear_18"  [label="(3072,)", style=solid];
-"392 scale_updated_constant19" -> "394 mul_tensor_19"  [label="(3072, 1)", style=solid];
-"393 compressed_weight_updated_constant19" -> "394 mul_tensor_19"  [label="(3072, 768)", style=solid];
-"394 mul_tensor_19" -> "396 sub_tensor_19"  [label="(3072, 768)", style=solid];
-"395 zero_point_updated_constant19" -> "396 sub_tensor_19"  [label="(3072, 1)", style=solid];
-"396 sub_tensor_19" -> "397 linear_18"  [label="(3072, 768)", style=solid];
-"397 linear_18" -> "398 gelu_4"  [label="(1, 197, 3072)", style=solid];
-"398 gelu_4" -> "399 dropout_14"  [label="(1, 197, 3072)", style=solid];
-"399 dropout_14" -> "400 dropout_14_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"400 dropout_14_0_0_nncf_smooth_quant_0" -> "401 quantize_per_tensor_default_30"  [label="(1, 197, 3072)", style=solid];
-"401 quantize_per_tensor_default_30" -> "402 dequantize_per_tensor_default_30"  [label="(1, 197, 3072)", style=solid];
-"402 dequantize_per_tensor_default_30" -> "409 linear_19"  [label="(1, 197, 3072)", style=solid];
-"403 _param_constant63_0_0" -> "409 linear_19"  [label="(768,)", style=solid];
-"404 scale_updated_constant20" -> "406 mul_tensor_20"  [label="(768, 1)", style=solid];
-"405 compressed_weight_updated_constant20" -> "406 mul_tensor_20"  [label="(768, 3072)", style=solid];
-"406 mul_tensor_20" -> "408 sub_tensor_20"  [label="(768, 3072)", style=solid];
-"407 zero_point_updated_constant20" -> "408 sub_tensor_20"  [label="(768, 1)", style=solid];
-"408 sub_tensor_20" -> "409 linear_19"  [label="(768, 3072)", style=solid];
-"409 linear_19" -> "410 dropout_15"  [label="(1, 197, 768)", style=solid];
-"410 dropout_15" -> "411 add_10"  [label="(1, 197, 768)", style=solid];
-"411 add_10" -> "414 layer_norm_10"  [label="(1, 197, 768)", style=solid];
-"411 add_10" -> "463 add_11"  [label="(1, 197, 768)", style=solid];
-"412 _param_constant64" -> "414 layer_norm_10"  [label="(768,)", style=solid];
-"413 _param_constant65" -> "414 layer_norm_10"  [label="(768,)", style=solid];
-"414 layer_norm_10" -> "415 transpose_30"  [label="(1, 197, 768)", style=solid];
-"415 transpose_30" -> "416 transpose_30_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"416 transpose_30_0_0_nncf_smooth_quant_0" -> "417 quantize_per_tensor_default_31"  [label="(197, 1, 768)", style=solid];
-"417 quantize_per_tensor_default_31" -> "418 dequantize_per_tensor_default_31"  [label="(197, 1, 768)", style=solid];
-"418 dequantize_per_tensor_default_31" -> "425 linear_20"  [label="(197, 1, 768)", style=solid];
-"419 _param_constant67_0_0" -> "425 linear_20"  [label="(2304,)", style=solid];
-"420 scale_updated_constant21" -> "422 mul_tensor_21"  [label="(2304, 1)", style=solid];
-"421 compressed_weight_updated_constant21" -> "422 mul_tensor_21"  [label="(2304, 768)", style=solid];
-"422 mul_tensor_21" -> "424 sub_tensor_21"  [label="(2304, 768)", style=solid];
-"423 zero_point_updated_constant21" -> "424 sub_tensor_21"  [label="(2304, 1)", style=solid];
-"424 sub_tensor_21" -> "425 linear_20"  [label="(2304, 768)", style=solid];
-"425 linear_20" -> "426 unflatten_5"  [label="(197, 1, 2304)", style=solid];
-"426 unflatten_5" -> "427 unsqueeze_5"  [label="(197, 1, 3, 768)", style=solid];
-"427 unsqueeze_5" -> "428 transpose_31"  [label="(1, 197, 1, 3, 768)", style=solid];
-"428 transpose_31" -> "429 squeeze_5"  [label="(3, 197, 1, 1, 768)", style=solid];
-"429 squeeze_5" -> "430 contiguous_5"  [label="(3, 197, 1, 768)", style=solid];
-"430 contiguous_5" -> "431 quantize_per_tensor_default_32"  [label="(3, 197, 1, 768)", style=solid];
-"430 contiguous_5" -> "434 quantize_per_tensor_default_33"  [label="(3, 197, 1, 768)", style=solid];
-"430 contiguous_5" -> "437 select_17"  [label="(3, 197, 1, 768)", style=solid];
-"431 quantize_per_tensor_default_32" -> "432 dequantize_per_tensor_default_32"  [label="(3, 197, 1, 768)", style=solid];
-"432 dequantize_per_tensor_default_32" -> "433 select_15"  [label="(3, 197, 1, 768)", style=solid];
-"433 select_15" -> "438 view_40"  [label="(197, 1, 768)", style=solid];
-"434 quantize_per_tensor_default_33" -> "435 dequantize_per_tensor_default_33"  [label="(3, 197, 1, 768)", style=solid];
-"435 dequantize_per_tensor_default_33" -> "436 select_16"  [label="(3, 197, 1, 768)", style=solid];
-"436 select_16" -> "440 view_41"  [label="(197, 1, 768)", style=solid];
-"437 select_17" -> "442 view_42"  [label="(197, 1, 768)", style=solid];
-"438 view_40" -> "439 transpose_32"  [label="(197, 12, 64)", style=solid];
-"439 transpose_32" -> "444 view_43"  [label="(12, 197, 64)", style=solid];
-"440 view_41" -> "441 transpose_33"  [label="(197, 12, 64)", style=solid];
-"441 transpose_33" -> "445 view_44"  [label="(12, 197, 64)", style=solid];
-"442 view_42" -> "443 transpose_34"  [label="(197, 12, 64)", style=solid];
-"443 transpose_34" -> "446 view_45"  [label="(12, 197, 64)", style=solid];
-"444 view_43" -> "447 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
-"445 view_44" -> "447 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
-"446 view_45" -> "447 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
-"447 scaled_dot_product_attention_5" -> "448 permute_6"  [label="(1, 12, 197, 64)", style=solid];
-"448 permute_6" -> "449 view_46"  [label="(197, 1, 12, 64)", style=solid];
-"449 view_46" -> "450 view_46_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"450 view_46_0_0_nncf_smooth_quant_0" -> "451 quantize_per_tensor_default_34"  [label="(197, 768)", style=solid];
-"451 quantize_per_tensor_default_34" -> "452 dequantize_per_tensor_default_34"  [label="(197, 768)", style=solid];
-"452 dequantize_per_tensor_default_34" -> "459 linear_21"  [label="(197, 768)", style=solid];
-"453 _param_constant69_0_0" -> "459 linear_21"  [label="(768,)", style=solid];
-"454 scale_updated_constant22" -> "456 mul_tensor_22"  [label="(768, 1)", style=solid];
-"455 compressed_weight_updated_constant22" -> "456 mul_tensor_22"  [label="(768, 768)", style=solid];
-"456 mul_tensor_22" -> "458 sub_tensor_22"  [label="(768, 768)", style=solid];
-"457 zero_point_updated_constant22" -> "458 sub_tensor_22"  [label="(768, 1)", style=solid];
-"458 sub_tensor_22" -> "459 linear_21"  [label="(768, 768)", style=solid];
-"459 linear_21" -> "460 view_47"  [label="(197, 768)", style=solid];
-"460 view_47" -> "461 transpose_35"  [label="(197, 1, 768)", style=solid];
-"461 transpose_35" -> "462 dropout_16"  [label="(1, 197, 768)", style=solid];
-"462 dropout_16" -> "463 add_11"  [label="(1, 197, 768)", style=solid];
-"463 add_11" -> "466 layer_norm_11"  [label="(1, 197, 768)", style=solid];
-"463 add_11" -> "490 add_12"  [label="(1, 197, 768)", style=solid];
-"464 _param_constant70" -> "466 layer_norm_11"  [label="(768,)", style=solid];
-"465 _param_constant71" -> "466 layer_norm_11"  [label="(768,)", style=solid];
-"466 layer_norm_11" -> "467 layer_norm_11_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"467 layer_norm_11_0_0_nncf_smooth_quant_0" -> "468 quantize_per_tensor_default_35"  [label="(1, 197, 768)", style=solid];
-"468 quantize_per_tensor_default_35" -> "469 dequantize_per_tensor_default_35"  [label="(1, 197, 768)", style=solid];
-"469 dequantize_per_tensor_default_35" -> "476 linear_22"  [label="(1, 197, 768)", style=solid];
-"470 _param_constant73_0_0" -> "476 linear_22"  [label="(3072,)", style=solid];
-"471 scale_updated_constant23" -> "473 mul_tensor_23"  [label="(3072, 1)", style=solid];
-"472 compressed_weight_updated_constant23" -> "473 mul_tensor_23"  [label="(3072, 768)", style=solid];
-"473 mul_tensor_23" -> "475 sub_tensor_23"  [label="(3072, 768)", style=solid];
-"474 zero_point_updated_constant23" -> "475 sub_tensor_23"  [label="(3072, 1)", style=solid];
-"475 sub_tensor_23" -> "476 linear_22"  [label="(3072, 768)", style=solid];
-"476 linear_22" -> "477 gelu_5"  [label="(1, 197, 3072)", style=solid];
-"477 gelu_5" -> "478 dropout_17"  [label="(1, 197, 3072)", style=solid];
-"478 dropout_17" -> "479 dropout_17_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"479 dropout_17_0_0_nncf_smooth_quant_0" -> "480 quantize_per_tensor_default_36"  [label="(1, 197, 3072)", style=solid];
-"480 quantize_per_tensor_default_36" -> "481 dequantize_per_tensor_default_36"  [label="(1, 197, 3072)", style=solid];
-"481 dequantize_per_tensor_default_36" -> "488 linear_23"  [label="(1, 197, 3072)", style=solid];
-"482 _param_constant75_0_0" -> "488 linear_23"  [label="(768,)", style=solid];
-"483 scale_updated_constant24" -> "485 mul_tensor_24"  [label="(768, 1)", style=solid];
-"484 compressed_weight_updated_constant24" -> "485 mul_tensor_24"  [label="(768, 3072)", style=solid];
-"485 mul_tensor_24" -> "487 sub_tensor_24"  [label="(768, 3072)", style=solid];
-"486 zero_point_updated_constant24" -> "487 sub_tensor_24"  [label="(768, 1)", style=solid];
-"487 sub_tensor_24" -> "488 linear_23"  [label="(768, 3072)", style=solid];
-"488 linear_23" -> "489 dropout_18"  [label="(1, 197, 768)", style=solid];
-"489 dropout_18" -> "490 add_12"  [label="(1, 197, 768)", style=solid];
-"490 add_12" -> "493 layer_norm_12"  [label="(1, 197, 768)", style=solid];
-"490 add_12" -> "542 add_13"  [label="(1, 197, 768)", style=solid];
-"491 _param_constant76" -> "493 layer_norm_12"  [label="(768,)", style=solid];
-"492 _param_constant77" -> "493 layer_norm_12"  [label="(768,)", style=solid];
-"493 layer_norm_12" -> "494 transpose_36"  [label="(1, 197, 768)", style=solid];
-"494 transpose_36" -> "495 transpose_36_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"495 transpose_36_0_0_nncf_smooth_quant_0" -> "496 quantize_per_tensor_default_37"  [label="(197, 1, 768)", style=solid];
-"496 quantize_per_tensor_default_37" -> "497 dequantize_per_tensor_default_37"  [label="(197, 1, 768)", style=solid];
-"497 dequantize_per_tensor_default_37" -> "504 linear_24"  [label="(197, 1, 768)", style=solid];
-"498 _param_constant79_0_0" -> "504 linear_24"  [label="(2304,)", style=solid];
-"499 scale_updated_constant25" -> "501 mul_tensor_25"  [label="(2304, 1)", style=solid];
-"500 compressed_weight_updated_constant25" -> "501 mul_tensor_25"  [label="(2304, 768)", style=solid];
-"501 mul_tensor_25" -> "503 sub_tensor_25"  [label="(2304, 768)", style=solid];
-"502 zero_point_updated_constant25" -> "503 sub_tensor_25"  [label="(2304, 1)", style=solid];
-"503 sub_tensor_25" -> "504 linear_24"  [label="(2304, 768)", style=solid];
-"504 linear_24" -> "505 unflatten_6"  [label="(197, 1, 2304)", style=solid];
-"505 unflatten_6" -> "506 unsqueeze_6"  [label="(197, 1, 3, 768)", style=solid];
-"506 unsqueeze_6" -> "507 transpose_37"  [label="(1, 197, 1, 3, 768)", style=solid];
-"507 transpose_37" -> "508 squeeze_6"  [label="(3, 197, 1, 1, 768)", style=solid];
-"508 squeeze_6" -> "509 contiguous_6"  [label="(3, 197, 1, 768)", style=solid];
-"509 contiguous_6" -> "510 quantize_per_tensor_default_38"  [label="(3, 197, 1, 768)", style=solid];
-"509 contiguous_6" -> "513 quantize_per_tensor_default_39"  [label="(3, 197, 1, 768)", style=solid];
-"509 contiguous_6" -> "516 select_20"  [label="(3, 197, 1, 768)", style=solid];
-"510 quantize_per_tensor_default_38" -> "511 dequantize_per_tensor_default_38"  [label="(3, 197, 1, 768)", style=solid];
-"511 dequantize_per_tensor_default_38" -> "512 select_18"  [label="(3, 197, 1, 768)", style=solid];
-"512 select_18" -> "517 view_48"  [label="(197, 1, 768)", style=solid];
-"513 quantize_per_tensor_default_39" -> "514 dequantize_per_tensor_default_39"  [label="(3, 197, 1, 768)", style=solid];
-"514 dequantize_per_tensor_default_39" -> "515 select_19"  [label="(3, 197, 1, 768)", style=solid];
-"515 select_19" -> "519 view_49"  [label="(197, 1, 768)", style=solid];
-"516 select_20" -> "521 view_50"  [label="(197, 1, 768)", style=solid];
-"517 view_48" -> "518 transpose_38"  [label="(197, 12, 64)", style=solid];
-"518 transpose_38" -> "523 view_51"  [label="(12, 197, 64)", style=solid];
-"519 view_49" -> "520 transpose_39"  [label="(197, 12, 64)", style=solid];
-"520 transpose_39" -> "524 view_52"  [label="(12, 197, 64)", style=solid];
-"521 view_50" -> "522 transpose_40"  [label="(197, 12, 64)", style=solid];
-"522 transpose_40" -> "525 view_53"  [label="(12, 197, 64)", style=solid];
-"523 view_51" -> "526 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
-"524 view_52" -> "526 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
-"525 view_53" -> "526 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
-"526 scaled_dot_product_attention_6" -> "527 permute_7"  [label="(1, 12, 197, 64)", style=solid];
-"527 permute_7" -> "528 view_54"  [label="(197, 1, 12, 64)", style=solid];
-"528 view_54" -> "529 view_54_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"529 view_54_0_0_nncf_smooth_quant_0" -> "530 quantize_per_tensor_default_40"  [label="(197, 768)", style=solid];
-"530 quantize_per_tensor_default_40" -> "531 dequantize_per_tensor_default_40"  [label="(197, 768)", style=solid];
-"531 dequantize_per_tensor_default_40" -> "538 linear_25"  [label="(197, 768)", style=solid];
-"532 _param_constant81_0_0" -> "538 linear_25"  [label="(768,)", style=solid];
-"533 scale_updated_constant26" -> "535 mul_tensor_26"  [label="(768, 1)", style=solid];
-"534 compressed_weight_updated_constant26" -> "535 mul_tensor_26"  [label="(768, 768)", style=solid];
-"535 mul_tensor_26" -> "537 sub_tensor_26"  [label="(768, 768)", style=solid];
-"536 zero_point_updated_constant26" -> "537 sub_tensor_26"  [label="(768, 1)", style=solid];
-"537 sub_tensor_26" -> "538 linear_25"  [label="(768, 768)", style=solid];
-"538 linear_25" -> "539 view_55"  [label="(197, 768)", style=solid];
-"539 view_55" -> "540 transpose_41"  [label="(197, 1, 768)", style=solid];
-"540 transpose_41" -> "541 dropout_19"  [label="(1, 197, 768)", style=solid];
-"541 dropout_19" -> "542 add_13"  [label="(1, 197, 768)", style=solid];
-"542 add_13" -> "545 layer_norm_13"  [label="(1, 197, 768)", style=solid];
-"542 add_13" -> "569 add_14"  [label="(1, 197, 768)", style=solid];
-"543 _param_constant82" -> "545 layer_norm_13"  [label="(768,)", style=solid];
-"544 _param_constant83" -> "545 layer_norm_13"  [label="(768,)", style=solid];
-"545 layer_norm_13" -> "546 layer_norm_13_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"546 layer_norm_13_0_0_nncf_smooth_quant_0" -> "547 quantize_per_tensor_default_41"  [label="(1, 197, 768)", style=solid];
-"547 quantize_per_tensor_default_41" -> "548 dequantize_per_tensor_default_41"  [label="(1, 197, 768)", style=solid];
-"548 dequantize_per_tensor_default_41" -> "555 linear_26"  [label="(1, 197, 768)", style=solid];
-"549 _param_constant85_0_0" -> "555 linear_26"  [label="(3072,)", style=solid];
-"550 scale_updated_constant27" -> "552 mul_tensor_27"  [label="(3072, 1)", style=solid];
-"551 compressed_weight_updated_constant27" -> "552 mul_tensor_27"  [label="(3072, 768)", style=solid];
-"552 mul_tensor_27" -> "554 sub_tensor_27"  [label="(3072, 768)", style=solid];
-"553 zero_point_updated_constant27" -> "554 sub_tensor_27"  [label="(3072, 1)", style=solid];
-"554 sub_tensor_27" -> "555 linear_26"  [label="(3072, 768)", style=solid];
-"555 linear_26" -> "556 gelu_6"  [label="(1, 197, 3072)", style=solid];
-"556 gelu_6" -> "557 dropout_20"  [label="(1, 197, 3072)", style=solid];
-"557 dropout_20" -> "558 dropout_20_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"558 dropout_20_0_0_nncf_smooth_quant_0" -> "559 quantize_per_tensor_default_42"  [label="(1, 197, 3072)", style=solid];
-"559 quantize_per_tensor_default_42" -> "560 dequantize_per_tensor_default_42"  [label="(1, 197, 3072)", style=solid];
-"560 dequantize_per_tensor_default_42" -> "567 linear_27"  [label="(1, 197, 3072)", style=solid];
-"561 _param_constant87_0_0" -> "567 linear_27"  [label="(768,)", style=solid];
-"562 scale_updated_constant28" -> "564 mul_tensor_28"  [label="(768, 1)", style=solid];
-"563 compressed_weight_updated_constant28" -> "564 mul_tensor_28"  [label="(768, 3072)", style=solid];
-"564 mul_tensor_28" -> "566 sub_tensor_28"  [label="(768, 3072)", style=solid];
-"565 zero_point_updated_constant28" -> "566 sub_tensor_28"  [label="(768, 1)", style=solid];
-"566 sub_tensor_28" -> "567 linear_27"  [label="(768, 3072)", style=solid];
-"567 linear_27" -> "568 dropout_21"  [label="(1, 197, 768)", style=solid];
-"568 dropout_21" -> "569 add_14"  [label="(1, 197, 768)", style=solid];
-"569 add_14" -> "572 layer_norm_14"  [label="(1, 197, 768)", style=solid];
-"569 add_14" -> "621 add_15"  [label="(1, 197, 768)", style=solid];
-"570 _param_constant88" -> "572 layer_norm_14"  [label="(768,)", style=solid];
-"571 _param_constant89" -> "572 layer_norm_14"  [label="(768,)", style=solid];
-"572 layer_norm_14" -> "573 transpose_42"  [label="(1, 197, 768)", style=solid];
-"573 transpose_42" -> "574 transpose_42_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"574 transpose_42_0_0_nncf_smooth_quant_0" -> "575 quantize_per_tensor_default_43"  [label="(197, 1, 768)", style=solid];
-"575 quantize_per_tensor_default_43" -> "576 dequantize_per_tensor_default_43"  [label="(197, 1, 768)", style=solid];
-"576 dequantize_per_tensor_default_43" -> "583 linear_28"  [label="(197, 1, 768)", style=solid];
-"577 _param_constant91_0_0" -> "583 linear_28"  [label="(2304,)", style=solid];
-"578 scale_updated_constant29" -> "580 mul_tensor_29"  [label="(2304, 1)", style=solid];
-"579 compressed_weight_updated_constant29" -> "580 mul_tensor_29"  [label="(2304, 768)", style=solid];
-"580 mul_tensor_29" -> "582 sub_tensor_29"  [label="(2304, 768)", style=solid];
-"581 zero_point_updated_constant29" -> "582 sub_tensor_29"  [label="(2304, 1)", style=solid];
-"582 sub_tensor_29" -> "583 linear_28"  [label="(2304, 768)", style=solid];
-"583 linear_28" -> "584 unflatten_7"  [label="(197, 1, 2304)", style=solid];
-"584 unflatten_7" -> "585 unsqueeze_7"  [label="(197, 1, 3, 768)", style=solid];
-"585 unsqueeze_7" -> "586 transpose_43"  [label="(1, 197, 1, 3, 768)", style=solid];
-"586 transpose_43" -> "587 squeeze_7"  [label="(3, 197, 1, 1, 768)", style=solid];
-"587 squeeze_7" -> "588 contiguous_7"  [label="(3, 197, 1, 768)", style=solid];
-"588 contiguous_7" -> "589 quantize_per_tensor_default_44"  [label="(3, 197, 1, 768)", style=solid];
-"588 contiguous_7" -> "592 quantize_per_tensor_default_45"  [label="(3, 197, 1, 768)", style=solid];
-"588 contiguous_7" -> "595 select_23"  [label="(3, 197, 1, 768)", style=solid];
-"589 quantize_per_tensor_default_44" -> "590 dequantize_per_tensor_default_44"  [label="(3, 197, 1, 768)", style=solid];
-"590 dequantize_per_tensor_default_44" -> "591 select_21"  [label="(3, 197, 1, 768)", style=solid];
-"591 select_21" -> "596 view_56"  [label="(197, 1, 768)", style=solid];
-"592 quantize_per_tensor_default_45" -> "593 dequantize_per_tensor_default_45"  [label="(3, 197, 1, 768)", style=solid];
-"593 dequantize_per_tensor_default_45" -> "594 select_22"  [label="(3, 197, 1, 768)", style=solid];
-"594 select_22" -> "598 view_57"  [label="(197, 1, 768)", style=solid];
-"595 select_23" -> "600 view_58"  [label="(197, 1, 768)", style=solid];
-"596 view_56" -> "597 transpose_44"  [label="(197, 12, 64)", style=solid];
-"597 transpose_44" -> "602 view_59"  [label="(12, 197, 64)", style=solid];
-"598 view_57" -> "599 transpose_45"  [label="(197, 12, 64)", style=solid];
-"599 transpose_45" -> "603 view_60"  [label="(12, 197, 64)", style=solid];
-"600 view_58" -> "601 transpose_46"  [label="(197, 12, 64)", style=solid];
-"601 transpose_46" -> "604 view_61"  [label="(12, 197, 64)", style=solid];
-"602 view_59" -> "605 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
-"603 view_60" -> "605 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
-"604 view_61" -> "605 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
-"605 scaled_dot_product_attention_7" -> "606 permute_8"  [label="(1, 12, 197, 64)", style=solid];
-"606 permute_8" -> "607 view_62"  [label="(197, 1, 12, 64)", style=solid];
-"607 view_62" -> "608 view_62_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"608 view_62_0_0_nncf_smooth_quant_0" -> "609 quantize_per_tensor_default_46"  [label="(197, 768)", style=solid];
-"609 quantize_per_tensor_default_46" -> "610 dequantize_per_tensor_default_46"  [label="(197, 768)", style=solid];
-"610 dequantize_per_tensor_default_46" -> "617 linear_29"  [label="(197, 768)", style=solid];
-"611 _param_constant93_0_0" -> "617 linear_29"  [label="(768,)", style=solid];
-"612 scale_updated_constant30" -> "614 mul_tensor_30"  [label="(768, 1)", style=solid];
-"613 compressed_weight_updated_constant30" -> "614 mul_tensor_30"  [label="(768, 768)", style=solid];
-"614 mul_tensor_30" -> "616 sub_tensor_30"  [label="(768, 768)", style=solid];
-"615 zero_point_updated_constant30" -> "616 sub_tensor_30"  [label="(768, 1)", style=solid];
-"616 sub_tensor_30" -> "617 linear_29"  [label="(768, 768)", style=solid];
-"617 linear_29" -> "618 view_63"  [label="(197, 768)", style=solid];
-"618 view_63" -> "619 transpose_47"  [label="(197, 1, 768)", style=solid];
-"619 transpose_47" -> "620 dropout_22"  [label="(1, 197, 768)", style=solid];
-"620 dropout_22" -> "621 add_15"  [label="(1, 197, 768)", style=solid];
-"621 add_15" -> "624 layer_norm_15"  [label="(1, 197, 768)", style=solid];
-"621 add_15" -> "648 add_16"  [label="(1, 197, 768)", style=solid];
-"622 _param_constant94" -> "624 layer_norm_15"  [label="(768,)", style=solid];
-"623 _param_constant95" -> "624 layer_norm_15"  [label="(768,)", style=solid];
-"624 layer_norm_15" -> "625 layer_norm_15_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"625 layer_norm_15_0_0_nncf_smooth_quant_0" -> "626 quantize_per_tensor_default_47"  [label="(1, 197, 768)", style=solid];
-"626 quantize_per_tensor_default_47" -> "627 dequantize_per_tensor_default_47"  [label="(1, 197, 768)", style=solid];
-"627 dequantize_per_tensor_default_47" -> "634 linear_30"  [label="(1, 197, 768)", style=solid];
-"628 _param_constant97_0_0" -> "634 linear_30"  [label="(3072,)", style=solid];
-"629 scale_updated_constant31" -> "631 mul_tensor_31"  [label="(3072, 1)", style=solid];
-"630 compressed_weight_updated_constant31" -> "631 mul_tensor_31"  [label="(3072, 768)", style=solid];
-"631 mul_tensor_31" -> "633 sub_tensor_31"  [label="(3072, 768)", style=solid];
-"632 zero_point_updated_constant31" -> "633 sub_tensor_31"  [label="(3072, 1)", style=solid];
-"633 sub_tensor_31" -> "634 linear_30"  [label="(3072, 768)", style=solid];
-"634 linear_30" -> "635 gelu_7"  [label="(1, 197, 3072)", style=solid];
-"635 gelu_7" -> "636 dropout_23"  [label="(1, 197, 3072)", style=solid];
-"636 dropout_23" -> "637 dropout_23_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"637 dropout_23_0_0_nncf_smooth_quant_0" -> "638 quantize_per_tensor_default_48"  [label="(1, 197, 3072)", style=solid];
-"638 quantize_per_tensor_default_48" -> "639 dequantize_per_tensor_default_48"  [label="(1, 197, 3072)", style=solid];
-"639 dequantize_per_tensor_default_48" -> "646 linear_31"  [label="(1, 197, 3072)", style=solid];
-"640 _param_constant99_0_0" -> "646 linear_31"  [label="(768,)", style=solid];
-"641 scale_updated_constant32" -> "643 mul_tensor_32"  [label="(768, 1)", style=solid];
-"642 compressed_weight_updated_constant32" -> "643 mul_tensor_32"  [label="(768, 3072)", style=solid];
-"643 mul_tensor_32" -> "645 sub_tensor_32"  [label="(768, 3072)", style=solid];
-"644 zero_point_updated_constant32" -> "645 sub_tensor_32"  [label="(768, 1)", style=solid];
-"645 sub_tensor_32" -> "646 linear_31"  [label="(768, 3072)", style=solid];
-"646 linear_31" -> "647 dropout_24"  [label="(1, 197, 768)", style=solid];
-"647 dropout_24" -> "648 add_16"  [label="(1, 197, 768)", style=solid];
-"648 add_16" -> "651 layer_norm_16"  [label="(1, 197, 768)", style=solid];
-"648 add_16" -> "700 add_17"  [label="(1, 197, 768)", style=solid];
-"649 _param_constant100" -> "651 layer_norm_16"  [label="(768,)", style=solid];
-"650 _param_constant101" -> "651 layer_norm_16"  [label="(768,)", style=solid];
-"651 layer_norm_16" -> "652 transpose_48"  [label="(1, 197, 768)", style=solid];
-"652 transpose_48" -> "653 transpose_48_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"653 transpose_48_0_0_nncf_smooth_quant_0" -> "654 quantize_per_tensor_default_49"  [label="(197, 1, 768)", style=solid];
-"654 quantize_per_tensor_default_49" -> "655 dequantize_per_tensor_default_49"  [label="(197, 1, 768)", style=solid];
-"655 dequantize_per_tensor_default_49" -> "662 linear_32"  [label="(197, 1, 768)", style=solid];
-"656 _param_constant103_0_0" -> "662 linear_32"  [label="(2304,)", style=solid];
-"657 scale_updated_constant33" -> "659 mul_tensor_33"  [label="(2304, 1)", style=solid];
-"658 compressed_weight_updated_constant33" -> "659 mul_tensor_33"  [label="(2304, 768)", style=solid];
-"659 mul_tensor_33" -> "661 sub_tensor_33"  [label="(2304, 768)", style=solid];
-"660 zero_point_updated_constant33" -> "661 sub_tensor_33"  [label="(2304, 1)", style=solid];
-"661 sub_tensor_33" -> "662 linear_32"  [label="(2304, 768)", style=solid];
-"662 linear_32" -> "663 unflatten_8"  [label="(197, 1, 2304)", style=solid];
-"663 unflatten_8" -> "664 unsqueeze_8"  [label="(197, 1, 3, 768)", style=solid];
-"664 unsqueeze_8" -> "665 transpose_49"  [label="(1, 197, 1, 3, 768)", style=solid];
-"665 transpose_49" -> "666 squeeze_8"  [label="(3, 197, 1, 1, 768)", style=solid];
-"666 squeeze_8" -> "667 contiguous_8"  [label="(3, 197, 1, 768)", style=solid];
-"667 contiguous_8" -> "668 quantize_per_tensor_default_50"  [label="(3, 197, 1, 768)", style=solid];
-"667 contiguous_8" -> "671 quantize_per_tensor_default_51"  [label="(3, 197, 1, 768)", style=solid];
-"667 contiguous_8" -> "674 select_26"  [label="(3, 197, 1, 768)", style=solid];
-"668 quantize_per_tensor_default_50" -> "669 dequantize_per_tensor_default_50"  [label="(3, 197, 1, 768)", style=solid];
-"669 dequantize_per_tensor_default_50" -> "670 select_24"  [label="(3, 197, 1, 768)", style=solid];
-"670 select_24" -> "675 view_64"  [label="(197, 1, 768)", style=solid];
-"671 quantize_per_tensor_default_51" -> "672 dequantize_per_tensor_default_51"  [label="(3, 197, 1, 768)", style=solid];
-"672 dequantize_per_tensor_default_51" -> "673 select_25"  [label="(3, 197, 1, 768)", style=solid];
-"673 select_25" -> "677 view_65"  [label="(197, 1, 768)", style=solid];
-"674 select_26" -> "679 view_66"  [label="(197, 1, 768)", style=solid];
-"675 view_64" -> "676 transpose_50"  [label="(197, 12, 64)", style=solid];
-"676 transpose_50" -> "681 view_67"  [label="(12, 197, 64)", style=solid];
-"677 view_65" -> "678 transpose_51"  [label="(197, 12, 64)", style=solid];
-"678 transpose_51" -> "682 view_68"  [label="(12, 197, 64)", style=solid];
-"679 view_66" -> "680 transpose_52"  [label="(197, 12, 64)", style=solid];
-"680 transpose_52" -> "683 view_69"  [label="(12, 197, 64)", style=solid];
-"681 view_67" -> "684 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
-"682 view_68" -> "684 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
-"683 view_69" -> "684 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
-"684 scaled_dot_product_attention_8" -> "685 permute_9"  [label="(1, 12, 197, 64)", style=solid];
-"685 permute_9" -> "686 view_70"  [label="(197, 1, 12, 64)", style=solid];
-"686 view_70" -> "687 view_70_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"687 view_70_0_0_nncf_smooth_quant_0" -> "688 quantize_per_tensor_default_52"  [label="(197, 768)", style=solid];
-"688 quantize_per_tensor_default_52" -> "689 dequantize_per_tensor_default_52"  [label="(197, 768)", style=solid];
-"689 dequantize_per_tensor_default_52" -> "696 linear_33"  [label="(197, 768)", style=solid];
-"690 _param_constant105_0_0" -> "696 linear_33"  [label="(768,)", style=solid];
-"691 scale_updated_constant34" -> "693 mul_tensor_34"  [label="(768, 1)", style=solid];
-"692 compressed_weight_updated_constant34" -> "693 mul_tensor_34"  [label="(768, 768)", style=solid];
-"693 mul_tensor_34" -> "695 sub_tensor_34"  [label="(768, 768)", style=solid];
-"694 zero_point_updated_constant34" -> "695 sub_tensor_34"  [label="(768, 1)", style=solid];
-"695 sub_tensor_34" -> "696 linear_33"  [label="(768, 768)", style=solid];
-"696 linear_33" -> "697 view_71"  [label="(197, 768)", style=solid];
-"697 view_71" -> "698 transpose_53"  [label="(197, 1, 768)", style=solid];
-"698 transpose_53" -> "699 dropout_25"  [label="(1, 197, 768)", style=solid];
-"699 dropout_25" -> "700 add_17"  [label="(1, 197, 768)", style=solid];
-"700 add_17" -> "703 layer_norm_17"  [label="(1, 197, 768)", style=solid];
-"700 add_17" -> "727 add_18"  [label="(1, 197, 768)", style=solid];
-"701 _param_constant106" -> "703 layer_norm_17"  [label="(768,)", style=solid];
-"702 _param_constant107" -> "703 layer_norm_17"  [label="(768,)", style=solid];
-"703 layer_norm_17" -> "704 layer_norm_17_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"704 layer_norm_17_0_0_nncf_smooth_quant_0" -> "705 quantize_per_tensor_default_53"  [label="(1, 197, 768)", style=solid];
-"705 quantize_per_tensor_default_53" -> "706 dequantize_per_tensor_default_53"  [label="(1, 197, 768)", style=solid];
-"706 dequantize_per_tensor_default_53" -> "713 linear_34"  [label="(1, 197, 768)", style=solid];
-"707 _param_constant109_0_0" -> "713 linear_34"  [label="(3072,)", style=solid];
-"708 scale_updated_constant35" -> "710 mul_tensor_35"  [label="(3072, 1)", style=solid];
-"709 compressed_weight_updated_constant35" -> "710 mul_tensor_35"  [label="(3072, 768)", style=solid];
-"710 mul_tensor_35" -> "712 sub_tensor_35"  [label="(3072, 768)", style=solid];
-"711 zero_point_updated_constant35" -> "712 sub_tensor_35"  [label="(3072, 1)", style=solid];
-"712 sub_tensor_35" -> "713 linear_34"  [label="(3072, 768)", style=solid];
-"713 linear_34" -> "714 gelu_8"  [label="(1, 197, 3072)", style=solid];
-"714 gelu_8" -> "715 dropout_26"  [label="(1, 197, 3072)", style=solid];
-"715 dropout_26" -> "716 dropout_26_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"716 dropout_26_0_0_nncf_smooth_quant_0" -> "717 quantize_per_tensor_default_54"  [label="(1, 197, 3072)", style=solid];
-"717 quantize_per_tensor_default_54" -> "718 dequantize_per_tensor_default_54"  [label="(1, 197, 3072)", style=solid];
-"718 dequantize_per_tensor_default_54" -> "725 linear_35"  [label="(1, 197, 3072)", style=solid];
-"719 _param_constant111_0_0" -> "725 linear_35"  [label="(768,)", style=solid];
-"720 scale_updated_constant36" -> "722 mul_tensor_36"  [label="(768, 1)", style=solid];
-"721 compressed_weight_updated_constant36" -> "722 mul_tensor_36"  [label="(768, 3072)", style=solid];
-"722 mul_tensor_36" -> "724 sub_tensor_36"  [label="(768, 3072)", style=solid];
-"723 zero_point_updated_constant36" -> "724 sub_tensor_36"  [label="(768, 1)", style=solid];
-"724 sub_tensor_36" -> "725 linear_35"  [label="(768, 3072)", style=solid];
-"725 linear_35" -> "726 dropout_27"  [label="(1, 197, 768)", style=solid];
-"726 dropout_27" -> "727 add_18"  [label="(1, 197, 768)", style=solid];
-"727 add_18" -> "730 layer_norm_18"  [label="(1, 197, 768)", style=solid];
-"727 add_18" -> "779 add_19"  [label="(1, 197, 768)", style=solid];
-"728 _param_constant112" -> "730 layer_norm_18"  [label="(768,)", style=solid];
-"729 _param_constant113" -> "730 layer_norm_18"  [label="(768,)", style=solid];
-"730 layer_norm_18" -> "731 transpose_54"  [label="(1, 197, 768)", style=solid];
-"731 transpose_54" -> "732 transpose_54_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"732 transpose_54_0_0_nncf_smooth_quant_0" -> "733 quantize_per_tensor_default_55"  [label="(197, 1, 768)", style=solid];
-"733 quantize_per_tensor_default_55" -> "734 dequantize_per_tensor_default_55"  [label="(197, 1, 768)", style=solid];
-"734 dequantize_per_tensor_default_55" -> "741 linear_36"  [label="(197, 1, 768)", style=solid];
-"735 _param_constant115_0_0" -> "741 linear_36"  [label="(2304,)", style=solid];
-"736 scale_updated_constant37" -> "738 mul_tensor_37"  [label="(2304, 1)", style=solid];
-"737 compressed_weight_updated_constant37" -> "738 mul_tensor_37"  [label="(2304, 768)", style=solid];
-"738 mul_tensor_37" -> "740 sub_tensor_37"  [label="(2304, 768)", style=solid];
-"739 zero_point_updated_constant37" -> "740 sub_tensor_37"  [label="(2304, 1)", style=solid];
-"740 sub_tensor_37" -> "741 linear_36"  [label="(2304, 768)", style=solid];
-"741 linear_36" -> "742 unflatten_9"  [label="(197, 1, 2304)", style=solid];
-"742 unflatten_9" -> "743 unsqueeze_9"  [label="(197, 1, 3, 768)", style=solid];
-"743 unsqueeze_9" -> "744 transpose_55"  [label="(1, 197, 1, 3, 768)", style=solid];
-"744 transpose_55" -> "745 squeeze_9"  [label="(3, 197, 1, 1, 768)", style=solid];
-"745 squeeze_9" -> "746 contiguous_9"  [label="(3, 197, 1, 768)", style=solid];
-"746 contiguous_9" -> "747 quantize_per_tensor_default_56"  [label="(3, 197, 1, 768)", style=solid];
-"746 contiguous_9" -> "750 quantize_per_tensor_default_57"  [label="(3, 197, 1, 768)", style=solid];
-"746 contiguous_9" -> "753 select_29"  [label="(3, 197, 1, 768)", style=solid];
-"747 quantize_per_tensor_default_56" -> "748 dequantize_per_tensor_default_56"  [label="(3, 197, 1, 768)", style=solid];
-"748 dequantize_per_tensor_default_56" -> "749 select_27"  [label="(3, 197, 1, 768)", style=solid];
-"749 select_27" -> "754 view_72"  [label="(197, 1, 768)", style=solid];
-"750 quantize_per_tensor_default_57" -> "751 dequantize_per_tensor_default_57"  [label="(3, 197, 1, 768)", style=solid];
-"751 dequantize_per_tensor_default_57" -> "752 select_28"  [label="(3, 197, 1, 768)", style=solid];
-"752 select_28" -> "756 view_73"  [label="(197, 1, 768)", style=solid];
-"753 select_29" -> "758 view_74"  [label="(197, 1, 768)", style=solid];
-"754 view_72" -> "755 transpose_56"  [label="(197, 12, 64)", style=solid];
-"755 transpose_56" -> "760 view_75"  [label="(12, 197, 64)", style=solid];
-"756 view_73" -> "757 transpose_57"  [label="(197, 12, 64)", style=solid];
-"757 transpose_57" -> "761 view_76"  [label="(12, 197, 64)", style=solid];
-"758 view_74" -> "759 transpose_58"  [label="(197, 12, 64)", style=solid];
-"759 transpose_58" -> "762 view_77"  [label="(12, 197, 64)", style=solid];
-"760 view_75" -> "763 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
-"761 view_76" -> "763 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
-"762 view_77" -> "763 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
-"763 scaled_dot_product_attention_9" -> "764 permute_10"  [label="(1, 12, 197, 64)", style=solid];
-"764 permute_10" -> "765 view_78"  [label="(197, 1, 12, 64)", style=solid];
-"765 view_78" -> "766 view_78_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"766 view_78_0_0_nncf_smooth_quant_0" -> "767 quantize_per_tensor_default_58"  [label="(197, 768)", style=solid];
-"767 quantize_per_tensor_default_58" -> "768 dequantize_per_tensor_default_58"  [label="(197, 768)", style=solid];
-"768 dequantize_per_tensor_default_58" -> "775 linear_37"  [label="(197, 768)", style=solid];
-"769 _param_constant117_0_0" -> "775 linear_37"  [label="(768,)", style=solid];
-"770 scale_updated_constant38" -> "772 mul_tensor_38"  [label="(768, 1)", style=solid];
-"771 compressed_weight_updated_constant38" -> "772 mul_tensor_38"  [label="(768, 768)", style=solid];
-"772 mul_tensor_38" -> "774 sub_tensor_38"  [label="(768, 768)", style=solid];
-"773 zero_point_updated_constant38" -> "774 sub_tensor_38"  [label="(768, 1)", style=solid];
-"774 sub_tensor_38" -> "775 linear_37"  [label="(768, 768)", style=solid];
-"775 linear_37" -> "776 view_79"  [label="(197, 768)", style=solid];
-"776 view_79" -> "777 transpose_59"  [label="(197, 1, 768)", style=solid];
-"777 transpose_59" -> "778 dropout_28"  [label="(1, 197, 768)", style=solid];
-"778 dropout_28" -> "779 add_19"  [label="(1, 197, 768)", style=solid];
-"779 add_19" -> "782 layer_norm_19"  [label="(1, 197, 768)", style=solid];
-"779 add_19" -> "806 add_20"  [label="(1, 197, 768)", style=solid];
-"780 _param_constant118" -> "782 layer_norm_19"  [label="(768,)", style=solid];
-"781 _param_constant119" -> "782 layer_norm_19"  [label="(768,)", style=solid];
-"782 layer_norm_19" -> "783 layer_norm_19_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"783 layer_norm_19_0_0_nncf_smooth_quant_0" -> "784 quantize_per_tensor_default_59"  [label="(1, 197, 768)", style=solid];
-"784 quantize_per_tensor_default_59" -> "785 dequantize_per_tensor_default_59"  [label="(1, 197, 768)", style=solid];
-"785 dequantize_per_tensor_default_59" -> "792 linear_38"  [label="(1, 197, 768)", style=solid];
-"786 _param_constant121_0_0" -> "792 linear_38"  [label="(3072,)", style=solid];
-"787 scale_updated_constant39" -> "789 mul_tensor_39"  [label="(3072, 1)", style=solid];
-"788 compressed_weight_updated_constant39" -> "789 mul_tensor_39"  [label="(3072, 768)", style=solid];
-"789 mul_tensor_39" -> "791 sub_tensor_39"  [label="(3072, 768)", style=solid];
-"790 zero_point_updated_constant39" -> "791 sub_tensor_39"  [label="(3072, 1)", style=solid];
-"791 sub_tensor_39" -> "792 linear_38"  [label="(3072, 768)", style=solid];
-"792 linear_38" -> "793 gelu_9"  [label="(1, 197, 3072)", style=solid];
-"793 gelu_9" -> "794 dropout_29"  [label="(1, 197, 3072)", style=solid];
-"794 dropout_29" -> "795 dropout_29_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"795 dropout_29_0_0_nncf_smooth_quant_0" -> "796 quantize_per_tensor_default_60"  [label="(1, 197, 3072)", style=solid];
-"796 quantize_per_tensor_default_60" -> "797 dequantize_per_tensor_default_60"  [label="(1, 197, 3072)", style=solid];
-"797 dequantize_per_tensor_default_60" -> "804 linear_39"  [label="(1, 197, 3072)", style=solid];
-"798 _param_constant123_0_0" -> "804 linear_39"  [label="(768,)", style=solid];
-"799 scale_updated_constant40" -> "801 mul_tensor_40"  [label="(768, 1)", style=solid];
-"800 compressed_weight_updated_constant40" -> "801 mul_tensor_40"  [label="(768, 3072)", style=solid];
-"801 mul_tensor_40" -> "803 sub_tensor_40"  [label="(768, 3072)", style=solid];
-"802 zero_point_updated_constant40" -> "803 sub_tensor_40"  [label="(768, 1)", style=solid];
-"803 sub_tensor_40" -> "804 linear_39"  [label="(768, 3072)", style=solid];
-"804 linear_39" -> "805 dropout_30"  [label="(1, 197, 768)", style=solid];
-"805 dropout_30" -> "806 add_20"  [label="(1, 197, 768)", style=solid];
-"806 add_20" -> "809 layer_norm_20"  [label="(1, 197, 768)", style=solid];
-"806 add_20" -> "858 add_21"  [label="(1, 197, 768)", style=solid];
-"807 _param_constant124" -> "809 layer_norm_20"  [label="(768,)", style=solid];
-"808 _param_constant125" -> "809 layer_norm_20"  [label="(768,)", style=solid];
-"809 layer_norm_20" -> "810 transpose_60"  [label="(1, 197, 768)", style=solid];
-"810 transpose_60" -> "811 transpose_60_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"811 transpose_60_0_0_nncf_smooth_quant_0" -> "812 quantize_per_tensor_default_61"  [label="(197, 1, 768)", style=solid];
-"812 quantize_per_tensor_default_61" -> "813 dequantize_per_tensor_default_61"  [label="(197, 1, 768)", style=solid];
-"813 dequantize_per_tensor_default_61" -> "820 linear_40"  [label="(197, 1, 768)", style=solid];
-"814 _param_constant127_0_0" -> "820 linear_40"  [label="(2304,)", style=solid];
-"815 scale_updated_constant41" -> "817 mul_tensor_41"  [label="(2304, 1)", style=solid];
-"816 compressed_weight_updated_constant41" -> "817 mul_tensor_41"  [label="(2304, 768)", style=solid];
-"817 mul_tensor_41" -> "819 sub_tensor_41"  [label="(2304, 768)", style=solid];
-"818 zero_point_updated_constant41" -> "819 sub_tensor_41"  [label="(2304, 1)", style=solid];
-"819 sub_tensor_41" -> "820 linear_40"  [label="(2304, 768)", style=solid];
-"820 linear_40" -> "821 unflatten_10"  [label="(197, 1, 2304)", style=solid];
-"821 unflatten_10" -> "822 unsqueeze_10"  [label="(197, 1, 3, 768)", style=solid];
-"822 unsqueeze_10" -> "823 transpose_61"  [label="(1, 197, 1, 3, 768)", style=solid];
-"823 transpose_61" -> "824 squeeze_10"  [label="(3, 197, 1, 1, 768)", style=solid];
-"824 squeeze_10" -> "825 contiguous_10"  [label="(3, 197, 1, 768)", style=solid];
-"825 contiguous_10" -> "826 quantize_per_tensor_default_62"  [label="(3, 197, 1, 768)", style=solid];
-"825 contiguous_10" -> "829 quantize_per_tensor_default_63"  [label="(3, 197, 1, 768)", style=solid];
-"825 contiguous_10" -> "832 select_32"  [label="(3, 197, 1, 768)", style=solid];
-"826 quantize_per_tensor_default_62" -> "827 dequantize_per_tensor_default_62"  [label="(3, 197, 1, 768)", style=solid];
-"827 dequantize_per_tensor_default_62" -> "828 select_30"  [label="(3, 197, 1, 768)", style=solid];
-"828 select_30" -> "833 view_80"  [label="(197, 1, 768)", style=solid];
-"829 quantize_per_tensor_default_63" -> "830 dequantize_per_tensor_default_63"  [label="(3, 197, 1, 768)", style=solid];
-"830 dequantize_per_tensor_default_63" -> "831 select_31"  [label="(3, 197, 1, 768)", style=solid];
-"831 select_31" -> "835 view_81"  [label="(197, 1, 768)", style=solid];
-"832 select_32" -> "837 view_82"  [label="(197, 1, 768)", style=solid];
-"833 view_80" -> "834 transpose_62"  [label="(197, 12, 64)", style=solid];
-"834 transpose_62" -> "839 view_83"  [label="(12, 197, 64)", style=solid];
-"835 view_81" -> "836 transpose_63"  [label="(197, 12, 64)", style=solid];
-"836 transpose_63" -> "840 view_84"  [label="(12, 197, 64)", style=solid];
-"837 view_82" -> "838 transpose_64"  [label="(197, 12, 64)", style=solid];
-"838 transpose_64" -> "841 view_85"  [label="(12, 197, 64)", style=solid];
-"839 view_83" -> "842 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
-"840 view_84" -> "842 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
-"841 view_85" -> "842 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
-"842 scaled_dot_product_attention_10" -> "843 permute_11"  [label="(1, 12, 197, 64)", style=solid];
-"843 permute_11" -> "844 view_86"  [label="(197, 1, 12, 64)", style=solid];
-"844 view_86" -> "845 view_86_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"845 view_86_0_0_nncf_smooth_quant_0" -> "846 quantize_per_tensor_default_64"  [label="(197, 768)", style=solid];
-"846 quantize_per_tensor_default_64" -> "847 dequantize_per_tensor_default_64"  [label="(197, 768)", style=solid];
-"847 dequantize_per_tensor_default_64" -> "854 linear_41"  [label="(197, 768)", style=solid];
-"848 _param_constant129_0_0" -> "854 linear_41"  [label="(768,)", style=solid];
-"849 scale_updated_constant42" -> "851 mul_tensor_42"  [label="(768, 1)", style=solid];
-"850 compressed_weight_updated_constant42" -> "851 mul_tensor_42"  [label="(768, 768)", style=solid];
-"851 mul_tensor_42" -> "853 sub_tensor_42"  [label="(768, 768)", style=solid];
-"852 zero_point_updated_constant42" -> "853 sub_tensor_42"  [label="(768, 1)", style=solid];
-"853 sub_tensor_42" -> "854 linear_41"  [label="(768, 768)", style=solid];
-"854 linear_41" -> "855 view_87"  [label="(197, 768)", style=solid];
-"855 view_87" -> "856 transpose_65"  [label="(197, 1, 768)", style=solid];
-"856 transpose_65" -> "857 dropout_31"  [label="(1, 197, 768)", style=solid];
-"857 dropout_31" -> "858 add_21"  [label="(1, 197, 768)", style=solid];
-"858 add_21" -> "861 layer_norm_21"  [label="(1, 197, 768)", style=solid];
-"858 add_21" -> "885 add_22"  [label="(1, 197, 768)", style=solid];
-"859 _param_constant130" -> "861 layer_norm_21"  [label="(768,)", style=solid];
-"860 _param_constant131" -> "861 layer_norm_21"  [label="(768,)", style=solid];
-"861 layer_norm_21" -> "862 layer_norm_21_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"862 layer_norm_21_0_0_nncf_smooth_quant_0" -> "863 quantize_per_tensor_default_65"  [label="(1, 197, 768)", style=solid];
-"863 quantize_per_tensor_default_65" -> "864 dequantize_per_tensor_default_65"  [label="(1, 197, 768)", style=solid];
-"864 dequantize_per_tensor_default_65" -> "871 linear_42"  [label="(1, 197, 768)", style=solid];
-"865 _param_constant133_0_0" -> "871 linear_42"  [label="(3072,)", style=solid];
-"866 scale_updated_constant43" -> "868 mul_tensor_43"  [label="(3072, 1)", style=solid];
-"867 compressed_weight_updated_constant43" -> "868 mul_tensor_43"  [label="(3072, 768)", style=solid];
-"868 mul_tensor_43" -> "870 sub_tensor_43"  [label="(3072, 768)", style=solid];
-"869 zero_point_updated_constant43" -> "870 sub_tensor_43"  [label="(3072, 1)", style=solid];
-"870 sub_tensor_43" -> "871 linear_42"  [label="(3072, 768)", style=solid];
-"871 linear_42" -> "872 gelu_10"  [label="(1, 197, 3072)", style=solid];
-"872 gelu_10" -> "873 dropout_32"  [label="(1, 197, 3072)", style=solid];
-"873 dropout_32" -> "874 dropout_32_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"874 dropout_32_0_0_nncf_smooth_quant_0" -> "875 quantize_per_tensor_default_66"  [label="(1, 197, 3072)", style=solid];
-"875 quantize_per_tensor_default_66" -> "876 dequantize_per_tensor_default_66"  [label="(1, 197, 3072)", style=solid];
-"876 dequantize_per_tensor_default_66" -> "883 linear_43"  [label="(1, 197, 3072)", style=solid];
-"877 _param_constant135_0_0" -> "883 linear_43"  [label="(768,)", style=solid];
-"878 scale_updated_constant44" -> "880 mul_tensor_44"  [label="(768, 1)", style=solid];
-"879 compressed_weight_updated_constant44" -> "880 mul_tensor_44"  [label="(768, 3072)", style=solid];
-"880 mul_tensor_44" -> "882 sub_tensor_44"  [label="(768, 3072)", style=solid];
-"881 zero_point_updated_constant44" -> "882 sub_tensor_44"  [label="(768, 1)", style=solid];
-"882 sub_tensor_44" -> "883 linear_43"  [label="(768, 3072)", style=solid];
-"883 linear_43" -> "884 dropout_33"  [label="(1, 197, 768)", style=solid];
-"884 dropout_33" -> "885 add_22"  [label="(1, 197, 768)", style=solid];
-"885 add_22" -> "888 layer_norm_22"  [label="(1, 197, 768)", style=solid];
-"885 add_22" -> "937 add_23"  [label="(1, 197, 768)", style=solid];
-"886 _param_constant136" -> "888 layer_norm_22"  [label="(768,)", style=solid];
-"887 _param_constant137" -> "888 layer_norm_22"  [label="(768,)", style=solid];
-"888 layer_norm_22" -> "889 transpose_66"  [label="(1, 197, 768)", style=solid];
-"889 transpose_66" -> "890 transpose_66_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"890 transpose_66_0_0_nncf_smooth_quant_0" -> "891 quantize_per_tensor_default_67"  [label="(197, 1, 768)", style=solid];
-"891 quantize_per_tensor_default_67" -> "892 dequantize_per_tensor_default_67"  [label="(197, 1, 768)", style=solid];
-"892 dequantize_per_tensor_default_67" -> "899 linear_44"  [label="(197, 1, 768)", style=solid];
-"893 _param_constant139_0_0" -> "899 linear_44"  [label="(2304,)", style=solid];
-"894 scale_updated_constant45" -> "896 mul_tensor_45"  [label="(2304, 1)", style=solid];
-"895 compressed_weight_updated_constant45" -> "896 mul_tensor_45"  [label="(2304, 768)", style=solid];
-"896 mul_tensor_45" -> "898 sub_tensor_45"  [label="(2304, 768)", style=solid];
-"897 zero_point_updated_constant45" -> "898 sub_tensor_45"  [label="(2304, 1)", style=solid];
-"898 sub_tensor_45" -> "899 linear_44"  [label="(2304, 768)", style=solid];
-"899 linear_44" -> "900 unflatten_11"  [label="(197, 1, 2304)", style=solid];
-"900 unflatten_11" -> "901 unsqueeze_11"  [label="(197, 1, 3, 768)", style=solid];
-"901 unsqueeze_11" -> "902 transpose_67"  [label="(1, 197, 1, 3, 768)", style=solid];
-"902 transpose_67" -> "903 squeeze_11"  [label="(3, 197, 1, 1, 768)", style=solid];
-"903 squeeze_11" -> "904 contiguous_11"  [label="(3, 197, 1, 768)", style=solid];
-"904 contiguous_11" -> "905 quantize_per_tensor_default_68"  [label="(3, 197, 1, 768)", style=solid];
-"904 contiguous_11" -> "908 quantize_per_tensor_default_69"  [label="(3, 197, 1, 768)", style=solid];
-"904 contiguous_11" -> "911 select_35"  [label="(3, 197, 1, 768)", style=solid];
-"905 quantize_per_tensor_default_68" -> "906 dequantize_per_tensor_default_68"  [label="(3, 197, 1, 768)", style=solid];
-"906 dequantize_per_tensor_default_68" -> "907 select_33"  [label="(3, 197, 1, 768)", style=solid];
-"907 select_33" -> "912 view_88"  [label="(197, 1, 768)", style=solid];
-"908 quantize_per_tensor_default_69" -> "909 dequantize_per_tensor_default_69"  [label="(3, 197, 1, 768)", style=solid];
-"909 dequantize_per_tensor_default_69" -> "910 select_34"  [label="(3, 197, 1, 768)", style=solid];
-"910 select_34" -> "914 view_89"  [label="(197, 1, 768)", style=solid];
-"911 select_35" -> "916 view_90"  [label="(197, 1, 768)", style=solid];
-"912 view_88" -> "913 transpose_68"  [label="(197, 12, 64)", style=solid];
-"913 transpose_68" -> "918 view_91"  [label="(12, 197, 64)", style=solid];
-"914 view_89" -> "915 transpose_69"  [label="(197, 12, 64)", style=solid];
-"915 transpose_69" -> "919 view_92"  [label="(12, 197, 64)", style=solid];
-"916 view_90" -> "917 transpose_70"  [label="(197, 12, 64)", style=solid];
-"917 transpose_70" -> "920 view_93"  [label="(12, 197, 64)", style=solid];
-"918 view_91" -> "921 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
-"919 view_92" -> "921 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
-"920 view_93" -> "921 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
-"921 scaled_dot_product_attention_11" -> "922 permute_12"  [label="(1, 12, 197, 64)", style=solid];
-"922 permute_12" -> "923 view_94"  [label="(197, 1, 12, 64)", style=solid];
-"923 view_94" -> "924 view_94_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"924 view_94_0_0_nncf_smooth_quant_0" -> "925 quantize_per_tensor_default_70"  [label="(197, 768)", style=solid];
-"925 quantize_per_tensor_default_70" -> "926 dequantize_per_tensor_default_70"  [label="(197, 768)", style=solid];
-"926 dequantize_per_tensor_default_70" -> "933 linear_45"  [label="(197, 768)", style=solid];
-"927 _param_constant141_0_0" -> "933 linear_45"  [label="(768,)", style=solid];
-"928 scale_updated_constant46" -> "930 mul_tensor_46"  [label="(768, 1)", style=solid];
-"929 compressed_weight_updated_constant46" -> "930 mul_tensor_46"  [label="(768, 768)", style=solid];
-"930 mul_tensor_46" -> "932 sub_tensor_46"  [label="(768, 768)", style=solid];
-"931 zero_point_updated_constant46" -> "932 sub_tensor_46"  [label="(768, 1)", style=solid];
-"932 sub_tensor_46" -> "933 linear_45"  [label="(768, 768)", style=solid];
-"933 linear_45" -> "934 view_95"  [label="(197, 768)", style=solid];
-"934 view_95" -> "935 transpose_71"  [label="(197, 1, 768)", style=solid];
-"935 transpose_71" -> "936 dropout_34"  [label="(1, 197, 768)", style=solid];
-"936 dropout_34" -> "937 add_23"  [label="(1, 197, 768)", style=solid];
-"937 add_23" -> "940 layer_norm_23"  [label="(1, 197, 768)", style=solid];
-"937 add_23" -> "964 add_24"  [label="(1, 197, 768)", style=solid];
-"938 _param_constant142" -> "940 layer_norm_23"  [label="(768,)", style=solid];
-"939 _param_constant143" -> "940 layer_norm_23"  [label="(768,)", style=solid];
-"940 layer_norm_23" -> "941 layer_norm_23_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"941 layer_norm_23_0_0_nncf_smooth_quant_0" -> "942 quantize_per_tensor_default_71"  [label="(1, 197, 768)", style=solid];
-"942 quantize_per_tensor_default_71" -> "943 dequantize_per_tensor_default_71"  [label="(1, 197, 768)", style=solid];
-"943 dequantize_per_tensor_default_71" -> "950 linear_46"  [label="(1, 197, 768)", style=solid];
-"944 _param_constant145_0_0" -> "950 linear_46"  [label="(3072,)", style=solid];
-"945 scale_updated_constant47" -> "947 mul_tensor_47"  [label="(3072, 1)", style=solid];
-"946 compressed_weight_updated_constant47" -> "947 mul_tensor_47"  [label="(3072, 768)", style=solid];
-"947 mul_tensor_47" -> "949 sub_tensor_47"  [label="(3072, 768)", style=solid];
-"948 zero_point_updated_constant47" -> "949 sub_tensor_47"  [label="(3072, 1)", style=solid];
-"949 sub_tensor_47" -> "950 linear_46"  [label="(3072, 768)", style=solid];
-"950 linear_46" -> "951 gelu_11"  [label="(1, 197, 3072)", style=solid];
-"951 gelu_11" -> "952 dropout_35"  [label="(1, 197, 3072)", style=solid];
-"952 dropout_35" -> "953 dropout_35_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"953 dropout_35_0_0_nncf_smooth_quant_0" -> "954 quantize_per_tensor_default_72"  [label="(1, 197, 3072)", style=solid];
-"954 quantize_per_tensor_default_72" -> "955 dequantize_per_tensor_default_72"  [label="(1, 197, 3072)", style=solid];
-"955 dequantize_per_tensor_default_72" -> "962 linear_47"  [label="(1, 197, 3072)", style=solid];
-"956 _param_constant147_0_0" -> "962 linear_47"  [label="(768,)", style=solid];
-"957 scale_updated_constant48" -> "959 mul_tensor_48"  [label="(768, 1)", style=solid];
-"958 compressed_weight_updated_constant48" -> "959 mul_tensor_48"  [label="(768, 3072)", style=solid];
-"959 mul_tensor_48" -> "961 sub_tensor_48"  [label="(768, 3072)", style=solid];
-"960 zero_point_updated_constant48" -> "961 sub_tensor_48"  [label="(768, 1)", style=solid];
-"961 sub_tensor_48" -> "962 linear_47"  [label="(768, 3072)", style=solid];
-"962 linear_47" -> "963 dropout_36"  [label="(1, 197, 768)", style=solid];
-"963 dropout_36" -> "964 add_24"  [label="(1, 197, 768)", style=solid];
-"964 add_24" -> "967 layer_norm_24"  [label="(1, 197, 768)", style=solid];
-"965 _param_constant148" -> "967 layer_norm_24"  [label="(768,)", style=solid];
-"966 _param_constant149" -> "967 layer_norm_24"  [label="(768,)", style=solid];
+"0 encoder_pos_embedding" -> "65 add"  [label="(1, 197, 768)", style=solid];
+"1 encoder_layers_encoder_layer_0_ln_1_weight" -> "67 layer_norm"  [label="(768,)", style=solid];
+"2 encoder_layers_encoder_layer_0_ln_1_bias" -> "67 layer_norm"  [label="(768,)", style=solid];
+"3 encoder_layers_encoder_layer_0_ln_2_weight" -> "117 layer_norm_1"  [label="(768,)", style=solid];
+"4 encoder_layers_encoder_layer_0_ln_2_bias" -> "117 layer_norm_1"  [label="(768,)", style=solid];
+"5 encoder_layers_encoder_layer_1_ln_1_weight" -> "142 layer_norm_2"  [label="(768,)", style=solid];
+"6 encoder_layers_encoder_layer_1_ln_1_bias" -> "142 layer_norm_2"  [label="(768,)", style=solid];
+"7 encoder_layers_encoder_layer_1_ln_2_weight" -> "192 layer_norm_3"  [label="(768,)", style=solid];
+"8 encoder_layers_encoder_layer_1_ln_2_bias" -> "192 layer_norm_3"  [label="(768,)", style=solid];
+"9 encoder_layers_encoder_layer_2_ln_1_weight" -> "217 layer_norm_4"  [label="(768,)", style=solid];
+"10 encoder_layers_encoder_layer_2_ln_1_bias" -> "217 layer_norm_4"  [label="(768,)", style=solid];
+"11 encoder_layers_encoder_layer_2_ln_2_weight" -> "267 layer_norm_5"  [label="(768,)", style=solid];
+"12 encoder_layers_encoder_layer_2_ln_2_bias" -> "267 layer_norm_5"  [label="(768,)", style=solid];
+"13 encoder_layers_encoder_layer_3_ln_1_weight" -> "292 layer_norm_6"  [label="(768,)", style=solid];
+"14 encoder_layers_encoder_layer_3_ln_1_bias" -> "292 layer_norm_6"  [label="(768,)", style=solid];
+"15 encoder_layers_encoder_layer_3_ln_2_weight" -> "342 layer_norm_7"  [label="(768,)", style=solid];
+"16 encoder_layers_encoder_layer_3_ln_2_bias" -> "342 layer_norm_7"  [label="(768,)", style=solid];
+"17 encoder_layers_encoder_layer_4_ln_1_weight" -> "367 layer_norm_8"  [label="(768,)", style=solid];
+"18 encoder_layers_encoder_layer_4_ln_1_bias" -> "367 layer_norm_8"  [label="(768,)", style=solid];
+"19 encoder_layers_encoder_layer_4_ln_2_weight" -> "417 layer_norm_9"  [label="(768,)", style=solid];
+"20 encoder_layers_encoder_layer_4_ln_2_bias" -> "417 layer_norm_9"  [label="(768,)", style=solid];
+"21 encoder_layers_encoder_layer_5_ln_1_weight" -> "442 layer_norm_10"  [label="(768,)", style=solid];
+"22 encoder_layers_encoder_layer_5_ln_1_bias" -> "442 layer_norm_10"  [label="(768,)", style=solid];
+"23 encoder_layers_encoder_layer_5_ln_2_weight" -> "492 layer_norm_11"  [label="(768,)", style=solid];
+"24 encoder_layers_encoder_layer_5_ln_2_bias" -> "492 layer_norm_11"  [label="(768,)", style=solid];
+"25 encoder_layers_encoder_layer_6_ln_1_weight" -> "517 layer_norm_12"  [label="(768,)", style=solid];
+"26 encoder_layers_encoder_layer_6_ln_1_bias" -> "517 layer_norm_12"  [label="(768,)", style=solid];
+"27 encoder_layers_encoder_layer_6_ln_2_weight" -> "567 layer_norm_13"  [label="(768,)", style=solid];
+"28 encoder_layers_encoder_layer_6_ln_2_bias" -> "567 layer_norm_13"  [label="(768,)", style=solid];
+"29 encoder_layers_encoder_layer_7_ln_1_weight" -> "592 layer_norm_14"  [label="(768,)", style=solid];
+"30 encoder_layers_encoder_layer_7_ln_1_bias" -> "592 layer_norm_14"  [label="(768,)", style=solid];
+"31 encoder_layers_encoder_layer_7_ln_2_weight" -> "642 layer_norm_15"  [label="(768,)", style=solid];
+"32 encoder_layers_encoder_layer_7_ln_2_bias" -> "642 layer_norm_15"  [label="(768,)", style=solid];
+"33 encoder_layers_encoder_layer_8_ln_1_weight" -> "667 layer_norm_16"  [label="(768,)", style=solid];
+"34 encoder_layers_encoder_layer_8_ln_1_bias" -> "667 layer_norm_16"  [label="(768,)", style=solid];
+"35 encoder_layers_encoder_layer_8_ln_2_weight" -> "717 layer_norm_17"  [label="(768,)", style=solid];
+"36 encoder_layers_encoder_layer_8_ln_2_bias" -> "717 layer_norm_17"  [label="(768,)", style=solid];
+"37 encoder_layers_encoder_layer_9_ln_1_weight" -> "742 layer_norm_18"  [label="(768,)", style=solid];
+"38 encoder_layers_encoder_layer_9_ln_1_bias" -> "742 layer_norm_18"  [label="(768,)", style=solid];
+"39 encoder_layers_encoder_layer_9_ln_2_weight" -> "792 layer_norm_19"  [label="(768,)", style=solid];
+"40 encoder_layers_encoder_layer_9_ln_2_bias" -> "792 layer_norm_19"  [label="(768,)", style=solid];
+"41 encoder_layers_encoder_layer_10_ln_1_weight" -> "817 layer_norm_20"  [label="(768,)", style=solid];
+"42 encoder_layers_encoder_layer_10_ln_1_bias" -> "817 layer_norm_20"  [label="(768,)", style=solid];
+"43 encoder_layers_encoder_layer_10_ln_2_weight" -> "867 layer_norm_21"  [label="(768,)", style=solid];
+"44 encoder_layers_encoder_layer_10_ln_2_bias" -> "867 layer_norm_21"  [label="(768,)", style=solid];
+"45 encoder_layers_encoder_layer_11_ln_1_weight" -> "892 layer_norm_22"  [label="(768,)", style=solid];
+"46 encoder_layers_encoder_layer_11_ln_1_bias" -> "892 layer_norm_22"  [label="(768,)", style=solid];
+"47 encoder_layers_encoder_layer_11_ln_2_weight" -> "942 layer_norm_23"  [label="(768,)", style=solid];
+"48 encoder_layers_encoder_layer_11_ln_2_bias" -> "942 layer_norm_23"  [label="(768,)", style=solid];
+"49 encoder_ln_weight" -> "967 layer_norm_24"  [label="(768,)", style=solid];
+"50 encoder_ln_bias" -> "967 layer_norm_24"  [label="(768,)", style=solid];
+"51 x" -> "52 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"52 quantize_per_tensor_default" -> "53 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"53 dequantize_per_tensor_default" -> "60 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"54 conv_proj_bias_0_0" -> "60 conv2d"  [label="(768,)", style=solid];
+"55 scale_updated_constant0" -> "57 mul_tensor"  [label="(768, 1, 1, 1)", style=solid];
+"56 compressed_weight_updated_constant0" -> "57 mul_tensor"  [label="(768, 3, 16, 16)", style=solid];
+"57 mul_tensor" -> "59 sub_tensor"  [label="(768, 3, 16, 16)", style=solid];
+"58 zero_point_updated_constant0" -> "59 sub_tensor"  [label="(768, 1, 1, 1)", style=solid];
+"59 sub_tensor" -> "60 conv2d"  [label="(768, 3, 16, 16)", style=solid];
+"60 conv2d" -> "61 reshape"  [label="(1, 768, 14, 14)", style=solid];
+"61 reshape" -> "62 permute"  [label="(1, 768, 196)", style=solid];
+"62 permute" -> "64 cat"  [label="(1, 196, 768)", style=solid];
+"63 _frozen_param0" -> "64 cat"  [label="(1, 1, 768)", style=solid];
+"64 cat" -> "65 add"  [label="(1, 197, 768)", style=solid];
+"65 add" -> "66 dropout"  [label="(1, 197, 768)", style=solid];
+"66 dropout" -> "67 layer_norm"  [label="(1, 197, 768)", style=solid];
+"66 dropout" -> "116 add_1"  [label="(1, 197, 768)", style=solid];
+"67 layer_norm" -> "68 transpose"  [label="(1, 197, 768)", style=solid];
+"68 transpose" -> "69 transpose_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"69 transpose_0_0_nncf_smooth_quant_0" -> "70 quantize_per_tensor_default_1"  [label="(197, 1, 768)", style=solid];
+"70 quantize_per_tensor_default_1" -> "71 dequantize_per_tensor_default_1"  [label="(197, 1, 768)", style=solid];
+"71 dequantize_per_tensor_default_1" -> "78 linear"  [label="(197, 1, 768)", style=solid];
+"72 encoder_layers_encoder_layer_0_self_attention_in_proj_bias_0_0" -> "78 linear"  [label="(2304,)", style=solid];
+"73 scale_updated_constant1" -> "75 mul_tensor_1"  [label="(2304, 1)", style=solid];
+"74 compressed_weight_updated_constant1" -> "75 mul_tensor_1"  [label="(2304, 768)", style=solid];
+"75 mul_tensor_1" -> "77 sub_tensor_1"  [label="(2304, 768)", style=solid];
+"76 zero_point_updated_constant1" -> "77 sub_tensor_1"  [label="(2304, 1)", style=solid];
+"77 sub_tensor_1" -> "78 linear"  [label="(2304, 768)", style=solid];
+"78 linear" -> "79 unflatten"  [label="(197, 1, 2304)", style=solid];
+"79 unflatten" -> "80 unsqueeze"  [label="(197, 1, 3, 768)", style=solid];
+"80 unsqueeze" -> "81 transpose_1"  [label="(1, 197, 1, 3, 768)", style=solid];
+"81 transpose_1" -> "82 squeeze"  [label="(3, 197, 1, 1, 768)", style=solid];
+"82 squeeze" -> "83 contiguous"  [label="(3, 197, 1, 768)", style=solid];
+"83 contiguous" -> "84 quantize_per_tensor_default_2"  [label="(3, 197, 1, 768)", style=solid];
+"83 contiguous" -> "87 quantize_per_tensor_default_3"  [label="(3, 197, 1, 768)", style=solid];
+"83 contiguous" -> "90 select_2"  [label="(3, 197, 1, 768)", style=solid];
+"84 quantize_per_tensor_default_2" -> "85 dequantize_per_tensor_default_2"  [label="(3, 197, 1, 768)", style=solid];
+"85 dequantize_per_tensor_default_2" -> "86 select"  [label="(3, 197, 1, 768)", style=solid];
+"86 select" -> "91 view"  [label="(197, 1, 768)", style=solid];
+"87 quantize_per_tensor_default_3" -> "88 dequantize_per_tensor_default_3"  [label="(3, 197, 1, 768)", style=solid];
+"88 dequantize_per_tensor_default_3" -> "89 select_1"  [label="(3, 197, 1, 768)", style=solid];
+"89 select_1" -> "93 view_1"  [label="(197, 1, 768)", style=solid];
+"90 select_2" -> "95 view_2"  [label="(197, 1, 768)", style=solid];
+"91 view" -> "92 transpose_2"  [label="(197, 12, 64)", style=solid];
+"92 transpose_2" -> "97 view_3"  [label="(12, 197, 64)", style=solid];
+"93 view_1" -> "94 transpose_3"  [label="(197, 12, 64)", style=solid];
+"94 transpose_3" -> "98 view_4"  [label="(12, 197, 64)", style=solid];
+"95 view_2" -> "96 transpose_4"  [label="(197, 12, 64)", style=solid];
+"96 transpose_4" -> "99 view_5"  [label="(12, 197, 64)", style=solid];
+"97 view_3" -> "100 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
+"98 view_4" -> "100 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
+"99 view_5" -> "100 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
+"100 scaled_dot_product_attention" -> "101 permute_1"  [label="(1, 12, 197, 64)", style=solid];
+"101 permute_1" -> "102 view_6"  [label="(197, 1, 12, 64)", style=solid];
+"102 view_6" -> "103 view_6_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"103 view_6_0_0_nncf_smooth_quant_0" -> "104 quantize_per_tensor_default_4"  [label="(197, 768)", style=solid];
+"104 quantize_per_tensor_default_4" -> "105 dequantize_per_tensor_default_4"  [label="(197, 768)", style=solid];
+"105 dequantize_per_tensor_default_4" -> "112 linear_1"  [label="(197, 768)", style=solid];
+"106 encoder_layers_encoder_layer_0_self_attention_out_proj_bias_0_0" -> "112 linear_1"  [label="(768,)", style=solid];
+"107 scale_updated_constant2" -> "109 mul_tensor_2"  [label="(768, 1)", style=solid];
+"108 compressed_weight_updated_constant2" -> "109 mul_tensor_2"  [label="(768, 768)", style=solid];
+"109 mul_tensor_2" -> "111 sub_tensor_2"  [label="(768, 768)", style=solid];
+"110 zero_point_updated_constant2" -> "111 sub_tensor_2"  [label="(768, 1)", style=solid];
+"111 sub_tensor_2" -> "112 linear_1"  [label="(768, 768)", style=solid];
+"112 linear_1" -> "113 view_7"  [label="(197, 768)", style=solid];
+"113 view_7" -> "114 transpose_5"  [label="(197, 1, 768)", style=solid];
+"114 transpose_5" -> "115 dropout_1"  [label="(1, 197, 768)", style=solid];
+"115 dropout_1" -> "116 add_1"  [label="(1, 197, 768)", style=solid];
+"116 add_1" -> "117 layer_norm_1"  [label="(1, 197, 768)", style=solid];
+"116 add_1" -> "141 add_2"  [label="(1, 197, 768)", style=solid];
+"117 layer_norm_1" -> "118 layer_norm_1_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"118 layer_norm_1_0_0_nncf_smooth_quant_0" -> "119 quantize_per_tensor_default_5"  [label="(1, 197, 768)", style=solid];
+"119 quantize_per_tensor_default_5" -> "120 dequantize_per_tensor_default_5"  [label="(1, 197, 768)", style=solid];
+"120 dequantize_per_tensor_default_5" -> "127 linear_2"  [label="(1, 197, 768)", style=solid];
+"121 encoder_layers_encoder_layer_0_mlp_0_bias_0_0" -> "127 linear_2"  [label="(3072,)", style=solid];
+"122 scale_updated_constant3" -> "124 mul_tensor_3"  [label="(3072, 1)", style=solid];
+"123 compressed_weight_updated_constant3" -> "124 mul_tensor_3"  [label="(3072, 768)", style=solid];
+"124 mul_tensor_3" -> "126 sub_tensor_3"  [label="(3072, 768)", style=solid];
+"125 zero_point_updated_constant3" -> "126 sub_tensor_3"  [label="(3072, 1)", style=solid];
+"126 sub_tensor_3" -> "127 linear_2"  [label="(3072, 768)", style=solid];
+"127 linear_2" -> "128 gelu"  [label="(1, 197, 3072)", style=solid];
+"128 gelu" -> "129 dropout_2"  [label="(1, 197, 3072)", style=solid];
+"129 dropout_2" -> "130 dropout_2_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"130 dropout_2_0_0_nncf_smooth_quant_0" -> "131 quantize_per_tensor_default_6"  [label="(1, 197, 3072)", style=solid];
+"131 quantize_per_tensor_default_6" -> "132 dequantize_per_tensor_default_6"  [label="(1, 197, 3072)", style=solid];
+"132 dequantize_per_tensor_default_6" -> "139 linear_3"  [label="(1, 197, 3072)", style=solid];
+"133 encoder_layers_encoder_layer_0_mlp_3_bias_0_0" -> "139 linear_3"  [label="(768,)", style=solid];
+"134 scale_updated_constant4" -> "136 mul_tensor_4"  [label="(768, 1)", style=solid];
+"135 compressed_weight_updated_constant4" -> "136 mul_tensor_4"  [label="(768, 3072)", style=solid];
+"136 mul_tensor_4" -> "138 sub_tensor_4"  [label="(768, 3072)", style=solid];
+"137 zero_point_updated_constant4" -> "138 sub_tensor_4"  [label="(768, 1)", style=solid];
+"138 sub_tensor_4" -> "139 linear_3"  [label="(768, 3072)", style=solid];
+"139 linear_3" -> "140 dropout_3"  [label="(1, 197, 768)", style=solid];
+"140 dropout_3" -> "141 add_2"  [label="(1, 197, 768)", style=solid];
+"141 add_2" -> "142 layer_norm_2"  [label="(1, 197, 768)", style=solid];
+"141 add_2" -> "191 add_3"  [label="(1, 197, 768)", style=solid];
+"142 layer_norm_2" -> "143 transpose_6"  [label="(1, 197, 768)", style=solid];
+"143 transpose_6" -> "144 transpose_6_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"144 transpose_6_0_0_nncf_smooth_quant_0" -> "145 quantize_per_tensor_default_7"  [label="(197, 1, 768)", style=solid];
+"145 quantize_per_tensor_default_7" -> "146 dequantize_per_tensor_default_7"  [label="(197, 1, 768)", style=solid];
+"146 dequantize_per_tensor_default_7" -> "153 linear_4"  [label="(197, 1, 768)", style=solid];
+"147 encoder_layers_encoder_layer_1_self_attention_in_proj_bias_0_0" -> "153 linear_4"  [label="(2304,)", style=solid];
+"148 scale_updated_constant5" -> "150 mul_tensor_5"  [label="(2304, 1)", style=solid];
+"149 compressed_weight_updated_constant5" -> "150 mul_tensor_5"  [label="(2304, 768)", style=solid];
+"150 mul_tensor_5" -> "152 sub_tensor_5"  [label="(2304, 768)", style=solid];
+"151 zero_point_updated_constant5" -> "152 sub_tensor_5"  [label="(2304, 1)", style=solid];
+"152 sub_tensor_5" -> "153 linear_4"  [label="(2304, 768)", style=solid];
+"153 linear_4" -> "154 unflatten_1"  [label="(197, 1, 2304)", style=solid];
+"154 unflatten_1" -> "155 unsqueeze_1"  [label="(197, 1, 3, 768)", style=solid];
+"155 unsqueeze_1" -> "156 transpose_7"  [label="(1, 197, 1, 3, 768)", style=solid];
+"156 transpose_7" -> "157 squeeze_1"  [label="(3, 197, 1, 1, 768)", style=solid];
+"157 squeeze_1" -> "158 contiguous_1"  [label="(3, 197, 1, 768)", style=solid];
+"158 contiguous_1" -> "159 quantize_per_tensor_default_8"  [label="(3, 197, 1, 768)", style=solid];
+"158 contiguous_1" -> "162 quantize_per_tensor_default_9"  [label="(3, 197, 1, 768)", style=solid];
+"158 contiguous_1" -> "165 select_5"  [label="(3, 197, 1, 768)", style=solid];
+"159 quantize_per_tensor_default_8" -> "160 dequantize_per_tensor_default_8"  [label="(3, 197, 1, 768)", style=solid];
+"160 dequantize_per_tensor_default_8" -> "161 select_3"  [label="(3, 197, 1, 768)", style=solid];
+"161 select_3" -> "166 view_8"  [label="(197, 1, 768)", style=solid];
+"162 quantize_per_tensor_default_9" -> "163 dequantize_per_tensor_default_9"  [label="(3, 197, 1, 768)", style=solid];
+"163 dequantize_per_tensor_default_9" -> "164 select_4"  [label="(3, 197, 1, 768)", style=solid];
+"164 select_4" -> "168 view_9"  [label="(197, 1, 768)", style=solid];
+"165 select_5" -> "170 view_10"  [label="(197, 1, 768)", style=solid];
+"166 view_8" -> "167 transpose_8"  [label="(197, 12, 64)", style=solid];
+"167 transpose_8" -> "172 view_11"  [label="(12, 197, 64)", style=solid];
+"168 view_9" -> "169 transpose_9"  [label="(197, 12, 64)", style=solid];
+"169 transpose_9" -> "173 view_12"  [label="(12, 197, 64)", style=solid];
+"170 view_10" -> "171 transpose_10"  [label="(197, 12, 64)", style=solid];
+"171 transpose_10" -> "174 view_13"  [label="(12, 197, 64)", style=solid];
+"172 view_11" -> "175 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
+"173 view_12" -> "175 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
+"174 view_13" -> "175 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
+"175 scaled_dot_product_attention_1" -> "176 permute_2"  [label="(1, 12, 197, 64)", style=solid];
+"176 permute_2" -> "177 view_14"  [label="(197, 1, 12, 64)", style=solid];
+"177 view_14" -> "178 view_14_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"178 view_14_0_0_nncf_smooth_quant_0" -> "179 quantize_per_tensor_default_10"  [label="(197, 768)", style=solid];
+"179 quantize_per_tensor_default_10" -> "180 dequantize_per_tensor_default_10"  [label="(197, 768)", style=solid];
+"180 dequantize_per_tensor_default_10" -> "187 linear_5"  [label="(197, 768)", style=solid];
+"181 encoder_layers_encoder_layer_1_self_attention_out_proj_bias_0_0" -> "187 linear_5"  [label="(768,)", style=solid];
+"182 scale_updated_constant6" -> "184 mul_tensor_6"  [label="(768, 1)", style=solid];
+"183 compressed_weight_updated_constant6" -> "184 mul_tensor_6"  [label="(768, 768)", style=solid];
+"184 mul_tensor_6" -> "186 sub_tensor_6"  [label="(768, 768)", style=solid];
+"185 zero_point_updated_constant6" -> "186 sub_tensor_6"  [label="(768, 1)", style=solid];
+"186 sub_tensor_6" -> "187 linear_5"  [label="(768, 768)", style=solid];
+"187 linear_5" -> "188 view_15"  [label="(197, 768)", style=solid];
+"188 view_15" -> "189 transpose_11"  [label="(197, 1, 768)", style=solid];
+"189 transpose_11" -> "190 dropout_4"  [label="(1, 197, 768)", style=solid];
+"190 dropout_4" -> "191 add_3"  [label="(1, 197, 768)", style=solid];
+"191 add_3" -> "192 layer_norm_3"  [label="(1, 197, 768)", style=solid];
+"191 add_3" -> "216 add_4"  [label="(1, 197, 768)", style=solid];
+"192 layer_norm_3" -> "193 layer_norm_3_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"193 layer_norm_3_0_0_nncf_smooth_quant_0" -> "194 quantize_per_tensor_default_11"  [label="(1, 197, 768)", style=solid];
+"194 quantize_per_tensor_default_11" -> "195 dequantize_per_tensor_default_11"  [label="(1, 197, 768)", style=solid];
+"195 dequantize_per_tensor_default_11" -> "202 linear_6"  [label="(1, 197, 768)", style=solid];
+"196 encoder_layers_encoder_layer_1_mlp_0_bias_0_0" -> "202 linear_6"  [label="(3072,)", style=solid];
+"197 scale_updated_constant7" -> "199 mul_tensor_7"  [label="(3072, 1)", style=solid];
+"198 compressed_weight_updated_constant7" -> "199 mul_tensor_7"  [label="(3072, 768)", style=solid];
+"199 mul_tensor_7" -> "201 sub_tensor_7"  [label="(3072, 768)", style=solid];
+"200 zero_point_updated_constant7" -> "201 sub_tensor_7"  [label="(3072, 1)", style=solid];
+"201 sub_tensor_7" -> "202 linear_6"  [label="(3072, 768)", style=solid];
+"202 linear_6" -> "203 gelu_1"  [label="(1, 197, 3072)", style=solid];
+"203 gelu_1" -> "204 dropout_5"  [label="(1, 197, 3072)", style=solid];
+"204 dropout_5" -> "205 dropout_5_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"205 dropout_5_0_0_nncf_smooth_quant_0" -> "206 quantize_per_tensor_default_12"  [label="(1, 197, 3072)", style=solid];
+"206 quantize_per_tensor_default_12" -> "207 dequantize_per_tensor_default_12"  [label="(1, 197, 3072)", style=solid];
+"207 dequantize_per_tensor_default_12" -> "214 linear_7"  [label="(1, 197, 3072)", style=solid];
+"208 encoder_layers_encoder_layer_1_mlp_3_bias_0_0" -> "214 linear_7"  [label="(768,)", style=solid];
+"209 scale_updated_constant8" -> "211 mul_tensor_8"  [label="(768, 1)", style=solid];
+"210 compressed_weight_updated_constant8" -> "211 mul_tensor_8"  [label="(768, 3072)", style=solid];
+"211 mul_tensor_8" -> "213 sub_tensor_8"  [label="(768, 3072)", style=solid];
+"212 zero_point_updated_constant8" -> "213 sub_tensor_8"  [label="(768, 1)", style=solid];
+"213 sub_tensor_8" -> "214 linear_7"  [label="(768, 3072)", style=solid];
+"214 linear_7" -> "215 dropout_6"  [label="(1, 197, 768)", style=solid];
+"215 dropout_6" -> "216 add_4"  [label="(1, 197, 768)", style=solid];
+"216 add_4" -> "217 layer_norm_4"  [label="(1, 197, 768)", style=solid];
+"216 add_4" -> "266 add_5"  [label="(1, 197, 768)", style=solid];
+"217 layer_norm_4" -> "218 transpose_12"  [label="(1, 197, 768)", style=solid];
+"218 transpose_12" -> "219 transpose_12_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"219 transpose_12_0_0_nncf_smooth_quant_0" -> "220 quantize_per_tensor_default_13"  [label="(197, 1, 768)", style=solid];
+"220 quantize_per_tensor_default_13" -> "221 dequantize_per_tensor_default_13"  [label="(197, 1, 768)", style=solid];
+"221 dequantize_per_tensor_default_13" -> "228 linear_8"  [label="(197, 1, 768)", style=solid];
+"222 encoder_layers_encoder_layer_2_self_attention_in_proj_bias_0_0" -> "228 linear_8"  [label="(2304,)", style=solid];
+"223 scale_updated_constant9" -> "225 mul_tensor_9"  [label="(2304, 1)", style=solid];
+"224 compressed_weight_updated_constant9" -> "225 mul_tensor_9"  [label="(2304, 768)", style=solid];
+"225 mul_tensor_9" -> "227 sub_tensor_9"  [label="(2304, 768)", style=solid];
+"226 zero_point_updated_constant9" -> "227 sub_tensor_9"  [label="(2304, 1)", style=solid];
+"227 sub_tensor_9" -> "228 linear_8"  [label="(2304, 768)", style=solid];
+"228 linear_8" -> "229 unflatten_2"  [label="(197, 1, 2304)", style=solid];
+"229 unflatten_2" -> "230 unsqueeze_2"  [label="(197, 1, 3, 768)", style=solid];
+"230 unsqueeze_2" -> "231 transpose_13"  [label="(1, 197, 1, 3, 768)", style=solid];
+"231 transpose_13" -> "232 squeeze_2"  [label="(3, 197, 1, 1, 768)", style=solid];
+"232 squeeze_2" -> "233 contiguous_2"  [label="(3, 197, 1, 768)", style=solid];
+"233 contiguous_2" -> "234 quantize_per_tensor_default_14"  [label="(3, 197, 1, 768)", style=solid];
+"233 contiguous_2" -> "237 quantize_per_tensor_default_15"  [label="(3, 197, 1, 768)", style=solid];
+"233 contiguous_2" -> "240 select_8"  [label="(3, 197, 1, 768)", style=solid];
+"234 quantize_per_tensor_default_14" -> "235 dequantize_per_tensor_default_14"  [label="(3, 197, 1, 768)", style=solid];
+"235 dequantize_per_tensor_default_14" -> "236 select_6"  [label="(3, 197, 1, 768)", style=solid];
+"236 select_6" -> "241 view_16"  [label="(197, 1, 768)", style=solid];
+"237 quantize_per_tensor_default_15" -> "238 dequantize_per_tensor_default_15"  [label="(3, 197, 1, 768)", style=solid];
+"238 dequantize_per_tensor_default_15" -> "239 select_7"  [label="(3, 197, 1, 768)", style=solid];
+"239 select_7" -> "243 view_17"  [label="(197, 1, 768)", style=solid];
+"240 select_8" -> "245 view_18"  [label="(197, 1, 768)", style=solid];
+"241 view_16" -> "242 transpose_14"  [label="(197, 12, 64)", style=solid];
+"242 transpose_14" -> "247 view_19"  [label="(12, 197, 64)", style=solid];
+"243 view_17" -> "244 transpose_15"  [label="(197, 12, 64)", style=solid];
+"244 transpose_15" -> "248 view_20"  [label="(12, 197, 64)", style=solid];
+"245 view_18" -> "246 transpose_16"  [label="(197, 12, 64)", style=solid];
+"246 transpose_16" -> "249 view_21"  [label="(12, 197, 64)", style=solid];
+"247 view_19" -> "250 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
+"248 view_20" -> "250 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
+"249 view_21" -> "250 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
+"250 scaled_dot_product_attention_2" -> "251 permute_3"  [label="(1, 12, 197, 64)", style=solid];
+"251 permute_3" -> "252 view_22"  [label="(197, 1, 12, 64)", style=solid];
+"252 view_22" -> "253 view_22_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"253 view_22_0_0_nncf_smooth_quant_0" -> "254 quantize_per_tensor_default_16"  [label="(197, 768)", style=solid];
+"254 quantize_per_tensor_default_16" -> "255 dequantize_per_tensor_default_16"  [label="(197, 768)", style=solid];
+"255 dequantize_per_tensor_default_16" -> "262 linear_9"  [label="(197, 768)", style=solid];
+"256 encoder_layers_encoder_layer_2_self_attention_out_proj_bias_0_0" -> "262 linear_9"  [label="(768,)", style=solid];
+"257 scale_updated_constant10" -> "259 mul_tensor_10"  [label="(768, 1)", style=solid];
+"258 compressed_weight_updated_constant10" -> "259 mul_tensor_10"  [label="(768, 768)", style=solid];
+"259 mul_tensor_10" -> "261 sub_tensor_10"  [label="(768, 768)", style=solid];
+"260 zero_point_updated_constant10" -> "261 sub_tensor_10"  [label="(768, 1)", style=solid];
+"261 sub_tensor_10" -> "262 linear_9"  [label="(768, 768)", style=solid];
+"262 linear_9" -> "263 view_23"  [label="(197, 768)", style=solid];
+"263 view_23" -> "264 transpose_17"  [label="(197, 1, 768)", style=solid];
+"264 transpose_17" -> "265 dropout_7"  [label="(1, 197, 768)", style=solid];
+"265 dropout_7" -> "266 add_5"  [label="(1, 197, 768)", style=solid];
+"266 add_5" -> "267 layer_norm_5"  [label="(1, 197, 768)", style=solid];
+"266 add_5" -> "291 add_6"  [label="(1, 197, 768)", style=solid];
+"267 layer_norm_5" -> "268 layer_norm_5_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"268 layer_norm_5_0_0_nncf_smooth_quant_0" -> "269 quantize_per_tensor_default_17"  [label="(1, 197, 768)", style=solid];
+"269 quantize_per_tensor_default_17" -> "270 dequantize_per_tensor_default_17"  [label="(1, 197, 768)", style=solid];
+"270 dequantize_per_tensor_default_17" -> "277 linear_10"  [label="(1, 197, 768)", style=solid];
+"271 encoder_layers_encoder_layer_2_mlp_0_bias_0_0" -> "277 linear_10"  [label="(3072,)", style=solid];
+"272 scale_updated_constant11" -> "274 mul_tensor_11"  [label="(3072, 1)", style=solid];
+"273 compressed_weight_updated_constant11" -> "274 mul_tensor_11"  [label="(3072, 768)", style=solid];
+"274 mul_tensor_11" -> "276 sub_tensor_11"  [label="(3072, 768)", style=solid];
+"275 zero_point_updated_constant11" -> "276 sub_tensor_11"  [label="(3072, 1)", style=solid];
+"276 sub_tensor_11" -> "277 linear_10"  [label="(3072, 768)", style=solid];
+"277 linear_10" -> "278 gelu_2"  [label="(1, 197, 3072)", style=solid];
+"278 gelu_2" -> "279 dropout_8"  [label="(1, 197, 3072)", style=solid];
+"279 dropout_8" -> "280 dropout_8_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"280 dropout_8_0_0_nncf_smooth_quant_0" -> "281 quantize_per_tensor_default_18"  [label="(1, 197, 3072)", style=solid];
+"281 quantize_per_tensor_default_18" -> "282 dequantize_per_tensor_default_18"  [label="(1, 197, 3072)", style=solid];
+"282 dequantize_per_tensor_default_18" -> "289 linear_11"  [label="(1, 197, 3072)", style=solid];
+"283 encoder_layers_encoder_layer_2_mlp_3_bias_0_0" -> "289 linear_11"  [label="(768,)", style=solid];
+"284 scale_updated_constant12" -> "286 mul_tensor_12"  [label="(768, 1)", style=solid];
+"285 compressed_weight_updated_constant12" -> "286 mul_tensor_12"  [label="(768, 3072)", style=solid];
+"286 mul_tensor_12" -> "288 sub_tensor_12"  [label="(768, 3072)", style=solid];
+"287 zero_point_updated_constant12" -> "288 sub_tensor_12"  [label="(768, 1)", style=solid];
+"288 sub_tensor_12" -> "289 linear_11"  [label="(768, 3072)", style=solid];
+"289 linear_11" -> "290 dropout_9"  [label="(1, 197, 768)", style=solid];
+"290 dropout_9" -> "291 add_6"  [label="(1, 197, 768)", style=solid];
+"291 add_6" -> "292 layer_norm_6"  [label="(1, 197, 768)", style=solid];
+"291 add_6" -> "341 add_7"  [label="(1, 197, 768)", style=solid];
+"292 layer_norm_6" -> "293 transpose_18"  [label="(1, 197, 768)", style=solid];
+"293 transpose_18" -> "294 transpose_18_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"294 transpose_18_0_0_nncf_smooth_quant_0" -> "295 quantize_per_tensor_default_19"  [label="(197, 1, 768)", style=solid];
+"295 quantize_per_tensor_default_19" -> "296 dequantize_per_tensor_default_19"  [label="(197, 1, 768)", style=solid];
+"296 dequantize_per_tensor_default_19" -> "303 linear_12"  [label="(197, 1, 768)", style=solid];
+"297 encoder_layers_encoder_layer_3_self_attention_in_proj_bias_0_0" -> "303 linear_12"  [label="(2304,)", style=solid];
+"298 scale_updated_constant13" -> "300 mul_tensor_13"  [label="(2304, 1)", style=solid];
+"299 compressed_weight_updated_constant13" -> "300 mul_tensor_13"  [label="(2304, 768)", style=solid];
+"300 mul_tensor_13" -> "302 sub_tensor_13"  [label="(2304, 768)", style=solid];
+"301 zero_point_updated_constant13" -> "302 sub_tensor_13"  [label="(2304, 1)", style=solid];
+"302 sub_tensor_13" -> "303 linear_12"  [label="(2304, 768)", style=solid];
+"303 linear_12" -> "304 unflatten_3"  [label="(197, 1, 2304)", style=solid];
+"304 unflatten_3" -> "305 unsqueeze_3"  [label="(197, 1, 3, 768)", style=solid];
+"305 unsqueeze_3" -> "306 transpose_19"  [label="(1, 197, 1, 3, 768)", style=solid];
+"306 transpose_19" -> "307 squeeze_3"  [label="(3, 197, 1, 1, 768)", style=solid];
+"307 squeeze_3" -> "308 contiguous_3"  [label="(3, 197, 1, 768)", style=solid];
+"308 contiguous_3" -> "309 quantize_per_tensor_default_20"  [label="(3, 197, 1, 768)", style=solid];
+"308 contiguous_3" -> "312 quantize_per_tensor_default_21"  [label="(3, 197, 1, 768)", style=solid];
+"308 contiguous_3" -> "315 select_11"  [label="(3, 197, 1, 768)", style=solid];
+"309 quantize_per_tensor_default_20" -> "310 dequantize_per_tensor_default_20"  [label="(3, 197, 1, 768)", style=solid];
+"310 dequantize_per_tensor_default_20" -> "311 select_9"  [label="(3, 197, 1, 768)", style=solid];
+"311 select_9" -> "316 view_24"  [label="(197, 1, 768)", style=solid];
+"312 quantize_per_tensor_default_21" -> "313 dequantize_per_tensor_default_21"  [label="(3, 197, 1, 768)", style=solid];
+"313 dequantize_per_tensor_default_21" -> "314 select_10"  [label="(3, 197, 1, 768)", style=solid];
+"314 select_10" -> "318 view_25"  [label="(197, 1, 768)", style=solid];
+"315 select_11" -> "320 view_26"  [label="(197, 1, 768)", style=solid];
+"316 view_24" -> "317 transpose_20"  [label="(197, 12, 64)", style=solid];
+"317 transpose_20" -> "322 view_27"  [label="(12, 197, 64)", style=solid];
+"318 view_25" -> "319 transpose_21"  [label="(197, 12, 64)", style=solid];
+"319 transpose_21" -> "323 view_28"  [label="(12, 197, 64)", style=solid];
+"320 view_26" -> "321 transpose_22"  [label="(197, 12, 64)", style=solid];
+"321 transpose_22" -> "324 view_29"  [label="(12, 197, 64)", style=solid];
+"322 view_27" -> "325 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
+"323 view_28" -> "325 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
+"324 view_29" -> "325 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
+"325 scaled_dot_product_attention_3" -> "326 permute_4"  [label="(1, 12, 197, 64)", style=solid];
+"326 permute_4" -> "327 view_30"  [label="(197, 1, 12, 64)", style=solid];
+"327 view_30" -> "328 view_30_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"328 view_30_0_0_nncf_smooth_quant_0" -> "329 quantize_per_tensor_default_22"  [label="(197, 768)", style=solid];
+"329 quantize_per_tensor_default_22" -> "330 dequantize_per_tensor_default_22"  [label="(197, 768)", style=solid];
+"330 dequantize_per_tensor_default_22" -> "337 linear_13"  [label="(197, 768)", style=solid];
+"331 encoder_layers_encoder_layer_3_self_attention_out_proj_bias_0_0" -> "337 linear_13"  [label="(768,)", style=solid];
+"332 scale_updated_constant14" -> "334 mul_tensor_14"  [label="(768, 1)", style=solid];
+"333 compressed_weight_updated_constant14" -> "334 mul_tensor_14"  [label="(768, 768)", style=solid];
+"334 mul_tensor_14" -> "336 sub_tensor_14"  [label="(768, 768)", style=solid];
+"335 zero_point_updated_constant14" -> "336 sub_tensor_14"  [label="(768, 1)", style=solid];
+"336 sub_tensor_14" -> "337 linear_13"  [label="(768, 768)", style=solid];
+"337 linear_13" -> "338 view_31"  [label="(197, 768)", style=solid];
+"338 view_31" -> "339 transpose_23"  [label="(197, 1, 768)", style=solid];
+"339 transpose_23" -> "340 dropout_10"  [label="(1, 197, 768)", style=solid];
+"340 dropout_10" -> "341 add_7"  [label="(1, 197, 768)", style=solid];
+"341 add_7" -> "342 layer_norm_7"  [label="(1, 197, 768)", style=solid];
+"341 add_7" -> "366 add_8"  [label="(1, 197, 768)", style=solid];
+"342 layer_norm_7" -> "343 layer_norm_7_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"343 layer_norm_7_0_0_nncf_smooth_quant_0" -> "344 quantize_per_tensor_default_23"  [label="(1, 197, 768)", style=solid];
+"344 quantize_per_tensor_default_23" -> "345 dequantize_per_tensor_default_23"  [label="(1, 197, 768)", style=solid];
+"345 dequantize_per_tensor_default_23" -> "352 linear_14"  [label="(1, 197, 768)", style=solid];
+"346 encoder_layers_encoder_layer_3_mlp_0_bias_0_0" -> "352 linear_14"  [label="(3072,)", style=solid];
+"347 scale_updated_constant15" -> "349 mul_tensor_15"  [label="(3072, 1)", style=solid];
+"348 compressed_weight_updated_constant15" -> "349 mul_tensor_15"  [label="(3072, 768)", style=solid];
+"349 mul_tensor_15" -> "351 sub_tensor_15"  [label="(3072, 768)", style=solid];
+"350 zero_point_updated_constant15" -> "351 sub_tensor_15"  [label="(3072, 1)", style=solid];
+"351 sub_tensor_15" -> "352 linear_14"  [label="(3072, 768)", style=solid];
+"352 linear_14" -> "353 gelu_3"  [label="(1, 197, 3072)", style=solid];
+"353 gelu_3" -> "354 dropout_11"  [label="(1, 197, 3072)", style=solid];
+"354 dropout_11" -> "355 dropout_11_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"355 dropout_11_0_0_nncf_smooth_quant_0" -> "356 quantize_per_tensor_default_24"  [label="(1, 197, 3072)", style=solid];
+"356 quantize_per_tensor_default_24" -> "357 dequantize_per_tensor_default_24"  [label="(1, 197, 3072)", style=solid];
+"357 dequantize_per_tensor_default_24" -> "364 linear_15"  [label="(1, 197, 3072)", style=solid];
+"358 encoder_layers_encoder_layer_3_mlp_3_bias_0_0" -> "364 linear_15"  [label="(768,)", style=solid];
+"359 scale_updated_constant16" -> "361 mul_tensor_16"  [label="(768, 1)", style=solid];
+"360 compressed_weight_updated_constant16" -> "361 mul_tensor_16"  [label="(768, 3072)", style=solid];
+"361 mul_tensor_16" -> "363 sub_tensor_16"  [label="(768, 3072)", style=solid];
+"362 zero_point_updated_constant16" -> "363 sub_tensor_16"  [label="(768, 1)", style=solid];
+"363 sub_tensor_16" -> "364 linear_15"  [label="(768, 3072)", style=solid];
+"364 linear_15" -> "365 dropout_12"  [label="(1, 197, 768)", style=solid];
+"365 dropout_12" -> "366 add_8"  [label="(1, 197, 768)", style=solid];
+"366 add_8" -> "367 layer_norm_8"  [label="(1, 197, 768)", style=solid];
+"366 add_8" -> "416 add_9"  [label="(1, 197, 768)", style=solid];
+"367 layer_norm_8" -> "368 transpose_24"  [label="(1, 197, 768)", style=solid];
+"368 transpose_24" -> "369 transpose_24_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"369 transpose_24_0_0_nncf_smooth_quant_0" -> "370 quantize_per_tensor_default_25"  [label="(197, 1, 768)", style=solid];
+"370 quantize_per_tensor_default_25" -> "371 dequantize_per_tensor_default_25"  [label="(197, 1, 768)", style=solid];
+"371 dequantize_per_tensor_default_25" -> "378 linear_16"  [label="(197, 1, 768)", style=solid];
+"372 encoder_layers_encoder_layer_4_self_attention_in_proj_bias_0_0" -> "378 linear_16"  [label="(2304,)", style=solid];
+"373 scale_updated_constant17" -> "375 mul_tensor_17"  [label="(2304, 1)", style=solid];
+"374 compressed_weight_updated_constant17" -> "375 mul_tensor_17"  [label="(2304, 768)", style=solid];
+"375 mul_tensor_17" -> "377 sub_tensor_17"  [label="(2304, 768)", style=solid];
+"376 zero_point_updated_constant17" -> "377 sub_tensor_17"  [label="(2304, 1)", style=solid];
+"377 sub_tensor_17" -> "378 linear_16"  [label="(2304, 768)", style=solid];
+"378 linear_16" -> "379 unflatten_4"  [label="(197, 1, 2304)", style=solid];
+"379 unflatten_4" -> "380 unsqueeze_4"  [label="(197, 1, 3, 768)", style=solid];
+"380 unsqueeze_4" -> "381 transpose_25"  [label="(1, 197, 1, 3, 768)", style=solid];
+"381 transpose_25" -> "382 squeeze_4"  [label="(3, 197, 1, 1, 768)", style=solid];
+"382 squeeze_4" -> "383 contiguous_4"  [label="(3, 197, 1, 768)", style=solid];
+"383 contiguous_4" -> "384 quantize_per_tensor_default_26"  [label="(3, 197, 1, 768)", style=solid];
+"383 contiguous_4" -> "387 quantize_per_tensor_default_27"  [label="(3, 197, 1, 768)", style=solid];
+"383 contiguous_4" -> "390 select_14"  [label="(3, 197, 1, 768)", style=solid];
+"384 quantize_per_tensor_default_26" -> "385 dequantize_per_tensor_default_26"  [label="(3, 197, 1, 768)", style=solid];
+"385 dequantize_per_tensor_default_26" -> "386 select_12"  [label="(3, 197, 1, 768)", style=solid];
+"386 select_12" -> "391 view_32"  [label="(197, 1, 768)", style=solid];
+"387 quantize_per_tensor_default_27" -> "388 dequantize_per_tensor_default_27"  [label="(3, 197, 1, 768)", style=solid];
+"388 dequantize_per_tensor_default_27" -> "389 select_13"  [label="(3, 197, 1, 768)", style=solid];
+"389 select_13" -> "393 view_33"  [label="(197, 1, 768)", style=solid];
+"390 select_14" -> "395 view_34"  [label="(197, 1, 768)", style=solid];
+"391 view_32" -> "392 transpose_26"  [label="(197, 12, 64)", style=solid];
+"392 transpose_26" -> "397 view_35"  [label="(12, 197, 64)", style=solid];
+"393 view_33" -> "394 transpose_27"  [label="(197, 12, 64)", style=solid];
+"394 transpose_27" -> "398 view_36"  [label="(12, 197, 64)", style=solid];
+"395 view_34" -> "396 transpose_28"  [label="(197, 12, 64)", style=solid];
+"396 transpose_28" -> "399 view_37"  [label="(12, 197, 64)", style=solid];
+"397 view_35" -> "400 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
+"398 view_36" -> "400 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
+"399 view_37" -> "400 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
+"400 scaled_dot_product_attention_4" -> "401 permute_5"  [label="(1, 12, 197, 64)", style=solid];
+"401 permute_5" -> "402 view_38"  [label="(197, 1, 12, 64)", style=solid];
+"402 view_38" -> "403 view_38_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"403 view_38_0_0_nncf_smooth_quant_0" -> "404 quantize_per_tensor_default_28"  [label="(197, 768)", style=solid];
+"404 quantize_per_tensor_default_28" -> "405 dequantize_per_tensor_default_28"  [label="(197, 768)", style=solid];
+"405 dequantize_per_tensor_default_28" -> "412 linear_17"  [label="(197, 768)", style=solid];
+"406 encoder_layers_encoder_layer_4_self_attention_out_proj_bias_0_0" -> "412 linear_17"  [label="(768,)", style=solid];
+"407 scale_updated_constant18" -> "409 mul_tensor_18"  [label="(768, 1)", style=solid];
+"408 compressed_weight_updated_constant18" -> "409 mul_tensor_18"  [label="(768, 768)", style=solid];
+"409 mul_tensor_18" -> "411 sub_tensor_18"  [label="(768, 768)", style=solid];
+"410 zero_point_updated_constant18" -> "411 sub_tensor_18"  [label="(768, 1)", style=solid];
+"411 sub_tensor_18" -> "412 linear_17"  [label="(768, 768)", style=solid];
+"412 linear_17" -> "413 view_39"  [label="(197, 768)", style=solid];
+"413 view_39" -> "414 transpose_29"  [label="(197, 1, 768)", style=solid];
+"414 transpose_29" -> "415 dropout_13"  [label="(1, 197, 768)", style=solid];
+"415 dropout_13" -> "416 add_9"  [label="(1, 197, 768)", style=solid];
+"416 add_9" -> "417 layer_norm_9"  [label="(1, 197, 768)", style=solid];
+"416 add_9" -> "441 add_10"  [label="(1, 197, 768)", style=solid];
+"417 layer_norm_9" -> "418 layer_norm_9_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"418 layer_norm_9_0_0_nncf_smooth_quant_0" -> "419 quantize_per_tensor_default_29"  [label="(1, 197, 768)", style=solid];
+"419 quantize_per_tensor_default_29" -> "420 dequantize_per_tensor_default_29"  [label="(1, 197, 768)", style=solid];
+"420 dequantize_per_tensor_default_29" -> "427 linear_18"  [label="(1, 197, 768)", style=solid];
+"421 encoder_layers_encoder_layer_4_mlp_0_bias_0_0" -> "427 linear_18"  [label="(3072,)", style=solid];
+"422 scale_updated_constant19" -> "424 mul_tensor_19"  [label="(3072, 1)", style=solid];
+"423 compressed_weight_updated_constant19" -> "424 mul_tensor_19"  [label="(3072, 768)", style=solid];
+"424 mul_tensor_19" -> "426 sub_tensor_19"  [label="(3072, 768)", style=solid];
+"425 zero_point_updated_constant19" -> "426 sub_tensor_19"  [label="(3072, 1)", style=solid];
+"426 sub_tensor_19" -> "427 linear_18"  [label="(3072, 768)", style=solid];
+"427 linear_18" -> "428 gelu_4"  [label="(1, 197, 3072)", style=solid];
+"428 gelu_4" -> "429 dropout_14"  [label="(1, 197, 3072)", style=solid];
+"429 dropout_14" -> "430 dropout_14_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"430 dropout_14_0_0_nncf_smooth_quant_0" -> "431 quantize_per_tensor_default_30"  [label="(1, 197, 3072)", style=solid];
+"431 quantize_per_tensor_default_30" -> "432 dequantize_per_tensor_default_30"  [label="(1, 197, 3072)", style=solid];
+"432 dequantize_per_tensor_default_30" -> "439 linear_19"  [label="(1, 197, 3072)", style=solid];
+"433 encoder_layers_encoder_layer_4_mlp_3_bias_0_0" -> "439 linear_19"  [label="(768,)", style=solid];
+"434 scale_updated_constant20" -> "436 mul_tensor_20"  [label="(768, 1)", style=solid];
+"435 compressed_weight_updated_constant20" -> "436 mul_tensor_20"  [label="(768, 3072)", style=solid];
+"436 mul_tensor_20" -> "438 sub_tensor_20"  [label="(768, 3072)", style=solid];
+"437 zero_point_updated_constant20" -> "438 sub_tensor_20"  [label="(768, 1)", style=solid];
+"438 sub_tensor_20" -> "439 linear_19"  [label="(768, 3072)", style=solid];
+"439 linear_19" -> "440 dropout_15"  [label="(1, 197, 768)", style=solid];
+"440 dropout_15" -> "441 add_10"  [label="(1, 197, 768)", style=solid];
+"441 add_10" -> "442 layer_norm_10"  [label="(1, 197, 768)", style=solid];
+"441 add_10" -> "491 add_11"  [label="(1, 197, 768)", style=solid];
+"442 layer_norm_10" -> "443 transpose_30"  [label="(1, 197, 768)", style=solid];
+"443 transpose_30" -> "444 transpose_30_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"444 transpose_30_0_0_nncf_smooth_quant_0" -> "445 quantize_per_tensor_default_31"  [label="(197, 1, 768)", style=solid];
+"445 quantize_per_tensor_default_31" -> "446 dequantize_per_tensor_default_31"  [label="(197, 1, 768)", style=solid];
+"446 dequantize_per_tensor_default_31" -> "453 linear_20"  [label="(197, 1, 768)", style=solid];
+"447 encoder_layers_encoder_layer_5_self_attention_in_proj_bias_0_0" -> "453 linear_20"  [label="(2304,)", style=solid];
+"448 scale_updated_constant21" -> "450 mul_tensor_21"  [label="(2304, 1)", style=solid];
+"449 compressed_weight_updated_constant21" -> "450 mul_tensor_21"  [label="(2304, 768)", style=solid];
+"450 mul_tensor_21" -> "452 sub_tensor_21"  [label="(2304, 768)", style=solid];
+"451 zero_point_updated_constant21" -> "452 sub_tensor_21"  [label="(2304, 1)", style=solid];
+"452 sub_tensor_21" -> "453 linear_20"  [label="(2304, 768)", style=solid];
+"453 linear_20" -> "454 unflatten_5"  [label="(197, 1, 2304)", style=solid];
+"454 unflatten_5" -> "455 unsqueeze_5"  [label="(197, 1, 3, 768)", style=solid];
+"455 unsqueeze_5" -> "456 transpose_31"  [label="(1, 197, 1, 3, 768)", style=solid];
+"456 transpose_31" -> "457 squeeze_5"  [label="(3, 197, 1, 1, 768)", style=solid];
+"457 squeeze_5" -> "458 contiguous_5"  [label="(3, 197, 1, 768)", style=solid];
+"458 contiguous_5" -> "459 quantize_per_tensor_default_32"  [label="(3, 197, 1, 768)", style=solid];
+"458 contiguous_5" -> "462 quantize_per_tensor_default_33"  [label="(3, 197, 1, 768)", style=solid];
+"458 contiguous_5" -> "465 select_17"  [label="(3, 197, 1, 768)", style=solid];
+"459 quantize_per_tensor_default_32" -> "460 dequantize_per_tensor_default_32"  [label="(3, 197, 1, 768)", style=solid];
+"460 dequantize_per_tensor_default_32" -> "461 select_15"  [label="(3, 197, 1, 768)", style=solid];
+"461 select_15" -> "466 view_40"  [label="(197, 1, 768)", style=solid];
+"462 quantize_per_tensor_default_33" -> "463 dequantize_per_tensor_default_33"  [label="(3, 197, 1, 768)", style=solid];
+"463 dequantize_per_tensor_default_33" -> "464 select_16"  [label="(3, 197, 1, 768)", style=solid];
+"464 select_16" -> "468 view_41"  [label="(197, 1, 768)", style=solid];
+"465 select_17" -> "470 view_42"  [label="(197, 1, 768)", style=solid];
+"466 view_40" -> "467 transpose_32"  [label="(197, 12, 64)", style=solid];
+"467 transpose_32" -> "472 view_43"  [label="(12, 197, 64)", style=solid];
+"468 view_41" -> "469 transpose_33"  [label="(197, 12, 64)", style=solid];
+"469 transpose_33" -> "473 view_44"  [label="(12, 197, 64)", style=solid];
+"470 view_42" -> "471 transpose_34"  [label="(197, 12, 64)", style=solid];
+"471 transpose_34" -> "474 view_45"  [label="(12, 197, 64)", style=solid];
+"472 view_43" -> "475 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
+"473 view_44" -> "475 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
+"474 view_45" -> "475 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
+"475 scaled_dot_product_attention_5" -> "476 permute_6"  [label="(1, 12, 197, 64)", style=solid];
+"476 permute_6" -> "477 view_46"  [label="(197, 1, 12, 64)", style=solid];
+"477 view_46" -> "478 view_46_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"478 view_46_0_0_nncf_smooth_quant_0" -> "479 quantize_per_tensor_default_34"  [label="(197, 768)", style=solid];
+"479 quantize_per_tensor_default_34" -> "480 dequantize_per_tensor_default_34"  [label="(197, 768)", style=solid];
+"480 dequantize_per_tensor_default_34" -> "487 linear_21"  [label="(197, 768)", style=solid];
+"481 encoder_layers_encoder_layer_5_self_attention_out_proj_bias_0_0" -> "487 linear_21"  [label="(768,)", style=solid];
+"482 scale_updated_constant22" -> "484 mul_tensor_22"  [label="(768, 1)", style=solid];
+"483 compressed_weight_updated_constant22" -> "484 mul_tensor_22"  [label="(768, 768)", style=solid];
+"484 mul_tensor_22" -> "486 sub_tensor_22"  [label="(768, 768)", style=solid];
+"485 zero_point_updated_constant22" -> "486 sub_tensor_22"  [label="(768, 1)", style=solid];
+"486 sub_tensor_22" -> "487 linear_21"  [label="(768, 768)", style=solid];
+"487 linear_21" -> "488 view_47"  [label="(197, 768)", style=solid];
+"488 view_47" -> "489 transpose_35"  [label="(197, 1, 768)", style=solid];
+"489 transpose_35" -> "490 dropout_16"  [label="(1, 197, 768)", style=solid];
+"490 dropout_16" -> "491 add_11"  [label="(1, 197, 768)", style=solid];
+"491 add_11" -> "492 layer_norm_11"  [label="(1, 197, 768)", style=solid];
+"491 add_11" -> "516 add_12"  [label="(1, 197, 768)", style=solid];
+"492 layer_norm_11" -> "493 layer_norm_11_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"493 layer_norm_11_0_0_nncf_smooth_quant_0" -> "494 quantize_per_tensor_default_35"  [label="(1, 197, 768)", style=solid];
+"494 quantize_per_tensor_default_35" -> "495 dequantize_per_tensor_default_35"  [label="(1, 197, 768)", style=solid];
+"495 dequantize_per_tensor_default_35" -> "502 linear_22"  [label="(1, 197, 768)", style=solid];
+"496 encoder_layers_encoder_layer_5_mlp_0_bias_0_0" -> "502 linear_22"  [label="(3072,)", style=solid];
+"497 scale_updated_constant23" -> "499 mul_tensor_23"  [label="(3072, 1)", style=solid];
+"498 compressed_weight_updated_constant23" -> "499 mul_tensor_23"  [label="(3072, 768)", style=solid];
+"499 mul_tensor_23" -> "501 sub_tensor_23"  [label="(3072, 768)", style=solid];
+"500 zero_point_updated_constant23" -> "501 sub_tensor_23"  [label="(3072, 1)", style=solid];
+"501 sub_tensor_23" -> "502 linear_22"  [label="(3072, 768)", style=solid];
+"502 linear_22" -> "503 gelu_5"  [label="(1, 197, 3072)", style=solid];
+"503 gelu_5" -> "504 dropout_17"  [label="(1, 197, 3072)", style=solid];
+"504 dropout_17" -> "505 dropout_17_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"505 dropout_17_0_0_nncf_smooth_quant_0" -> "506 quantize_per_tensor_default_36"  [label="(1, 197, 3072)", style=solid];
+"506 quantize_per_tensor_default_36" -> "507 dequantize_per_tensor_default_36"  [label="(1, 197, 3072)", style=solid];
+"507 dequantize_per_tensor_default_36" -> "514 linear_23"  [label="(1, 197, 3072)", style=solid];
+"508 encoder_layers_encoder_layer_5_mlp_3_bias_0_0" -> "514 linear_23"  [label="(768,)", style=solid];
+"509 scale_updated_constant24" -> "511 mul_tensor_24"  [label="(768, 1)", style=solid];
+"510 compressed_weight_updated_constant24" -> "511 mul_tensor_24"  [label="(768, 3072)", style=solid];
+"511 mul_tensor_24" -> "513 sub_tensor_24"  [label="(768, 3072)", style=solid];
+"512 zero_point_updated_constant24" -> "513 sub_tensor_24"  [label="(768, 1)", style=solid];
+"513 sub_tensor_24" -> "514 linear_23"  [label="(768, 3072)", style=solid];
+"514 linear_23" -> "515 dropout_18"  [label="(1, 197, 768)", style=solid];
+"515 dropout_18" -> "516 add_12"  [label="(1, 197, 768)", style=solid];
+"516 add_12" -> "517 layer_norm_12"  [label="(1, 197, 768)", style=solid];
+"516 add_12" -> "566 add_13"  [label="(1, 197, 768)", style=solid];
+"517 layer_norm_12" -> "518 transpose_36"  [label="(1, 197, 768)", style=solid];
+"518 transpose_36" -> "519 transpose_36_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"519 transpose_36_0_0_nncf_smooth_quant_0" -> "520 quantize_per_tensor_default_37"  [label="(197, 1, 768)", style=solid];
+"520 quantize_per_tensor_default_37" -> "521 dequantize_per_tensor_default_37"  [label="(197, 1, 768)", style=solid];
+"521 dequantize_per_tensor_default_37" -> "528 linear_24"  [label="(197, 1, 768)", style=solid];
+"522 encoder_layers_encoder_layer_6_self_attention_in_proj_bias_0_0" -> "528 linear_24"  [label="(2304,)", style=solid];
+"523 scale_updated_constant25" -> "525 mul_tensor_25"  [label="(2304, 1)", style=solid];
+"524 compressed_weight_updated_constant25" -> "525 mul_tensor_25"  [label="(2304, 768)", style=solid];
+"525 mul_tensor_25" -> "527 sub_tensor_25"  [label="(2304, 768)", style=solid];
+"526 zero_point_updated_constant25" -> "527 sub_tensor_25"  [label="(2304, 1)", style=solid];
+"527 sub_tensor_25" -> "528 linear_24"  [label="(2304, 768)", style=solid];
+"528 linear_24" -> "529 unflatten_6"  [label="(197, 1, 2304)", style=solid];
+"529 unflatten_6" -> "530 unsqueeze_6"  [label="(197, 1, 3, 768)", style=solid];
+"530 unsqueeze_6" -> "531 transpose_37"  [label="(1, 197, 1, 3, 768)", style=solid];
+"531 transpose_37" -> "532 squeeze_6"  [label="(3, 197, 1, 1, 768)", style=solid];
+"532 squeeze_6" -> "533 contiguous_6"  [label="(3, 197, 1, 768)", style=solid];
+"533 contiguous_6" -> "534 quantize_per_tensor_default_38"  [label="(3, 197, 1, 768)", style=solid];
+"533 contiguous_6" -> "537 quantize_per_tensor_default_39"  [label="(3, 197, 1, 768)", style=solid];
+"533 contiguous_6" -> "540 select_20"  [label="(3, 197, 1, 768)", style=solid];
+"534 quantize_per_tensor_default_38" -> "535 dequantize_per_tensor_default_38"  [label="(3, 197, 1, 768)", style=solid];
+"535 dequantize_per_tensor_default_38" -> "536 select_18"  [label="(3, 197, 1, 768)", style=solid];
+"536 select_18" -> "541 view_48"  [label="(197, 1, 768)", style=solid];
+"537 quantize_per_tensor_default_39" -> "538 dequantize_per_tensor_default_39"  [label="(3, 197, 1, 768)", style=solid];
+"538 dequantize_per_tensor_default_39" -> "539 select_19"  [label="(3, 197, 1, 768)", style=solid];
+"539 select_19" -> "543 view_49"  [label="(197, 1, 768)", style=solid];
+"540 select_20" -> "545 view_50"  [label="(197, 1, 768)", style=solid];
+"541 view_48" -> "542 transpose_38"  [label="(197, 12, 64)", style=solid];
+"542 transpose_38" -> "547 view_51"  [label="(12, 197, 64)", style=solid];
+"543 view_49" -> "544 transpose_39"  [label="(197, 12, 64)", style=solid];
+"544 transpose_39" -> "548 view_52"  [label="(12, 197, 64)", style=solid];
+"545 view_50" -> "546 transpose_40"  [label="(197, 12, 64)", style=solid];
+"546 transpose_40" -> "549 view_53"  [label="(12, 197, 64)", style=solid];
+"547 view_51" -> "550 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
+"548 view_52" -> "550 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
+"549 view_53" -> "550 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
+"550 scaled_dot_product_attention_6" -> "551 permute_7"  [label="(1, 12, 197, 64)", style=solid];
+"551 permute_7" -> "552 view_54"  [label="(197, 1, 12, 64)", style=solid];
+"552 view_54" -> "553 view_54_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"553 view_54_0_0_nncf_smooth_quant_0" -> "554 quantize_per_tensor_default_40"  [label="(197, 768)", style=solid];
+"554 quantize_per_tensor_default_40" -> "555 dequantize_per_tensor_default_40"  [label="(197, 768)", style=solid];
+"555 dequantize_per_tensor_default_40" -> "562 linear_25"  [label="(197, 768)", style=solid];
+"556 encoder_layers_encoder_layer_6_self_attention_out_proj_bias_0_0" -> "562 linear_25"  [label="(768,)", style=solid];
+"557 scale_updated_constant26" -> "559 mul_tensor_26"  [label="(768, 1)", style=solid];
+"558 compressed_weight_updated_constant26" -> "559 mul_tensor_26"  [label="(768, 768)", style=solid];
+"559 mul_tensor_26" -> "561 sub_tensor_26"  [label="(768, 768)", style=solid];
+"560 zero_point_updated_constant26" -> "561 sub_tensor_26"  [label="(768, 1)", style=solid];
+"561 sub_tensor_26" -> "562 linear_25"  [label="(768, 768)", style=solid];
+"562 linear_25" -> "563 view_55"  [label="(197, 768)", style=solid];
+"563 view_55" -> "564 transpose_41"  [label="(197, 1, 768)", style=solid];
+"564 transpose_41" -> "565 dropout_19"  [label="(1, 197, 768)", style=solid];
+"565 dropout_19" -> "566 add_13"  [label="(1, 197, 768)", style=solid];
+"566 add_13" -> "567 layer_norm_13"  [label="(1, 197, 768)", style=solid];
+"566 add_13" -> "591 add_14"  [label="(1, 197, 768)", style=solid];
+"567 layer_norm_13" -> "568 layer_norm_13_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"568 layer_norm_13_0_0_nncf_smooth_quant_0" -> "569 quantize_per_tensor_default_41"  [label="(1, 197, 768)", style=solid];
+"569 quantize_per_tensor_default_41" -> "570 dequantize_per_tensor_default_41"  [label="(1, 197, 768)", style=solid];
+"570 dequantize_per_tensor_default_41" -> "577 linear_26"  [label="(1, 197, 768)", style=solid];
+"571 encoder_layers_encoder_layer_6_mlp_0_bias_0_0" -> "577 linear_26"  [label="(3072,)", style=solid];
+"572 scale_updated_constant27" -> "574 mul_tensor_27"  [label="(3072, 1)", style=solid];
+"573 compressed_weight_updated_constant27" -> "574 mul_tensor_27"  [label="(3072, 768)", style=solid];
+"574 mul_tensor_27" -> "576 sub_tensor_27"  [label="(3072, 768)", style=solid];
+"575 zero_point_updated_constant27" -> "576 sub_tensor_27"  [label="(3072, 1)", style=solid];
+"576 sub_tensor_27" -> "577 linear_26"  [label="(3072, 768)", style=solid];
+"577 linear_26" -> "578 gelu_6"  [label="(1, 197, 3072)", style=solid];
+"578 gelu_6" -> "579 dropout_20"  [label="(1, 197, 3072)", style=solid];
+"579 dropout_20" -> "580 dropout_20_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"580 dropout_20_0_0_nncf_smooth_quant_0" -> "581 quantize_per_tensor_default_42"  [label="(1, 197, 3072)", style=solid];
+"581 quantize_per_tensor_default_42" -> "582 dequantize_per_tensor_default_42"  [label="(1, 197, 3072)", style=solid];
+"582 dequantize_per_tensor_default_42" -> "589 linear_27"  [label="(1, 197, 3072)", style=solid];
+"583 encoder_layers_encoder_layer_6_mlp_3_bias_0_0" -> "589 linear_27"  [label="(768,)", style=solid];
+"584 scale_updated_constant28" -> "586 mul_tensor_28"  [label="(768, 1)", style=solid];
+"585 compressed_weight_updated_constant28" -> "586 mul_tensor_28"  [label="(768, 3072)", style=solid];
+"586 mul_tensor_28" -> "588 sub_tensor_28"  [label="(768, 3072)", style=solid];
+"587 zero_point_updated_constant28" -> "588 sub_tensor_28"  [label="(768, 1)", style=solid];
+"588 sub_tensor_28" -> "589 linear_27"  [label="(768, 3072)", style=solid];
+"589 linear_27" -> "590 dropout_21"  [label="(1, 197, 768)", style=solid];
+"590 dropout_21" -> "591 add_14"  [label="(1, 197, 768)", style=solid];
+"591 add_14" -> "592 layer_norm_14"  [label="(1, 197, 768)", style=solid];
+"591 add_14" -> "641 add_15"  [label="(1, 197, 768)", style=solid];
+"592 layer_norm_14" -> "593 transpose_42"  [label="(1, 197, 768)", style=solid];
+"593 transpose_42" -> "594 transpose_42_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"594 transpose_42_0_0_nncf_smooth_quant_0" -> "595 quantize_per_tensor_default_43"  [label="(197, 1, 768)", style=solid];
+"595 quantize_per_tensor_default_43" -> "596 dequantize_per_tensor_default_43"  [label="(197, 1, 768)", style=solid];
+"596 dequantize_per_tensor_default_43" -> "603 linear_28"  [label="(197, 1, 768)", style=solid];
+"597 encoder_layers_encoder_layer_7_self_attention_in_proj_bias_0_0" -> "603 linear_28"  [label="(2304,)", style=solid];
+"598 scale_updated_constant29" -> "600 mul_tensor_29"  [label="(2304, 1)", style=solid];
+"599 compressed_weight_updated_constant29" -> "600 mul_tensor_29"  [label="(2304, 768)", style=solid];
+"600 mul_tensor_29" -> "602 sub_tensor_29"  [label="(2304, 768)", style=solid];
+"601 zero_point_updated_constant29" -> "602 sub_tensor_29"  [label="(2304, 1)", style=solid];
+"602 sub_tensor_29" -> "603 linear_28"  [label="(2304, 768)", style=solid];
+"603 linear_28" -> "604 unflatten_7"  [label="(197, 1, 2304)", style=solid];
+"604 unflatten_7" -> "605 unsqueeze_7"  [label="(197, 1, 3, 768)", style=solid];
+"605 unsqueeze_7" -> "606 transpose_43"  [label="(1, 197, 1, 3, 768)", style=solid];
+"606 transpose_43" -> "607 squeeze_7"  [label="(3, 197, 1, 1, 768)", style=solid];
+"607 squeeze_7" -> "608 contiguous_7"  [label="(3, 197, 1, 768)", style=solid];
+"608 contiguous_7" -> "609 quantize_per_tensor_default_44"  [label="(3, 197, 1, 768)", style=solid];
+"608 contiguous_7" -> "612 quantize_per_tensor_default_45"  [label="(3, 197, 1, 768)", style=solid];
+"608 contiguous_7" -> "615 select_23"  [label="(3, 197, 1, 768)", style=solid];
+"609 quantize_per_tensor_default_44" -> "610 dequantize_per_tensor_default_44"  [label="(3, 197, 1, 768)", style=solid];
+"610 dequantize_per_tensor_default_44" -> "611 select_21"  [label="(3, 197, 1, 768)", style=solid];
+"611 select_21" -> "616 view_56"  [label="(197, 1, 768)", style=solid];
+"612 quantize_per_tensor_default_45" -> "613 dequantize_per_tensor_default_45"  [label="(3, 197, 1, 768)", style=solid];
+"613 dequantize_per_tensor_default_45" -> "614 select_22"  [label="(3, 197, 1, 768)", style=solid];
+"614 select_22" -> "618 view_57"  [label="(197, 1, 768)", style=solid];
+"615 select_23" -> "620 view_58"  [label="(197, 1, 768)", style=solid];
+"616 view_56" -> "617 transpose_44"  [label="(197, 12, 64)", style=solid];
+"617 transpose_44" -> "622 view_59"  [label="(12, 197, 64)", style=solid];
+"618 view_57" -> "619 transpose_45"  [label="(197, 12, 64)", style=solid];
+"619 transpose_45" -> "623 view_60"  [label="(12, 197, 64)", style=solid];
+"620 view_58" -> "621 transpose_46"  [label="(197, 12, 64)", style=solid];
+"621 transpose_46" -> "624 view_61"  [label="(12, 197, 64)", style=solid];
+"622 view_59" -> "625 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
+"623 view_60" -> "625 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
+"624 view_61" -> "625 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
+"625 scaled_dot_product_attention_7" -> "626 permute_8"  [label="(1, 12, 197, 64)", style=solid];
+"626 permute_8" -> "627 view_62"  [label="(197, 1, 12, 64)", style=solid];
+"627 view_62" -> "628 view_62_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"628 view_62_0_0_nncf_smooth_quant_0" -> "629 quantize_per_tensor_default_46"  [label="(197, 768)", style=solid];
+"629 quantize_per_tensor_default_46" -> "630 dequantize_per_tensor_default_46"  [label="(197, 768)", style=solid];
+"630 dequantize_per_tensor_default_46" -> "637 linear_29"  [label="(197, 768)", style=solid];
+"631 encoder_layers_encoder_layer_7_self_attention_out_proj_bias_0_0" -> "637 linear_29"  [label="(768,)", style=solid];
+"632 scale_updated_constant30" -> "634 mul_tensor_30"  [label="(768, 1)", style=solid];
+"633 compressed_weight_updated_constant30" -> "634 mul_tensor_30"  [label="(768, 768)", style=solid];
+"634 mul_tensor_30" -> "636 sub_tensor_30"  [label="(768, 768)", style=solid];
+"635 zero_point_updated_constant30" -> "636 sub_tensor_30"  [label="(768, 1)", style=solid];
+"636 sub_tensor_30" -> "637 linear_29"  [label="(768, 768)", style=solid];
+"637 linear_29" -> "638 view_63"  [label="(197, 768)", style=solid];
+"638 view_63" -> "639 transpose_47"  [label="(197, 1, 768)", style=solid];
+"639 transpose_47" -> "640 dropout_22"  [label="(1, 197, 768)", style=solid];
+"640 dropout_22" -> "641 add_15"  [label="(1, 197, 768)", style=solid];
+"641 add_15" -> "642 layer_norm_15"  [label="(1, 197, 768)", style=solid];
+"641 add_15" -> "666 add_16"  [label="(1, 197, 768)", style=solid];
+"642 layer_norm_15" -> "643 layer_norm_15_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"643 layer_norm_15_0_0_nncf_smooth_quant_0" -> "644 quantize_per_tensor_default_47"  [label="(1, 197, 768)", style=solid];
+"644 quantize_per_tensor_default_47" -> "645 dequantize_per_tensor_default_47"  [label="(1, 197, 768)", style=solid];
+"645 dequantize_per_tensor_default_47" -> "652 linear_30"  [label="(1, 197, 768)", style=solid];
+"646 encoder_layers_encoder_layer_7_mlp_0_bias_0_0" -> "652 linear_30"  [label="(3072,)", style=solid];
+"647 scale_updated_constant31" -> "649 mul_tensor_31"  [label="(3072, 1)", style=solid];
+"648 compressed_weight_updated_constant31" -> "649 mul_tensor_31"  [label="(3072, 768)", style=solid];
+"649 mul_tensor_31" -> "651 sub_tensor_31"  [label="(3072, 768)", style=solid];
+"650 zero_point_updated_constant31" -> "651 sub_tensor_31"  [label="(3072, 1)", style=solid];
+"651 sub_tensor_31" -> "652 linear_30"  [label="(3072, 768)", style=solid];
+"652 linear_30" -> "653 gelu_7"  [label="(1, 197, 3072)", style=solid];
+"653 gelu_7" -> "654 dropout_23"  [label="(1, 197, 3072)", style=solid];
+"654 dropout_23" -> "655 dropout_23_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"655 dropout_23_0_0_nncf_smooth_quant_0" -> "656 quantize_per_tensor_default_48"  [label="(1, 197, 3072)", style=solid];
+"656 quantize_per_tensor_default_48" -> "657 dequantize_per_tensor_default_48"  [label="(1, 197, 3072)", style=solid];
+"657 dequantize_per_tensor_default_48" -> "664 linear_31"  [label="(1, 197, 3072)", style=solid];
+"658 encoder_layers_encoder_layer_7_mlp_3_bias_0_0" -> "664 linear_31"  [label="(768,)", style=solid];
+"659 scale_updated_constant32" -> "661 mul_tensor_32"  [label="(768, 1)", style=solid];
+"660 compressed_weight_updated_constant32" -> "661 mul_tensor_32"  [label="(768, 3072)", style=solid];
+"661 mul_tensor_32" -> "663 sub_tensor_32"  [label="(768, 3072)", style=solid];
+"662 zero_point_updated_constant32" -> "663 sub_tensor_32"  [label="(768, 1)", style=solid];
+"663 sub_tensor_32" -> "664 linear_31"  [label="(768, 3072)", style=solid];
+"664 linear_31" -> "665 dropout_24"  [label="(1, 197, 768)", style=solid];
+"665 dropout_24" -> "666 add_16"  [label="(1, 197, 768)", style=solid];
+"666 add_16" -> "667 layer_norm_16"  [label="(1, 197, 768)", style=solid];
+"666 add_16" -> "716 add_17"  [label="(1, 197, 768)", style=solid];
+"667 layer_norm_16" -> "668 transpose_48"  [label="(1, 197, 768)", style=solid];
+"668 transpose_48" -> "669 transpose_48_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"669 transpose_48_0_0_nncf_smooth_quant_0" -> "670 quantize_per_tensor_default_49"  [label="(197, 1, 768)", style=solid];
+"670 quantize_per_tensor_default_49" -> "671 dequantize_per_tensor_default_49"  [label="(197, 1, 768)", style=solid];
+"671 dequantize_per_tensor_default_49" -> "678 linear_32"  [label="(197, 1, 768)", style=solid];
+"672 encoder_layers_encoder_layer_8_self_attention_in_proj_bias_0_0" -> "678 linear_32"  [label="(2304,)", style=solid];
+"673 scale_updated_constant33" -> "675 mul_tensor_33"  [label="(2304, 1)", style=solid];
+"674 compressed_weight_updated_constant33" -> "675 mul_tensor_33"  [label="(2304, 768)", style=solid];
+"675 mul_tensor_33" -> "677 sub_tensor_33"  [label="(2304, 768)", style=solid];
+"676 zero_point_updated_constant33" -> "677 sub_tensor_33"  [label="(2304, 1)", style=solid];
+"677 sub_tensor_33" -> "678 linear_32"  [label="(2304, 768)", style=solid];
+"678 linear_32" -> "679 unflatten_8"  [label="(197, 1, 2304)", style=solid];
+"679 unflatten_8" -> "680 unsqueeze_8"  [label="(197, 1, 3, 768)", style=solid];
+"680 unsqueeze_8" -> "681 transpose_49"  [label="(1, 197, 1, 3, 768)", style=solid];
+"681 transpose_49" -> "682 squeeze_8"  [label="(3, 197, 1, 1, 768)", style=solid];
+"682 squeeze_8" -> "683 contiguous_8"  [label="(3, 197, 1, 768)", style=solid];
+"683 contiguous_8" -> "684 quantize_per_tensor_default_50"  [label="(3, 197, 1, 768)", style=solid];
+"683 contiguous_8" -> "687 quantize_per_tensor_default_51"  [label="(3, 197, 1, 768)", style=solid];
+"683 contiguous_8" -> "690 select_26"  [label="(3, 197, 1, 768)", style=solid];
+"684 quantize_per_tensor_default_50" -> "685 dequantize_per_tensor_default_50"  [label="(3, 197, 1, 768)", style=solid];
+"685 dequantize_per_tensor_default_50" -> "686 select_24"  [label="(3, 197, 1, 768)", style=solid];
+"686 select_24" -> "691 view_64"  [label="(197, 1, 768)", style=solid];
+"687 quantize_per_tensor_default_51" -> "688 dequantize_per_tensor_default_51"  [label="(3, 197, 1, 768)", style=solid];
+"688 dequantize_per_tensor_default_51" -> "689 select_25"  [label="(3, 197, 1, 768)", style=solid];
+"689 select_25" -> "693 view_65"  [label="(197, 1, 768)", style=solid];
+"690 select_26" -> "695 view_66"  [label="(197, 1, 768)", style=solid];
+"691 view_64" -> "692 transpose_50"  [label="(197, 12, 64)", style=solid];
+"692 transpose_50" -> "697 view_67"  [label="(12, 197, 64)", style=solid];
+"693 view_65" -> "694 transpose_51"  [label="(197, 12, 64)", style=solid];
+"694 transpose_51" -> "698 view_68"  [label="(12, 197, 64)", style=solid];
+"695 view_66" -> "696 transpose_52"  [label="(197, 12, 64)", style=solid];
+"696 transpose_52" -> "699 view_69"  [label="(12, 197, 64)", style=solid];
+"697 view_67" -> "700 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
+"698 view_68" -> "700 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
+"699 view_69" -> "700 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
+"700 scaled_dot_product_attention_8" -> "701 permute_9"  [label="(1, 12, 197, 64)", style=solid];
+"701 permute_9" -> "702 view_70"  [label="(197, 1, 12, 64)", style=solid];
+"702 view_70" -> "703 view_70_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"703 view_70_0_0_nncf_smooth_quant_0" -> "704 quantize_per_tensor_default_52"  [label="(197, 768)", style=solid];
+"704 quantize_per_tensor_default_52" -> "705 dequantize_per_tensor_default_52"  [label="(197, 768)", style=solid];
+"705 dequantize_per_tensor_default_52" -> "712 linear_33"  [label="(197, 768)", style=solid];
+"706 encoder_layers_encoder_layer_8_self_attention_out_proj_bias_0_0" -> "712 linear_33"  [label="(768,)", style=solid];
+"707 scale_updated_constant34" -> "709 mul_tensor_34"  [label="(768, 1)", style=solid];
+"708 compressed_weight_updated_constant34" -> "709 mul_tensor_34"  [label="(768, 768)", style=solid];
+"709 mul_tensor_34" -> "711 sub_tensor_34"  [label="(768, 768)", style=solid];
+"710 zero_point_updated_constant34" -> "711 sub_tensor_34"  [label="(768, 1)", style=solid];
+"711 sub_tensor_34" -> "712 linear_33"  [label="(768, 768)", style=solid];
+"712 linear_33" -> "713 view_71"  [label="(197, 768)", style=solid];
+"713 view_71" -> "714 transpose_53"  [label="(197, 1, 768)", style=solid];
+"714 transpose_53" -> "715 dropout_25"  [label="(1, 197, 768)", style=solid];
+"715 dropout_25" -> "716 add_17"  [label="(1, 197, 768)", style=solid];
+"716 add_17" -> "717 layer_norm_17"  [label="(1, 197, 768)", style=solid];
+"716 add_17" -> "741 add_18"  [label="(1, 197, 768)", style=solid];
+"717 layer_norm_17" -> "718 layer_norm_17_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"718 layer_norm_17_0_0_nncf_smooth_quant_0" -> "719 quantize_per_tensor_default_53"  [label="(1, 197, 768)", style=solid];
+"719 quantize_per_tensor_default_53" -> "720 dequantize_per_tensor_default_53"  [label="(1, 197, 768)", style=solid];
+"720 dequantize_per_tensor_default_53" -> "727 linear_34"  [label="(1, 197, 768)", style=solid];
+"721 encoder_layers_encoder_layer_8_mlp_0_bias_0_0" -> "727 linear_34"  [label="(3072,)", style=solid];
+"722 scale_updated_constant35" -> "724 mul_tensor_35"  [label="(3072, 1)", style=solid];
+"723 compressed_weight_updated_constant35" -> "724 mul_tensor_35"  [label="(3072, 768)", style=solid];
+"724 mul_tensor_35" -> "726 sub_tensor_35"  [label="(3072, 768)", style=solid];
+"725 zero_point_updated_constant35" -> "726 sub_tensor_35"  [label="(3072, 1)", style=solid];
+"726 sub_tensor_35" -> "727 linear_34"  [label="(3072, 768)", style=solid];
+"727 linear_34" -> "728 gelu_8"  [label="(1, 197, 3072)", style=solid];
+"728 gelu_8" -> "729 dropout_26"  [label="(1, 197, 3072)", style=solid];
+"729 dropout_26" -> "730 dropout_26_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"730 dropout_26_0_0_nncf_smooth_quant_0" -> "731 quantize_per_tensor_default_54"  [label="(1, 197, 3072)", style=solid];
+"731 quantize_per_tensor_default_54" -> "732 dequantize_per_tensor_default_54"  [label="(1, 197, 3072)", style=solid];
+"732 dequantize_per_tensor_default_54" -> "739 linear_35"  [label="(1, 197, 3072)", style=solid];
+"733 encoder_layers_encoder_layer_8_mlp_3_bias_0_0" -> "739 linear_35"  [label="(768,)", style=solid];
+"734 scale_updated_constant36" -> "736 mul_tensor_36"  [label="(768, 1)", style=solid];
+"735 compressed_weight_updated_constant36" -> "736 mul_tensor_36"  [label="(768, 3072)", style=solid];
+"736 mul_tensor_36" -> "738 sub_tensor_36"  [label="(768, 3072)", style=solid];
+"737 zero_point_updated_constant36" -> "738 sub_tensor_36"  [label="(768, 1)", style=solid];
+"738 sub_tensor_36" -> "739 linear_35"  [label="(768, 3072)", style=solid];
+"739 linear_35" -> "740 dropout_27"  [label="(1, 197, 768)", style=solid];
+"740 dropout_27" -> "741 add_18"  [label="(1, 197, 768)", style=solid];
+"741 add_18" -> "742 layer_norm_18"  [label="(1, 197, 768)", style=solid];
+"741 add_18" -> "791 add_19"  [label="(1, 197, 768)", style=solid];
+"742 layer_norm_18" -> "743 transpose_54"  [label="(1, 197, 768)", style=solid];
+"743 transpose_54" -> "744 transpose_54_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"744 transpose_54_0_0_nncf_smooth_quant_0" -> "745 quantize_per_tensor_default_55"  [label="(197, 1, 768)", style=solid];
+"745 quantize_per_tensor_default_55" -> "746 dequantize_per_tensor_default_55"  [label="(197, 1, 768)", style=solid];
+"746 dequantize_per_tensor_default_55" -> "753 linear_36"  [label="(197, 1, 768)", style=solid];
+"747 encoder_layers_encoder_layer_9_self_attention_in_proj_bias_0_0" -> "753 linear_36"  [label="(2304,)", style=solid];
+"748 scale_updated_constant37" -> "750 mul_tensor_37"  [label="(2304, 1)", style=solid];
+"749 compressed_weight_updated_constant37" -> "750 mul_tensor_37"  [label="(2304, 768)", style=solid];
+"750 mul_tensor_37" -> "752 sub_tensor_37"  [label="(2304, 768)", style=solid];
+"751 zero_point_updated_constant37" -> "752 sub_tensor_37"  [label="(2304, 1)", style=solid];
+"752 sub_tensor_37" -> "753 linear_36"  [label="(2304, 768)", style=solid];
+"753 linear_36" -> "754 unflatten_9"  [label="(197, 1, 2304)", style=solid];
+"754 unflatten_9" -> "755 unsqueeze_9"  [label="(197, 1, 3, 768)", style=solid];
+"755 unsqueeze_9" -> "756 transpose_55"  [label="(1, 197, 1, 3, 768)", style=solid];
+"756 transpose_55" -> "757 squeeze_9"  [label="(3, 197, 1, 1, 768)", style=solid];
+"757 squeeze_9" -> "758 contiguous_9"  [label="(3, 197, 1, 768)", style=solid];
+"758 contiguous_9" -> "759 quantize_per_tensor_default_56"  [label="(3, 197, 1, 768)", style=solid];
+"758 contiguous_9" -> "762 quantize_per_tensor_default_57"  [label="(3, 197, 1, 768)", style=solid];
+"758 contiguous_9" -> "765 select_29"  [label="(3, 197, 1, 768)", style=solid];
+"759 quantize_per_tensor_default_56" -> "760 dequantize_per_tensor_default_56"  [label="(3, 197, 1, 768)", style=solid];
+"760 dequantize_per_tensor_default_56" -> "761 select_27"  [label="(3, 197, 1, 768)", style=solid];
+"761 select_27" -> "766 view_72"  [label="(197, 1, 768)", style=solid];
+"762 quantize_per_tensor_default_57" -> "763 dequantize_per_tensor_default_57"  [label="(3, 197, 1, 768)", style=solid];
+"763 dequantize_per_tensor_default_57" -> "764 select_28"  [label="(3, 197, 1, 768)", style=solid];
+"764 select_28" -> "768 view_73"  [label="(197, 1, 768)", style=solid];
+"765 select_29" -> "770 view_74"  [label="(197, 1, 768)", style=solid];
+"766 view_72" -> "767 transpose_56"  [label="(197, 12, 64)", style=solid];
+"767 transpose_56" -> "772 view_75"  [label="(12, 197, 64)", style=solid];
+"768 view_73" -> "769 transpose_57"  [label="(197, 12, 64)", style=solid];
+"769 transpose_57" -> "773 view_76"  [label="(12, 197, 64)", style=solid];
+"770 view_74" -> "771 transpose_58"  [label="(197, 12, 64)", style=solid];
+"771 transpose_58" -> "774 view_77"  [label="(12, 197, 64)", style=solid];
+"772 view_75" -> "775 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
+"773 view_76" -> "775 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
+"774 view_77" -> "775 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
+"775 scaled_dot_product_attention_9" -> "776 permute_10"  [label="(1, 12, 197, 64)", style=solid];
+"776 permute_10" -> "777 view_78"  [label="(197, 1, 12, 64)", style=solid];
+"777 view_78" -> "778 view_78_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"778 view_78_0_0_nncf_smooth_quant_0" -> "779 quantize_per_tensor_default_58"  [label="(197, 768)", style=solid];
+"779 quantize_per_tensor_default_58" -> "780 dequantize_per_tensor_default_58"  [label="(197, 768)", style=solid];
+"780 dequantize_per_tensor_default_58" -> "787 linear_37"  [label="(197, 768)", style=solid];
+"781 encoder_layers_encoder_layer_9_self_attention_out_proj_bias_0_0" -> "787 linear_37"  [label="(768,)", style=solid];
+"782 scale_updated_constant38" -> "784 mul_tensor_38"  [label="(768, 1)", style=solid];
+"783 compressed_weight_updated_constant38" -> "784 mul_tensor_38"  [label="(768, 768)", style=solid];
+"784 mul_tensor_38" -> "786 sub_tensor_38"  [label="(768, 768)", style=solid];
+"785 zero_point_updated_constant38" -> "786 sub_tensor_38"  [label="(768, 1)", style=solid];
+"786 sub_tensor_38" -> "787 linear_37"  [label="(768, 768)", style=solid];
+"787 linear_37" -> "788 view_79"  [label="(197, 768)", style=solid];
+"788 view_79" -> "789 transpose_59"  [label="(197, 1, 768)", style=solid];
+"789 transpose_59" -> "790 dropout_28"  [label="(1, 197, 768)", style=solid];
+"790 dropout_28" -> "791 add_19"  [label="(1, 197, 768)", style=solid];
+"791 add_19" -> "792 layer_norm_19"  [label="(1, 197, 768)", style=solid];
+"791 add_19" -> "816 add_20"  [label="(1, 197, 768)", style=solid];
+"792 layer_norm_19" -> "793 layer_norm_19_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"793 layer_norm_19_0_0_nncf_smooth_quant_0" -> "794 quantize_per_tensor_default_59"  [label="(1, 197, 768)", style=solid];
+"794 quantize_per_tensor_default_59" -> "795 dequantize_per_tensor_default_59"  [label="(1, 197, 768)", style=solid];
+"795 dequantize_per_tensor_default_59" -> "802 linear_38"  [label="(1, 197, 768)", style=solid];
+"796 encoder_layers_encoder_layer_9_mlp_0_bias_0_0" -> "802 linear_38"  [label="(3072,)", style=solid];
+"797 scale_updated_constant39" -> "799 mul_tensor_39"  [label="(3072, 1)", style=solid];
+"798 compressed_weight_updated_constant39" -> "799 mul_tensor_39"  [label="(3072, 768)", style=solid];
+"799 mul_tensor_39" -> "801 sub_tensor_39"  [label="(3072, 768)", style=solid];
+"800 zero_point_updated_constant39" -> "801 sub_tensor_39"  [label="(3072, 1)", style=solid];
+"801 sub_tensor_39" -> "802 linear_38"  [label="(3072, 768)", style=solid];
+"802 linear_38" -> "803 gelu_9"  [label="(1, 197, 3072)", style=solid];
+"803 gelu_9" -> "804 dropout_29"  [label="(1, 197, 3072)", style=solid];
+"804 dropout_29" -> "805 dropout_29_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"805 dropout_29_0_0_nncf_smooth_quant_0" -> "806 quantize_per_tensor_default_60"  [label="(1, 197, 3072)", style=solid];
+"806 quantize_per_tensor_default_60" -> "807 dequantize_per_tensor_default_60"  [label="(1, 197, 3072)", style=solid];
+"807 dequantize_per_tensor_default_60" -> "814 linear_39"  [label="(1, 197, 3072)", style=solid];
+"808 encoder_layers_encoder_layer_9_mlp_3_bias_0_0" -> "814 linear_39"  [label="(768,)", style=solid];
+"809 scale_updated_constant40" -> "811 mul_tensor_40"  [label="(768, 1)", style=solid];
+"810 compressed_weight_updated_constant40" -> "811 mul_tensor_40"  [label="(768, 3072)", style=solid];
+"811 mul_tensor_40" -> "813 sub_tensor_40"  [label="(768, 3072)", style=solid];
+"812 zero_point_updated_constant40" -> "813 sub_tensor_40"  [label="(768, 1)", style=solid];
+"813 sub_tensor_40" -> "814 linear_39"  [label="(768, 3072)", style=solid];
+"814 linear_39" -> "815 dropout_30"  [label="(1, 197, 768)", style=solid];
+"815 dropout_30" -> "816 add_20"  [label="(1, 197, 768)", style=solid];
+"816 add_20" -> "817 layer_norm_20"  [label="(1, 197, 768)", style=solid];
+"816 add_20" -> "866 add_21"  [label="(1, 197, 768)", style=solid];
+"817 layer_norm_20" -> "818 transpose_60"  [label="(1, 197, 768)", style=solid];
+"818 transpose_60" -> "819 transpose_60_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"819 transpose_60_0_0_nncf_smooth_quant_0" -> "820 quantize_per_tensor_default_61"  [label="(197, 1, 768)", style=solid];
+"820 quantize_per_tensor_default_61" -> "821 dequantize_per_tensor_default_61"  [label="(197, 1, 768)", style=solid];
+"821 dequantize_per_tensor_default_61" -> "828 linear_40"  [label="(197, 1, 768)", style=solid];
+"822 encoder_layers_encoder_layer_10_self_attention_in_proj_bias_0_0" -> "828 linear_40"  [label="(2304,)", style=solid];
+"823 scale_updated_constant41" -> "825 mul_tensor_41"  [label="(2304, 1)", style=solid];
+"824 compressed_weight_updated_constant41" -> "825 mul_tensor_41"  [label="(2304, 768)", style=solid];
+"825 mul_tensor_41" -> "827 sub_tensor_41"  [label="(2304, 768)", style=solid];
+"826 zero_point_updated_constant41" -> "827 sub_tensor_41"  [label="(2304, 1)", style=solid];
+"827 sub_tensor_41" -> "828 linear_40"  [label="(2304, 768)", style=solid];
+"828 linear_40" -> "829 unflatten_10"  [label="(197, 1, 2304)", style=solid];
+"829 unflatten_10" -> "830 unsqueeze_10"  [label="(197, 1, 3, 768)", style=solid];
+"830 unsqueeze_10" -> "831 transpose_61"  [label="(1, 197, 1, 3, 768)", style=solid];
+"831 transpose_61" -> "832 squeeze_10"  [label="(3, 197, 1, 1, 768)", style=solid];
+"832 squeeze_10" -> "833 contiguous_10"  [label="(3, 197, 1, 768)", style=solid];
+"833 contiguous_10" -> "834 quantize_per_tensor_default_62"  [label="(3, 197, 1, 768)", style=solid];
+"833 contiguous_10" -> "837 quantize_per_tensor_default_63"  [label="(3, 197, 1, 768)", style=solid];
+"833 contiguous_10" -> "840 select_32"  [label="(3, 197, 1, 768)", style=solid];
+"834 quantize_per_tensor_default_62" -> "835 dequantize_per_tensor_default_62"  [label="(3, 197, 1, 768)", style=solid];
+"835 dequantize_per_tensor_default_62" -> "836 select_30"  [label="(3, 197, 1, 768)", style=solid];
+"836 select_30" -> "841 view_80"  [label="(197, 1, 768)", style=solid];
+"837 quantize_per_tensor_default_63" -> "838 dequantize_per_tensor_default_63"  [label="(3, 197, 1, 768)", style=solid];
+"838 dequantize_per_tensor_default_63" -> "839 select_31"  [label="(3, 197, 1, 768)", style=solid];
+"839 select_31" -> "843 view_81"  [label="(197, 1, 768)", style=solid];
+"840 select_32" -> "845 view_82"  [label="(197, 1, 768)", style=solid];
+"841 view_80" -> "842 transpose_62"  [label="(197, 12, 64)", style=solid];
+"842 transpose_62" -> "847 view_83"  [label="(12, 197, 64)", style=solid];
+"843 view_81" -> "844 transpose_63"  [label="(197, 12, 64)", style=solid];
+"844 transpose_63" -> "848 view_84"  [label="(12, 197, 64)", style=solid];
+"845 view_82" -> "846 transpose_64"  [label="(197, 12, 64)", style=solid];
+"846 transpose_64" -> "849 view_85"  [label="(12, 197, 64)", style=solid];
+"847 view_83" -> "850 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
+"848 view_84" -> "850 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
+"849 view_85" -> "850 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
+"850 scaled_dot_product_attention_10" -> "851 permute_11"  [label="(1, 12, 197, 64)", style=solid];
+"851 permute_11" -> "852 view_86"  [label="(197, 1, 12, 64)", style=solid];
+"852 view_86" -> "853 view_86_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"853 view_86_0_0_nncf_smooth_quant_0" -> "854 quantize_per_tensor_default_64"  [label="(197, 768)", style=solid];
+"854 quantize_per_tensor_default_64" -> "855 dequantize_per_tensor_default_64"  [label="(197, 768)", style=solid];
+"855 dequantize_per_tensor_default_64" -> "862 linear_41"  [label="(197, 768)", style=solid];
+"856 encoder_layers_encoder_layer_10_self_attention_out_proj_bias_0_0" -> "862 linear_41"  [label="(768,)", style=solid];
+"857 scale_updated_constant42" -> "859 mul_tensor_42"  [label="(768, 1)", style=solid];
+"858 compressed_weight_updated_constant42" -> "859 mul_tensor_42"  [label="(768, 768)", style=solid];
+"859 mul_tensor_42" -> "861 sub_tensor_42"  [label="(768, 768)", style=solid];
+"860 zero_point_updated_constant42" -> "861 sub_tensor_42"  [label="(768, 1)", style=solid];
+"861 sub_tensor_42" -> "862 linear_41"  [label="(768, 768)", style=solid];
+"862 linear_41" -> "863 view_87"  [label="(197, 768)", style=solid];
+"863 view_87" -> "864 transpose_65"  [label="(197, 1, 768)", style=solid];
+"864 transpose_65" -> "865 dropout_31"  [label="(1, 197, 768)", style=solid];
+"865 dropout_31" -> "866 add_21"  [label="(1, 197, 768)", style=solid];
+"866 add_21" -> "867 layer_norm_21"  [label="(1, 197, 768)", style=solid];
+"866 add_21" -> "891 add_22"  [label="(1, 197, 768)", style=solid];
+"867 layer_norm_21" -> "868 layer_norm_21_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"868 layer_norm_21_0_0_nncf_smooth_quant_0" -> "869 quantize_per_tensor_default_65"  [label="(1, 197, 768)", style=solid];
+"869 quantize_per_tensor_default_65" -> "870 dequantize_per_tensor_default_65"  [label="(1, 197, 768)", style=solid];
+"870 dequantize_per_tensor_default_65" -> "877 linear_42"  [label="(1, 197, 768)", style=solid];
+"871 encoder_layers_encoder_layer_10_mlp_0_bias_0_0" -> "877 linear_42"  [label="(3072,)", style=solid];
+"872 scale_updated_constant43" -> "874 mul_tensor_43"  [label="(3072, 1)", style=solid];
+"873 compressed_weight_updated_constant43" -> "874 mul_tensor_43"  [label="(3072, 768)", style=solid];
+"874 mul_tensor_43" -> "876 sub_tensor_43"  [label="(3072, 768)", style=solid];
+"875 zero_point_updated_constant43" -> "876 sub_tensor_43"  [label="(3072, 1)", style=solid];
+"876 sub_tensor_43" -> "877 linear_42"  [label="(3072, 768)", style=solid];
+"877 linear_42" -> "878 gelu_10"  [label="(1, 197, 3072)", style=solid];
+"878 gelu_10" -> "879 dropout_32"  [label="(1, 197, 3072)", style=solid];
+"879 dropout_32" -> "880 dropout_32_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"880 dropout_32_0_0_nncf_smooth_quant_0" -> "881 quantize_per_tensor_default_66"  [label="(1, 197, 3072)", style=solid];
+"881 quantize_per_tensor_default_66" -> "882 dequantize_per_tensor_default_66"  [label="(1, 197, 3072)", style=solid];
+"882 dequantize_per_tensor_default_66" -> "889 linear_43"  [label="(1, 197, 3072)", style=solid];
+"883 encoder_layers_encoder_layer_10_mlp_3_bias_0_0" -> "889 linear_43"  [label="(768,)", style=solid];
+"884 scale_updated_constant44" -> "886 mul_tensor_44"  [label="(768, 1)", style=solid];
+"885 compressed_weight_updated_constant44" -> "886 mul_tensor_44"  [label="(768, 3072)", style=solid];
+"886 mul_tensor_44" -> "888 sub_tensor_44"  [label="(768, 3072)", style=solid];
+"887 zero_point_updated_constant44" -> "888 sub_tensor_44"  [label="(768, 1)", style=solid];
+"888 sub_tensor_44" -> "889 linear_43"  [label="(768, 3072)", style=solid];
+"889 linear_43" -> "890 dropout_33"  [label="(1, 197, 768)", style=solid];
+"890 dropout_33" -> "891 add_22"  [label="(1, 197, 768)", style=solid];
+"891 add_22" -> "892 layer_norm_22"  [label="(1, 197, 768)", style=solid];
+"891 add_22" -> "941 add_23"  [label="(1, 197, 768)", style=solid];
+"892 layer_norm_22" -> "893 transpose_66"  [label="(1, 197, 768)", style=solid];
+"893 transpose_66" -> "894 transpose_66_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"894 transpose_66_0_0_nncf_smooth_quant_0" -> "895 quantize_per_tensor_default_67"  [label="(197, 1, 768)", style=solid];
+"895 quantize_per_tensor_default_67" -> "896 dequantize_per_tensor_default_67"  [label="(197, 1, 768)", style=solid];
+"896 dequantize_per_tensor_default_67" -> "903 linear_44"  [label="(197, 1, 768)", style=solid];
+"897 encoder_layers_encoder_layer_11_self_attention_in_proj_bias_0_0" -> "903 linear_44"  [label="(2304,)", style=solid];
+"898 scale_updated_constant45" -> "900 mul_tensor_45"  [label="(2304, 1)", style=solid];
+"899 compressed_weight_updated_constant45" -> "900 mul_tensor_45"  [label="(2304, 768)", style=solid];
+"900 mul_tensor_45" -> "902 sub_tensor_45"  [label="(2304, 768)", style=solid];
+"901 zero_point_updated_constant45" -> "902 sub_tensor_45"  [label="(2304, 1)", style=solid];
+"902 sub_tensor_45" -> "903 linear_44"  [label="(2304, 768)", style=solid];
+"903 linear_44" -> "904 unflatten_11"  [label="(197, 1, 2304)", style=solid];
+"904 unflatten_11" -> "905 unsqueeze_11"  [label="(197, 1, 3, 768)", style=solid];
+"905 unsqueeze_11" -> "906 transpose_67"  [label="(1, 197, 1, 3, 768)", style=solid];
+"906 transpose_67" -> "907 squeeze_11"  [label="(3, 197, 1, 1, 768)", style=solid];
+"907 squeeze_11" -> "908 contiguous_11"  [label="(3, 197, 1, 768)", style=solid];
+"908 contiguous_11" -> "909 quantize_per_tensor_default_68"  [label="(3, 197, 1, 768)", style=solid];
+"908 contiguous_11" -> "912 quantize_per_tensor_default_69"  [label="(3, 197, 1, 768)", style=solid];
+"908 contiguous_11" -> "915 select_35"  [label="(3, 197, 1, 768)", style=solid];
+"909 quantize_per_tensor_default_68" -> "910 dequantize_per_tensor_default_68"  [label="(3, 197, 1, 768)", style=solid];
+"910 dequantize_per_tensor_default_68" -> "911 select_33"  [label="(3, 197, 1, 768)", style=solid];
+"911 select_33" -> "916 view_88"  [label="(197, 1, 768)", style=solid];
+"912 quantize_per_tensor_default_69" -> "913 dequantize_per_tensor_default_69"  [label="(3, 197, 1, 768)", style=solid];
+"913 dequantize_per_tensor_default_69" -> "914 select_34"  [label="(3, 197, 1, 768)", style=solid];
+"914 select_34" -> "918 view_89"  [label="(197, 1, 768)", style=solid];
+"915 select_35" -> "920 view_90"  [label="(197, 1, 768)", style=solid];
+"916 view_88" -> "917 transpose_68"  [label="(197, 12, 64)", style=solid];
+"917 transpose_68" -> "922 view_91"  [label="(12, 197, 64)", style=solid];
+"918 view_89" -> "919 transpose_69"  [label="(197, 12, 64)", style=solid];
+"919 transpose_69" -> "923 view_92"  [label="(12, 197, 64)", style=solid];
+"920 view_90" -> "921 transpose_70"  [label="(197, 12, 64)", style=solid];
+"921 transpose_70" -> "924 view_93"  [label="(12, 197, 64)", style=solid];
+"922 view_91" -> "925 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
+"923 view_92" -> "925 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
+"924 view_93" -> "925 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
+"925 scaled_dot_product_attention_11" -> "926 permute_12"  [label="(1, 12, 197, 64)", style=solid];
+"926 permute_12" -> "927 view_94"  [label="(197, 1, 12, 64)", style=solid];
+"927 view_94" -> "928 view_94_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"928 view_94_0_0_nncf_smooth_quant_0" -> "929 quantize_per_tensor_default_70"  [label="(197, 768)", style=solid];
+"929 quantize_per_tensor_default_70" -> "930 dequantize_per_tensor_default_70"  [label="(197, 768)", style=solid];
+"930 dequantize_per_tensor_default_70" -> "937 linear_45"  [label="(197, 768)", style=solid];
+"931 encoder_layers_encoder_layer_11_self_attention_out_proj_bias_0_0" -> "937 linear_45"  [label="(768,)", style=solid];
+"932 scale_updated_constant46" -> "934 mul_tensor_46"  [label="(768, 1)", style=solid];
+"933 compressed_weight_updated_constant46" -> "934 mul_tensor_46"  [label="(768, 768)", style=solid];
+"934 mul_tensor_46" -> "936 sub_tensor_46"  [label="(768, 768)", style=solid];
+"935 zero_point_updated_constant46" -> "936 sub_tensor_46"  [label="(768, 1)", style=solid];
+"936 sub_tensor_46" -> "937 linear_45"  [label="(768, 768)", style=solid];
+"937 linear_45" -> "938 view_95"  [label="(197, 768)", style=solid];
+"938 view_95" -> "939 transpose_71"  [label="(197, 1, 768)", style=solid];
+"939 transpose_71" -> "940 dropout_34"  [label="(1, 197, 768)", style=solid];
+"940 dropout_34" -> "941 add_23"  [label="(1, 197, 768)", style=solid];
+"941 add_23" -> "942 layer_norm_23"  [label="(1, 197, 768)", style=solid];
+"941 add_23" -> "966 add_24"  [label="(1, 197, 768)", style=solid];
+"942 layer_norm_23" -> "943 layer_norm_23_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"943 layer_norm_23_0_0_nncf_smooth_quant_0" -> "944 quantize_per_tensor_default_71"  [label="(1, 197, 768)", style=solid];
+"944 quantize_per_tensor_default_71" -> "945 dequantize_per_tensor_default_71"  [label="(1, 197, 768)", style=solid];
+"945 dequantize_per_tensor_default_71" -> "952 linear_46"  [label="(1, 197, 768)", style=solid];
+"946 encoder_layers_encoder_layer_11_mlp_0_bias_0_0" -> "952 linear_46"  [label="(3072,)", style=solid];
+"947 scale_updated_constant47" -> "949 mul_tensor_47"  [label="(3072, 1)", style=solid];
+"948 compressed_weight_updated_constant47" -> "949 mul_tensor_47"  [label="(3072, 768)", style=solid];
+"949 mul_tensor_47" -> "951 sub_tensor_47"  [label="(3072, 768)", style=solid];
+"950 zero_point_updated_constant47" -> "951 sub_tensor_47"  [label="(3072, 1)", style=solid];
+"951 sub_tensor_47" -> "952 linear_46"  [label="(3072, 768)", style=solid];
+"952 linear_46" -> "953 gelu_11"  [label="(1, 197, 3072)", style=solid];
+"953 gelu_11" -> "954 dropout_35"  [label="(1, 197, 3072)", style=solid];
+"954 dropout_35" -> "955 dropout_35_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"955 dropout_35_0_0_nncf_smooth_quant_0" -> "956 quantize_per_tensor_default_72"  [label="(1, 197, 3072)", style=solid];
+"956 quantize_per_tensor_default_72" -> "957 dequantize_per_tensor_default_72"  [label="(1, 197, 3072)", style=solid];
+"957 dequantize_per_tensor_default_72" -> "964 linear_47"  [label="(1, 197, 3072)", style=solid];
+"958 encoder_layers_encoder_layer_11_mlp_3_bias_0_0" -> "964 linear_47"  [label="(768,)", style=solid];
+"959 scale_updated_constant48" -> "961 mul_tensor_48"  [label="(768, 1)", style=solid];
+"960 compressed_weight_updated_constant48" -> "961 mul_tensor_48"  [label="(768, 3072)", style=solid];
+"961 mul_tensor_48" -> "963 sub_tensor_48"  [label="(768, 3072)", style=solid];
+"962 zero_point_updated_constant48" -> "963 sub_tensor_48"  [label="(768, 1)", style=solid];
+"963 sub_tensor_48" -> "964 linear_47"  [label="(768, 3072)", style=solid];
+"964 linear_47" -> "965 dropout_36"  [label="(1, 197, 768)", style=solid];
+"965 dropout_36" -> "966 add_24"  [label="(1, 197, 768)", style=solid];
+"966 add_24" -> "967 layer_norm_24"  [label="(1, 197, 768)", style=solid];
 "967 layer_norm_24" -> "968 slice_1"  [label="(1, 197, 768)", style=solid];
 "968 slice_1" -> "969 select_36"  [label="(1, 197, 768)", style=solid];
 "969 select_36" -> "970 select_36_0_0_nncf_smooth_quant_0"  [label="(1, 768)", style=solid];
 "970 select_36_0_0_nncf_smooth_quant_0" -> "971 quantize_per_tensor_default_73"  [label="(1, 768)", style=solid];
 "971 quantize_per_tensor_default_73" -> "972 dequantize_per_tensor_default_73"  [label="(1, 768)", style=solid];
 "972 dequantize_per_tensor_default_73" -> "979 linear_48"  [label="(1, 768)", style=solid];
-"973 _param_constant151_0_0" -> "979 linear_48"  [label="(1000,)", style=solid];
+"973 heads_head_bias_0_0" -> "979 linear_48"  [label="(1000,)", style=solid];
 "974 scale_updated_constant49" -> "976 mul_tensor_49"  [label="(1000, 1)", style=solid];
 "975 compressed_weight_updated_constant49" -> "976 mul_tensor_49"  [label="(1000, 768)", style=solid];
 "976 mul_tensor_49" -> "978 sub_tensor_49"  [label="(1000, 768)", style=solid];
diff --git a/tests/torch/data/reference_graphs/fx/quantized/mobilenet_v3_small.dot b/tests/torch/data/reference_graphs/fx/quantized/mobilenet_v3_small.dot
index f29084b18c6..bbba2af8408 100644
--- a/tests/torch/data/reference_graphs/fx/quantized/mobilenet_v3_small.dot
+++ b/tests/torch/data/reference_graphs/fx/quantized/mobilenet_v3_small.dot
@@ -1,488 +1,488 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 quantize_per_tensor_default" [id=1, type=quantize_per_tensor];
-"2 dequantize_per_tensor_default" [id=2, type=dequantize_per_tensor];
-"3 conv2d_scale_0" [id=3, type=get_attr];
-"4 conv2d_zero_point_0" [id=4, type=get_attr];
-"5 compressed_weight_updated_constant0" [id=5, type=get_attr];
-"6 quantize_per_channel_default" [id=6, type=quantize_per_channel];
-"7 dequantize_per_channel_default" [id=7, type=dequantize_per_channel];
-"8 features_0_0_weight_bias_0_0" [id=8, type=get_attr];
-"9 conv2d" [id=9, type=conv2d];
-"10 hardswish__scale_0" [id=10, type=get_attr];
-"11 hardswish__zero_point_0" [id=11, type=get_attr];
-"12 hardswish_" [id=12, type=hardswish_];
-"13 quantize_per_channel_default_1" [id=13, type=quantize_per_channel];
-"14 dequantize_per_channel_default_1" [id=14, type=dequantize_per_channel];
-"15 conv2d_1_scale_0" [id=15, type=get_attr];
-"16 conv2d_1_zero_point_0" [id=16, type=get_attr];
-"17 compressed_weight_updated_constant1" [id=17, type=get_attr];
-"18 quantize_per_channel_default_2" [id=18, type=quantize_per_channel];
-"19 dequantize_per_channel_default_2" [id=19, type=dequantize_per_channel];
-"20 features_1_block_0_0_weight_bias_0_0" [id=20, type=get_attr];
-"21 conv2d_1" [id=21, type=conv2d];
-"22 relu_" [id=22, type=relu_];
-"23 quantize_per_tensor_default_1" [id=23, type=quantize_per_tensor];
-"24 dequantize_per_tensor_default_2" [id=24, type=dequantize_per_tensor];
-"25 dequantize_per_tensor_default_1" [id=25, type=dequantize_per_tensor];
-"26 adaptive_avg_pool2d" [id=26, type=adaptive_avg_pool2d];
-"27 _param_constant6" [id=27, type=get_attr];
-"28 _param_constant7_0_0" [id=28, type=get_attr];
-"29 conv2d_2" [id=29, type=conv2d];
-"30 relu" [id=30, type=relu];
-"31 _param_constant8" [id=31, type=get_attr];
-"32 _param_constant9_0_0" [id=32, type=get_attr];
-"33 conv2d_3" [id=33, type=conv2d];
-"34 hardsigmoid" [id=34, type=hardsigmoid];
-"35 quantize_per_tensor_default_2" [id=35, type=quantize_per_tensor];
-"36 dequantize_per_tensor_default_3" [id=36, type=dequantize_per_tensor];
-"37 mul" [id=37, type=mul];
-"38 quantize_per_tensor_default_3" [id=38, type=quantize_per_tensor];
-"39 dequantize_per_tensor_default_4" [id=39, type=dequantize_per_tensor];
-"40 conv2d_4_scale_0" [id=40, type=get_attr];
-"41 conv2d_4_zero_point_0" [id=41, type=get_attr];
-"42 compressed_weight_updated_constant2" [id=42, type=get_attr];
-"43 quantize_per_channel_default_3" [id=43, type=quantize_per_channel];
-"44 dequantize_per_channel_default_3" [id=44, type=dequantize_per_channel];
-"45 features_1_block_2_0_weight_bias_0_0" [id=45, type=get_attr];
-"46 conv2d_4" [id=46, type=conv2d];
-"47 quantize_per_tensor_default_4" [id=47, type=quantize_per_tensor];
-"48 dequantize_per_tensor_default_5" [id=48, type=dequantize_per_tensor];
-"49 conv2d_5_scale_0" [id=49, type=get_attr];
-"50 conv2d_5_zero_point_0" [id=50, type=get_attr];
-"51 compressed_weight_updated_constant3" [id=51, type=get_attr];
-"52 quantize_per_channel_default_4" [id=52, type=quantize_per_channel];
-"53 dequantize_per_channel_default_4" [id=53, type=dequantize_per_channel];
-"54 features_2_block_0_0_weight_bias_0_0" [id=54, type=get_attr];
-"55 conv2d_5" [id=55, type=conv2d];
-"56 relu__1_scale_0" [id=56, type=get_attr];
-"57 relu__1_zero_point_0" [id=57, type=get_attr];
-"58 relu__1" [id=58, type=relu_];
-"59 quantize_per_channel_default_5" [id=59, type=quantize_per_channel];
-"60 dequantize_per_channel_default_5" [id=60, type=dequantize_per_channel];
-"61 conv2d_6_scale_0" [id=61, type=get_attr];
-"62 conv2d_6_zero_point_0" [id=62, type=get_attr];
-"63 compressed_weight_updated_constant4" [id=63, type=get_attr];
-"64 quantize_per_channel_default_6" [id=64, type=quantize_per_channel];
-"65 dequantize_per_channel_default_6" [id=65, type=dequantize_per_channel];
-"66 features_2_block_1_0_weight_bias_0_0" [id=66, type=get_attr];
-"67 conv2d_6" [id=67, type=conv2d];
-"68 relu__2" [id=68, type=relu_];
-"69 quantize_per_tensor_default_5" [id=69, type=quantize_per_tensor];
-"70 dequantize_per_tensor_default_6" [id=70, type=dequantize_per_tensor];
-"71 conv2d_7_scale_0" [id=71, type=get_attr];
-"72 conv2d_7_zero_point_0" [id=72, type=get_attr];
-"73 compressed_weight_updated_constant5" [id=73, type=get_attr];
-"74 quantize_per_channel_default_7" [id=74, type=quantize_per_channel];
-"75 dequantize_per_channel_default_7" [id=75, type=dequantize_per_channel];
-"76 features_2_block_2_0_weight_bias_0_0" [id=76, type=get_attr];
-"77 conv2d_7" [id=77, type=conv2d];
-"78 quantize_per_tensor_default_6" [id=78, type=quantize_per_tensor];
-"79 dequantize_per_tensor_default_8" [id=79, type=dequantize_per_tensor];
-"80 dequantize_per_tensor_default_7" [id=80, type=dequantize_per_tensor];
-"81 conv2d_8_scale_0" [id=81, type=get_attr];
-"82 conv2d_8_zero_point_0" [id=82, type=get_attr];
-"83 compressed_weight_updated_constant6" [id=83, type=get_attr];
-"84 quantize_per_channel_default_8" [id=84, type=quantize_per_channel];
-"85 dequantize_per_channel_default_8" [id=85, type=dequantize_per_channel];
-"86 features_3_block_0_0_weight_bias_0_0" [id=86, type=get_attr];
-"87 conv2d_8" [id=87, type=conv2d];
-"88 relu__3_scale_0" [id=88, type=get_attr];
-"89 relu__3_zero_point_0" [id=89, type=get_attr];
-"90 relu__3" [id=90, type=relu_];
-"91 quantize_per_channel_default_9" [id=91, type=quantize_per_channel];
-"92 dequantize_per_channel_default_9" [id=92, type=dequantize_per_channel];
-"93 conv2d_9_scale_0" [id=93, type=get_attr];
-"94 conv2d_9_zero_point_0" [id=94, type=get_attr];
-"95 compressed_weight_updated_constant7" [id=95, type=get_attr];
-"96 quantize_per_channel_default_10" [id=96, type=quantize_per_channel];
-"97 dequantize_per_channel_default_10" [id=97, type=dequantize_per_channel];
-"98 features_3_block_1_0_weight_bias_0_0" [id=98, type=get_attr];
-"99 conv2d_9" [id=99, type=conv2d];
-"100 relu__4" [id=100, type=relu_];
-"101 quantize_per_tensor_default_7" [id=101, type=quantize_per_tensor];
-"102 dequantize_per_tensor_default_9" [id=102, type=dequantize_per_tensor];
-"103 conv2d_10_scale_0" [id=103, type=get_attr];
-"104 conv2d_10_zero_point_0" [id=104, type=get_attr];
-"105 compressed_weight_updated_constant8" [id=105, type=get_attr];
-"106 quantize_per_channel_default_11" [id=106, type=quantize_per_channel];
-"107 dequantize_per_channel_default_11" [id=107, type=dequantize_per_channel];
-"108 features_3_block_2_0_weight_bias_0_0" [id=108, type=get_attr];
-"109 conv2d_10" [id=109, type=conv2d];
-"110 quantize_per_tensor_default_8" [id=110, type=quantize_per_tensor];
-"111 dequantize_per_tensor_default_10" [id=111, type=dequantize_per_tensor];
-"112 add_" [id=112, type=add_];
-"113 quantize_per_tensor_default_9" [id=113, type=quantize_per_tensor];
-"114 dequantize_per_tensor_default_11" [id=114, type=dequantize_per_tensor];
-"115 conv2d_11_scale_0" [id=115, type=get_attr];
-"116 conv2d_11_zero_point_0" [id=116, type=get_attr];
-"117 compressed_weight_updated_constant9" [id=117, type=get_attr];
-"118 quantize_per_channel_default_12" [id=118, type=quantize_per_channel];
-"119 dequantize_per_channel_default_12" [id=119, type=dequantize_per_channel];
-"120 features_4_block_0_0_weight_bias_0_0" [id=120, type=get_attr];
-"121 conv2d_11" [id=121, type=conv2d];
-"122 hardswish__1_scale_0" [id=122, type=get_attr];
-"123 hardswish__1_zero_point_0" [id=123, type=get_attr];
-"124 hardswish__1" [id=124, type=hardswish_];
-"125 quantize_per_channel_default_13" [id=125, type=quantize_per_channel];
-"126 dequantize_per_channel_default_13" [id=126, type=dequantize_per_channel];
-"127 conv2d_12_scale_0" [id=127, type=get_attr];
-"128 conv2d_12_zero_point_0" [id=128, type=get_attr];
-"129 compressed_weight_updated_constant10" [id=129, type=get_attr];
-"130 quantize_per_channel_default_14" [id=130, type=quantize_per_channel];
-"131 dequantize_per_channel_default_14" [id=131, type=dequantize_per_channel];
-"132 features_4_block_1_0_weight_bias_0_0" [id=132, type=get_attr];
-"133 conv2d_12" [id=133, type=conv2d];
-"134 hardswish__2" [id=134, type=hardswish_];
-"135 quantize_per_tensor_default_10" [id=135, type=quantize_per_tensor];
-"136 dequantize_per_tensor_default_13" [id=136, type=dequantize_per_tensor];
-"137 dequantize_per_tensor_default_12" [id=137, type=dequantize_per_tensor];
-"138 adaptive_avg_pool2d_1" [id=138, type=adaptive_avg_pool2d];
-"139 _param_constant37" [id=139, type=get_attr];
-"140 _param_constant38_0_0" [id=140, type=get_attr];
-"141 conv2d_13" [id=141, type=conv2d];
-"142 relu_1" [id=142, type=relu];
-"143 _param_constant39" [id=143, type=get_attr];
-"144 _param_constant40_0_0" [id=144, type=get_attr];
-"145 conv2d_14" [id=145, type=conv2d];
-"146 hardsigmoid_1" [id=146, type=hardsigmoid];
-"147 quantize_per_tensor_default_11" [id=147, type=quantize_per_tensor];
-"148 dequantize_per_tensor_default_14" [id=148, type=dequantize_per_tensor];
-"149 mul_1" [id=149, type=mul];
-"150 quantize_per_tensor_default_12" [id=150, type=quantize_per_tensor];
-"151 dequantize_per_tensor_default_15" [id=151, type=dequantize_per_tensor];
-"152 conv2d_15_scale_0" [id=152, type=get_attr];
-"153 conv2d_15_zero_point_0" [id=153, type=get_attr];
-"154 compressed_weight_updated_constant11" [id=154, type=get_attr];
-"155 quantize_per_channel_default_15" [id=155, type=quantize_per_channel];
-"156 dequantize_per_channel_default_15" [id=156, type=dequantize_per_channel];
-"157 features_4_block_3_0_weight_bias_0_0" [id=157, type=get_attr];
-"158 conv2d_15" [id=158, type=conv2d];
-"159 quantize_per_tensor_default_13" [id=159, type=quantize_per_tensor];
-"160 dequantize_per_tensor_default_17" [id=160, type=dequantize_per_tensor];
-"161 dequantize_per_tensor_default_16" [id=161, type=dequantize_per_tensor];
-"162 conv2d_16_scale_0" [id=162, type=get_attr];
-"163 conv2d_16_zero_point_0" [id=163, type=get_attr];
-"164 compressed_weight_updated_constant12" [id=164, type=get_attr];
-"165 quantize_per_channel_default_16" [id=165, type=quantize_per_channel];
-"166 dequantize_per_channel_default_16" [id=166, type=dequantize_per_channel];
-"167 features_5_block_0_0_weight_bias_0_0" [id=167, type=get_attr];
-"168 conv2d_16" [id=168, type=conv2d];
-"169 hardswish__3_scale_0" [id=169, type=get_attr];
-"170 hardswish__3_zero_point_0" [id=170, type=get_attr];
-"171 hardswish__3" [id=171, type=hardswish_];
-"172 quantize_per_channel_default_17" [id=172, type=quantize_per_channel];
-"173 dequantize_per_channel_default_17" [id=173, type=dequantize_per_channel];
-"174 conv2d_17_scale_0" [id=174, type=get_attr];
-"175 conv2d_17_zero_point_0" [id=175, type=get_attr];
-"176 compressed_weight_updated_constant13" [id=176, type=get_attr];
-"177 quantize_per_channel_default_18" [id=177, type=quantize_per_channel];
-"178 dequantize_per_channel_default_18" [id=178, type=dequantize_per_channel];
-"179 features_5_block_1_0_weight_bias_0_0" [id=179, type=get_attr];
-"180 conv2d_17" [id=180, type=conv2d];
-"181 hardswish__4" [id=181, type=hardswish_];
-"182 quantize_per_tensor_default_14" [id=182, type=quantize_per_tensor];
-"183 dequantize_per_tensor_default_19" [id=183, type=dequantize_per_tensor];
-"184 dequantize_per_tensor_default_18" [id=184, type=dequantize_per_tensor];
-"185 adaptive_avg_pool2d_2" [id=185, type=adaptive_avg_pool2d];
-"186 _param_constant50" [id=186, type=get_attr];
-"187 _param_constant51_0_0" [id=187, type=get_attr];
-"188 conv2d_18" [id=188, type=conv2d];
-"189 relu_2" [id=189, type=relu];
-"190 _param_constant52" [id=190, type=get_attr];
-"191 _param_constant53_0_0" [id=191, type=get_attr];
-"192 conv2d_19" [id=192, type=conv2d];
-"193 hardsigmoid_2" [id=193, type=hardsigmoid];
-"194 quantize_per_tensor_default_15" [id=194, type=quantize_per_tensor];
-"195 dequantize_per_tensor_default_20" [id=195, type=dequantize_per_tensor];
-"196 mul_2" [id=196, type=mul];
-"197 quantize_per_tensor_default_16" [id=197, type=quantize_per_tensor];
-"198 dequantize_per_tensor_default_21" [id=198, type=dequantize_per_tensor];
-"199 conv2d_20_scale_0" [id=199, type=get_attr];
-"200 conv2d_20_zero_point_0" [id=200, type=get_attr];
-"201 compressed_weight_updated_constant14" [id=201, type=get_attr];
-"202 quantize_per_channel_default_19" [id=202, type=quantize_per_channel];
-"203 dequantize_per_channel_default_19" [id=203, type=dequantize_per_channel];
-"204 features_5_block_3_0_weight_bias_0_0" [id=204, type=get_attr];
-"205 conv2d_20" [id=205, type=conv2d];
-"206 quantize_per_tensor_default_17" [id=206, type=quantize_per_tensor];
-"207 dequantize_per_tensor_default_22" [id=207, type=dequantize_per_tensor];
-"208 add__1" [id=208, type=add_];
-"209 quantize_per_tensor_default_18" [id=209, type=quantize_per_tensor];
-"210 dequantize_per_tensor_default_24" [id=210, type=dequantize_per_tensor];
-"211 dequantize_per_tensor_default_23" [id=211, type=dequantize_per_tensor];
-"212 conv2d_21_scale_0" [id=212, type=get_attr];
-"213 conv2d_21_zero_point_0" [id=213, type=get_attr];
-"214 compressed_weight_updated_constant15" [id=214, type=get_attr];
-"215 quantize_per_channel_default_20" [id=215, type=quantize_per_channel];
-"216 dequantize_per_channel_default_20" [id=216, type=dequantize_per_channel];
-"217 features_6_block_0_0_weight_bias_0_0" [id=217, type=get_attr];
-"218 conv2d_21" [id=218, type=conv2d];
-"219 hardswish__5_scale_0" [id=219, type=get_attr];
-"220 hardswish__5_zero_point_0" [id=220, type=get_attr];
-"221 hardswish__5" [id=221, type=hardswish_];
-"222 quantize_per_channel_default_21" [id=222, type=quantize_per_channel];
-"223 dequantize_per_channel_default_21" [id=223, type=dequantize_per_channel];
-"224 conv2d_22_scale_0" [id=224, type=get_attr];
-"225 conv2d_22_zero_point_0" [id=225, type=get_attr];
-"226 compressed_weight_updated_constant16" [id=226, type=get_attr];
-"227 quantize_per_channel_default_22" [id=227, type=quantize_per_channel];
-"228 dequantize_per_channel_default_22" [id=228, type=dequantize_per_channel];
-"229 features_6_block_1_0_weight_bias_0_0" [id=229, type=get_attr];
-"230 conv2d_22" [id=230, type=conv2d];
-"231 hardswish__6" [id=231, type=hardswish_];
-"232 quantize_per_tensor_default_19" [id=232, type=quantize_per_tensor];
-"233 dequantize_per_tensor_default_26" [id=233, type=dequantize_per_tensor];
-"234 dequantize_per_tensor_default_25" [id=234, type=dequantize_per_tensor];
-"235 adaptive_avg_pool2d_3" [id=235, type=adaptive_avg_pool2d];
-"236 _param_constant63" [id=236, type=get_attr];
-"237 _param_constant64_0_0" [id=237, type=get_attr];
-"238 conv2d_23" [id=238, type=conv2d];
-"239 relu_3" [id=239, type=relu];
-"240 _param_constant65" [id=240, type=get_attr];
-"241 _param_constant66_0_0" [id=241, type=get_attr];
-"242 conv2d_24" [id=242, type=conv2d];
-"243 hardsigmoid_3" [id=243, type=hardsigmoid];
-"244 quantize_per_tensor_default_20" [id=244, type=quantize_per_tensor];
-"245 dequantize_per_tensor_default_27" [id=245, type=dequantize_per_tensor];
-"246 mul_3" [id=246, type=mul];
-"247 quantize_per_tensor_default_21" [id=247, type=quantize_per_tensor];
-"248 dequantize_per_tensor_default_28" [id=248, type=dequantize_per_tensor];
-"249 conv2d_25_scale_0" [id=249, type=get_attr];
-"250 conv2d_25_zero_point_0" [id=250, type=get_attr];
-"251 compressed_weight_updated_constant17" [id=251, type=get_attr];
-"252 quantize_per_channel_default_23" [id=252, type=quantize_per_channel];
-"253 dequantize_per_channel_default_23" [id=253, type=dequantize_per_channel];
-"254 features_6_block_3_0_weight_bias_0_0" [id=254, type=get_attr];
-"255 conv2d_25" [id=255, type=conv2d];
-"256 quantize_per_tensor_default_22" [id=256, type=quantize_per_tensor];
-"257 dequantize_per_tensor_default_29" [id=257, type=dequantize_per_tensor];
-"258 add__2" [id=258, type=add_];
-"259 quantize_per_tensor_default_23" [id=259, type=quantize_per_tensor];
-"260 dequantize_per_tensor_default_30" [id=260, type=dequantize_per_tensor];
-"261 conv2d_26_scale_0" [id=261, type=get_attr];
-"262 conv2d_26_zero_point_0" [id=262, type=get_attr];
-"263 compressed_weight_updated_constant18" [id=263, type=get_attr];
-"264 quantize_per_channel_default_24" [id=264, type=quantize_per_channel];
-"265 dequantize_per_channel_default_24" [id=265, type=dequantize_per_channel];
-"266 features_7_block_0_0_weight_bias_0_0" [id=266, type=get_attr];
-"267 conv2d_26" [id=267, type=conv2d];
-"268 hardswish__7_scale_0" [id=268, type=get_attr];
-"269 hardswish__7_zero_point_0" [id=269, type=get_attr];
-"270 hardswish__7" [id=270, type=hardswish_];
-"271 quantize_per_channel_default_25" [id=271, type=quantize_per_channel];
-"272 dequantize_per_channel_default_25" [id=272, type=dequantize_per_channel];
-"273 conv2d_27_scale_0" [id=273, type=get_attr];
-"274 conv2d_27_zero_point_0" [id=274, type=get_attr];
-"275 compressed_weight_updated_constant19" [id=275, type=get_attr];
-"276 quantize_per_channel_default_26" [id=276, type=quantize_per_channel];
-"277 dequantize_per_channel_default_26" [id=277, type=dequantize_per_channel];
-"278 features_7_block_1_0_weight_bias_0_0" [id=278, type=get_attr];
-"279 conv2d_27" [id=279, type=conv2d];
-"280 hardswish__8" [id=280, type=hardswish_];
-"281 quantize_per_tensor_default_24" [id=281, type=quantize_per_tensor];
-"282 dequantize_per_tensor_default_32" [id=282, type=dequantize_per_tensor];
-"283 dequantize_per_tensor_default_31" [id=283, type=dequantize_per_tensor];
-"284 adaptive_avg_pool2d_4" [id=284, type=adaptive_avg_pool2d];
-"285 _param_constant76" [id=285, type=get_attr];
-"286 _param_constant77_0_0" [id=286, type=get_attr];
-"287 conv2d_28" [id=287, type=conv2d];
-"288 relu_4" [id=288, type=relu];
-"289 _param_constant78" [id=289, type=get_attr];
-"290 _param_constant79_0_0" [id=290, type=get_attr];
-"291 conv2d_29" [id=291, type=conv2d];
-"292 hardsigmoid_4" [id=292, type=hardsigmoid];
-"293 quantize_per_tensor_default_25" [id=293, type=quantize_per_tensor];
-"294 dequantize_per_tensor_default_33" [id=294, type=dequantize_per_tensor];
-"295 mul_4" [id=295, type=mul];
-"296 quantize_per_tensor_default_26" [id=296, type=quantize_per_tensor];
-"297 dequantize_per_tensor_default_34" [id=297, type=dequantize_per_tensor];
-"298 conv2d_30_scale_0" [id=298, type=get_attr];
-"299 conv2d_30_zero_point_0" [id=299, type=get_attr];
-"300 compressed_weight_updated_constant20" [id=300, type=get_attr];
-"301 quantize_per_channel_default_27" [id=301, type=quantize_per_channel];
-"302 dequantize_per_channel_default_27" [id=302, type=dequantize_per_channel];
-"303 features_7_block_3_0_weight_bias_0_0" [id=303, type=get_attr];
-"304 conv2d_30" [id=304, type=conv2d];
-"305 quantize_per_tensor_default_27" [id=305, type=quantize_per_tensor];
-"306 dequantize_per_tensor_default_36" [id=306, type=dequantize_per_tensor];
-"307 dequantize_per_tensor_default_35" [id=307, type=dequantize_per_tensor];
-"308 conv2d_31_scale_0" [id=308, type=get_attr];
-"309 conv2d_31_zero_point_0" [id=309, type=get_attr];
-"310 compressed_weight_updated_constant21" [id=310, type=get_attr];
-"311 quantize_per_channel_default_28" [id=311, type=quantize_per_channel];
-"312 dequantize_per_channel_default_28" [id=312, type=dequantize_per_channel];
-"313 features_8_block_0_0_weight_bias_0_0" [id=313, type=get_attr];
-"314 conv2d_31" [id=314, type=conv2d];
-"315 hardswish__9_scale_0" [id=315, type=get_attr];
-"316 hardswish__9_zero_point_0" [id=316, type=get_attr];
-"317 hardswish__9" [id=317, type=hardswish_];
-"318 quantize_per_channel_default_29" [id=318, type=quantize_per_channel];
-"319 dequantize_per_channel_default_29" [id=319, type=dequantize_per_channel];
-"320 conv2d_32_scale_0" [id=320, type=get_attr];
-"321 conv2d_32_zero_point_0" [id=321, type=get_attr];
-"322 compressed_weight_updated_constant22" [id=322, type=get_attr];
-"323 quantize_per_channel_default_30" [id=323, type=quantize_per_channel];
-"324 dequantize_per_channel_default_30" [id=324, type=dequantize_per_channel];
-"325 features_8_block_1_0_weight_bias_0_0" [id=325, type=get_attr];
-"326 conv2d_32" [id=326, type=conv2d];
-"327 hardswish__10" [id=327, type=hardswish_];
-"328 quantize_per_tensor_default_28" [id=328, type=quantize_per_tensor];
-"329 dequantize_per_tensor_default_38" [id=329, type=dequantize_per_tensor];
-"330 dequantize_per_tensor_default_37" [id=330, type=dequantize_per_tensor];
-"331 adaptive_avg_pool2d_5" [id=331, type=adaptive_avg_pool2d];
-"332 _param_constant89" [id=332, type=get_attr];
-"333 _param_constant90_0_0" [id=333, type=get_attr];
-"334 conv2d_33" [id=334, type=conv2d];
-"335 relu_5" [id=335, type=relu];
-"336 _param_constant91" [id=336, type=get_attr];
-"337 _param_constant92_0_0" [id=337, type=get_attr];
-"338 conv2d_34" [id=338, type=conv2d];
-"339 hardsigmoid_5" [id=339, type=hardsigmoid];
-"340 quantize_per_tensor_default_29" [id=340, type=quantize_per_tensor];
-"341 dequantize_per_tensor_default_39" [id=341, type=dequantize_per_tensor];
-"342 mul_5" [id=342, type=mul];
-"343 quantize_per_tensor_default_30" [id=343, type=quantize_per_tensor];
-"344 dequantize_per_tensor_default_40" [id=344, type=dequantize_per_tensor];
-"345 conv2d_35_scale_0" [id=345, type=get_attr];
-"346 conv2d_35_zero_point_0" [id=346, type=get_attr];
-"347 compressed_weight_updated_constant23" [id=347, type=get_attr];
-"348 quantize_per_channel_default_31" [id=348, type=quantize_per_channel];
-"349 dequantize_per_channel_default_31" [id=349, type=dequantize_per_channel];
-"350 features_8_block_3_0_weight_bias_0_0" [id=350, type=get_attr];
-"351 conv2d_35" [id=351, type=conv2d];
-"352 quantize_per_tensor_default_31" [id=352, type=quantize_per_tensor];
-"353 dequantize_per_tensor_default_41" [id=353, type=dequantize_per_tensor];
-"354 add__3" [id=354, type=add_];
-"355 quantize_per_tensor_default_32" [id=355, type=quantize_per_tensor];
-"356 dequantize_per_tensor_default_42" [id=356, type=dequantize_per_tensor];
-"357 conv2d_36_scale_0" [id=357, type=get_attr];
-"358 conv2d_36_zero_point_0" [id=358, type=get_attr];
-"359 compressed_weight_updated_constant24" [id=359, type=get_attr];
-"360 quantize_per_channel_default_32" [id=360, type=quantize_per_channel];
-"361 dequantize_per_channel_default_32" [id=361, type=dequantize_per_channel];
-"362 features_9_block_0_0_weight_bias_0_0" [id=362, type=get_attr];
-"363 conv2d_36" [id=363, type=conv2d];
-"364 hardswish__11_scale_0" [id=364, type=get_attr];
-"365 hardswish__11_zero_point_0" [id=365, type=get_attr];
-"366 hardswish__11" [id=366, type=hardswish_];
-"367 quantize_per_channel_default_33" [id=367, type=quantize_per_channel];
-"368 dequantize_per_channel_default_33" [id=368, type=dequantize_per_channel];
-"369 conv2d_37_scale_0" [id=369, type=get_attr];
-"370 conv2d_37_zero_point_0" [id=370, type=get_attr];
-"371 compressed_weight_updated_constant25" [id=371, type=get_attr];
-"372 quantize_per_channel_default_34" [id=372, type=quantize_per_channel];
-"373 dequantize_per_channel_default_34" [id=373, type=dequantize_per_channel];
-"374 features_9_block_1_0_weight_bias_0_0" [id=374, type=get_attr];
-"375 conv2d_37" [id=375, type=conv2d];
-"376 hardswish__12" [id=376, type=hardswish_];
-"377 quantize_per_tensor_default_33" [id=377, type=quantize_per_tensor];
-"378 dequantize_per_tensor_default_44" [id=378, type=dequantize_per_tensor];
-"379 dequantize_per_tensor_default_43" [id=379, type=dequantize_per_tensor];
-"380 adaptive_avg_pool2d_6" [id=380, type=adaptive_avg_pool2d];
-"381 _param_constant102" [id=381, type=get_attr];
-"382 _param_constant103_0_0" [id=382, type=get_attr];
-"383 conv2d_38" [id=383, type=conv2d];
-"384 relu_6" [id=384, type=relu];
-"385 _param_constant104" [id=385, type=get_attr];
-"386 _param_constant105_0_0" [id=386, type=get_attr];
-"387 conv2d_39" [id=387, type=conv2d];
-"388 hardsigmoid_6" [id=388, type=hardsigmoid];
-"389 quantize_per_tensor_default_34" [id=389, type=quantize_per_tensor];
-"390 dequantize_per_tensor_default_45" [id=390, type=dequantize_per_tensor];
-"391 mul_6" [id=391, type=mul];
-"392 quantize_per_tensor_default_35" [id=392, type=quantize_per_tensor];
-"393 dequantize_per_tensor_default_46" [id=393, type=dequantize_per_tensor];
-"394 conv2d_40_scale_0" [id=394, type=get_attr];
-"395 conv2d_40_zero_point_0" [id=395, type=get_attr];
-"396 compressed_weight_updated_constant26" [id=396, type=get_attr];
-"397 quantize_per_channel_default_35" [id=397, type=quantize_per_channel];
-"398 dequantize_per_channel_default_35" [id=398, type=dequantize_per_channel];
-"399 features_9_block_3_0_weight_bias_0_0" [id=399, type=get_attr];
-"400 conv2d_40" [id=400, type=conv2d];
-"401 quantize_per_tensor_default_36" [id=401, type=quantize_per_tensor];
-"402 dequantize_per_tensor_default_48" [id=402, type=dequantize_per_tensor];
-"403 dequantize_per_tensor_default_47" [id=403, type=dequantize_per_tensor];
-"404 conv2d_41_scale_0" [id=404, type=get_attr];
-"405 conv2d_41_zero_point_0" [id=405, type=get_attr];
-"406 compressed_weight_updated_constant27" [id=406, type=get_attr];
-"407 quantize_per_channel_default_36" [id=407, type=quantize_per_channel];
-"408 dequantize_per_channel_default_36" [id=408, type=dequantize_per_channel];
-"409 features_10_block_0_0_weight_bias_0_0" [id=409, type=get_attr];
-"410 conv2d_41" [id=410, type=conv2d];
-"411 hardswish__13_scale_0" [id=411, type=get_attr];
-"412 hardswish__13_zero_point_0" [id=412, type=get_attr];
-"413 hardswish__13" [id=413, type=hardswish_];
-"414 quantize_per_channel_default_37" [id=414, type=quantize_per_channel];
-"415 dequantize_per_channel_default_37" [id=415, type=dequantize_per_channel];
-"416 conv2d_42_scale_0" [id=416, type=get_attr];
-"417 conv2d_42_zero_point_0" [id=417, type=get_attr];
-"418 compressed_weight_updated_constant28" [id=418, type=get_attr];
-"419 quantize_per_channel_default_38" [id=419, type=quantize_per_channel];
-"420 dequantize_per_channel_default_38" [id=420, type=dequantize_per_channel];
-"421 features_10_block_1_0_weight_bias_0_0" [id=421, type=get_attr];
-"422 conv2d_42" [id=422, type=conv2d];
-"423 hardswish__14" [id=423, type=hardswish_];
-"424 quantize_per_tensor_default_37" [id=424, type=quantize_per_tensor];
-"425 dequantize_per_tensor_default_50" [id=425, type=dequantize_per_tensor];
-"426 dequantize_per_tensor_default_49" [id=426, type=dequantize_per_tensor];
-"427 adaptive_avg_pool2d_7" [id=427, type=adaptive_avg_pool2d];
-"428 _param_constant115" [id=428, type=get_attr];
-"429 _param_constant116_0_0" [id=429, type=get_attr];
-"430 conv2d_43" [id=430, type=conv2d];
-"431 relu_7" [id=431, type=relu];
-"432 _param_constant117" [id=432, type=get_attr];
-"433 _param_constant118_0_0" [id=433, type=get_attr];
-"434 conv2d_44" [id=434, type=conv2d];
-"435 hardsigmoid_7" [id=435, type=hardsigmoid];
-"436 quantize_per_tensor_default_38" [id=436, type=quantize_per_tensor];
-"437 dequantize_per_tensor_default_51" [id=437, type=dequantize_per_tensor];
-"438 mul_7" [id=438, type=mul];
-"439 quantize_per_tensor_default_39" [id=439, type=quantize_per_tensor];
-"440 dequantize_per_tensor_default_52" [id=440, type=dequantize_per_tensor];
-"441 conv2d_45_scale_0" [id=441, type=get_attr];
-"442 conv2d_45_zero_point_0" [id=442, type=get_attr];
-"443 compressed_weight_updated_constant29" [id=443, type=get_attr];
-"444 quantize_per_channel_default_39" [id=444, type=quantize_per_channel];
-"445 dequantize_per_channel_default_39" [id=445, type=dequantize_per_channel];
-"446 features_10_block_3_0_weight_bias_0_0" [id=446, type=get_attr];
-"447 conv2d_45" [id=447, type=conv2d];
-"448 quantize_per_tensor_default_40" [id=448, type=quantize_per_tensor];
-"449 dequantize_per_tensor_default_53" [id=449, type=dequantize_per_tensor];
-"450 add__4" [id=450, type=add_];
-"451 quantize_per_tensor_default_41" [id=451, type=quantize_per_tensor];
-"452 dequantize_per_tensor_default_55" [id=452, type=dequantize_per_tensor];
-"453 dequantize_per_tensor_default_54" [id=453, type=dequantize_per_tensor];
-"454 conv2d_46_scale_0" [id=454, type=get_attr];
-"455 conv2d_46_zero_point_0" [id=455, type=get_attr];
-"456 compressed_weight_updated_constant30" [id=456, type=get_attr];
-"457 quantize_per_channel_default_40" [id=457, type=quantize_per_channel];
-"458 dequantize_per_channel_default_40" [id=458, type=dequantize_per_channel];
-"459 features_11_block_0_0_weight_bias_0_0" [id=459, type=get_attr];
-"460 conv2d_46" [id=460, type=conv2d];
-"461 hardswish__15_scale_0" [id=461, type=get_attr];
-"462 hardswish__15_zero_point_0" [id=462, type=get_attr];
-"463 hardswish__15" [id=463, type=hardswish_];
-"464 quantize_per_channel_default_41" [id=464, type=quantize_per_channel];
-"465 dequantize_per_channel_default_41" [id=465, type=dequantize_per_channel];
-"466 conv2d_47_scale_0" [id=466, type=get_attr];
-"467 conv2d_47_zero_point_0" [id=467, type=get_attr];
-"468 compressed_weight_updated_constant31" [id=468, type=get_attr];
-"469 quantize_per_channel_default_42" [id=469, type=quantize_per_channel];
-"470 dequantize_per_channel_default_42" [id=470, type=dequantize_per_channel];
-"471 features_11_block_1_0_weight_bias_0_0" [id=471, type=get_attr];
-"472 conv2d_47" [id=472, type=conv2d];
-"473 hardswish__16" [id=473, type=hardswish_];
-"474 quantize_per_tensor_default_42" [id=474, type=quantize_per_tensor];
-"475 dequantize_per_tensor_default_57" [id=475, type=dequantize_per_tensor];
-"476 dequantize_per_tensor_default_56" [id=476, type=dequantize_per_tensor];
-"477 adaptive_avg_pool2d_8" [id=477, type=adaptive_avg_pool2d];
-"478 _param_constant128" [id=478, type=get_attr];
-"479 _param_constant129_0_0" [id=479, type=get_attr];
-"480 conv2d_48" [id=480, type=conv2d];
-"481 relu_8" [id=481, type=relu];
-"482 _param_constant130" [id=482, type=get_attr];
-"483 _param_constant131_0_0" [id=483, type=get_attr];
+"0 features_1_block_1_fc1_weight" [id=0, type=get_attr];
+"1 features_1_block_1_fc2_weight" [id=1, type=get_attr];
+"2 features_4_block_2_fc1_weight" [id=2, type=get_attr];
+"3 features_4_block_2_fc2_weight" [id=3, type=get_attr];
+"4 features_5_block_2_fc1_weight" [id=4, type=get_attr];
+"5 features_5_block_2_fc2_weight" [id=5, type=get_attr];
+"6 features_6_block_2_fc1_weight" [id=6, type=get_attr];
+"7 features_6_block_2_fc2_weight" [id=7, type=get_attr];
+"8 features_7_block_2_fc1_weight" [id=8, type=get_attr];
+"9 features_7_block_2_fc2_weight" [id=9, type=get_attr];
+"10 features_8_block_2_fc1_weight" [id=10, type=get_attr];
+"11 features_8_block_2_fc2_weight" [id=11, type=get_attr];
+"12 features_9_block_2_fc1_weight" [id=12, type=get_attr];
+"13 features_9_block_2_fc2_weight" [id=13, type=get_attr];
+"14 features_10_block_2_fc1_weight" [id=14, type=get_attr];
+"15 features_10_block_2_fc2_weight" [id=15, type=get_attr];
+"16 features_11_block_2_fc1_weight" [id=16, type=get_attr];
+"17 features_11_block_2_fc2_weight" [id=17, type=get_attr];
+"18 x" [id=18, type=input];
+"19 quantize_per_tensor_default" [id=19, type=quantize_per_tensor];
+"20 dequantize_per_tensor_default" [id=20, type=dequantize_per_tensor];
+"21 conv2d_scale_0" [id=21, type=get_attr];
+"22 conv2d_zero_point_0" [id=22, type=get_attr];
+"23 compressed_weight_updated_constant0" [id=23, type=get_attr];
+"24 quantize_per_channel_default" [id=24, type=quantize_per_channel];
+"25 dequantize_per_channel_default" [id=25, type=dequantize_per_channel];
+"26 features_0_0_weight_bias_0_0" [id=26, type=get_attr];
+"27 conv2d" [id=27, type=conv2d];
+"28 hardswish__scale_0" [id=28, type=get_attr];
+"29 hardswish__zero_point_0" [id=29, type=get_attr];
+"30 hardswish_" [id=30, type=hardswish_];
+"31 quantize_per_channel_default_1" [id=31, type=quantize_per_channel];
+"32 dequantize_per_channel_default_1" [id=32, type=dequantize_per_channel];
+"33 conv2d_1_scale_0" [id=33, type=get_attr];
+"34 conv2d_1_zero_point_0" [id=34, type=get_attr];
+"35 compressed_weight_updated_constant1" [id=35, type=get_attr];
+"36 quantize_per_channel_default_2" [id=36, type=quantize_per_channel];
+"37 dequantize_per_channel_default_2" [id=37, type=dequantize_per_channel];
+"38 features_1_block_0_0_weight_bias_0_0" [id=38, type=get_attr];
+"39 conv2d_1" [id=39, type=conv2d];
+"40 relu_" [id=40, type=relu_];
+"41 quantize_per_tensor_default_1" [id=41, type=quantize_per_tensor];
+"42 dequantize_per_tensor_default_2" [id=42, type=dequantize_per_tensor];
+"43 dequantize_per_tensor_default_1" [id=43, type=dequantize_per_tensor];
+"44 adaptive_avg_pool2d" [id=44, type=adaptive_avg_pool2d];
+"45 features_1_block_1_fc1_bias_0_0" [id=45, type=get_attr];
+"46 conv2d_2" [id=46, type=conv2d];
+"47 relu" [id=47, type=relu];
+"48 features_1_block_1_fc2_bias_0_0" [id=48, type=get_attr];
+"49 conv2d_3" [id=49, type=conv2d];
+"50 hardsigmoid" [id=50, type=hardsigmoid];
+"51 quantize_per_tensor_default_2" [id=51, type=quantize_per_tensor];
+"52 dequantize_per_tensor_default_3" [id=52, type=dequantize_per_tensor];
+"53 mul" [id=53, type=mul];
+"54 quantize_per_tensor_default_3" [id=54, type=quantize_per_tensor];
+"55 dequantize_per_tensor_default_4" [id=55, type=dequantize_per_tensor];
+"56 conv2d_4_scale_0" [id=56, type=get_attr];
+"57 conv2d_4_zero_point_0" [id=57, type=get_attr];
+"58 compressed_weight_updated_constant2" [id=58, type=get_attr];
+"59 quantize_per_channel_default_3" [id=59, type=quantize_per_channel];
+"60 dequantize_per_channel_default_3" [id=60, type=dequantize_per_channel];
+"61 features_1_block_2_0_weight_bias_0_0" [id=61, type=get_attr];
+"62 conv2d_4" [id=62, type=conv2d];
+"63 quantize_per_tensor_default_4" [id=63, type=quantize_per_tensor];
+"64 dequantize_per_tensor_default_5" [id=64, type=dequantize_per_tensor];
+"65 conv2d_5_scale_0" [id=65, type=get_attr];
+"66 conv2d_5_zero_point_0" [id=66, type=get_attr];
+"67 compressed_weight_updated_constant3" [id=67, type=get_attr];
+"68 quantize_per_channel_default_4" [id=68, type=quantize_per_channel];
+"69 dequantize_per_channel_default_4" [id=69, type=dequantize_per_channel];
+"70 features_2_block_0_0_weight_bias_0_0" [id=70, type=get_attr];
+"71 conv2d_5" [id=71, type=conv2d];
+"72 relu__1_scale_0" [id=72, type=get_attr];
+"73 relu__1_zero_point_0" [id=73, type=get_attr];
+"74 relu__1" [id=74, type=relu_];
+"75 quantize_per_channel_default_5" [id=75, type=quantize_per_channel];
+"76 dequantize_per_channel_default_5" [id=76, type=dequantize_per_channel];
+"77 conv2d_6_scale_0" [id=77, type=get_attr];
+"78 conv2d_6_zero_point_0" [id=78, type=get_attr];
+"79 compressed_weight_updated_constant4" [id=79, type=get_attr];
+"80 quantize_per_channel_default_6" [id=80, type=quantize_per_channel];
+"81 dequantize_per_channel_default_6" [id=81, type=dequantize_per_channel];
+"82 features_2_block_1_0_weight_bias_0_0" [id=82, type=get_attr];
+"83 conv2d_6" [id=83, type=conv2d];
+"84 relu__2" [id=84, type=relu_];
+"85 quantize_per_tensor_default_5" [id=85, type=quantize_per_tensor];
+"86 dequantize_per_tensor_default_6" [id=86, type=dequantize_per_tensor];
+"87 conv2d_7_scale_0" [id=87, type=get_attr];
+"88 conv2d_7_zero_point_0" [id=88, type=get_attr];
+"89 compressed_weight_updated_constant5" [id=89, type=get_attr];
+"90 quantize_per_channel_default_7" [id=90, type=quantize_per_channel];
+"91 dequantize_per_channel_default_7" [id=91, type=dequantize_per_channel];
+"92 features_2_block_2_0_weight_bias_0_0" [id=92, type=get_attr];
+"93 conv2d_7" [id=93, type=conv2d];
+"94 quantize_per_tensor_default_6" [id=94, type=quantize_per_tensor];
+"95 dequantize_per_tensor_default_8" [id=95, type=dequantize_per_tensor];
+"96 dequantize_per_tensor_default_7" [id=96, type=dequantize_per_tensor];
+"97 conv2d_8_scale_0" [id=97, type=get_attr];
+"98 conv2d_8_zero_point_0" [id=98, type=get_attr];
+"99 compressed_weight_updated_constant6" [id=99, type=get_attr];
+"100 quantize_per_channel_default_8" [id=100, type=quantize_per_channel];
+"101 dequantize_per_channel_default_8" [id=101, type=dequantize_per_channel];
+"102 features_3_block_0_0_weight_bias_0_0" [id=102, type=get_attr];
+"103 conv2d_8" [id=103, type=conv2d];
+"104 relu__3_scale_0" [id=104, type=get_attr];
+"105 relu__3_zero_point_0" [id=105, type=get_attr];
+"106 relu__3" [id=106, type=relu_];
+"107 quantize_per_channel_default_9" [id=107, type=quantize_per_channel];
+"108 dequantize_per_channel_default_9" [id=108, type=dequantize_per_channel];
+"109 conv2d_9_scale_0" [id=109, type=get_attr];
+"110 conv2d_9_zero_point_0" [id=110, type=get_attr];
+"111 compressed_weight_updated_constant7" [id=111, type=get_attr];
+"112 quantize_per_channel_default_10" [id=112, type=quantize_per_channel];
+"113 dequantize_per_channel_default_10" [id=113, type=dequantize_per_channel];
+"114 features_3_block_1_0_weight_bias_0_0" [id=114, type=get_attr];
+"115 conv2d_9" [id=115, type=conv2d];
+"116 relu__4" [id=116, type=relu_];
+"117 quantize_per_tensor_default_7" [id=117, type=quantize_per_tensor];
+"118 dequantize_per_tensor_default_9" [id=118, type=dequantize_per_tensor];
+"119 conv2d_10_scale_0" [id=119, type=get_attr];
+"120 conv2d_10_zero_point_0" [id=120, type=get_attr];
+"121 compressed_weight_updated_constant8" [id=121, type=get_attr];
+"122 quantize_per_channel_default_11" [id=122, type=quantize_per_channel];
+"123 dequantize_per_channel_default_11" [id=123, type=dequantize_per_channel];
+"124 features_3_block_2_0_weight_bias_0_0" [id=124, type=get_attr];
+"125 conv2d_10" [id=125, type=conv2d];
+"126 quantize_per_tensor_default_8" [id=126, type=quantize_per_tensor];
+"127 dequantize_per_tensor_default_10" [id=127, type=dequantize_per_tensor];
+"128 add_" [id=128, type=add_];
+"129 quantize_per_tensor_default_9" [id=129, type=quantize_per_tensor];
+"130 dequantize_per_tensor_default_11" [id=130, type=dequantize_per_tensor];
+"131 conv2d_11_scale_0" [id=131, type=get_attr];
+"132 conv2d_11_zero_point_0" [id=132, type=get_attr];
+"133 compressed_weight_updated_constant9" [id=133, type=get_attr];
+"134 quantize_per_channel_default_12" [id=134, type=quantize_per_channel];
+"135 dequantize_per_channel_default_12" [id=135, type=dequantize_per_channel];
+"136 features_4_block_0_0_weight_bias_0_0" [id=136, type=get_attr];
+"137 conv2d_11" [id=137, type=conv2d];
+"138 hardswish__1_scale_0" [id=138, type=get_attr];
+"139 hardswish__1_zero_point_0" [id=139, type=get_attr];
+"140 hardswish__1" [id=140, type=hardswish_];
+"141 quantize_per_channel_default_13" [id=141, type=quantize_per_channel];
+"142 dequantize_per_channel_default_13" [id=142, type=dequantize_per_channel];
+"143 conv2d_12_scale_0" [id=143, type=get_attr];
+"144 conv2d_12_zero_point_0" [id=144, type=get_attr];
+"145 compressed_weight_updated_constant10" [id=145, type=get_attr];
+"146 quantize_per_channel_default_14" [id=146, type=quantize_per_channel];
+"147 dequantize_per_channel_default_14" [id=147, type=dequantize_per_channel];
+"148 features_4_block_1_0_weight_bias_0_0" [id=148, type=get_attr];
+"149 conv2d_12" [id=149, type=conv2d];
+"150 hardswish__2" [id=150, type=hardswish_];
+"151 quantize_per_tensor_default_10" [id=151, type=quantize_per_tensor];
+"152 dequantize_per_tensor_default_13" [id=152, type=dequantize_per_tensor];
+"153 dequantize_per_tensor_default_12" [id=153, type=dequantize_per_tensor];
+"154 adaptive_avg_pool2d_1" [id=154, type=adaptive_avg_pool2d];
+"155 features_4_block_2_fc1_bias_0_0" [id=155, type=get_attr];
+"156 conv2d_13" [id=156, type=conv2d];
+"157 relu_1" [id=157, type=relu];
+"158 features_4_block_2_fc2_bias_0_0" [id=158, type=get_attr];
+"159 conv2d_14" [id=159, type=conv2d];
+"160 hardsigmoid_1" [id=160, type=hardsigmoid];
+"161 quantize_per_tensor_default_11" [id=161, type=quantize_per_tensor];
+"162 dequantize_per_tensor_default_14" [id=162, type=dequantize_per_tensor];
+"163 mul_1" [id=163, type=mul];
+"164 quantize_per_tensor_default_12" [id=164, type=quantize_per_tensor];
+"165 dequantize_per_tensor_default_15" [id=165, type=dequantize_per_tensor];
+"166 conv2d_15_scale_0" [id=166, type=get_attr];
+"167 conv2d_15_zero_point_0" [id=167, type=get_attr];
+"168 compressed_weight_updated_constant11" [id=168, type=get_attr];
+"169 quantize_per_channel_default_15" [id=169, type=quantize_per_channel];
+"170 dequantize_per_channel_default_15" [id=170, type=dequantize_per_channel];
+"171 features_4_block_3_0_weight_bias_0_0" [id=171, type=get_attr];
+"172 conv2d_15" [id=172, type=conv2d];
+"173 quantize_per_tensor_default_13" [id=173, type=quantize_per_tensor];
+"174 dequantize_per_tensor_default_17" [id=174, type=dequantize_per_tensor];
+"175 dequantize_per_tensor_default_16" [id=175, type=dequantize_per_tensor];
+"176 conv2d_16_scale_0" [id=176, type=get_attr];
+"177 conv2d_16_zero_point_0" [id=177, type=get_attr];
+"178 compressed_weight_updated_constant12" [id=178, type=get_attr];
+"179 quantize_per_channel_default_16" [id=179, type=quantize_per_channel];
+"180 dequantize_per_channel_default_16" [id=180, type=dequantize_per_channel];
+"181 features_5_block_0_0_weight_bias_0_0" [id=181, type=get_attr];
+"182 conv2d_16" [id=182, type=conv2d];
+"183 hardswish__3_scale_0" [id=183, type=get_attr];
+"184 hardswish__3_zero_point_0" [id=184, type=get_attr];
+"185 hardswish__3" [id=185, type=hardswish_];
+"186 quantize_per_channel_default_17" [id=186, type=quantize_per_channel];
+"187 dequantize_per_channel_default_17" [id=187, type=dequantize_per_channel];
+"188 conv2d_17_scale_0" [id=188, type=get_attr];
+"189 conv2d_17_zero_point_0" [id=189, type=get_attr];
+"190 compressed_weight_updated_constant13" [id=190, type=get_attr];
+"191 quantize_per_channel_default_18" [id=191, type=quantize_per_channel];
+"192 dequantize_per_channel_default_18" [id=192, type=dequantize_per_channel];
+"193 features_5_block_1_0_weight_bias_0_0" [id=193, type=get_attr];
+"194 conv2d_17" [id=194, type=conv2d];
+"195 hardswish__4" [id=195, type=hardswish_];
+"196 quantize_per_tensor_default_14" [id=196, type=quantize_per_tensor];
+"197 dequantize_per_tensor_default_19" [id=197, type=dequantize_per_tensor];
+"198 dequantize_per_tensor_default_18" [id=198, type=dequantize_per_tensor];
+"199 adaptive_avg_pool2d_2" [id=199, type=adaptive_avg_pool2d];
+"200 features_5_block_2_fc1_bias_0_0" [id=200, type=get_attr];
+"201 conv2d_18" [id=201, type=conv2d];
+"202 relu_2" [id=202, type=relu];
+"203 features_5_block_2_fc2_bias_0_0" [id=203, type=get_attr];
+"204 conv2d_19" [id=204, type=conv2d];
+"205 hardsigmoid_2" [id=205, type=hardsigmoid];
+"206 quantize_per_tensor_default_15" [id=206, type=quantize_per_tensor];
+"207 dequantize_per_tensor_default_20" [id=207, type=dequantize_per_tensor];
+"208 mul_2" [id=208, type=mul];
+"209 quantize_per_tensor_default_16" [id=209, type=quantize_per_tensor];
+"210 dequantize_per_tensor_default_21" [id=210, type=dequantize_per_tensor];
+"211 conv2d_20_scale_0" [id=211, type=get_attr];
+"212 conv2d_20_zero_point_0" [id=212, type=get_attr];
+"213 compressed_weight_updated_constant14" [id=213, type=get_attr];
+"214 quantize_per_channel_default_19" [id=214, type=quantize_per_channel];
+"215 dequantize_per_channel_default_19" [id=215, type=dequantize_per_channel];
+"216 features_5_block_3_0_weight_bias_0_0" [id=216, type=get_attr];
+"217 conv2d_20" [id=217, type=conv2d];
+"218 quantize_per_tensor_default_17" [id=218, type=quantize_per_tensor];
+"219 dequantize_per_tensor_default_22" [id=219, type=dequantize_per_tensor];
+"220 add__1" [id=220, type=add_];
+"221 quantize_per_tensor_default_18" [id=221, type=quantize_per_tensor];
+"222 dequantize_per_tensor_default_24" [id=222, type=dequantize_per_tensor];
+"223 dequantize_per_tensor_default_23" [id=223, type=dequantize_per_tensor];
+"224 conv2d_21_scale_0" [id=224, type=get_attr];
+"225 conv2d_21_zero_point_0" [id=225, type=get_attr];
+"226 compressed_weight_updated_constant15" [id=226, type=get_attr];
+"227 quantize_per_channel_default_20" [id=227, type=quantize_per_channel];
+"228 dequantize_per_channel_default_20" [id=228, type=dequantize_per_channel];
+"229 features_6_block_0_0_weight_bias_0_0" [id=229, type=get_attr];
+"230 conv2d_21" [id=230, type=conv2d];
+"231 hardswish__5_scale_0" [id=231, type=get_attr];
+"232 hardswish__5_zero_point_0" [id=232, type=get_attr];
+"233 hardswish__5" [id=233, type=hardswish_];
+"234 quantize_per_channel_default_21" [id=234, type=quantize_per_channel];
+"235 dequantize_per_channel_default_21" [id=235, type=dequantize_per_channel];
+"236 conv2d_22_scale_0" [id=236, type=get_attr];
+"237 conv2d_22_zero_point_0" [id=237, type=get_attr];
+"238 compressed_weight_updated_constant16" [id=238, type=get_attr];
+"239 quantize_per_channel_default_22" [id=239, type=quantize_per_channel];
+"240 dequantize_per_channel_default_22" [id=240, type=dequantize_per_channel];
+"241 features_6_block_1_0_weight_bias_0_0" [id=241, type=get_attr];
+"242 conv2d_22" [id=242, type=conv2d];
+"243 hardswish__6" [id=243, type=hardswish_];
+"244 quantize_per_tensor_default_19" [id=244, type=quantize_per_tensor];
+"245 dequantize_per_tensor_default_26" [id=245, type=dequantize_per_tensor];
+"246 dequantize_per_tensor_default_25" [id=246, type=dequantize_per_tensor];
+"247 adaptive_avg_pool2d_3" [id=247, type=adaptive_avg_pool2d];
+"248 features_6_block_2_fc1_bias_0_0" [id=248, type=get_attr];
+"249 conv2d_23" [id=249, type=conv2d];
+"250 relu_3" [id=250, type=relu];
+"251 features_6_block_2_fc2_bias_0_0" [id=251, type=get_attr];
+"252 conv2d_24" [id=252, type=conv2d];
+"253 hardsigmoid_3" [id=253, type=hardsigmoid];
+"254 quantize_per_tensor_default_20" [id=254, type=quantize_per_tensor];
+"255 dequantize_per_tensor_default_27" [id=255, type=dequantize_per_tensor];
+"256 mul_3" [id=256, type=mul];
+"257 quantize_per_tensor_default_21" [id=257, type=quantize_per_tensor];
+"258 dequantize_per_tensor_default_28" [id=258, type=dequantize_per_tensor];
+"259 conv2d_25_scale_0" [id=259, type=get_attr];
+"260 conv2d_25_zero_point_0" [id=260, type=get_attr];
+"261 compressed_weight_updated_constant17" [id=261, type=get_attr];
+"262 quantize_per_channel_default_23" [id=262, type=quantize_per_channel];
+"263 dequantize_per_channel_default_23" [id=263, type=dequantize_per_channel];
+"264 features_6_block_3_0_weight_bias_0_0" [id=264, type=get_attr];
+"265 conv2d_25" [id=265, type=conv2d];
+"266 quantize_per_tensor_default_22" [id=266, type=quantize_per_tensor];
+"267 dequantize_per_tensor_default_29" [id=267, type=dequantize_per_tensor];
+"268 add__2" [id=268, type=add_];
+"269 quantize_per_tensor_default_23" [id=269, type=quantize_per_tensor];
+"270 dequantize_per_tensor_default_30" [id=270, type=dequantize_per_tensor];
+"271 conv2d_26_scale_0" [id=271, type=get_attr];
+"272 conv2d_26_zero_point_0" [id=272, type=get_attr];
+"273 compressed_weight_updated_constant18" [id=273, type=get_attr];
+"274 quantize_per_channel_default_24" [id=274, type=quantize_per_channel];
+"275 dequantize_per_channel_default_24" [id=275, type=dequantize_per_channel];
+"276 features_7_block_0_0_weight_bias_0_0" [id=276, type=get_attr];
+"277 conv2d_26" [id=277, type=conv2d];
+"278 hardswish__7_scale_0" [id=278, type=get_attr];
+"279 hardswish__7_zero_point_0" [id=279, type=get_attr];
+"280 hardswish__7" [id=280, type=hardswish_];
+"281 quantize_per_channel_default_25" [id=281, type=quantize_per_channel];
+"282 dequantize_per_channel_default_25" [id=282, type=dequantize_per_channel];
+"283 conv2d_27_scale_0" [id=283, type=get_attr];
+"284 conv2d_27_zero_point_0" [id=284, type=get_attr];
+"285 compressed_weight_updated_constant19" [id=285, type=get_attr];
+"286 quantize_per_channel_default_26" [id=286, type=quantize_per_channel];
+"287 dequantize_per_channel_default_26" [id=287, type=dequantize_per_channel];
+"288 features_7_block_1_0_weight_bias_0_0" [id=288, type=get_attr];
+"289 conv2d_27" [id=289, type=conv2d];
+"290 hardswish__8" [id=290, type=hardswish_];
+"291 quantize_per_tensor_default_24" [id=291, type=quantize_per_tensor];
+"292 dequantize_per_tensor_default_32" [id=292, type=dequantize_per_tensor];
+"293 dequantize_per_tensor_default_31" [id=293, type=dequantize_per_tensor];
+"294 adaptive_avg_pool2d_4" [id=294, type=adaptive_avg_pool2d];
+"295 features_7_block_2_fc1_bias_0_0" [id=295, type=get_attr];
+"296 conv2d_28" [id=296, type=conv2d];
+"297 relu_4" [id=297, type=relu];
+"298 features_7_block_2_fc2_bias_0_0" [id=298, type=get_attr];
+"299 conv2d_29" [id=299, type=conv2d];
+"300 hardsigmoid_4" [id=300, type=hardsigmoid];
+"301 quantize_per_tensor_default_25" [id=301, type=quantize_per_tensor];
+"302 dequantize_per_tensor_default_33" [id=302, type=dequantize_per_tensor];
+"303 mul_4" [id=303, type=mul];
+"304 quantize_per_tensor_default_26" [id=304, type=quantize_per_tensor];
+"305 dequantize_per_tensor_default_34" [id=305, type=dequantize_per_tensor];
+"306 conv2d_30_scale_0" [id=306, type=get_attr];
+"307 conv2d_30_zero_point_0" [id=307, type=get_attr];
+"308 compressed_weight_updated_constant20" [id=308, type=get_attr];
+"309 quantize_per_channel_default_27" [id=309, type=quantize_per_channel];
+"310 dequantize_per_channel_default_27" [id=310, type=dequantize_per_channel];
+"311 features_7_block_3_0_weight_bias_0_0" [id=311, type=get_attr];
+"312 conv2d_30" [id=312, type=conv2d];
+"313 quantize_per_tensor_default_27" [id=313, type=quantize_per_tensor];
+"314 dequantize_per_tensor_default_36" [id=314, type=dequantize_per_tensor];
+"315 dequantize_per_tensor_default_35" [id=315, type=dequantize_per_tensor];
+"316 conv2d_31_scale_0" [id=316, type=get_attr];
+"317 conv2d_31_zero_point_0" [id=317, type=get_attr];
+"318 compressed_weight_updated_constant21" [id=318, type=get_attr];
+"319 quantize_per_channel_default_28" [id=319, type=quantize_per_channel];
+"320 dequantize_per_channel_default_28" [id=320, type=dequantize_per_channel];
+"321 features_8_block_0_0_weight_bias_0_0" [id=321, type=get_attr];
+"322 conv2d_31" [id=322, type=conv2d];
+"323 hardswish__9_scale_0" [id=323, type=get_attr];
+"324 hardswish__9_zero_point_0" [id=324, type=get_attr];
+"325 hardswish__9" [id=325, type=hardswish_];
+"326 quantize_per_channel_default_29" [id=326, type=quantize_per_channel];
+"327 dequantize_per_channel_default_29" [id=327, type=dequantize_per_channel];
+"328 conv2d_32_scale_0" [id=328, type=get_attr];
+"329 conv2d_32_zero_point_0" [id=329, type=get_attr];
+"330 compressed_weight_updated_constant22" [id=330, type=get_attr];
+"331 quantize_per_channel_default_30" [id=331, type=quantize_per_channel];
+"332 dequantize_per_channel_default_30" [id=332, type=dequantize_per_channel];
+"333 features_8_block_1_0_weight_bias_0_0" [id=333, type=get_attr];
+"334 conv2d_32" [id=334, type=conv2d];
+"335 hardswish__10" [id=335, type=hardswish_];
+"336 quantize_per_tensor_default_28" [id=336, type=quantize_per_tensor];
+"337 dequantize_per_tensor_default_38" [id=337, type=dequantize_per_tensor];
+"338 dequantize_per_tensor_default_37" [id=338, type=dequantize_per_tensor];
+"339 adaptive_avg_pool2d_5" [id=339, type=adaptive_avg_pool2d];
+"340 features_8_block_2_fc1_bias_0_0" [id=340, type=get_attr];
+"341 conv2d_33" [id=341, type=conv2d];
+"342 relu_5" [id=342, type=relu];
+"343 features_8_block_2_fc2_bias_0_0" [id=343, type=get_attr];
+"344 conv2d_34" [id=344, type=conv2d];
+"345 hardsigmoid_5" [id=345, type=hardsigmoid];
+"346 quantize_per_tensor_default_29" [id=346, type=quantize_per_tensor];
+"347 dequantize_per_tensor_default_39" [id=347, type=dequantize_per_tensor];
+"348 mul_5" [id=348, type=mul];
+"349 quantize_per_tensor_default_30" [id=349, type=quantize_per_tensor];
+"350 dequantize_per_tensor_default_40" [id=350, type=dequantize_per_tensor];
+"351 conv2d_35_scale_0" [id=351, type=get_attr];
+"352 conv2d_35_zero_point_0" [id=352, type=get_attr];
+"353 compressed_weight_updated_constant23" [id=353, type=get_attr];
+"354 quantize_per_channel_default_31" [id=354, type=quantize_per_channel];
+"355 dequantize_per_channel_default_31" [id=355, type=dequantize_per_channel];
+"356 features_8_block_3_0_weight_bias_0_0" [id=356, type=get_attr];
+"357 conv2d_35" [id=357, type=conv2d];
+"358 quantize_per_tensor_default_31" [id=358, type=quantize_per_tensor];
+"359 dequantize_per_tensor_default_41" [id=359, type=dequantize_per_tensor];
+"360 add__3" [id=360, type=add_];
+"361 quantize_per_tensor_default_32" [id=361, type=quantize_per_tensor];
+"362 dequantize_per_tensor_default_42" [id=362, type=dequantize_per_tensor];
+"363 conv2d_36_scale_0" [id=363, type=get_attr];
+"364 conv2d_36_zero_point_0" [id=364, type=get_attr];
+"365 compressed_weight_updated_constant24" [id=365, type=get_attr];
+"366 quantize_per_channel_default_32" [id=366, type=quantize_per_channel];
+"367 dequantize_per_channel_default_32" [id=367, type=dequantize_per_channel];
+"368 features_9_block_0_0_weight_bias_0_0" [id=368, type=get_attr];
+"369 conv2d_36" [id=369, type=conv2d];
+"370 hardswish__11_scale_0" [id=370, type=get_attr];
+"371 hardswish__11_zero_point_0" [id=371, type=get_attr];
+"372 hardswish__11" [id=372, type=hardswish_];
+"373 quantize_per_channel_default_33" [id=373, type=quantize_per_channel];
+"374 dequantize_per_channel_default_33" [id=374, type=dequantize_per_channel];
+"375 conv2d_37_scale_0" [id=375, type=get_attr];
+"376 conv2d_37_zero_point_0" [id=376, type=get_attr];
+"377 compressed_weight_updated_constant25" [id=377, type=get_attr];
+"378 quantize_per_channel_default_34" [id=378, type=quantize_per_channel];
+"379 dequantize_per_channel_default_34" [id=379, type=dequantize_per_channel];
+"380 features_9_block_1_0_weight_bias_0_0" [id=380, type=get_attr];
+"381 conv2d_37" [id=381, type=conv2d];
+"382 hardswish__12" [id=382, type=hardswish_];
+"383 quantize_per_tensor_default_33" [id=383, type=quantize_per_tensor];
+"384 dequantize_per_tensor_default_44" [id=384, type=dequantize_per_tensor];
+"385 dequantize_per_tensor_default_43" [id=385, type=dequantize_per_tensor];
+"386 adaptive_avg_pool2d_6" [id=386, type=adaptive_avg_pool2d];
+"387 features_9_block_2_fc1_bias_0_0" [id=387, type=get_attr];
+"388 conv2d_38" [id=388, type=conv2d];
+"389 relu_6" [id=389, type=relu];
+"390 features_9_block_2_fc2_bias_0_0" [id=390, type=get_attr];
+"391 conv2d_39" [id=391, type=conv2d];
+"392 hardsigmoid_6" [id=392, type=hardsigmoid];
+"393 quantize_per_tensor_default_34" [id=393, type=quantize_per_tensor];
+"394 dequantize_per_tensor_default_45" [id=394, type=dequantize_per_tensor];
+"395 mul_6" [id=395, type=mul];
+"396 quantize_per_tensor_default_35" [id=396, type=quantize_per_tensor];
+"397 dequantize_per_tensor_default_46" [id=397, type=dequantize_per_tensor];
+"398 conv2d_40_scale_0" [id=398, type=get_attr];
+"399 conv2d_40_zero_point_0" [id=399, type=get_attr];
+"400 compressed_weight_updated_constant26" [id=400, type=get_attr];
+"401 quantize_per_channel_default_35" [id=401, type=quantize_per_channel];
+"402 dequantize_per_channel_default_35" [id=402, type=dequantize_per_channel];
+"403 features_9_block_3_0_weight_bias_0_0" [id=403, type=get_attr];
+"404 conv2d_40" [id=404, type=conv2d];
+"405 quantize_per_tensor_default_36" [id=405, type=quantize_per_tensor];
+"406 dequantize_per_tensor_default_48" [id=406, type=dequantize_per_tensor];
+"407 dequantize_per_tensor_default_47" [id=407, type=dequantize_per_tensor];
+"408 conv2d_41_scale_0" [id=408, type=get_attr];
+"409 conv2d_41_zero_point_0" [id=409, type=get_attr];
+"410 compressed_weight_updated_constant27" [id=410, type=get_attr];
+"411 quantize_per_channel_default_36" [id=411, type=quantize_per_channel];
+"412 dequantize_per_channel_default_36" [id=412, type=dequantize_per_channel];
+"413 features_10_block_0_0_weight_bias_0_0" [id=413, type=get_attr];
+"414 conv2d_41" [id=414, type=conv2d];
+"415 hardswish__13_scale_0" [id=415, type=get_attr];
+"416 hardswish__13_zero_point_0" [id=416, type=get_attr];
+"417 hardswish__13" [id=417, type=hardswish_];
+"418 quantize_per_channel_default_37" [id=418, type=quantize_per_channel];
+"419 dequantize_per_channel_default_37" [id=419, type=dequantize_per_channel];
+"420 conv2d_42_scale_0" [id=420, type=get_attr];
+"421 conv2d_42_zero_point_0" [id=421, type=get_attr];
+"422 compressed_weight_updated_constant28" [id=422, type=get_attr];
+"423 quantize_per_channel_default_38" [id=423, type=quantize_per_channel];
+"424 dequantize_per_channel_default_38" [id=424, type=dequantize_per_channel];
+"425 features_10_block_1_0_weight_bias_0_0" [id=425, type=get_attr];
+"426 conv2d_42" [id=426, type=conv2d];
+"427 hardswish__14" [id=427, type=hardswish_];
+"428 quantize_per_tensor_default_37" [id=428, type=quantize_per_tensor];
+"429 dequantize_per_tensor_default_50" [id=429, type=dequantize_per_tensor];
+"430 dequantize_per_tensor_default_49" [id=430, type=dequantize_per_tensor];
+"431 adaptive_avg_pool2d_7" [id=431, type=adaptive_avg_pool2d];
+"432 features_10_block_2_fc1_bias_0_0" [id=432, type=get_attr];
+"433 conv2d_43" [id=433, type=conv2d];
+"434 relu_7" [id=434, type=relu];
+"435 features_10_block_2_fc2_bias_0_0" [id=435, type=get_attr];
+"436 conv2d_44" [id=436, type=conv2d];
+"437 hardsigmoid_7" [id=437, type=hardsigmoid];
+"438 quantize_per_tensor_default_38" [id=438, type=quantize_per_tensor];
+"439 dequantize_per_tensor_default_51" [id=439, type=dequantize_per_tensor];
+"440 mul_7" [id=440, type=mul];
+"441 quantize_per_tensor_default_39" [id=441, type=quantize_per_tensor];
+"442 dequantize_per_tensor_default_52" [id=442, type=dequantize_per_tensor];
+"443 conv2d_45_scale_0" [id=443, type=get_attr];
+"444 conv2d_45_zero_point_0" [id=444, type=get_attr];
+"445 compressed_weight_updated_constant29" [id=445, type=get_attr];
+"446 quantize_per_channel_default_39" [id=446, type=quantize_per_channel];
+"447 dequantize_per_channel_default_39" [id=447, type=dequantize_per_channel];
+"448 features_10_block_3_0_weight_bias_0_0" [id=448, type=get_attr];
+"449 conv2d_45" [id=449, type=conv2d];
+"450 quantize_per_tensor_default_40" [id=450, type=quantize_per_tensor];
+"451 dequantize_per_tensor_default_53" [id=451, type=dequantize_per_tensor];
+"452 add__4" [id=452, type=add_];
+"453 quantize_per_tensor_default_41" [id=453, type=quantize_per_tensor];
+"454 dequantize_per_tensor_default_55" [id=454, type=dequantize_per_tensor];
+"455 dequantize_per_tensor_default_54" [id=455, type=dequantize_per_tensor];
+"456 conv2d_46_scale_0" [id=456, type=get_attr];
+"457 conv2d_46_zero_point_0" [id=457, type=get_attr];
+"458 compressed_weight_updated_constant30" [id=458, type=get_attr];
+"459 quantize_per_channel_default_40" [id=459, type=quantize_per_channel];
+"460 dequantize_per_channel_default_40" [id=460, type=dequantize_per_channel];
+"461 features_11_block_0_0_weight_bias_0_0" [id=461, type=get_attr];
+"462 conv2d_46" [id=462, type=conv2d];
+"463 hardswish__15_scale_0" [id=463, type=get_attr];
+"464 hardswish__15_zero_point_0" [id=464, type=get_attr];
+"465 hardswish__15" [id=465, type=hardswish_];
+"466 quantize_per_channel_default_41" [id=466, type=quantize_per_channel];
+"467 dequantize_per_channel_default_41" [id=467, type=dequantize_per_channel];
+"468 conv2d_47_scale_0" [id=468, type=get_attr];
+"469 conv2d_47_zero_point_0" [id=469, type=get_attr];
+"470 compressed_weight_updated_constant31" [id=470, type=get_attr];
+"471 quantize_per_channel_default_42" [id=471, type=quantize_per_channel];
+"472 dequantize_per_channel_default_42" [id=472, type=dequantize_per_channel];
+"473 features_11_block_1_0_weight_bias_0_0" [id=473, type=get_attr];
+"474 conv2d_47" [id=474, type=conv2d];
+"475 hardswish__16" [id=475, type=hardswish_];
+"476 quantize_per_tensor_default_42" [id=476, type=quantize_per_tensor];
+"477 dequantize_per_tensor_default_57" [id=477, type=dequantize_per_tensor];
+"478 dequantize_per_tensor_default_56" [id=478, type=dequantize_per_tensor];
+"479 adaptive_avg_pool2d_8" [id=479, type=adaptive_avg_pool2d];
+"480 features_11_block_2_fc1_bias_0_0" [id=480, type=get_attr];
+"481 conv2d_48" [id=481, type=conv2d];
+"482 relu_8" [id=482, type=relu];
+"483 features_11_block_2_fc2_bias_0_0" [id=483, type=get_attr];
 "484 conv2d_49" [id=484, type=conv2d];
 "485 hardsigmoid_8" [id=485, type=hardsigmoid];
 "486 quantize_per_tensor_default_43" [id=486, type=quantize_per_tensor];
@@ -521,7 +521,7 @@ strict digraph  {
 "519 compressed_weight_updated_constant34" [id=519, type=get_attr];
 "520 quantize_per_channel_default_45" [id=520, type=quantize_per_channel];
 "521 dequantize_per_channel_default_45" [id=521, type=dequantize_per_channel];
-"522 _param_constant139_0_0" [id=522, type=get_attr];
+"522 classifier_0_bias_0_0" [id=522, type=get_attr];
 "523 linear" [id=523, type=linear];
 "524 hardswish__18" [id=524, type=hardswish_];
 "525 quantize_per_tensor_default_49" [id=525, type=quantize_per_tensor];
@@ -532,594 +532,594 @@ strict digraph  {
 "530 compressed_weight_updated_constant35" [id=530, type=get_attr];
 "531 quantize_per_channel_default_46" [id=531, type=quantize_per_channel];
 "532 dequantize_per_channel_default_46" [id=532, type=dequantize_per_channel];
-"533 _param_constant141_0_0" [id=533, type=get_attr];
+"533 classifier_3_bias_0_0" [id=533, type=get_attr];
 "534 linear_1" [id=534, type=linear];
 "535 output" [id=535, type=output];
-"0 arg0_1" -> "1 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"1 quantize_per_tensor_default" -> "2 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"2 dequantize_per_tensor_default" -> "9 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"3 conv2d_scale_0" -> "6 quantize_per_channel_default"  [label="(16,)", style=solid];
-"3 conv2d_scale_0" -> "7 dequantize_per_channel_default"  [label="(16,)", style=solid];
-"4 conv2d_zero_point_0" -> "6 quantize_per_channel_default"  [label="(16,)", style=solid];
-"4 conv2d_zero_point_0" -> "7 dequantize_per_channel_default"  [label="(16,)", style=solid];
-"5 compressed_weight_updated_constant0" -> "6 quantize_per_channel_default"  [label="(16, 3, 3, 3)", style=solid];
-"6 quantize_per_channel_default" -> "7 dequantize_per_channel_default"  [label="(16, 3, 3, 3)", style=solid];
-"7 dequantize_per_channel_default" -> "9 conv2d"  [label="(16, 3, 3, 3)", style=solid];
-"8 features_0_0_weight_bias_0_0" -> "9 conv2d"  [label="(16,)", style=solid];
-"9 conv2d" -> "12 hardswish_"  [label="(1, 16, 112, 112)", style=solid];
-"10 hardswish__scale_0" -> "13 quantize_per_channel_default_1"  [label="(16,)", style=solid];
-"10 hardswish__scale_0" -> "14 dequantize_per_channel_default_1"  [label="(16,)", style=solid];
-"11 hardswish__zero_point_0" -> "13 quantize_per_channel_default_1"  [label="(16,)", style=solid];
-"11 hardswish__zero_point_0" -> "14 dequantize_per_channel_default_1"  [label="(16,)", style=solid];
-"12 hardswish_" -> "13 quantize_per_channel_default_1"  [label="(1, 16, 112, 112)", style=solid];
-"13 quantize_per_channel_default_1" -> "14 dequantize_per_channel_default_1"  [label="(1, 16, 112, 112)", style=solid];
-"14 dequantize_per_channel_default_1" -> "21 conv2d_1"  [label="(1, 16, 112, 112)", style=solid];
-"15 conv2d_1_scale_0" -> "18 quantize_per_channel_default_2"  [label="(16,)", style=solid];
-"15 conv2d_1_scale_0" -> "19 dequantize_per_channel_default_2"  [label="(16,)", style=solid];
-"16 conv2d_1_zero_point_0" -> "18 quantize_per_channel_default_2"  [label="(16,)", style=solid];
-"16 conv2d_1_zero_point_0" -> "19 dequantize_per_channel_default_2"  [label="(16,)", style=solid];
-"17 compressed_weight_updated_constant1" -> "18 quantize_per_channel_default_2"  [label="(16, 1, 3, 3)", style=solid];
-"18 quantize_per_channel_default_2" -> "19 dequantize_per_channel_default_2"  [label="(16, 1, 3, 3)", style=solid];
-"19 dequantize_per_channel_default_2" -> "21 conv2d_1"  [label="(16, 1, 3, 3)", style=solid];
-"20 features_1_block_0_0_weight_bias_0_0" -> "21 conv2d_1"  [label="(16,)", style=solid];
-"21 conv2d_1" -> "22 relu_"  [label="(1, 16, 56, 56)", style=solid];
-"22 relu_" -> "23 quantize_per_tensor_default_1"  [label="(1, 16, 56, 56)", style=solid];
-"23 quantize_per_tensor_default_1" -> "24 dequantize_per_tensor_default_2"  [label="(1, 16, 56, 56)", style=solid];
-"23 quantize_per_tensor_default_1" -> "25 dequantize_per_tensor_default_1"  [label="(1, 16, 56, 56)", style=solid];
-"24 dequantize_per_tensor_default_2" -> "37 mul"  [label="(1, 16, 56, 56)", style=solid];
-"25 dequantize_per_tensor_default_1" -> "26 adaptive_avg_pool2d"  [label="(1, 16, 56, 56)", style=solid];
-"26 adaptive_avg_pool2d" -> "29 conv2d_2"  [label="(1, 16, 1, 1)", style=solid];
-"27 _param_constant6" -> "29 conv2d_2"  [label="(8, 16, 1, 1)", style=solid];
-"28 _param_constant7_0_0" -> "29 conv2d_2"  [label="(8,)", style=solid];
-"29 conv2d_2" -> "30 relu"  [label="(1, 8, 1, 1)", style=solid];
-"30 relu" -> "33 conv2d_3"  [label="(1, 8, 1, 1)", style=solid];
-"31 _param_constant8" -> "33 conv2d_3"  [label="(16, 8, 1, 1)", style=solid];
-"32 _param_constant9_0_0" -> "33 conv2d_3"  [label="(16,)", style=solid];
-"33 conv2d_3" -> "34 hardsigmoid"  [label="(1, 16, 1, 1)", style=solid];
-"34 hardsigmoid" -> "35 quantize_per_tensor_default_2"  [label="(1, 16, 1, 1)", style=solid];
-"35 quantize_per_tensor_default_2" -> "36 dequantize_per_tensor_default_3"  [label="(1, 16, 1, 1)", style=solid];
-"36 dequantize_per_tensor_default_3" -> "37 mul"  [label="(1, 16, 1, 1)", style=solid];
-"37 mul" -> "38 quantize_per_tensor_default_3"  [label="(1, 16, 56, 56)", style=solid];
-"38 quantize_per_tensor_default_3" -> "39 dequantize_per_tensor_default_4"  [label="(1, 16, 56, 56)", style=solid];
-"39 dequantize_per_tensor_default_4" -> "46 conv2d_4"  [label="(1, 16, 56, 56)", style=solid];
-"40 conv2d_4_scale_0" -> "43 quantize_per_channel_default_3"  [label="(16,)", style=solid];
-"40 conv2d_4_scale_0" -> "44 dequantize_per_channel_default_3"  [label="(16,)", style=solid];
-"41 conv2d_4_zero_point_0" -> "43 quantize_per_channel_default_3"  [label="(16,)", style=solid];
-"41 conv2d_4_zero_point_0" -> "44 dequantize_per_channel_default_3"  [label="(16,)", style=solid];
-"42 compressed_weight_updated_constant2" -> "43 quantize_per_channel_default_3"  [label="(16, 16, 1, 1)", style=solid];
-"43 quantize_per_channel_default_3" -> "44 dequantize_per_channel_default_3"  [label="(16, 16, 1, 1)", style=solid];
-"44 dequantize_per_channel_default_3" -> "46 conv2d_4"  [label="(16, 16, 1, 1)", style=solid];
-"45 features_1_block_2_0_weight_bias_0_0" -> "46 conv2d_4"  [label="(16,)", style=solid];
-"46 conv2d_4" -> "47 quantize_per_tensor_default_4"  [label="(1, 16, 56, 56)", style=solid];
-"47 quantize_per_tensor_default_4" -> "48 dequantize_per_tensor_default_5"  [label="(1, 16, 56, 56)", style=solid];
-"48 dequantize_per_tensor_default_5" -> "55 conv2d_5"  [label="(1, 16, 56, 56)", style=solid];
-"49 conv2d_5_scale_0" -> "52 quantize_per_channel_default_4"  [label="(72,)", style=solid];
-"49 conv2d_5_scale_0" -> "53 dequantize_per_channel_default_4"  [label="(72,)", style=solid];
-"50 conv2d_5_zero_point_0" -> "52 quantize_per_channel_default_4"  [label="(72,)", style=solid];
-"50 conv2d_5_zero_point_0" -> "53 dequantize_per_channel_default_4"  [label="(72,)", style=solid];
-"51 compressed_weight_updated_constant3" -> "52 quantize_per_channel_default_4"  [label="(72, 16, 1, 1)", style=solid];
-"52 quantize_per_channel_default_4" -> "53 dequantize_per_channel_default_4"  [label="(72, 16, 1, 1)", style=solid];
-"53 dequantize_per_channel_default_4" -> "55 conv2d_5"  [label="(72, 16, 1, 1)", style=solid];
-"54 features_2_block_0_0_weight_bias_0_0" -> "55 conv2d_5"  [label="(72,)", style=solid];
-"55 conv2d_5" -> "58 relu__1"  [label="(1, 72, 56, 56)", style=solid];
-"56 relu__1_scale_0" -> "59 quantize_per_channel_default_5"  [label="(72,)", style=solid];
-"56 relu__1_scale_0" -> "60 dequantize_per_channel_default_5"  [label="(72,)", style=solid];
-"57 relu__1_zero_point_0" -> "59 quantize_per_channel_default_5"  [label="(72,)", style=solid];
-"57 relu__1_zero_point_0" -> "60 dequantize_per_channel_default_5"  [label="(72,)", style=solid];
-"58 relu__1" -> "59 quantize_per_channel_default_5"  [label="(1, 72, 56, 56)", style=solid];
-"59 quantize_per_channel_default_5" -> "60 dequantize_per_channel_default_5"  [label="(1, 72, 56, 56)", style=solid];
-"60 dequantize_per_channel_default_5" -> "67 conv2d_6"  [label="(1, 72, 56, 56)", style=solid];
-"61 conv2d_6_scale_0" -> "64 quantize_per_channel_default_6"  [label="(72,)", style=solid];
-"61 conv2d_6_scale_0" -> "65 dequantize_per_channel_default_6"  [label="(72,)", style=solid];
-"62 conv2d_6_zero_point_0" -> "64 quantize_per_channel_default_6"  [label="(72,)", style=solid];
-"62 conv2d_6_zero_point_0" -> "65 dequantize_per_channel_default_6"  [label="(72,)", style=solid];
-"63 compressed_weight_updated_constant4" -> "64 quantize_per_channel_default_6"  [label="(72, 1, 3, 3)", style=solid];
-"64 quantize_per_channel_default_6" -> "65 dequantize_per_channel_default_6"  [label="(72, 1, 3, 3)", style=solid];
-"65 dequantize_per_channel_default_6" -> "67 conv2d_6"  [label="(72, 1, 3, 3)", style=solid];
-"66 features_2_block_1_0_weight_bias_0_0" -> "67 conv2d_6"  [label="(72,)", style=solid];
-"67 conv2d_6" -> "68 relu__2"  [label="(1, 72, 28, 28)", style=solid];
-"68 relu__2" -> "69 quantize_per_tensor_default_5"  [label="(1, 72, 28, 28)", style=solid];
-"69 quantize_per_tensor_default_5" -> "70 dequantize_per_tensor_default_6"  [label="(1, 72, 28, 28)", style=solid];
-"70 dequantize_per_tensor_default_6" -> "77 conv2d_7"  [label="(1, 72, 28, 28)", style=solid];
-"71 conv2d_7_scale_0" -> "74 quantize_per_channel_default_7"  [label="(24,)", style=solid];
-"71 conv2d_7_scale_0" -> "75 dequantize_per_channel_default_7"  [label="(24,)", style=solid];
-"72 conv2d_7_zero_point_0" -> "74 quantize_per_channel_default_7"  [label="(24,)", style=solid];
-"72 conv2d_7_zero_point_0" -> "75 dequantize_per_channel_default_7"  [label="(24,)", style=solid];
-"73 compressed_weight_updated_constant5" -> "74 quantize_per_channel_default_7"  [label="(24, 72, 1, 1)", style=solid];
-"74 quantize_per_channel_default_7" -> "75 dequantize_per_channel_default_7"  [label="(24, 72, 1, 1)", style=solid];
-"75 dequantize_per_channel_default_7" -> "77 conv2d_7"  [label="(24, 72, 1, 1)", style=solid];
-"76 features_2_block_2_0_weight_bias_0_0" -> "77 conv2d_7"  [label="(24,)", style=solid];
-"77 conv2d_7" -> "78 quantize_per_tensor_default_6"  [label="(1, 24, 28, 28)", style=solid];
-"78 quantize_per_tensor_default_6" -> "79 dequantize_per_tensor_default_8"  [label="(1, 24, 28, 28)", style=solid];
-"78 quantize_per_tensor_default_6" -> "80 dequantize_per_tensor_default_7"  [label="(1, 24, 28, 28)", style=solid];
-"79 dequantize_per_tensor_default_8" -> "112 add_"  [label="(1, 24, 28, 28)", style=solid];
-"80 dequantize_per_tensor_default_7" -> "87 conv2d_8"  [label="(1, 24, 28, 28)", style=solid];
-"81 conv2d_8_scale_0" -> "84 quantize_per_channel_default_8"  [label="(88,)", style=solid];
-"81 conv2d_8_scale_0" -> "85 dequantize_per_channel_default_8"  [label="(88,)", style=solid];
-"82 conv2d_8_zero_point_0" -> "84 quantize_per_channel_default_8"  [label="(88,)", style=solid];
-"82 conv2d_8_zero_point_0" -> "85 dequantize_per_channel_default_8"  [label="(88,)", style=solid];
-"83 compressed_weight_updated_constant6" -> "84 quantize_per_channel_default_8"  [label="(88, 24, 1, 1)", style=solid];
-"84 quantize_per_channel_default_8" -> "85 dequantize_per_channel_default_8"  [label="(88, 24, 1, 1)", style=solid];
-"85 dequantize_per_channel_default_8" -> "87 conv2d_8"  [label="(88, 24, 1, 1)", style=solid];
-"86 features_3_block_0_0_weight_bias_0_0" -> "87 conv2d_8"  [label="(88,)", style=solid];
-"87 conv2d_8" -> "90 relu__3"  [label="(1, 88, 28, 28)", style=solid];
-"88 relu__3_scale_0" -> "91 quantize_per_channel_default_9"  [label="(88,)", style=solid];
-"88 relu__3_scale_0" -> "92 dequantize_per_channel_default_9"  [label="(88,)", style=solid];
-"89 relu__3_zero_point_0" -> "91 quantize_per_channel_default_9"  [label="(88,)", style=solid];
-"89 relu__3_zero_point_0" -> "92 dequantize_per_channel_default_9"  [label="(88,)", style=solid];
-"90 relu__3" -> "91 quantize_per_channel_default_9"  [label="(1, 88, 28, 28)", style=solid];
-"91 quantize_per_channel_default_9" -> "92 dequantize_per_channel_default_9"  [label="(1, 88, 28, 28)", style=solid];
-"92 dequantize_per_channel_default_9" -> "99 conv2d_9"  [label="(1, 88, 28, 28)", style=solid];
-"93 conv2d_9_scale_0" -> "96 quantize_per_channel_default_10"  [label="(88,)", style=solid];
-"93 conv2d_9_scale_0" -> "97 dequantize_per_channel_default_10"  [label="(88,)", style=solid];
-"94 conv2d_9_zero_point_0" -> "96 quantize_per_channel_default_10"  [label="(88,)", style=solid];
-"94 conv2d_9_zero_point_0" -> "97 dequantize_per_channel_default_10"  [label="(88,)", style=solid];
-"95 compressed_weight_updated_constant7" -> "96 quantize_per_channel_default_10"  [label="(88, 1, 3, 3)", style=solid];
-"96 quantize_per_channel_default_10" -> "97 dequantize_per_channel_default_10"  [label="(88, 1, 3, 3)", style=solid];
-"97 dequantize_per_channel_default_10" -> "99 conv2d_9"  [label="(88, 1, 3, 3)", style=solid];
-"98 features_3_block_1_0_weight_bias_0_0" -> "99 conv2d_9"  [label="(88,)", style=solid];
-"99 conv2d_9" -> "100 relu__4"  [label="(1, 88, 28, 28)", style=solid];
-"100 relu__4" -> "101 quantize_per_tensor_default_7"  [label="(1, 88, 28, 28)", style=solid];
-"101 quantize_per_tensor_default_7" -> "102 dequantize_per_tensor_default_9"  [label="(1, 88, 28, 28)", style=solid];
-"102 dequantize_per_tensor_default_9" -> "109 conv2d_10"  [label="(1, 88, 28, 28)", style=solid];
-"103 conv2d_10_scale_0" -> "106 quantize_per_channel_default_11"  [label="(24,)", style=solid];
-"103 conv2d_10_scale_0" -> "107 dequantize_per_channel_default_11"  [label="(24,)", style=solid];
-"104 conv2d_10_zero_point_0" -> "106 quantize_per_channel_default_11"  [label="(24,)", style=solid];
-"104 conv2d_10_zero_point_0" -> "107 dequantize_per_channel_default_11"  [label="(24,)", style=solid];
-"105 compressed_weight_updated_constant8" -> "106 quantize_per_channel_default_11"  [label="(24, 88, 1, 1)", style=solid];
-"106 quantize_per_channel_default_11" -> "107 dequantize_per_channel_default_11"  [label="(24, 88, 1, 1)", style=solid];
-"107 dequantize_per_channel_default_11" -> "109 conv2d_10"  [label="(24, 88, 1, 1)", style=solid];
-"108 features_3_block_2_0_weight_bias_0_0" -> "109 conv2d_10"  [label="(24,)", style=solid];
-"109 conv2d_10" -> "110 quantize_per_tensor_default_8"  [label="(1, 24, 28, 28)", style=solid];
-"110 quantize_per_tensor_default_8" -> "111 dequantize_per_tensor_default_10"  [label="(1, 24, 28, 28)", style=solid];
-"111 dequantize_per_tensor_default_10" -> "112 add_"  [label="(1, 24, 28, 28)", style=solid];
-"112 add_" -> "113 quantize_per_tensor_default_9"  [label="(1, 24, 28, 28)", style=solid];
-"113 quantize_per_tensor_default_9" -> "114 dequantize_per_tensor_default_11"  [label="(1, 24, 28, 28)", style=solid];
-"114 dequantize_per_tensor_default_11" -> "121 conv2d_11"  [label="(1, 24, 28, 28)", style=solid];
-"115 conv2d_11_scale_0" -> "118 quantize_per_channel_default_12"  [label="(96,)", style=solid];
-"115 conv2d_11_scale_0" -> "119 dequantize_per_channel_default_12"  [label="(96,)", style=solid];
-"116 conv2d_11_zero_point_0" -> "118 quantize_per_channel_default_12"  [label="(96,)", style=solid];
-"116 conv2d_11_zero_point_0" -> "119 dequantize_per_channel_default_12"  [label="(96,)", style=solid];
-"117 compressed_weight_updated_constant9" -> "118 quantize_per_channel_default_12"  [label="(96, 24, 1, 1)", style=solid];
-"118 quantize_per_channel_default_12" -> "119 dequantize_per_channel_default_12"  [label="(96, 24, 1, 1)", style=solid];
-"119 dequantize_per_channel_default_12" -> "121 conv2d_11"  [label="(96, 24, 1, 1)", style=solid];
-"120 features_4_block_0_0_weight_bias_0_0" -> "121 conv2d_11"  [label="(96,)", style=solid];
-"121 conv2d_11" -> "124 hardswish__1"  [label="(1, 96, 28, 28)", style=solid];
-"122 hardswish__1_scale_0" -> "125 quantize_per_channel_default_13"  [label="(96,)", style=solid];
-"122 hardswish__1_scale_0" -> "126 dequantize_per_channel_default_13"  [label="(96,)", style=solid];
-"123 hardswish__1_zero_point_0" -> "125 quantize_per_channel_default_13"  [label="(96,)", style=solid];
-"123 hardswish__1_zero_point_0" -> "126 dequantize_per_channel_default_13"  [label="(96,)", style=solid];
-"124 hardswish__1" -> "125 quantize_per_channel_default_13"  [label="(1, 96, 28, 28)", style=solid];
-"125 quantize_per_channel_default_13" -> "126 dequantize_per_channel_default_13"  [label="(1, 96, 28, 28)", style=solid];
-"126 dequantize_per_channel_default_13" -> "133 conv2d_12"  [label="(1, 96, 28, 28)", style=solid];
-"127 conv2d_12_scale_0" -> "130 quantize_per_channel_default_14"  [label="(96,)", style=solid];
-"127 conv2d_12_scale_0" -> "131 dequantize_per_channel_default_14"  [label="(96,)", style=solid];
-"128 conv2d_12_zero_point_0" -> "130 quantize_per_channel_default_14"  [label="(96,)", style=solid];
-"128 conv2d_12_zero_point_0" -> "131 dequantize_per_channel_default_14"  [label="(96,)", style=solid];
-"129 compressed_weight_updated_constant10" -> "130 quantize_per_channel_default_14"  [label="(96, 1, 5, 5)", style=solid];
-"130 quantize_per_channel_default_14" -> "131 dequantize_per_channel_default_14"  [label="(96, 1, 5, 5)", style=solid];
-"131 dequantize_per_channel_default_14" -> "133 conv2d_12"  [label="(96, 1, 5, 5)", style=solid];
-"132 features_4_block_1_0_weight_bias_0_0" -> "133 conv2d_12"  [label="(96,)", style=solid];
-"133 conv2d_12" -> "134 hardswish__2"  [label="(1, 96, 14, 14)", style=solid];
-"134 hardswish__2" -> "135 quantize_per_tensor_default_10"  [label="(1, 96, 14, 14)", style=solid];
-"135 quantize_per_tensor_default_10" -> "136 dequantize_per_tensor_default_13"  [label="(1, 96, 14, 14)", style=solid];
-"135 quantize_per_tensor_default_10" -> "137 dequantize_per_tensor_default_12"  [label="(1, 96, 14, 14)", style=solid];
-"136 dequantize_per_tensor_default_13" -> "149 mul_1"  [label="(1, 96, 14, 14)", style=solid];
-"137 dequantize_per_tensor_default_12" -> "138 adaptive_avg_pool2d_1"  [label="(1, 96, 14, 14)", style=solid];
-"138 adaptive_avg_pool2d_1" -> "141 conv2d_13"  [label="(1, 96, 1, 1)", style=solid];
-"139 _param_constant37" -> "141 conv2d_13"  [label="(24, 96, 1, 1)", style=solid];
-"140 _param_constant38_0_0" -> "141 conv2d_13"  [label="(24,)", style=solid];
-"141 conv2d_13" -> "142 relu_1"  [label="(1, 24, 1, 1)", style=solid];
-"142 relu_1" -> "145 conv2d_14"  [label="(1, 24, 1, 1)", style=solid];
-"143 _param_constant39" -> "145 conv2d_14"  [label="(96, 24, 1, 1)", style=solid];
-"144 _param_constant40_0_0" -> "145 conv2d_14"  [label="(96,)", style=solid];
-"145 conv2d_14" -> "146 hardsigmoid_1"  [label="(1, 96, 1, 1)", style=solid];
-"146 hardsigmoid_1" -> "147 quantize_per_tensor_default_11"  [label="(1, 96, 1, 1)", style=solid];
-"147 quantize_per_tensor_default_11" -> "148 dequantize_per_tensor_default_14"  [label="(1, 96, 1, 1)", style=solid];
-"148 dequantize_per_tensor_default_14" -> "149 mul_1"  [label="(1, 96, 1, 1)", style=solid];
-"149 mul_1" -> "150 quantize_per_tensor_default_12"  [label="(1, 96, 14, 14)", style=solid];
-"150 quantize_per_tensor_default_12" -> "151 dequantize_per_tensor_default_15"  [label="(1, 96, 14, 14)", style=solid];
-"151 dequantize_per_tensor_default_15" -> "158 conv2d_15"  [label="(1, 96, 14, 14)", style=solid];
-"152 conv2d_15_scale_0" -> "155 quantize_per_channel_default_15"  [label="(40,)", style=solid];
-"152 conv2d_15_scale_0" -> "156 dequantize_per_channel_default_15"  [label="(40,)", style=solid];
-"153 conv2d_15_zero_point_0" -> "155 quantize_per_channel_default_15"  [label="(40,)", style=solid];
-"153 conv2d_15_zero_point_0" -> "156 dequantize_per_channel_default_15"  [label="(40,)", style=solid];
-"154 compressed_weight_updated_constant11" -> "155 quantize_per_channel_default_15"  [label="(40, 96, 1, 1)", style=solid];
-"155 quantize_per_channel_default_15" -> "156 dequantize_per_channel_default_15"  [label="(40, 96, 1, 1)", style=solid];
-"156 dequantize_per_channel_default_15" -> "158 conv2d_15"  [label="(40, 96, 1, 1)", style=solid];
-"157 features_4_block_3_0_weight_bias_0_0" -> "158 conv2d_15"  [label="(40,)", style=solid];
-"158 conv2d_15" -> "159 quantize_per_tensor_default_13"  [label="(1, 40, 14, 14)", style=solid];
-"159 quantize_per_tensor_default_13" -> "160 dequantize_per_tensor_default_17"  [label="(1, 40, 14, 14)", style=solid];
-"159 quantize_per_tensor_default_13" -> "161 dequantize_per_tensor_default_16"  [label="(1, 40, 14, 14)", style=solid];
-"160 dequantize_per_tensor_default_17" -> "208 add__1"  [label="(1, 40, 14, 14)", style=solid];
-"161 dequantize_per_tensor_default_16" -> "168 conv2d_16"  [label="(1, 40, 14, 14)", style=solid];
-"162 conv2d_16_scale_0" -> "165 quantize_per_channel_default_16"  [label="(240,)", style=solid];
-"162 conv2d_16_scale_0" -> "166 dequantize_per_channel_default_16"  [label="(240,)", style=solid];
-"163 conv2d_16_zero_point_0" -> "165 quantize_per_channel_default_16"  [label="(240,)", style=solid];
-"163 conv2d_16_zero_point_0" -> "166 dequantize_per_channel_default_16"  [label="(240,)", style=solid];
-"164 compressed_weight_updated_constant12" -> "165 quantize_per_channel_default_16"  [label="(240, 40, 1, 1)", style=solid];
-"165 quantize_per_channel_default_16" -> "166 dequantize_per_channel_default_16"  [label="(240, 40, 1, 1)", style=solid];
-"166 dequantize_per_channel_default_16" -> "168 conv2d_16"  [label="(240, 40, 1, 1)", style=solid];
-"167 features_5_block_0_0_weight_bias_0_0" -> "168 conv2d_16"  [label="(240,)", style=solid];
-"168 conv2d_16" -> "171 hardswish__3"  [label="(1, 240, 14, 14)", style=solid];
-"169 hardswish__3_scale_0" -> "172 quantize_per_channel_default_17"  [label="(240,)", style=solid];
-"169 hardswish__3_scale_0" -> "173 dequantize_per_channel_default_17"  [label="(240,)", style=solid];
-"170 hardswish__3_zero_point_0" -> "172 quantize_per_channel_default_17"  [label="(240,)", style=solid];
-"170 hardswish__3_zero_point_0" -> "173 dequantize_per_channel_default_17"  [label="(240,)", style=solid];
-"171 hardswish__3" -> "172 quantize_per_channel_default_17"  [label="(1, 240, 14, 14)", style=solid];
-"172 quantize_per_channel_default_17" -> "173 dequantize_per_channel_default_17"  [label="(1, 240, 14, 14)", style=solid];
-"173 dequantize_per_channel_default_17" -> "180 conv2d_17"  [label="(1, 240, 14, 14)", style=solid];
-"174 conv2d_17_scale_0" -> "177 quantize_per_channel_default_18"  [label="(240,)", style=solid];
-"174 conv2d_17_scale_0" -> "178 dequantize_per_channel_default_18"  [label="(240,)", style=solid];
-"175 conv2d_17_zero_point_0" -> "177 quantize_per_channel_default_18"  [label="(240,)", style=solid];
-"175 conv2d_17_zero_point_0" -> "178 dequantize_per_channel_default_18"  [label="(240,)", style=solid];
-"176 compressed_weight_updated_constant13" -> "177 quantize_per_channel_default_18"  [label="(240, 1, 5, 5)", style=solid];
-"177 quantize_per_channel_default_18" -> "178 dequantize_per_channel_default_18"  [label="(240, 1, 5, 5)", style=solid];
-"178 dequantize_per_channel_default_18" -> "180 conv2d_17"  [label="(240, 1, 5, 5)", style=solid];
-"179 features_5_block_1_0_weight_bias_0_0" -> "180 conv2d_17"  [label="(240,)", style=solid];
-"180 conv2d_17" -> "181 hardswish__4"  [label="(1, 240, 14, 14)", style=solid];
-"181 hardswish__4" -> "182 quantize_per_tensor_default_14"  [label="(1, 240, 14, 14)", style=solid];
-"182 quantize_per_tensor_default_14" -> "183 dequantize_per_tensor_default_19"  [label="(1, 240, 14, 14)", style=solid];
-"182 quantize_per_tensor_default_14" -> "184 dequantize_per_tensor_default_18"  [label="(1, 240, 14, 14)", style=solid];
-"183 dequantize_per_tensor_default_19" -> "196 mul_2"  [label="(1, 240, 14, 14)", style=solid];
-"184 dequantize_per_tensor_default_18" -> "185 adaptive_avg_pool2d_2"  [label="(1, 240, 14, 14)", style=solid];
-"185 adaptive_avg_pool2d_2" -> "188 conv2d_18"  [label="(1, 240, 1, 1)", style=solid];
-"186 _param_constant50" -> "188 conv2d_18"  [label="(64, 240, 1, 1)", style=solid];
-"187 _param_constant51_0_0" -> "188 conv2d_18"  [label="(64,)", style=solid];
-"188 conv2d_18" -> "189 relu_2"  [label="(1, 64, 1, 1)", style=solid];
-"189 relu_2" -> "192 conv2d_19"  [label="(1, 64, 1, 1)", style=solid];
-"190 _param_constant52" -> "192 conv2d_19"  [label="(240, 64, 1, 1)", style=solid];
-"191 _param_constant53_0_0" -> "192 conv2d_19"  [label="(240,)", style=solid];
-"192 conv2d_19" -> "193 hardsigmoid_2"  [label="(1, 240, 1, 1)", style=solid];
-"193 hardsigmoid_2" -> "194 quantize_per_tensor_default_15"  [label="(1, 240, 1, 1)", style=solid];
-"194 quantize_per_tensor_default_15" -> "195 dequantize_per_tensor_default_20"  [label="(1, 240, 1, 1)", style=solid];
-"195 dequantize_per_tensor_default_20" -> "196 mul_2"  [label="(1, 240, 1, 1)", style=solid];
-"196 mul_2" -> "197 quantize_per_tensor_default_16"  [label="(1, 240, 14, 14)", style=solid];
-"197 quantize_per_tensor_default_16" -> "198 dequantize_per_tensor_default_21"  [label="(1, 240, 14, 14)", style=solid];
-"198 dequantize_per_tensor_default_21" -> "205 conv2d_20"  [label="(1, 240, 14, 14)", style=solid];
-"199 conv2d_20_scale_0" -> "202 quantize_per_channel_default_19"  [label="(40,)", style=solid];
-"199 conv2d_20_scale_0" -> "203 dequantize_per_channel_default_19"  [label="(40,)", style=solid];
-"200 conv2d_20_zero_point_0" -> "202 quantize_per_channel_default_19"  [label="(40,)", style=solid];
-"200 conv2d_20_zero_point_0" -> "203 dequantize_per_channel_default_19"  [label="(40,)", style=solid];
-"201 compressed_weight_updated_constant14" -> "202 quantize_per_channel_default_19"  [label="(40, 240, 1, 1)", style=solid];
-"202 quantize_per_channel_default_19" -> "203 dequantize_per_channel_default_19"  [label="(40, 240, 1, 1)", style=solid];
-"203 dequantize_per_channel_default_19" -> "205 conv2d_20"  [label="(40, 240, 1, 1)", style=solid];
-"204 features_5_block_3_0_weight_bias_0_0" -> "205 conv2d_20"  [label="(40,)", style=solid];
-"205 conv2d_20" -> "206 quantize_per_tensor_default_17"  [label="(1, 40, 14, 14)", style=solid];
-"206 quantize_per_tensor_default_17" -> "207 dequantize_per_tensor_default_22"  [label="(1, 40, 14, 14)", style=solid];
-"207 dequantize_per_tensor_default_22" -> "208 add__1"  [label="(1, 40, 14, 14)", style=solid];
-"208 add__1" -> "209 quantize_per_tensor_default_18"  [label="(1, 40, 14, 14)", style=solid];
-"209 quantize_per_tensor_default_18" -> "210 dequantize_per_tensor_default_24"  [label="(1, 40, 14, 14)", style=solid];
-"209 quantize_per_tensor_default_18" -> "211 dequantize_per_tensor_default_23"  [label="(1, 40, 14, 14)", style=solid];
-"210 dequantize_per_tensor_default_24" -> "258 add__2"  [label="(1, 40, 14, 14)", style=solid];
-"211 dequantize_per_tensor_default_23" -> "218 conv2d_21"  [label="(1, 40, 14, 14)", style=solid];
-"212 conv2d_21_scale_0" -> "215 quantize_per_channel_default_20"  [label="(240,)", style=solid];
-"212 conv2d_21_scale_0" -> "216 dequantize_per_channel_default_20"  [label="(240,)", style=solid];
-"213 conv2d_21_zero_point_0" -> "215 quantize_per_channel_default_20"  [label="(240,)", style=solid];
-"213 conv2d_21_zero_point_0" -> "216 dequantize_per_channel_default_20"  [label="(240,)", style=solid];
-"214 compressed_weight_updated_constant15" -> "215 quantize_per_channel_default_20"  [label="(240, 40, 1, 1)", style=solid];
-"215 quantize_per_channel_default_20" -> "216 dequantize_per_channel_default_20"  [label="(240, 40, 1, 1)", style=solid];
-"216 dequantize_per_channel_default_20" -> "218 conv2d_21"  [label="(240, 40, 1, 1)", style=solid];
-"217 features_6_block_0_0_weight_bias_0_0" -> "218 conv2d_21"  [label="(240,)", style=solid];
-"218 conv2d_21" -> "221 hardswish__5"  [label="(1, 240, 14, 14)", style=solid];
-"219 hardswish__5_scale_0" -> "222 quantize_per_channel_default_21"  [label="(240,)", style=solid];
-"219 hardswish__5_scale_0" -> "223 dequantize_per_channel_default_21"  [label="(240,)", style=solid];
-"220 hardswish__5_zero_point_0" -> "222 quantize_per_channel_default_21"  [label="(240,)", style=solid];
-"220 hardswish__5_zero_point_0" -> "223 dequantize_per_channel_default_21"  [label="(240,)", style=solid];
-"221 hardswish__5" -> "222 quantize_per_channel_default_21"  [label="(1, 240, 14, 14)", style=solid];
-"222 quantize_per_channel_default_21" -> "223 dequantize_per_channel_default_21"  [label="(1, 240, 14, 14)", style=solid];
-"223 dequantize_per_channel_default_21" -> "230 conv2d_22"  [label="(1, 240, 14, 14)", style=solid];
-"224 conv2d_22_scale_0" -> "227 quantize_per_channel_default_22"  [label="(240,)", style=solid];
-"224 conv2d_22_scale_0" -> "228 dequantize_per_channel_default_22"  [label="(240,)", style=solid];
-"225 conv2d_22_zero_point_0" -> "227 quantize_per_channel_default_22"  [label="(240,)", style=solid];
-"225 conv2d_22_zero_point_0" -> "228 dequantize_per_channel_default_22"  [label="(240,)", style=solid];
-"226 compressed_weight_updated_constant16" -> "227 quantize_per_channel_default_22"  [label="(240, 1, 5, 5)", style=solid];
-"227 quantize_per_channel_default_22" -> "228 dequantize_per_channel_default_22"  [label="(240, 1, 5, 5)", style=solid];
-"228 dequantize_per_channel_default_22" -> "230 conv2d_22"  [label="(240, 1, 5, 5)", style=solid];
-"229 features_6_block_1_0_weight_bias_0_0" -> "230 conv2d_22"  [label="(240,)", style=solid];
-"230 conv2d_22" -> "231 hardswish__6"  [label="(1, 240, 14, 14)", style=solid];
-"231 hardswish__6" -> "232 quantize_per_tensor_default_19"  [label="(1, 240, 14, 14)", style=solid];
-"232 quantize_per_tensor_default_19" -> "233 dequantize_per_tensor_default_26"  [label="(1, 240, 14, 14)", style=solid];
-"232 quantize_per_tensor_default_19" -> "234 dequantize_per_tensor_default_25"  [label="(1, 240, 14, 14)", style=solid];
-"233 dequantize_per_tensor_default_26" -> "246 mul_3"  [label="(1, 240, 14, 14)", style=solid];
-"234 dequantize_per_tensor_default_25" -> "235 adaptive_avg_pool2d_3"  [label="(1, 240, 14, 14)", style=solid];
-"235 adaptive_avg_pool2d_3" -> "238 conv2d_23"  [label="(1, 240, 1, 1)", style=solid];
-"236 _param_constant63" -> "238 conv2d_23"  [label="(64, 240, 1, 1)", style=solid];
-"237 _param_constant64_0_0" -> "238 conv2d_23"  [label="(64,)", style=solid];
-"238 conv2d_23" -> "239 relu_3"  [label="(1, 64, 1, 1)", style=solid];
-"239 relu_3" -> "242 conv2d_24"  [label="(1, 64, 1, 1)", style=solid];
-"240 _param_constant65" -> "242 conv2d_24"  [label="(240, 64, 1, 1)", style=solid];
-"241 _param_constant66_0_0" -> "242 conv2d_24"  [label="(240,)", style=solid];
-"242 conv2d_24" -> "243 hardsigmoid_3"  [label="(1, 240, 1, 1)", style=solid];
-"243 hardsigmoid_3" -> "244 quantize_per_tensor_default_20"  [label="(1, 240, 1, 1)", style=solid];
-"244 quantize_per_tensor_default_20" -> "245 dequantize_per_tensor_default_27"  [label="(1, 240, 1, 1)", style=solid];
-"245 dequantize_per_tensor_default_27" -> "246 mul_3"  [label="(1, 240, 1, 1)", style=solid];
-"246 mul_3" -> "247 quantize_per_tensor_default_21"  [label="(1, 240, 14, 14)", style=solid];
-"247 quantize_per_tensor_default_21" -> "248 dequantize_per_tensor_default_28"  [label="(1, 240, 14, 14)", style=solid];
-"248 dequantize_per_tensor_default_28" -> "255 conv2d_25"  [label="(1, 240, 14, 14)", style=solid];
-"249 conv2d_25_scale_0" -> "252 quantize_per_channel_default_23"  [label="(40,)", style=solid];
-"249 conv2d_25_scale_0" -> "253 dequantize_per_channel_default_23"  [label="(40,)", style=solid];
-"250 conv2d_25_zero_point_0" -> "252 quantize_per_channel_default_23"  [label="(40,)", style=solid];
-"250 conv2d_25_zero_point_0" -> "253 dequantize_per_channel_default_23"  [label="(40,)", style=solid];
-"251 compressed_weight_updated_constant17" -> "252 quantize_per_channel_default_23"  [label="(40, 240, 1, 1)", style=solid];
-"252 quantize_per_channel_default_23" -> "253 dequantize_per_channel_default_23"  [label="(40, 240, 1, 1)", style=solid];
-"253 dequantize_per_channel_default_23" -> "255 conv2d_25"  [label="(40, 240, 1, 1)", style=solid];
-"254 features_6_block_3_0_weight_bias_0_0" -> "255 conv2d_25"  [label="(40,)", style=solid];
-"255 conv2d_25" -> "256 quantize_per_tensor_default_22"  [label="(1, 40, 14, 14)", style=solid];
-"256 quantize_per_tensor_default_22" -> "257 dequantize_per_tensor_default_29"  [label="(1, 40, 14, 14)", style=solid];
-"257 dequantize_per_tensor_default_29" -> "258 add__2"  [label="(1, 40, 14, 14)", style=solid];
-"258 add__2" -> "259 quantize_per_tensor_default_23"  [label="(1, 40, 14, 14)", style=solid];
-"259 quantize_per_tensor_default_23" -> "260 dequantize_per_tensor_default_30"  [label="(1, 40, 14, 14)", style=solid];
-"260 dequantize_per_tensor_default_30" -> "267 conv2d_26"  [label="(1, 40, 14, 14)", style=solid];
-"261 conv2d_26_scale_0" -> "264 quantize_per_channel_default_24"  [label="(120,)", style=solid];
-"261 conv2d_26_scale_0" -> "265 dequantize_per_channel_default_24"  [label="(120,)", style=solid];
-"262 conv2d_26_zero_point_0" -> "264 quantize_per_channel_default_24"  [label="(120,)", style=solid];
-"262 conv2d_26_zero_point_0" -> "265 dequantize_per_channel_default_24"  [label="(120,)", style=solid];
-"263 compressed_weight_updated_constant18" -> "264 quantize_per_channel_default_24"  [label="(120, 40, 1, 1)", style=solid];
-"264 quantize_per_channel_default_24" -> "265 dequantize_per_channel_default_24"  [label="(120, 40, 1, 1)", style=solid];
-"265 dequantize_per_channel_default_24" -> "267 conv2d_26"  [label="(120, 40, 1, 1)", style=solid];
-"266 features_7_block_0_0_weight_bias_0_0" -> "267 conv2d_26"  [label="(120,)", style=solid];
-"267 conv2d_26" -> "270 hardswish__7"  [label="(1, 120, 14, 14)", style=solid];
-"268 hardswish__7_scale_0" -> "271 quantize_per_channel_default_25"  [label="(120,)", style=solid];
-"268 hardswish__7_scale_0" -> "272 dequantize_per_channel_default_25"  [label="(120,)", style=solid];
-"269 hardswish__7_zero_point_0" -> "271 quantize_per_channel_default_25"  [label="(120,)", style=solid];
-"269 hardswish__7_zero_point_0" -> "272 dequantize_per_channel_default_25"  [label="(120,)", style=solid];
-"270 hardswish__7" -> "271 quantize_per_channel_default_25"  [label="(1, 120, 14, 14)", style=solid];
-"271 quantize_per_channel_default_25" -> "272 dequantize_per_channel_default_25"  [label="(1, 120, 14, 14)", style=solid];
-"272 dequantize_per_channel_default_25" -> "279 conv2d_27"  [label="(1, 120, 14, 14)", style=solid];
-"273 conv2d_27_scale_0" -> "276 quantize_per_channel_default_26"  [label="(120,)", style=solid];
-"273 conv2d_27_scale_0" -> "277 dequantize_per_channel_default_26"  [label="(120,)", style=solid];
-"274 conv2d_27_zero_point_0" -> "276 quantize_per_channel_default_26"  [label="(120,)", style=solid];
-"274 conv2d_27_zero_point_0" -> "277 dequantize_per_channel_default_26"  [label="(120,)", style=solid];
-"275 compressed_weight_updated_constant19" -> "276 quantize_per_channel_default_26"  [label="(120, 1, 5, 5)", style=solid];
-"276 quantize_per_channel_default_26" -> "277 dequantize_per_channel_default_26"  [label="(120, 1, 5, 5)", style=solid];
-"277 dequantize_per_channel_default_26" -> "279 conv2d_27"  [label="(120, 1, 5, 5)", style=solid];
-"278 features_7_block_1_0_weight_bias_0_0" -> "279 conv2d_27"  [label="(120,)", style=solid];
-"279 conv2d_27" -> "280 hardswish__8"  [label="(1, 120, 14, 14)", style=solid];
-"280 hardswish__8" -> "281 quantize_per_tensor_default_24"  [label="(1, 120, 14, 14)", style=solid];
-"281 quantize_per_tensor_default_24" -> "282 dequantize_per_tensor_default_32"  [label="(1, 120, 14, 14)", style=solid];
-"281 quantize_per_tensor_default_24" -> "283 dequantize_per_tensor_default_31"  [label="(1, 120, 14, 14)", style=solid];
-"282 dequantize_per_tensor_default_32" -> "295 mul_4"  [label="(1, 120, 14, 14)", style=solid];
-"283 dequantize_per_tensor_default_31" -> "284 adaptive_avg_pool2d_4"  [label="(1, 120, 14, 14)", style=solid];
-"284 adaptive_avg_pool2d_4" -> "287 conv2d_28"  [label="(1, 120, 1, 1)", style=solid];
-"285 _param_constant76" -> "287 conv2d_28"  [label="(32, 120, 1, 1)", style=solid];
-"286 _param_constant77_0_0" -> "287 conv2d_28"  [label="(32,)", style=solid];
-"287 conv2d_28" -> "288 relu_4"  [label="(1, 32, 1, 1)", style=solid];
-"288 relu_4" -> "291 conv2d_29"  [label="(1, 32, 1, 1)", style=solid];
-"289 _param_constant78" -> "291 conv2d_29"  [label="(120, 32, 1, 1)", style=solid];
-"290 _param_constant79_0_0" -> "291 conv2d_29"  [label="(120,)", style=solid];
-"291 conv2d_29" -> "292 hardsigmoid_4"  [label="(1, 120, 1, 1)", style=solid];
-"292 hardsigmoid_4" -> "293 quantize_per_tensor_default_25"  [label="(1, 120, 1, 1)", style=solid];
-"293 quantize_per_tensor_default_25" -> "294 dequantize_per_tensor_default_33"  [label="(1, 120, 1, 1)", style=solid];
-"294 dequantize_per_tensor_default_33" -> "295 mul_4"  [label="(1, 120, 1, 1)", style=solid];
-"295 mul_4" -> "296 quantize_per_tensor_default_26"  [label="(1, 120, 14, 14)", style=solid];
-"296 quantize_per_tensor_default_26" -> "297 dequantize_per_tensor_default_34"  [label="(1, 120, 14, 14)", style=solid];
-"297 dequantize_per_tensor_default_34" -> "304 conv2d_30"  [label="(1, 120, 14, 14)", style=solid];
-"298 conv2d_30_scale_0" -> "301 quantize_per_channel_default_27"  [label="(48,)", style=solid];
-"298 conv2d_30_scale_0" -> "302 dequantize_per_channel_default_27"  [label="(48,)", style=solid];
-"299 conv2d_30_zero_point_0" -> "301 quantize_per_channel_default_27"  [label="(48,)", style=solid];
-"299 conv2d_30_zero_point_0" -> "302 dequantize_per_channel_default_27"  [label="(48,)", style=solid];
-"300 compressed_weight_updated_constant20" -> "301 quantize_per_channel_default_27"  [label="(48, 120, 1, 1)", style=solid];
-"301 quantize_per_channel_default_27" -> "302 dequantize_per_channel_default_27"  [label="(48, 120, 1, 1)", style=solid];
-"302 dequantize_per_channel_default_27" -> "304 conv2d_30"  [label="(48, 120, 1, 1)", style=solid];
-"303 features_7_block_3_0_weight_bias_0_0" -> "304 conv2d_30"  [label="(48,)", style=solid];
-"304 conv2d_30" -> "305 quantize_per_tensor_default_27"  [label="(1, 48, 14, 14)", style=solid];
-"305 quantize_per_tensor_default_27" -> "306 dequantize_per_tensor_default_36"  [label="(1, 48, 14, 14)", style=solid];
-"305 quantize_per_tensor_default_27" -> "307 dequantize_per_tensor_default_35"  [label="(1, 48, 14, 14)", style=solid];
-"306 dequantize_per_tensor_default_36" -> "354 add__3"  [label="(1, 48, 14, 14)", style=solid];
-"307 dequantize_per_tensor_default_35" -> "314 conv2d_31"  [label="(1, 48, 14, 14)", style=solid];
-"308 conv2d_31_scale_0" -> "311 quantize_per_channel_default_28"  [label="(144,)", style=solid];
-"308 conv2d_31_scale_0" -> "312 dequantize_per_channel_default_28"  [label="(144,)", style=solid];
-"309 conv2d_31_zero_point_0" -> "311 quantize_per_channel_default_28"  [label="(144,)", style=solid];
-"309 conv2d_31_zero_point_0" -> "312 dequantize_per_channel_default_28"  [label="(144,)", style=solid];
-"310 compressed_weight_updated_constant21" -> "311 quantize_per_channel_default_28"  [label="(144, 48, 1, 1)", style=solid];
-"311 quantize_per_channel_default_28" -> "312 dequantize_per_channel_default_28"  [label="(144, 48, 1, 1)", style=solid];
-"312 dequantize_per_channel_default_28" -> "314 conv2d_31"  [label="(144, 48, 1, 1)", style=solid];
-"313 features_8_block_0_0_weight_bias_0_0" -> "314 conv2d_31"  [label="(144,)", style=solid];
-"314 conv2d_31" -> "317 hardswish__9"  [label="(1, 144, 14, 14)", style=solid];
-"315 hardswish__9_scale_0" -> "318 quantize_per_channel_default_29"  [label="(144,)", style=solid];
-"315 hardswish__9_scale_0" -> "319 dequantize_per_channel_default_29"  [label="(144,)", style=solid];
-"316 hardswish__9_zero_point_0" -> "318 quantize_per_channel_default_29"  [label="(144,)", style=solid];
-"316 hardswish__9_zero_point_0" -> "319 dequantize_per_channel_default_29"  [label="(144,)", style=solid];
-"317 hardswish__9" -> "318 quantize_per_channel_default_29"  [label="(1, 144, 14, 14)", style=solid];
-"318 quantize_per_channel_default_29" -> "319 dequantize_per_channel_default_29"  [label="(1, 144, 14, 14)", style=solid];
-"319 dequantize_per_channel_default_29" -> "326 conv2d_32"  [label="(1, 144, 14, 14)", style=solid];
-"320 conv2d_32_scale_0" -> "323 quantize_per_channel_default_30"  [label="(144,)", style=solid];
-"320 conv2d_32_scale_0" -> "324 dequantize_per_channel_default_30"  [label="(144,)", style=solid];
-"321 conv2d_32_zero_point_0" -> "323 quantize_per_channel_default_30"  [label="(144,)", style=solid];
-"321 conv2d_32_zero_point_0" -> "324 dequantize_per_channel_default_30"  [label="(144,)", style=solid];
-"322 compressed_weight_updated_constant22" -> "323 quantize_per_channel_default_30"  [label="(144, 1, 5, 5)", style=solid];
-"323 quantize_per_channel_default_30" -> "324 dequantize_per_channel_default_30"  [label="(144, 1, 5, 5)", style=solid];
-"324 dequantize_per_channel_default_30" -> "326 conv2d_32"  [label="(144, 1, 5, 5)", style=solid];
-"325 features_8_block_1_0_weight_bias_0_0" -> "326 conv2d_32"  [label="(144,)", style=solid];
-"326 conv2d_32" -> "327 hardswish__10"  [label="(1, 144, 14, 14)", style=solid];
-"327 hardswish__10" -> "328 quantize_per_tensor_default_28"  [label="(1, 144, 14, 14)", style=solid];
-"328 quantize_per_tensor_default_28" -> "329 dequantize_per_tensor_default_38"  [label="(1, 144, 14, 14)", style=solid];
-"328 quantize_per_tensor_default_28" -> "330 dequantize_per_tensor_default_37"  [label="(1, 144, 14, 14)", style=solid];
-"329 dequantize_per_tensor_default_38" -> "342 mul_5"  [label="(1, 144, 14, 14)", style=solid];
-"330 dequantize_per_tensor_default_37" -> "331 adaptive_avg_pool2d_5"  [label="(1, 144, 14, 14)", style=solid];
-"331 adaptive_avg_pool2d_5" -> "334 conv2d_33"  [label="(1, 144, 1, 1)", style=solid];
-"332 _param_constant89" -> "334 conv2d_33"  [label="(40, 144, 1, 1)", style=solid];
-"333 _param_constant90_0_0" -> "334 conv2d_33"  [label="(40,)", style=solid];
-"334 conv2d_33" -> "335 relu_5"  [label="(1, 40, 1, 1)", style=solid];
-"335 relu_5" -> "338 conv2d_34"  [label="(1, 40, 1, 1)", style=solid];
-"336 _param_constant91" -> "338 conv2d_34"  [label="(144, 40, 1, 1)", style=solid];
-"337 _param_constant92_0_0" -> "338 conv2d_34"  [label="(144,)", style=solid];
-"338 conv2d_34" -> "339 hardsigmoid_5"  [label="(1, 144, 1, 1)", style=solid];
-"339 hardsigmoid_5" -> "340 quantize_per_tensor_default_29"  [label="(1, 144, 1, 1)", style=solid];
-"340 quantize_per_tensor_default_29" -> "341 dequantize_per_tensor_default_39"  [label="(1, 144, 1, 1)", style=solid];
-"341 dequantize_per_tensor_default_39" -> "342 mul_5"  [label="(1, 144, 1, 1)", style=solid];
-"342 mul_5" -> "343 quantize_per_tensor_default_30"  [label="(1, 144, 14, 14)", style=solid];
-"343 quantize_per_tensor_default_30" -> "344 dequantize_per_tensor_default_40"  [label="(1, 144, 14, 14)", style=solid];
-"344 dequantize_per_tensor_default_40" -> "351 conv2d_35"  [label="(1, 144, 14, 14)", style=solid];
-"345 conv2d_35_scale_0" -> "348 quantize_per_channel_default_31"  [label="(48,)", style=solid];
-"345 conv2d_35_scale_0" -> "349 dequantize_per_channel_default_31"  [label="(48,)", style=solid];
-"346 conv2d_35_zero_point_0" -> "348 quantize_per_channel_default_31"  [label="(48,)", style=solid];
-"346 conv2d_35_zero_point_0" -> "349 dequantize_per_channel_default_31"  [label="(48,)", style=solid];
-"347 compressed_weight_updated_constant23" -> "348 quantize_per_channel_default_31"  [label="(48, 144, 1, 1)", style=solid];
-"348 quantize_per_channel_default_31" -> "349 dequantize_per_channel_default_31"  [label="(48, 144, 1, 1)", style=solid];
-"349 dequantize_per_channel_default_31" -> "351 conv2d_35"  [label="(48, 144, 1, 1)", style=solid];
-"350 features_8_block_3_0_weight_bias_0_0" -> "351 conv2d_35"  [label="(48,)", style=solid];
-"351 conv2d_35" -> "352 quantize_per_tensor_default_31"  [label="(1, 48, 14, 14)", style=solid];
-"352 quantize_per_tensor_default_31" -> "353 dequantize_per_tensor_default_41"  [label="(1, 48, 14, 14)", style=solid];
-"353 dequantize_per_tensor_default_41" -> "354 add__3"  [label="(1, 48, 14, 14)", style=solid];
-"354 add__3" -> "355 quantize_per_tensor_default_32"  [label="(1, 48, 14, 14)", style=solid];
-"355 quantize_per_tensor_default_32" -> "356 dequantize_per_tensor_default_42"  [label="(1, 48, 14, 14)", style=solid];
-"356 dequantize_per_tensor_default_42" -> "363 conv2d_36"  [label="(1, 48, 14, 14)", style=solid];
-"357 conv2d_36_scale_0" -> "360 quantize_per_channel_default_32"  [label="(288,)", style=solid];
-"357 conv2d_36_scale_0" -> "361 dequantize_per_channel_default_32"  [label="(288,)", style=solid];
-"358 conv2d_36_zero_point_0" -> "360 quantize_per_channel_default_32"  [label="(288,)", style=solid];
-"358 conv2d_36_zero_point_0" -> "361 dequantize_per_channel_default_32"  [label="(288,)", style=solid];
-"359 compressed_weight_updated_constant24" -> "360 quantize_per_channel_default_32"  [label="(288, 48, 1, 1)", style=solid];
-"360 quantize_per_channel_default_32" -> "361 dequantize_per_channel_default_32"  [label="(288, 48, 1, 1)", style=solid];
-"361 dequantize_per_channel_default_32" -> "363 conv2d_36"  [label="(288, 48, 1, 1)", style=solid];
-"362 features_9_block_0_0_weight_bias_0_0" -> "363 conv2d_36"  [label="(288,)", style=solid];
-"363 conv2d_36" -> "366 hardswish__11"  [label="(1, 288, 14, 14)", style=solid];
-"364 hardswish__11_scale_0" -> "367 quantize_per_channel_default_33"  [label="(288,)", style=solid];
-"364 hardswish__11_scale_0" -> "368 dequantize_per_channel_default_33"  [label="(288,)", style=solid];
-"365 hardswish__11_zero_point_0" -> "367 quantize_per_channel_default_33"  [label="(288,)", style=solid];
-"365 hardswish__11_zero_point_0" -> "368 dequantize_per_channel_default_33"  [label="(288,)", style=solid];
-"366 hardswish__11" -> "367 quantize_per_channel_default_33"  [label="(1, 288, 14, 14)", style=solid];
-"367 quantize_per_channel_default_33" -> "368 dequantize_per_channel_default_33"  [label="(1, 288, 14, 14)", style=solid];
-"368 dequantize_per_channel_default_33" -> "375 conv2d_37"  [label="(1, 288, 14, 14)", style=solid];
-"369 conv2d_37_scale_0" -> "372 quantize_per_channel_default_34"  [label="(288,)", style=solid];
-"369 conv2d_37_scale_0" -> "373 dequantize_per_channel_default_34"  [label="(288,)", style=solid];
-"370 conv2d_37_zero_point_0" -> "372 quantize_per_channel_default_34"  [label="(288,)", style=solid];
-"370 conv2d_37_zero_point_0" -> "373 dequantize_per_channel_default_34"  [label="(288,)", style=solid];
-"371 compressed_weight_updated_constant25" -> "372 quantize_per_channel_default_34"  [label="(288, 1, 5, 5)", style=solid];
-"372 quantize_per_channel_default_34" -> "373 dequantize_per_channel_default_34"  [label="(288, 1, 5, 5)", style=solid];
-"373 dequantize_per_channel_default_34" -> "375 conv2d_37"  [label="(288, 1, 5, 5)", style=solid];
-"374 features_9_block_1_0_weight_bias_0_0" -> "375 conv2d_37"  [label="(288,)", style=solid];
-"375 conv2d_37" -> "376 hardswish__12"  [label="(1, 288, 7, 7)", style=solid];
-"376 hardswish__12" -> "377 quantize_per_tensor_default_33"  [label="(1, 288, 7, 7)", style=solid];
-"377 quantize_per_tensor_default_33" -> "378 dequantize_per_tensor_default_44"  [label="(1, 288, 7, 7)", style=solid];
-"377 quantize_per_tensor_default_33" -> "379 dequantize_per_tensor_default_43"  [label="(1, 288, 7, 7)", style=solid];
-"378 dequantize_per_tensor_default_44" -> "391 mul_6"  [label="(1, 288, 7, 7)", style=solid];
-"379 dequantize_per_tensor_default_43" -> "380 adaptive_avg_pool2d_6"  [label="(1, 288, 7, 7)", style=solid];
-"380 adaptive_avg_pool2d_6" -> "383 conv2d_38"  [label="(1, 288, 1, 1)", style=solid];
-"381 _param_constant102" -> "383 conv2d_38"  [label="(72, 288, 1, 1)", style=solid];
-"382 _param_constant103_0_0" -> "383 conv2d_38"  [label="(72,)", style=solid];
-"383 conv2d_38" -> "384 relu_6"  [label="(1, 72, 1, 1)", style=solid];
-"384 relu_6" -> "387 conv2d_39"  [label="(1, 72, 1, 1)", style=solid];
-"385 _param_constant104" -> "387 conv2d_39"  [label="(288, 72, 1, 1)", style=solid];
-"386 _param_constant105_0_0" -> "387 conv2d_39"  [label="(288,)", style=solid];
-"387 conv2d_39" -> "388 hardsigmoid_6"  [label="(1, 288, 1, 1)", style=solid];
-"388 hardsigmoid_6" -> "389 quantize_per_tensor_default_34"  [label="(1, 288, 1, 1)", style=solid];
-"389 quantize_per_tensor_default_34" -> "390 dequantize_per_tensor_default_45"  [label="(1, 288, 1, 1)", style=solid];
-"390 dequantize_per_tensor_default_45" -> "391 mul_6"  [label="(1, 288, 1, 1)", style=solid];
-"391 mul_6" -> "392 quantize_per_tensor_default_35"  [label="(1, 288, 7, 7)", style=solid];
-"392 quantize_per_tensor_default_35" -> "393 dequantize_per_tensor_default_46"  [label="(1, 288, 7, 7)", style=solid];
-"393 dequantize_per_tensor_default_46" -> "400 conv2d_40"  [label="(1, 288, 7, 7)", style=solid];
-"394 conv2d_40_scale_0" -> "397 quantize_per_channel_default_35"  [label="(96,)", style=solid];
-"394 conv2d_40_scale_0" -> "398 dequantize_per_channel_default_35"  [label="(96,)", style=solid];
-"395 conv2d_40_zero_point_0" -> "397 quantize_per_channel_default_35"  [label="(96,)", style=solid];
-"395 conv2d_40_zero_point_0" -> "398 dequantize_per_channel_default_35"  [label="(96,)", style=solid];
-"396 compressed_weight_updated_constant26" -> "397 quantize_per_channel_default_35"  [label="(96, 288, 1, 1)", style=solid];
-"397 quantize_per_channel_default_35" -> "398 dequantize_per_channel_default_35"  [label="(96, 288, 1, 1)", style=solid];
-"398 dequantize_per_channel_default_35" -> "400 conv2d_40"  [label="(96, 288, 1, 1)", style=solid];
-"399 features_9_block_3_0_weight_bias_0_0" -> "400 conv2d_40"  [label="(96,)", style=solid];
-"400 conv2d_40" -> "401 quantize_per_tensor_default_36"  [label="(1, 96, 7, 7)", style=solid];
-"401 quantize_per_tensor_default_36" -> "402 dequantize_per_tensor_default_48"  [label="(1, 96, 7, 7)", style=solid];
-"401 quantize_per_tensor_default_36" -> "403 dequantize_per_tensor_default_47"  [label="(1, 96, 7, 7)", style=solid];
-"402 dequantize_per_tensor_default_48" -> "450 add__4"  [label="(1, 96, 7, 7)", style=solid];
-"403 dequantize_per_tensor_default_47" -> "410 conv2d_41"  [label="(1, 96, 7, 7)", style=solid];
-"404 conv2d_41_scale_0" -> "407 quantize_per_channel_default_36"  [label="(576,)", style=solid];
-"404 conv2d_41_scale_0" -> "408 dequantize_per_channel_default_36"  [label="(576,)", style=solid];
-"405 conv2d_41_zero_point_0" -> "407 quantize_per_channel_default_36"  [label="(576,)", style=solid];
-"405 conv2d_41_zero_point_0" -> "408 dequantize_per_channel_default_36"  [label="(576,)", style=solid];
-"406 compressed_weight_updated_constant27" -> "407 quantize_per_channel_default_36"  [label="(576, 96, 1, 1)", style=solid];
-"407 quantize_per_channel_default_36" -> "408 dequantize_per_channel_default_36"  [label="(576, 96, 1, 1)", style=solid];
-"408 dequantize_per_channel_default_36" -> "410 conv2d_41"  [label="(576, 96, 1, 1)", style=solid];
-"409 features_10_block_0_0_weight_bias_0_0" -> "410 conv2d_41"  [label="(576,)", style=solid];
-"410 conv2d_41" -> "413 hardswish__13"  [label="(1, 576, 7, 7)", style=solid];
-"411 hardswish__13_scale_0" -> "414 quantize_per_channel_default_37"  [label="(576,)", style=solid];
-"411 hardswish__13_scale_0" -> "415 dequantize_per_channel_default_37"  [label="(576,)", style=solid];
-"412 hardswish__13_zero_point_0" -> "414 quantize_per_channel_default_37"  [label="(576,)", style=solid];
-"412 hardswish__13_zero_point_0" -> "415 dequantize_per_channel_default_37"  [label="(576,)", style=solid];
-"413 hardswish__13" -> "414 quantize_per_channel_default_37"  [label="(1, 576, 7, 7)", style=solid];
-"414 quantize_per_channel_default_37" -> "415 dequantize_per_channel_default_37"  [label="(1, 576, 7, 7)", style=solid];
-"415 dequantize_per_channel_default_37" -> "422 conv2d_42"  [label="(1, 576, 7, 7)", style=solid];
-"416 conv2d_42_scale_0" -> "419 quantize_per_channel_default_38"  [label="(576,)", style=solid];
-"416 conv2d_42_scale_0" -> "420 dequantize_per_channel_default_38"  [label="(576,)", style=solid];
-"417 conv2d_42_zero_point_0" -> "419 quantize_per_channel_default_38"  [label="(576,)", style=solid];
-"417 conv2d_42_zero_point_0" -> "420 dequantize_per_channel_default_38"  [label="(576,)", style=solid];
-"418 compressed_weight_updated_constant28" -> "419 quantize_per_channel_default_38"  [label="(576, 1, 5, 5)", style=solid];
-"419 quantize_per_channel_default_38" -> "420 dequantize_per_channel_default_38"  [label="(576, 1, 5, 5)", style=solid];
-"420 dequantize_per_channel_default_38" -> "422 conv2d_42"  [label="(576, 1, 5, 5)", style=solid];
-"421 features_10_block_1_0_weight_bias_0_0" -> "422 conv2d_42"  [label="(576,)", style=solid];
-"422 conv2d_42" -> "423 hardswish__14"  [label="(1, 576, 7, 7)", style=solid];
-"423 hardswish__14" -> "424 quantize_per_tensor_default_37"  [label="(1, 576, 7, 7)", style=solid];
-"424 quantize_per_tensor_default_37" -> "425 dequantize_per_tensor_default_50"  [label="(1, 576, 7, 7)", style=solid];
-"424 quantize_per_tensor_default_37" -> "426 dequantize_per_tensor_default_49"  [label="(1, 576, 7, 7)", style=solid];
-"425 dequantize_per_tensor_default_50" -> "438 mul_7"  [label="(1, 576, 7, 7)", style=solid];
-"426 dequantize_per_tensor_default_49" -> "427 adaptive_avg_pool2d_7"  [label="(1, 576, 7, 7)", style=solid];
-"427 adaptive_avg_pool2d_7" -> "430 conv2d_43"  [label="(1, 576, 1, 1)", style=solid];
-"428 _param_constant115" -> "430 conv2d_43"  [label="(144, 576, 1, 1)", style=solid];
-"429 _param_constant116_0_0" -> "430 conv2d_43"  [label="(144,)", style=solid];
-"430 conv2d_43" -> "431 relu_7"  [label="(1, 144, 1, 1)", style=solid];
-"431 relu_7" -> "434 conv2d_44"  [label="(1, 144, 1, 1)", style=solid];
-"432 _param_constant117" -> "434 conv2d_44"  [label="(576, 144, 1, 1)", style=solid];
-"433 _param_constant118_0_0" -> "434 conv2d_44"  [label="(576,)", style=solid];
-"434 conv2d_44" -> "435 hardsigmoid_7"  [label="(1, 576, 1, 1)", style=solid];
-"435 hardsigmoid_7" -> "436 quantize_per_tensor_default_38"  [label="(1, 576, 1, 1)", style=solid];
-"436 quantize_per_tensor_default_38" -> "437 dequantize_per_tensor_default_51"  [label="(1, 576, 1, 1)", style=solid];
-"437 dequantize_per_tensor_default_51" -> "438 mul_7"  [label="(1, 576, 1, 1)", style=solid];
-"438 mul_7" -> "439 quantize_per_tensor_default_39"  [label="(1, 576, 7, 7)", style=solid];
-"439 quantize_per_tensor_default_39" -> "440 dequantize_per_tensor_default_52"  [label="(1, 576, 7, 7)", style=solid];
-"440 dequantize_per_tensor_default_52" -> "447 conv2d_45"  [label="(1, 576, 7, 7)", style=solid];
-"441 conv2d_45_scale_0" -> "444 quantize_per_channel_default_39"  [label="(96,)", style=solid];
-"441 conv2d_45_scale_0" -> "445 dequantize_per_channel_default_39"  [label="(96,)", style=solid];
-"442 conv2d_45_zero_point_0" -> "444 quantize_per_channel_default_39"  [label="(96,)", style=solid];
-"442 conv2d_45_zero_point_0" -> "445 dequantize_per_channel_default_39"  [label="(96,)", style=solid];
-"443 compressed_weight_updated_constant29" -> "444 quantize_per_channel_default_39"  [label="(96, 576, 1, 1)", style=solid];
-"444 quantize_per_channel_default_39" -> "445 dequantize_per_channel_default_39"  [label="(96, 576, 1, 1)", style=solid];
-"445 dequantize_per_channel_default_39" -> "447 conv2d_45"  [label="(96, 576, 1, 1)", style=solid];
-"446 features_10_block_3_0_weight_bias_0_0" -> "447 conv2d_45"  [label="(96,)", style=solid];
-"447 conv2d_45" -> "448 quantize_per_tensor_default_40"  [label="(1, 96, 7, 7)", style=solid];
-"448 quantize_per_tensor_default_40" -> "449 dequantize_per_tensor_default_53"  [label="(1, 96, 7, 7)", style=solid];
-"449 dequantize_per_tensor_default_53" -> "450 add__4"  [label="(1, 96, 7, 7)", style=solid];
-"450 add__4" -> "451 quantize_per_tensor_default_41"  [label="(1, 96, 7, 7)", style=solid];
-"451 quantize_per_tensor_default_41" -> "452 dequantize_per_tensor_default_55"  [label="(1, 96, 7, 7)", style=solid];
-"451 quantize_per_tensor_default_41" -> "453 dequantize_per_tensor_default_54"  [label="(1, 96, 7, 7)", style=solid];
-"452 dequantize_per_tensor_default_55" -> "500 add__5"  [label="(1, 96, 7, 7)", style=solid];
-"453 dequantize_per_tensor_default_54" -> "460 conv2d_46"  [label="(1, 96, 7, 7)", style=solid];
-"454 conv2d_46_scale_0" -> "457 quantize_per_channel_default_40"  [label="(576,)", style=solid];
-"454 conv2d_46_scale_0" -> "458 dequantize_per_channel_default_40"  [label="(576,)", style=solid];
-"455 conv2d_46_zero_point_0" -> "457 quantize_per_channel_default_40"  [label="(576,)", style=solid];
-"455 conv2d_46_zero_point_0" -> "458 dequantize_per_channel_default_40"  [label="(576,)", style=solid];
-"456 compressed_weight_updated_constant30" -> "457 quantize_per_channel_default_40"  [label="(576, 96, 1, 1)", style=solid];
-"457 quantize_per_channel_default_40" -> "458 dequantize_per_channel_default_40"  [label="(576, 96, 1, 1)", style=solid];
-"458 dequantize_per_channel_default_40" -> "460 conv2d_46"  [label="(576, 96, 1, 1)", style=solid];
-"459 features_11_block_0_0_weight_bias_0_0" -> "460 conv2d_46"  [label="(576,)", style=solid];
-"460 conv2d_46" -> "463 hardswish__15"  [label="(1, 576, 7, 7)", style=solid];
-"461 hardswish__15_scale_0" -> "464 quantize_per_channel_default_41"  [label="(576,)", style=solid];
-"461 hardswish__15_scale_0" -> "465 dequantize_per_channel_default_41"  [label="(576,)", style=solid];
-"462 hardswish__15_zero_point_0" -> "464 quantize_per_channel_default_41"  [label="(576,)", style=solid];
-"462 hardswish__15_zero_point_0" -> "465 dequantize_per_channel_default_41"  [label="(576,)", style=solid];
-"463 hardswish__15" -> "464 quantize_per_channel_default_41"  [label="(1, 576, 7, 7)", style=solid];
-"464 quantize_per_channel_default_41" -> "465 dequantize_per_channel_default_41"  [label="(1, 576, 7, 7)", style=solid];
-"465 dequantize_per_channel_default_41" -> "472 conv2d_47"  [label="(1, 576, 7, 7)", style=solid];
-"466 conv2d_47_scale_0" -> "469 quantize_per_channel_default_42"  [label="(576,)", style=solid];
-"466 conv2d_47_scale_0" -> "470 dequantize_per_channel_default_42"  [label="(576,)", style=solid];
-"467 conv2d_47_zero_point_0" -> "469 quantize_per_channel_default_42"  [label="(576,)", style=solid];
-"467 conv2d_47_zero_point_0" -> "470 dequantize_per_channel_default_42"  [label="(576,)", style=solid];
-"468 compressed_weight_updated_constant31" -> "469 quantize_per_channel_default_42"  [label="(576, 1, 5, 5)", style=solid];
-"469 quantize_per_channel_default_42" -> "470 dequantize_per_channel_default_42"  [label="(576, 1, 5, 5)", style=solid];
-"470 dequantize_per_channel_default_42" -> "472 conv2d_47"  [label="(576, 1, 5, 5)", style=solid];
-"471 features_11_block_1_0_weight_bias_0_0" -> "472 conv2d_47"  [label="(576,)", style=solid];
-"472 conv2d_47" -> "473 hardswish__16"  [label="(1, 576, 7, 7)", style=solid];
-"473 hardswish__16" -> "474 quantize_per_tensor_default_42"  [label="(1, 576, 7, 7)", style=solid];
-"474 quantize_per_tensor_default_42" -> "475 dequantize_per_tensor_default_57"  [label="(1, 576, 7, 7)", style=solid];
-"474 quantize_per_tensor_default_42" -> "476 dequantize_per_tensor_default_56"  [label="(1, 576, 7, 7)", style=solid];
-"475 dequantize_per_tensor_default_57" -> "488 mul_8"  [label="(1, 576, 7, 7)", style=solid];
-"476 dequantize_per_tensor_default_56" -> "477 adaptive_avg_pool2d_8"  [label="(1, 576, 7, 7)", style=solid];
-"477 adaptive_avg_pool2d_8" -> "480 conv2d_48"  [label="(1, 576, 1, 1)", style=solid];
-"478 _param_constant128" -> "480 conv2d_48"  [label="(144, 576, 1, 1)", style=solid];
-"479 _param_constant129_0_0" -> "480 conv2d_48"  [label="(144,)", style=solid];
-"480 conv2d_48" -> "481 relu_8"  [label="(1, 144, 1, 1)", style=solid];
-"481 relu_8" -> "484 conv2d_49"  [label="(1, 144, 1, 1)", style=solid];
-"482 _param_constant130" -> "484 conv2d_49"  [label="(576, 144, 1, 1)", style=solid];
-"483 _param_constant131_0_0" -> "484 conv2d_49"  [label="(576,)", style=solid];
+"0 features_1_block_1_fc1_weight" -> "46 conv2d_2"  [label="(8, 16, 1, 1)", style=solid];
+"1 features_1_block_1_fc2_weight" -> "49 conv2d_3"  [label="(16, 8, 1, 1)", style=solid];
+"2 features_4_block_2_fc1_weight" -> "156 conv2d_13"  [label="(24, 96, 1, 1)", style=solid];
+"3 features_4_block_2_fc2_weight" -> "159 conv2d_14"  [label="(96, 24, 1, 1)", style=solid];
+"4 features_5_block_2_fc1_weight" -> "201 conv2d_18"  [label="(64, 240, 1, 1)", style=solid];
+"5 features_5_block_2_fc2_weight" -> "204 conv2d_19"  [label="(240, 64, 1, 1)", style=solid];
+"6 features_6_block_2_fc1_weight" -> "249 conv2d_23"  [label="(64, 240, 1, 1)", style=solid];
+"7 features_6_block_2_fc2_weight" -> "252 conv2d_24"  [label="(240, 64, 1, 1)", style=solid];
+"8 features_7_block_2_fc1_weight" -> "296 conv2d_28"  [label="(32, 120, 1, 1)", style=solid];
+"9 features_7_block_2_fc2_weight" -> "299 conv2d_29"  [label="(120, 32, 1, 1)", style=solid];
+"10 features_8_block_2_fc1_weight" -> "341 conv2d_33"  [label="(40, 144, 1, 1)", style=solid];
+"11 features_8_block_2_fc2_weight" -> "344 conv2d_34"  [label="(144, 40, 1, 1)", style=solid];
+"12 features_9_block_2_fc1_weight" -> "388 conv2d_38"  [label="(72, 288, 1, 1)", style=solid];
+"13 features_9_block_2_fc2_weight" -> "391 conv2d_39"  [label="(288, 72, 1, 1)", style=solid];
+"14 features_10_block_2_fc1_weight" -> "433 conv2d_43"  [label="(144, 576, 1, 1)", style=solid];
+"15 features_10_block_2_fc2_weight" -> "436 conv2d_44"  [label="(576, 144, 1, 1)", style=solid];
+"16 features_11_block_2_fc1_weight" -> "481 conv2d_48"  [label="(144, 576, 1, 1)", style=solid];
+"17 features_11_block_2_fc2_weight" -> "484 conv2d_49"  [label="(576, 144, 1, 1)", style=solid];
+"18 x" -> "19 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"19 quantize_per_tensor_default" -> "20 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"20 dequantize_per_tensor_default" -> "27 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"21 conv2d_scale_0" -> "24 quantize_per_channel_default"  [label="(16,)", style=solid];
+"21 conv2d_scale_0" -> "25 dequantize_per_channel_default"  [label="(16,)", style=solid];
+"22 conv2d_zero_point_0" -> "24 quantize_per_channel_default"  [label="(16,)", style=solid];
+"22 conv2d_zero_point_0" -> "25 dequantize_per_channel_default"  [label="(16,)", style=solid];
+"23 compressed_weight_updated_constant0" -> "24 quantize_per_channel_default"  [label="(16, 3, 3, 3)", style=solid];
+"24 quantize_per_channel_default" -> "25 dequantize_per_channel_default"  [label="(16, 3, 3, 3)", style=solid];
+"25 dequantize_per_channel_default" -> "27 conv2d"  [label="(16, 3, 3, 3)", style=solid];
+"26 features_0_0_weight_bias_0_0" -> "27 conv2d"  [label="(16,)", style=solid];
+"27 conv2d" -> "30 hardswish_"  [label="(1, 16, 112, 112)", style=solid];
+"28 hardswish__scale_0" -> "31 quantize_per_channel_default_1"  [label="(16,)", style=solid];
+"28 hardswish__scale_0" -> "32 dequantize_per_channel_default_1"  [label="(16,)", style=solid];
+"29 hardswish__zero_point_0" -> "31 quantize_per_channel_default_1"  [label="(16,)", style=solid];
+"29 hardswish__zero_point_0" -> "32 dequantize_per_channel_default_1"  [label="(16,)", style=solid];
+"30 hardswish_" -> "31 quantize_per_channel_default_1"  [label="(1, 16, 112, 112)", style=solid];
+"31 quantize_per_channel_default_1" -> "32 dequantize_per_channel_default_1"  [label="(1, 16, 112, 112)", style=solid];
+"32 dequantize_per_channel_default_1" -> "39 conv2d_1"  [label="(1, 16, 112, 112)", style=solid];
+"33 conv2d_1_scale_0" -> "36 quantize_per_channel_default_2"  [label="(16,)", style=solid];
+"33 conv2d_1_scale_0" -> "37 dequantize_per_channel_default_2"  [label="(16,)", style=solid];
+"34 conv2d_1_zero_point_0" -> "36 quantize_per_channel_default_2"  [label="(16,)", style=solid];
+"34 conv2d_1_zero_point_0" -> "37 dequantize_per_channel_default_2"  [label="(16,)", style=solid];
+"35 compressed_weight_updated_constant1" -> "36 quantize_per_channel_default_2"  [label="(16, 1, 3, 3)", style=solid];
+"36 quantize_per_channel_default_2" -> "37 dequantize_per_channel_default_2"  [label="(16, 1, 3, 3)", style=solid];
+"37 dequantize_per_channel_default_2" -> "39 conv2d_1"  [label="(16, 1, 3, 3)", style=solid];
+"38 features_1_block_0_0_weight_bias_0_0" -> "39 conv2d_1"  [label="(16,)", style=solid];
+"39 conv2d_1" -> "40 relu_"  [label="(1, 16, 56, 56)", style=solid];
+"40 relu_" -> "41 quantize_per_tensor_default_1"  [label="(1, 16, 56, 56)", style=solid];
+"41 quantize_per_tensor_default_1" -> "42 dequantize_per_tensor_default_2"  [label="(1, 16, 56, 56)", style=solid];
+"41 quantize_per_tensor_default_1" -> "43 dequantize_per_tensor_default_1"  [label="(1, 16, 56, 56)", style=solid];
+"42 dequantize_per_tensor_default_2" -> "53 mul"  [label="(1, 16, 56, 56)", style=solid];
+"43 dequantize_per_tensor_default_1" -> "44 adaptive_avg_pool2d"  [label="(1, 16, 56, 56)", style=solid];
+"44 adaptive_avg_pool2d" -> "46 conv2d_2"  [label="(1, 16, 1, 1)", style=solid];
+"45 features_1_block_1_fc1_bias_0_0" -> "46 conv2d_2"  [label="(8,)", style=solid];
+"46 conv2d_2" -> "47 relu"  [label="(1, 8, 1, 1)", style=solid];
+"47 relu" -> "49 conv2d_3"  [label="(1, 8, 1, 1)", style=solid];
+"48 features_1_block_1_fc2_bias_0_0" -> "49 conv2d_3"  [label="(16,)", style=solid];
+"49 conv2d_3" -> "50 hardsigmoid"  [label="(1, 16, 1, 1)", style=solid];
+"50 hardsigmoid" -> "51 quantize_per_tensor_default_2"  [label="(1, 16, 1, 1)", style=solid];
+"51 quantize_per_tensor_default_2" -> "52 dequantize_per_tensor_default_3"  [label="(1, 16, 1, 1)", style=solid];
+"52 dequantize_per_tensor_default_3" -> "53 mul"  [label="(1, 16, 1, 1)", style=solid];
+"53 mul" -> "54 quantize_per_tensor_default_3"  [label="(1, 16, 56, 56)", style=solid];
+"54 quantize_per_tensor_default_3" -> "55 dequantize_per_tensor_default_4"  [label="(1, 16, 56, 56)", style=solid];
+"55 dequantize_per_tensor_default_4" -> "62 conv2d_4"  [label="(1, 16, 56, 56)", style=solid];
+"56 conv2d_4_scale_0" -> "59 quantize_per_channel_default_3"  [label="(16,)", style=solid];
+"56 conv2d_4_scale_0" -> "60 dequantize_per_channel_default_3"  [label="(16,)", style=solid];
+"57 conv2d_4_zero_point_0" -> "59 quantize_per_channel_default_3"  [label="(16,)", style=solid];
+"57 conv2d_4_zero_point_0" -> "60 dequantize_per_channel_default_3"  [label="(16,)", style=solid];
+"58 compressed_weight_updated_constant2" -> "59 quantize_per_channel_default_3"  [label="(16, 16, 1, 1)", style=solid];
+"59 quantize_per_channel_default_3" -> "60 dequantize_per_channel_default_3"  [label="(16, 16, 1, 1)", style=solid];
+"60 dequantize_per_channel_default_3" -> "62 conv2d_4"  [label="(16, 16, 1, 1)", style=solid];
+"61 features_1_block_2_0_weight_bias_0_0" -> "62 conv2d_4"  [label="(16,)", style=solid];
+"62 conv2d_4" -> "63 quantize_per_tensor_default_4"  [label="(1, 16, 56, 56)", style=solid];
+"63 quantize_per_tensor_default_4" -> "64 dequantize_per_tensor_default_5"  [label="(1, 16, 56, 56)", style=solid];
+"64 dequantize_per_tensor_default_5" -> "71 conv2d_5"  [label="(1, 16, 56, 56)", style=solid];
+"65 conv2d_5_scale_0" -> "68 quantize_per_channel_default_4"  [label="(72,)", style=solid];
+"65 conv2d_5_scale_0" -> "69 dequantize_per_channel_default_4"  [label="(72,)", style=solid];
+"66 conv2d_5_zero_point_0" -> "68 quantize_per_channel_default_4"  [label="(72,)", style=solid];
+"66 conv2d_5_zero_point_0" -> "69 dequantize_per_channel_default_4"  [label="(72,)", style=solid];
+"67 compressed_weight_updated_constant3" -> "68 quantize_per_channel_default_4"  [label="(72, 16, 1, 1)", style=solid];
+"68 quantize_per_channel_default_4" -> "69 dequantize_per_channel_default_4"  [label="(72, 16, 1, 1)", style=solid];
+"69 dequantize_per_channel_default_4" -> "71 conv2d_5"  [label="(72, 16, 1, 1)", style=solid];
+"70 features_2_block_0_0_weight_bias_0_0" -> "71 conv2d_5"  [label="(72,)", style=solid];
+"71 conv2d_5" -> "74 relu__1"  [label="(1, 72, 56, 56)", style=solid];
+"72 relu__1_scale_0" -> "75 quantize_per_channel_default_5"  [label="(72,)", style=solid];
+"72 relu__1_scale_0" -> "76 dequantize_per_channel_default_5"  [label="(72,)", style=solid];
+"73 relu__1_zero_point_0" -> "75 quantize_per_channel_default_5"  [label="(72,)", style=solid];
+"73 relu__1_zero_point_0" -> "76 dequantize_per_channel_default_5"  [label="(72,)", style=solid];
+"74 relu__1" -> "75 quantize_per_channel_default_5"  [label="(1, 72, 56, 56)", style=solid];
+"75 quantize_per_channel_default_5" -> "76 dequantize_per_channel_default_5"  [label="(1, 72, 56, 56)", style=solid];
+"76 dequantize_per_channel_default_5" -> "83 conv2d_6"  [label="(1, 72, 56, 56)", style=solid];
+"77 conv2d_6_scale_0" -> "80 quantize_per_channel_default_6"  [label="(72,)", style=solid];
+"77 conv2d_6_scale_0" -> "81 dequantize_per_channel_default_6"  [label="(72,)", style=solid];
+"78 conv2d_6_zero_point_0" -> "80 quantize_per_channel_default_6"  [label="(72,)", style=solid];
+"78 conv2d_6_zero_point_0" -> "81 dequantize_per_channel_default_6"  [label="(72,)", style=solid];
+"79 compressed_weight_updated_constant4" -> "80 quantize_per_channel_default_6"  [label="(72, 1, 3, 3)", style=solid];
+"80 quantize_per_channel_default_6" -> "81 dequantize_per_channel_default_6"  [label="(72, 1, 3, 3)", style=solid];
+"81 dequantize_per_channel_default_6" -> "83 conv2d_6"  [label="(72, 1, 3, 3)", style=solid];
+"82 features_2_block_1_0_weight_bias_0_0" -> "83 conv2d_6"  [label="(72,)", style=solid];
+"83 conv2d_6" -> "84 relu__2"  [label="(1, 72, 28, 28)", style=solid];
+"84 relu__2" -> "85 quantize_per_tensor_default_5"  [label="(1, 72, 28, 28)", style=solid];
+"85 quantize_per_tensor_default_5" -> "86 dequantize_per_tensor_default_6"  [label="(1, 72, 28, 28)", style=solid];
+"86 dequantize_per_tensor_default_6" -> "93 conv2d_7"  [label="(1, 72, 28, 28)", style=solid];
+"87 conv2d_7_scale_0" -> "90 quantize_per_channel_default_7"  [label="(24,)", style=solid];
+"87 conv2d_7_scale_0" -> "91 dequantize_per_channel_default_7"  [label="(24,)", style=solid];
+"88 conv2d_7_zero_point_0" -> "90 quantize_per_channel_default_7"  [label="(24,)", style=solid];
+"88 conv2d_7_zero_point_0" -> "91 dequantize_per_channel_default_7"  [label="(24,)", style=solid];
+"89 compressed_weight_updated_constant5" -> "90 quantize_per_channel_default_7"  [label="(24, 72, 1, 1)", style=solid];
+"90 quantize_per_channel_default_7" -> "91 dequantize_per_channel_default_7"  [label="(24, 72, 1, 1)", style=solid];
+"91 dequantize_per_channel_default_7" -> "93 conv2d_7"  [label="(24, 72, 1, 1)", style=solid];
+"92 features_2_block_2_0_weight_bias_0_0" -> "93 conv2d_7"  [label="(24,)", style=solid];
+"93 conv2d_7" -> "94 quantize_per_tensor_default_6"  [label="(1, 24, 28, 28)", style=solid];
+"94 quantize_per_tensor_default_6" -> "95 dequantize_per_tensor_default_8"  [label="(1, 24, 28, 28)", style=solid];
+"94 quantize_per_tensor_default_6" -> "96 dequantize_per_tensor_default_7"  [label="(1, 24, 28, 28)", style=solid];
+"95 dequantize_per_tensor_default_8" -> "128 add_"  [label="(1, 24, 28, 28)", style=solid];
+"96 dequantize_per_tensor_default_7" -> "103 conv2d_8"  [label="(1, 24, 28, 28)", style=solid];
+"97 conv2d_8_scale_0" -> "100 quantize_per_channel_default_8"  [label="(88,)", style=solid];
+"97 conv2d_8_scale_0" -> "101 dequantize_per_channel_default_8"  [label="(88,)", style=solid];
+"98 conv2d_8_zero_point_0" -> "100 quantize_per_channel_default_8"  [label="(88,)", style=solid];
+"98 conv2d_8_zero_point_0" -> "101 dequantize_per_channel_default_8"  [label="(88,)", style=solid];
+"99 compressed_weight_updated_constant6" -> "100 quantize_per_channel_default_8"  [label="(88, 24, 1, 1)", style=solid];
+"100 quantize_per_channel_default_8" -> "101 dequantize_per_channel_default_8"  [label="(88, 24, 1, 1)", style=solid];
+"101 dequantize_per_channel_default_8" -> "103 conv2d_8"  [label="(88, 24, 1, 1)", style=solid];
+"102 features_3_block_0_0_weight_bias_0_0" -> "103 conv2d_8"  [label="(88,)", style=solid];
+"103 conv2d_8" -> "106 relu__3"  [label="(1, 88, 28, 28)", style=solid];
+"104 relu__3_scale_0" -> "107 quantize_per_channel_default_9"  [label="(88,)", style=solid];
+"104 relu__3_scale_0" -> "108 dequantize_per_channel_default_9"  [label="(88,)", style=solid];
+"105 relu__3_zero_point_0" -> "107 quantize_per_channel_default_9"  [label="(88,)", style=solid];
+"105 relu__3_zero_point_0" -> "108 dequantize_per_channel_default_9"  [label="(88,)", style=solid];
+"106 relu__3" -> "107 quantize_per_channel_default_9"  [label="(1, 88, 28, 28)", style=solid];
+"107 quantize_per_channel_default_9" -> "108 dequantize_per_channel_default_9"  [label="(1, 88, 28, 28)", style=solid];
+"108 dequantize_per_channel_default_9" -> "115 conv2d_9"  [label="(1, 88, 28, 28)", style=solid];
+"109 conv2d_9_scale_0" -> "112 quantize_per_channel_default_10"  [label="(88,)", style=solid];
+"109 conv2d_9_scale_0" -> "113 dequantize_per_channel_default_10"  [label="(88,)", style=solid];
+"110 conv2d_9_zero_point_0" -> "112 quantize_per_channel_default_10"  [label="(88,)", style=solid];
+"110 conv2d_9_zero_point_0" -> "113 dequantize_per_channel_default_10"  [label="(88,)", style=solid];
+"111 compressed_weight_updated_constant7" -> "112 quantize_per_channel_default_10"  [label="(88, 1, 3, 3)", style=solid];
+"112 quantize_per_channel_default_10" -> "113 dequantize_per_channel_default_10"  [label="(88, 1, 3, 3)", style=solid];
+"113 dequantize_per_channel_default_10" -> "115 conv2d_9"  [label="(88, 1, 3, 3)", style=solid];
+"114 features_3_block_1_0_weight_bias_0_0" -> "115 conv2d_9"  [label="(88,)", style=solid];
+"115 conv2d_9" -> "116 relu__4"  [label="(1, 88, 28, 28)", style=solid];
+"116 relu__4" -> "117 quantize_per_tensor_default_7"  [label="(1, 88, 28, 28)", style=solid];
+"117 quantize_per_tensor_default_7" -> "118 dequantize_per_tensor_default_9"  [label="(1, 88, 28, 28)", style=solid];
+"118 dequantize_per_tensor_default_9" -> "125 conv2d_10"  [label="(1, 88, 28, 28)", style=solid];
+"119 conv2d_10_scale_0" -> "122 quantize_per_channel_default_11"  [label="(24,)", style=solid];
+"119 conv2d_10_scale_0" -> "123 dequantize_per_channel_default_11"  [label="(24,)", style=solid];
+"120 conv2d_10_zero_point_0" -> "122 quantize_per_channel_default_11"  [label="(24,)", style=solid];
+"120 conv2d_10_zero_point_0" -> "123 dequantize_per_channel_default_11"  [label="(24,)", style=solid];
+"121 compressed_weight_updated_constant8" -> "122 quantize_per_channel_default_11"  [label="(24, 88, 1, 1)", style=solid];
+"122 quantize_per_channel_default_11" -> "123 dequantize_per_channel_default_11"  [label="(24, 88, 1, 1)", style=solid];
+"123 dequantize_per_channel_default_11" -> "125 conv2d_10"  [label="(24, 88, 1, 1)", style=solid];
+"124 features_3_block_2_0_weight_bias_0_0" -> "125 conv2d_10"  [label="(24,)", style=solid];
+"125 conv2d_10" -> "126 quantize_per_tensor_default_8"  [label="(1, 24, 28, 28)", style=solid];
+"126 quantize_per_tensor_default_8" -> "127 dequantize_per_tensor_default_10"  [label="(1, 24, 28, 28)", style=solid];
+"127 dequantize_per_tensor_default_10" -> "128 add_"  [label="(1, 24, 28, 28)", style=solid];
+"128 add_" -> "129 quantize_per_tensor_default_9"  [label="(1, 24, 28, 28)", style=solid];
+"129 quantize_per_tensor_default_9" -> "130 dequantize_per_tensor_default_11"  [label="(1, 24, 28, 28)", style=solid];
+"130 dequantize_per_tensor_default_11" -> "137 conv2d_11"  [label="(1, 24, 28, 28)", style=solid];
+"131 conv2d_11_scale_0" -> "134 quantize_per_channel_default_12"  [label="(96,)", style=solid];
+"131 conv2d_11_scale_0" -> "135 dequantize_per_channel_default_12"  [label="(96,)", style=solid];
+"132 conv2d_11_zero_point_0" -> "134 quantize_per_channel_default_12"  [label="(96,)", style=solid];
+"132 conv2d_11_zero_point_0" -> "135 dequantize_per_channel_default_12"  [label="(96,)", style=solid];
+"133 compressed_weight_updated_constant9" -> "134 quantize_per_channel_default_12"  [label="(96, 24, 1, 1)", style=solid];
+"134 quantize_per_channel_default_12" -> "135 dequantize_per_channel_default_12"  [label="(96, 24, 1, 1)", style=solid];
+"135 dequantize_per_channel_default_12" -> "137 conv2d_11"  [label="(96, 24, 1, 1)", style=solid];
+"136 features_4_block_0_0_weight_bias_0_0" -> "137 conv2d_11"  [label="(96,)", style=solid];
+"137 conv2d_11" -> "140 hardswish__1"  [label="(1, 96, 28, 28)", style=solid];
+"138 hardswish__1_scale_0" -> "141 quantize_per_channel_default_13"  [label="(96,)", style=solid];
+"138 hardswish__1_scale_0" -> "142 dequantize_per_channel_default_13"  [label="(96,)", style=solid];
+"139 hardswish__1_zero_point_0" -> "141 quantize_per_channel_default_13"  [label="(96,)", style=solid];
+"139 hardswish__1_zero_point_0" -> "142 dequantize_per_channel_default_13"  [label="(96,)", style=solid];
+"140 hardswish__1" -> "141 quantize_per_channel_default_13"  [label="(1, 96, 28, 28)", style=solid];
+"141 quantize_per_channel_default_13" -> "142 dequantize_per_channel_default_13"  [label="(1, 96, 28, 28)", style=solid];
+"142 dequantize_per_channel_default_13" -> "149 conv2d_12"  [label="(1, 96, 28, 28)", style=solid];
+"143 conv2d_12_scale_0" -> "146 quantize_per_channel_default_14"  [label="(96,)", style=solid];
+"143 conv2d_12_scale_0" -> "147 dequantize_per_channel_default_14"  [label="(96,)", style=solid];
+"144 conv2d_12_zero_point_0" -> "146 quantize_per_channel_default_14"  [label="(96,)", style=solid];
+"144 conv2d_12_zero_point_0" -> "147 dequantize_per_channel_default_14"  [label="(96,)", style=solid];
+"145 compressed_weight_updated_constant10" -> "146 quantize_per_channel_default_14"  [label="(96, 1, 5, 5)", style=solid];
+"146 quantize_per_channel_default_14" -> "147 dequantize_per_channel_default_14"  [label="(96, 1, 5, 5)", style=solid];
+"147 dequantize_per_channel_default_14" -> "149 conv2d_12"  [label="(96, 1, 5, 5)", style=solid];
+"148 features_4_block_1_0_weight_bias_0_0" -> "149 conv2d_12"  [label="(96,)", style=solid];
+"149 conv2d_12" -> "150 hardswish__2"  [label="(1, 96, 14, 14)", style=solid];
+"150 hardswish__2" -> "151 quantize_per_tensor_default_10"  [label="(1, 96, 14, 14)", style=solid];
+"151 quantize_per_tensor_default_10" -> "152 dequantize_per_tensor_default_13"  [label="(1, 96, 14, 14)", style=solid];
+"151 quantize_per_tensor_default_10" -> "153 dequantize_per_tensor_default_12"  [label="(1, 96, 14, 14)", style=solid];
+"152 dequantize_per_tensor_default_13" -> "163 mul_1"  [label="(1, 96, 14, 14)", style=solid];
+"153 dequantize_per_tensor_default_12" -> "154 adaptive_avg_pool2d_1"  [label="(1, 96, 14, 14)", style=solid];
+"154 adaptive_avg_pool2d_1" -> "156 conv2d_13"  [label="(1, 96, 1, 1)", style=solid];
+"155 features_4_block_2_fc1_bias_0_0" -> "156 conv2d_13"  [label="(24,)", style=solid];
+"156 conv2d_13" -> "157 relu_1"  [label="(1, 24, 1, 1)", style=solid];
+"157 relu_1" -> "159 conv2d_14"  [label="(1, 24, 1, 1)", style=solid];
+"158 features_4_block_2_fc2_bias_0_0" -> "159 conv2d_14"  [label="(96,)", style=solid];
+"159 conv2d_14" -> "160 hardsigmoid_1"  [label="(1, 96, 1, 1)", style=solid];
+"160 hardsigmoid_1" -> "161 quantize_per_tensor_default_11"  [label="(1, 96, 1, 1)", style=solid];
+"161 quantize_per_tensor_default_11" -> "162 dequantize_per_tensor_default_14"  [label="(1, 96, 1, 1)", style=solid];
+"162 dequantize_per_tensor_default_14" -> "163 mul_1"  [label="(1, 96, 1, 1)", style=solid];
+"163 mul_1" -> "164 quantize_per_tensor_default_12"  [label="(1, 96, 14, 14)", style=solid];
+"164 quantize_per_tensor_default_12" -> "165 dequantize_per_tensor_default_15"  [label="(1, 96, 14, 14)", style=solid];
+"165 dequantize_per_tensor_default_15" -> "172 conv2d_15"  [label="(1, 96, 14, 14)", style=solid];
+"166 conv2d_15_scale_0" -> "169 quantize_per_channel_default_15"  [label="(40,)", style=solid];
+"166 conv2d_15_scale_0" -> "170 dequantize_per_channel_default_15"  [label="(40,)", style=solid];
+"167 conv2d_15_zero_point_0" -> "169 quantize_per_channel_default_15"  [label="(40,)", style=solid];
+"167 conv2d_15_zero_point_0" -> "170 dequantize_per_channel_default_15"  [label="(40,)", style=solid];
+"168 compressed_weight_updated_constant11" -> "169 quantize_per_channel_default_15"  [label="(40, 96, 1, 1)", style=solid];
+"169 quantize_per_channel_default_15" -> "170 dequantize_per_channel_default_15"  [label="(40, 96, 1, 1)", style=solid];
+"170 dequantize_per_channel_default_15" -> "172 conv2d_15"  [label="(40, 96, 1, 1)", style=solid];
+"171 features_4_block_3_0_weight_bias_0_0" -> "172 conv2d_15"  [label="(40,)", style=solid];
+"172 conv2d_15" -> "173 quantize_per_tensor_default_13"  [label="(1, 40, 14, 14)", style=solid];
+"173 quantize_per_tensor_default_13" -> "174 dequantize_per_tensor_default_17"  [label="(1, 40, 14, 14)", style=solid];
+"173 quantize_per_tensor_default_13" -> "175 dequantize_per_tensor_default_16"  [label="(1, 40, 14, 14)", style=solid];
+"174 dequantize_per_tensor_default_17" -> "220 add__1"  [label="(1, 40, 14, 14)", style=solid];
+"175 dequantize_per_tensor_default_16" -> "182 conv2d_16"  [label="(1, 40, 14, 14)", style=solid];
+"176 conv2d_16_scale_0" -> "179 quantize_per_channel_default_16"  [label="(240,)", style=solid];
+"176 conv2d_16_scale_0" -> "180 dequantize_per_channel_default_16"  [label="(240,)", style=solid];
+"177 conv2d_16_zero_point_0" -> "179 quantize_per_channel_default_16"  [label="(240,)", style=solid];
+"177 conv2d_16_zero_point_0" -> "180 dequantize_per_channel_default_16"  [label="(240,)", style=solid];
+"178 compressed_weight_updated_constant12" -> "179 quantize_per_channel_default_16"  [label="(240, 40, 1, 1)", style=solid];
+"179 quantize_per_channel_default_16" -> "180 dequantize_per_channel_default_16"  [label="(240, 40, 1, 1)", style=solid];
+"180 dequantize_per_channel_default_16" -> "182 conv2d_16"  [label="(240, 40, 1, 1)", style=solid];
+"181 features_5_block_0_0_weight_bias_0_0" -> "182 conv2d_16"  [label="(240,)", style=solid];
+"182 conv2d_16" -> "185 hardswish__3"  [label="(1, 240, 14, 14)", style=solid];
+"183 hardswish__3_scale_0" -> "186 quantize_per_channel_default_17"  [label="(240,)", style=solid];
+"183 hardswish__3_scale_0" -> "187 dequantize_per_channel_default_17"  [label="(240,)", style=solid];
+"184 hardswish__3_zero_point_0" -> "186 quantize_per_channel_default_17"  [label="(240,)", style=solid];
+"184 hardswish__3_zero_point_0" -> "187 dequantize_per_channel_default_17"  [label="(240,)", style=solid];
+"185 hardswish__3" -> "186 quantize_per_channel_default_17"  [label="(1, 240, 14, 14)", style=solid];
+"186 quantize_per_channel_default_17" -> "187 dequantize_per_channel_default_17"  [label="(1, 240, 14, 14)", style=solid];
+"187 dequantize_per_channel_default_17" -> "194 conv2d_17"  [label="(1, 240, 14, 14)", style=solid];
+"188 conv2d_17_scale_0" -> "191 quantize_per_channel_default_18"  [label="(240,)", style=solid];
+"188 conv2d_17_scale_0" -> "192 dequantize_per_channel_default_18"  [label="(240,)", style=solid];
+"189 conv2d_17_zero_point_0" -> "191 quantize_per_channel_default_18"  [label="(240,)", style=solid];
+"189 conv2d_17_zero_point_0" -> "192 dequantize_per_channel_default_18"  [label="(240,)", style=solid];
+"190 compressed_weight_updated_constant13" -> "191 quantize_per_channel_default_18"  [label="(240, 1, 5, 5)", style=solid];
+"191 quantize_per_channel_default_18" -> "192 dequantize_per_channel_default_18"  [label="(240, 1, 5, 5)", style=solid];
+"192 dequantize_per_channel_default_18" -> "194 conv2d_17"  [label="(240, 1, 5, 5)", style=solid];
+"193 features_5_block_1_0_weight_bias_0_0" -> "194 conv2d_17"  [label="(240,)", style=solid];
+"194 conv2d_17" -> "195 hardswish__4"  [label="(1, 240, 14, 14)", style=solid];
+"195 hardswish__4" -> "196 quantize_per_tensor_default_14"  [label="(1, 240, 14, 14)", style=solid];
+"196 quantize_per_tensor_default_14" -> "197 dequantize_per_tensor_default_19"  [label="(1, 240, 14, 14)", style=solid];
+"196 quantize_per_tensor_default_14" -> "198 dequantize_per_tensor_default_18"  [label="(1, 240, 14, 14)", style=solid];
+"197 dequantize_per_tensor_default_19" -> "208 mul_2"  [label="(1, 240, 14, 14)", style=solid];
+"198 dequantize_per_tensor_default_18" -> "199 adaptive_avg_pool2d_2"  [label="(1, 240, 14, 14)", style=solid];
+"199 adaptive_avg_pool2d_2" -> "201 conv2d_18"  [label="(1, 240, 1, 1)", style=solid];
+"200 features_5_block_2_fc1_bias_0_0" -> "201 conv2d_18"  [label="(64,)", style=solid];
+"201 conv2d_18" -> "202 relu_2"  [label="(1, 64, 1, 1)", style=solid];
+"202 relu_2" -> "204 conv2d_19"  [label="(1, 64, 1, 1)", style=solid];
+"203 features_5_block_2_fc2_bias_0_0" -> "204 conv2d_19"  [label="(240,)", style=solid];
+"204 conv2d_19" -> "205 hardsigmoid_2"  [label="(1, 240, 1, 1)", style=solid];
+"205 hardsigmoid_2" -> "206 quantize_per_tensor_default_15"  [label="(1, 240, 1, 1)", style=solid];
+"206 quantize_per_tensor_default_15" -> "207 dequantize_per_tensor_default_20"  [label="(1, 240, 1, 1)", style=solid];
+"207 dequantize_per_tensor_default_20" -> "208 mul_2"  [label="(1, 240, 1, 1)", style=solid];
+"208 mul_2" -> "209 quantize_per_tensor_default_16"  [label="(1, 240, 14, 14)", style=solid];
+"209 quantize_per_tensor_default_16" -> "210 dequantize_per_tensor_default_21"  [label="(1, 240, 14, 14)", style=solid];
+"210 dequantize_per_tensor_default_21" -> "217 conv2d_20"  [label="(1, 240, 14, 14)", style=solid];
+"211 conv2d_20_scale_0" -> "214 quantize_per_channel_default_19"  [label="(40,)", style=solid];
+"211 conv2d_20_scale_0" -> "215 dequantize_per_channel_default_19"  [label="(40,)", style=solid];
+"212 conv2d_20_zero_point_0" -> "214 quantize_per_channel_default_19"  [label="(40,)", style=solid];
+"212 conv2d_20_zero_point_0" -> "215 dequantize_per_channel_default_19"  [label="(40,)", style=solid];
+"213 compressed_weight_updated_constant14" -> "214 quantize_per_channel_default_19"  [label="(40, 240, 1, 1)", style=solid];
+"214 quantize_per_channel_default_19" -> "215 dequantize_per_channel_default_19"  [label="(40, 240, 1, 1)", style=solid];
+"215 dequantize_per_channel_default_19" -> "217 conv2d_20"  [label="(40, 240, 1, 1)", style=solid];
+"216 features_5_block_3_0_weight_bias_0_0" -> "217 conv2d_20"  [label="(40,)", style=solid];
+"217 conv2d_20" -> "218 quantize_per_tensor_default_17"  [label="(1, 40, 14, 14)", style=solid];
+"218 quantize_per_tensor_default_17" -> "219 dequantize_per_tensor_default_22"  [label="(1, 40, 14, 14)", style=solid];
+"219 dequantize_per_tensor_default_22" -> "220 add__1"  [label="(1, 40, 14, 14)", style=solid];
+"220 add__1" -> "221 quantize_per_tensor_default_18"  [label="(1, 40, 14, 14)", style=solid];
+"221 quantize_per_tensor_default_18" -> "222 dequantize_per_tensor_default_24"  [label="(1, 40, 14, 14)", style=solid];
+"221 quantize_per_tensor_default_18" -> "223 dequantize_per_tensor_default_23"  [label="(1, 40, 14, 14)", style=solid];
+"222 dequantize_per_tensor_default_24" -> "268 add__2"  [label="(1, 40, 14, 14)", style=solid];
+"223 dequantize_per_tensor_default_23" -> "230 conv2d_21"  [label="(1, 40, 14, 14)", style=solid];
+"224 conv2d_21_scale_0" -> "227 quantize_per_channel_default_20"  [label="(240,)", style=solid];
+"224 conv2d_21_scale_0" -> "228 dequantize_per_channel_default_20"  [label="(240,)", style=solid];
+"225 conv2d_21_zero_point_0" -> "227 quantize_per_channel_default_20"  [label="(240,)", style=solid];
+"225 conv2d_21_zero_point_0" -> "228 dequantize_per_channel_default_20"  [label="(240,)", style=solid];
+"226 compressed_weight_updated_constant15" -> "227 quantize_per_channel_default_20"  [label="(240, 40, 1, 1)", style=solid];
+"227 quantize_per_channel_default_20" -> "228 dequantize_per_channel_default_20"  [label="(240, 40, 1, 1)", style=solid];
+"228 dequantize_per_channel_default_20" -> "230 conv2d_21"  [label="(240, 40, 1, 1)", style=solid];
+"229 features_6_block_0_0_weight_bias_0_0" -> "230 conv2d_21"  [label="(240,)", style=solid];
+"230 conv2d_21" -> "233 hardswish__5"  [label="(1, 240, 14, 14)", style=solid];
+"231 hardswish__5_scale_0" -> "234 quantize_per_channel_default_21"  [label="(240,)", style=solid];
+"231 hardswish__5_scale_0" -> "235 dequantize_per_channel_default_21"  [label="(240,)", style=solid];
+"232 hardswish__5_zero_point_0" -> "234 quantize_per_channel_default_21"  [label="(240,)", style=solid];
+"232 hardswish__5_zero_point_0" -> "235 dequantize_per_channel_default_21"  [label="(240,)", style=solid];
+"233 hardswish__5" -> "234 quantize_per_channel_default_21"  [label="(1, 240, 14, 14)", style=solid];
+"234 quantize_per_channel_default_21" -> "235 dequantize_per_channel_default_21"  [label="(1, 240, 14, 14)", style=solid];
+"235 dequantize_per_channel_default_21" -> "242 conv2d_22"  [label="(1, 240, 14, 14)", style=solid];
+"236 conv2d_22_scale_0" -> "239 quantize_per_channel_default_22"  [label="(240,)", style=solid];
+"236 conv2d_22_scale_0" -> "240 dequantize_per_channel_default_22"  [label="(240,)", style=solid];
+"237 conv2d_22_zero_point_0" -> "239 quantize_per_channel_default_22"  [label="(240,)", style=solid];
+"237 conv2d_22_zero_point_0" -> "240 dequantize_per_channel_default_22"  [label="(240,)", style=solid];
+"238 compressed_weight_updated_constant16" -> "239 quantize_per_channel_default_22"  [label="(240, 1, 5, 5)", style=solid];
+"239 quantize_per_channel_default_22" -> "240 dequantize_per_channel_default_22"  [label="(240, 1, 5, 5)", style=solid];
+"240 dequantize_per_channel_default_22" -> "242 conv2d_22"  [label="(240, 1, 5, 5)", style=solid];
+"241 features_6_block_1_0_weight_bias_0_0" -> "242 conv2d_22"  [label="(240,)", style=solid];
+"242 conv2d_22" -> "243 hardswish__6"  [label="(1, 240, 14, 14)", style=solid];
+"243 hardswish__6" -> "244 quantize_per_tensor_default_19"  [label="(1, 240, 14, 14)", style=solid];
+"244 quantize_per_tensor_default_19" -> "245 dequantize_per_tensor_default_26"  [label="(1, 240, 14, 14)", style=solid];
+"244 quantize_per_tensor_default_19" -> "246 dequantize_per_tensor_default_25"  [label="(1, 240, 14, 14)", style=solid];
+"245 dequantize_per_tensor_default_26" -> "256 mul_3"  [label="(1, 240, 14, 14)", style=solid];
+"246 dequantize_per_tensor_default_25" -> "247 adaptive_avg_pool2d_3"  [label="(1, 240, 14, 14)", style=solid];
+"247 adaptive_avg_pool2d_3" -> "249 conv2d_23"  [label="(1, 240, 1, 1)", style=solid];
+"248 features_6_block_2_fc1_bias_0_0" -> "249 conv2d_23"  [label="(64,)", style=solid];
+"249 conv2d_23" -> "250 relu_3"  [label="(1, 64, 1, 1)", style=solid];
+"250 relu_3" -> "252 conv2d_24"  [label="(1, 64, 1, 1)", style=solid];
+"251 features_6_block_2_fc2_bias_0_0" -> "252 conv2d_24"  [label="(240,)", style=solid];
+"252 conv2d_24" -> "253 hardsigmoid_3"  [label="(1, 240, 1, 1)", style=solid];
+"253 hardsigmoid_3" -> "254 quantize_per_tensor_default_20"  [label="(1, 240, 1, 1)", style=solid];
+"254 quantize_per_tensor_default_20" -> "255 dequantize_per_tensor_default_27"  [label="(1, 240, 1, 1)", style=solid];
+"255 dequantize_per_tensor_default_27" -> "256 mul_3"  [label="(1, 240, 1, 1)", style=solid];
+"256 mul_3" -> "257 quantize_per_tensor_default_21"  [label="(1, 240, 14, 14)", style=solid];
+"257 quantize_per_tensor_default_21" -> "258 dequantize_per_tensor_default_28"  [label="(1, 240, 14, 14)", style=solid];
+"258 dequantize_per_tensor_default_28" -> "265 conv2d_25"  [label="(1, 240, 14, 14)", style=solid];
+"259 conv2d_25_scale_0" -> "262 quantize_per_channel_default_23"  [label="(40,)", style=solid];
+"259 conv2d_25_scale_0" -> "263 dequantize_per_channel_default_23"  [label="(40,)", style=solid];
+"260 conv2d_25_zero_point_0" -> "262 quantize_per_channel_default_23"  [label="(40,)", style=solid];
+"260 conv2d_25_zero_point_0" -> "263 dequantize_per_channel_default_23"  [label="(40,)", style=solid];
+"261 compressed_weight_updated_constant17" -> "262 quantize_per_channel_default_23"  [label="(40, 240, 1, 1)", style=solid];
+"262 quantize_per_channel_default_23" -> "263 dequantize_per_channel_default_23"  [label="(40, 240, 1, 1)", style=solid];
+"263 dequantize_per_channel_default_23" -> "265 conv2d_25"  [label="(40, 240, 1, 1)", style=solid];
+"264 features_6_block_3_0_weight_bias_0_0" -> "265 conv2d_25"  [label="(40,)", style=solid];
+"265 conv2d_25" -> "266 quantize_per_tensor_default_22"  [label="(1, 40, 14, 14)", style=solid];
+"266 quantize_per_tensor_default_22" -> "267 dequantize_per_tensor_default_29"  [label="(1, 40, 14, 14)", style=solid];
+"267 dequantize_per_tensor_default_29" -> "268 add__2"  [label="(1, 40, 14, 14)", style=solid];
+"268 add__2" -> "269 quantize_per_tensor_default_23"  [label="(1, 40, 14, 14)", style=solid];
+"269 quantize_per_tensor_default_23" -> "270 dequantize_per_tensor_default_30"  [label="(1, 40, 14, 14)", style=solid];
+"270 dequantize_per_tensor_default_30" -> "277 conv2d_26"  [label="(1, 40, 14, 14)", style=solid];
+"271 conv2d_26_scale_0" -> "274 quantize_per_channel_default_24"  [label="(120,)", style=solid];
+"271 conv2d_26_scale_0" -> "275 dequantize_per_channel_default_24"  [label="(120,)", style=solid];
+"272 conv2d_26_zero_point_0" -> "274 quantize_per_channel_default_24"  [label="(120,)", style=solid];
+"272 conv2d_26_zero_point_0" -> "275 dequantize_per_channel_default_24"  [label="(120,)", style=solid];
+"273 compressed_weight_updated_constant18" -> "274 quantize_per_channel_default_24"  [label="(120, 40, 1, 1)", style=solid];
+"274 quantize_per_channel_default_24" -> "275 dequantize_per_channel_default_24"  [label="(120, 40, 1, 1)", style=solid];
+"275 dequantize_per_channel_default_24" -> "277 conv2d_26"  [label="(120, 40, 1, 1)", style=solid];
+"276 features_7_block_0_0_weight_bias_0_0" -> "277 conv2d_26"  [label="(120,)", style=solid];
+"277 conv2d_26" -> "280 hardswish__7"  [label="(1, 120, 14, 14)", style=solid];
+"278 hardswish__7_scale_0" -> "281 quantize_per_channel_default_25"  [label="(120,)", style=solid];
+"278 hardswish__7_scale_0" -> "282 dequantize_per_channel_default_25"  [label="(120,)", style=solid];
+"279 hardswish__7_zero_point_0" -> "281 quantize_per_channel_default_25"  [label="(120,)", style=solid];
+"279 hardswish__7_zero_point_0" -> "282 dequantize_per_channel_default_25"  [label="(120,)", style=solid];
+"280 hardswish__7" -> "281 quantize_per_channel_default_25"  [label="(1, 120, 14, 14)", style=solid];
+"281 quantize_per_channel_default_25" -> "282 dequantize_per_channel_default_25"  [label="(1, 120, 14, 14)", style=solid];
+"282 dequantize_per_channel_default_25" -> "289 conv2d_27"  [label="(1, 120, 14, 14)", style=solid];
+"283 conv2d_27_scale_0" -> "286 quantize_per_channel_default_26"  [label="(120,)", style=solid];
+"283 conv2d_27_scale_0" -> "287 dequantize_per_channel_default_26"  [label="(120,)", style=solid];
+"284 conv2d_27_zero_point_0" -> "286 quantize_per_channel_default_26"  [label="(120,)", style=solid];
+"284 conv2d_27_zero_point_0" -> "287 dequantize_per_channel_default_26"  [label="(120,)", style=solid];
+"285 compressed_weight_updated_constant19" -> "286 quantize_per_channel_default_26"  [label="(120, 1, 5, 5)", style=solid];
+"286 quantize_per_channel_default_26" -> "287 dequantize_per_channel_default_26"  [label="(120, 1, 5, 5)", style=solid];
+"287 dequantize_per_channel_default_26" -> "289 conv2d_27"  [label="(120, 1, 5, 5)", style=solid];
+"288 features_7_block_1_0_weight_bias_0_0" -> "289 conv2d_27"  [label="(120,)", style=solid];
+"289 conv2d_27" -> "290 hardswish__8"  [label="(1, 120, 14, 14)", style=solid];
+"290 hardswish__8" -> "291 quantize_per_tensor_default_24"  [label="(1, 120, 14, 14)", style=solid];
+"291 quantize_per_tensor_default_24" -> "292 dequantize_per_tensor_default_32"  [label="(1, 120, 14, 14)", style=solid];
+"291 quantize_per_tensor_default_24" -> "293 dequantize_per_tensor_default_31"  [label="(1, 120, 14, 14)", style=solid];
+"292 dequantize_per_tensor_default_32" -> "303 mul_4"  [label="(1, 120, 14, 14)", style=solid];
+"293 dequantize_per_tensor_default_31" -> "294 adaptive_avg_pool2d_4"  [label="(1, 120, 14, 14)", style=solid];
+"294 adaptive_avg_pool2d_4" -> "296 conv2d_28"  [label="(1, 120, 1, 1)", style=solid];
+"295 features_7_block_2_fc1_bias_0_0" -> "296 conv2d_28"  [label="(32,)", style=solid];
+"296 conv2d_28" -> "297 relu_4"  [label="(1, 32, 1, 1)", style=solid];
+"297 relu_4" -> "299 conv2d_29"  [label="(1, 32, 1, 1)", style=solid];
+"298 features_7_block_2_fc2_bias_0_0" -> "299 conv2d_29"  [label="(120,)", style=solid];
+"299 conv2d_29" -> "300 hardsigmoid_4"  [label="(1, 120, 1, 1)", style=solid];
+"300 hardsigmoid_4" -> "301 quantize_per_tensor_default_25"  [label="(1, 120, 1, 1)", style=solid];
+"301 quantize_per_tensor_default_25" -> "302 dequantize_per_tensor_default_33"  [label="(1, 120, 1, 1)", style=solid];
+"302 dequantize_per_tensor_default_33" -> "303 mul_4"  [label="(1, 120, 1, 1)", style=solid];
+"303 mul_4" -> "304 quantize_per_tensor_default_26"  [label="(1, 120, 14, 14)", style=solid];
+"304 quantize_per_tensor_default_26" -> "305 dequantize_per_tensor_default_34"  [label="(1, 120, 14, 14)", style=solid];
+"305 dequantize_per_tensor_default_34" -> "312 conv2d_30"  [label="(1, 120, 14, 14)", style=solid];
+"306 conv2d_30_scale_0" -> "309 quantize_per_channel_default_27"  [label="(48,)", style=solid];
+"306 conv2d_30_scale_0" -> "310 dequantize_per_channel_default_27"  [label="(48,)", style=solid];
+"307 conv2d_30_zero_point_0" -> "309 quantize_per_channel_default_27"  [label="(48,)", style=solid];
+"307 conv2d_30_zero_point_0" -> "310 dequantize_per_channel_default_27"  [label="(48,)", style=solid];
+"308 compressed_weight_updated_constant20" -> "309 quantize_per_channel_default_27"  [label="(48, 120, 1, 1)", style=solid];
+"309 quantize_per_channel_default_27" -> "310 dequantize_per_channel_default_27"  [label="(48, 120, 1, 1)", style=solid];
+"310 dequantize_per_channel_default_27" -> "312 conv2d_30"  [label="(48, 120, 1, 1)", style=solid];
+"311 features_7_block_3_0_weight_bias_0_0" -> "312 conv2d_30"  [label="(48,)", style=solid];
+"312 conv2d_30" -> "313 quantize_per_tensor_default_27"  [label="(1, 48, 14, 14)", style=solid];
+"313 quantize_per_tensor_default_27" -> "314 dequantize_per_tensor_default_36"  [label="(1, 48, 14, 14)", style=solid];
+"313 quantize_per_tensor_default_27" -> "315 dequantize_per_tensor_default_35"  [label="(1, 48, 14, 14)", style=solid];
+"314 dequantize_per_tensor_default_36" -> "360 add__3"  [label="(1, 48, 14, 14)", style=solid];
+"315 dequantize_per_tensor_default_35" -> "322 conv2d_31"  [label="(1, 48, 14, 14)", style=solid];
+"316 conv2d_31_scale_0" -> "319 quantize_per_channel_default_28"  [label="(144,)", style=solid];
+"316 conv2d_31_scale_0" -> "320 dequantize_per_channel_default_28"  [label="(144,)", style=solid];
+"317 conv2d_31_zero_point_0" -> "319 quantize_per_channel_default_28"  [label="(144,)", style=solid];
+"317 conv2d_31_zero_point_0" -> "320 dequantize_per_channel_default_28"  [label="(144,)", style=solid];
+"318 compressed_weight_updated_constant21" -> "319 quantize_per_channel_default_28"  [label="(144, 48, 1, 1)", style=solid];
+"319 quantize_per_channel_default_28" -> "320 dequantize_per_channel_default_28"  [label="(144, 48, 1, 1)", style=solid];
+"320 dequantize_per_channel_default_28" -> "322 conv2d_31"  [label="(144, 48, 1, 1)", style=solid];
+"321 features_8_block_0_0_weight_bias_0_0" -> "322 conv2d_31"  [label="(144,)", style=solid];
+"322 conv2d_31" -> "325 hardswish__9"  [label="(1, 144, 14, 14)", style=solid];
+"323 hardswish__9_scale_0" -> "326 quantize_per_channel_default_29"  [label="(144,)", style=solid];
+"323 hardswish__9_scale_0" -> "327 dequantize_per_channel_default_29"  [label="(144,)", style=solid];
+"324 hardswish__9_zero_point_0" -> "326 quantize_per_channel_default_29"  [label="(144,)", style=solid];
+"324 hardswish__9_zero_point_0" -> "327 dequantize_per_channel_default_29"  [label="(144,)", style=solid];
+"325 hardswish__9" -> "326 quantize_per_channel_default_29"  [label="(1, 144, 14, 14)", style=solid];
+"326 quantize_per_channel_default_29" -> "327 dequantize_per_channel_default_29"  [label="(1, 144, 14, 14)", style=solid];
+"327 dequantize_per_channel_default_29" -> "334 conv2d_32"  [label="(1, 144, 14, 14)", style=solid];
+"328 conv2d_32_scale_0" -> "331 quantize_per_channel_default_30"  [label="(144,)", style=solid];
+"328 conv2d_32_scale_0" -> "332 dequantize_per_channel_default_30"  [label="(144,)", style=solid];
+"329 conv2d_32_zero_point_0" -> "331 quantize_per_channel_default_30"  [label="(144,)", style=solid];
+"329 conv2d_32_zero_point_0" -> "332 dequantize_per_channel_default_30"  [label="(144,)", style=solid];
+"330 compressed_weight_updated_constant22" -> "331 quantize_per_channel_default_30"  [label="(144, 1, 5, 5)", style=solid];
+"331 quantize_per_channel_default_30" -> "332 dequantize_per_channel_default_30"  [label="(144, 1, 5, 5)", style=solid];
+"332 dequantize_per_channel_default_30" -> "334 conv2d_32"  [label="(144, 1, 5, 5)", style=solid];
+"333 features_8_block_1_0_weight_bias_0_0" -> "334 conv2d_32"  [label="(144,)", style=solid];
+"334 conv2d_32" -> "335 hardswish__10"  [label="(1, 144, 14, 14)", style=solid];
+"335 hardswish__10" -> "336 quantize_per_tensor_default_28"  [label="(1, 144, 14, 14)", style=solid];
+"336 quantize_per_tensor_default_28" -> "337 dequantize_per_tensor_default_38"  [label="(1, 144, 14, 14)", style=solid];
+"336 quantize_per_tensor_default_28" -> "338 dequantize_per_tensor_default_37"  [label="(1, 144, 14, 14)", style=solid];
+"337 dequantize_per_tensor_default_38" -> "348 mul_5"  [label="(1, 144, 14, 14)", style=solid];
+"338 dequantize_per_tensor_default_37" -> "339 adaptive_avg_pool2d_5"  [label="(1, 144, 14, 14)", style=solid];
+"339 adaptive_avg_pool2d_5" -> "341 conv2d_33"  [label="(1, 144, 1, 1)", style=solid];
+"340 features_8_block_2_fc1_bias_0_0" -> "341 conv2d_33"  [label="(40,)", style=solid];
+"341 conv2d_33" -> "342 relu_5"  [label="(1, 40, 1, 1)", style=solid];
+"342 relu_5" -> "344 conv2d_34"  [label="(1, 40, 1, 1)", style=solid];
+"343 features_8_block_2_fc2_bias_0_0" -> "344 conv2d_34"  [label="(144,)", style=solid];
+"344 conv2d_34" -> "345 hardsigmoid_5"  [label="(1, 144, 1, 1)", style=solid];
+"345 hardsigmoid_5" -> "346 quantize_per_tensor_default_29"  [label="(1, 144, 1, 1)", style=solid];
+"346 quantize_per_tensor_default_29" -> "347 dequantize_per_tensor_default_39"  [label="(1, 144, 1, 1)", style=solid];
+"347 dequantize_per_tensor_default_39" -> "348 mul_5"  [label="(1, 144, 1, 1)", style=solid];
+"348 mul_5" -> "349 quantize_per_tensor_default_30"  [label="(1, 144, 14, 14)", style=solid];
+"349 quantize_per_tensor_default_30" -> "350 dequantize_per_tensor_default_40"  [label="(1, 144, 14, 14)", style=solid];
+"350 dequantize_per_tensor_default_40" -> "357 conv2d_35"  [label="(1, 144, 14, 14)", style=solid];
+"351 conv2d_35_scale_0" -> "354 quantize_per_channel_default_31"  [label="(48,)", style=solid];
+"351 conv2d_35_scale_0" -> "355 dequantize_per_channel_default_31"  [label="(48,)", style=solid];
+"352 conv2d_35_zero_point_0" -> "354 quantize_per_channel_default_31"  [label="(48,)", style=solid];
+"352 conv2d_35_zero_point_0" -> "355 dequantize_per_channel_default_31"  [label="(48,)", style=solid];
+"353 compressed_weight_updated_constant23" -> "354 quantize_per_channel_default_31"  [label="(48, 144, 1, 1)", style=solid];
+"354 quantize_per_channel_default_31" -> "355 dequantize_per_channel_default_31"  [label="(48, 144, 1, 1)", style=solid];
+"355 dequantize_per_channel_default_31" -> "357 conv2d_35"  [label="(48, 144, 1, 1)", style=solid];
+"356 features_8_block_3_0_weight_bias_0_0" -> "357 conv2d_35"  [label="(48,)", style=solid];
+"357 conv2d_35" -> "358 quantize_per_tensor_default_31"  [label="(1, 48, 14, 14)", style=solid];
+"358 quantize_per_tensor_default_31" -> "359 dequantize_per_tensor_default_41"  [label="(1, 48, 14, 14)", style=solid];
+"359 dequantize_per_tensor_default_41" -> "360 add__3"  [label="(1, 48, 14, 14)", style=solid];
+"360 add__3" -> "361 quantize_per_tensor_default_32"  [label="(1, 48, 14, 14)", style=solid];
+"361 quantize_per_tensor_default_32" -> "362 dequantize_per_tensor_default_42"  [label="(1, 48, 14, 14)", style=solid];
+"362 dequantize_per_tensor_default_42" -> "369 conv2d_36"  [label="(1, 48, 14, 14)", style=solid];
+"363 conv2d_36_scale_0" -> "366 quantize_per_channel_default_32"  [label="(288,)", style=solid];
+"363 conv2d_36_scale_0" -> "367 dequantize_per_channel_default_32"  [label="(288,)", style=solid];
+"364 conv2d_36_zero_point_0" -> "366 quantize_per_channel_default_32"  [label="(288,)", style=solid];
+"364 conv2d_36_zero_point_0" -> "367 dequantize_per_channel_default_32"  [label="(288,)", style=solid];
+"365 compressed_weight_updated_constant24" -> "366 quantize_per_channel_default_32"  [label="(288, 48, 1, 1)", style=solid];
+"366 quantize_per_channel_default_32" -> "367 dequantize_per_channel_default_32"  [label="(288, 48, 1, 1)", style=solid];
+"367 dequantize_per_channel_default_32" -> "369 conv2d_36"  [label="(288, 48, 1, 1)", style=solid];
+"368 features_9_block_0_0_weight_bias_0_0" -> "369 conv2d_36"  [label="(288,)", style=solid];
+"369 conv2d_36" -> "372 hardswish__11"  [label="(1, 288, 14, 14)", style=solid];
+"370 hardswish__11_scale_0" -> "373 quantize_per_channel_default_33"  [label="(288,)", style=solid];
+"370 hardswish__11_scale_0" -> "374 dequantize_per_channel_default_33"  [label="(288,)", style=solid];
+"371 hardswish__11_zero_point_0" -> "373 quantize_per_channel_default_33"  [label="(288,)", style=solid];
+"371 hardswish__11_zero_point_0" -> "374 dequantize_per_channel_default_33"  [label="(288,)", style=solid];
+"372 hardswish__11" -> "373 quantize_per_channel_default_33"  [label="(1, 288, 14, 14)", style=solid];
+"373 quantize_per_channel_default_33" -> "374 dequantize_per_channel_default_33"  [label="(1, 288, 14, 14)", style=solid];
+"374 dequantize_per_channel_default_33" -> "381 conv2d_37"  [label="(1, 288, 14, 14)", style=solid];
+"375 conv2d_37_scale_0" -> "378 quantize_per_channel_default_34"  [label="(288,)", style=solid];
+"375 conv2d_37_scale_0" -> "379 dequantize_per_channel_default_34"  [label="(288,)", style=solid];
+"376 conv2d_37_zero_point_0" -> "378 quantize_per_channel_default_34"  [label="(288,)", style=solid];
+"376 conv2d_37_zero_point_0" -> "379 dequantize_per_channel_default_34"  [label="(288,)", style=solid];
+"377 compressed_weight_updated_constant25" -> "378 quantize_per_channel_default_34"  [label="(288, 1, 5, 5)", style=solid];
+"378 quantize_per_channel_default_34" -> "379 dequantize_per_channel_default_34"  [label="(288, 1, 5, 5)", style=solid];
+"379 dequantize_per_channel_default_34" -> "381 conv2d_37"  [label="(288, 1, 5, 5)", style=solid];
+"380 features_9_block_1_0_weight_bias_0_0" -> "381 conv2d_37"  [label="(288,)", style=solid];
+"381 conv2d_37" -> "382 hardswish__12"  [label="(1, 288, 7, 7)", style=solid];
+"382 hardswish__12" -> "383 quantize_per_tensor_default_33"  [label="(1, 288, 7, 7)", style=solid];
+"383 quantize_per_tensor_default_33" -> "384 dequantize_per_tensor_default_44"  [label="(1, 288, 7, 7)", style=solid];
+"383 quantize_per_tensor_default_33" -> "385 dequantize_per_tensor_default_43"  [label="(1, 288, 7, 7)", style=solid];
+"384 dequantize_per_tensor_default_44" -> "395 mul_6"  [label="(1, 288, 7, 7)", style=solid];
+"385 dequantize_per_tensor_default_43" -> "386 adaptive_avg_pool2d_6"  [label="(1, 288, 7, 7)", style=solid];
+"386 adaptive_avg_pool2d_6" -> "388 conv2d_38"  [label="(1, 288, 1, 1)", style=solid];
+"387 features_9_block_2_fc1_bias_0_0" -> "388 conv2d_38"  [label="(72,)", style=solid];
+"388 conv2d_38" -> "389 relu_6"  [label="(1, 72, 1, 1)", style=solid];
+"389 relu_6" -> "391 conv2d_39"  [label="(1, 72, 1, 1)", style=solid];
+"390 features_9_block_2_fc2_bias_0_0" -> "391 conv2d_39"  [label="(288,)", style=solid];
+"391 conv2d_39" -> "392 hardsigmoid_6"  [label="(1, 288, 1, 1)", style=solid];
+"392 hardsigmoid_6" -> "393 quantize_per_tensor_default_34"  [label="(1, 288, 1, 1)", style=solid];
+"393 quantize_per_tensor_default_34" -> "394 dequantize_per_tensor_default_45"  [label="(1, 288, 1, 1)", style=solid];
+"394 dequantize_per_tensor_default_45" -> "395 mul_6"  [label="(1, 288, 1, 1)", style=solid];
+"395 mul_6" -> "396 quantize_per_tensor_default_35"  [label="(1, 288, 7, 7)", style=solid];
+"396 quantize_per_tensor_default_35" -> "397 dequantize_per_tensor_default_46"  [label="(1, 288, 7, 7)", style=solid];
+"397 dequantize_per_tensor_default_46" -> "404 conv2d_40"  [label="(1, 288, 7, 7)", style=solid];
+"398 conv2d_40_scale_0" -> "401 quantize_per_channel_default_35"  [label="(96,)", style=solid];
+"398 conv2d_40_scale_0" -> "402 dequantize_per_channel_default_35"  [label="(96,)", style=solid];
+"399 conv2d_40_zero_point_0" -> "401 quantize_per_channel_default_35"  [label="(96,)", style=solid];
+"399 conv2d_40_zero_point_0" -> "402 dequantize_per_channel_default_35"  [label="(96,)", style=solid];
+"400 compressed_weight_updated_constant26" -> "401 quantize_per_channel_default_35"  [label="(96, 288, 1, 1)", style=solid];
+"401 quantize_per_channel_default_35" -> "402 dequantize_per_channel_default_35"  [label="(96, 288, 1, 1)", style=solid];
+"402 dequantize_per_channel_default_35" -> "404 conv2d_40"  [label="(96, 288, 1, 1)", style=solid];
+"403 features_9_block_3_0_weight_bias_0_0" -> "404 conv2d_40"  [label="(96,)", style=solid];
+"404 conv2d_40" -> "405 quantize_per_tensor_default_36"  [label="(1, 96, 7, 7)", style=solid];
+"405 quantize_per_tensor_default_36" -> "406 dequantize_per_tensor_default_48"  [label="(1, 96, 7, 7)", style=solid];
+"405 quantize_per_tensor_default_36" -> "407 dequantize_per_tensor_default_47"  [label="(1, 96, 7, 7)", style=solid];
+"406 dequantize_per_tensor_default_48" -> "452 add__4"  [label="(1, 96, 7, 7)", style=solid];
+"407 dequantize_per_tensor_default_47" -> "414 conv2d_41"  [label="(1, 96, 7, 7)", style=solid];
+"408 conv2d_41_scale_0" -> "411 quantize_per_channel_default_36"  [label="(576,)", style=solid];
+"408 conv2d_41_scale_0" -> "412 dequantize_per_channel_default_36"  [label="(576,)", style=solid];
+"409 conv2d_41_zero_point_0" -> "411 quantize_per_channel_default_36"  [label="(576,)", style=solid];
+"409 conv2d_41_zero_point_0" -> "412 dequantize_per_channel_default_36"  [label="(576,)", style=solid];
+"410 compressed_weight_updated_constant27" -> "411 quantize_per_channel_default_36"  [label="(576, 96, 1, 1)", style=solid];
+"411 quantize_per_channel_default_36" -> "412 dequantize_per_channel_default_36"  [label="(576, 96, 1, 1)", style=solid];
+"412 dequantize_per_channel_default_36" -> "414 conv2d_41"  [label="(576, 96, 1, 1)", style=solid];
+"413 features_10_block_0_0_weight_bias_0_0" -> "414 conv2d_41"  [label="(576,)", style=solid];
+"414 conv2d_41" -> "417 hardswish__13"  [label="(1, 576, 7, 7)", style=solid];
+"415 hardswish__13_scale_0" -> "418 quantize_per_channel_default_37"  [label="(576,)", style=solid];
+"415 hardswish__13_scale_0" -> "419 dequantize_per_channel_default_37"  [label="(576,)", style=solid];
+"416 hardswish__13_zero_point_0" -> "418 quantize_per_channel_default_37"  [label="(576,)", style=solid];
+"416 hardswish__13_zero_point_0" -> "419 dequantize_per_channel_default_37"  [label="(576,)", style=solid];
+"417 hardswish__13" -> "418 quantize_per_channel_default_37"  [label="(1, 576, 7, 7)", style=solid];
+"418 quantize_per_channel_default_37" -> "419 dequantize_per_channel_default_37"  [label="(1, 576, 7, 7)", style=solid];
+"419 dequantize_per_channel_default_37" -> "426 conv2d_42"  [label="(1, 576, 7, 7)", style=solid];
+"420 conv2d_42_scale_0" -> "423 quantize_per_channel_default_38"  [label="(576,)", style=solid];
+"420 conv2d_42_scale_0" -> "424 dequantize_per_channel_default_38"  [label="(576,)", style=solid];
+"421 conv2d_42_zero_point_0" -> "423 quantize_per_channel_default_38"  [label="(576,)", style=solid];
+"421 conv2d_42_zero_point_0" -> "424 dequantize_per_channel_default_38"  [label="(576,)", style=solid];
+"422 compressed_weight_updated_constant28" -> "423 quantize_per_channel_default_38"  [label="(576, 1, 5, 5)", style=solid];
+"423 quantize_per_channel_default_38" -> "424 dequantize_per_channel_default_38"  [label="(576, 1, 5, 5)", style=solid];
+"424 dequantize_per_channel_default_38" -> "426 conv2d_42"  [label="(576, 1, 5, 5)", style=solid];
+"425 features_10_block_1_0_weight_bias_0_0" -> "426 conv2d_42"  [label="(576,)", style=solid];
+"426 conv2d_42" -> "427 hardswish__14"  [label="(1, 576, 7, 7)", style=solid];
+"427 hardswish__14" -> "428 quantize_per_tensor_default_37"  [label="(1, 576, 7, 7)", style=solid];
+"428 quantize_per_tensor_default_37" -> "429 dequantize_per_tensor_default_50"  [label="(1, 576, 7, 7)", style=solid];
+"428 quantize_per_tensor_default_37" -> "430 dequantize_per_tensor_default_49"  [label="(1, 576, 7, 7)", style=solid];
+"429 dequantize_per_tensor_default_50" -> "440 mul_7"  [label="(1, 576, 7, 7)", style=solid];
+"430 dequantize_per_tensor_default_49" -> "431 adaptive_avg_pool2d_7"  [label="(1, 576, 7, 7)", style=solid];
+"431 adaptive_avg_pool2d_7" -> "433 conv2d_43"  [label="(1, 576, 1, 1)", style=solid];
+"432 features_10_block_2_fc1_bias_0_0" -> "433 conv2d_43"  [label="(144,)", style=solid];
+"433 conv2d_43" -> "434 relu_7"  [label="(1, 144, 1, 1)", style=solid];
+"434 relu_7" -> "436 conv2d_44"  [label="(1, 144, 1, 1)", style=solid];
+"435 features_10_block_2_fc2_bias_0_0" -> "436 conv2d_44"  [label="(576,)", style=solid];
+"436 conv2d_44" -> "437 hardsigmoid_7"  [label="(1, 576, 1, 1)", style=solid];
+"437 hardsigmoid_7" -> "438 quantize_per_tensor_default_38"  [label="(1, 576, 1, 1)", style=solid];
+"438 quantize_per_tensor_default_38" -> "439 dequantize_per_tensor_default_51"  [label="(1, 576, 1, 1)", style=solid];
+"439 dequantize_per_tensor_default_51" -> "440 mul_7"  [label="(1, 576, 1, 1)", style=solid];
+"440 mul_7" -> "441 quantize_per_tensor_default_39"  [label="(1, 576, 7, 7)", style=solid];
+"441 quantize_per_tensor_default_39" -> "442 dequantize_per_tensor_default_52"  [label="(1, 576, 7, 7)", style=solid];
+"442 dequantize_per_tensor_default_52" -> "449 conv2d_45"  [label="(1, 576, 7, 7)", style=solid];
+"443 conv2d_45_scale_0" -> "446 quantize_per_channel_default_39"  [label="(96,)", style=solid];
+"443 conv2d_45_scale_0" -> "447 dequantize_per_channel_default_39"  [label="(96,)", style=solid];
+"444 conv2d_45_zero_point_0" -> "446 quantize_per_channel_default_39"  [label="(96,)", style=solid];
+"444 conv2d_45_zero_point_0" -> "447 dequantize_per_channel_default_39"  [label="(96,)", style=solid];
+"445 compressed_weight_updated_constant29" -> "446 quantize_per_channel_default_39"  [label="(96, 576, 1, 1)", style=solid];
+"446 quantize_per_channel_default_39" -> "447 dequantize_per_channel_default_39"  [label="(96, 576, 1, 1)", style=solid];
+"447 dequantize_per_channel_default_39" -> "449 conv2d_45"  [label="(96, 576, 1, 1)", style=solid];
+"448 features_10_block_3_0_weight_bias_0_0" -> "449 conv2d_45"  [label="(96,)", style=solid];
+"449 conv2d_45" -> "450 quantize_per_tensor_default_40"  [label="(1, 96, 7, 7)", style=solid];
+"450 quantize_per_tensor_default_40" -> "451 dequantize_per_tensor_default_53"  [label="(1, 96, 7, 7)", style=solid];
+"451 dequantize_per_tensor_default_53" -> "452 add__4"  [label="(1, 96, 7, 7)", style=solid];
+"452 add__4" -> "453 quantize_per_tensor_default_41"  [label="(1, 96, 7, 7)", style=solid];
+"453 quantize_per_tensor_default_41" -> "454 dequantize_per_tensor_default_55"  [label="(1, 96, 7, 7)", style=solid];
+"453 quantize_per_tensor_default_41" -> "455 dequantize_per_tensor_default_54"  [label="(1, 96, 7, 7)", style=solid];
+"454 dequantize_per_tensor_default_55" -> "500 add__5"  [label="(1, 96, 7, 7)", style=solid];
+"455 dequantize_per_tensor_default_54" -> "462 conv2d_46"  [label="(1, 96, 7, 7)", style=solid];
+"456 conv2d_46_scale_0" -> "459 quantize_per_channel_default_40"  [label="(576,)", style=solid];
+"456 conv2d_46_scale_0" -> "460 dequantize_per_channel_default_40"  [label="(576,)", style=solid];
+"457 conv2d_46_zero_point_0" -> "459 quantize_per_channel_default_40"  [label="(576,)", style=solid];
+"457 conv2d_46_zero_point_0" -> "460 dequantize_per_channel_default_40"  [label="(576,)", style=solid];
+"458 compressed_weight_updated_constant30" -> "459 quantize_per_channel_default_40"  [label="(576, 96, 1, 1)", style=solid];
+"459 quantize_per_channel_default_40" -> "460 dequantize_per_channel_default_40"  [label="(576, 96, 1, 1)", style=solid];
+"460 dequantize_per_channel_default_40" -> "462 conv2d_46"  [label="(576, 96, 1, 1)", style=solid];
+"461 features_11_block_0_0_weight_bias_0_0" -> "462 conv2d_46"  [label="(576,)", style=solid];
+"462 conv2d_46" -> "465 hardswish__15"  [label="(1, 576, 7, 7)", style=solid];
+"463 hardswish__15_scale_0" -> "466 quantize_per_channel_default_41"  [label="(576,)", style=solid];
+"463 hardswish__15_scale_0" -> "467 dequantize_per_channel_default_41"  [label="(576,)", style=solid];
+"464 hardswish__15_zero_point_0" -> "466 quantize_per_channel_default_41"  [label="(576,)", style=solid];
+"464 hardswish__15_zero_point_0" -> "467 dequantize_per_channel_default_41"  [label="(576,)", style=solid];
+"465 hardswish__15" -> "466 quantize_per_channel_default_41"  [label="(1, 576, 7, 7)", style=solid];
+"466 quantize_per_channel_default_41" -> "467 dequantize_per_channel_default_41"  [label="(1, 576, 7, 7)", style=solid];
+"467 dequantize_per_channel_default_41" -> "474 conv2d_47"  [label="(1, 576, 7, 7)", style=solid];
+"468 conv2d_47_scale_0" -> "471 quantize_per_channel_default_42"  [label="(576,)", style=solid];
+"468 conv2d_47_scale_0" -> "472 dequantize_per_channel_default_42"  [label="(576,)", style=solid];
+"469 conv2d_47_zero_point_0" -> "471 quantize_per_channel_default_42"  [label="(576,)", style=solid];
+"469 conv2d_47_zero_point_0" -> "472 dequantize_per_channel_default_42"  [label="(576,)", style=solid];
+"470 compressed_weight_updated_constant31" -> "471 quantize_per_channel_default_42"  [label="(576, 1, 5, 5)", style=solid];
+"471 quantize_per_channel_default_42" -> "472 dequantize_per_channel_default_42"  [label="(576, 1, 5, 5)", style=solid];
+"472 dequantize_per_channel_default_42" -> "474 conv2d_47"  [label="(576, 1, 5, 5)", style=solid];
+"473 features_11_block_1_0_weight_bias_0_0" -> "474 conv2d_47"  [label="(576,)", style=solid];
+"474 conv2d_47" -> "475 hardswish__16"  [label="(1, 576, 7, 7)", style=solid];
+"475 hardswish__16" -> "476 quantize_per_tensor_default_42"  [label="(1, 576, 7, 7)", style=solid];
+"476 quantize_per_tensor_default_42" -> "477 dequantize_per_tensor_default_57"  [label="(1, 576, 7, 7)", style=solid];
+"476 quantize_per_tensor_default_42" -> "478 dequantize_per_tensor_default_56"  [label="(1, 576, 7, 7)", style=solid];
+"477 dequantize_per_tensor_default_57" -> "488 mul_8"  [label="(1, 576, 7, 7)", style=solid];
+"478 dequantize_per_tensor_default_56" -> "479 adaptive_avg_pool2d_8"  [label="(1, 576, 7, 7)", style=solid];
+"479 adaptive_avg_pool2d_8" -> "481 conv2d_48"  [label="(1, 576, 1, 1)", style=solid];
+"480 features_11_block_2_fc1_bias_0_0" -> "481 conv2d_48"  [label="(144,)", style=solid];
+"481 conv2d_48" -> "482 relu_8"  [label="(1, 144, 1, 1)", style=solid];
+"482 relu_8" -> "484 conv2d_49"  [label="(1, 144, 1, 1)", style=solid];
+"483 features_11_block_2_fc2_bias_0_0" -> "484 conv2d_49"  [label="(576,)", style=solid];
 "484 conv2d_49" -> "485 hardsigmoid_8"  [label="(1, 576, 1, 1)", style=solid];
 "485 hardsigmoid_8" -> "486 quantize_per_tensor_default_43"  [label="(1, 576, 1, 1)", style=solid];
 "486 quantize_per_tensor_default_43" -> "487 dequantize_per_tensor_default_58"  [label="(1, 576, 1, 1)", style=solid];
@@ -1164,7 +1164,7 @@ strict digraph  {
 "519 compressed_weight_updated_constant34" -> "520 quantize_per_channel_default_45"  [label="(1024, 576)", style=solid];
 "520 quantize_per_channel_default_45" -> "521 dequantize_per_channel_default_45"  [label="(1024, 576)", style=solid];
 "521 dequantize_per_channel_default_45" -> "523 linear"  [label="(1024, 576)", style=solid];
-"522 _param_constant139_0_0" -> "523 linear"  [label="(1024,)", style=solid];
+"522 classifier_0_bias_0_0" -> "523 linear"  [label="(1024,)", style=solid];
 "523 linear" -> "524 hardswish__18"  [label="(1, 1024)", style=solid];
 "524 hardswish__18" -> "525 quantize_per_tensor_default_49"  [label="(1, 1024)", style=solid];
 "525 quantize_per_tensor_default_49" -> "526 dequantize_per_tensor_default_64"  [label="(1, 1024)", style=solid];
@@ -1177,6 +1177,6 @@ strict digraph  {
 "530 compressed_weight_updated_constant35" -> "531 quantize_per_channel_default_46"  [label="(1000, 1024)", style=solid];
 "531 quantize_per_channel_default_46" -> "532 dequantize_per_channel_default_46"  [label="(1000, 1024)", style=solid];
 "532 dequantize_per_channel_default_46" -> "534 linear_1"  [label="(1000, 1024)", style=solid];
-"533 _param_constant141_0_0" -> "534 linear_1"  [label="(1000,)", style=solid];
+"533 classifier_3_bias_0_0" -> "534 linear_1"  [label="(1000,)", style=solid];
 "534 linear_1" -> "535 output"  [label="(1, 1000)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/quantized/resnet18.dot b/tests/torch/data/reference_graphs/fx/quantized/resnet18.dot
index e2c32658bbb..e6875129c47 100644
--- a/tests/torch/data/reference_graphs/fx/quantized/resnet18.dot
+++ b/tests/torch/data/reference_graphs/fx/quantized/resnet18.dot
@@ -1,5 +1,5 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
+"0 x" [id=0, type=input];
 "1 quantize_per_tensor_default" [id=1, type=quantize_per_tensor];
 "2 dequantize_per_tensor_default" [id=2, type=dequantize_per_tensor];
 "3 conv2d_scale_0" [id=3, type=get_attr];
@@ -240,10 +240,10 @@ strict digraph  {
 "238 compressed_weight_updated_constant20" [id=238, type=get_attr];
 "239 quantize_per_channel_default_20" [id=239, type=quantize_per_channel];
 "240 dequantize_per_channel_default_20" [id=240, type=dequantize_per_channel];
-"241 _param_constant61_0_0" [id=241, type=get_attr];
+"241 fc_bias_0_0" [id=241, type=get_attr];
 "242 linear" [id=242, type=linear];
 "243 output" [id=243, type=output];
-"0 arg0_1" -> "1 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"0 x" -> "1 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
 "1 quantize_per_tensor_default" -> "2 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
 "2 dequantize_per_tensor_default" -> "9 conv2d"  [label="(1, 3, 224, 224)", style=solid];
 "3 conv2d_scale_0" -> "6 quantize_per_channel_default"  [label="(64,)", style=solid];
@@ -534,6 +534,6 @@ strict digraph  {
 "238 compressed_weight_updated_constant20" -> "239 quantize_per_channel_default_20"  [label="(1000, 512)", style=solid];
 "239 quantize_per_channel_default_20" -> "240 dequantize_per_channel_default_20"  [label="(1000, 512)", style=solid];
 "240 dequantize_per_channel_default_20" -> "242 linear"  [label="(1000, 512)", style=solid];
-"241 _param_constant61_0_0" -> "242 linear"  [label="(1000,)", style=solid];
+"241 fc_bias_0_0" -> "242 linear"  [label="(1000,)", style=solid];
 "242 linear" -> "243 output"  [label="(1, 1000)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/quantized/swin_v2_s.dot b/tests/torch/data/reference_graphs/fx/quantized/swin_v2_s.dot
index 1a0f64ecfe5..e45747bd6fa 100644
--- a/tests/torch/data/reference_graphs/fx/quantized/swin_v2_s.dot
+++ b/tests/torch/data/reference_graphs/fx/quantized/swin_v2_s.dot
@@ -1,2498 +1,2498 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 quantize_per_tensor_default" [id=1, type=quantize_per_tensor];
-"2 dequantize_per_tensor_default" [id=2, type=dequantize_per_tensor];
-"3 conv2d_scale_0" [id=3, type=get_attr];
-"4 conv2d_zero_point_0" [id=4, type=get_attr];
-"5 compressed_weight_updated_constant0" [id=5, type=get_attr];
-"6 quantize_per_channel_default" [id=6, type=quantize_per_channel];
-"7 dequantize_per_channel_default" [id=7, type=dequantize_per_channel];
-"8 _param_constant1_0_0" [id=8, type=get_attr];
-"9 conv2d" [id=9, type=conv2d];
-"10 permute" [id=10, type=permute];
-"11 _param_constant2" [id=11, type=get_attr];
-"12 _param_constant3" [id=12, type=get_attr];
-"13 layer_norm" [id=13, type=layer_norm];
-"14 _frozen_param0" [id=14, type=get_attr];
-"15 pad" [id=15, type=pad];
-"16 view_2" [id=16, type=view];
-"17 permute_2" [id=17, type=permute];
-"18 reshape" [id=18, type=reshape];
-"19 reshape_0_0_nncf_smooth_quant_0" [id=19, type=call_module];
-"20 quantize_per_tensor_default_1" [id=20, type=quantize_per_tensor];
-"21 dequantize_per_tensor_default_1" [id=21, type=dequantize_per_tensor];
-"22 linear_2_scale_0" [id=22, type=get_attr];
-"23 linear_2_zero_point_0" [id=23, type=get_attr];
-"24 compressed_weight_updated_constant1" [id=24, type=get_attr];
-"25 quantize_per_channel_default_2" [id=25, type=quantize_per_channel];
-"26 dequantize_per_channel_default_2" [id=26, type=dequantize_per_channel];
-"27 _frozen_param1_0_0" [id=27, type=get_attr];
-"28 linear_2" [id=28, type=linear];
-"29 reshape_1" [id=29, type=reshape];
-"30 permute_3" [id=30, type=permute];
-"31 select" [id=31, type=select];
-"32 select_1" [id=32, type=select];
-"33 select_2" [id=33, type=select];
-"34 linalg_vector_norm" [id=34, type=linalg_vector_norm];
-"35 clamp_min" [id=35, type=clamp_min];
-"36 expand_as" [id=36, type=expand_as];
-"37 div" [id=37, type=div];
-"38 quantize_per_tensor_default_2" [id=38, type=quantize_per_tensor];
-"39 dequantize_per_tensor_default_2" [id=39, type=dequantize_per_tensor];
-"40 linalg_vector_norm_1" [id=40, type=linalg_vector_norm];
-"41 clamp_min_1" [id=41, type=clamp_min];
-"42 expand_as_1" [id=42, type=expand_as];
-"43 div_1" [id=43, type=div];
-"44 quantize_per_tensor_default_3" [id=44, type=quantize_per_tensor];
-"45 dequantize_per_tensor_default_3" [id=45, type=dequantize_per_tensor];
-"46 transpose" [id=46, type=transpose];
-"47 matmul" [id=47, type=matmul];
-"48 _frozen_param3" [id=48, type=get_attr];
-"49 mul_1" [id=49, type=mul];
-"50 add" [id=50, type=add];
-"51 softmax" [id=51, type=softmax];
-"52 dropout" [id=52, type=dropout];
-"53 matmul_1" [id=53, type=matmul];
-"54 transpose_1" [id=54, type=transpose];
-"55 reshape_2" [id=55, type=reshape];
-"56 reshape_2_0_0_nncf_smooth_quant_0" [id=56, type=call_module];
-"57 quantize_per_tensor_default_4" [id=57, type=quantize_per_tensor];
-"58 dequantize_per_tensor_default_4" [id=58, type=dequantize_per_tensor];
-"59 linear_3_scale_0" [id=59, type=get_attr];
-"60 linear_3_zero_point_0" [id=60, type=get_attr];
-"61 compressed_weight_updated_constant2" [id=61, type=get_attr];
-"62 quantize_per_channel_default_3" [id=62, type=quantize_per_channel];
-"63 dequantize_per_channel_default_3" [id=63, type=dequantize_per_channel];
-"64 _param_constant11_0_0" [id=64, type=get_attr];
-"65 linear_3" [id=65, type=linear];
-"66 dropout_1" [id=66, type=dropout];
-"67 view_3" [id=67, type=view];
-"68 permute_4" [id=68, type=permute];
-"69 reshape_3" [id=69, type=reshape];
-"70 slice_2" [id=70, type=slice];
-"71 slice_3" [id=71, type=slice];
-"72 _param_constant12" [id=72, type=get_attr];
-"73 _param_constant13" [id=73, type=get_attr];
-"74 layer_norm_1" [id=74, type=layer_norm];
-"75 add_1" [id=75, type=add];
-"76 add_1_0_0_nncf_smooth_quant_0" [id=76, type=call_module];
-"77 quantize_per_tensor_default_5" [id=77, type=quantize_per_tensor];
-"78 dequantize_per_tensor_default_5" [id=78, type=dequantize_per_tensor];
-"79 linear_4_scale_0" [id=79, type=get_attr];
-"80 linear_4_zero_point_0" [id=80, type=get_attr];
-"81 compressed_weight_updated_constant3" [id=81, type=get_attr];
-"82 quantize_per_channel_default_4" [id=82, type=quantize_per_channel];
-"83 dequantize_per_channel_default_4" [id=83, type=dequantize_per_channel];
-"84 _param_constant15_0_0" [id=84, type=get_attr];
-"85 linear_4" [id=85, type=linear];
-"86 gelu" [id=86, type=gelu];
-"87 dropout_2" [id=87, type=dropout];
-"88 dropout_2_0_0_nncf_smooth_quant_0" [id=88, type=call_module];
-"89 quantize_per_tensor_default_6" [id=89, type=quantize_per_tensor];
-"90 dequantize_per_tensor_default_6" [id=90, type=dequantize_per_tensor];
-"91 linear_5_scale_0" [id=91, type=get_attr];
-"92 linear_5_zero_point_0" [id=92, type=get_attr];
-"93 compressed_weight_updated_constant4" [id=93, type=get_attr];
-"94 quantize_per_channel_default_5" [id=94, type=quantize_per_channel];
-"95 dequantize_per_channel_default_5" [id=95, type=dequantize_per_channel];
-"96 _param_constant17_0_0" [id=96, type=get_attr];
-"97 linear_5" [id=97, type=linear];
-"98 dropout_3" [id=98, type=dropout];
-"99 _param_constant18" [id=99, type=get_attr];
-"100 _param_constant19" [id=100, type=get_attr];
-"101 layer_norm_2" [id=101, type=layer_norm];
-"102 add_2" [id=102, type=add];
-"103 _frozen_param4" [id=103, type=get_attr];
-"104 pad_1" [id=104, type=pad];
-"105 roll" [id=105, type=roll];
-"106 view_6" [id=106, type=view];
-"107 permute_6" [id=107, type=permute];
-"108 reshape_4" [id=108, type=reshape];
-"109 reshape_4_0_0_nncf_smooth_quant_0" [id=109, type=call_module];
-"110 quantize_per_tensor_default_7" [id=110, type=quantize_per_tensor];
-"111 dequantize_per_tensor_default_7" [id=111, type=dequantize_per_tensor];
-"112 linear_8_scale_0" [id=112, type=get_attr];
-"113 linear_8_zero_point_0" [id=113, type=get_attr];
-"114 compressed_weight_updated_constant5" [id=114, type=get_attr];
-"115 quantize_per_channel_default_7" [id=115, type=quantize_per_channel];
-"116 dequantize_per_channel_default_7" [id=116, type=dequantize_per_channel];
-"117 _frozen_param5_0_0" [id=117, type=get_attr];
-"118 linear_8" [id=118, type=linear];
-"119 reshape_5" [id=119, type=reshape];
-"120 permute_7" [id=120, type=permute];
-"121 select_3" [id=121, type=select];
-"122 select_4" [id=122, type=select];
-"123 select_5" [id=123, type=select];
-"124 linalg_vector_norm_2" [id=124, type=linalg_vector_norm];
-"125 clamp_min_2" [id=125, type=clamp_min];
-"126 expand_as_2" [id=126, type=expand_as];
-"127 div_2" [id=127, type=div];
-"128 quantize_per_tensor_default_8" [id=128, type=quantize_per_tensor];
-"129 dequantize_per_tensor_default_8" [id=129, type=dequantize_per_tensor];
-"130 linalg_vector_norm_3" [id=130, type=linalg_vector_norm];
-"131 clamp_min_3" [id=131, type=clamp_min];
-"132 expand_as_3" [id=132, type=expand_as];
-"133 div_3" [id=133, type=div];
-"134 quantize_per_tensor_default_9" [id=134, type=quantize_per_tensor];
-"135 dequantize_per_tensor_default_9" [id=135, type=dequantize_per_tensor];
-"136 transpose_2" [id=136, type=transpose];
-"137 matmul_2" [id=137, type=matmul];
-"138 _frozen_param7" [id=138, type=get_attr];
-"139 mul_3" [id=139, type=mul];
-"140 add_3" [id=140, type=add];
-"141 new_zeros" [id=141, type=new_zeros];
-"142 view_7" [id=142, type=view];
-"143 permute_8" [id=143, type=permute];
-"144 reshape_6" [id=144, type=reshape];
-"145 unsqueeze_2" [id=145, type=unsqueeze];
-"146 unsqueeze_3" [id=146, type=unsqueeze];
-"147 sub" [id=147, type=sub];
-"148 ne" [id=148, type=ne];
-"149 masked_fill" [id=149, type=masked_fill];
-"150 eq" [id=150, type=eq];
-"151 masked_fill_1" [id=151, type=masked_fill];
-"152 view_8" [id=152, type=view];
-"153 unsqueeze_4" [id=153, type=unsqueeze];
-"154 unsqueeze_5" [id=154, type=unsqueeze];
-"155 add_4" [id=155, type=add];
-"156 view_9" [id=156, type=view];
-"157 softmax_1" [id=157, type=softmax];
-"158 dropout_4" [id=158, type=dropout];
-"159 matmul_3" [id=159, type=matmul];
-"160 transpose_3" [id=160, type=transpose];
-"161 reshape_7" [id=161, type=reshape];
-"162 reshape_7_0_0_nncf_smooth_quant_0" [id=162, type=call_module];
-"163 quantize_per_tensor_default_10" [id=163, type=quantize_per_tensor];
-"164 dequantize_per_tensor_default_10" [id=164, type=dequantize_per_tensor];
-"165 linear_9_scale_0" [id=165, type=get_attr];
-"166 linear_9_zero_point_0" [id=166, type=get_attr];
-"167 compressed_weight_updated_constant6" [id=167, type=get_attr];
-"168 quantize_per_channel_default_8" [id=168, type=quantize_per_channel];
-"169 dequantize_per_channel_default_8" [id=169, type=dequantize_per_channel];
-"170 _param_constant27_0_0" [id=170, type=get_attr];
-"171 linear_9" [id=171, type=linear];
-"172 dropout_5" [id=172, type=dropout];
-"173 view_10" [id=173, type=view];
-"174 permute_9" [id=174, type=permute];
-"175 reshape_8" [id=175, type=reshape];
-"176 roll_1" [id=176, type=roll];
-"177 slice_23" [id=177, type=slice];
-"178 slice_24" [id=178, type=slice];
-"179 _param_constant28" [id=179, type=get_attr];
-"180 _param_constant29" [id=180, type=get_attr];
-"181 layer_norm_3" [id=181, type=layer_norm];
-"182 add_5" [id=182, type=add];
-"183 add_5_0_0_nncf_smooth_quant_0" [id=183, type=call_module];
-"184 quantize_per_tensor_default_11" [id=184, type=quantize_per_tensor];
-"185 dequantize_per_tensor_default_11" [id=185, type=dequantize_per_tensor];
-"186 linear_10_scale_0" [id=186, type=get_attr];
-"187 linear_10_zero_point_0" [id=187, type=get_attr];
-"188 compressed_weight_updated_constant7" [id=188, type=get_attr];
-"189 quantize_per_channel_default_9" [id=189, type=quantize_per_channel];
-"190 dequantize_per_channel_default_9" [id=190, type=dequantize_per_channel];
-"191 _param_constant31_0_0" [id=191, type=get_attr];
-"192 linear_10" [id=192, type=linear];
-"193 gelu_1" [id=193, type=gelu];
-"194 dropout_6" [id=194, type=dropout];
-"195 dropout_6_0_0_nncf_smooth_quant_0" [id=195, type=call_module];
-"196 quantize_per_tensor_default_12" [id=196, type=quantize_per_tensor];
-"197 dequantize_per_tensor_default_12" [id=197, type=dequantize_per_tensor];
-"198 linear_11_scale_0" [id=198, type=get_attr];
-"199 linear_11_zero_point_0" [id=199, type=get_attr];
-"200 compressed_weight_updated_constant8" [id=200, type=get_attr];
-"201 quantize_per_channel_default_10" [id=201, type=quantize_per_channel];
-"202 dequantize_per_channel_default_10" [id=202, type=dequantize_per_channel];
-"203 _param_constant33_0_0" [id=203, type=get_attr];
-"204 linear_11" [id=204, type=linear];
-"205 dropout_7" [id=205, type=dropout];
-"206 _param_constant34" [id=206, type=get_attr];
-"207 _param_constant35" [id=207, type=get_attr];
-"208 layer_norm_4" [id=208, type=layer_norm];
-"209 add_6" [id=209, type=add];
-"210 pad_2" [id=210, type=pad];
-"211 slice_25" [id=211, type=slice];
-"212 slice_26" [id=212, type=slice];
-"213 slice_27" [id=213, type=slice];
-"214 slice_28" [id=214, type=slice];
-"215 slice_29" [id=215, type=slice];
-"216 slice_30" [id=216, type=slice];
-"217 slice_31" [id=217, type=slice];
-"218 slice_32" [id=218, type=slice];
-"219 slice_33" [id=219, type=slice];
-"220 slice_34" [id=220, type=slice];
-"221 slice_35" [id=221, type=slice];
-"222 slice_36" [id=222, type=slice];
-"223 cat" [id=223, type=cat];
-"224 cat_0_0_nncf_smooth_quant_0" [id=224, type=call_module];
-"225 quantize_per_tensor_default_13" [id=225, type=quantize_per_tensor];
-"226 dequantize_per_tensor_default_13" [id=226, type=dequantize_per_tensor];
-"227 linear_12_scale_0" [id=227, type=get_attr];
-"228 linear_12_zero_point_0" [id=228, type=get_attr];
-"229 compressed_weight_updated_constant9" [id=229, type=get_attr];
-"230 quantize_per_channel_default_11" [id=230, type=quantize_per_channel];
-"231 dequantize_per_channel_default_11" [id=231, type=dequantize_per_channel];
-"232 linear_12" [id=232, type=linear];
-"233 _param_constant37" [id=233, type=get_attr];
-"234 _param_constant38" [id=234, type=get_attr];
-"235 layer_norm_5" [id=235, type=layer_norm];
-"236 _frozen_param17" [id=236, type=get_attr];
-"237 pad_3" [id=237, type=pad];
-"238 view_13" [id=238, type=view];
-"239 permute_11" [id=239, type=permute];
-"240 reshape_9" [id=240, type=reshape];
-"241 reshape_9_0_0_nncf_smooth_quant_0" [id=241, type=call_module];
-"242 quantize_per_tensor_default_14" [id=242, type=quantize_per_tensor];
-"243 dequantize_per_tensor_default_14" [id=243, type=dequantize_per_tensor];
-"244 linear_15_scale_0" [id=244, type=get_attr];
-"245 linear_15_zero_point_0" [id=245, type=get_attr];
-"246 compressed_weight_updated_constant10" [id=246, type=get_attr];
-"247 quantize_per_channel_default_13" [id=247, type=quantize_per_channel];
-"248 dequantize_per_channel_default_13" [id=248, type=dequantize_per_channel];
-"249 _frozen_param18_0_0" [id=249, type=get_attr];
-"250 linear_15" [id=250, type=linear];
-"251 reshape_10" [id=251, type=reshape];
-"252 permute_12" [id=252, type=permute];
-"253 select_6" [id=253, type=select];
-"254 select_7" [id=254, type=select];
-"255 select_8" [id=255, type=select];
-"256 linalg_vector_norm_4" [id=256, type=linalg_vector_norm];
-"257 clamp_min_4" [id=257, type=clamp_min];
-"258 expand_as_4" [id=258, type=expand_as];
-"259 div_4" [id=259, type=div];
-"260 quantize_per_tensor_default_15" [id=260, type=quantize_per_tensor];
-"261 dequantize_per_tensor_default_15" [id=261, type=dequantize_per_tensor];
-"262 linalg_vector_norm_5" [id=262, type=linalg_vector_norm];
-"263 clamp_min_5" [id=263, type=clamp_min];
-"264 expand_as_5" [id=264, type=expand_as];
-"265 div_5" [id=265, type=div];
-"266 quantize_per_tensor_default_16" [id=266, type=quantize_per_tensor];
-"267 dequantize_per_tensor_default_16" [id=267, type=dequantize_per_tensor];
-"268 transpose_4" [id=268, type=transpose];
-"269 matmul_4" [id=269, type=matmul];
-"270 _frozen_param20" [id=270, type=get_attr];
-"271 mul_5" [id=271, type=mul];
-"272 add_7" [id=272, type=add];
-"273 softmax_2" [id=273, type=softmax];
-"274 dropout_8" [id=274, type=dropout];
-"275 matmul_5" [id=275, type=matmul];
-"276 transpose_5" [id=276, type=transpose];
-"277 reshape_11" [id=277, type=reshape];
-"278 reshape_11_0_0_nncf_smooth_quant_0" [id=278, type=call_module];
-"279 quantize_per_tensor_default_17" [id=279, type=quantize_per_tensor];
-"280 dequantize_per_tensor_default_17" [id=280, type=dequantize_per_tensor];
-"281 linear_16_scale_0" [id=281, type=get_attr];
-"282 linear_16_zero_point_0" [id=282, type=get_attr];
-"283 compressed_weight_updated_constant11" [id=283, type=get_attr];
-"284 quantize_per_channel_default_14" [id=284, type=quantize_per_channel];
-"285 dequantize_per_channel_default_14" [id=285, type=dequantize_per_channel];
-"286 _param_constant46_0_0" [id=286, type=get_attr];
-"287 linear_16" [id=287, type=linear];
-"288 dropout_9" [id=288, type=dropout];
-"289 view_14" [id=289, type=view];
-"290 permute_13" [id=290, type=permute];
-"291 reshape_12" [id=291, type=reshape];
-"292 slice_38" [id=292, type=slice];
-"293 slice_39" [id=293, type=slice];
-"294 slice_40" [id=294, type=slice];
-"295 slice_41" [id=295, type=slice];
-"296 contiguous_3" [id=296, type=contiguous];
-"297 _param_constant47" [id=297, type=get_attr];
-"298 _param_constant48" [id=298, type=get_attr];
-"299 layer_norm_6" [id=299, type=layer_norm];
-"300 add_8" [id=300, type=add];
-"301 add_8_0_0_nncf_smooth_quant_0" [id=301, type=call_module];
-"302 quantize_per_tensor_default_18" [id=302, type=quantize_per_tensor];
-"303 dequantize_per_tensor_default_18" [id=303, type=dequantize_per_tensor];
-"304 linear_17_scale_0" [id=304, type=get_attr];
-"305 linear_17_zero_point_0" [id=305, type=get_attr];
-"306 compressed_weight_updated_constant12" [id=306, type=get_attr];
-"307 quantize_per_channel_default_15" [id=307, type=quantize_per_channel];
-"308 dequantize_per_channel_default_15" [id=308, type=dequantize_per_channel];
-"309 _param_constant50_0_0" [id=309, type=get_attr];
-"310 linear_17" [id=310, type=linear];
-"311 gelu_2" [id=311, type=gelu];
-"312 dropout_10" [id=312, type=dropout];
-"313 dropout_10_0_0_nncf_smooth_quant_0" [id=313, type=call_module];
-"314 quantize_per_tensor_default_19" [id=314, type=quantize_per_tensor];
-"315 dequantize_per_tensor_default_19" [id=315, type=dequantize_per_tensor];
-"316 linear_18_scale_0" [id=316, type=get_attr];
-"317 linear_18_zero_point_0" [id=317, type=get_attr];
-"318 compressed_weight_updated_constant13" [id=318, type=get_attr];
-"319 quantize_per_channel_default_16" [id=319, type=quantize_per_channel];
-"320 dequantize_per_channel_default_16" [id=320, type=dequantize_per_channel];
-"321 _param_constant52_0_0" [id=321, type=get_attr];
-"322 linear_18" [id=322, type=linear];
-"323 dropout_11" [id=323, type=dropout];
-"324 _param_constant53" [id=324, type=get_attr];
-"325 _param_constant54" [id=325, type=get_attr];
-"326 layer_norm_7" [id=326, type=layer_norm];
-"327 add_9" [id=327, type=add];
-"328 _frozen_param21" [id=328, type=get_attr];
-"329 pad_4" [id=329, type=pad];
-"330 roll_2" [id=330, type=roll];
-"331 view_17" [id=331, type=view];
-"332 permute_15" [id=332, type=permute];
-"333 reshape_13" [id=333, type=reshape];
-"334 reshape_13_0_0_nncf_smooth_quant_0" [id=334, type=call_module];
-"335 quantize_per_tensor_default_20" [id=335, type=quantize_per_tensor];
-"336 dequantize_per_tensor_default_20" [id=336, type=dequantize_per_tensor];
-"337 linear_21_scale_0" [id=337, type=get_attr];
-"338 linear_21_zero_point_0" [id=338, type=get_attr];
-"339 compressed_weight_updated_constant14" [id=339, type=get_attr];
-"340 quantize_per_channel_default_18" [id=340, type=quantize_per_channel];
-"341 dequantize_per_channel_default_18" [id=341, type=dequantize_per_channel];
-"342 _frozen_param22_0_0" [id=342, type=get_attr];
-"343 linear_21" [id=343, type=linear];
-"344 reshape_14" [id=344, type=reshape];
-"345 permute_16" [id=345, type=permute];
-"346 select_9" [id=346, type=select];
-"347 select_10" [id=347, type=select];
-"348 select_11" [id=348, type=select];
-"349 linalg_vector_norm_6" [id=349, type=linalg_vector_norm];
-"350 clamp_min_6" [id=350, type=clamp_min];
-"351 expand_as_6" [id=351, type=expand_as];
-"352 div_6" [id=352, type=div];
-"353 quantize_per_tensor_default_21" [id=353, type=quantize_per_tensor];
-"354 dequantize_per_tensor_default_21" [id=354, type=dequantize_per_tensor];
-"355 linalg_vector_norm_7" [id=355, type=linalg_vector_norm];
-"356 clamp_min_7" [id=356, type=clamp_min];
-"357 expand_as_7" [id=357, type=expand_as];
-"358 div_7" [id=358, type=div];
-"359 quantize_per_tensor_default_22" [id=359, type=quantize_per_tensor];
-"360 dequantize_per_tensor_default_22" [id=360, type=dequantize_per_tensor];
-"361 transpose_6" [id=361, type=transpose];
-"362 matmul_6" [id=362, type=matmul];
-"363 _frozen_param24" [id=363, type=get_attr];
-"364 mul_7" [id=364, type=mul];
-"365 add_10" [id=365, type=add];
-"366 new_zeros_1" [id=366, type=new_zeros];
-"367 view_18" [id=367, type=view];
-"368 permute_17" [id=368, type=permute];
-"369 reshape_15" [id=369, type=reshape];
-"370 unsqueeze_8" [id=370, type=unsqueeze];
-"371 unsqueeze_9" [id=371, type=unsqueeze];
-"372 sub_1" [id=372, type=sub];
-"373 ne_1" [id=373, type=ne];
-"374 masked_fill_2" [id=374, type=masked_fill];
-"375 eq_1" [id=375, type=eq];
-"376 masked_fill_3" [id=376, type=masked_fill];
-"377 view_19" [id=377, type=view];
-"378 unsqueeze_10" [id=378, type=unsqueeze];
-"379 unsqueeze_11" [id=379, type=unsqueeze];
-"380 add_11" [id=380, type=add];
-"381 view_20" [id=381, type=view];
-"382 softmax_3" [id=382, type=softmax];
-"383 dropout_12" [id=383, type=dropout];
-"384 matmul_7" [id=384, type=matmul];
-"385 transpose_7" [id=385, type=transpose];
-"386 reshape_16" [id=386, type=reshape];
-"387 reshape_16_0_0_nncf_smooth_quant_0" [id=387, type=call_module];
-"388 quantize_per_tensor_default_23" [id=388, type=quantize_per_tensor];
-"389 dequantize_per_tensor_default_23" [id=389, type=dequantize_per_tensor];
-"390 linear_22_scale_0" [id=390, type=get_attr];
-"391 linear_22_zero_point_0" [id=391, type=get_attr];
-"392 compressed_weight_updated_constant15" [id=392, type=get_attr];
-"393 quantize_per_channel_default_19" [id=393, type=quantize_per_channel];
-"394 dequantize_per_channel_default_19" [id=394, type=dequantize_per_channel];
-"395 _param_constant62_0_0" [id=395, type=get_attr];
-"396 linear_22" [id=396, type=linear];
-"397 dropout_13" [id=397, type=dropout];
-"398 view_21" [id=398, type=view];
-"399 permute_18" [id=399, type=permute];
-"400 reshape_17" [id=400, type=reshape];
-"401 roll_3" [id=401, type=roll];
-"402 slice_61" [id=402, type=slice];
-"403 slice_62" [id=403, type=slice];
-"404 slice_63" [id=404, type=slice];
-"405 slice_64" [id=405, type=slice];
-"406 contiguous_5" [id=406, type=contiguous];
-"407 _param_constant63" [id=407, type=get_attr];
-"408 _param_constant64" [id=408, type=get_attr];
-"409 layer_norm_8" [id=409, type=layer_norm];
-"410 add_12" [id=410, type=add];
-"411 add_12_0_0_nncf_smooth_quant_0" [id=411, type=call_module];
-"412 quantize_per_tensor_default_24" [id=412, type=quantize_per_tensor];
-"413 dequantize_per_tensor_default_24" [id=413, type=dequantize_per_tensor];
-"414 linear_23_scale_0" [id=414, type=get_attr];
-"415 linear_23_zero_point_0" [id=415, type=get_attr];
-"416 compressed_weight_updated_constant16" [id=416, type=get_attr];
-"417 quantize_per_channel_default_20" [id=417, type=quantize_per_channel];
-"418 dequantize_per_channel_default_20" [id=418, type=dequantize_per_channel];
-"419 _param_constant66_0_0" [id=419, type=get_attr];
-"420 linear_23" [id=420, type=linear];
-"421 gelu_3" [id=421, type=gelu];
-"422 dropout_14" [id=422, type=dropout];
-"423 dropout_14_0_0_nncf_smooth_quant_0" [id=423, type=call_module];
-"424 quantize_per_tensor_default_25" [id=424, type=quantize_per_tensor];
-"425 dequantize_per_tensor_default_25" [id=425, type=dequantize_per_tensor];
-"426 linear_24_scale_0" [id=426, type=get_attr];
-"427 linear_24_zero_point_0" [id=427, type=get_attr];
-"428 compressed_weight_updated_constant17" [id=428, type=get_attr];
-"429 quantize_per_channel_default_21" [id=429, type=quantize_per_channel];
-"430 dequantize_per_channel_default_21" [id=430, type=dequantize_per_channel];
-"431 _param_constant68_0_0" [id=431, type=get_attr];
-"432 linear_24" [id=432, type=linear];
-"433 dropout_15" [id=433, type=dropout];
-"434 _param_constant69" [id=434, type=get_attr];
-"435 _param_constant70" [id=435, type=get_attr];
-"436 layer_norm_9" [id=436, type=layer_norm];
-"437 add_13" [id=437, type=add];
-"438 pad_5" [id=438, type=pad];
-"439 slice_65" [id=439, type=slice];
-"440 slice_66" [id=440, type=slice];
-"441 slice_67" [id=441, type=slice];
-"442 slice_68" [id=442, type=slice];
-"443 slice_69" [id=443, type=slice];
-"444 slice_70" [id=444, type=slice];
-"445 slice_71" [id=445, type=slice];
-"446 slice_72" [id=446, type=slice];
-"447 slice_73" [id=447, type=slice];
-"448 slice_74" [id=448, type=slice];
-"449 slice_75" [id=449, type=slice];
-"450 slice_76" [id=450, type=slice];
-"451 cat_1" [id=451, type=cat];
-"452 cat_1_0_0_nncf_smooth_quant_0" [id=452, type=call_module];
-"453 quantize_per_tensor_default_26" [id=453, type=quantize_per_tensor];
-"454 dequantize_per_tensor_default_26" [id=454, type=dequantize_per_tensor];
-"455 linear_25_scale_0" [id=455, type=get_attr];
-"456 linear_25_zero_point_0" [id=456, type=get_attr];
-"457 compressed_weight_updated_constant18" [id=457, type=get_attr];
-"458 quantize_per_channel_default_22" [id=458, type=quantize_per_channel];
-"459 dequantize_per_channel_default_22" [id=459, type=dequantize_per_channel];
-"460 linear_25" [id=460, type=linear];
-"461 _param_constant72" [id=461, type=get_attr];
-"462 _param_constant73" [id=462, type=get_attr];
-"463 layer_norm_10" [id=463, type=layer_norm];
-"464 _frozen_param34" [id=464, type=get_attr];
-"465 pad_6" [id=465, type=pad];
-"466 view_24" [id=466, type=view];
-"467 permute_20" [id=467, type=permute];
-"468 reshape_18" [id=468, type=reshape];
-"469 reshape_18_0_0_nncf_smooth_quant_0" [id=469, type=call_module];
-"470 quantize_per_tensor_default_27" [id=470, type=quantize_per_tensor];
-"471 dequantize_per_tensor_default_27" [id=471, type=dequantize_per_tensor];
-"472 linear_28_scale_0" [id=472, type=get_attr];
-"473 linear_28_zero_point_0" [id=473, type=get_attr];
-"474 compressed_weight_updated_constant19" [id=474, type=get_attr];
-"475 quantize_per_channel_default_24" [id=475, type=quantize_per_channel];
-"476 dequantize_per_channel_default_24" [id=476, type=dequantize_per_channel];
-"477 _frozen_param35_0_0" [id=477, type=get_attr];
-"478 linear_28" [id=478, type=linear];
-"479 reshape_19" [id=479, type=reshape];
-"480 permute_21" [id=480, type=permute];
-"481 select_12" [id=481, type=select];
-"482 select_13" [id=482, type=select];
-"483 select_14" [id=483, type=select];
-"484 linalg_vector_norm_8" [id=484, type=linalg_vector_norm];
-"485 clamp_min_8" [id=485, type=clamp_min];
-"486 expand_as_8" [id=486, type=expand_as];
-"487 div_8" [id=487, type=div];
-"488 quantize_per_tensor_default_28" [id=488, type=quantize_per_tensor];
-"489 dequantize_per_tensor_default_28" [id=489, type=dequantize_per_tensor];
-"490 linalg_vector_norm_9" [id=490, type=linalg_vector_norm];
-"491 clamp_min_9" [id=491, type=clamp_min];
-"492 expand_as_9" [id=492, type=expand_as];
-"493 div_9" [id=493, type=div];
-"494 quantize_per_tensor_default_29" [id=494, type=quantize_per_tensor];
-"495 dequantize_per_tensor_default_29" [id=495, type=dequantize_per_tensor];
-"496 transpose_8" [id=496, type=transpose];
-"497 matmul_8" [id=497, type=matmul];
-"498 _frozen_param37" [id=498, type=get_attr];
-"499 mul_9" [id=499, type=mul];
-"500 add_14" [id=500, type=add];
-"501 softmax_4" [id=501, type=softmax];
-"502 dropout_16" [id=502, type=dropout];
-"503 matmul_9" [id=503, type=matmul];
-"504 transpose_9" [id=504, type=transpose];
-"505 reshape_20" [id=505, type=reshape];
-"506 reshape_20_0_0_nncf_smooth_quant_0" [id=506, type=call_module];
-"507 quantize_per_tensor_default_30" [id=507, type=quantize_per_tensor];
-"508 dequantize_per_tensor_default_30" [id=508, type=dequantize_per_tensor];
-"509 linear_29_scale_0" [id=509, type=get_attr];
-"510 linear_29_zero_point_0" [id=510, type=get_attr];
-"511 compressed_weight_updated_constant20" [id=511, type=get_attr];
-"512 quantize_per_channel_default_25" [id=512, type=quantize_per_channel];
-"513 dequantize_per_channel_default_25" [id=513, type=dequantize_per_channel];
-"514 _param_constant81_0_0" [id=514, type=get_attr];
-"515 linear_29" [id=515, type=linear];
-"516 dropout_17" [id=516, type=dropout];
-"517 view_25" [id=517, type=view];
-"518 permute_22" [id=518, type=permute];
-"519 reshape_21" [id=519, type=reshape];
-"520 slice_78" [id=520, type=slice];
-"521 slice_79" [id=521, type=slice];
-"522 slice_80" [id=522, type=slice];
-"523 slice_81" [id=523, type=slice];
-"524 contiguous_7" [id=524, type=contiguous];
-"525 _param_constant82" [id=525, type=get_attr];
-"526 _param_constant83" [id=526, type=get_attr];
-"527 layer_norm_11" [id=527, type=layer_norm];
-"528 add_15" [id=528, type=add];
-"529 add_15_0_0_nncf_smooth_quant_0" [id=529, type=call_module];
-"530 quantize_per_tensor_default_31" [id=530, type=quantize_per_tensor];
-"531 dequantize_per_tensor_default_31" [id=531, type=dequantize_per_tensor];
-"532 linear_30_scale_0" [id=532, type=get_attr];
-"533 linear_30_zero_point_0" [id=533, type=get_attr];
-"534 compressed_weight_updated_constant21" [id=534, type=get_attr];
-"535 quantize_per_channel_default_26" [id=535, type=quantize_per_channel];
-"536 dequantize_per_channel_default_26" [id=536, type=dequantize_per_channel];
-"537 _param_constant85_0_0" [id=537, type=get_attr];
-"538 linear_30" [id=538, type=linear];
-"539 gelu_4" [id=539, type=gelu];
-"540 dropout_18" [id=540, type=dropout];
-"541 dropout_18_0_0_nncf_smooth_quant_0" [id=541, type=call_module];
-"542 quantize_per_tensor_default_32" [id=542, type=quantize_per_tensor];
-"543 dequantize_per_tensor_default_32" [id=543, type=dequantize_per_tensor];
-"544 linear_31_scale_0" [id=544, type=get_attr];
-"545 linear_31_zero_point_0" [id=545, type=get_attr];
-"546 compressed_weight_updated_constant22" [id=546, type=get_attr];
-"547 quantize_per_channel_default_27" [id=547, type=quantize_per_channel];
-"548 dequantize_per_channel_default_27" [id=548, type=dequantize_per_channel];
-"549 _param_constant87_0_0" [id=549, type=get_attr];
-"550 linear_31" [id=550, type=linear];
-"551 dropout_19" [id=551, type=dropout];
-"552 _param_constant88" [id=552, type=get_attr];
-"553 _param_constant89" [id=553, type=get_attr];
-"554 layer_norm_12" [id=554, type=layer_norm];
-"555 add_16" [id=555, type=add];
-"556 _frozen_param38" [id=556, type=get_attr];
-"557 pad_7" [id=557, type=pad];
-"558 roll_4" [id=558, type=roll];
-"559 view_28" [id=559, type=view];
-"560 permute_24" [id=560, type=permute];
-"561 reshape_22" [id=561, type=reshape];
-"562 reshape_22_0_0_nncf_smooth_quant_0" [id=562, type=call_module];
-"563 quantize_per_tensor_default_33" [id=563, type=quantize_per_tensor];
-"564 dequantize_per_tensor_default_33" [id=564, type=dequantize_per_tensor];
-"565 linear_34_scale_0" [id=565, type=get_attr];
-"566 linear_34_zero_point_0" [id=566, type=get_attr];
-"567 compressed_weight_updated_constant23" [id=567, type=get_attr];
-"568 quantize_per_channel_default_29" [id=568, type=quantize_per_channel];
-"569 dequantize_per_channel_default_29" [id=569, type=dequantize_per_channel];
-"570 _frozen_param39_0_0" [id=570, type=get_attr];
-"571 linear_34" [id=571, type=linear];
-"572 reshape_23" [id=572, type=reshape];
-"573 permute_25" [id=573, type=permute];
-"574 select_15" [id=574, type=select];
-"575 select_16" [id=575, type=select];
-"576 select_17" [id=576, type=select];
-"577 linalg_vector_norm_10" [id=577, type=linalg_vector_norm];
-"578 clamp_min_10" [id=578, type=clamp_min];
-"579 expand_as_10" [id=579, type=expand_as];
-"580 div_10" [id=580, type=div];
-"581 quantize_per_tensor_default_34" [id=581, type=quantize_per_tensor];
-"582 dequantize_per_tensor_default_34" [id=582, type=dequantize_per_tensor];
-"583 linalg_vector_norm_11" [id=583, type=linalg_vector_norm];
-"584 clamp_min_11" [id=584, type=clamp_min];
-"585 expand_as_11" [id=585, type=expand_as];
-"586 div_11" [id=586, type=div];
-"587 quantize_per_tensor_default_35" [id=587, type=quantize_per_tensor];
-"588 dequantize_per_tensor_default_35" [id=588, type=dequantize_per_tensor];
-"589 transpose_10" [id=589, type=transpose];
-"590 matmul_10" [id=590, type=matmul];
-"591 _frozen_param41" [id=591, type=get_attr];
-"592 mul_11" [id=592, type=mul];
-"593 add_17" [id=593, type=add];
-"594 new_zeros_2" [id=594, type=new_zeros];
-"595 view_29" [id=595, type=view];
-"596 permute_26" [id=596, type=permute];
-"597 reshape_24" [id=597, type=reshape];
-"598 unsqueeze_14" [id=598, type=unsqueeze];
-"599 unsqueeze_15" [id=599, type=unsqueeze];
-"600 sub_2" [id=600, type=sub];
-"601 ne_2" [id=601, type=ne];
-"602 masked_fill_4" [id=602, type=masked_fill];
-"603 eq_2" [id=603, type=eq];
-"604 masked_fill_5" [id=604, type=masked_fill];
-"605 view_30" [id=605, type=view];
-"606 unsqueeze_16" [id=606, type=unsqueeze];
-"607 unsqueeze_17" [id=607, type=unsqueeze];
-"608 add_18" [id=608, type=add];
-"609 view_31" [id=609, type=view];
-"610 softmax_5" [id=610, type=softmax];
-"611 dropout_20" [id=611, type=dropout];
-"612 matmul_11" [id=612, type=matmul];
-"613 transpose_11" [id=613, type=transpose];
-"614 reshape_25" [id=614, type=reshape];
-"615 reshape_25_0_0_nncf_smooth_quant_0" [id=615, type=call_module];
-"616 quantize_per_tensor_default_36" [id=616, type=quantize_per_tensor];
-"617 dequantize_per_tensor_default_36" [id=617, type=dequantize_per_tensor];
-"618 linear_35_scale_0" [id=618, type=get_attr];
-"619 linear_35_zero_point_0" [id=619, type=get_attr];
-"620 compressed_weight_updated_constant24" [id=620, type=get_attr];
-"621 quantize_per_channel_default_30" [id=621, type=quantize_per_channel];
-"622 dequantize_per_channel_default_30" [id=622, type=dequantize_per_channel];
-"623 _param_constant97_0_0" [id=623, type=get_attr];
-"624 linear_35" [id=624, type=linear];
-"625 dropout_21" [id=625, type=dropout];
-"626 view_32" [id=626, type=view];
-"627 permute_27" [id=627, type=permute];
-"628 reshape_26" [id=628, type=reshape];
-"629 roll_5" [id=629, type=roll];
-"630 slice_101" [id=630, type=slice];
-"631 slice_102" [id=631, type=slice];
-"632 slice_103" [id=632, type=slice];
-"633 slice_104" [id=633, type=slice];
-"634 contiguous_9" [id=634, type=contiguous];
-"635 _param_constant98" [id=635, type=get_attr];
-"636 _param_constant99" [id=636, type=get_attr];
-"637 layer_norm_13" [id=637, type=layer_norm];
-"638 add_19" [id=638, type=add];
-"639 add_19_0_0_nncf_smooth_quant_0" [id=639, type=call_module];
-"640 quantize_per_tensor_default_37" [id=640, type=quantize_per_tensor];
-"641 dequantize_per_tensor_default_37" [id=641, type=dequantize_per_tensor];
-"642 linear_36_scale_0" [id=642, type=get_attr];
-"643 linear_36_zero_point_0" [id=643, type=get_attr];
-"644 compressed_weight_updated_constant25" [id=644, type=get_attr];
-"645 quantize_per_channel_default_31" [id=645, type=quantize_per_channel];
-"646 dequantize_per_channel_default_31" [id=646, type=dequantize_per_channel];
-"647 _param_constant101_0_0" [id=647, type=get_attr];
-"648 linear_36" [id=648, type=linear];
-"649 gelu_5" [id=649, type=gelu];
-"650 dropout_22" [id=650, type=dropout];
-"651 dropout_22_0_0_nncf_smooth_quant_0" [id=651, type=call_module];
-"652 quantize_per_tensor_default_38" [id=652, type=quantize_per_tensor];
-"653 dequantize_per_tensor_default_38" [id=653, type=dequantize_per_tensor];
-"654 linear_37_scale_0" [id=654, type=get_attr];
-"655 linear_37_zero_point_0" [id=655, type=get_attr];
-"656 compressed_weight_updated_constant26" [id=656, type=get_attr];
-"657 quantize_per_channel_default_32" [id=657, type=quantize_per_channel];
-"658 dequantize_per_channel_default_32" [id=658, type=dequantize_per_channel];
-"659 _param_constant103_0_0" [id=659, type=get_attr];
-"660 linear_37" [id=660, type=linear];
-"661 dropout_23" [id=661, type=dropout];
-"662 _param_constant104" [id=662, type=get_attr];
-"663 _param_constant105" [id=663, type=get_attr];
-"664 layer_norm_14" [id=664, type=layer_norm];
-"665 add_20" [id=665, type=add];
-"666 _frozen_param51" [id=666, type=get_attr];
-"667 pad_8" [id=667, type=pad];
-"668 view_35" [id=668, type=view];
-"669 permute_29" [id=669, type=permute];
-"670 reshape_27" [id=670, type=reshape];
-"671 reshape_27_0_0_nncf_smooth_quant_0" [id=671, type=call_module];
-"672 quantize_per_tensor_default_39" [id=672, type=quantize_per_tensor];
-"673 dequantize_per_tensor_default_39" [id=673, type=dequantize_per_tensor];
-"674 linear_40_scale_0" [id=674, type=get_attr];
-"675 linear_40_zero_point_0" [id=675, type=get_attr];
-"676 compressed_weight_updated_constant27" [id=676, type=get_attr];
-"677 quantize_per_channel_default_34" [id=677, type=quantize_per_channel];
-"678 dequantize_per_channel_default_34" [id=678, type=dequantize_per_channel];
-"679 _frozen_param52_0_0" [id=679, type=get_attr];
-"680 linear_40" [id=680, type=linear];
-"681 reshape_28" [id=681, type=reshape];
-"682 permute_30" [id=682, type=permute];
-"683 select_18" [id=683, type=select];
-"684 select_19" [id=684, type=select];
-"685 select_20" [id=685, type=select];
-"686 linalg_vector_norm_12" [id=686, type=linalg_vector_norm];
-"687 clamp_min_12" [id=687, type=clamp_min];
-"688 expand_as_12" [id=688, type=expand_as];
-"689 div_12" [id=689, type=div];
-"690 quantize_per_tensor_default_40" [id=690, type=quantize_per_tensor];
-"691 dequantize_per_tensor_default_40" [id=691, type=dequantize_per_tensor];
-"692 linalg_vector_norm_13" [id=692, type=linalg_vector_norm];
-"693 clamp_min_13" [id=693, type=clamp_min];
-"694 expand_as_13" [id=694, type=expand_as];
-"695 div_13" [id=695, type=div];
-"696 quantize_per_tensor_default_41" [id=696, type=quantize_per_tensor];
-"697 dequantize_per_tensor_default_41" [id=697, type=dequantize_per_tensor];
-"698 transpose_12" [id=698, type=transpose];
-"699 matmul_12" [id=699, type=matmul];
-"700 _frozen_param54" [id=700, type=get_attr];
-"701 mul_13" [id=701, type=mul];
-"702 add_21" [id=702, type=add];
-"703 softmax_6" [id=703, type=softmax];
-"704 dropout_24" [id=704, type=dropout];
-"705 matmul_13" [id=705, type=matmul];
-"706 transpose_13" [id=706, type=transpose];
-"707 reshape_29" [id=707, type=reshape];
-"708 reshape_29_0_0_nncf_smooth_quant_0" [id=708, type=call_module];
-"709 quantize_per_tensor_default_42" [id=709, type=quantize_per_tensor];
-"710 dequantize_per_tensor_default_42" [id=710, type=dequantize_per_tensor];
-"711 linear_41_scale_0" [id=711, type=get_attr];
-"712 linear_41_zero_point_0" [id=712, type=get_attr];
-"713 compressed_weight_updated_constant28" [id=713, type=get_attr];
-"714 quantize_per_channel_default_35" [id=714, type=quantize_per_channel];
-"715 dequantize_per_channel_default_35" [id=715, type=dequantize_per_channel];
-"716 _param_constant113_0_0" [id=716, type=get_attr];
-"717 linear_41" [id=717, type=linear];
-"718 dropout_25" [id=718, type=dropout];
-"719 view_36" [id=719, type=view];
-"720 permute_31" [id=720, type=permute];
-"721 reshape_30" [id=721, type=reshape];
-"722 slice_106" [id=722, type=slice];
-"723 slice_107" [id=723, type=slice];
-"724 slice_108" [id=724, type=slice];
-"725 slice_109" [id=725, type=slice];
-"726 contiguous_11" [id=726, type=contiguous];
-"727 _param_constant114" [id=727, type=get_attr];
-"728 _param_constant115" [id=728, type=get_attr];
-"729 layer_norm_15" [id=729, type=layer_norm];
-"730 add_22" [id=730, type=add];
-"731 add_22_0_0_nncf_smooth_quant_0" [id=731, type=call_module];
-"732 quantize_per_tensor_default_43" [id=732, type=quantize_per_tensor];
-"733 dequantize_per_tensor_default_43" [id=733, type=dequantize_per_tensor];
-"734 linear_42_scale_0" [id=734, type=get_attr];
-"735 linear_42_zero_point_0" [id=735, type=get_attr];
-"736 compressed_weight_updated_constant29" [id=736, type=get_attr];
-"737 quantize_per_channel_default_36" [id=737, type=quantize_per_channel];
-"738 dequantize_per_channel_default_36" [id=738, type=dequantize_per_channel];
-"739 _param_constant117_0_0" [id=739, type=get_attr];
-"740 linear_42" [id=740, type=linear];
-"741 gelu_6" [id=741, type=gelu];
-"742 dropout_26" [id=742, type=dropout];
-"743 dropout_26_0_0_nncf_smooth_quant_0" [id=743, type=call_module];
-"744 quantize_per_tensor_default_44" [id=744, type=quantize_per_tensor];
-"745 dequantize_per_tensor_default_44" [id=745, type=dequantize_per_tensor];
-"746 linear_43_scale_0" [id=746, type=get_attr];
-"747 linear_43_zero_point_0" [id=747, type=get_attr];
-"748 compressed_weight_updated_constant30" [id=748, type=get_attr];
-"749 quantize_per_channel_default_37" [id=749, type=quantize_per_channel];
-"750 dequantize_per_channel_default_37" [id=750, type=dequantize_per_channel];
-"751 _param_constant119_0_0" [id=751, type=get_attr];
-"752 linear_43" [id=752, type=linear];
-"753 dropout_27" [id=753, type=dropout];
-"754 _param_constant120" [id=754, type=get_attr];
-"755 _param_constant121" [id=755, type=get_attr];
-"756 layer_norm_16" [id=756, type=layer_norm];
-"757 add_23" [id=757, type=add];
-"758 _frozen_param55" [id=758, type=get_attr];
-"759 pad_9" [id=759, type=pad];
-"760 roll_6" [id=760, type=roll];
-"761 view_39" [id=761, type=view];
-"762 permute_33" [id=762, type=permute];
-"763 reshape_31" [id=763, type=reshape];
-"764 reshape_31_0_0_nncf_smooth_quant_0" [id=764, type=call_module];
-"765 quantize_per_tensor_default_45" [id=765, type=quantize_per_tensor];
-"766 dequantize_per_tensor_default_45" [id=766, type=dequantize_per_tensor];
-"767 linear_46_scale_0" [id=767, type=get_attr];
-"768 linear_46_zero_point_0" [id=768, type=get_attr];
-"769 compressed_weight_updated_constant31" [id=769, type=get_attr];
-"770 quantize_per_channel_default_39" [id=770, type=quantize_per_channel];
-"771 dequantize_per_channel_default_39" [id=771, type=dequantize_per_channel];
-"772 _frozen_param56_0_0" [id=772, type=get_attr];
-"773 linear_46" [id=773, type=linear];
-"774 reshape_32" [id=774, type=reshape];
-"775 permute_34" [id=775, type=permute];
-"776 select_21" [id=776, type=select];
-"777 select_22" [id=777, type=select];
-"778 select_23" [id=778, type=select];
-"779 linalg_vector_norm_14" [id=779, type=linalg_vector_norm];
-"780 clamp_min_14" [id=780, type=clamp_min];
-"781 expand_as_14" [id=781, type=expand_as];
-"782 div_14" [id=782, type=div];
-"783 quantize_per_tensor_default_46" [id=783, type=quantize_per_tensor];
-"784 dequantize_per_tensor_default_46" [id=784, type=dequantize_per_tensor];
-"785 linalg_vector_norm_15" [id=785, type=linalg_vector_norm];
-"786 clamp_min_15" [id=786, type=clamp_min];
-"787 expand_as_15" [id=787, type=expand_as];
-"788 div_15" [id=788, type=div];
-"789 quantize_per_tensor_default_47" [id=789, type=quantize_per_tensor];
-"790 dequantize_per_tensor_default_47" [id=790, type=dequantize_per_tensor];
-"791 transpose_14" [id=791, type=transpose];
-"792 matmul_14" [id=792, type=matmul];
-"793 _frozen_param58" [id=793, type=get_attr];
-"794 mul_15" [id=794, type=mul];
-"795 add_24" [id=795, type=add];
-"796 new_zeros_3" [id=796, type=new_zeros];
-"797 view_40" [id=797, type=view];
-"798 permute_35" [id=798, type=permute];
-"799 reshape_33" [id=799, type=reshape];
-"800 unsqueeze_20" [id=800, type=unsqueeze];
-"801 unsqueeze_21" [id=801, type=unsqueeze];
-"802 sub_3" [id=802, type=sub];
-"803 ne_3" [id=803, type=ne];
-"804 masked_fill_6" [id=804, type=masked_fill];
-"805 eq_3" [id=805, type=eq];
-"806 masked_fill_7" [id=806, type=masked_fill];
-"807 view_41" [id=807, type=view];
-"808 unsqueeze_22" [id=808, type=unsqueeze];
-"809 unsqueeze_23" [id=809, type=unsqueeze];
-"810 add_25" [id=810, type=add];
-"811 view_42" [id=811, type=view];
-"812 softmax_7" [id=812, type=softmax];
-"813 dropout_28" [id=813, type=dropout];
-"814 matmul_15" [id=814, type=matmul];
-"815 transpose_15" [id=815, type=transpose];
-"816 reshape_34" [id=816, type=reshape];
-"817 reshape_34_0_0_nncf_smooth_quant_0" [id=817, type=call_module];
-"818 quantize_per_tensor_default_48" [id=818, type=quantize_per_tensor];
-"819 dequantize_per_tensor_default_48" [id=819, type=dequantize_per_tensor];
-"820 linear_47_scale_0" [id=820, type=get_attr];
-"821 linear_47_zero_point_0" [id=821, type=get_attr];
-"822 compressed_weight_updated_constant32" [id=822, type=get_attr];
-"823 quantize_per_channel_default_40" [id=823, type=quantize_per_channel];
-"824 dequantize_per_channel_default_40" [id=824, type=dequantize_per_channel];
-"825 _param_constant129_0_0" [id=825, type=get_attr];
-"826 linear_47" [id=826, type=linear];
-"827 dropout_29" [id=827, type=dropout];
-"828 view_43" [id=828, type=view];
-"829 permute_36" [id=829, type=permute];
-"830 reshape_35" [id=830, type=reshape];
-"831 roll_7" [id=831, type=roll];
-"832 slice_129" [id=832, type=slice];
-"833 slice_130" [id=833, type=slice];
-"834 slice_131" [id=834, type=slice];
-"835 slice_132" [id=835, type=slice];
-"836 contiguous_13" [id=836, type=contiguous];
-"837 _param_constant130" [id=837, type=get_attr];
-"838 _param_constant131" [id=838, type=get_attr];
-"839 layer_norm_17" [id=839, type=layer_norm];
-"840 add_26" [id=840, type=add];
-"841 add_26_0_0_nncf_smooth_quant_0" [id=841, type=call_module];
-"842 quantize_per_tensor_default_49" [id=842, type=quantize_per_tensor];
-"843 dequantize_per_tensor_default_49" [id=843, type=dequantize_per_tensor];
-"844 linear_48_scale_0" [id=844, type=get_attr];
-"845 linear_48_zero_point_0" [id=845, type=get_attr];
-"846 compressed_weight_updated_constant33" [id=846, type=get_attr];
-"847 quantize_per_channel_default_41" [id=847, type=quantize_per_channel];
-"848 dequantize_per_channel_default_41" [id=848, type=dequantize_per_channel];
-"849 _param_constant133_0_0" [id=849, type=get_attr];
-"850 linear_48" [id=850, type=linear];
-"851 gelu_7" [id=851, type=gelu];
-"852 dropout_30" [id=852, type=dropout];
-"853 dropout_30_0_0_nncf_smooth_quant_0" [id=853, type=call_module];
-"854 quantize_per_tensor_default_50" [id=854, type=quantize_per_tensor];
-"855 dequantize_per_tensor_default_50" [id=855, type=dequantize_per_tensor];
-"856 linear_49_scale_0" [id=856, type=get_attr];
-"857 linear_49_zero_point_0" [id=857, type=get_attr];
-"858 compressed_weight_updated_constant34" [id=858, type=get_attr];
-"859 quantize_per_channel_default_42" [id=859, type=quantize_per_channel];
-"860 dequantize_per_channel_default_42" [id=860, type=dequantize_per_channel];
-"861 _param_constant135_0_0" [id=861, type=get_attr];
-"862 linear_49" [id=862, type=linear];
-"863 dropout_31" [id=863, type=dropout];
-"864 _param_constant136" [id=864, type=get_attr];
-"865 _param_constant137" [id=865, type=get_attr];
-"866 layer_norm_18" [id=866, type=layer_norm];
-"867 add_27" [id=867, type=add];
-"868 _frozen_param68" [id=868, type=get_attr];
-"869 pad_10" [id=869, type=pad];
-"870 view_46" [id=870, type=view];
-"871 permute_38" [id=871, type=permute];
-"872 reshape_36" [id=872, type=reshape];
-"873 reshape_36_0_0_nncf_smooth_quant_0" [id=873, type=call_module];
-"874 quantize_per_tensor_default_51" [id=874, type=quantize_per_tensor];
-"875 dequantize_per_tensor_default_51" [id=875, type=dequantize_per_tensor];
-"876 linear_52_scale_0" [id=876, type=get_attr];
-"877 linear_52_zero_point_0" [id=877, type=get_attr];
-"878 compressed_weight_updated_constant35" [id=878, type=get_attr];
-"879 quantize_per_channel_default_44" [id=879, type=quantize_per_channel];
-"880 dequantize_per_channel_default_44" [id=880, type=dequantize_per_channel];
-"881 _frozen_param69_0_0" [id=881, type=get_attr];
-"882 linear_52" [id=882, type=linear];
-"883 reshape_37" [id=883, type=reshape];
-"884 permute_39" [id=884, type=permute];
-"885 select_24" [id=885, type=select];
-"886 select_25" [id=886, type=select];
-"887 select_26" [id=887, type=select];
-"888 linalg_vector_norm_16" [id=888, type=linalg_vector_norm];
-"889 clamp_min_16" [id=889, type=clamp_min];
-"890 expand_as_16" [id=890, type=expand_as];
-"891 div_16" [id=891, type=div];
-"892 quantize_per_tensor_default_52" [id=892, type=quantize_per_tensor];
-"893 dequantize_per_tensor_default_52" [id=893, type=dequantize_per_tensor];
-"894 linalg_vector_norm_17" [id=894, type=linalg_vector_norm];
-"895 clamp_min_17" [id=895, type=clamp_min];
-"896 expand_as_17" [id=896, type=expand_as];
-"897 div_17" [id=897, type=div];
-"898 quantize_per_tensor_default_53" [id=898, type=quantize_per_tensor];
-"899 dequantize_per_tensor_default_53" [id=899, type=dequantize_per_tensor];
-"900 transpose_16" [id=900, type=transpose];
-"901 matmul_16" [id=901, type=matmul];
-"902 _frozen_param71" [id=902, type=get_attr];
-"903 mul_17" [id=903, type=mul];
-"904 add_28" [id=904, type=add];
-"905 softmax_8" [id=905, type=softmax];
-"906 dropout_32" [id=906, type=dropout];
-"907 matmul_17" [id=907, type=matmul];
-"908 transpose_17" [id=908, type=transpose];
-"909 reshape_38" [id=909, type=reshape];
-"910 reshape_38_0_0_nncf_smooth_quant_0" [id=910, type=call_module];
-"911 quantize_per_tensor_default_54" [id=911, type=quantize_per_tensor];
-"912 dequantize_per_tensor_default_54" [id=912, type=dequantize_per_tensor];
-"913 linear_53_scale_0" [id=913, type=get_attr];
-"914 linear_53_zero_point_0" [id=914, type=get_attr];
-"915 compressed_weight_updated_constant36" [id=915, type=get_attr];
-"916 quantize_per_channel_default_45" [id=916, type=quantize_per_channel];
-"917 dequantize_per_channel_default_45" [id=917, type=dequantize_per_channel];
-"918 _param_constant145_0_0" [id=918, type=get_attr];
-"919 linear_53" [id=919, type=linear];
-"920 dropout_33" [id=920, type=dropout];
-"921 view_47" [id=921, type=view];
-"922 permute_40" [id=922, type=permute];
-"923 reshape_39" [id=923, type=reshape];
-"924 slice_134" [id=924, type=slice];
-"925 slice_135" [id=925, type=slice];
-"926 slice_136" [id=926, type=slice];
-"927 slice_137" [id=927, type=slice];
-"928 contiguous_15" [id=928, type=contiguous];
-"929 _param_constant146" [id=929, type=get_attr];
-"930 _param_constant147" [id=930, type=get_attr];
-"931 layer_norm_19" [id=931, type=layer_norm];
-"932 add_29" [id=932, type=add];
-"933 add_29_0_0_nncf_smooth_quant_0" [id=933, type=call_module];
-"934 quantize_per_tensor_default_55" [id=934, type=quantize_per_tensor];
-"935 dequantize_per_tensor_default_55" [id=935, type=dequantize_per_tensor];
-"936 linear_54_scale_0" [id=936, type=get_attr];
-"937 linear_54_zero_point_0" [id=937, type=get_attr];
-"938 compressed_weight_updated_constant37" [id=938, type=get_attr];
-"939 quantize_per_channel_default_46" [id=939, type=quantize_per_channel];
-"940 dequantize_per_channel_default_46" [id=940, type=dequantize_per_channel];
-"941 _param_constant149_0_0" [id=941, type=get_attr];
-"942 linear_54" [id=942, type=linear];
-"943 gelu_8" [id=943, type=gelu];
-"944 dropout_34" [id=944, type=dropout];
-"945 dropout_34_0_0_nncf_smooth_quant_0" [id=945, type=call_module];
-"946 quantize_per_tensor_default_56" [id=946, type=quantize_per_tensor];
-"947 dequantize_per_tensor_default_56" [id=947, type=dequantize_per_tensor];
-"948 linear_55_scale_0" [id=948, type=get_attr];
-"949 linear_55_zero_point_0" [id=949, type=get_attr];
-"950 compressed_weight_updated_constant38" [id=950, type=get_attr];
-"951 quantize_per_channel_default_47" [id=951, type=quantize_per_channel];
-"952 dequantize_per_channel_default_47" [id=952, type=dequantize_per_channel];
-"953 _param_constant151_0_0" [id=953, type=get_attr];
-"954 linear_55" [id=954, type=linear];
-"955 dropout_35" [id=955, type=dropout];
-"956 _param_constant152" [id=956, type=get_attr];
-"957 _param_constant153" [id=957, type=get_attr];
-"958 layer_norm_20" [id=958, type=layer_norm];
-"959 add_30" [id=959, type=add];
-"960 _frozen_param72" [id=960, type=get_attr];
-"961 pad_11" [id=961, type=pad];
-"962 roll_8" [id=962, type=roll];
-"963 view_50" [id=963, type=view];
-"964 permute_42" [id=964, type=permute];
-"965 reshape_40" [id=965, type=reshape];
-"966 reshape_40_0_0_nncf_smooth_quant_0" [id=966, type=call_module];
-"967 quantize_per_tensor_default_57" [id=967, type=quantize_per_tensor];
-"968 dequantize_per_tensor_default_57" [id=968, type=dequantize_per_tensor];
-"969 linear_58_scale_0" [id=969, type=get_attr];
-"970 linear_58_zero_point_0" [id=970, type=get_attr];
-"971 compressed_weight_updated_constant39" [id=971, type=get_attr];
-"972 quantize_per_channel_default_49" [id=972, type=quantize_per_channel];
-"973 dequantize_per_channel_default_49" [id=973, type=dequantize_per_channel];
-"974 _frozen_param73_0_0" [id=974, type=get_attr];
-"975 linear_58" [id=975, type=linear];
-"976 reshape_41" [id=976, type=reshape];
-"977 permute_43" [id=977, type=permute];
-"978 select_27" [id=978, type=select];
-"979 select_28" [id=979, type=select];
-"980 select_29" [id=980, type=select];
-"981 linalg_vector_norm_18" [id=981, type=linalg_vector_norm];
-"982 clamp_min_18" [id=982, type=clamp_min];
-"983 expand_as_18" [id=983, type=expand_as];
-"984 div_18" [id=984, type=div];
-"985 quantize_per_tensor_default_58" [id=985, type=quantize_per_tensor];
-"986 dequantize_per_tensor_default_58" [id=986, type=dequantize_per_tensor];
-"987 linalg_vector_norm_19" [id=987, type=linalg_vector_norm];
-"988 clamp_min_19" [id=988, type=clamp_min];
-"989 expand_as_19" [id=989, type=expand_as];
-"990 div_19" [id=990, type=div];
-"991 quantize_per_tensor_default_59" [id=991, type=quantize_per_tensor];
-"992 dequantize_per_tensor_default_59" [id=992, type=dequantize_per_tensor];
-"993 transpose_18" [id=993, type=transpose];
-"994 matmul_18" [id=994, type=matmul];
-"995 _frozen_param75" [id=995, type=get_attr];
-"996 mul_19" [id=996, type=mul];
-"997 add_31" [id=997, type=add];
-"998 new_zeros_4" [id=998, type=new_zeros];
-"999 view_51" [id=999, type=view];
-"1000 permute_44" [id=1000, type=permute];
-"1001 reshape_42" [id=1001, type=reshape];
-"1002 unsqueeze_26" [id=1002, type=unsqueeze];
-"1003 unsqueeze_27" [id=1003, type=unsqueeze];
-"1004 sub_4" [id=1004, type=sub];
-"1005 ne_4" [id=1005, type=ne];
-"1006 masked_fill_8" [id=1006, type=masked_fill];
-"1007 eq_4" [id=1007, type=eq];
-"1008 masked_fill_9" [id=1008, type=masked_fill];
-"1009 view_52" [id=1009, type=view];
-"1010 unsqueeze_28" [id=1010, type=unsqueeze];
-"1011 unsqueeze_29" [id=1011, type=unsqueeze];
-"1012 add_32" [id=1012, type=add];
-"1013 view_53" [id=1013, type=view];
-"1014 softmax_9" [id=1014, type=softmax];
-"1015 dropout_36" [id=1015, type=dropout];
-"1016 matmul_19" [id=1016, type=matmul];
-"1017 transpose_19" [id=1017, type=transpose];
-"1018 reshape_43" [id=1018, type=reshape];
-"1019 reshape_43_0_0_nncf_smooth_quant_0" [id=1019, type=call_module];
-"1020 quantize_per_tensor_default_60" [id=1020, type=quantize_per_tensor];
-"1021 dequantize_per_tensor_default_60" [id=1021, type=dequantize_per_tensor];
-"1022 linear_59_scale_0" [id=1022, type=get_attr];
-"1023 linear_59_zero_point_0" [id=1023, type=get_attr];
-"1024 compressed_weight_updated_constant40" [id=1024, type=get_attr];
-"1025 quantize_per_channel_default_50" [id=1025, type=quantize_per_channel];
-"1026 dequantize_per_channel_default_50" [id=1026, type=dequantize_per_channel];
-"1027 _param_constant161_0_0" [id=1027, type=get_attr];
-"1028 linear_59" [id=1028, type=linear];
-"1029 dropout_37" [id=1029, type=dropout];
-"1030 view_54" [id=1030, type=view];
-"1031 permute_45" [id=1031, type=permute];
-"1032 reshape_44" [id=1032, type=reshape];
-"1033 roll_9" [id=1033, type=roll];
-"1034 slice_157" [id=1034, type=slice];
-"1035 slice_158" [id=1035, type=slice];
-"1036 slice_159" [id=1036, type=slice];
-"1037 slice_160" [id=1037, type=slice];
-"1038 contiguous_17" [id=1038, type=contiguous];
-"1039 _param_constant162" [id=1039, type=get_attr];
-"1040 _param_constant163" [id=1040, type=get_attr];
-"1041 layer_norm_21" [id=1041, type=layer_norm];
-"1042 add_33" [id=1042, type=add];
-"1043 add_33_0_0_nncf_smooth_quant_0" [id=1043, type=call_module];
-"1044 quantize_per_tensor_default_61" [id=1044, type=quantize_per_tensor];
-"1045 dequantize_per_tensor_default_61" [id=1045, type=dequantize_per_tensor];
-"1046 linear_60_scale_0" [id=1046, type=get_attr];
-"1047 linear_60_zero_point_0" [id=1047, type=get_attr];
-"1048 compressed_weight_updated_constant41" [id=1048, type=get_attr];
-"1049 quantize_per_channel_default_51" [id=1049, type=quantize_per_channel];
-"1050 dequantize_per_channel_default_51" [id=1050, type=dequantize_per_channel];
-"1051 _param_constant165_0_0" [id=1051, type=get_attr];
-"1052 linear_60" [id=1052, type=linear];
-"1053 gelu_9" [id=1053, type=gelu];
-"1054 dropout_38" [id=1054, type=dropout];
-"1055 dropout_38_0_0_nncf_smooth_quant_0" [id=1055, type=call_module];
-"1056 quantize_per_tensor_default_62" [id=1056, type=quantize_per_tensor];
-"1057 dequantize_per_tensor_default_62" [id=1057, type=dequantize_per_tensor];
-"1058 linear_61_scale_0" [id=1058, type=get_attr];
-"1059 linear_61_zero_point_0" [id=1059, type=get_attr];
-"1060 compressed_weight_updated_constant42" [id=1060, type=get_attr];
-"1061 quantize_per_channel_default_52" [id=1061, type=quantize_per_channel];
-"1062 dequantize_per_channel_default_52" [id=1062, type=dequantize_per_channel];
-"1063 _param_constant167_0_0" [id=1063, type=get_attr];
-"1064 linear_61" [id=1064, type=linear];
-"1065 dropout_39" [id=1065, type=dropout];
-"1066 _param_constant168" [id=1066, type=get_attr];
-"1067 _param_constant169" [id=1067, type=get_attr];
-"1068 layer_norm_22" [id=1068, type=layer_norm];
-"1069 add_34" [id=1069, type=add];
-"1070 _frozen_param85" [id=1070, type=get_attr];
-"1071 pad_12" [id=1071, type=pad];
-"1072 view_57" [id=1072, type=view];
-"1073 permute_47" [id=1073, type=permute];
-"1074 reshape_45" [id=1074, type=reshape];
-"1075 reshape_45_0_0_nncf_smooth_quant_0" [id=1075, type=call_module];
-"1076 quantize_per_tensor_default_63" [id=1076, type=quantize_per_tensor];
-"1077 dequantize_per_tensor_default_63" [id=1077, type=dequantize_per_tensor];
-"1078 linear_64_scale_0" [id=1078, type=get_attr];
-"1079 linear_64_zero_point_0" [id=1079, type=get_attr];
-"1080 compressed_weight_updated_constant43" [id=1080, type=get_attr];
-"1081 quantize_per_channel_default_54" [id=1081, type=quantize_per_channel];
-"1082 dequantize_per_channel_default_54" [id=1082, type=dequantize_per_channel];
-"1083 _frozen_param86_0_0" [id=1083, type=get_attr];
-"1084 linear_64" [id=1084, type=linear];
-"1085 reshape_46" [id=1085, type=reshape];
-"1086 permute_48" [id=1086, type=permute];
-"1087 select_30" [id=1087, type=select];
-"1088 select_31" [id=1088, type=select];
-"1089 select_32" [id=1089, type=select];
-"1090 linalg_vector_norm_20" [id=1090, type=linalg_vector_norm];
-"1091 clamp_min_20" [id=1091, type=clamp_min];
-"1092 expand_as_20" [id=1092, type=expand_as];
-"1093 div_20" [id=1093, type=div];
-"1094 quantize_per_tensor_default_64" [id=1094, type=quantize_per_tensor];
-"1095 dequantize_per_tensor_default_64" [id=1095, type=dequantize_per_tensor];
-"1096 linalg_vector_norm_21" [id=1096, type=linalg_vector_norm];
-"1097 clamp_min_21" [id=1097, type=clamp_min];
-"1098 expand_as_21" [id=1098, type=expand_as];
-"1099 div_21" [id=1099, type=div];
-"1100 quantize_per_tensor_default_65" [id=1100, type=quantize_per_tensor];
-"1101 dequantize_per_tensor_default_65" [id=1101, type=dequantize_per_tensor];
-"1102 transpose_20" [id=1102, type=transpose];
-"1103 matmul_20" [id=1103, type=matmul];
-"1104 _frozen_param88" [id=1104, type=get_attr];
-"1105 mul_21" [id=1105, type=mul];
-"1106 add_35" [id=1106, type=add];
-"1107 softmax_10" [id=1107, type=softmax];
-"1108 dropout_40" [id=1108, type=dropout];
-"1109 matmul_21" [id=1109, type=matmul];
-"1110 transpose_21" [id=1110, type=transpose];
-"1111 reshape_47" [id=1111, type=reshape];
-"1112 reshape_47_0_0_nncf_smooth_quant_0" [id=1112, type=call_module];
-"1113 quantize_per_tensor_default_66" [id=1113, type=quantize_per_tensor];
-"1114 dequantize_per_tensor_default_66" [id=1114, type=dequantize_per_tensor];
-"1115 linear_65_scale_0" [id=1115, type=get_attr];
-"1116 linear_65_zero_point_0" [id=1116, type=get_attr];
-"1117 compressed_weight_updated_constant44" [id=1117, type=get_attr];
-"1118 quantize_per_channel_default_55" [id=1118, type=quantize_per_channel];
-"1119 dequantize_per_channel_default_55" [id=1119, type=dequantize_per_channel];
-"1120 _param_constant177_0_0" [id=1120, type=get_attr];
-"1121 linear_65" [id=1121, type=linear];
-"1122 dropout_41" [id=1122, type=dropout];
-"1123 view_58" [id=1123, type=view];
-"1124 permute_49" [id=1124, type=permute];
-"1125 reshape_48" [id=1125, type=reshape];
-"1126 slice_162" [id=1126, type=slice];
-"1127 slice_163" [id=1127, type=slice];
-"1128 slice_164" [id=1128, type=slice];
-"1129 slice_165" [id=1129, type=slice];
-"1130 contiguous_19" [id=1130, type=contiguous];
-"1131 _param_constant178" [id=1131, type=get_attr];
-"1132 _param_constant179" [id=1132, type=get_attr];
-"1133 layer_norm_23" [id=1133, type=layer_norm];
-"1134 add_36" [id=1134, type=add];
-"1135 add_36_0_0_nncf_smooth_quant_0" [id=1135, type=call_module];
-"1136 quantize_per_tensor_default_67" [id=1136, type=quantize_per_tensor];
-"1137 dequantize_per_tensor_default_67" [id=1137, type=dequantize_per_tensor];
-"1138 linear_66_scale_0" [id=1138, type=get_attr];
-"1139 linear_66_zero_point_0" [id=1139, type=get_attr];
-"1140 compressed_weight_updated_constant45" [id=1140, type=get_attr];
-"1141 quantize_per_channel_default_56" [id=1141, type=quantize_per_channel];
-"1142 dequantize_per_channel_default_56" [id=1142, type=dequantize_per_channel];
-"1143 _param_constant181_0_0" [id=1143, type=get_attr];
-"1144 linear_66" [id=1144, type=linear];
-"1145 gelu_10" [id=1145, type=gelu];
-"1146 dropout_42" [id=1146, type=dropout];
-"1147 dropout_42_0_0_nncf_smooth_quant_0" [id=1147, type=call_module];
-"1148 quantize_per_tensor_default_68" [id=1148, type=quantize_per_tensor];
-"1149 dequantize_per_tensor_default_68" [id=1149, type=dequantize_per_tensor];
-"1150 linear_67_scale_0" [id=1150, type=get_attr];
-"1151 linear_67_zero_point_0" [id=1151, type=get_attr];
-"1152 compressed_weight_updated_constant46" [id=1152, type=get_attr];
-"1153 quantize_per_channel_default_57" [id=1153, type=quantize_per_channel];
-"1154 dequantize_per_channel_default_57" [id=1154, type=dequantize_per_channel];
-"1155 _param_constant183_0_0" [id=1155, type=get_attr];
-"1156 linear_67" [id=1156, type=linear];
-"1157 dropout_43" [id=1157, type=dropout];
-"1158 _param_constant184" [id=1158, type=get_attr];
-"1159 _param_constant185" [id=1159, type=get_attr];
-"1160 layer_norm_24" [id=1160, type=layer_norm];
-"1161 add_37" [id=1161, type=add];
-"1162 _frozen_param89" [id=1162, type=get_attr];
-"1163 pad_13" [id=1163, type=pad];
-"1164 roll_10" [id=1164, type=roll];
-"1165 view_61" [id=1165, type=view];
-"1166 permute_51" [id=1166, type=permute];
-"1167 reshape_49" [id=1167, type=reshape];
-"1168 reshape_49_0_0_nncf_smooth_quant_0" [id=1168, type=call_module];
-"1169 quantize_per_tensor_default_69" [id=1169, type=quantize_per_tensor];
-"1170 dequantize_per_tensor_default_69" [id=1170, type=dequantize_per_tensor];
-"1171 linear_70_scale_0" [id=1171, type=get_attr];
-"1172 linear_70_zero_point_0" [id=1172, type=get_attr];
-"1173 compressed_weight_updated_constant47" [id=1173, type=get_attr];
-"1174 quantize_per_channel_default_59" [id=1174, type=quantize_per_channel];
-"1175 dequantize_per_channel_default_59" [id=1175, type=dequantize_per_channel];
-"1176 _frozen_param90_0_0" [id=1176, type=get_attr];
-"1177 linear_70" [id=1177, type=linear];
-"1178 reshape_50" [id=1178, type=reshape];
-"1179 permute_52" [id=1179, type=permute];
-"1180 select_33" [id=1180, type=select];
-"1181 select_34" [id=1181, type=select];
-"1182 select_35" [id=1182, type=select];
-"1183 linalg_vector_norm_22" [id=1183, type=linalg_vector_norm];
-"1184 clamp_min_22" [id=1184, type=clamp_min];
-"1185 expand_as_22" [id=1185, type=expand_as];
-"1186 div_22" [id=1186, type=div];
-"1187 quantize_per_tensor_default_70" [id=1187, type=quantize_per_tensor];
-"1188 dequantize_per_tensor_default_70" [id=1188, type=dequantize_per_tensor];
-"1189 linalg_vector_norm_23" [id=1189, type=linalg_vector_norm];
-"1190 clamp_min_23" [id=1190, type=clamp_min];
-"1191 expand_as_23" [id=1191, type=expand_as];
-"1192 div_23" [id=1192, type=div];
-"1193 quantize_per_tensor_default_71" [id=1193, type=quantize_per_tensor];
-"1194 dequantize_per_tensor_default_71" [id=1194, type=dequantize_per_tensor];
-"1195 transpose_22" [id=1195, type=transpose];
-"1196 matmul_22" [id=1196, type=matmul];
-"1197 _frozen_param92" [id=1197, type=get_attr];
-"1198 mul_23" [id=1198, type=mul];
-"1199 add_38" [id=1199, type=add];
-"1200 new_zeros_5" [id=1200, type=new_zeros];
-"1201 view_62" [id=1201, type=view];
-"1202 permute_53" [id=1202, type=permute];
-"1203 reshape_51" [id=1203, type=reshape];
-"1204 unsqueeze_32" [id=1204, type=unsqueeze];
-"1205 unsqueeze_33" [id=1205, type=unsqueeze];
-"1206 sub_5" [id=1206, type=sub];
-"1207 ne_5" [id=1207, type=ne];
-"1208 masked_fill_10" [id=1208, type=masked_fill];
-"1209 eq_5" [id=1209, type=eq];
-"1210 masked_fill_11" [id=1210, type=masked_fill];
-"1211 view_63" [id=1211, type=view];
-"1212 unsqueeze_34" [id=1212, type=unsqueeze];
-"1213 unsqueeze_35" [id=1213, type=unsqueeze];
-"1214 add_39" [id=1214, type=add];
-"1215 view_64" [id=1215, type=view];
-"1216 softmax_11" [id=1216, type=softmax];
-"1217 dropout_44" [id=1217, type=dropout];
-"1218 matmul_23" [id=1218, type=matmul];
-"1219 transpose_23" [id=1219, type=transpose];
-"1220 reshape_52" [id=1220, type=reshape];
-"1221 reshape_52_0_0_nncf_smooth_quant_0" [id=1221, type=call_module];
-"1222 quantize_per_tensor_default_72" [id=1222, type=quantize_per_tensor];
-"1223 dequantize_per_tensor_default_72" [id=1223, type=dequantize_per_tensor];
-"1224 linear_71_scale_0" [id=1224, type=get_attr];
-"1225 linear_71_zero_point_0" [id=1225, type=get_attr];
-"1226 compressed_weight_updated_constant48" [id=1226, type=get_attr];
-"1227 quantize_per_channel_default_60" [id=1227, type=quantize_per_channel];
-"1228 dequantize_per_channel_default_60" [id=1228, type=dequantize_per_channel];
-"1229 _param_constant193_0_0" [id=1229, type=get_attr];
-"1230 linear_71" [id=1230, type=linear];
-"1231 dropout_45" [id=1231, type=dropout];
-"1232 view_65" [id=1232, type=view];
-"1233 permute_54" [id=1233, type=permute];
-"1234 reshape_53" [id=1234, type=reshape];
-"1235 roll_11" [id=1235, type=roll];
-"1236 slice_185" [id=1236, type=slice];
-"1237 slice_186" [id=1237, type=slice];
-"1238 slice_187" [id=1238, type=slice];
-"1239 slice_188" [id=1239, type=slice];
-"1240 contiguous_21" [id=1240, type=contiguous];
-"1241 _param_constant194" [id=1241, type=get_attr];
-"1242 _param_constant195" [id=1242, type=get_attr];
-"1243 layer_norm_25" [id=1243, type=layer_norm];
-"1244 add_40" [id=1244, type=add];
-"1245 add_40_0_0_nncf_smooth_quant_0" [id=1245, type=call_module];
-"1246 quantize_per_tensor_default_73" [id=1246, type=quantize_per_tensor];
-"1247 dequantize_per_tensor_default_73" [id=1247, type=dequantize_per_tensor];
-"1248 linear_72_scale_0" [id=1248, type=get_attr];
-"1249 linear_72_zero_point_0" [id=1249, type=get_attr];
-"1250 compressed_weight_updated_constant49" [id=1250, type=get_attr];
-"1251 quantize_per_channel_default_61" [id=1251, type=quantize_per_channel];
-"1252 dequantize_per_channel_default_61" [id=1252, type=dequantize_per_channel];
-"1253 _param_constant197_0_0" [id=1253, type=get_attr];
-"1254 linear_72" [id=1254, type=linear];
-"1255 gelu_11" [id=1255, type=gelu];
-"1256 dropout_46" [id=1256, type=dropout];
-"1257 dropout_46_0_0_nncf_smooth_quant_0" [id=1257, type=call_module];
-"1258 quantize_per_tensor_default_74" [id=1258, type=quantize_per_tensor];
-"1259 dequantize_per_tensor_default_74" [id=1259, type=dequantize_per_tensor];
-"1260 linear_73_scale_0" [id=1260, type=get_attr];
-"1261 linear_73_zero_point_0" [id=1261, type=get_attr];
-"1262 compressed_weight_updated_constant50" [id=1262, type=get_attr];
-"1263 quantize_per_channel_default_62" [id=1263, type=quantize_per_channel];
-"1264 dequantize_per_channel_default_62" [id=1264, type=dequantize_per_channel];
-"1265 _param_constant199_0_0" [id=1265, type=get_attr];
-"1266 linear_73" [id=1266, type=linear];
-"1267 dropout_47" [id=1267, type=dropout];
-"1268 _param_constant200" [id=1268, type=get_attr];
-"1269 _param_constant201" [id=1269, type=get_attr];
-"1270 layer_norm_26" [id=1270, type=layer_norm];
-"1271 add_41" [id=1271, type=add];
-"1272 _frozen_param102" [id=1272, type=get_attr];
-"1273 pad_14" [id=1273, type=pad];
-"1274 view_68" [id=1274, type=view];
-"1275 permute_56" [id=1275, type=permute];
-"1276 reshape_54" [id=1276, type=reshape];
-"1277 reshape_54_0_0_nncf_smooth_quant_0" [id=1277, type=call_module];
-"1278 quantize_per_tensor_default_75" [id=1278, type=quantize_per_tensor];
-"1279 dequantize_per_tensor_default_75" [id=1279, type=dequantize_per_tensor];
-"1280 linear_76_scale_0" [id=1280, type=get_attr];
-"1281 linear_76_zero_point_0" [id=1281, type=get_attr];
-"1282 compressed_weight_updated_constant51" [id=1282, type=get_attr];
-"1283 quantize_per_channel_default_64" [id=1283, type=quantize_per_channel];
-"1284 dequantize_per_channel_default_64" [id=1284, type=dequantize_per_channel];
-"1285 _frozen_param103_0_0" [id=1285, type=get_attr];
-"1286 linear_76" [id=1286, type=linear];
-"1287 reshape_55" [id=1287, type=reshape];
-"1288 permute_57" [id=1288, type=permute];
-"1289 select_36" [id=1289, type=select];
-"1290 select_37" [id=1290, type=select];
-"1291 select_38" [id=1291, type=select];
-"1292 linalg_vector_norm_24" [id=1292, type=linalg_vector_norm];
-"1293 clamp_min_24" [id=1293, type=clamp_min];
-"1294 expand_as_24" [id=1294, type=expand_as];
-"1295 div_24" [id=1295, type=div];
-"1296 quantize_per_tensor_default_76" [id=1296, type=quantize_per_tensor];
-"1297 dequantize_per_tensor_default_76" [id=1297, type=dequantize_per_tensor];
-"1298 linalg_vector_norm_25" [id=1298, type=linalg_vector_norm];
-"1299 clamp_min_25" [id=1299, type=clamp_min];
-"1300 expand_as_25" [id=1300, type=expand_as];
-"1301 div_25" [id=1301, type=div];
-"1302 quantize_per_tensor_default_77" [id=1302, type=quantize_per_tensor];
-"1303 dequantize_per_tensor_default_77" [id=1303, type=dequantize_per_tensor];
-"1304 transpose_24" [id=1304, type=transpose];
-"1305 matmul_24" [id=1305, type=matmul];
-"1306 _frozen_param105" [id=1306, type=get_attr];
-"1307 mul_25" [id=1307, type=mul];
-"1308 add_42" [id=1308, type=add];
-"1309 softmax_12" [id=1309, type=softmax];
-"1310 dropout_48" [id=1310, type=dropout];
-"1311 matmul_25" [id=1311, type=matmul];
-"1312 transpose_25" [id=1312, type=transpose];
-"1313 reshape_56" [id=1313, type=reshape];
-"1314 reshape_56_0_0_nncf_smooth_quant_0" [id=1314, type=call_module];
-"1315 quantize_per_tensor_default_78" [id=1315, type=quantize_per_tensor];
-"1316 dequantize_per_tensor_default_78" [id=1316, type=dequantize_per_tensor];
-"1317 linear_77_scale_0" [id=1317, type=get_attr];
-"1318 linear_77_zero_point_0" [id=1318, type=get_attr];
-"1319 compressed_weight_updated_constant52" [id=1319, type=get_attr];
-"1320 quantize_per_channel_default_65" [id=1320, type=quantize_per_channel];
-"1321 dequantize_per_channel_default_65" [id=1321, type=dequantize_per_channel];
-"1322 _param_constant209_0_0" [id=1322, type=get_attr];
-"1323 linear_77" [id=1323, type=linear];
-"1324 dropout_49" [id=1324, type=dropout];
-"1325 view_69" [id=1325, type=view];
-"1326 permute_58" [id=1326, type=permute];
-"1327 reshape_57" [id=1327, type=reshape];
-"1328 slice_190" [id=1328, type=slice];
-"1329 slice_191" [id=1329, type=slice];
-"1330 slice_192" [id=1330, type=slice];
-"1331 slice_193" [id=1331, type=slice];
-"1332 contiguous_23" [id=1332, type=contiguous];
-"1333 _param_constant210" [id=1333, type=get_attr];
-"1334 _param_constant211" [id=1334, type=get_attr];
-"1335 layer_norm_27" [id=1335, type=layer_norm];
-"1336 add_43" [id=1336, type=add];
-"1337 add_43_0_0_nncf_smooth_quant_0" [id=1337, type=call_module];
-"1338 quantize_per_tensor_default_79" [id=1338, type=quantize_per_tensor];
-"1339 dequantize_per_tensor_default_79" [id=1339, type=dequantize_per_tensor];
-"1340 linear_78_scale_0" [id=1340, type=get_attr];
-"1341 linear_78_zero_point_0" [id=1341, type=get_attr];
-"1342 compressed_weight_updated_constant53" [id=1342, type=get_attr];
-"1343 quantize_per_channel_default_66" [id=1343, type=quantize_per_channel];
-"1344 dequantize_per_channel_default_66" [id=1344, type=dequantize_per_channel];
-"1345 _param_constant213_0_0" [id=1345, type=get_attr];
-"1346 linear_78" [id=1346, type=linear];
-"1347 gelu_12" [id=1347, type=gelu];
-"1348 dropout_50" [id=1348, type=dropout];
-"1349 dropout_50_0_0_nncf_smooth_quant_0" [id=1349, type=call_module];
-"1350 quantize_per_tensor_default_80" [id=1350, type=quantize_per_tensor];
-"1351 dequantize_per_tensor_default_80" [id=1351, type=dequantize_per_tensor];
-"1352 linear_79_scale_0" [id=1352, type=get_attr];
-"1353 linear_79_zero_point_0" [id=1353, type=get_attr];
-"1354 compressed_weight_updated_constant54" [id=1354, type=get_attr];
-"1355 quantize_per_channel_default_67" [id=1355, type=quantize_per_channel];
-"1356 dequantize_per_channel_default_67" [id=1356, type=dequantize_per_channel];
-"1357 _param_constant215_0_0" [id=1357, type=get_attr];
-"1358 linear_79" [id=1358, type=linear];
-"1359 dropout_51" [id=1359, type=dropout];
-"1360 _param_constant216" [id=1360, type=get_attr];
-"1361 _param_constant217" [id=1361, type=get_attr];
-"1362 layer_norm_28" [id=1362, type=layer_norm];
-"1363 add_44" [id=1363, type=add];
-"1364 _frozen_param106" [id=1364, type=get_attr];
-"1365 pad_15" [id=1365, type=pad];
-"1366 roll_12" [id=1366, type=roll];
-"1367 view_72" [id=1367, type=view];
-"1368 permute_60" [id=1368, type=permute];
-"1369 reshape_58" [id=1369, type=reshape];
-"1370 reshape_58_0_0_nncf_smooth_quant_0" [id=1370, type=call_module];
-"1371 quantize_per_tensor_default_81" [id=1371, type=quantize_per_tensor];
-"1372 dequantize_per_tensor_default_81" [id=1372, type=dequantize_per_tensor];
-"1373 linear_82_scale_0" [id=1373, type=get_attr];
-"1374 linear_82_zero_point_0" [id=1374, type=get_attr];
-"1375 compressed_weight_updated_constant55" [id=1375, type=get_attr];
-"1376 quantize_per_channel_default_69" [id=1376, type=quantize_per_channel];
-"1377 dequantize_per_channel_default_69" [id=1377, type=dequantize_per_channel];
-"1378 _frozen_param107_0_0" [id=1378, type=get_attr];
-"1379 linear_82" [id=1379, type=linear];
-"1380 reshape_59" [id=1380, type=reshape];
-"1381 permute_61" [id=1381, type=permute];
-"1382 select_39" [id=1382, type=select];
-"1383 select_40" [id=1383, type=select];
-"1384 select_41" [id=1384, type=select];
-"1385 linalg_vector_norm_26" [id=1385, type=linalg_vector_norm];
-"1386 clamp_min_26" [id=1386, type=clamp_min];
-"1387 expand_as_26" [id=1387, type=expand_as];
-"1388 div_26" [id=1388, type=div];
-"1389 quantize_per_tensor_default_82" [id=1389, type=quantize_per_tensor];
-"1390 dequantize_per_tensor_default_82" [id=1390, type=dequantize_per_tensor];
-"1391 linalg_vector_norm_27" [id=1391, type=linalg_vector_norm];
-"1392 clamp_min_27" [id=1392, type=clamp_min];
-"1393 expand_as_27" [id=1393, type=expand_as];
-"1394 div_27" [id=1394, type=div];
-"1395 quantize_per_tensor_default_83" [id=1395, type=quantize_per_tensor];
-"1396 dequantize_per_tensor_default_83" [id=1396, type=dequantize_per_tensor];
-"1397 transpose_26" [id=1397, type=transpose];
-"1398 matmul_26" [id=1398, type=matmul];
-"1399 _frozen_param109" [id=1399, type=get_attr];
-"1400 mul_27" [id=1400, type=mul];
-"1401 add_45" [id=1401, type=add];
-"1402 new_zeros_6" [id=1402, type=new_zeros];
-"1403 view_73" [id=1403, type=view];
-"1404 permute_62" [id=1404, type=permute];
-"1405 reshape_60" [id=1405, type=reshape];
-"1406 unsqueeze_38" [id=1406, type=unsqueeze];
-"1407 unsqueeze_39" [id=1407, type=unsqueeze];
-"1408 sub_6" [id=1408, type=sub];
-"1409 ne_6" [id=1409, type=ne];
-"1410 masked_fill_12" [id=1410, type=masked_fill];
-"1411 eq_6" [id=1411, type=eq];
-"1412 masked_fill_13" [id=1412, type=masked_fill];
-"1413 view_74" [id=1413, type=view];
-"1414 unsqueeze_40" [id=1414, type=unsqueeze];
-"1415 unsqueeze_41" [id=1415, type=unsqueeze];
-"1416 add_46" [id=1416, type=add];
-"1417 view_75" [id=1417, type=view];
-"1418 softmax_13" [id=1418, type=softmax];
-"1419 dropout_52" [id=1419, type=dropout];
-"1420 matmul_27" [id=1420, type=matmul];
-"1421 transpose_27" [id=1421, type=transpose];
-"1422 reshape_61" [id=1422, type=reshape];
-"1423 reshape_61_0_0_nncf_smooth_quant_0" [id=1423, type=call_module];
-"1424 quantize_per_tensor_default_84" [id=1424, type=quantize_per_tensor];
-"1425 dequantize_per_tensor_default_84" [id=1425, type=dequantize_per_tensor];
-"1426 linear_83_scale_0" [id=1426, type=get_attr];
-"1427 linear_83_zero_point_0" [id=1427, type=get_attr];
-"1428 compressed_weight_updated_constant56" [id=1428, type=get_attr];
-"1429 quantize_per_channel_default_70" [id=1429, type=quantize_per_channel];
-"1430 dequantize_per_channel_default_70" [id=1430, type=dequantize_per_channel];
-"1431 _param_constant225_0_0" [id=1431, type=get_attr];
-"1432 linear_83" [id=1432, type=linear];
-"1433 dropout_53" [id=1433, type=dropout];
-"1434 view_76" [id=1434, type=view];
-"1435 permute_63" [id=1435, type=permute];
-"1436 reshape_62" [id=1436, type=reshape];
-"1437 roll_13" [id=1437, type=roll];
-"1438 slice_213" [id=1438, type=slice];
-"1439 slice_214" [id=1439, type=slice];
-"1440 slice_215" [id=1440, type=slice];
-"1441 slice_216" [id=1441, type=slice];
-"1442 contiguous_25" [id=1442, type=contiguous];
-"1443 _param_constant226" [id=1443, type=get_attr];
-"1444 _param_constant227" [id=1444, type=get_attr];
-"1445 layer_norm_29" [id=1445, type=layer_norm];
-"1446 add_47" [id=1446, type=add];
-"1447 add_47_0_0_nncf_smooth_quant_0" [id=1447, type=call_module];
-"1448 quantize_per_tensor_default_85" [id=1448, type=quantize_per_tensor];
-"1449 dequantize_per_tensor_default_85" [id=1449, type=dequantize_per_tensor];
-"1450 linear_84_scale_0" [id=1450, type=get_attr];
-"1451 linear_84_zero_point_0" [id=1451, type=get_attr];
-"1452 compressed_weight_updated_constant57" [id=1452, type=get_attr];
-"1453 quantize_per_channel_default_71" [id=1453, type=quantize_per_channel];
-"1454 dequantize_per_channel_default_71" [id=1454, type=dequantize_per_channel];
-"1455 _param_constant229_0_0" [id=1455, type=get_attr];
-"1456 linear_84" [id=1456, type=linear];
-"1457 gelu_13" [id=1457, type=gelu];
-"1458 dropout_54" [id=1458, type=dropout];
-"1459 dropout_54_0_0_nncf_smooth_quant_0" [id=1459, type=call_module];
-"1460 quantize_per_tensor_default_86" [id=1460, type=quantize_per_tensor];
-"1461 dequantize_per_tensor_default_86" [id=1461, type=dequantize_per_tensor];
-"1462 linear_85_scale_0" [id=1462, type=get_attr];
-"1463 linear_85_zero_point_0" [id=1463, type=get_attr];
-"1464 compressed_weight_updated_constant58" [id=1464, type=get_attr];
-"1465 quantize_per_channel_default_72" [id=1465, type=quantize_per_channel];
-"1466 dequantize_per_channel_default_72" [id=1466, type=dequantize_per_channel];
-"1467 _param_constant231_0_0" [id=1467, type=get_attr];
-"1468 linear_85" [id=1468, type=linear];
-"1469 dropout_55" [id=1469, type=dropout];
-"1470 _param_constant232" [id=1470, type=get_attr];
-"1471 _param_constant233" [id=1471, type=get_attr];
-"1472 layer_norm_30" [id=1472, type=layer_norm];
-"1473 add_48" [id=1473, type=add];
-"1474 _frozen_param119" [id=1474, type=get_attr];
-"1475 pad_16" [id=1475, type=pad];
-"1476 view_79" [id=1476, type=view];
-"1477 permute_65" [id=1477, type=permute];
-"1478 reshape_63" [id=1478, type=reshape];
-"1479 reshape_63_0_0_nncf_smooth_quant_0" [id=1479, type=call_module];
-"1480 quantize_per_tensor_default_87" [id=1480, type=quantize_per_tensor];
-"1481 dequantize_per_tensor_default_87" [id=1481, type=dequantize_per_tensor];
-"1482 linear_88_scale_0" [id=1482, type=get_attr];
-"1483 linear_88_zero_point_0" [id=1483, type=get_attr];
-"1484 compressed_weight_updated_constant59" [id=1484, type=get_attr];
-"1485 quantize_per_channel_default_74" [id=1485, type=quantize_per_channel];
-"1486 dequantize_per_channel_default_74" [id=1486, type=dequantize_per_channel];
-"1487 _frozen_param120_0_0" [id=1487, type=get_attr];
-"1488 linear_88" [id=1488, type=linear];
-"1489 reshape_64" [id=1489, type=reshape];
-"1490 permute_66" [id=1490, type=permute];
-"1491 select_42" [id=1491, type=select];
-"1492 select_43" [id=1492, type=select];
-"1493 select_44" [id=1493, type=select];
-"1494 linalg_vector_norm_28" [id=1494, type=linalg_vector_norm];
-"1495 clamp_min_28" [id=1495, type=clamp_min];
-"1496 expand_as_28" [id=1496, type=expand_as];
-"1497 div_28" [id=1497, type=div];
-"1498 quantize_per_tensor_default_88" [id=1498, type=quantize_per_tensor];
-"1499 dequantize_per_tensor_default_88" [id=1499, type=dequantize_per_tensor];
-"1500 linalg_vector_norm_29" [id=1500, type=linalg_vector_norm];
-"1501 clamp_min_29" [id=1501, type=clamp_min];
-"1502 expand_as_29" [id=1502, type=expand_as];
-"1503 div_29" [id=1503, type=div];
-"1504 quantize_per_tensor_default_89" [id=1504, type=quantize_per_tensor];
-"1505 dequantize_per_tensor_default_89" [id=1505, type=dequantize_per_tensor];
-"1506 transpose_28" [id=1506, type=transpose];
-"1507 matmul_28" [id=1507, type=matmul];
-"1508 _frozen_param122" [id=1508, type=get_attr];
-"1509 mul_29" [id=1509, type=mul];
-"1510 add_49" [id=1510, type=add];
-"1511 softmax_14" [id=1511, type=softmax];
-"1512 dropout_56" [id=1512, type=dropout];
-"1513 matmul_29" [id=1513, type=matmul];
-"1514 transpose_29" [id=1514, type=transpose];
-"1515 reshape_65" [id=1515, type=reshape];
-"1516 reshape_65_0_0_nncf_smooth_quant_0" [id=1516, type=call_module];
-"1517 quantize_per_tensor_default_90" [id=1517, type=quantize_per_tensor];
-"1518 dequantize_per_tensor_default_90" [id=1518, type=dequantize_per_tensor];
-"1519 linear_89_scale_0" [id=1519, type=get_attr];
-"1520 linear_89_zero_point_0" [id=1520, type=get_attr];
-"1521 compressed_weight_updated_constant60" [id=1521, type=get_attr];
-"1522 quantize_per_channel_default_75" [id=1522, type=quantize_per_channel];
-"1523 dequantize_per_channel_default_75" [id=1523, type=dequantize_per_channel];
-"1524 _param_constant241_0_0" [id=1524, type=get_attr];
-"1525 linear_89" [id=1525, type=linear];
-"1526 dropout_57" [id=1526, type=dropout];
-"1527 view_80" [id=1527, type=view];
-"1528 permute_67" [id=1528, type=permute];
-"1529 reshape_66" [id=1529, type=reshape];
-"1530 slice_218" [id=1530, type=slice];
-"1531 slice_219" [id=1531, type=slice];
-"1532 slice_220" [id=1532, type=slice];
-"1533 slice_221" [id=1533, type=slice];
-"1534 contiguous_27" [id=1534, type=contiguous];
-"1535 _param_constant242" [id=1535, type=get_attr];
-"1536 _param_constant243" [id=1536, type=get_attr];
-"1537 layer_norm_31" [id=1537, type=layer_norm];
-"1538 add_50" [id=1538, type=add];
-"1539 add_50_0_0_nncf_smooth_quant_0" [id=1539, type=call_module];
-"1540 quantize_per_tensor_default_91" [id=1540, type=quantize_per_tensor];
-"1541 dequantize_per_tensor_default_91" [id=1541, type=dequantize_per_tensor];
-"1542 linear_90_scale_0" [id=1542, type=get_attr];
-"1543 linear_90_zero_point_0" [id=1543, type=get_attr];
-"1544 compressed_weight_updated_constant61" [id=1544, type=get_attr];
-"1545 quantize_per_channel_default_76" [id=1545, type=quantize_per_channel];
-"1546 dequantize_per_channel_default_76" [id=1546, type=dequantize_per_channel];
-"1547 _param_constant245_0_0" [id=1547, type=get_attr];
-"1548 linear_90" [id=1548, type=linear];
-"1549 gelu_14" [id=1549, type=gelu];
-"1550 dropout_58" [id=1550, type=dropout];
-"1551 dropout_58_0_0_nncf_smooth_quant_0" [id=1551, type=call_module];
-"1552 quantize_per_tensor_default_92" [id=1552, type=quantize_per_tensor];
-"1553 dequantize_per_tensor_default_92" [id=1553, type=dequantize_per_tensor];
-"1554 linear_91_scale_0" [id=1554, type=get_attr];
-"1555 linear_91_zero_point_0" [id=1555, type=get_attr];
-"1556 compressed_weight_updated_constant62" [id=1556, type=get_attr];
-"1557 quantize_per_channel_default_77" [id=1557, type=quantize_per_channel];
-"1558 dequantize_per_channel_default_77" [id=1558, type=dequantize_per_channel];
-"1559 _param_constant247_0_0" [id=1559, type=get_attr];
-"1560 linear_91" [id=1560, type=linear];
-"1561 dropout_59" [id=1561, type=dropout];
-"1562 _param_constant248" [id=1562, type=get_attr];
-"1563 _param_constant249" [id=1563, type=get_attr];
-"1564 layer_norm_32" [id=1564, type=layer_norm];
-"1565 add_51" [id=1565, type=add];
-"1566 _frozen_param123" [id=1566, type=get_attr];
-"1567 pad_17" [id=1567, type=pad];
-"1568 roll_14" [id=1568, type=roll];
-"1569 view_83" [id=1569, type=view];
-"1570 permute_69" [id=1570, type=permute];
-"1571 reshape_67" [id=1571, type=reshape];
-"1572 reshape_67_0_0_nncf_smooth_quant_0" [id=1572, type=call_module];
-"1573 quantize_per_tensor_default_93" [id=1573, type=quantize_per_tensor];
-"1574 dequantize_per_tensor_default_93" [id=1574, type=dequantize_per_tensor];
-"1575 linear_94_scale_0" [id=1575, type=get_attr];
-"1576 linear_94_zero_point_0" [id=1576, type=get_attr];
-"1577 compressed_weight_updated_constant63" [id=1577, type=get_attr];
-"1578 quantize_per_channel_default_79" [id=1578, type=quantize_per_channel];
-"1579 dequantize_per_channel_default_79" [id=1579, type=dequantize_per_channel];
-"1580 _frozen_param124_0_0" [id=1580, type=get_attr];
-"1581 linear_94" [id=1581, type=linear];
-"1582 reshape_68" [id=1582, type=reshape];
-"1583 permute_70" [id=1583, type=permute];
-"1584 select_45" [id=1584, type=select];
-"1585 select_46" [id=1585, type=select];
-"1586 select_47" [id=1586, type=select];
-"1587 linalg_vector_norm_30" [id=1587, type=linalg_vector_norm];
-"1588 clamp_min_30" [id=1588, type=clamp_min];
-"1589 expand_as_30" [id=1589, type=expand_as];
-"1590 div_30" [id=1590, type=div];
-"1591 quantize_per_tensor_default_94" [id=1591, type=quantize_per_tensor];
-"1592 dequantize_per_tensor_default_94" [id=1592, type=dequantize_per_tensor];
-"1593 linalg_vector_norm_31" [id=1593, type=linalg_vector_norm];
-"1594 clamp_min_31" [id=1594, type=clamp_min];
-"1595 expand_as_31" [id=1595, type=expand_as];
-"1596 div_31" [id=1596, type=div];
-"1597 quantize_per_tensor_default_95" [id=1597, type=quantize_per_tensor];
-"1598 dequantize_per_tensor_default_95" [id=1598, type=dequantize_per_tensor];
-"1599 transpose_30" [id=1599, type=transpose];
-"1600 matmul_30" [id=1600, type=matmul];
-"1601 _frozen_param126" [id=1601, type=get_attr];
-"1602 mul_31" [id=1602, type=mul];
-"1603 add_52" [id=1603, type=add];
-"1604 new_zeros_7" [id=1604, type=new_zeros];
-"1605 view_84" [id=1605, type=view];
-"1606 permute_71" [id=1606, type=permute];
-"1607 reshape_69" [id=1607, type=reshape];
-"1608 unsqueeze_44" [id=1608, type=unsqueeze];
-"1609 unsqueeze_45" [id=1609, type=unsqueeze];
-"1610 sub_7" [id=1610, type=sub];
-"1611 ne_7" [id=1611, type=ne];
-"1612 masked_fill_14" [id=1612, type=masked_fill];
-"1613 eq_7" [id=1613, type=eq];
-"1614 masked_fill_15" [id=1614, type=masked_fill];
-"1615 view_85" [id=1615, type=view];
-"1616 unsqueeze_46" [id=1616, type=unsqueeze];
-"1617 unsqueeze_47" [id=1617, type=unsqueeze];
-"1618 add_53" [id=1618, type=add];
-"1619 view_86" [id=1619, type=view];
-"1620 softmax_15" [id=1620, type=softmax];
-"1621 dropout_60" [id=1621, type=dropout];
-"1622 matmul_31" [id=1622, type=matmul];
-"1623 transpose_31" [id=1623, type=transpose];
-"1624 reshape_70" [id=1624, type=reshape];
-"1625 reshape_70_0_0_nncf_smooth_quant_0" [id=1625, type=call_module];
-"1626 quantize_per_tensor_default_96" [id=1626, type=quantize_per_tensor];
-"1627 dequantize_per_tensor_default_96" [id=1627, type=dequantize_per_tensor];
-"1628 linear_95_scale_0" [id=1628, type=get_attr];
-"1629 linear_95_zero_point_0" [id=1629, type=get_attr];
-"1630 compressed_weight_updated_constant64" [id=1630, type=get_attr];
-"1631 quantize_per_channel_default_80" [id=1631, type=quantize_per_channel];
-"1632 dequantize_per_channel_default_80" [id=1632, type=dequantize_per_channel];
-"1633 _param_constant257_0_0" [id=1633, type=get_attr];
-"1634 linear_95" [id=1634, type=linear];
-"1635 dropout_61" [id=1635, type=dropout];
-"1636 view_87" [id=1636, type=view];
-"1637 permute_72" [id=1637, type=permute];
-"1638 reshape_71" [id=1638, type=reshape];
-"1639 roll_15" [id=1639, type=roll];
-"1640 slice_241" [id=1640, type=slice];
-"1641 slice_242" [id=1641, type=slice];
-"1642 slice_243" [id=1642, type=slice];
-"1643 slice_244" [id=1643, type=slice];
-"1644 contiguous_29" [id=1644, type=contiguous];
-"1645 _param_constant258" [id=1645, type=get_attr];
-"1646 _param_constant259" [id=1646, type=get_attr];
-"1647 layer_norm_33" [id=1647, type=layer_norm];
-"1648 add_54" [id=1648, type=add];
-"1649 add_54_0_0_nncf_smooth_quant_0" [id=1649, type=call_module];
-"1650 quantize_per_tensor_default_97" [id=1650, type=quantize_per_tensor];
-"1651 dequantize_per_tensor_default_97" [id=1651, type=dequantize_per_tensor];
-"1652 linear_96_scale_0" [id=1652, type=get_attr];
-"1653 linear_96_zero_point_0" [id=1653, type=get_attr];
-"1654 compressed_weight_updated_constant65" [id=1654, type=get_attr];
-"1655 quantize_per_channel_default_81" [id=1655, type=quantize_per_channel];
-"1656 dequantize_per_channel_default_81" [id=1656, type=dequantize_per_channel];
-"1657 _param_constant261_0_0" [id=1657, type=get_attr];
-"1658 linear_96" [id=1658, type=linear];
-"1659 gelu_15" [id=1659, type=gelu];
-"1660 dropout_62" [id=1660, type=dropout];
-"1661 dropout_62_0_0_nncf_smooth_quant_0" [id=1661, type=call_module];
-"1662 quantize_per_tensor_default_98" [id=1662, type=quantize_per_tensor];
-"1663 dequantize_per_tensor_default_98" [id=1663, type=dequantize_per_tensor];
-"1664 linear_97_scale_0" [id=1664, type=get_attr];
-"1665 linear_97_zero_point_0" [id=1665, type=get_attr];
-"1666 compressed_weight_updated_constant66" [id=1666, type=get_attr];
-"1667 quantize_per_channel_default_82" [id=1667, type=quantize_per_channel];
-"1668 dequantize_per_channel_default_82" [id=1668, type=dequantize_per_channel];
-"1669 _param_constant263_0_0" [id=1669, type=get_attr];
-"1670 linear_97" [id=1670, type=linear];
-"1671 dropout_63" [id=1671, type=dropout];
-"1672 _param_constant264" [id=1672, type=get_attr];
-"1673 _param_constant265" [id=1673, type=get_attr];
-"1674 layer_norm_34" [id=1674, type=layer_norm];
-"1675 add_55" [id=1675, type=add];
-"1676 _frozen_param136" [id=1676, type=get_attr];
-"1677 pad_18" [id=1677, type=pad];
-"1678 view_90" [id=1678, type=view];
-"1679 permute_74" [id=1679, type=permute];
-"1680 reshape_72" [id=1680, type=reshape];
-"1681 reshape_72_0_0_nncf_smooth_quant_0" [id=1681, type=call_module];
-"1682 quantize_per_tensor_default_99" [id=1682, type=quantize_per_tensor];
-"1683 dequantize_per_tensor_default_99" [id=1683, type=dequantize_per_tensor];
-"1684 linear_100_scale_0" [id=1684, type=get_attr];
-"1685 linear_100_zero_point_0" [id=1685, type=get_attr];
-"1686 compressed_weight_updated_constant67" [id=1686, type=get_attr];
-"1687 quantize_per_channel_default_84" [id=1687, type=quantize_per_channel];
-"1688 dequantize_per_channel_default_84" [id=1688, type=dequantize_per_channel];
-"1689 _frozen_param137_0_0" [id=1689, type=get_attr];
-"1690 linear_100" [id=1690, type=linear];
-"1691 reshape_73" [id=1691, type=reshape];
-"1692 permute_75" [id=1692, type=permute];
-"1693 select_48" [id=1693, type=select];
-"1694 select_49" [id=1694, type=select];
-"1695 select_50" [id=1695, type=select];
-"1696 linalg_vector_norm_32" [id=1696, type=linalg_vector_norm];
-"1697 clamp_min_32" [id=1697, type=clamp_min];
-"1698 expand_as_32" [id=1698, type=expand_as];
-"1699 div_32" [id=1699, type=div];
-"1700 quantize_per_tensor_default_100" [id=1700, type=quantize_per_tensor];
-"1701 dequantize_per_tensor_default_100" [id=1701, type=dequantize_per_tensor];
-"1702 linalg_vector_norm_33" [id=1702, type=linalg_vector_norm];
-"1703 clamp_min_33" [id=1703, type=clamp_min];
-"1704 expand_as_33" [id=1704, type=expand_as];
-"1705 div_33" [id=1705, type=div];
-"1706 quantize_per_tensor_default_101" [id=1706, type=quantize_per_tensor];
-"1707 dequantize_per_tensor_default_101" [id=1707, type=dequantize_per_tensor];
-"1708 transpose_32" [id=1708, type=transpose];
-"1709 matmul_32" [id=1709, type=matmul];
-"1710 _frozen_param139" [id=1710, type=get_attr];
-"1711 mul_33" [id=1711, type=mul];
-"1712 add_56" [id=1712, type=add];
-"1713 softmax_16" [id=1713, type=softmax];
-"1714 dropout_64" [id=1714, type=dropout];
-"1715 matmul_33" [id=1715, type=matmul];
-"1716 transpose_33" [id=1716, type=transpose];
-"1717 reshape_74" [id=1717, type=reshape];
-"1718 reshape_74_0_0_nncf_smooth_quant_0" [id=1718, type=call_module];
-"1719 quantize_per_tensor_default_102" [id=1719, type=quantize_per_tensor];
-"1720 dequantize_per_tensor_default_102" [id=1720, type=dequantize_per_tensor];
-"1721 linear_101_scale_0" [id=1721, type=get_attr];
-"1722 linear_101_zero_point_0" [id=1722, type=get_attr];
-"1723 compressed_weight_updated_constant68" [id=1723, type=get_attr];
-"1724 quantize_per_channel_default_85" [id=1724, type=quantize_per_channel];
-"1725 dequantize_per_channel_default_85" [id=1725, type=dequantize_per_channel];
-"1726 _param_constant273_0_0" [id=1726, type=get_attr];
-"1727 linear_101" [id=1727, type=linear];
-"1728 dropout_65" [id=1728, type=dropout];
-"1729 view_91" [id=1729, type=view];
-"1730 permute_76" [id=1730, type=permute];
-"1731 reshape_75" [id=1731, type=reshape];
-"1732 slice_246" [id=1732, type=slice];
-"1733 slice_247" [id=1733, type=slice];
-"1734 slice_248" [id=1734, type=slice];
-"1735 slice_249" [id=1735, type=slice];
-"1736 contiguous_31" [id=1736, type=contiguous];
-"1737 _param_constant274" [id=1737, type=get_attr];
-"1738 _param_constant275" [id=1738, type=get_attr];
-"1739 layer_norm_35" [id=1739, type=layer_norm];
-"1740 add_57" [id=1740, type=add];
-"1741 add_57_0_0_nncf_smooth_quant_0" [id=1741, type=call_module];
-"1742 quantize_per_tensor_default_103" [id=1742, type=quantize_per_tensor];
-"1743 dequantize_per_tensor_default_103" [id=1743, type=dequantize_per_tensor];
-"1744 linear_102_scale_0" [id=1744, type=get_attr];
-"1745 linear_102_zero_point_0" [id=1745, type=get_attr];
-"1746 compressed_weight_updated_constant69" [id=1746, type=get_attr];
-"1747 quantize_per_channel_default_86" [id=1747, type=quantize_per_channel];
-"1748 dequantize_per_channel_default_86" [id=1748, type=dequantize_per_channel];
-"1749 _param_constant277_0_0" [id=1749, type=get_attr];
-"1750 linear_102" [id=1750, type=linear];
-"1751 gelu_16" [id=1751, type=gelu];
-"1752 dropout_66" [id=1752, type=dropout];
-"1753 dropout_66_0_0_nncf_smooth_quant_0" [id=1753, type=call_module];
-"1754 quantize_per_tensor_default_104" [id=1754, type=quantize_per_tensor];
-"1755 dequantize_per_tensor_default_104" [id=1755, type=dequantize_per_tensor];
-"1756 linear_103_scale_0" [id=1756, type=get_attr];
-"1757 linear_103_zero_point_0" [id=1757, type=get_attr];
-"1758 compressed_weight_updated_constant70" [id=1758, type=get_attr];
-"1759 quantize_per_channel_default_87" [id=1759, type=quantize_per_channel];
-"1760 dequantize_per_channel_default_87" [id=1760, type=dequantize_per_channel];
-"1761 _param_constant279_0_0" [id=1761, type=get_attr];
-"1762 linear_103" [id=1762, type=linear];
-"1763 dropout_67" [id=1763, type=dropout];
-"1764 _param_constant280" [id=1764, type=get_attr];
-"1765 _param_constant281" [id=1765, type=get_attr];
-"1766 layer_norm_36" [id=1766, type=layer_norm];
-"1767 add_58" [id=1767, type=add];
-"1768 _frozen_param140" [id=1768, type=get_attr];
-"1769 pad_19" [id=1769, type=pad];
-"1770 roll_16" [id=1770, type=roll];
-"1771 view_94" [id=1771, type=view];
-"1772 permute_78" [id=1772, type=permute];
-"1773 reshape_76" [id=1773, type=reshape];
-"1774 reshape_76_0_0_nncf_smooth_quant_0" [id=1774, type=call_module];
-"1775 quantize_per_tensor_default_105" [id=1775, type=quantize_per_tensor];
-"1776 dequantize_per_tensor_default_105" [id=1776, type=dequantize_per_tensor];
-"1777 linear_106_scale_0" [id=1777, type=get_attr];
-"1778 linear_106_zero_point_0" [id=1778, type=get_attr];
-"1779 compressed_weight_updated_constant71" [id=1779, type=get_attr];
-"1780 quantize_per_channel_default_89" [id=1780, type=quantize_per_channel];
-"1781 dequantize_per_channel_default_89" [id=1781, type=dequantize_per_channel];
-"1782 _frozen_param141_0_0" [id=1782, type=get_attr];
-"1783 linear_106" [id=1783, type=linear];
-"1784 reshape_77" [id=1784, type=reshape];
-"1785 permute_79" [id=1785, type=permute];
-"1786 select_51" [id=1786, type=select];
-"1787 select_52" [id=1787, type=select];
-"1788 select_53" [id=1788, type=select];
-"1789 linalg_vector_norm_34" [id=1789, type=linalg_vector_norm];
-"1790 clamp_min_34" [id=1790, type=clamp_min];
-"1791 expand_as_34" [id=1791, type=expand_as];
-"1792 div_34" [id=1792, type=div];
-"1793 quantize_per_tensor_default_106" [id=1793, type=quantize_per_tensor];
-"1794 dequantize_per_tensor_default_106" [id=1794, type=dequantize_per_tensor];
-"1795 linalg_vector_norm_35" [id=1795, type=linalg_vector_norm];
-"1796 clamp_min_35" [id=1796, type=clamp_min];
-"1797 expand_as_35" [id=1797, type=expand_as];
-"1798 div_35" [id=1798, type=div];
-"1799 quantize_per_tensor_default_107" [id=1799, type=quantize_per_tensor];
-"1800 dequantize_per_tensor_default_107" [id=1800, type=dequantize_per_tensor];
-"1801 transpose_34" [id=1801, type=transpose];
-"1802 matmul_34" [id=1802, type=matmul];
-"1803 _frozen_param143" [id=1803, type=get_attr];
-"1804 mul_35" [id=1804, type=mul];
-"1805 add_59" [id=1805, type=add];
-"1806 new_zeros_8" [id=1806, type=new_zeros];
-"1807 view_95" [id=1807, type=view];
-"1808 permute_80" [id=1808, type=permute];
-"1809 reshape_78" [id=1809, type=reshape];
-"1810 unsqueeze_50" [id=1810, type=unsqueeze];
-"1811 unsqueeze_51" [id=1811, type=unsqueeze];
-"1812 sub_8" [id=1812, type=sub];
-"1813 ne_8" [id=1813, type=ne];
-"1814 masked_fill_16" [id=1814, type=masked_fill];
-"1815 eq_8" [id=1815, type=eq];
-"1816 masked_fill_17" [id=1816, type=masked_fill];
-"1817 view_96" [id=1817, type=view];
-"1818 unsqueeze_52" [id=1818, type=unsqueeze];
-"1819 unsqueeze_53" [id=1819, type=unsqueeze];
-"1820 add_60" [id=1820, type=add];
-"1821 view_97" [id=1821, type=view];
-"1822 softmax_17" [id=1822, type=softmax];
-"1823 dropout_68" [id=1823, type=dropout];
-"1824 matmul_35" [id=1824, type=matmul];
-"1825 transpose_35" [id=1825, type=transpose];
-"1826 reshape_79" [id=1826, type=reshape];
-"1827 reshape_79_0_0_nncf_smooth_quant_0" [id=1827, type=call_module];
-"1828 quantize_per_tensor_default_108" [id=1828, type=quantize_per_tensor];
-"1829 dequantize_per_tensor_default_108" [id=1829, type=dequantize_per_tensor];
-"1830 linear_107_scale_0" [id=1830, type=get_attr];
-"1831 linear_107_zero_point_0" [id=1831, type=get_attr];
-"1832 compressed_weight_updated_constant72" [id=1832, type=get_attr];
-"1833 quantize_per_channel_default_90" [id=1833, type=quantize_per_channel];
-"1834 dequantize_per_channel_default_90" [id=1834, type=dequantize_per_channel];
-"1835 _param_constant289_0_0" [id=1835, type=get_attr];
-"1836 linear_107" [id=1836, type=linear];
-"1837 dropout_69" [id=1837, type=dropout];
-"1838 view_98" [id=1838, type=view];
-"1839 permute_81" [id=1839, type=permute];
-"1840 reshape_80" [id=1840, type=reshape];
-"1841 roll_17" [id=1841, type=roll];
-"1842 slice_269" [id=1842, type=slice];
-"1843 slice_270" [id=1843, type=slice];
-"1844 slice_271" [id=1844, type=slice];
-"1845 slice_272" [id=1845, type=slice];
-"1846 contiguous_33" [id=1846, type=contiguous];
-"1847 _param_constant290" [id=1847, type=get_attr];
-"1848 _param_constant291" [id=1848, type=get_attr];
-"1849 layer_norm_37" [id=1849, type=layer_norm];
-"1850 add_61" [id=1850, type=add];
-"1851 add_61_0_0_nncf_smooth_quant_0" [id=1851, type=call_module];
-"1852 quantize_per_tensor_default_109" [id=1852, type=quantize_per_tensor];
-"1853 dequantize_per_tensor_default_109" [id=1853, type=dequantize_per_tensor];
-"1854 linear_108_scale_0" [id=1854, type=get_attr];
-"1855 linear_108_zero_point_0" [id=1855, type=get_attr];
-"1856 compressed_weight_updated_constant73" [id=1856, type=get_attr];
-"1857 quantize_per_channel_default_91" [id=1857, type=quantize_per_channel];
-"1858 dequantize_per_channel_default_91" [id=1858, type=dequantize_per_channel];
-"1859 _param_constant293_0_0" [id=1859, type=get_attr];
-"1860 linear_108" [id=1860, type=linear];
-"1861 gelu_17" [id=1861, type=gelu];
-"1862 dropout_70" [id=1862, type=dropout];
-"1863 dropout_70_0_0_nncf_smooth_quant_0" [id=1863, type=call_module];
-"1864 quantize_per_tensor_default_110" [id=1864, type=quantize_per_tensor];
-"1865 dequantize_per_tensor_default_110" [id=1865, type=dequantize_per_tensor];
-"1866 linear_109_scale_0" [id=1866, type=get_attr];
-"1867 linear_109_zero_point_0" [id=1867, type=get_attr];
-"1868 compressed_weight_updated_constant74" [id=1868, type=get_attr];
-"1869 quantize_per_channel_default_92" [id=1869, type=quantize_per_channel];
-"1870 dequantize_per_channel_default_92" [id=1870, type=dequantize_per_channel];
-"1871 _param_constant295_0_0" [id=1871, type=get_attr];
-"1872 linear_109" [id=1872, type=linear];
-"1873 dropout_71" [id=1873, type=dropout];
-"1874 _param_constant296" [id=1874, type=get_attr];
-"1875 _param_constant297" [id=1875, type=get_attr];
-"1876 layer_norm_38" [id=1876, type=layer_norm];
-"1877 add_62" [id=1877, type=add];
-"1878 _frozen_param153" [id=1878, type=get_attr];
-"1879 pad_20" [id=1879, type=pad];
-"1880 view_101" [id=1880, type=view];
-"1881 permute_83" [id=1881, type=permute];
-"1882 reshape_81" [id=1882, type=reshape];
-"1883 reshape_81_0_0_nncf_smooth_quant_0" [id=1883, type=call_module];
-"1884 quantize_per_tensor_default_111" [id=1884, type=quantize_per_tensor];
-"1885 dequantize_per_tensor_default_111" [id=1885, type=dequantize_per_tensor];
-"1886 linear_112_scale_0" [id=1886, type=get_attr];
-"1887 linear_112_zero_point_0" [id=1887, type=get_attr];
-"1888 compressed_weight_updated_constant75" [id=1888, type=get_attr];
-"1889 quantize_per_channel_default_94" [id=1889, type=quantize_per_channel];
-"1890 dequantize_per_channel_default_94" [id=1890, type=dequantize_per_channel];
-"1891 _frozen_param154_0_0" [id=1891, type=get_attr];
-"1892 linear_112" [id=1892, type=linear];
-"1893 reshape_82" [id=1893, type=reshape];
-"1894 permute_84" [id=1894, type=permute];
-"1895 select_54" [id=1895, type=select];
-"1896 select_55" [id=1896, type=select];
-"1897 select_56" [id=1897, type=select];
-"1898 linalg_vector_norm_36" [id=1898, type=linalg_vector_norm];
-"1899 clamp_min_36" [id=1899, type=clamp_min];
-"1900 expand_as_36" [id=1900, type=expand_as];
-"1901 div_36" [id=1901, type=div];
-"1902 quantize_per_tensor_default_112" [id=1902, type=quantize_per_tensor];
-"1903 dequantize_per_tensor_default_112" [id=1903, type=dequantize_per_tensor];
-"1904 linalg_vector_norm_37" [id=1904, type=linalg_vector_norm];
-"1905 clamp_min_37" [id=1905, type=clamp_min];
-"1906 expand_as_37" [id=1906, type=expand_as];
-"1907 div_37" [id=1907, type=div];
-"1908 quantize_per_tensor_default_113" [id=1908, type=quantize_per_tensor];
-"1909 dequantize_per_tensor_default_113" [id=1909, type=dequantize_per_tensor];
-"1910 transpose_36" [id=1910, type=transpose];
-"1911 matmul_36" [id=1911, type=matmul];
-"1912 _frozen_param156" [id=1912, type=get_attr];
-"1913 mul_37" [id=1913, type=mul];
-"1914 add_63" [id=1914, type=add];
-"1915 softmax_18" [id=1915, type=softmax];
-"1916 dropout_72" [id=1916, type=dropout];
-"1917 matmul_37" [id=1917, type=matmul];
-"1918 transpose_37" [id=1918, type=transpose];
-"1919 reshape_83" [id=1919, type=reshape];
-"1920 reshape_83_0_0_nncf_smooth_quant_0" [id=1920, type=call_module];
-"1921 quantize_per_tensor_default_114" [id=1921, type=quantize_per_tensor];
-"1922 dequantize_per_tensor_default_114" [id=1922, type=dequantize_per_tensor];
-"1923 linear_113_scale_0" [id=1923, type=get_attr];
-"1924 linear_113_zero_point_0" [id=1924, type=get_attr];
-"1925 compressed_weight_updated_constant76" [id=1925, type=get_attr];
-"1926 quantize_per_channel_default_95" [id=1926, type=quantize_per_channel];
-"1927 dequantize_per_channel_default_95" [id=1927, type=dequantize_per_channel];
-"1928 _param_constant305_0_0" [id=1928, type=get_attr];
-"1929 linear_113" [id=1929, type=linear];
-"1930 dropout_73" [id=1930, type=dropout];
-"1931 view_102" [id=1931, type=view];
-"1932 permute_85" [id=1932, type=permute];
-"1933 reshape_84" [id=1933, type=reshape];
-"1934 slice_274" [id=1934, type=slice];
-"1935 slice_275" [id=1935, type=slice];
-"1936 slice_276" [id=1936, type=slice];
-"1937 slice_277" [id=1937, type=slice];
-"1938 contiguous_35" [id=1938, type=contiguous];
-"1939 _param_constant306" [id=1939, type=get_attr];
-"1940 _param_constant307" [id=1940, type=get_attr];
-"1941 layer_norm_39" [id=1941, type=layer_norm];
-"1942 add_64" [id=1942, type=add];
-"1943 add_64_0_0_nncf_smooth_quant_0" [id=1943, type=call_module];
-"1944 quantize_per_tensor_default_115" [id=1944, type=quantize_per_tensor];
-"1945 dequantize_per_tensor_default_115" [id=1945, type=dequantize_per_tensor];
-"1946 linear_114_scale_0" [id=1946, type=get_attr];
-"1947 linear_114_zero_point_0" [id=1947, type=get_attr];
-"1948 compressed_weight_updated_constant77" [id=1948, type=get_attr];
-"1949 quantize_per_channel_default_96" [id=1949, type=quantize_per_channel];
-"1950 dequantize_per_channel_default_96" [id=1950, type=dequantize_per_channel];
-"1951 _param_constant309_0_0" [id=1951, type=get_attr];
-"1952 linear_114" [id=1952, type=linear];
-"1953 gelu_18" [id=1953, type=gelu];
-"1954 dropout_74" [id=1954, type=dropout];
-"1955 dropout_74_0_0_nncf_smooth_quant_0" [id=1955, type=call_module];
-"1956 quantize_per_tensor_default_116" [id=1956, type=quantize_per_tensor];
-"1957 dequantize_per_tensor_default_116" [id=1957, type=dequantize_per_tensor];
-"1958 linear_115_scale_0" [id=1958, type=get_attr];
-"1959 linear_115_zero_point_0" [id=1959, type=get_attr];
-"1960 compressed_weight_updated_constant78" [id=1960, type=get_attr];
-"1961 quantize_per_channel_default_97" [id=1961, type=quantize_per_channel];
-"1962 dequantize_per_channel_default_97" [id=1962, type=dequantize_per_channel];
-"1963 _param_constant311_0_0" [id=1963, type=get_attr];
-"1964 linear_115" [id=1964, type=linear];
-"1965 dropout_75" [id=1965, type=dropout];
-"1966 _param_constant312" [id=1966, type=get_attr];
-"1967 _param_constant313" [id=1967, type=get_attr];
-"1968 layer_norm_40" [id=1968, type=layer_norm];
-"1969 add_65" [id=1969, type=add];
-"1970 _frozen_param157" [id=1970, type=get_attr];
-"1971 pad_21" [id=1971, type=pad];
-"1972 roll_18" [id=1972, type=roll];
-"1973 view_105" [id=1973, type=view];
-"1974 permute_87" [id=1974, type=permute];
-"1975 reshape_85" [id=1975, type=reshape];
-"1976 reshape_85_0_0_nncf_smooth_quant_0" [id=1976, type=call_module];
-"1977 quantize_per_tensor_default_117" [id=1977, type=quantize_per_tensor];
-"1978 dequantize_per_tensor_default_117" [id=1978, type=dequantize_per_tensor];
-"1979 linear_118_scale_0" [id=1979, type=get_attr];
-"1980 linear_118_zero_point_0" [id=1980, type=get_attr];
-"1981 compressed_weight_updated_constant79" [id=1981, type=get_attr];
-"1982 quantize_per_channel_default_99" [id=1982, type=quantize_per_channel];
-"1983 dequantize_per_channel_default_99" [id=1983, type=dequantize_per_channel];
-"1984 _frozen_param158_0_0" [id=1984, type=get_attr];
-"1985 linear_118" [id=1985, type=linear];
-"1986 reshape_86" [id=1986, type=reshape];
-"1987 permute_88" [id=1987, type=permute];
-"1988 select_57" [id=1988, type=select];
-"1989 select_58" [id=1989, type=select];
-"1990 select_59" [id=1990, type=select];
-"1991 linalg_vector_norm_38" [id=1991, type=linalg_vector_norm];
-"1992 clamp_min_38" [id=1992, type=clamp_min];
-"1993 expand_as_38" [id=1993, type=expand_as];
-"1994 div_38" [id=1994, type=div];
-"1995 quantize_per_tensor_default_118" [id=1995, type=quantize_per_tensor];
-"1996 dequantize_per_tensor_default_118" [id=1996, type=dequantize_per_tensor];
-"1997 linalg_vector_norm_39" [id=1997, type=linalg_vector_norm];
-"1998 clamp_min_39" [id=1998, type=clamp_min];
-"1999 expand_as_39" [id=1999, type=expand_as];
-"2000 div_39" [id=2000, type=div];
-"2001 quantize_per_tensor_default_119" [id=2001, type=quantize_per_tensor];
-"2002 dequantize_per_tensor_default_119" [id=2002, type=dequantize_per_tensor];
-"2003 transpose_38" [id=2003, type=transpose];
-"2004 matmul_38" [id=2004, type=matmul];
-"2005 _frozen_param160" [id=2005, type=get_attr];
-"2006 mul_39" [id=2006, type=mul];
-"2007 add_66" [id=2007, type=add];
-"2008 new_zeros_9" [id=2008, type=new_zeros];
-"2009 view_106" [id=2009, type=view];
-"2010 permute_89" [id=2010, type=permute];
-"2011 reshape_87" [id=2011, type=reshape];
-"2012 unsqueeze_56" [id=2012, type=unsqueeze];
-"2013 unsqueeze_57" [id=2013, type=unsqueeze];
-"2014 sub_9" [id=2014, type=sub];
-"2015 ne_9" [id=2015, type=ne];
-"2016 masked_fill_18" [id=2016, type=masked_fill];
-"2017 eq_9" [id=2017, type=eq];
-"2018 masked_fill_19" [id=2018, type=masked_fill];
-"2019 view_107" [id=2019, type=view];
-"2020 unsqueeze_58" [id=2020, type=unsqueeze];
-"2021 unsqueeze_59" [id=2021, type=unsqueeze];
-"2022 add_67" [id=2022, type=add];
-"2023 view_108" [id=2023, type=view];
-"2024 softmax_19" [id=2024, type=softmax];
-"2025 dropout_76" [id=2025, type=dropout];
-"2026 matmul_39" [id=2026, type=matmul];
-"2027 transpose_39" [id=2027, type=transpose];
-"2028 reshape_88" [id=2028, type=reshape];
-"2029 reshape_88_0_0_nncf_smooth_quant_0" [id=2029, type=call_module];
-"2030 quantize_per_tensor_default_120" [id=2030, type=quantize_per_tensor];
-"2031 dequantize_per_tensor_default_120" [id=2031, type=dequantize_per_tensor];
-"2032 linear_119_scale_0" [id=2032, type=get_attr];
-"2033 linear_119_zero_point_0" [id=2033, type=get_attr];
-"2034 compressed_weight_updated_constant80" [id=2034, type=get_attr];
-"2035 quantize_per_channel_default_100" [id=2035, type=quantize_per_channel];
-"2036 dequantize_per_channel_default_100" [id=2036, type=dequantize_per_channel];
-"2037 _param_constant321_0_0" [id=2037, type=get_attr];
-"2038 linear_119" [id=2038, type=linear];
-"2039 dropout_77" [id=2039, type=dropout];
-"2040 view_109" [id=2040, type=view];
-"2041 permute_90" [id=2041, type=permute];
-"2042 reshape_89" [id=2042, type=reshape];
-"2043 roll_19" [id=2043, type=roll];
-"2044 slice_297" [id=2044, type=slice];
-"2045 slice_298" [id=2045, type=slice];
-"2046 slice_299" [id=2046, type=slice];
-"2047 slice_300" [id=2047, type=slice];
-"2048 contiguous_37" [id=2048, type=contiguous];
-"2049 _param_constant322" [id=2049, type=get_attr];
-"2050 _param_constant323" [id=2050, type=get_attr];
-"2051 layer_norm_41" [id=2051, type=layer_norm];
-"2052 add_68" [id=2052, type=add];
-"2053 add_68_0_0_nncf_smooth_quant_0" [id=2053, type=call_module];
-"2054 quantize_per_tensor_default_121" [id=2054, type=quantize_per_tensor];
-"2055 dequantize_per_tensor_default_121" [id=2055, type=dequantize_per_tensor];
-"2056 linear_120_scale_0" [id=2056, type=get_attr];
-"2057 linear_120_zero_point_0" [id=2057, type=get_attr];
-"2058 compressed_weight_updated_constant81" [id=2058, type=get_attr];
-"2059 quantize_per_channel_default_101" [id=2059, type=quantize_per_channel];
-"2060 dequantize_per_channel_default_101" [id=2060, type=dequantize_per_channel];
-"2061 _param_constant325_0_0" [id=2061, type=get_attr];
-"2062 linear_120" [id=2062, type=linear];
-"2063 gelu_19" [id=2063, type=gelu];
-"2064 dropout_78" [id=2064, type=dropout];
-"2065 dropout_78_0_0_nncf_smooth_quant_0" [id=2065, type=call_module];
-"2066 quantize_per_tensor_default_122" [id=2066, type=quantize_per_tensor];
-"2067 dequantize_per_tensor_default_122" [id=2067, type=dequantize_per_tensor];
-"2068 linear_121_scale_0" [id=2068, type=get_attr];
-"2069 linear_121_zero_point_0" [id=2069, type=get_attr];
-"2070 compressed_weight_updated_constant82" [id=2070, type=get_attr];
-"2071 quantize_per_channel_default_102" [id=2071, type=quantize_per_channel];
-"2072 dequantize_per_channel_default_102" [id=2072, type=dequantize_per_channel];
-"2073 _param_constant327_0_0" [id=2073, type=get_attr];
-"2074 linear_121" [id=2074, type=linear];
-"2075 dropout_79" [id=2075, type=dropout];
-"2076 _param_constant328" [id=2076, type=get_attr];
-"2077 _param_constant329" [id=2077, type=get_attr];
-"2078 layer_norm_42" [id=2078, type=layer_norm];
-"2079 add_69" [id=2079, type=add];
-"2080 _frozen_param170" [id=2080, type=get_attr];
-"2081 pad_22" [id=2081, type=pad];
-"2082 view_112" [id=2082, type=view];
-"2083 permute_92" [id=2083, type=permute];
-"2084 reshape_90" [id=2084, type=reshape];
-"2085 reshape_90_0_0_nncf_smooth_quant_0" [id=2085, type=call_module];
-"2086 quantize_per_tensor_default_123" [id=2086, type=quantize_per_tensor];
-"2087 dequantize_per_tensor_default_123" [id=2087, type=dequantize_per_tensor];
-"2088 linear_124_scale_0" [id=2088, type=get_attr];
-"2089 linear_124_zero_point_0" [id=2089, type=get_attr];
-"2090 compressed_weight_updated_constant83" [id=2090, type=get_attr];
-"2091 quantize_per_channel_default_104" [id=2091, type=quantize_per_channel];
-"2092 dequantize_per_channel_default_104" [id=2092, type=dequantize_per_channel];
-"2093 _frozen_param171_0_0" [id=2093, type=get_attr];
-"2094 linear_124" [id=2094, type=linear];
-"2095 reshape_91" [id=2095, type=reshape];
-"2096 permute_93" [id=2096, type=permute];
-"2097 select_60" [id=2097, type=select];
-"2098 select_61" [id=2098, type=select];
-"2099 select_62" [id=2099, type=select];
-"2100 linalg_vector_norm_40" [id=2100, type=linalg_vector_norm];
-"2101 clamp_min_40" [id=2101, type=clamp_min];
-"2102 expand_as_40" [id=2102, type=expand_as];
-"2103 div_40" [id=2103, type=div];
-"2104 quantize_per_tensor_default_124" [id=2104, type=quantize_per_tensor];
-"2105 dequantize_per_tensor_default_124" [id=2105, type=dequantize_per_tensor];
-"2106 linalg_vector_norm_41" [id=2106, type=linalg_vector_norm];
-"2107 clamp_min_41" [id=2107, type=clamp_min];
-"2108 expand_as_41" [id=2108, type=expand_as];
-"2109 div_41" [id=2109, type=div];
-"2110 quantize_per_tensor_default_125" [id=2110, type=quantize_per_tensor];
-"2111 dequantize_per_tensor_default_125" [id=2111, type=dequantize_per_tensor];
-"2112 transpose_40" [id=2112, type=transpose];
-"2113 matmul_40" [id=2113, type=matmul];
-"2114 _frozen_param173" [id=2114, type=get_attr];
-"2115 mul_41" [id=2115, type=mul];
-"2116 add_70" [id=2116, type=add];
-"2117 softmax_20" [id=2117, type=softmax];
-"2118 dropout_80" [id=2118, type=dropout];
-"2119 matmul_41" [id=2119, type=matmul];
-"2120 transpose_41" [id=2120, type=transpose];
-"2121 reshape_92" [id=2121, type=reshape];
-"2122 reshape_92_0_0_nncf_smooth_quant_0" [id=2122, type=call_module];
-"2123 quantize_per_tensor_default_126" [id=2123, type=quantize_per_tensor];
-"2124 dequantize_per_tensor_default_126" [id=2124, type=dequantize_per_tensor];
-"2125 linear_125_scale_0" [id=2125, type=get_attr];
-"2126 linear_125_zero_point_0" [id=2126, type=get_attr];
-"2127 compressed_weight_updated_constant84" [id=2127, type=get_attr];
-"2128 quantize_per_channel_default_105" [id=2128, type=quantize_per_channel];
-"2129 dequantize_per_channel_default_105" [id=2129, type=dequantize_per_channel];
-"2130 _param_constant337_0_0" [id=2130, type=get_attr];
-"2131 linear_125" [id=2131, type=linear];
-"2132 dropout_81" [id=2132, type=dropout];
-"2133 view_113" [id=2133, type=view];
-"2134 permute_94" [id=2134, type=permute];
-"2135 reshape_93" [id=2135, type=reshape];
-"2136 slice_302" [id=2136, type=slice];
-"2137 slice_303" [id=2137, type=slice];
-"2138 slice_304" [id=2138, type=slice];
-"2139 slice_305" [id=2139, type=slice];
-"2140 contiguous_39" [id=2140, type=contiguous];
-"2141 _param_constant338" [id=2141, type=get_attr];
-"2142 _param_constant339" [id=2142, type=get_attr];
-"2143 layer_norm_43" [id=2143, type=layer_norm];
-"2144 add_71" [id=2144, type=add];
-"2145 add_71_0_0_nncf_smooth_quant_0" [id=2145, type=call_module];
-"2146 quantize_per_tensor_default_127" [id=2146, type=quantize_per_tensor];
-"2147 dequantize_per_tensor_default_127" [id=2147, type=dequantize_per_tensor];
-"2148 linear_126_scale_0" [id=2148, type=get_attr];
-"2149 linear_126_zero_point_0" [id=2149, type=get_attr];
-"2150 compressed_weight_updated_constant85" [id=2150, type=get_attr];
-"2151 quantize_per_channel_default_106" [id=2151, type=quantize_per_channel];
-"2152 dequantize_per_channel_default_106" [id=2152, type=dequantize_per_channel];
-"2153 _param_constant341_0_0" [id=2153, type=get_attr];
-"2154 linear_126" [id=2154, type=linear];
-"2155 gelu_20" [id=2155, type=gelu];
-"2156 dropout_82" [id=2156, type=dropout];
-"2157 dropout_82_0_0_nncf_smooth_quant_0" [id=2157, type=call_module];
-"2158 quantize_per_tensor_default_128" [id=2158, type=quantize_per_tensor];
-"2159 dequantize_per_tensor_default_128" [id=2159, type=dequantize_per_tensor];
-"2160 linear_127_scale_0" [id=2160, type=get_attr];
-"2161 linear_127_zero_point_0" [id=2161, type=get_attr];
-"2162 compressed_weight_updated_constant86" [id=2162, type=get_attr];
-"2163 quantize_per_channel_default_107" [id=2163, type=quantize_per_channel];
-"2164 dequantize_per_channel_default_107" [id=2164, type=dequantize_per_channel];
-"2165 _param_constant343_0_0" [id=2165, type=get_attr];
-"2166 linear_127" [id=2166, type=linear];
-"2167 dropout_83" [id=2167, type=dropout];
-"2168 _param_constant344" [id=2168, type=get_attr];
-"2169 _param_constant345" [id=2169, type=get_attr];
-"2170 layer_norm_44" [id=2170, type=layer_norm];
-"2171 add_72" [id=2171, type=add];
-"2172 _frozen_param174" [id=2172, type=get_attr];
-"2173 pad_23" [id=2173, type=pad];
-"2174 roll_20" [id=2174, type=roll];
-"2175 view_116" [id=2175, type=view];
-"2176 permute_96" [id=2176, type=permute];
-"2177 reshape_94" [id=2177, type=reshape];
-"2178 reshape_94_0_0_nncf_smooth_quant_0" [id=2178, type=call_module];
-"2179 quantize_per_tensor_default_129" [id=2179, type=quantize_per_tensor];
-"2180 dequantize_per_tensor_default_129" [id=2180, type=dequantize_per_tensor];
-"2181 linear_130_scale_0" [id=2181, type=get_attr];
-"2182 linear_130_zero_point_0" [id=2182, type=get_attr];
-"2183 compressed_weight_updated_constant87" [id=2183, type=get_attr];
-"2184 quantize_per_channel_default_109" [id=2184, type=quantize_per_channel];
-"2185 dequantize_per_channel_default_109" [id=2185, type=dequantize_per_channel];
-"2186 _frozen_param175_0_0" [id=2186, type=get_attr];
-"2187 linear_130" [id=2187, type=linear];
-"2188 reshape_95" [id=2188, type=reshape];
-"2189 permute_97" [id=2189, type=permute];
-"2190 select_63" [id=2190, type=select];
-"2191 select_64" [id=2191, type=select];
-"2192 select_65" [id=2192, type=select];
-"2193 linalg_vector_norm_42" [id=2193, type=linalg_vector_norm];
-"2194 clamp_min_42" [id=2194, type=clamp_min];
-"2195 expand_as_42" [id=2195, type=expand_as];
-"2196 div_42" [id=2196, type=div];
-"2197 quantize_per_tensor_default_130" [id=2197, type=quantize_per_tensor];
-"2198 dequantize_per_tensor_default_130" [id=2198, type=dequantize_per_tensor];
-"2199 linalg_vector_norm_43" [id=2199, type=linalg_vector_norm];
-"2200 clamp_min_43" [id=2200, type=clamp_min];
-"2201 expand_as_43" [id=2201, type=expand_as];
-"2202 div_43" [id=2202, type=div];
-"2203 quantize_per_tensor_default_131" [id=2203, type=quantize_per_tensor];
-"2204 dequantize_per_tensor_default_131" [id=2204, type=dequantize_per_tensor];
-"2205 transpose_42" [id=2205, type=transpose];
-"2206 matmul_42" [id=2206, type=matmul];
-"2207 _frozen_param177" [id=2207, type=get_attr];
-"2208 mul_43" [id=2208, type=mul];
-"2209 add_73" [id=2209, type=add];
-"2210 new_zeros_10" [id=2210, type=new_zeros];
-"2211 view_117" [id=2211, type=view];
-"2212 permute_98" [id=2212, type=permute];
-"2213 reshape_96" [id=2213, type=reshape];
-"2214 unsqueeze_62" [id=2214, type=unsqueeze];
-"2215 unsqueeze_63" [id=2215, type=unsqueeze];
-"2216 sub_10" [id=2216, type=sub];
-"2217 ne_10" [id=2217, type=ne];
-"2218 masked_fill_20" [id=2218, type=masked_fill];
-"2219 eq_10" [id=2219, type=eq];
-"2220 masked_fill_21" [id=2220, type=masked_fill];
-"2221 view_118" [id=2221, type=view];
-"2222 unsqueeze_64" [id=2222, type=unsqueeze];
-"2223 unsqueeze_65" [id=2223, type=unsqueeze];
-"2224 add_74" [id=2224, type=add];
-"2225 view_119" [id=2225, type=view];
-"2226 softmax_21" [id=2226, type=softmax];
-"2227 dropout_84" [id=2227, type=dropout];
-"2228 matmul_43" [id=2228, type=matmul];
-"2229 transpose_43" [id=2229, type=transpose];
-"2230 reshape_97" [id=2230, type=reshape];
-"2231 reshape_97_0_0_nncf_smooth_quant_0" [id=2231, type=call_module];
-"2232 quantize_per_tensor_default_132" [id=2232, type=quantize_per_tensor];
-"2233 dequantize_per_tensor_default_132" [id=2233, type=dequantize_per_tensor];
-"2234 linear_131_scale_0" [id=2234, type=get_attr];
-"2235 linear_131_zero_point_0" [id=2235, type=get_attr];
-"2236 compressed_weight_updated_constant88" [id=2236, type=get_attr];
-"2237 quantize_per_channel_default_110" [id=2237, type=quantize_per_channel];
-"2238 dequantize_per_channel_default_110" [id=2238, type=dequantize_per_channel];
-"2239 _param_constant353_0_0" [id=2239, type=get_attr];
-"2240 linear_131" [id=2240, type=linear];
-"2241 dropout_85" [id=2241, type=dropout];
-"2242 view_120" [id=2242, type=view];
-"2243 permute_99" [id=2243, type=permute];
-"2244 reshape_98" [id=2244, type=reshape];
-"2245 roll_21" [id=2245, type=roll];
-"2246 slice_325" [id=2246, type=slice];
-"2247 slice_326" [id=2247, type=slice];
-"2248 slice_327" [id=2248, type=slice];
-"2249 slice_328" [id=2249, type=slice];
-"2250 contiguous_41" [id=2250, type=contiguous];
-"2251 _param_constant354" [id=2251, type=get_attr];
-"2252 _param_constant355" [id=2252, type=get_attr];
-"2253 layer_norm_45" [id=2253, type=layer_norm];
-"2254 add_75" [id=2254, type=add];
-"2255 add_75_0_0_nncf_smooth_quant_0" [id=2255, type=call_module];
-"2256 quantize_per_tensor_default_133" [id=2256, type=quantize_per_tensor];
-"2257 dequantize_per_tensor_default_133" [id=2257, type=dequantize_per_tensor];
-"2258 linear_132_scale_0" [id=2258, type=get_attr];
-"2259 linear_132_zero_point_0" [id=2259, type=get_attr];
-"2260 compressed_weight_updated_constant89" [id=2260, type=get_attr];
-"2261 quantize_per_channel_default_111" [id=2261, type=quantize_per_channel];
-"2262 dequantize_per_channel_default_111" [id=2262, type=dequantize_per_channel];
-"2263 _param_constant357_0_0" [id=2263, type=get_attr];
-"2264 linear_132" [id=2264, type=linear];
-"2265 gelu_21" [id=2265, type=gelu];
-"2266 dropout_86" [id=2266, type=dropout];
-"2267 dropout_86_0_0_nncf_smooth_quant_0" [id=2267, type=call_module];
-"2268 quantize_per_tensor_default_134" [id=2268, type=quantize_per_tensor];
-"2269 dequantize_per_tensor_default_134" [id=2269, type=dequantize_per_tensor];
-"2270 linear_133_scale_0" [id=2270, type=get_attr];
-"2271 linear_133_zero_point_0" [id=2271, type=get_attr];
-"2272 compressed_weight_updated_constant90" [id=2272, type=get_attr];
-"2273 quantize_per_channel_default_112" [id=2273, type=quantize_per_channel];
-"2274 dequantize_per_channel_default_112" [id=2274, type=dequantize_per_channel];
-"2275 _param_constant359_0_0" [id=2275, type=get_attr];
-"2276 linear_133" [id=2276, type=linear];
-"2277 dropout_87" [id=2277, type=dropout];
-"2278 _param_constant360" [id=2278, type=get_attr];
-"2279 _param_constant361" [id=2279, type=get_attr];
-"2280 layer_norm_46" [id=2280, type=layer_norm];
-"2281 add_76" [id=2281, type=add];
-"2282 pad_24" [id=2282, type=pad];
-"2283 slice_329" [id=2283, type=slice];
-"2284 slice_330" [id=2284, type=slice];
-"2285 slice_331" [id=2285, type=slice];
-"2286 slice_332" [id=2286, type=slice];
-"2287 slice_333" [id=2287, type=slice];
-"2288 slice_334" [id=2288, type=slice];
-"2289 slice_335" [id=2289, type=slice];
-"2290 slice_336" [id=2290, type=slice];
-"2291 slice_337" [id=2291, type=slice];
-"2292 slice_338" [id=2292, type=slice];
-"2293 slice_339" [id=2293, type=slice];
-"2294 slice_340" [id=2294, type=slice];
-"2295 cat_2" [id=2295, type=cat];
-"2296 cat_2_0_0_nncf_smooth_quant_0" [id=2296, type=call_module];
-"2297 quantize_per_tensor_default_135" [id=2297, type=quantize_per_tensor];
-"2298 dequantize_per_tensor_default_135" [id=2298, type=dequantize_per_tensor];
-"2299 linear_134_scale_0" [id=2299, type=get_attr];
-"2300 linear_134_zero_point_0" [id=2300, type=get_attr];
-"2301 compressed_weight_updated_constant91" [id=2301, type=get_attr];
-"2302 quantize_per_channel_default_113" [id=2302, type=quantize_per_channel];
-"2303 dequantize_per_channel_default_113" [id=2303, type=dequantize_per_channel];
-"2304 linear_134" [id=2304, type=linear];
-"2305 _param_constant363" [id=2305, type=get_attr];
-"2306 _param_constant364" [id=2306, type=get_attr];
-"2307 layer_norm_47" [id=2307, type=layer_norm];
-"2308 _frozen_param187" [id=2308, type=get_attr];
-"2309 pad_25" [id=2309, type=pad];
-"2310 view_123" [id=2310, type=view];
-"2311 permute_101" [id=2311, type=permute];
-"2312 reshape_99" [id=2312, type=reshape];
-"2313 reshape_99_0_0_nncf_smooth_quant_0" [id=2313, type=call_module];
-"2314 quantize_per_tensor_default_136" [id=2314, type=quantize_per_tensor];
-"2315 dequantize_per_tensor_default_136" [id=2315, type=dequantize_per_tensor];
-"2316 linear_137_scale_0" [id=2316, type=get_attr];
-"2317 linear_137_zero_point_0" [id=2317, type=get_attr];
-"2318 compressed_weight_updated_constant92" [id=2318, type=get_attr];
-"2319 quantize_per_channel_default_115" [id=2319, type=quantize_per_channel];
-"2320 dequantize_per_channel_default_115" [id=2320, type=dequantize_per_channel];
-"2321 _frozen_param188_0_0" [id=2321, type=get_attr];
-"2322 linear_137" [id=2322, type=linear];
-"2323 reshape_100" [id=2323, type=reshape];
-"2324 permute_102" [id=2324, type=permute];
-"2325 select_66" [id=2325, type=select];
-"2326 select_67" [id=2326, type=select];
-"2327 select_68" [id=2327, type=select];
-"2328 linalg_vector_norm_44" [id=2328, type=linalg_vector_norm];
-"2329 clamp_min_44" [id=2329, type=clamp_min];
-"2330 expand_as_44" [id=2330, type=expand_as];
-"2331 div_44" [id=2331, type=div];
-"2332 quantize_per_tensor_default_137" [id=2332, type=quantize_per_tensor];
-"2333 dequantize_per_tensor_default_137" [id=2333, type=dequantize_per_tensor];
-"2334 linalg_vector_norm_45" [id=2334, type=linalg_vector_norm];
-"2335 clamp_min_45" [id=2335, type=clamp_min];
-"2336 expand_as_45" [id=2336, type=expand_as];
-"2337 div_45" [id=2337, type=div];
-"2338 quantize_per_tensor_default_138" [id=2338, type=quantize_per_tensor];
-"2339 dequantize_per_tensor_default_138" [id=2339, type=dequantize_per_tensor];
-"2340 transpose_44" [id=2340, type=transpose];
-"2341 matmul_44" [id=2341, type=matmul];
-"2342 _frozen_param190" [id=2342, type=get_attr];
-"2343 mul_45" [id=2343, type=mul];
-"2344 add_77" [id=2344, type=add];
-"2345 softmax_22" [id=2345, type=softmax];
-"2346 dropout_88" [id=2346, type=dropout];
-"2347 matmul_45" [id=2347, type=matmul];
-"2348 transpose_45" [id=2348, type=transpose];
-"2349 reshape_101" [id=2349, type=reshape];
-"2350 reshape_101_0_0_nncf_smooth_quant_0" [id=2350, type=call_module];
-"2351 quantize_per_tensor_default_139" [id=2351, type=quantize_per_tensor];
-"2352 dequantize_per_tensor_default_139" [id=2352, type=dequantize_per_tensor];
-"2353 linear_138_scale_0" [id=2353, type=get_attr];
-"2354 linear_138_zero_point_0" [id=2354, type=get_attr];
-"2355 compressed_weight_updated_constant93" [id=2355, type=get_attr];
-"2356 quantize_per_channel_default_116" [id=2356, type=quantize_per_channel];
-"2357 dequantize_per_channel_default_116" [id=2357, type=dequantize_per_channel];
-"2358 _param_constant372_0_0" [id=2358, type=get_attr];
-"2359 linear_138" [id=2359, type=linear];
-"2360 dropout_89" [id=2360, type=dropout];
-"2361 view_124" [id=2361, type=view];
-"2362 permute_103" [id=2362, type=permute];
-"2363 reshape_102" [id=2363, type=reshape];
-"2364 slice_342" [id=2364, type=slice];
-"2365 slice_343" [id=2365, type=slice];
-"2366 slice_344" [id=2366, type=slice];
-"2367 slice_345" [id=2367, type=slice];
-"2368 contiguous_43" [id=2368, type=contiguous];
-"2369 _param_constant373" [id=2369, type=get_attr];
-"2370 _param_constant374" [id=2370, type=get_attr];
-"2371 layer_norm_48" [id=2371, type=layer_norm];
-"2372 add_78" [id=2372, type=add];
-"2373 add_78_0_0_nncf_smooth_quant_0" [id=2373, type=call_module];
-"2374 quantize_per_tensor_default_140" [id=2374, type=quantize_per_tensor];
-"2375 dequantize_per_tensor_default_140" [id=2375, type=dequantize_per_tensor];
-"2376 linear_139_scale_0" [id=2376, type=get_attr];
-"2377 linear_139_zero_point_0" [id=2377, type=get_attr];
-"2378 compressed_weight_updated_constant94" [id=2378, type=get_attr];
-"2379 quantize_per_channel_default_117" [id=2379, type=quantize_per_channel];
-"2380 dequantize_per_channel_default_117" [id=2380, type=dequantize_per_channel];
-"2381 _param_constant376_0_0" [id=2381, type=get_attr];
-"2382 linear_139" [id=2382, type=linear];
-"2383 gelu_22" [id=2383, type=gelu];
-"2384 dropout_90" [id=2384, type=dropout];
-"2385 dropout_90_0_0_nncf_smooth_quant_0" [id=2385, type=call_module];
-"2386 quantize_per_tensor_default_141" [id=2386, type=quantize_per_tensor];
-"2387 dequantize_per_tensor_default_141" [id=2387, type=dequantize_per_tensor];
-"2388 linear_140_scale_0" [id=2388, type=get_attr];
-"2389 linear_140_zero_point_0" [id=2389, type=get_attr];
-"2390 compressed_weight_updated_constant95" [id=2390, type=get_attr];
-"2391 quantize_per_channel_default_118" [id=2391, type=quantize_per_channel];
-"2392 dequantize_per_channel_default_118" [id=2392, type=dequantize_per_channel];
-"2393 _param_constant378_0_0" [id=2393, type=get_attr];
-"2394 linear_140" [id=2394, type=linear];
-"2395 dropout_91" [id=2395, type=dropout];
-"2396 _param_constant379" [id=2396, type=get_attr];
-"2397 _param_constant380" [id=2397, type=get_attr];
-"2398 layer_norm_49" [id=2398, type=layer_norm];
-"2399 add_79" [id=2399, type=add];
-"2400 _frozen_param191" [id=2400, type=get_attr];
-"2401 pad_26" [id=2401, type=pad];
-"2402 view_127" [id=2402, type=view];
-"2403 permute_105" [id=2403, type=permute];
-"2404 reshape_103" [id=2404, type=reshape];
-"2405 reshape_103_0_0_nncf_smooth_quant_0" [id=2405, type=call_module];
-"2406 quantize_per_tensor_default_142" [id=2406, type=quantize_per_tensor];
-"2407 dequantize_per_tensor_default_142" [id=2407, type=dequantize_per_tensor];
-"2408 linear_143_scale_0" [id=2408, type=get_attr];
-"2409 linear_143_zero_point_0" [id=2409, type=get_attr];
-"2410 compressed_weight_updated_constant96" [id=2410, type=get_attr];
-"2411 quantize_per_channel_default_120" [id=2411, type=quantize_per_channel];
-"2412 dequantize_per_channel_default_120" [id=2412, type=dequantize_per_channel];
-"2413 _frozen_param192_0_0" [id=2413, type=get_attr];
-"2414 linear_143" [id=2414, type=linear];
-"2415 reshape_104" [id=2415, type=reshape];
-"2416 permute_106" [id=2416, type=permute];
-"2417 select_69" [id=2417, type=select];
-"2418 select_70" [id=2418, type=select];
-"2419 select_71" [id=2419, type=select];
-"2420 linalg_vector_norm_46" [id=2420, type=linalg_vector_norm];
-"2421 clamp_min_46" [id=2421, type=clamp_min];
-"2422 expand_as_46" [id=2422, type=expand_as];
-"2423 div_46" [id=2423, type=div];
-"2424 quantize_per_tensor_default_143" [id=2424, type=quantize_per_tensor];
-"2425 dequantize_per_tensor_default_143" [id=2425, type=dequantize_per_tensor];
-"2426 linalg_vector_norm_47" [id=2426, type=linalg_vector_norm];
-"2427 clamp_min_47" [id=2427, type=clamp_min];
-"2428 expand_as_47" [id=2428, type=expand_as];
-"2429 div_47" [id=2429, type=div];
-"2430 quantize_per_tensor_default_144" [id=2430, type=quantize_per_tensor];
-"2431 dequantize_per_tensor_default_144" [id=2431, type=dequantize_per_tensor];
-"2432 transpose_46" [id=2432, type=transpose];
-"2433 matmul_46" [id=2433, type=matmul];
-"2434 _frozen_param194" [id=2434, type=get_attr];
-"2435 mul_47" [id=2435, type=mul];
-"2436 add_80" [id=2436, type=add];
-"2437 softmax_23" [id=2437, type=softmax];
-"2438 dropout_92" [id=2438, type=dropout];
-"2439 matmul_47" [id=2439, type=matmul];
-"2440 transpose_47" [id=2440, type=transpose];
-"2441 reshape_105" [id=2441, type=reshape];
-"2442 reshape_105_0_0_nncf_smooth_quant_0" [id=2442, type=call_module];
-"2443 quantize_per_tensor_default_145" [id=2443, type=quantize_per_tensor];
-"2444 dequantize_per_tensor_default_145" [id=2444, type=dequantize_per_tensor];
-"2445 linear_144_scale_0" [id=2445, type=get_attr];
-"2446 linear_144_zero_point_0" [id=2446, type=get_attr];
-"2447 compressed_weight_updated_constant97" [id=2447, type=get_attr];
-"2448 quantize_per_channel_default_121" [id=2448, type=quantize_per_channel];
-"2449 dequantize_per_channel_default_121" [id=2449, type=dequantize_per_channel];
-"2450 _param_constant388_0_0" [id=2450, type=get_attr];
-"2451 linear_144" [id=2451, type=linear];
-"2452 dropout_93" [id=2452, type=dropout];
-"2453 view_128" [id=2453, type=view];
-"2454 permute_107" [id=2454, type=permute];
-"2455 reshape_106" [id=2455, type=reshape];
-"2456 slice_347" [id=2456, type=slice];
-"2457 slice_348" [id=2457, type=slice];
-"2458 slice_349" [id=2458, type=slice];
-"2459 slice_350" [id=2459, type=slice];
-"2460 contiguous_45" [id=2460, type=contiguous];
-"2461 _param_constant389" [id=2461, type=get_attr];
-"2462 _param_constant390" [id=2462, type=get_attr];
-"2463 layer_norm_50" [id=2463, type=layer_norm];
-"2464 add_81" [id=2464, type=add];
-"2465 add_81_0_0_nncf_smooth_quant_0" [id=2465, type=call_module];
-"2466 quantize_per_tensor_default_146" [id=2466, type=quantize_per_tensor];
-"2467 dequantize_per_tensor_default_146" [id=2467, type=dequantize_per_tensor];
-"2468 linear_145_scale_0" [id=2468, type=get_attr];
-"2469 linear_145_zero_point_0" [id=2469, type=get_attr];
-"2470 compressed_weight_updated_constant98" [id=2470, type=get_attr];
-"2471 quantize_per_channel_default_122" [id=2471, type=quantize_per_channel];
-"2472 dequantize_per_channel_default_122" [id=2472, type=dequantize_per_channel];
-"2473 _param_constant392_0_0" [id=2473, type=get_attr];
-"2474 linear_145" [id=2474, type=linear];
-"2475 gelu_23" [id=2475, type=gelu];
-"2476 dropout_94" [id=2476, type=dropout];
-"2477 dropout_94_0_0_nncf_smooth_quant_0" [id=2477, type=call_module];
-"2478 quantize_per_tensor_default_147" [id=2478, type=quantize_per_tensor];
-"2479 dequantize_per_tensor_default_147" [id=2479, type=dequantize_per_tensor];
-"2480 linear_146_scale_0" [id=2480, type=get_attr];
-"2481 linear_146_zero_point_0" [id=2481, type=get_attr];
-"2482 compressed_weight_updated_constant99" [id=2482, type=get_attr];
-"2483 quantize_per_channel_default_123" [id=2483, type=quantize_per_channel];
-"2484 dequantize_per_channel_default_123" [id=2484, type=dequantize_per_channel];
-"2485 _param_constant394_0_0" [id=2485, type=get_attr];
-"2486 linear_146" [id=2486, type=linear];
-"2487 dropout_95" [id=2487, type=dropout];
-"2488 _param_constant395" [id=2488, type=get_attr];
-"2489 _param_constant396" [id=2489, type=get_attr];
-"2490 layer_norm_51" [id=2490, type=layer_norm];
-"2491 add_82" [id=2491, type=add];
-"2492 _param_constant397" [id=2492, type=get_attr];
-"2493 _param_constant398" [id=2493, type=get_attr];
+"0 features_0_2_weight" [id=0, type=get_attr];
+"1 features_0_2_bias" [id=1, type=get_attr];
+"2 features_1_0_norm1_weight" [id=2, type=get_attr];
+"3 features_1_0_norm1_bias" [id=3, type=get_attr];
+"4 features_1_0_norm2_weight" [id=4, type=get_attr];
+"5 features_1_0_norm2_bias" [id=5, type=get_attr];
+"6 features_1_1_norm1_weight" [id=6, type=get_attr];
+"7 features_1_1_norm1_bias" [id=7, type=get_attr];
+"8 features_1_1_norm2_weight" [id=8, type=get_attr];
+"9 features_1_1_norm2_bias" [id=9, type=get_attr];
+"10 features_2_norm_weight" [id=10, type=get_attr];
+"11 features_2_norm_bias" [id=11, type=get_attr];
+"12 features_3_0_norm1_weight" [id=12, type=get_attr];
+"13 features_3_0_norm1_bias" [id=13, type=get_attr];
+"14 features_3_0_norm2_weight" [id=14, type=get_attr];
+"15 features_3_0_norm2_bias" [id=15, type=get_attr];
+"16 features_3_1_norm1_weight" [id=16, type=get_attr];
+"17 features_3_1_norm1_bias" [id=17, type=get_attr];
+"18 features_3_1_norm2_weight" [id=18, type=get_attr];
+"19 features_3_1_norm2_bias" [id=19, type=get_attr];
+"20 features_4_norm_weight" [id=20, type=get_attr];
+"21 features_4_norm_bias" [id=21, type=get_attr];
+"22 features_5_0_norm1_weight" [id=22, type=get_attr];
+"23 features_5_0_norm1_bias" [id=23, type=get_attr];
+"24 features_5_0_norm2_weight" [id=24, type=get_attr];
+"25 features_5_0_norm2_bias" [id=25, type=get_attr];
+"26 features_5_1_norm1_weight" [id=26, type=get_attr];
+"27 features_5_1_norm1_bias" [id=27, type=get_attr];
+"28 features_5_1_norm2_weight" [id=28, type=get_attr];
+"29 features_5_1_norm2_bias" [id=29, type=get_attr];
+"30 features_5_2_norm1_weight" [id=30, type=get_attr];
+"31 features_5_2_norm1_bias" [id=31, type=get_attr];
+"32 features_5_2_norm2_weight" [id=32, type=get_attr];
+"33 features_5_2_norm2_bias" [id=33, type=get_attr];
+"34 features_5_3_norm1_weight" [id=34, type=get_attr];
+"35 features_5_3_norm1_bias" [id=35, type=get_attr];
+"36 features_5_3_norm2_weight" [id=36, type=get_attr];
+"37 features_5_3_norm2_bias" [id=37, type=get_attr];
+"38 features_5_4_norm1_weight" [id=38, type=get_attr];
+"39 features_5_4_norm1_bias" [id=39, type=get_attr];
+"40 features_5_4_norm2_weight" [id=40, type=get_attr];
+"41 features_5_4_norm2_bias" [id=41, type=get_attr];
+"42 features_5_5_norm1_weight" [id=42, type=get_attr];
+"43 features_5_5_norm1_bias" [id=43, type=get_attr];
+"44 features_5_5_norm2_weight" [id=44, type=get_attr];
+"45 features_5_5_norm2_bias" [id=45, type=get_attr];
+"46 features_5_6_norm1_weight" [id=46, type=get_attr];
+"47 features_5_6_norm1_bias" [id=47, type=get_attr];
+"48 features_5_6_norm2_weight" [id=48, type=get_attr];
+"49 features_5_6_norm2_bias" [id=49, type=get_attr];
+"50 features_5_7_norm1_weight" [id=50, type=get_attr];
+"51 features_5_7_norm1_bias" [id=51, type=get_attr];
+"52 features_5_7_norm2_weight" [id=52, type=get_attr];
+"53 features_5_7_norm2_bias" [id=53, type=get_attr];
+"54 features_5_8_norm1_weight" [id=54, type=get_attr];
+"55 features_5_8_norm1_bias" [id=55, type=get_attr];
+"56 features_5_8_norm2_weight" [id=56, type=get_attr];
+"57 features_5_8_norm2_bias" [id=57, type=get_attr];
+"58 features_5_9_norm1_weight" [id=58, type=get_attr];
+"59 features_5_9_norm1_bias" [id=59, type=get_attr];
+"60 features_5_9_norm2_weight" [id=60, type=get_attr];
+"61 features_5_9_norm2_bias" [id=61, type=get_attr];
+"62 features_5_10_norm1_weight" [id=62, type=get_attr];
+"63 features_5_10_norm1_bias" [id=63, type=get_attr];
+"64 features_5_10_norm2_weight" [id=64, type=get_attr];
+"65 features_5_10_norm2_bias" [id=65, type=get_attr];
+"66 features_5_11_norm1_weight" [id=66, type=get_attr];
+"67 features_5_11_norm1_bias" [id=67, type=get_attr];
+"68 features_5_11_norm2_weight" [id=68, type=get_attr];
+"69 features_5_11_norm2_bias" [id=69, type=get_attr];
+"70 features_5_12_norm1_weight" [id=70, type=get_attr];
+"71 features_5_12_norm1_bias" [id=71, type=get_attr];
+"72 features_5_12_norm2_weight" [id=72, type=get_attr];
+"73 features_5_12_norm2_bias" [id=73, type=get_attr];
+"74 features_5_13_norm1_weight" [id=74, type=get_attr];
+"75 features_5_13_norm1_bias" [id=75, type=get_attr];
+"76 features_5_13_norm2_weight" [id=76, type=get_attr];
+"77 features_5_13_norm2_bias" [id=77, type=get_attr];
+"78 features_5_14_norm1_weight" [id=78, type=get_attr];
+"79 features_5_14_norm1_bias" [id=79, type=get_attr];
+"80 features_5_14_norm2_weight" [id=80, type=get_attr];
+"81 features_5_14_norm2_bias" [id=81, type=get_attr];
+"82 features_5_15_norm1_weight" [id=82, type=get_attr];
+"83 features_5_15_norm1_bias" [id=83, type=get_attr];
+"84 features_5_15_norm2_weight" [id=84, type=get_attr];
+"85 features_5_15_norm2_bias" [id=85, type=get_attr];
+"86 features_5_16_norm1_weight" [id=86, type=get_attr];
+"87 features_5_16_norm1_bias" [id=87, type=get_attr];
+"88 features_5_16_norm2_weight" [id=88, type=get_attr];
+"89 features_5_16_norm2_bias" [id=89, type=get_attr];
+"90 features_5_17_norm1_weight" [id=90, type=get_attr];
+"91 features_5_17_norm1_bias" [id=91, type=get_attr];
+"92 features_5_17_norm2_weight" [id=92, type=get_attr];
+"93 features_5_17_norm2_bias" [id=93, type=get_attr];
+"94 features_6_norm_weight" [id=94, type=get_attr];
+"95 features_6_norm_bias" [id=95, type=get_attr];
+"96 features_7_0_norm1_weight" [id=96, type=get_attr];
+"97 features_7_0_norm1_bias" [id=97, type=get_attr];
+"98 features_7_0_norm2_weight" [id=98, type=get_attr];
+"99 features_7_0_norm2_bias" [id=99, type=get_attr];
+"100 features_7_1_norm1_weight" [id=100, type=get_attr];
+"101 features_7_1_norm1_bias" [id=101, type=get_attr];
+"102 features_7_1_norm2_weight" [id=102, type=get_attr];
+"103 features_7_1_norm2_bias" [id=103, type=get_attr];
+"104 norm_weight" [id=104, type=get_attr];
+"105 norm_bias" [id=105, type=get_attr];
+"106 x" [id=106, type=input];
+"107 quantize_per_tensor_default" [id=107, type=quantize_per_tensor];
+"108 dequantize_per_tensor_default" [id=108, type=dequantize_per_tensor];
+"109 conv2d_scale_0" [id=109, type=get_attr];
+"110 conv2d_zero_point_0" [id=110, type=get_attr];
+"111 compressed_weight_updated_constant0" [id=111, type=get_attr];
+"112 quantize_per_channel_default" [id=112, type=quantize_per_channel];
+"113 dequantize_per_channel_default" [id=113, type=dequantize_per_channel];
+"114 features_0_0_bias_0_0" [id=114, type=get_attr];
+"115 conv2d" [id=115, type=conv2d];
+"116 permute" [id=116, type=permute];
+"117 layer_norm" [id=117, type=layer_norm];
+"118 _frozen_param0" [id=118, type=get_attr];
+"119 pad" [id=119, type=pad];
+"120 view_2" [id=120, type=view];
+"121 permute_2" [id=121, type=permute];
+"122 reshape" [id=122, type=reshape];
+"123 reshape_0_0_nncf_smooth_quant_0" [id=123, type=call_module];
+"124 quantize_per_tensor_default_1" [id=124, type=quantize_per_tensor];
+"125 dequantize_per_tensor_default_1" [id=125, type=dequantize_per_tensor];
+"126 linear_2_scale_0" [id=126, type=get_attr];
+"127 linear_2_zero_point_0" [id=127, type=get_attr];
+"128 compressed_weight_updated_constant1" [id=128, type=get_attr];
+"129 quantize_per_channel_default_2" [id=129, type=quantize_per_channel];
+"130 dequantize_per_channel_default_2" [id=130, type=dequantize_per_channel];
+"131 _frozen_param1_0_0" [id=131, type=get_attr];
+"132 linear_2" [id=132, type=linear];
+"133 reshape_1" [id=133, type=reshape];
+"134 permute_3" [id=134, type=permute];
+"135 select" [id=135, type=select];
+"136 select_1" [id=136, type=select];
+"137 select_2" [id=137, type=select];
+"138 linalg_vector_norm" [id=138, type=linalg_vector_norm];
+"139 clamp_min" [id=139, type=clamp_min];
+"140 expand_as" [id=140, type=expand_as];
+"141 div" [id=141, type=div];
+"142 quantize_per_tensor_default_2" [id=142, type=quantize_per_tensor];
+"143 dequantize_per_tensor_default_2" [id=143, type=dequantize_per_tensor];
+"144 linalg_vector_norm_1" [id=144, type=linalg_vector_norm];
+"145 clamp_min_1" [id=145, type=clamp_min];
+"146 expand_as_1" [id=146, type=expand_as];
+"147 div_1" [id=147, type=div];
+"148 quantize_per_tensor_default_3" [id=148, type=quantize_per_tensor];
+"149 dequantize_per_tensor_default_3" [id=149, type=dequantize_per_tensor];
+"150 transpose" [id=150, type=transpose];
+"151 matmul" [id=151, type=matmul];
+"152 _frozen_param3" [id=152, type=get_attr];
+"153 mul_1" [id=153, type=mul];
+"154 add" [id=154, type=add];
+"155 softmax" [id=155, type=softmax];
+"156 dropout" [id=156, type=dropout];
+"157 matmul_1" [id=157, type=matmul];
+"158 transpose_1" [id=158, type=transpose];
+"159 reshape_2" [id=159, type=reshape];
+"160 reshape_2_0_0_nncf_smooth_quant_0" [id=160, type=call_module];
+"161 quantize_per_tensor_default_4" [id=161, type=quantize_per_tensor];
+"162 dequantize_per_tensor_default_4" [id=162, type=dequantize_per_tensor];
+"163 linear_3_scale_0" [id=163, type=get_attr];
+"164 linear_3_zero_point_0" [id=164, type=get_attr];
+"165 compressed_weight_updated_constant2" [id=165, type=get_attr];
+"166 quantize_per_channel_default_3" [id=166, type=quantize_per_channel];
+"167 dequantize_per_channel_default_3" [id=167, type=dequantize_per_channel];
+"168 features_1_0_attn_proj_bias_0_0" [id=168, type=get_attr];
+"169 linear_3" [id=169, type=linear];
+"170 dropout_1" [id=170, type=dropout];
+"171 view_3" [id=171, type=view];
+"172 permute_4" [id=172, type=permute];
+"173 reshape_3" [id=173, type=reshape];
+"174 slice_2" [id=174, type=slice];
+"175 slice_3" [id=175, type=slice];
+"176 layer_norm_1" [id=176, type=layer_norm];
+"177 add_1" [id=177, type=add];
+"178 add_1_0_0_nncf_smooth_quant_0" [id=178, type=call_module];
+"179 quantize_per_tensor_default_5" [id=179, type=quantize_per_tensor];
+"180 dequantize_per_tensor_default_5" [id=180, type=dequantize_per_tensor];
+"181 linear_4_scale_0" [id=181, type=get_attr];
+"182 linear_4_zero_point_0" [id=182, type=get_attr];
+"183 compressed_weight_updated_constant3" [id=183, type=get_attr];
+"184 quantize_per_channel_default_4" [id=184, type=quantize_per_channel];
+"185 dequantize_per_channel_default_4" [id=185, type=dequantize_per_channel];
+"186 features_1_0_mlp_0_bias_0_0" [id=186, type=get_attr];
+"187 linear_4" [id=187, type=linear];
+"188 gelu" [id=188, type=gelu];
+"189 dropout_2" [id=189, type=dropout];
+"190 dropout_2_0_0_nncf_smooth_quant_0" [id=190, type=call_module];
+"191 quantize_per_tensor_default_6" [id=191, type=quantize_per_tensor];
+"192 dequantize_per_tensor_default_6" [id=192, type=dequantize_per_tensor];
+"193 linear_5_scale_0" [id=193, type=get_attr];
+"194 linear_5_zero_point_0" [id=194, type=get_attr];
+"195 compressed_weight_updated_constant4" [id=195, type=get_attr];
+"196 quantize_per_channel_default_5" [id=196, type=quantize_per_channel];
+"197 dequantize_per_channel_default_5" [id=197, type=dequantize_per_channel];
+"198 features_1_0_mlp_3_bias_0_0" [id=198, type=get_attr];
+"199 linear_5" [id=199, type=linear];
+"200 dropout_3" [id=200, type=dropout];
+"201 layer_norm_2" [id=201, type=layer_norm];
+"202 add_2" [id=202, type=add];
+"203 _frozen_param4" [id=203, type=get_attr];
+"204 pad_1" [id=204, type=pad];
+"205 roll" [id=205, type=roll];
+"206 view_6" [id=206, type=view];
+"207 permute_6" [id=207, type=permute];
+"208 reshape_4" [id=208, type=reshape];
+"209 reshape_4_0_0_nncf_smooth_quant_0" [id=209, type=call_module];
+"210 quantize_per_tensor_default_7" [id=210, type=quantize_per_tensor];
+"211 dequantize_per_tensor_default_7" [id=211, type=dequantize_per_tensor];
+"212 linear_8_scale_0" [id=212, type=get_attr];
+"213 linear_8_zero_point_0" [id=213, type=get_attr];
+"214 compressed_weight_updated_constant5" [id=214, type=get_attr];
+"215 quantize_per_channel_default_7" [id=215, type=quantize_per_channel];
+"216 dequantize_per_channel_default_7" [id=216, type=dequantize_per_channel];
+"217 _frozen_param5_0_0" [id=217, type=get_attr];
+"218 linear_8" [id=218, type=linear];
+"219 reshape_5" [id=219, type=reshape];
+"220 permute_7" [id=220, type=permute];
+"221 select_3" [id=221, type=select];
+"222 select_4" [id=222, type=select];
+"223 select_5" [id=223, type=select];
+"224 linalg_vector_norm_2" [id=224, type=linalg_vector_norm];
+"225 clamp_min_2" [id=225, type=clamp_min];
+"226 expand_as_2" [id=226, type=expand_as];
+"227 div_2" [id=227, type=div];
+"228 quantize_per_tensor_default_8" [id=228, type=quantize_per_tensor];
+"229 dequantize_per_tensor_default_8" [id=229, type=dequantize_per_tensor];
+"230 linalg_vector_norm_3" [id=230, type=linalg_vector_norm];
+"231 clamp_min_3" [id=231, type=clamp_min];
+"232 expand_as_3" [id=232, type=expand_as];
+"233 div_3" [id=233, type=div];
+"234 quantize_per_tensor_default_9" [id=234, type=quantize_per_tensor];
+"235 dequantize_per_tensor_default_9" [id=235, type=dequantize_per_tensor];
+"236 transpose_2" [id=236, type=transpose];
+"237 matmul_2" [id=237, type=matmul];
+"238 _frozen_param7" [id=238, type=get_attr];
+"239 mul_3" [id=239, type=mul];
+"240 add_3" [id=240, type=add];
+"241 new_zeros" [id=241, type=new_zeros];
+"242 view_7" [id=242, type=view];
+"243 permute_8" [id=243, type=permute];
+"244 reshape_6" [id=244, type=reshape];
+"245 unsqueeze_2" [id=245, type=unsqueeze];
+"246 unsqueeze_3" [id=246, type=unsqueeze];
+"247 sub" [id=247, type=sub];
+"248 ne" [id=248, type=ne];
+"249 masked_fill" [id=249, type=masked_fill];
+"250 eq" [id=250, type=eq];
+"251 masked_fill_1" [id=251, type=masked_fill];
+"252 view_8" [id=252, type=view];
+"253 unsqueeze_4" [id=253, type=unsqueeze];
+"254 unsqueeze_5" [id=254, type=unsqueeze];
+"255 add_4" [id=255, type=add];
+"256 view_9" [id=256, type=view];
+"257 softmax_1" [id=257, type=softmax];
+"258 dropout_4" [id=258, type=dropout];
+"259 matmul_3" [id=259, type=matmul];
+"260 transpose_3" [id=260, type=transpose];
+"261 reshape_7" [id=261, type=reshape];
+"262 reshape_7_0_0_nncf_smooth_quant_0" [id=262, type=call_module];
+"263 quantize_per_tensor_default_10" [id=263, type=quantize_per_tensor];
+"264 dequantize_per_tensor_default_10" [id=264, type=dequantize_per_tensor];
+"265 linear_9_scale_0" [id=265, type=get_attr];
+"266 linear_9_zero_point_0" [id=266, type=get_attr];
+"267 compressed_weight_updated_constant6" [id=267, type=get_attr];
+"268 quantize_per_channel_default_8" [id=268, type=quantize_per_channel];
+"269 dequantize_per_channel_default_8" [id=269, type=dequantize_per_channel];
+"270 features_1_1_attn_proj_bias_0_0" [id=270, type=get_attr];
+"271 linear_9" [id=271, type=linear];
+"272 dropout_5" [id=272, type=dropout];
+"273 view_10" [id=273, type=view];
+"274 permute_9" [id=274, type=permute];
+"275 reshape_8" [id=275, type=reshape];
+"276 roll_1" [id=276, type=roll];
+"277 slice_23" [id=277, type=slice];
+"278 slice_24" [id=278, type=slice];
+"279 layer_norm_3" [id=279, type=layer_norm];
+"280 add_5" [id=280, type=add];
+"281 add_5_0_0_nncf_smooth_quant_0" [id=281, type=call_module];
+"282 quantize_per_tensor_default_11" [id=282, type=quantize_per_tensor];
+"283 dequantize_per_tensor_default_11" [id=283, type=dequantize_per_tensor];
+"284 linear_10_scale_0" [id=284, type=get_attr];
+"285 linear_10_zero_point_0" [id=285, type=get_attr];
+"286 compressed_weight_updated_constant7" [id=286, type=get_attr];
+"287 quantize_per_channel_default_9" [id=287, type=quantize_per_channel];
+"288 dequantize_per_channel_default_9" [id=288, type=dequantize_per_channel];
+"289 features_1_1_mlp_0_bias_0_0" [id=289, type=get_attr];
+"290 linear_10" [id=290, type=linear];
+"291 gelu_1" [id=291, type=gelu];
+"292 dropout_6" [id=292, type=dropout];
+"293 dropout_6_0_0_nncf_smooth_quant_0" [id=293, type=call_module];
+"294 quantize_per_tensor_default_12" [id=294, type=quantize_per_tensor];
+"295 dequantize_per_tensor_default_12" [id=295, type=dequantize_per_tensor];
+"296 linear_11_scale_0" [id=296, type=get_attr];
+"297 linear_11_zero_point_0" [id=297, type=get_attr];
+"298 compressed_weight_updated_constant8" [id=298, type=get_attr];
+"299 quantize_per_channel_default_10" [id=299, type=quantize_per_channel];
+"300 dequantize_per_channel_default_10" [id=300, type=dequantize_per_channel];
+"301 features_1_1_mlp_3_bias_0_0" [id=301, type=get_attr];
+"302 linear_11" [id=302, type=linear];
+"303 dropout_7" [id=303, type=dropout];
+"304 layer_norm_4" [id=304, type=layer_norm];
+"305 add_6" [id=305, type=add];
+"306 pad_2" [id=306, type=pad];
+"307 slice_25" [id=307, type=slice];
+"308 slice_26" [id=308, type=slice];
+"309 slice_27" [id=309, type=slice];
+"310 slice_28" [id=310, type=slice];
+"311 slice_29" [id=311, type=slice];
+"312 slice_30" [id=312, type=slice];
+"313 slice_31" [id=313, type=slice];
+"314 slice_32" [id=314, type=slice];
+"315 slice_33" [id=315, type=slice];
+"316 slice_34" [id=316, type=slice];
+"317 slice_35" [id=317, type=slice];
+"318 slice_36" [id=318, type=slice];
+"319 cat" [id=319, type=cat];
+"320 cat_0_0_nncf_smooth_quant_0" [id=320, type=call_module];
+"321 quantize_per_tensor_default_13" [id=321, type=quantize_per_tensor];
+"322 dequantize_per_tensor_default_13" [id=322, type=dequantize_per_tensor];
+"323 linear_12_scale_0" [id=323, type=get_attr];
+"324 linear_12_zero_point_0" [id=324, type=get_attr];
+"325 compressed_weight_updated_constant9" [id=325, type=get_attr];
+"326 quantize_per_channel_default_11" [id=326, type=quantize_per_channel];
+"327 dequantize_per_channel_default_11" [id=327, type=dequantize_per_channel];
+"328 linear_12" [id=328, type=linear];
+"329 layer_norm_5" [id=329, type=layer_norm];
+"330 _frozen_param17" [id=330, type=get_attr];
+"331 pad_3" [id=331, type=pad];
+"332 view_13" [id=332, type=view];
+"333 permute_11" [id=333, type=permute];
+"334 reshape_9" [id=334, type=reshape];
+"335 reshape_9_0_0_nncf_smooth_quant_0" [id=335, type=call_module];
+"336 quantize_per_tensor_default_14" [id=336, type=quantize_per_tensor];
+"337 dequantize_per_tensor_default_14" [id=337, type=dequantize_per_tensor];
+"338 linear_15_scale_0" [id=338, type=get_attr];
+"339 linear_15_zero_point_0" [id=339, type=get_attr];
+"340 compressed_weight_updated_constant10" [id=340, type=get_attr];
+"341 quantize_per_channel_default_13" [id=341, type=quantize_per_channel];
+"342 dequantize_per_channel_default_13" [id=342, type=dequantize_per_channel];
+"343 _frozen_param18_0_0" [id=343, type=get_attr];
+"344 linear_15" [id=344, type=linear];
+"345 reshape_10" [id=345, type=reshape];
+"346 permute_12" [id=346, type=permute];
+"347 select_6" [id=347, type=select];
+"348 select_7" [id=348, type=select];
+"349 select_8" [id=349, type=select];
+"350 linalg_vector_norm_4" [id=350, type=linalg_vector_norm];
+"351 clamp_min_4" [id=351, type=clamp_min];
+"352 expand_as_4" [id=352, type=expand_as];
+"353 div_4" [id=353, type=div];
+"354 quantize_per_tensor_default_15" [id=354, type=quantize_per_tensor];
+"355 dequantize_per_tensor_default_15" [id=355, type=dequantize_per_tensor];
+"356 linalg_vector_norm_5" [id=356, type=linalg_vector_norm];
+"357 clamp_min_5" [id=357, type=clamp_min];
+"358 expand_as_5" [id=358, type=expand_as];
+"359 div_5" [id=359, type=div];
+"360 quantize_per_tensor_default_16" [id=360, type=quantize_per_tensor];
+"361 dequantize_per_tensor_default_16" [id=361, type=dequantize_per_tensor];
+"362 transpose_4" [id=362, type=transpose];
+"363 matmul_4" [id=363, type=matmul];
+"364 _frozen_param20" [id=364, type=get_attr];
+"365 mul_5" [id=365, type=mul];
+"366 add_7" [id=366, type=add];
+"367 softmax_2" [id=367, type=softmax];
+"368 dropout_8" [id=368, type=dropout];
+"369 matmul_5" [id=369, type=matmul];
+"370 transpose_5" [id=370, type=transpose];
+"371 reshape_11" [id=371, type=reshape];
+"372 reshape_11_0_0_nncf_smooth_quant_0" [id=372, type=call_module];
+"373 quantize_per_tensor_default_17" [id=373, type=quantize_per_tensor];
+"374 dequantize_per_tensor_default_17" [id=374, type=dequantize_per_tensor];
+"375 linear_16_scale_0" [id=375, type=get_attr];
+"376 linear_16_zero_point_0" [id=376, type=get_attr];
+"377 compressed_weight_updated_constant11" [id=377, type=get_attr];
+"378 quantize_per_channel_default_14" [id=378, type=quantize_per_channel];
+"379 dequantize_per_channel_default_14" [id=379, type=dequantize_per_channel];
+"380 features_3_0_attn_proj_bias_0_0" [id=380, type=get_attr];
+"381 linear_16" [id=381, type=linear];
+"382 dropout_9" [id=382, type=dropout];
+"383 view_14" [id=383, type=view];
+"384 permute_13" [id=384, type=permute];
+"385 reshape_12" [id=385, type=reshape];
+"386 slice_38" [id=386, type=slice];
+"387 slice_39" [id=387, type=slice];
+"388 slice_40" [id=388, type=slice];
+"389 slice_41" [id=389, type=slice];
+"390 contiguous_3" [id=390, type=contiguous];
+"391 layer_norm_6" [id=391, type=layer_norm];
+"392 add_8" [id=392, type=add];
+"393 add_8_0_0_nncf_smooth_quant_0" [id=393, type=call_module];
+"394 quantize_per_tensor_default_18" [id=394, type=quantize_per_tensor];
+"395 dequantize_per_tensor_default_18" [id=395, type=dequantize_per_tensor];
+"396 linear_17_scale_0" [id=396, type=get_attr];
+"397 linear_17_zero_point_0" [id=397, type=get_attr];
+"398 compressed_weight_updated_constant12" [id=398, type=get_attr];
+"399 quantize_per_channel_default_15" [id=399, type=quantize_per_channel];
+"400 dequantize_per_channel_default_15" [id=400, type=dequantize_per_channel];
+"401 features_3_0_mlp_0_bias_0_0" [id=401, type=get_attr];
+"402 linear_17" [id=402, type=linear];
+"403 gelu_2" [id=403, type=gelu];
+"404 dropout_10" [id=404, type=dropout];
+"405 dropout_10_0_0_nncf_smooth_quant_0" [id=405, type=call_module];
+"406 quantize_per_tensor_default_19" [id=406, type=quantize_per_tensor];
+"407 dequantize_per_tensor_default_19" [id=407, type=dequantize_per_tensor];
+"408 linear_18_scale_0" [id=408, type=get_attr];
+"409 linear_18_zero_point_0" [id=409, type=get_attr];
+"410 compressed_weight_updated_constant13" [id=410, type=get_attr];
+"411 quantize_per_channel_default_16" [id=411, type=quantize_per_channel];
+"412 dequantize_per_channel_default_16" [id=412, type=dequantize_per_channel];
+"413 features_3_0_mlp_3_bias_0_0" [id=413, type=get_attr];
+"414 linear_18" [id=414, type=linear];
+"415 dropout_11" [id=415, type=dropout];
+"416 layer_norm_7" [id=416, type=layer_norm];
+"417 add_9" [id=417, type=add];
+"418 _frozen_param21" [id=418, type=get_attr];
+"419 pad_4" [id=419, type=pad];
+"420 roll_2" [id=420, type=roll];
+"421 view_17" [id=421, type=view];
+"422 permute_15" [id=422, type=permute];
+"423 reshape_13" [id=423, type=reshape];
+"424 reshape_13_0_0_nncf_smooth_quant_0" [id=424, type=call_module];
+"425 quantize_per_tensor_default_20" [id=425, type=quantize_per_tensor];
+"426 dequantize_per_tensor_default_20" [id=426, type=dequantize_per_tensor];
+"427 linear_21_scale_0" [id=427, type=get_attr];
+"428 linear_21_zero_point_0" [id=428, type=get_attr];
+"429 compressed_weight_updated_constant14" [id=429, type=get_attr];
+"430 quantize_per_channel_default_18" [id=430, type=quantize_per_channel];
+"431 dequantize_per_channel_default_18" [id=431, type=dequantize_per_channel];
+"432 _frozen_param22_0_0" [id=432, type=get_attr];
+"433 linear_21" [id=433, type=linear];
+"434 reshape_14" [id=434, type=reshape];
+"435 permute_16" [id=435, type=permute];
+"436 select_9" [id=436, type=select];
+"437 select_10" [id=437, type=select];
+"438 select_11" [id=438, type=select];
+"439 linalg_vector_norm_6" [id=439, type=linalg_vector_norm];
+"440 clamp_min_6" [id=440, type=clamp_min];
+"441 expand_as_6" [id=441, type=expand_as];
+"442 div_6" [id=442, type=div];
+"443 quantize_per_tensor_default_21" [id=443, type=quantize_per_tensor];
+"444 dequantize_per_tensor_default_21" [id=444, type=dequantize_per_tensor];
+"445 linalg_vector_norm_7" [id=445, type=linalg_vector_norm];
+"446 clamp_min_7" [id=446, type=clamp_min];
+"447 expand_as_7" [id=447, type=expand_as];
+"448 div_7" [id=448, type=div];
+"449 quantize_per_tensor_default_22" [id=449, type=quantize_per_tensor];
+"450 dequantize_per_tensor_default_22" [id=450, type=dequantize_per_tensor];
+"451 transpose_6" [id=451, type=transpose];
+"452 matmul_6" [id=452, type=matmul];
+"453 _frozen_param24" [id=453, type=get_attr];
+"454 mul_7" [id=454, type=mul];
+"455 add_10" [id=455, type=add];
+"456 new_zeros_1" [id=456, type=new_zeros];
+"457 view_18" [id=457, type=view];
+"458 permute_17" [id=458, type=permute];
+"459 reshape_15" [id=459, type=reshape];
+"460 unsqueeze_8" [id=460, type=unsqueeze];
+"461 unsqueeze_9" [id=461, type=unsqueeze];
+"462 sub_1" [id=462, type=sub];
+"463 ne_1" [id=463, type=ne];
+"464 masked_fill_2" [id=464, type=masked_fill];
+"465 eq_1" [id=465, type=eq];
+"466 masked_fill_3" [id=466, type=masked_fill];
+"467 view_19" [id=467, type=view];
+"468 unsqueeze_10" [id=468, type=unsqueeze];
+"469 unsqueeze_11" [id=469, type=unsqueeze];
+"470 add_11" [id=470, type=add];
+"471 view_20" [id=471, type=view];
+"472 softmax_3" [id=472, type=softmax];
+"473 dropout_12" [id=473, type=dropout];
+"474 matmul_7" [id=474, type=matmul];
+"475 transpose_7" [id=475, type=transpose];
+"476 reshape_16" [id=476, type=reshape];
+"477 reshape_16_0_0_nncf_smooth_quant_0" [id=477, type=call_module];
+"478 quantize_per_tensor_default_23" [id=478, type=quantize_per_tensor];
+"479 dequantize_per_tensor_default_23" [id=479, type=dequantize_per_tensor];
+"480 linear_22_scale_0" [id=480, type=get_attr];
+"481 linear_22_zero_point_0" [id=481, type=get_attr];
+"482 compressed_weight_updated_constant15" [id=482, type=get_attr];
+"483 quantize_per_channel_default_19" [id=483, type=quantize_per_channel];
+"484 dequantize_per_channel_default_19" [id=484, type=dequantize_per_channel];
+"485 features_3_1_attn_proj_bias_0_0" [id=485, type=get_attr];
+"486 linear_22" [id=486, type=linear];
+"487 dropout_13" [id=487, type=dropout];
+"488 view_21" [id=488, type=view];
+"489 permute_18" [id=489, type=permute];
+"490 reshape_17" [id=490, type=reshape];
+"491 roll_3" [id=491, type=roll];
+"492 slice_61" [id=492, type=slice];
+"493 slice_62" [id=493, type=slice];
+"494 slice_63" [id=494, type=slice];
+"495 slice_64" [id=495, type=slice];
+"496 contiguous_5" [id=496, type=contiguous];
+"497 layer_norm_8" [id=497, type=layer_norm];
+"498 add_12" [id=498, type=add];
+"499 add_12_0_0_nncf_smooth_quant_0" [id=499, type=call_module];
+"500 quantize_per_tensor_default_24" [id=500, type=quantize_per_tensor];
+"501 dequantize_per_tensor_default_24" [id=501, type=dequantize_per_tensor];
+"502 linear_23_scale_0" [id=502, type=get_attr];
+"503 linear_23_zero_point_0" [id=503, type=get_attr];
+"504 compressed_weight_updated_constant16" [id=504, type=get_attr];
+"505 quantize_per_channel_default_20" [id=505, type=quantize_per_channel];
+"506 dequantize_per_channel_default_20" [id=506, type=dequantize_per_channel];
+"507 features_3_1_mlp_0_bias_0_0" [id=507, type=get_attr];
+"508 linear_23" [id=508, type=linear];
+"509 gelu_3" [id=509, type=gelu];
+"510 dropout_14" [id=510, type=dropout];
+"511 dropout_14_0_0_nncf_smooth_quant_0" [id=511, type=call_module];
+"512 quantize_per_tensor_default_25" [id=512, type=quantize_per_tensor];
+"513 dequantize_per_tensor_default_25" [id=513, type=dequantize_per_tensor];
+"514 linear_24_scale_0" [id=514, type=get_attr];
+"515 linear_24_zero_point_0" [id=515, type=get_attr];
+"516 compressed_weight_updated_constant17" [id=516, type=get_attr];
+"517 quantize_per_channel_default_21" [id=517, type=quantize_per_channel];
+"518 dequantize_per_channel_default_21" [id=518, type=dequantize_per_channel];
+"519 features_3_1_mlp_3_bias_0_0" [id=519, type=get_attr];
+"520 linear_24" [id=520, type=linear];
+"521 dropout_15" [id=521, type=dropout];
+"522 layer_norm_9" [id=522, type=layer_norm];
+"523 add_13" [id=523, type=add];
+"524 pad_5" [id=524, type=pad];
+"525 slice_65" [id=525, type=slice];
+"526 slice_66" [id=526, type=slice];
+"527 slice_67" [id=527, type=slice];
+"528 slice_68" [id=528, type=slice];
+"529 slice_69" [id=529, type=slice];
+"530 slice_70" [id=530, type=slice];
+"531 slice_71" [id=531, type=slice];
+"532 slice_72" [id=532, type=slice];
+"533 slice_73" [id=533, type=slice];
+"534 slice_74" [id=534, type=slice];
+"535 slice_75" [id=535, type=slice];
+"536 slice_76" [id=536, type=slice];
+"537 cat_1" [id=537, type=cat];
+"538 cat_1_0_0_nncf_smooth_quant_0" [id=538, type=call_module];
+"539 quantize_per_tensor_default_26" [id=539, type=quantize_per_tensor];
+"540 dequantize_per_tensor_default_26" [id=540, type=dequantize_per_tensor];
+"541 linear_25_scale_0" [id=541, type=get_attr];
+"542 linear_25_zero_point_0" [id=542, type=get_attr];
+"543 compressed_weight_updated_constant18" [id=543, type=get_attr];
+"544 quantize_per_channel_default_22" [id=544, type=quantize_per_channel];
+"545 dequantize_per_channel_default_22" [id=545, type=dequantize_per_channel];
+"546 linear_25" [id=546, type=linear];
+"547 layer_norm_10" [id=547, type=layer_norm];
+"548 _frozen_param34" [id=548, type=get_attr];
+"549 pad_6" [id=549, type=pad];
+"550 view_24" [id=550, type=view];
+"551 permute_20" [id=551, type=permute];
+"552 reshape_18" [id=552, type=reshape];
+"553 reshape_18_0_0_nncf_smooth_quant_0" [id=553, type=call_module];
+"554 quantize_per_tensor_default_27" [id=554, type=quantize_per_tensor];
+"555 dequantize_per_tensor_default_27" [id=555, type=dequantize_per_tensor];
+"556 linear_28_scale_0" [id=556, type=get_attr];
+"557 linear_28_zero_point_0" [id=557, type=get_attr];
+"558 compressed_weight_updated_constant19" [id=558, type=get_attr];
+"559 quantize_per_channel_default_24" [id=559, type=quantize_per_channel];
+"560 dequantize_per_channel_default_24" [id=560, type=dequantize_per_channel];
+"561 _frozen_param35_0_0" [id=561, type=get_attr];
+"562 linear_28" [id=562, type=linear];
+"563 reshape_19" [id=563, type=reshape];
+"564 permute_21" [id=564, type=permute];
+"565 select_12" [id=565, type=select];
+"566 select_13" [id=566, type=select];
+"567 select_14" [id=567, type=select];
+"568 linalg_vector_norm_8" [id=568, type=linalg_vector_norm];
+"569 clamp_min_8" [id=569, type=clamp_min];
+"570 expand_as_8" [id=570, type=expand_as];
+"571 div_8" [id=571, type=div];
+"572 quantize_per_tensor_default_28" [id=572, type=quantize_per_tensor];
+"573 dequantize_per_tensor_default_28" [id=573, type=dequantize_per_tensor];
+"574 linalg_vector_norm_9" [id=574, type=linalg_vector_norm];
+"575 clamp_min_9" [id=575, type=clamp_min];
+"576 expand_as_9" [id=576, type=expand_as];
+"577 div_9" [id=577, type=div];
+"578 quantize_per_tensor_default_29" [id=578, type=quantize_per_tensor];
+"579 dequantize_per_tensor_default_29" [id=579, type=dequantize_per_tensor];
+"580 transpose_8" [id=580, type=transpose];
+"581 matmul_8" [id=581, type=matmul];
+"582 _frozen_param37" [id=582, type=get_attr];
+"583 mul_9" [id=583, type=mul];
+"584 add_14" [id=584, type=add];
+"585 softmax_4" [id=585, type=softmax];
+"586 dropout_16" [id=586, type=dropout];
+"587 matmul_9" [id=587, type=matmul];
+"588 transpose_9" [id=588, type=transpose];
+"589 reshape_20" [id=589, type=reshape];
+"590 reshape_20_0_0_nncf_smooth_quant_0" [id=590, type=call_module];
+"591 quantize_per_tensor_default_30" [id=591, type=quantize_per_tensor];
+"592 dequantize_per_tensor_default_30" [id=592, type=dequantize_per_tensor];
+"593 linear_29_scale_0" [id=593, type=get_attr];
+"594 linear_29_zero_point_0" [id=594, type=get_attr];
+"595 compressed_weight_updated_constant20" [id=595, type=get_attr];
+"596 quantize_per_channel_default_25" [id=596, type=quantize_per_channel];
+"597 dequantize_per_channel_default_25" [id=597, type=dequantize_per_channel];
+"598 features_5_0_attn_proj_bias_0_0" [id=598, type=get_attr];
+"599 linear_29" [id=599, type=linear];
+"600 dropout_17" [id=600, type=dropout];
+"601 view_25" [id=601, type=view];
+"602 permute_22" [id=602, type=permute];
+"603 reshape_21" [id=603, type=reshape];
+"604 slice_78" [id=604, type=slice];
+"605 slice_79" [id=605, type=slice];
+"606 slice_80" [id=606, type=slice];
+"607 slice_81" [id=607, type=slice];
+"608 contiguous_7" [id=608, type=contiguous];
+"609 layer_norm_11" [id=609, type=layer_norm];
+"610 add_15" [id=610, type=add];
+"611 add_15_0_0_nncf_smooth_quant_0" [id=611, type=call_module];
+"612 quantize_per_tensor_default_31" [id=612, type=quantize_per_tensor];
+"613 dequantize_per_tensor_default_31" [id=613, type=dequantize_per_tensor];
+"614 linear_30_scale_0" [id=614, type=get_attr];
+"615 linear_30_zero_point_0" [id=615, type=get_attr];
+"616 compressed_weight_updated_constant21" [id=616, type=get_attr];
+"617 quantize_per_channel_default_26" [id=617, type=quantize_per_channel];
+"618 dequantize_per_channel_default_26" [id=618, type=dequantize_per_channel];
+"619 features_5_0_mlp_0_bias_0_0" [id=619, type=get_attr];
+"620 linear_30" [id=620, type=linear];
+"621 gelu_4" [id=621, type=gelu];
+"622 dropout_18" [id=622, type=dropout];
+"623 dropout_18_0_0_nncf_smooth_quant_0" [id=623, type=call_module];
+"624 quantize_per_tensor_default_32" [id=624, type=quantize_per_tensor];
+"625 dequantize_per_tensor_default_32" [id=625, type=dequantize_per_tensor];
+"626 linear_31_scale_0" [id=626, type=get_attr];
+"627 linear_31_zero_point_0" [id=627, type=get_attr];
+"628 compressed_weight_updated_constant22" [id=628, type=get_attr];
+"629 quantize_per_channel_default_27" [id=629, type=quantize_per_channel];
+"630 dequantize_per_channel_default_27" [id=630, type=dequantize_per_channel];
+"631 features_5_0_mlp_3_bias_0_0" [id=631, type=get_attr];
+"632 linear_31" [id=632, type=linear];
+"633 dropout_19" [id=633, type=dropout];
+"634 layer_norm_12" [id=634, type=layer_norm];
+"635 add_16" [id=635, type=add];
+"636 _frozen_param38" [id=636, type=get_attr];
+"637 pad_7" [id=637, type=pad];
+"638 roll_4" [id=638, type=roll];
+"639 view_28" [id=639, type=view];
+"640 permute_24" [id=640, type=permute];
+"641 reshape_22" [id=641, type=reshape];
+"642 reshape_22_0_0_nncf_smooth_quant_0" [id=642, type=call_module];
+"643 quantize_per_tensor_default_33" [id=643, type=quantize_per_tensor];
+"644 dequantize_per_tensor_default_33" [id=644, type=dequantize_per_tensor];
+"645 linear_34_scale_0" [id=645, type=get_attr];
+"646 linear_34_zero_point_0" [id=646, type=get_attr];
+"647 compressed_weight_updated_constant23" [id=647, type=get_attr];
+"648 quantize_per_channel_default_29" [id=648, type=quantize_per_channel];
+"649 dequantize_per_channel_default_29" [id=649, type=dequantize_per_channel];
+"650 _frozen_param39_0_0" [id=650, type=get_attr];
+"651 linear_34" [id=651, type=linear];
+"652 reshape_23" [id=652, type=reshape];
+"653 permute_25" [id=653, type=permute];
+"654 select_15" [id=654, type=select];
+"655 select_16" [id=655, type=select];
+"656 select_17" [id=656, type=select];
+"657 linalg_vector_norm_10" [id=657, type=linalg_vector_norm];
+"658 clamp_min_10" [id=658, type=clamp_min];
+"659 expand_as_10" [id=659, type=expand_as];
+"660 div_10" [id=660, type=div];
+"661 quantize_per_tensor_default_34" [id=661, type=quantize_per_tensor];
+"662 dequantize_per_tensor_default_34" [id=662, type=dequantize_per_tensor];
+"663 linalg_vector_norm_11" [id=663, type=linalg_vector_norm];
+"664 clamp_min_11" [id=664, type=clamp_min];
+"665 expand_as_11" [id=665, type=expand_as];
+"666 div_11" [id=666, type=div];
+"667 quantize_per_tensor_default_35" [id=667, type=quantize_per_tensor];
+"668 dequantize_per_tensor_default_35" [id=668, type=dequantize_per_tensor];
+"669 transpose_10" [id=669, type=transpose];
+"670 matmul_10" [id=670, type=matmul];
+"671 _frozen_param41" [id=671, type=get_attr];
+"672 mul_11" [id=672, type=mul];
+"673 add_17" [id=673, type=add];
+"674 new_zeros_2" [id=674, type=new_zeros];
+"675 view_29" [id=675, type=view];
+"676 permute_26" [id=676, type=permute];
+"677 reshape_24" [id=677, type=reshape];
+"678 unsqueeze_14" [id=678, type=unsqueeze];
+"679 unsqueeze_15" [id=679, type=unsqueeze];
+"680 sub_2" [id=680, type=sub];
+"681 ne_2" [id=681, type=ne];
+"682 masked_fill_4" [id=682, type=masked_fill];
+"683 eq_2" [id=683, type=eq];
+"684 masked_fill_5" [id=684, type=masked_fill];
+"685 view_30" [id=685, type=view];
+"686 unsqueeze_16" [id=686, type=unsqueeze];
+"687 unsqueeze_17" [id=687, type=unsqueeze];
+"688 add_18" [id=688, type=add];
+"689 view_31" [id=689, type=view];
+"690 softmax_5" [id=690, type=softmax];
+"691 dropout_20" [id=691, type=dropout];
+"692 matmul_11" [id=692, type=matmul];
+"693 transpose_11" [id=693, type=transpose];
+"694 reshape_25" [id=694, type=reshape];
+"695 reshape_25_0_0_nncf_smooth_quant_0" [id=695, type=call_module];
+"696 quantize_per_tensor_default_36" [id=696, type=quantize_per_tensor];
+"697 dequantize_per_tensor_default_36" [id=697, type=dequantize_per_tensor];
+"698 linear_35_scale_0" [id=698, type=get_attr];
+"699 linear_35_zero_point_0" [id=699, type=get_attr];
+"700 compressed_weight_updated_constant24" [id=700, type=get_attr];
+"701 quantize_per_channel_default_30" [id=701, type=quantize_per_channel];
+"702 dequantize_per_channel_default_30" [id=702, type=dequantize_per_channel];
+"703 features_5_1_attn_proj_bias_0_0" [id=703, type=get_attr];
+"704 linear_35" [id=704, type=linear];
+"705 dropout_21" [id=705, type=dropout];
+"706 view_32" [id=706, type=view];
+"707 permute_27" [id=707, type=permute];
+"708 reshape_26" [id=708, type=reshape];
+"709 roll_5" [id=709, type=roll];
+"710 slice_101" [id=710, type=slice];
+"711 slice_102" [id=711, type=slice];
+"712 slice_103" [id=712, type=slice];
+"713 slice_104" [id=713, type=slice];
+"714 contiguous_9" [id=714, type=contiguous];
+"715 layer_norm_13" [id=715, type=layer_norm];
+"716 add_19" [id=716, type=add];
+"717 add_19_0_0_nncf_smooth_quant_0" [id=717, type=call_module];
+"718 quantize_per_tensor_default_37" [id=718, type=quantize_per_tensor];
+"719 dequantize_per_tensor_default_37" [id=719, type=dequantize_per_tensor];
+"720 linear_36_scale_0" [id=720, type=get_attr];
+"721 linear_36_zero_point_0" [id=721, type=get_attr];
+"722 compressed_weight_updated_constant25" [id=722, type=get_attr];
+"723 quantize_per_channel_default_31" [id=723, type=quantize_per_channel];
+"724 dequantize_per_channel_default_31" [id=724, type=dequantize_per_channel];
+"725 features_5_1_mlp_0_bias_0_0" [id=725, type=get_attr];
+"726 linear_36" [id=726, type=linear];
+"727 gelu_5" [id=727, type=gelu];
+"728 dropout_22" [id=728, type=dropout];
+"729 dropout_22_0_0_nncf_smooth_quant_0" [id=729, type=call_module];
+"730 quantize_per_tensor_default_38" [id=730, type=quantize_per_tensor];
+"731 dequantize_per_tensor_default_38" [id=731, type=dequantize_per_tensor];
+"732 linear_37_scale_0" [id=732, type=get_attr];
+"733 linear_37_zero_point_0" [id=733, type=get_attr];
+"734 compressed_weight_updated_constant26" [id=734, type=get_attr];
+"735 quantize_per_channel_default_32" [id=735, type=quantize_per_channel];
+"736 dequantize_per_channel_default_32" [id=736, type=dequantize_per_channel];
+"737 features_5_1_mlp_3_bias_0_0" [id=737, type=get_attr];
+"738 linear_37" [id=738, type=linear];
+"739 dropout_23" [id=739, type=dropout];
+"740 layer_norm_14" [id=740, type=layer_norm];
+"741 add_20" [id=741, type=add];
+"742 _frozen_param51" [id=742, type=get_attr];
+"743 pad_8" [id=743, type=pad];
+"744 view_35" [id=744, type=view];
+"745 permute_29" [id=745, type=permute];
+"746 reshape_27" [id=746, type=reshape];
+"747 reshape_27_0_0_nncf_smooth_quant_0" [id=747, type=call_module];
+"748 quantize_per_tensor_default_39" [id=748, type=quantize_per_tensor];
+"749 dequantize_per_tensor_default_39" [id=749, type=dequantize_per_tensor];
+"750 linear_40_scale_0" [id=750, type=get_attr];
+"751 linear_40_zero_point_0" [id=751, type=get_attr];
+"752 compressed_weight_updated_constant27" [id=752, type=get_attr];
+"753 quantize_per_channel_default_34" [id=753, type=quantize_per_channel];
+"754 dequantize_per_channel_default_34" [id=754, type=dequantize_per_channel];
+"755 _frozen_param52_0_0" [id=755, type=get_attr];
+"756 linear_40" [id=756, type=linear];
+"757 reshape_28" [id=757, type=reshape];
+"758 permute_30" [id=758, type=permute];
+"759 select_18" [id=759, type=select];
+"760 select_19" [id=760, type=select];
+"761 select_20" [id=761, type=select];
+"762 linalg_vector_norm_12" [id=762, type=linalg_vector_norm];
+"763 clamp_min_12" [id=763, type=clamp_min];
+"764 expand_as_12" [id=764, type=expand_as];
+"765 div_12" [id=765, type=div];
+"766 quantize_per_tensor_default_40" [id=766, type=quantize_per_tensor];
+"767 dequantize_per_tensor_default_40" [id=767, type=dequantize_per_tensor];
+"768 linalg_vector_norm_13" [id=768, type=linalg_vector_norm];
+"769 clamp_min_13" [id=769, type=clamp_min];
+"770 expand_as_13" [id=770, type=expand_as];
+"771 div_13" [id=771, type=div];
+"772 quantize_per_tensor_default_41" [id=772, type=quantize_per_tensor];
+"773 dequantize_per_tensor_default_41" [id=773, type=dequantize_per_tensor];
+"774 transpose_12" [id=774, type=transpose];
+"775 matmul_12" [id=775, type=matmul];
+"776 _frozen_param54" [id=776, type=get_attr];
+"777 mul_13" [id=777, type=mul];
+"778 add_21" [id=778, type=add];
+"779 softmax_6" [id=779, type=softmax];
+"780 dropout_24" [id=780, type=dropout];
+"781 matmul_13" [id=781, type=matmul];
+"782 transpose_13" [id=782, type=transpose];
+"783 reshape_29" [id=783, type=reshape];
+"784 reshape_29_0_0_nncf_smooth_quant_0" [id=784, type=call_module];
+"785 quantize_per_tensor_default_42" [id=785, type=quantize_per_tensor];
+"786 dequantize_per_tensor_default_42" [id=786, type=dequantize_per_tensor];
+"787 linear_41_scale_0" [id=787, type=get_attr];
+"788 linear_41_zero_point_0" [id=788, type=get_attr];
+"789 compressed_weight_updated_constant28" [id=789, type=get_attr];
+"790 quantize_per_channel_default_35" [id=790, type=quantize_per_channel];
+"791 dequantize_per_channel_default_35" [id=791, type=dequantize_per_channel];
+"792 features_5_2_attn_proj_bias_0_0" [id=792, type=get_attr];
+"793 linear_41" [id=793, type=linear];
+"794 dropout_25" [id=794, type=dropout];
+"795 view_36" [id=795, type=view];
+"796 permute_31" [id=796, type=permute];
+"797 reshape_30" [id=797, type=reshape];
+"798 slice_106" [id=798, type=slice];
+"799 slice_107" [id=799, type=slice];
+"800 slice_108" [id=800, type=slice];
+"801 slice_109" [id=801, type=slice];
+"802 contiguous_11" [id=802, type=contiguous];
+"803 layer_norm_15" [id=803, type=layer_norm];
+"804 add_22" [id=804, type=add];
+"805 add_22_0_0_nncf_smooth_quant_0" [id=805, type=call_module];
+"806 quantize_per_tensor_default_43" [id=806, type=quantize_per_tensor];
+"807 dequantize_per_tensor_default_43" [id=807, type=dequantize_per_tensor];
+"808 linear_42_scale_0" [id=808, type=get_attr];
+"809 linear_42_zero_point_0" [id=809, type=get_attr];
+"810 compressed_weight_updated_constant29" [id=810, type=get_attr];
+"811 quantize_per_channel_default_36" [id=811, type=quantize_per_channel];
+"812 dequantize_per_channel_default_36" [id=812, type=dequantize_per_channel];
+"813 features_5_2_mlp_0_bias_0_0" [id=813, type=get_attr];
+"814 linear_42" [id=814, type=linear];
+"815 gelu_6" [id=815, type=gelu];
+"816 dropout_26" [id=816, type=dropout];
+"817 dropout_26_0_0_nncf_smooth_quant_0" [id=817, type=call_module];
+"818 quantize_per_tensor_default_44" [id=818, type=quantize_per_tensor];
+"819 dequantize_per_tensor_default_44" [id=819, type=dequantize_per_tensor];
+"820 linear_43_scale_0" [id=820, type=get_attr];
+"821 linear_43_zero_point_0" [id=821, type=get_attr];
+"822 compressed_weight_updated_constant30" [id=822, type=get_attr];
+"823 quantize_per_channel_default_37" [id=823, type=quantize_per_channel];
+"824 dequantize_per_channel_default_37" [id=824, type=dequantize_per_channel];
+"825 features_5_2_mlp_3_bias_0_0" [id=825, type=get_attr];
+"826 linear_43" [id=826, type=linear];
+"827 dropout_27" [id=827, type=dropout];
+"828 layer_norm_16" [id=828, type=layer_norm];
+"829 add_23" [id=829, type=add];
+"830 _frozen_param55" [id=830, type=get_attr];
+"831 pad_9" [id=831, type=pad];
+"832 roll_6" [id=832, type=roll];
+"833 view_39" [id=833, type=view];
+"834 permute_33" [id=834, type=permute];
+"835 reshape_31" [id=835, type=reshape];
+"836 reshape_31_0_0_nncf_smooth_quant_0" [id=836, type=call_module];
+"837 quantize_per_tensor_default_45" [id=837, type=quantize_per_tensor];
+"838 dequantize_per_tensor_default_45" [id=838, type=dequantize_per_tensor];
+"839 linear_46_scale_0" [id=839, type=get_attr];
+"840 linear_46_zero_point_0" [id=840, type=get_attr];
+"841 compressed_weight_updated_constant31" [id=841, type=get_attr];
+"842 quantize_per_channel_default_39" [id=842, type=quantize_per_channel];
+"843 dequantize_per_channel_default_39" [id=843, type=dequantize_per_channel];
+"844 _frozen_param56_0_0" [id=844, type=get_attr];
+"845 linear_46" [id=845, type=linear];
+"846 reshape_32" [id=846, type=reshape];
+"847 permute_34" [id=847, type=permute];
+"848 select_21" [id=848, type=select];
+"849 select_22" [id=849, type=select];
+"850 select_23" [id=850, type=select];
+"851 linalg_vector_norm_14" [id=851, type=linalg_vector_norm];
+"852 clamp_min_14" [id=852, type=clamp_min];
+"853 expand_as_14" [id=853, type=expand_as];
+"854 div_14" [id=854, type=div];
+"855 quantize_per_tensor_default_46" [id=855, type=quantize_per_tensor];
+"856 dequantize_per_tensor_default_46" [id=856, type=dequantize_per_tensor];
+"857 linalg_vector_norm_15" [id=857, type=linalg_vector_norm];
+"858 clamp_min_15" [id=858, type=clamp_min];
+"859 expand_as_15" [id=859, type=expand_as];
+"860 div_15" [id=860, type=div];
+"861 quantize_per_tensor_default_47" [id=861, type=quantize_per_tensor];
+"862 dequantize_per_tensor_default_47" [id=862, type=dequantize_per_tensor];
+"863 transpose_14" [id=863, type=transpose];
+"864 matmul_14" [id=864, type=matmul];
+"865 _frozen_param58" [id=865, type=get_attr];
+"866 mul_15" [id=866, type=mul];
+"867 add_24" [id=867, type=add];
+"868 new_zeros_3" [id=868, type=new_zeros];
+"869 view_40" [id=869, type=view];
+"870 permute_35" [id=870, type=permute];
+"871 reshape_33" [id=871, type=reshape];
+"872 unsqueeze_20" [id=872, type=unsqueeze];
+"873 unsqueeze_21" [id=873, type=unsqueeze];
+"874 sub_3" [id=874, type=sub];
+"875 ne_3" [id=875, type=ne];
+"876 masked_fill_6" [id=876, type=masked_fill];
+"877 eq_3" [id=877, type=eq];
+"878 masked_fill_7" [id=878, type=masked_fill];
+"879 view_41" [id=879, type=view];
+"880 unsqueeze_22" [id=880, type=unsqueeze];
+"881 unsqueeze_23" [id=881, type=unsqueeze];
+"882 add_25" [id=882, type=add];
+"883 view_42" [id=883, type=view];
+"884 softmax_7" [id=884, type=softmax];
+"885 dropout_28" [id=885, type=dropout];
+"886 matmul_15" [id=886, type=matmul];
+"887 transpose_15" [id=887, type=transpose];
+"888 reshape_34" [id=888, type=reshape];
+"889 reshape_34_0_0_nncf_smooth_quant_0" [id=889, type=call_module];
+"890 quantize_per_tensor_default_48" [id=890, type=quantize_per_tensor];
+"891 dequantize_per_tensor_default_48" [id=891, type=dequantize_per_tensor];
+"892 linear_47_scale_0" [id=892, type=get_attr];
+"893 linear_47_zero_point_0" [id=893, type=get_attr];
+"894 compressed_weight_updated_constant32" [id=894, type=get_attr];
+"895 quantize_per_channel_default_40" [id=895, type=quantize_per_channel];
+"896 dequantize_per_channel_default_40" [id=896, type=dequantize_per_channel];
+"897 features_5_3_attn_proj_bias_0_0" [id=897, type=get_attr];
+"898 linear_47" [id=898, type=linear];
+"899 dropout_29" [id=899, type=dropout];
+"900 view_43" [id=900, type=view];
+"901 permute_36" [id=901, type=permute];
+"902 reshape_35" [id=902, type=reshape];
+"903 roll_7" [id=903, type=roll];
+"904 slice_129" [id=904, type=slice];
+"905 slice_130" [id=905, type=slice];
+"906 slice_131" [id=906, type=slice];
+"907 slice_132" [id=907, type=slice];
+"908 contiguous_13" [id=908, type=contiguous];
+"909 layer_norm_17" [id=909, type=layer_norm];
+"910 add_26" [id=910, type=add];
+"911 add_26_0_0_nncf_smooth_quant_0" [id=911, type=call_module];
+"912 quantize_per_tensor_default_49" [id=912, type=quantize_per_tensor];
+"913 dequantize_per_tensor_default_49" [id=913, type=dequantize_per_tensor];
+"914 linear_48_scale_0" [id=914, type=get_attr];
+"915 linear_48_zero_point_0" [id=915, type=get_attr];
+"916 compressed_weight_updated_constant33" [id=916, type=get_attr];
+"917 quantize_per_channel_default_41" [id=917, type=quantize_per_channel];
+"918 dequantize_per_channel_default_41" [id=918, type=dequantize_per_channel];
+"919 features_5_3_mlp_0_bias_0_0" [id=919, type=get_attr];
+"920 linear_48" [id=920, type=linear];
+"921 gelu_7" [id=921, type=gelu];
+"922 dropout_30" [id=922, type=dropout];
+"923 dropout_30_0_0_nncf_smooth_quant_0" [id=923, type=call_module];
+"924 quantize_per_tensor_default_50" [id=924, type=quantize_per_tensor];
+"925 dequantize_per_tensor_default_50" [id=925, type=dequantize_per_tensor];
+"926 linear_49_scale_0" [id=926, type=get_attr];
+"927 linear_49_zero_point_0" [id=927, type=get_attr];
+"928 compressed_weight_updated_constant34" [id=928, type=get_attr];
+"929 quantize_per_channel_default_42" [id=929, type=quantize_per_channel];
+"930 dequantize_per_channel_default_42" [id=930, type=dequantize_per_channel];
+"931 features_5_3_mlp_3_bias_0_0" [id=931, type=get_attr];
+"932 linear_49" [id=932, type=linear];
+"933 dropout_31" [id=933, type=dropout];
+"934 layer_norm_18" [id=934, type=layer_norm];
+"935 add_27" [id=935, type=add];
+"936 _frozen_param68" [id=936, type=get_attr];
+"937 pad_10" [id=937, type=pad];
+"938 view_46" [id=938, type=view];
+"939 permute_38" [id=939, type=permute];
+"940 reshape_36" [id=940, type=reshape];
+"941 reshape_36_0_0_nncf_smooth_quant_0" [id=941, type=call_module];
+"942 quantize_per_tensor_default_51" [id=942, type=quantize_per_tensor];
+"943 dequantize_per_tensor_default_51" [id=943, type=dequantize_per_tensor];
+"944 linear_52_scale_0" [id=944, type=get_attr];
+"945 linear_52_zero_point_0" [id=945, type=get_attr];
+"946 compressed_weight_updated_constant35" [id=946, type=get_attr];
+"947 quantize_per_channel_default_44" [id=947, type=quantize_per_channel];
+"948 dequantize_per_channel_default_44" [id=948, type=dequantize_per_channel];
+"949 _frozen_param69_0_0" [id=949, type=get_attr];
+"950 linear_52" [id=950, type=linear];
+"951 reshape_37" [id=951, type=reshape];
+"952 permute_39" [id=952, type=permute];
+"953 select_24" [id=953, type=select];
+"954 select_25" [id=954, type=select];
+"955 select_26" [id=955, type=select];
+"956 linalg_vector_norm_16" [id=956, type=linalg_vector_norm];
+"957 clamp_min_16" [id=957, type=clamp_min];
+"958 expand_as_16" [id=958, type=expand_as];
+"959 div_16" [id=959, type=div];
+"960 quantize_per_tensor_default_52" [id=960, type=quantize_per_tensor];
+"961 dequantize_per_tensor_default_52" [id=961, type=dequantize_per_tensor];
+"962 linalg_vector_norm_17" [id=962, type=linalg_vector_norm];
+"963 clamp_min_17" [id=963, type=clamp_min];
+"964 expand_as_17" [id=964, type=expand_as];
+"965 div_17" [id=965, type=div];
+"966 quantize_per_tensor_default_53" [id=966, type=quantize_per_tensor];
+"967 dequantize_per_tensor_default_53" [id=967, type=dequantize_per_tensor];
+"968 transpose_16" [id=968, type=transpose];
+"969 matmul_16" [id=969, type=matmul];
+"970 _frozen_param71" [id=970, type=get_attr];
+"971 mul_17" [id=971, type=mul];
+"972 add_28" [id=972, type=add];
+"973 softmax_8" [id=973, type=softmax];
+"974 dropout_32" [id=974, type=dropout];
+"975 matmul_17" [id=975, type=matmul];
+"976 transpose_17" [id=976, type=transpose];
+"977 reshape_38" [id=977, type=reshape];
+"978 reshape_38_0_0_nncf_smooth_quant_0" [id=978, type=call_module];
+"979 quantize_per_tensor_default_54" [id=979, type=quantize_per_tensor];
+"980 dequantize_per_tensor_default_54" [id=980, type=dequantize_per_tensor];
+"981 linear_53_scale_0" [id=981, type=get_attr];
+"982 linear_53_zero_point_0" [id=982, type=get_attr];
+"983 compressed_weight_updated_constant36" [id=983, type=get_attr];
+"984 quantize_per_channel_default_45" [id=984, type=quantize_per_channel];
+"985 dequantize_per_channel_default_45" [id=985, type=dequantize_per_channel];
+"986 features_5_4_attn_proj_bias_0_0" [id=986, type=get_attr];
+"987 linear_53" [id=987, type=linear];
+"988 dropout_33" [id=988, type=dropout];
+"989 view_47" [id=989, type=view];
+"990 permute_40" [id=990, type=permute];
+"991 reshape_39" [id=991, type=reshape];
+"992 slice_134" [id=992, type=slice];
+"993 slice_135" [id=993, type=slice];
+"994 slice_136" [id=994, type=slice];
+"995 slice_137" [id=995, type=slice];
+"996 contiguous_15" [id=996, type=contiguous];
+"997 layer_norm_19" [id=997, type=layer_norm];
+"998 add_29" [id=998, type=add];
+"999 add_29_0_0_nncf_smooth_quant_0" [id=999, type=call_module];
+"1000 quantize_per_tensor_default_55" [id=1000, type=quantize_per_tensor];
+"1001 dequantize_per_tensor_default_55" [id=1001, type=dequantize_per_tensor];
+"1002 linear_54_scale_0" [id=1002, type=get_attr];
+"1003 linear_54_zero_point_0" [id=1003, type=get_attr];
+"1004 compressed_weight_updated_constant37" [id=1004, type=get_attr];
+"1005 quantize_per_channel_default_46" [id=1005, type=quantize_per_channel];
+"1006 dequantize_per_channel_default_46" [id=1006, type=dequantize_per_channel];
+"1007 features_5_4_mlp_0_bias_0_0" [id=1007, type=get_attr];
+"1008 linear_54" [id=1008, type=linear];
+"1009 gelu_8" [id=1009, type=gelu];
+"1010 dropout_34" [id=1010, type=dropout];
+"1011 dropout_34_0_0_nncf_smooth_quant_0" [id=1011, type=call_module];
+"1012 quantize_per_tensor_default_56" [id=1012, type=quantize_per_tensor];
+"1013 dequantize_per_tensor_default_56" [id=1013, type=dequantize_per_tensor];
+"1014 linear_55_scale_0" [id=1014, type=get_attr];
+"1015 linear_55_zero_point_0" [id=1015, type=get_attr];
+"1016 compressed_weight_updated_constant38" [id=1016, type=get_attr];
+"1017 quantize_per_channel_default_47" [id=1017, type=quantize_per_channel];
+"1018 dequantize_per_channel_default_47" [id=1018, type=dequantize_per_channel];
+"1019 features_5_4_mlp_3_bias_0_0" [id=1019, type=get_attr];
+"1020 linear_55" [id=1020, type=linear];
+"1021 dropout_35" [id=1021, type=dropout];
+"1022 layer_norm_20" [id=1022, type=layer_norm];
+"1023 add_30" [id=1023, type=add];
+"1024 _frozen_param72" [id=1024, type=get_attr];
+"1025 pad_11" [id=1025, type=pad];
+"1026 roll_8" [id=1026, type=roll];
+"1027 view_50" [id=1027, type=view];
+"1028 permute_42" [id=1028, type=permute];
+"1029 reshape_40" [id=1029, type=reshape];
+"1030 reshape_40_0_0_nncf_smooth_quant_0" [id=1030, type=call_module];
+"1031 quantize_per_tensor_default_57" [id=1031, type=quantize_per_tensor];
+"1032 dequantize_per_tensor_default_57" [id=1032, type=dequantize_per_tensor];
+"1033 linear_58_scale_0" [id=1033, type=get_attr];
+"1034 linear_58_zero_point_0" [id=1034, type=get_attr];
+"1035 compressed_weight_updated_constant39" [id=1035, type=get_attr];
+"1036 quantize_per_channel_default_49" [id=1036, type=quantize_per_channel];
+"1037 dequantize_per_channel_default_49" [id=1037, type=dequantize_per_channel];
+"1038 _frozen_param73_0_0" [id=1038, type=get_attr];
+"1039 linear_58" [id=1039, type=linear];
+"1040 reshape_41" [id=1040, type=reshape];
+"1041 permute_43" [id=1041, type=permute];
+"1042 select_27" [id=1042, type=select];
+"1043 select_28" [id=1043, type=select];
+"1044 select_29" [id=1044, type=select];
+"1045 linalg_vector_norm_18" [id=1045, type=linalg_vector_norm];
+"1046 clamp_min_18" [id=1046, type=clamp_min];
+"1047 expand_as_18" [id=1047, type=expand_as];
+"1048 div_18" [id=1048, type=div];
+"1049 quantize_per_tensor_default_58" [id=1049, type=quantize_per_tensor];
+"1050 dequantize_per_tensor_default_58" [id=1050, type=dequantize_per_tensor];
+"1051 linalg_vector_norm_19" [id=1051, type=linalg_vector_norm];
+"1052 clamp_min_19" [id=1052, type=clamp_min];
+"1053 expand_as_19" [id=1053, type=expand_as];
+"1054 div_19" [id=1054, type=div];
+"1055 quantize_per_tensor_default_59" [id=1055, type=quantize_per_tensor];
+"1056 dequantize_per_tensor_default_59" [id=1056, type=dequantize_per_tensor];
+"1057 transpose_18" [id=1057, type=transpose];
+"1058 matmul_18" [id=1058, type=matmul];
+"1059 _frozen_param75" [id=1059, type=get_attr];
+"1060 mul_19" [id=1060, type=mul];
+"1061 add_31" [id=1061, type=add];
+"1062 new_zeros_4" [id=1062, type=new_zeros];
+"1063 view_51" [id=1063, type=view];
+"1064 permute_44" [id=1064, type=permute];
+"1065 reshape_42" [id=1065, type=reshape];
+"1066 unsqueeze_26" [id=1066, type=unsqueeze];
+"1067 unsqueeze_27" [id=1067, type=unsqueeze];
+"1068 sub_4" [id=1068, type=sub];
+"1069 ne_4" [id=1069, type=ne];
+"1070 masked_fill_8" [id=1070, type=masked_fill];
+"1071 eq_4" [id=1071, type=eq];
+"1072 masked_fill_9" [id=1072, type=masked_fill];
+"1073 view_52" [id=1073, type=view];
+"1074 unsqueeze_28" [id=1074, type=unsqueeze];
+"1075 unsqueeze_29" [id=1075, type=unsqueeze];
+"1076 add_32" [id=1076, type=add];
+"1077 view_53" [id=1077, type=view];
+"1078 softmax_9" [id=1078, type=softmax];
+"1079 dropout_36" [id=1079, type=dropout];
+"1080 matmul_19" [id=1080, type=matmul];
+"1081 transpose_19" [id=1081, type=transpose];
+"1082 reshape_43" [id=1082, type=reshape];
+"1083 reshape_43_0_0_nncf_smooth_quant_0" [id=1083, type=call_module];
+"1084 quantize_per_tensor_default_60" [id=1084, type=quantize_per_tensor];
+"1085 dequantize_per_tensor_default_60" [id=1085, type=dequantize_per_tensor];
+"1086 linear_59_scale_0" [id=1086, type=get_attr];
+"1087 linear_59_zero_point_0" [id=1087, type=get_attr];
+"1088 compressed_weight_updated_constant40" [id=1088, type=get_attr];
+"1089 quantize_per_channel_default_50" [id=1089, type=quantize_per_channel];
+"1090 dequantize_per_channel_default_50" [id=1090, type=dequantize_per_channel];
+"1091 features_5_5_attn_proj_bias_0_0" [id=1091, type=get_attr];
+"1092 linear_59" [id=1092, type=linear];
+"1093 dropout_37" [id=1093, type=dropout];
+"1094 view_54" [id=1094, type=view];
+"1095 permute_45" [id=1095, type=permute];
+"1096 reshape_44" [id=1096, type=reshape];
+"1097 roll_9" [id=1097, type=roll];
+"1098 slice_157" [id=1098, type=slice];
+"1099 slice_158" [id=1099, type=slice];
+"1100 slice_159" [id=1100, type=slice];
+"1101 slice_160" [id=1101, type=slice];
+"1102 contiguous_17" [id=1102, type=contiguous];
+"1103 layer_norm_21" [id=1103, type=layer_norm];
+"1104 add_33" [id=1104, type=add];
+"1105 add_33_0_0_nncf_smooth_quant_0" [id=1105, type=call_module];
+"1106 quantize_per_tensor_default_61" [id=1106, type=quantize_per_tensor];
+"1107 dequantize_per_tensor_default_61" [id=1107, type=dequantize_per_tensor];
+"1108 linear_60_scale_0" [id=1108, type=get_attr];
+"1109 linear_60_zero_point_0" [id=1109, type=get_attr];
+"1110 compressed_weight_updated_constant41" [id=1110, type=get_attr];
+"1111 quantize_per_channel_default_51" [id=1111, type=quantize_per_channel];
+"1112 dequantize_per_channel_default_51" [id=1112, type=dequantize_per_channel];
+"1113 features_5_5_mlp_0_bias_0_0" [id=1113, type=get_attr];
+"1114 linear_60" [id=1114, type=linear];
+"1115 gelu_9" [id=1115, type=gelu];
+"1116 dropout_38" [id=1116, type=dropout];
+"1117 dropout_38_0_0_nncf_smooth_quant_0" [id=1117, type=call_module];
+"1118 quantize_per_tensor_default_62" [id=1118, type=quantize_per_tensor];
+"1119 dequantize_per_tensor_default_62" [id=1119, type=dequantize_per_tensor];
+"1120 linear_61_scale_0" [id=1120, type=get_attr];
+"1121 linear_61_zero_point_0" [id=1121, type=get_attr];
+"1122 compressed_weight_updated_constant42" [id=1122, type=get_attr];
+"1123 quantize_per_channel_default_52" [id=1123, type=quantize_per_channel];
+"1124 dequantize_per_channel_default_52" [id=1124, type=dequantize_per_channel];
+"1125 features_5_5_mlp_3_bias_0_0" [id=1125, type=get_attr];
+"1126 linear_61" [id=1126, type=linear];
+"1127 dropout_39" [id=1127, type=dropout];
+"1128 layer_norm_22" [id=1128, type=layer_norm];
+"1129 add_34" [id=1129, type=add];
+"1130 _frozen_param85" [id=1130, type=get_attr];
+"1131 pad_12" [id=1131, type=pad];
+"1132 view_57" [id=1132, type=view];
+"1133 permute_47" [id=1133, type=permute];
+"1134 reshape_45" [id=1134, type=reshape];
+"1135 reshape_45_0_0_nncf_smooth_quant_0" [id=1135, type=call_module];
+"1136 quantize_per_tensor_default_63" [id=1136, type=quantize_per_tensor];
+"1137 dequantize_per_tensor_default_63" [id=1137, type=dequantize_per_tensor];
+"1138 linear_64_scale_0" [id=1138, type=get_attr];
+"1139 linear_64_zero_point_0" [id=1139, type=get_attr];
+"1140 compressed_weight_updated_constant43" [id=1140, type=get_attr];
+"1141 quantize_per_channel_default_54" [id=1141, type=quantize_per_channel];
+"1142 dequantize_per_channel_default_54" [id=1142, type=dequantize_per_channel];
+"1143 _frozen_param86_0_0" [id=1143, type=get_attr];
+"1144 linear_64" [id=1144, type=linear];
+"1145 reshape_46" [id=1145, type=reshape];
+"1146 permute_48" [id=1146, type=permute];
+"1147 select_30" [id=1147, type=select];
+"1148 select_31" [id=1148, type=select];
+"1149 select_32" [id=1149, type=select];
+"1150 linalg_vector_norm_20" [id=1150, type=linalg_vector_norm];
+"1151 clamp_min_20" [id=1151, type=clamp_min];
+"1152 expand_as_20" [id=1152, type=expand_as];
+"1153 div_20" [id=1153, type=div];
+"1154 quantize_per_tensor_default_64" [id=1154, type=quantize_per_tensor];
+"1155 dequantize_per_tensor_default_64" [id=1155, type=dequantize_per_tensor];
+"1156 linalg_vector_norm_21" [id=1156, type=linalg_vector_norm];
+"1157 clamp_min_21" [id=1157, type=clamp_min];
+"1158 expand_as_21" [id=1158, type=expand_as];
+"1159 div_21" [id=1159, type=div];
+"1160 quantize_per_tensor_default_65" [id=1160, type=quantize_per_tensor];
+"1161 dequantize_per_tensor_default_65" [id=1161, type=dequantize_per_tensor];
+"1162 transpose_20" [id=1162, type=transpose];
+"1163 matmul_20" [id=1163, type=matmul];
+"1164 _frozen_param88" [id=1164, type=get_attr];
+"1165 mul_21" [id=1165, type=mul];
+"1166 add_35" [id=1166, type=add];
+"1167 softmax_10" [id=1167, type=softmax];
+"1168 dropout_40" [id=1168, type=dropout];
+"1169 matmul_21" [id=1169, type=matmul];
+"1170 transpose_21" [id=1170, type=transpose];
+"1171 reshape_47" [id=1171, type=reshape];
+"1172 reshape_47_0_0_nncf_smooth_quant_0" [id=1172, type=call_module];
+"1173 quantize_per_tensor_default_66" [id=1173, type=quantize_per_tensor];
+"1174 dequantize_per_tensor_default_66" [id=1174, type=dequantize_per_tensor];
+"1175 linear_65_scale_0" [id=1175, type=get_attr];
+"1176 linear_65_zero_point_0" [id=1176, type=get_attr];
+"1177 compressed_weight_updated_constant44" [id=1177, type=get_attr];
+"1178 quantize_per_channel_default_55" [id=1178, type=quantize_per_channel];
+"1179 dequantize_per_channel_default_55" [id=1179, type=dequantize_per_channel];
+"1180 features_5_6_attn_proj_bias_0_0" [id=1180, type=get_attr];
+"1181 linear_65" [id=1181, type=linear];
+"1182 dropout_41" [id=1182, type=dropout];
+"1183 view_58" [id=1183, type=view];
+"1184 permute_49" [id=1184, type=permute];
+"1185 reshape_48" [id=1185, type=reshape];
+"1186 slice_162" [id=1186, type=slice];
+"1187 slice_163" [id=1187, type=slice];
+"1188 slice_164" [id=1188, type=slice];
+"1189 slice_165" [id=1189, type=slice];
+"1190 contiguous_19" [id=1190, type=contiguous];
+"1191 layer_norm_23" [id=1191, type=layer_norm];
+"1192 add_36" [id=1192, type=add];
+"1193 add_36_0_0_nncf_smooth_quant_0" [id=1193, type=call_module];
+"1194 quantize_per_tensor_default_67" [id=1194, type=quantize_per_tensor];
+"1195 dequantize_per_tensor_default_67" [id=1195, type=dequantize_per_tensor];
+"1196 linear_66_scale_0" [id=1196, type=get_attr];
+"1197 linear_66_zero_point_0" [id=1197, type=get_attr];
+"1198 compressed_weight_updated_constant45" [id=1198, type=get_attr];
+"1199 quantize_per_channel_default_56" [id=1199, type=quantize_per_channel];
+"1200 dequantize_per_channel_default_56" [id=1200, type=dequantize_per_channel];
+"1201 features_5_6_mlp_0_bias_0_0" [id=1201, type=get_attr];
+"1202 linear_66" [id=1202, type=linear];
+"1203 gelu_10" [id=1203, type=gelu];
+"1204 dropout_42" [id=1204, type=dropout];
+"1205 dropout_42_0_0_nncf_smooth_quant_0" [id=1205, type=call_module];
+"1206 quantize_per_tensor_default_68" [id=1206, type=quantize_per_tensor];
+"1207 dequantize_per_tensor_default_68" [id=1207, type=dequantize_per_tensor];
+"1208 linear_67_scale_0" [id=1208, type=get_attr];
+"1209 linear_67_zero_point_0" [id=1209, type=get_attr];
+"1210 compressed_weight_updated_constant46" [id=1210, type=get_attr];
+"1211 quantize_per_channel_default_57" [id=1211, type=quantize_per_channel];
+"1212 dequantize_per_channel_default_57" [id=1212, type=dequantize_per_channel];
+"1213 features_5_6_mlp_3_bias_0_0" [id=1213, type=get_attr];
+"1214 linear_67" [id=1214, type=linear];
+"1215 dropout_43" [id=1215, type=dropout];
+"1216 layer_norm_24" [id=1216, type=layer_norm];
+"1217 add_37" [id=1217, type=add];
+"1218 _frozen_param89" [id=1218, type=get_attr];
+"1219 pad_13" [id=1219, type=pad];
+"1220 roll_10" [id=1220, type=roll];
+"1221 view_61" [id=1221, type=view];
+"1222 permute_51" [id=1222, type=permute];
+"1223 reshape_49" [id=1223, type=reshape];
+"1224 reshape_49_0_0_nncf_smooth_quant_0" [id=1224, type=call_module];
+"1225 quantize_per_tensor_default_69" [id=1225, type=quantize_per_tensor];
+"1226 dequantize_per_tensor_default_69" [id=1226, type=dequantize_per_tensor];
+"1227 linear_70_scale_0" [id=1227, type=get_attr];
+"1228 linear_70_zero_point_0" [id=1228, type=get_attr];
+"1229 compressed_weight_updated_constant47" [id=1229, type=get_attr];
+"1230 quantize_per_channel_default_59" [id=1230, type=quantize_per_channel];
+"1231 dequantize_per_channel_default_59" [id=1231, type=dequantize_per_channel];
+"1232 _frozen_param90_0_0" [id=1232, type=get_attr];
+"1233 linear_70" [id=1233, type=linear];
+"1234 reshape_50" [id=1234, type=reshape];
+"1235 permute_52" [id=1235, type=permute];
+"1236 select_33" [id=1236, type=select];
+"1237 select_34" [id=1237, type=select];
+"1238 select_35" [id=1238, type=select];
+"1239 linalg_vector_norm_22" [id=1239, type=linalg_vector_norm];
+"1240 clamp_min_22" [id=1240, type=clamp_min];
+"1241 expand_as_22" [id=1241, type=expand_as];
+"1242 div_22" [id=1242, type=div];
+"1243 quantize_per_tensor_default_70" [id=1243, type=quantize_per_tensor];
+"1244 dequantize_per_tensor_default_70" [id=1244, type=dequantize_per_tensor];
+"1245 linalg_vector_norm_23" [id=1245, type=linalg_vector_norm];
+"1246 clamp_min_23" [id=1246, type=clamp_min];
+"1247 expand_as_23" [id=1247, type=expand_as];
+"1248 div_23" [id=1248, type=div];
+"1249 quantize_per_tensor_default_71" [id=1249, type=quantize_per_tensor];
+"1250 dequantize_per_tensor_default_71" [id=1250, type=dequantize_per_tensor];
+"1251 transpose_22" [id=1251, type=transpose];
+"1252 matmul_22" [id=1252, type=matmul];
+"1253 _frozen_param92" [id=1253, type=get_attr];
+"1254 mul_23" [id=1254, type=mul];
+"1255 add_38" [id=1255, type=add];
+"1256 new_zeros_5" [id=1256, type=new_zeros];
+"1257 view_62" [id=1257, type=view];
+"1258 permute_53" [id=1258, type=permute];
+"1259 reshape_51" [id=1259, type=reshape];
+"1260 unsqueeze_32" [id=1260, type=unsqueeze];
+"1261 unsqueeze_33" [id=1261, type=unsqueeze];
+"1262 sub_5" [id=1262, type=sub];
+"1263 ne_5" [id=1263, type=ne];
+"1264 masked_fill_10" [id=1264, type=masked_fill];
+"1265 eq_5" [id=1265, type=eq];
+"1266 masked_fill_11" [id=1266, type=masked_fill];
+"1267 view_63" [id=1267, type=view];
+"1268 unsqueeze_34" [id=1268, type=unsqueeze];
+"1269 unsqueeze_35" [id=1269, type=unsqueeze];
+"1270 add_39" [id=1270, type=add];
+"1271 view_64" [id=1271, type=view];
+"1272 softmax_11" [id=1272, type=softmax];
+"1273 dropout_44" [id=1273, type=dropout];
+"1274 matmul_23" [id=1274, type=matmul];
+"1275 transpose_23" [id=1275, type=transpose];
+"1276 reshape_52" [id=1276, type=reshape];
+"1277 reshape_52_0_0_nncf_smooth_quant_0" [id=1277, type=call_module];
+"1278 quantize_per_tensor_default_72" [id=1278, type=quantize_per_tensor];
+"1279 dequantize_per_tensor_default_72" [id=1279, type=dequantize_per_tensor];
+"1280 linear_71_scale_0" [id=1280, type=get_attr];
+"1281 linear_71_zero_point_0" [id=1281, type=get_attr];
+"1282 compressed_weight_updated_constant48" [id=1282, type=get_attr];
+"1283 quantize_per_channel_default_60" [id=1283, type=quantize_per_channel];
+"1284 dequantize_per_channel_default_60" [id=1284, type=dequantize_per_channel];
+"1285 features_5_7_attn_proj_bias_0_0" [id=1285, type=get_attr];
+"1286 linear_71" [id=1286, type=linear];
+"1287 dropout_45" [id=1287, type=dropout];
+"1288 view_65" [id=1288, type=view];
+"1289 permute_54" [id=1289, type=permute];
+"1290 reshape_53" [id=1290, type=reshape];
+"1291 roll_11" [id=1291, type=roll];
+"1292 slice_185" [id=1292, type=slice];
+"1293 slice_186" [id=1293, type=slice];
+"1294 slice_187" [id=1294, type=slice];
+"1295 slice_188" [id=1295, type=slice];
+"1296 contiguous_21" [id=1296, type=contiguous];
+"1297 layer_norm_25" [id=1297, type=layer_norm];
+"1298 add_40" [id=1298, type=add];
+"1299 add_40_0_0_nncf_smooth_quant_0" [id=1299, type=call_module];
+"1300 quantize_per_tensor_default_73" [id=1300, type=quantize_per_tensor];
+"1301 dequantize_per_tensor_default_73" [id=1301, type=dequantize_per_tensor];
+"1302 linear_72_scale_0" [id=1302, type=get_attr];
+"1303 linear_72_zero_point_0" [id=1303, type=get_attr];
+"1304 compressed_weight_updated_constant49" [id=1304, type=get_attr];
+"1305 quantize_per_channel_default_61" [id=1305, type=quantize_per_channel];
+"1306 dequantize_per_channel_default_61" [id=1306, type=dequantize_per_channel];
+"1307 features_5_7_mlp_0_bias_0_0" [id=1307, type=get_attr];
+"1308 linear_72" [id=1308, type=linear];
+"1309 gelu_11" [id=1309, type=gelu];
+"1310 dropout_46" [id=1310, type=dropout];
+"1311 dropout_46_0_0_nncf_smooth_quant_0" [id=1311, type=call_module];
+"1312 quantize_per_tensor_default_74" [id=1312, type=quantize_per_tensor];
+"1313 dequantize_per_tensor_default_74" [id=1313, type=dequantize_per_tensor];
+"1314 linear_73_scale_0" [id=1314, type=get_attr];
+"1315 linear_73_zero_point_0" [id=1315, type=get_attr];
+"1316 compressed_weight_updated_constant50" [id=1316, type=get_attr];
+"1317 quantize_per_channel_default_62" [id=1317, type=quantize_per_channel];
+"1318 dequantize_per_channel_default_62" [id=1318, type=dequantize_per_channel];
+"1319 features_5_7_mlp_3_bias_0_0" [id=1319, type=get_attr];
+"1320 linear_73" [id=1320, type=linear];
+"1321 dropout_47" [id=1321, type=dropout];
+"1322 layer_norm_26" [id=1322, type=layer_norm];
+"1323 add_41" [id=1323, type=add];
+"1324 _frozen_param102" [id=1324, type=get_attr];
+"1325 pad_14" [id=1325, type=pad];
+"1326 view_68" [id=1326, type=view];
+"1327 permute_56" [id=1327, type=permute];
+"1328 reshape_54" [id=1328, type=reshape];
+"1329 reshape_54_0_0_nncf_smooth_quant_0" [id=1329, type=call_module];
+"1330 quantize_per_tensor_default_75" [id=1330, type=quantize_per_tensor];
+"1331 dequantize_per_tensor_default_75" [id=1331, type=dequantize_per_tensor];
+"1332 linear_76_scale_0" [id=1332, type=get_attr];
+"1333 linear_76_zero_point_0" [id=1333, type=get_attr];
+"1334 compressed_weight_updated_constant51" [id=1334, type=get_attr];
+"1335 quantize_per_channel_default_64" [id=1335, type=quantize_per_channel];
+"1336 dequantize_per_channel_default_64" [id=1336, type=dequantize_per_channel];
+"1337 _frozen_param103_0_0" [id=1337, type=get_attr];
+"1338 linear_76" [id=1338, type=linear];
+"1339 reshape_55" [id=1339, type=reshape];
+"1340 permute_57" [id=1340, type=permute];
+"1341 select_36" [id=1341, type=select];
+"1342 select_37" [id=1342, type=select];
+"1343 select_38" [id=1343, type=select];
+"1344 linalg_vector_norm_24" [id=1344, type=linalg_vector_norm];
+"1345 clamp_min_24" [id=1345, type=clamp_min];
+"1346 expand_as_24" [id=1346, type=expand_as];
+"1347 div_24" [id=1347, type=div];
+"1348 quantize_per_tensor_default_76" [id=1348, type=quantize_per_tensor];
+"1349 dequantize_per_tensor_default_76" [id=1349, type=dequantize_per_tensor];
+"1350 linalg_vector_norm_25" [id=1350, type=linalg_vector_norm];
+"1351 clamp_min_25" [id=1351, type=clamp_min];
+"1352 expand_as_25" [id=1352, type=expand_as];
+"1353 div_25" [id=1353, type=div];
+"1354 quantize_per_tensor_default_77" [id=1354, type=quantize_per_tensor];
+"1355 dequantize_per_tensor_default_77" [id=1355, type=dequantize_per_tensor];
+"1356 transpose_24" [id=1356, type=transpose];
+"1357 matmul_24" [id=1357, type=matmul];
+"1358 _frozen_param105" [id=1358, type=get_attr];
+"1359 mul_25" [id=1359, type=mul];
+"1360 add_42" [id=1360, type=add];
+"1361 softmax_12" [id=1361, type=softmax];
+"1362 dropout_48" [id=1362, type=dropout];
+"1363 matmul_25" [id=1363, type=matmul];
+"1364 transpose_25" [id=1364, type=transpose];
+"1365 reshape_56" [id=1365, type=reshape];
+"1366 reshape_56_0_0_nncf_smooth_quant_0" [id=1366, type=call_module];
+"1367 quantize_per_tensor_default_78" [id=1367, type=quantize_per_tensor];
+"1368 dequantize_per_tensor_default_78" [id=1368, type=dequantize_per_tensor];
+"1369 linear_77_scale_0" [id=1369, type=get_attr];
+"1370 linear_77_zero_point_0" [id=1370, type=get_attr];
+"1371 compressed_weight_updated_constant52" [id=1371, type=get_attr];
+"1372 quantize_per_channel_default_65" [id=1372, type=quantize_per_channel];
+"1373 dequantize_per_channel_default_65" [id=1373, type=dequantize_per_channel];
+"1374 features_5_8_attn_proj_bias_0_0" [id=1374, type=get_attr];
+"1375 linear_77" [id=1375, type=linear];
+"1376 dropout_49" [id=1376, type=dropout];
+"1377 view_69" [id=1377, type=view];
+"1378 permute_58" [id=1378, type=permute];
+"1379 reshape_57" [id=1379, type=reshape];
+"1380 slice_190" [id=1380, type=slice];
+"1381 slice_191" [id=1381, type=slice];
+"1382 slice_192" [id=1382, type=slice];
+"1383 slice_193" [id=1383, type=slice];
+"1384 contiguous_23" [id=1384, type=contiguous];
+"1385 layer_norm_27" [id=1385, type=layer_norm];
+"1386 add_43" [id=1386, type=add];
+"1387 add_43_0_0_nncf_smooth_quant_0" [id=1387, type=call_module];
+"1388 quantize_per_tensor_default_79" [id=1388, type=quantize_per_tensor];
+"1389 dequantize_per_tensor_default_79" [id=1389, type=dequantize_per_tensor];
+"1390 linear_78_scale_0" [id=1390, type=get_attr];
+"1391 linear_78_zero_point_0" [id=1391, type=get_attr];
+"1392 compressed_weight_updated_constant53" [id=1392, type=get_attr];
+"1393 quantize_per_channel_default_66" [id=1393, type=quantize_per_channel];
+"1394 dequantize_per_channel_default_66" [id=1394, type=dequantize_per_channel];
+"1395 features_5_8_mlp_0_bias_0_0" [id=1395, type=get_attr];
+"1396 linear_78" [id=1396, type=linear];
+"1397 gelu_12" [id=1397, type=gelu];
+"1398 dropout_50" [id=1398, type=dropout];
+"1399 dropout_50_0_0_nncf_smooth_quant_0" [id=1399, type=call_module];
+"1400 quantize_per_tensor_default_80" [id=1400, type=quantize_per_tensor];
+"1401 dequantize_per_tensor_default_80" [id=1401, type=dequantize_per_tensor];
+"1402 linear_79_scale_0" [id=1402, type=get_attr];
+"1403 linear_79_zero_point_0" [id=1403, type=get_attr];
+"1404 compressed_weight_updated_constant54" [id=1404, type=get_attr];
+"1405 quantize_per_channel_default_67" [id=1405, type=quantize_per_channel];
+"1406 dequantize_per_channel_default_67" [id=1406, type=dequantize_per_channel];
+"1407 features_5_8_mlp_3_bias_0_0" [id=1407, type=get_attr];
+"1408 linear_79" [id=1408, type=linear];
+"1409 dropout_51" [id=1409, type=dropout];
+"1410 layer_norm_28" [id=1410, type=layer_norm];
+"1411 add_44" [id=1411, type=add];
+"1412 _frozen_param106" [id=1412, type=get_attr];
+"1413 pad_15" [id=1413, type=pad];
+"1414 roll_12" [id=1414, type=roll];
+"1415 view_72" [id=1415, type=view];
+"1416 permute_60" [id=1416, type=permute];
+"1417 reshape_58" [id=1417, type=reshape];
+"1418 reshape_58_0_0_nncf_smooth_quant_0" [id=1418, type=call_module];
+"1419 quantize_per_tensor_default_81" [id=1419, type=quantize_per_tensor];
+"1420 dequantize_per_tensor_default_81" [id=1420, type=dequantize_per_tensor];
+"1421 linear_82_scale_0" [id=1421, type=get_attr];
+"1422 linear_82_zero_point_0" [id=1422, type=get_attr];
+"1423 compressed_weight_updated_constant55" [id=1423, type=get_attr];
+"1424 quantize_per_channel_default_69" [id=1424, type=quantize_per_channel];
+"1425 dequantize_per_channel_default_69" [id=1425, type=dequantize_per_channel];
+"1426 _frozen_param107_0_0" [id=1426, type=get_attr];
+"1427 linear_82" [id=1427, type=linear];
+"1428 reshape_59" [id=1428, type=reshape];
+"1429 permute_61" [id=1429, type=permute];
+"1430 select_39" [id=1430, type=select];
+"1431 select_40" [id=1431, type=select];
+"1432 select_41" [id=1432, type=select];
+"1433 linalg_vector_norm_26" [id=1433, type=linalg_vector_norm];
+"1434 clamp_min_26" [id=1434, type=clamp_min];
+"1435 expand_as_26" [id=1435, type=expand_as];
+"1436 div_26" [id=1436, type=div];
+"1437 quantize_per_tensor_default_82" [id=1437, type=quantize_per_tensor];
+"1438 dequantize_per_tensor_default_82" [id=1438, type=dequantize_per_tensor];
+"1439 linalg_vector_norm_27" [id=1439, type=linalg_vector_norm];
+"1440 clamp_min_27" [id=1440, type=clamp_min];
+"1441 expand_as_27" [id=1441, type=expand_as];
+"1442 div_27" [id=1442, type=div];
+"1443 quantize_per_tensor_default_83" [id=1443, type=quantize_per_tensor];
+"1444 dequantize_per_tensor_default_83" [id=1444, type=dequantize_per_tensor];
+"1445 transpose_26" [id=1445, type=transpose];
+"1446 matmul_26" [id=1446, type=matmul];
+"1447 _frozen_param109" [id=1447, type=get_attr];
+"1448 mul_27" [id=1448, type=mul];
+"1449 add_45" [id=1449, type=add];
+"1450 new_zeros_6" [id=1450, type=new_zeros];
+"1451 view_73" [id=1451, type=view];
+"1452 permute_62" [id=1452, type=permute];
+"1453 reshape_60" [id=1453, type=reshape];
+"1454 unsqueeze_38" [id=1454, type=unsqueeze];
+"1455 unsqueeze_39" [id=1455, type=unsqueeze];
+"1456 sub_6" [id=1456, type=sub];
+"1457 ne_6" [id=1457, type=ne];
+"1458 masked_fill_12" [id=1458, type=masked_fill];
+"1459 eq_6" [id=1459, type=eq];
+"1460 masked_fill_13" [id=1460, type=masked_fill];
+"1461 view_74" [id=1461, type=view];
+"1462 unsqueeze_40" [id=1462, type=unsqueeze];
+"1463 unsqueeze_41" [id=1463, type=unsqueeze];
+"1464 add_46" [id=1464, type=add];
+"1465 view_75" [id=1465, type=view];
+"1466 softmax_13" [id=1466, type=softmax];
+"1467 dropout_52" [id=1467, type=dropout];
+"1468 matmul_27" [id=1468, type=matmul];
+"1469 transpose_27" [id=1469, type=transpose];
+"1470 reshape_61" [id=1470, type=reshape];
+"1471 reshape_61_0_0_nncf_smooth_quant_0" [id=1471, type=call_module];
+"1472 quantize_per_tensor_default_84" [id=1472, type=quantize_per_tensor];
+"1473 dequantize_per_tensor_default_84" [id=1473, type=dequantize_per_tensor];
+"1474 linear_83_scale_0" [id=1474, type=get_attr];
+"1475 linear_83_zero_point_0" [id=1475, type=get_attr];
+"1476 compressed_weight_updated_constant56" [id=1476, type=get_attr];
+"1477 quantize_per_channel_default_70" [id=1477, type=quantize_per_channel];
+"1478 dequantize_per_channel_default_70" [id=1478, type=dequantize_per_channel];
+"1479 features_5_9_attn_proj_bias_0_0" [id=1479, type=get_attr];
+"1480 linear_83" [id=1480, type=linear];
+"1481 dropout_53" [id=1481, type=dropout];
+"1482 view_76" [id=1482, type=view];
+"1483 permute_63" [id=1483, type=permute];
+"1484 reshape_62" [id=1484, type=reshape];
+"1485 roll_13" [id=1485, type=roll];
+"1486 slice_213" [id=1486, type=slice];
+"1487 slice_214" [id=1487, type=slice];
+"1488 slice_215" [id=1488, type=slice];
+"1489 slice_216" [id=1489, type=slice];
+"1490 contiguous_25" [id=1490, type=contiguous];
+"1491 layer_norm_29" [id=1491, type=layer_norm];
+"1492 add_47" [id=1492, type=add];
+"1493 add_47_0_0_nncf_smooth_quant_0" [id=1493, type=call_module];
+"1494 quantize_per_tensor_default_85" [id=1494, type=quantize_per_tensor];
+"1495 dequantize_per_tensor_default_85" [id=1495, type=dequantize_per_tensor];
+"1496 linear_84_scale_0" [id=1496, type=get_attr];
+"1497 linear_84_zero_point_0" [id=1497, type=get_attr];
+"1498 compressed_weight_updated_constant57" [id=1498, type=get_attr];
+"1499 quantize_per_channel_default_71" [id=1499, type=quantize_per_channel];
+"1500 dequantize_per_channel_default_71" [id=1500, type=dequantize_per_channel];
+"1501 features_5_9_mlp_0_bias_0_0" [id=1501, type=get_attr];
+"1502 linear_84" [id=1502, type=linear];
+"1503 gelu_13" [id=1503, type=gelu];
+"1504 dropout_54" [id=1504, type=dropout];
+"1505 dropout_54_0_0_nncf_smooth_quant_0" [id=1505, type=call_module];
+"1506 quantize_per_tensor_default_86" [id=1506, type=quantize_per_tensor];
+"1507 dequantize_per_tensor_default_86" [id=1507, type=dequantize_per_tensor];
+"1508 linear_85_scale_0" [id=1508, type=get_attr];
+"1509 linear_85_zero_point_0" [id=1509, type=get_attr];
+"1510 compressed_weight_updated_constant58" [id=1510, type=get_attr];
+"1511 quantize_per_channel_default_72" [id=1511, type=quantize_per_channel];
+"1512 dequantize_per_channel_default_72" [id=1512, type=dequantize_per_channel];
+"1513 features_5_9_mlp_3_bias_0_0" [id=1513, type=get_attr];
+"1514 linear_85" [id=1514, type=linear];
+"1515 dropout_55" [id=1515, type=dropout];
+"1516 layer_norm_30" [id=1516, type=layer_norm];
+"1517 add_48" [id=1517, type=add];
+"1518 _frozen_param119" [id=1518, type=get_attr];
+"1519 pad_16" [id=1519, type=pad];
+"1520 view_79" [id=1520, type=view];
+"1521 permute_65" [id=1521, type=permute];
+"1522 reshape_63" [id=1522, type=reshape];
+"1523 reshape_63_0_0_nncf_smooth_quant_0" [id=1523, type=call_module];
+"1524 quantize_per_tensor_default_87" [id=1524, type=quantize_per_tensor];
+"1525 dequantize_per_tensor_default_87" [id=1525, type=dequantize_per_tensor];
+"1526 linear_88_scale_0" [id=1526, type=get_attr];
+"1527 linear_88_zero_point_0" [id=1527, type=get_attr];
+"1528 compressed_weight_updated_constant59" [id=1528, type=get_attr];
+"1529 quantize_per_channel_default_74" [id=1529, type=quantize_per_channel];
+"1530 dequantize_per_channel_default_74" [id=1530, type=dequantize_per_channel];
+"1531 _frozen_param120_0_0" [id=1531, type=get_attr];
+"1532 linear_88" [id=1532, type=linear];
+"1533 reshape_64" [id=1533, type=reshape];
+"1534 permute_66" [id=1534, type=permute];
+"1535 select_42" [id=1535, type=select];
+"1536 select_43" [id=1536, type=select];
+"1537 select_44" [id=1537, type=select];
+"1538 linalg_vector_norm_28" [id=1538, type=linalg_vector_norm];
+"1539 clamp_min_28" [id=1539, type=clamp_min];
+"1540 expand_as_28" [id=1540, type=expand_as];
+"1541 div_28" [id=1541, type=div];
+"1542 quantize_per_tensor_default_88" [id=1542, type=quantize_per_tensor];
+"1543 dequantize_per_tensor_default_88" [id=1543, type=dequantize_per_tensor];
+"1544 linalg_vector_norm_29" [id=1544, type=linalg_vector_norm];
+"1545 clamp_min_29" [id=1545, type=clamp_min];
+"1546 expand_as_29" [id=1546, type=expand_as];
+"1547 div_29" [id=1547, type=div];
+"1548 quantize_per_tensor_default_89" [id=1548, type=quantize_per_tensor];
+"1549 dequantize_per_tensor_default_89" [id=1549, type=dequantize_per_tensor];
+"1550 transpose_28" [id=1550, type=transpose];
+"1551 matmul_28" [id=1551, type=matmul];
+"1552 _frozen_param122" [id=1552, type=get_attr];
+"1553 mul_29" [id=1553, type=mul];
+"1554 add_49" [id=1554, type=add];
+"1555 softmax_14" [id=1555, type=softmax];
+"1556 dropout_56" [id=1556, type=dropout];
+"1557 matmul_29" [id=1557, type=matmul];
+"1558 transpose_29" [id=1558, type=transpose];
+"1559 reshape_65" [id=1559, type=reshape];
+"1560 reshape_65_0_0_nncf_smooth_quant_0" [id=1560, type=call_module];
+"1561 quantize_per_tensor_default_90" [id=1561, type=quantize_per_tensor];
+"1562 dequantize_per_tensor_default_90" [id=1562, type=dequantize_per_tensor];
+"1563 linear_89_scale_0" [id=1563, type=get_attr];
+"1564 linear_89_zero_point_0" [id=1564, type=get_attr];
+"1565 compressed_weight_updated_constant60" [id=1565, type=get_attr];
+"1566 quantize_per_channel_default_75" [id=1566, type=quantize_per_channel];
+"1567 dequantize_per_channel_default_75" [id=1567, type=dequantize_per_channel];
+"1568 features_5_10_attn_proj_bias_0_0" [id=1568, type=get_attr];
+"1569 linear_89" [id=1569, type=linear];
+"1570 dropout_57" [id=1570, type=dropout];
+"1571 view_80" [id=1571, type=view];
+"1572 permute_67" [id=1572, type=permute];
+"1573 reshape_66" [id=1573, type=reshape];
+"1574 slice_218" [id=1574, type=slice];
+"1575 slice_219" [id=1575, type=slice];
+"1576 slice_220" [id=1576, type=slice];
+"1577 slice_221" [id=1577, type=slice];
+"1578 contiguous_27" [id=1578, type=contiguous];
+"1579 layer_norm_31" [id=1579, type=layer_norm];
+"1580 add_50" [id=1580, type=add];
+"1581 add_50_0_0_nncf_smooth_quant_0" [id=1581, type=call_module];
+"1582 quantize_per_tensor_default_91" [id=1582, type=quantize_per_tensor];
+"1583 dequantize_per_tensor_default_91" [id=1583, type=dequantize_per_tensor];
+"1584 linear_90_scale_0" [id=1584, type=get_attr];
+"1585 linear_90_zero_point_0" [id=1585, type=get_attr];
+"1586 compressed_weight_updated_constant61" [id=1586, type=get_attr];
+"1587 quantize_per_channel_default_76" [id=1587, type=quantize_per_channel];
+"1588 dequantize_per_channel_default_76" [id=1588, type=dequantize_per_channel];
+"1589 features_5_10_mlp_0_bias_0_0" [id=1589, type=get_attr];
+"1590 linear_90" [id=1590, type=linear];
+"1591 gelu_14" [id=1591, type=gelu];
+"1592 dropout_58" [id=1592, type=dropout];
+"1593 dropout_58_0_0_nncf_smooth_quant_0" [id=1593, type=call_module];
+"1594 quantize_per_tensor_default_92" [id=1594, type=quantize_per_tensor];
+"1595 dequantize_per_tensor_default_92" [id=1595, type=dequantize_per_tensor];
+"1596 linear_91_scale_0" [id=1596, type=get_attr];
+"1597 linear_91_zero_point_0" [id=1597, type=get_attr];
+"1598 compressed_weight_updated_constant62" [id=1598, type=get_attr];
+"1599 quantize_per_channel_default_77" [id=1599, type=quantize_per_channel];
+"1600 dequantize_per_channel_default_77" [id=1600, type=dequantize_per_channel];
+"1601 features_5_10_mlp_3_bias_0_0" [id=1601, type=get_attr];
+"1602 linear_91" [id=1602, type=linear];
+"1603 dropout_59" [id=1603, type=dropout];
+"1604 layer_norm_32" [id=1604, type=layer_norm];
+"1605 add_51" [id=1605, type=add];
+"1606 _frozen_param123" [id=1606, type=get_attr];
+"1607 pad_17" [id=1607, type=pad];
+"1608 roll_14" [id=1608, type=roll];
+"1609 view_83" [id=1609, type=view];
+"1610 permute_69" [id=1610, type=permute];
+"1611 reshape_67" [id=1611, type=reshape];
+"1612 reshape_67_0_0_nncf_smooth_quant_0" [id=1612, type=call_module];
+"1613 quantize_per_tensor_default_93" [id=1613, type=quantize_per_tensor];
+"1614 dequantize_per_tensor_default_93" [id=1614, type=dequantize_per_tensor];
+"1615 linear_94_scale_0" [id=1615, type=get_attr];
+"1616 linear_94_zero_point_0" [id=1616, type=get_attr];
+"1617 compressed_weight_updated_constant63" [id=1617, type=get_attr];
+"1618 quantize_per_channel_default_79" [id=1618, type=quantize_per_channel];
+"1619 dequantize_per_channel_default_79" [id=1619, type=dequantize_per_channel];
+"1620 _frozen_param124_0_0" [id=1620, type=get_attr];
+"1621 linear_94" [id=1621, type=linear];
+"1622 reshape_68" [id=1622, type=reshape];
+"1623 permute_70" [id=1623, type=permute];
+"1624 select_45" [id=1624, type=select];
+"1625 select_46" [id=1625, type=select];
+"1626 select_47" [id=1626, type=select];
+"1627 linalg_vector_norm_30" [id=1627, type=linalg_vector_norm];
+"1628 clamp_min_30" [id=1628, type=clamp_min];
+"1629 expand_as_30" [id=1629, type=expand_as];
+"1630 div_30" [id=1630, type=div];
+"1631 quantize_per_tensor_default_94" [id=1631, type=quantize_per_tensor];
+"1632 dequantize_per_tensor_default_94" [id=1632, type=dequantize_per_tensor];
+"1633 linalg_vector_norm_31" [id=1633, type=linalg_vector_norm];
+"1634 clamp_min_31" [id=1634, type=clamp_min];
+"1635 expand_as_31" [id=1635, type=expand_as];
+"1636 div_31" [id=1636, type=div];
+"1637 quantize_per_tensor_default_95" [id=1637, type=quantize_per_tensor];
+"1638 dequantize_per_tensor_default_95" [id=1638, type=dequantize_per_tensor];
+"1639 transpose_30" [id=1639, type=transpose];
+"1640 matmul_30" [id=1640, type=matmul];
+"1641 _frozen_param126" [id=1641, type=get_attr];
+"1642 mul_31" [id=1642, type=mul];
+"1643 add_52" [id=1643, type=add];
+"1644 new_zeros_7" [id=1644, type=new_zeros];
+"1645 view_84" [id=1645, type=view];
+"1646 permute_71" [id=1646, type=permute];
+"1647 reshape_69" [id=1647, type=reshape];
+"1648 unsqueeze_44" [id=1648, type=unsqueeze];
+"1649 unsqueeze_45" [id=1649, type=unsqueeze];
+"1650 sub_7" [id=1650, type=sub];
+"1651 ne_7" [id=1651, type=ne];
+"1652 masked_fill_14" [id=1652, type=masked_fill];
+"1653 eq_7" [id=1653, type=eq];
+"1654 masked_fill_15" [id=1654, type=masked_fill];
+"1655 view_85" [id=1655, type=view];
+"1656 unsqueeze_46" [id=1656, type=unsqueeze];
+"1657 unsqueeze_47" [id=1657, type=unsqueeze];
+"1658 add_53" [id=1658, type=add];
+"1659 view_86" [id=1659, type=view];
+"1660 softmax_15" [id=1660, type=softmax];
+"1661 dropout_60" [id=1661, type=dropout];
+"1662 matmul_31" [id=1662, type=matmul];
+"1663 transpose_31" [id=1663, type=transpose];
+"1664 reshape_70" [id=1664, type=reshape];
+"1665 reshape_70_0_0_nncf_smooth_quant_0" [id=1665, type=call_module];
+"1666 quantize_per_tensor_default_96" [id=1666, type=quantize_per_tensor];
+"1667 dequantize_per_tensor_default_96" [id=1667, type=dequantize_per_tensor];
+"1668 linear_95_scale_0" [id=1668, type=get_attr];
+"1669 linear_95_zero_point_0" [id=1669, type=get_attr];
+"1670 compressed_weight_updated_constant64" [id=1670, type=get_attr];
+"1671 quantize_per_channel_default_80" [id=1671, type=quantize_per_channel];
+"1672 dequantize_per_channel_default_80" [id=1672, type=dequantize_per_channel];
+"1673 features_5_11_attn_proj_bias_0_0" [id=1673, type=get_attr];
+"1674 linear_95" [id=1674, type=linear];
+"1675 dropout_61" [id=1675, type=dropout];
+"1676 view_87" [id=1676, type=view];
+"1677 permute_72" [id=1677, type=permute];
+"1678 reshape_71" [id=1678, type=reshape];
+"1679 roll_15" [id=1679, type=roll];
+"1680 slice_241" [id=1680, type=slice];
+"1681 slice_242" [id=1681, type=slice];
+"1682 slice_243" [id=1682, type=slice];
+"1683 slice_244" [id=1683, type=slice];
+"1684 contiguous_29" [id=1684, type=contiguous];
+"1685 layer_norm_33" [id=1685, type=layer_norm];
+"1686 add_54" [id=1686, type=add];
+"1687 add_54_0_0_nncf_smooth_quant_0" [id=1687, type=call_module];
+"1688 quantize_per_tensor_default_97" [id=1688, type=quantize_per_tensor];
+"1689 dequantize_per_tensor_default_97" [id=1689, type=dequantize_per_tensor];
+"1690 linear_96_scale_0" [id=1690, type=get_attr];
+"1691 linear_96_zero_point_0" [id=1691, type=get_attr];
+"1692 compressed_weight_updated_constant65" [id=1692, type=get_attr];
+"1693 quantize_per_channel_default_81" [id=1693, type=quantize_per_channel];
+"1694 dequantize_per_channel_default_81" [id=1694, type=dequantize_per_channel];
+"1695 features_5_11_mlp_0_bias_0_0" [id=1695, type=get_attr];
+"1696 linear_96" [id=1696, type=linear];
+"1697 gelu_15" [id=1697, type=gelu];
+"1698 dropout_62" [id=1698, type=dropout];
+"1699 dropout_62_0_0_nncf_smooth_quant_0" [id=1699, type=call_module];
+"1700 quantize_per_tensor_default_98" [id=1700, type=quantize_per_tensor];
+"1701 dequantize_per_tensor_default_98" [id=1701, type=dequantize_per_tensor];
+"1702 linear_97_scale_0" [id=1702, type=get_attr];
+"1703 linear_97_zero_point_0" [id=1703, type=get_attr];
+"1704 compressed_weight_updated_constant66" [id=1704, type=get_attr];
+"1705 quantize_per_channel_default_82" [id=1705, type=quantize_per_channel];
+"1706 dequantize_per_channel_default_82" [id=1706, type=dequantize_per_channel];
+"1707 features_5_11_mlp_3_bias_0_0" [id=1707, type=get_attr];
+"1708 linear_97" [id=1708, type=linear];
+"1709 dropout_63" [id=1709, type=dropout];
+"1710 layer_norm_34" [id=1710, type=layer_norm];
+"1711 add_55" [id=1711, type=add];
+"1712 _frozen_param136" [id=1712, type=get_attr];
+"1713 pad_18" [id=1713, type=pad];
+"1714 view_90" [id=1714, type=view];
+"1715 permute_74" [id=1715, type=permute];
+"1716 reshape_72" [id=1716, type=reshape];
+"1717 reshape_72_0_0_nncf_smooth_quant_0" [id=1717, type=call_module];
+"1718 quantize_per_tensor_default_99" [id=1718, type=quantize_per_tensor];
+"1719 dequantize_per_tensor_default_99" [id=1719, type=dequantize_per_tensor];
+"1720 linear_100_scale_0" [id=1720, type=get_attr];
+"1721 linear_100_zero_point_0" [id=1721, type=get_attr];
+"1722 compressed_weight_updated_constant67" [id=1722, type=get_attr];
+"1723 quantize_per_channel_default_84" [id=1723, type=quantize_per_channel];
+"1724 dequantize_per_channel_default_84" [id=1724, type=dequantize_per_channel];
+"1725 _frozen_param137_0_0" [id=1725, type=get_attr];
+"1726 linear_100" [id=1726, type=linear];
+"1727 reshape_73" [id=1727, type=reshape];
+"1728 permute_75" [id=1728, type=permute];
+"1729 select_48" [id=1729, type=select];
+"1730 select_49" [id=1730, type=select];
+"1731 select_50" [id=1731, type=select];
+"1732 linalg_vector_norm_32" [id=1732, type=linalg_vector_norm];
+"1733 clamp_min_32" [id=1733, type=clamp_min];
+"1734 expand_as_32" [id=1734, type=expand_as];
+"1735 div_32" [id=1735, type=div];
+"1736 quantize_per_tensor_default_100" [id=1736, type=quantize_per_tensor];
+"1737 dequantize_per_tensor_default_100" [id=1737, type=dequantize_per_tensor];
+"1738 linalg_vector_norm_33" [id=1738, type=linalg_vector_norm];
+"1739 clamp_min_33" [id=1739, type=clamp_min];
+"1740 expand_as_33" [id=1740, type=expand_as];
+"1741 div_33" [id=1741, type=div];
+"1742 quantize_per_tensor_default_101" [id=1742, type=quantize_per_tensor];
+"1743 dequantize_per_tensor_default_101" [id=1743, type=dequantize_per_tensor];
+"1744 transpose_32" [id=1744, type=transpose];
+"1745 matmul_32" [id=1745, type=matmul];
+"1746 _frozen_param139" [id=1746, type=get_attr];
+"1747 mul_33" [id=1747, type=mul];
+"1748 add_56" [id=1748, type=add];
+"1749 softmax_16" [id=1749, type=softmax];
+"1750 dropout_64" [id=1750, type=dropout];
+"1751 matmul_33" [id=1751, type=matmul];
+"1752 transpose_33" [id=1752, type=transpose];
+"1753 reshape_74" [id=1753, type=reshape];
+"1754 reshape_74_0_0_nncf_smooth_quant_0" [id=1754, type=call_module];
+"1755 quantize_per_tensor_default_102" [id=1755, type=quantize_per_tensor];
+"1756 dequantize_per_tensor_default_102" [id=1756, type=dequantize_per_tensor];
+"1757 linear_101_scale_0" [id=1757, type=get_attr];
+"1758 linear_101_zero_point_0" [id=1758, type=get_attr];
+"1759 compressed_weight_updated_constant68" [id=1759, type=get_attr];
+"1760 quantize_per_channel_default_85" [id=1760, type=quantize_per_channel];
+"1761 dequantize_per_channel_default_85" [id=1761, type=dequantize_per_channel];
+"1762 features_5_12_attn_proj_bias_0_0" [id=1762, type=get_attr];
+"1763 linear_101" [id=1763, type=linear];
+"1764 dropout_65" [id=1764, type=dropout];
+"1765 view_91" [id=1765, type=view];
+"1766 permute_76" [id=1766, type=permute];
+"1767 reshape_75" [id=1767, type=reshape];
+"1768 slice_246" [id=1768, type=slice];
+"1769 slice_247" [id=1769, type=slice];
+"1770 slice_248" [id=1770, type=slice];
+"1771 slice_249" [id=1771, type=slice];
+"1772 contiguous_31" [id=1772, type=contiguous];
+"1773 layer_norm_35" [id=1773, type=layer_norm];
+"1774 add_57" [id=1774, type=add];
+"1775 add_57_0_0_nncf_smooth_quant_0" [id=1775, type=call_module];
+"1776 quantize_per_tensor_default_103" [id=1776, type=quantize_per_tensor];
+"1777 dequantize_per_tensor_default_103" [id=1777, type=dequantize_per_tensor];
+"1778 linear_102_scale_0" [id=1778, type=get_attr];
+"1779 linear_102_zero_point_0" [id=1779, type=get_attr];
+"1780 compressed_weight_updated_constant69" [id=1780, type=get_attr];
+"1781 quantize_per_channel_default_86" [id=1781, type=quantize_per_channel];
+"1782 dequantize_per_channel_default_86" [id=1782, type=dequantize_per_channel];
+"1783 features_5_12_mlp_0_bias_0_0" [id=1783, type=get_attr];
+"1784 linear_102" [id=1784, type=linear];
+"1785 gelu_16" [id=1785, type=gelu];
+"1786 dropout_66" [id=1786, type=dropout];
+"1787 dropout_66_0_0_nncf_smooth_quant_0" [id=1787, type=call_module];
+"1788 quantize_per_tensor_default_104" [id=1788, type=quantize_per_tensor];
+"1789 dequantize_per_tensor_default_104" [id=1789, type=dequantize_per_tensor];
+"1790 linear_103_scale_0" [id=1790, type=get_attr];
+"1791 linear_103_zero_point_0" [id=1791, type=get_attr];
+"1792 compressed_weight_updated_constant70" [id=1792, type=get_attr];
+"1793 quantize_per_channel_default_87" [id=1793, type=quantize_per_channel];
+"1794 dequantize_per_channel_default_87" [id=1794, type=dequantize_per_channel];
+"1795 features_5_12_mlp_3_bias_0_0" [id=1795, type=get_attr];
+"1796 linear_103" [id=1796, type=linear];
+"1797 dropout_67" [id=1797, type=dropout];
+"1798 layer_norm_36" [id=1798, type=layer_norm];
+"1799 add_58" [id=1799, type=add];
+"1800 _frozen_param140" [id=1800, type=get_attr];
+"1801 pad_19" [id=1801, type=pad];
+"1802 roll_16" [id=1802, type=roll];
+"1803 view_94" [id=1803, type=view];
+"1804 permute_78" [id=1804, type=permute];
+"1805 reshape_76" [id=1805, type=reshape];
+"1806 reshape_76_0_0_nncf_smooth_quant_0" [id=1806, type=call_module];
+"1807 quantize_per_tensor_default_105" [id=1807, type=quantize_per_tensor];
+"1808 dequantize_per_tensor_default_105" [id=1808, type=dequantize_per_tensor];
+"1809 linear_106_scale_0" [id=1809, type=get_attr];
+"1810 linear_106_zero_point_0" [id=1810, type=get_attr];
+"1811 compressed_weight_updated_constant71" [id=1811, type=get_attr];
+"1812 quantize_per_channel_default_89" [id=1812, type=quantize_per_channel];
+"1813 dequantize_per_channel_default_89" [id=1813, type=dequantize_per_channel];
+"1814 _frozen_param141_0_0" [id=1814, type=get_attr];
+"1815 linear_106" [id=1815, type=linear];
+"1816 reshape_77" [id=1816, type=reshape];
+"1817 permute_79" [id=1817, type=permute];
+"1818 select_51" [id=1818, type=select];
+"1819 select_52" [id=1819, type=select];
+"1820 select_53" [id=1820, type=select];
+"1821 linalg_vector_norm_34" [id=1821, type=linalg_vector_norm];
+"1822 clamp_min_34" [id=1822, type=clamp_min];
+"1823 expand_as_34" [id=1823, type=expand_as];
+"1824 div_34" [id=1824, type=div];
+"1825 quantize_per_tensor_default_106" [id=1825, type=quantize_per_tensor];
+"1826 dequantize_per_tensor_default_106" [id=1826, type=dequantize_per_tensor];
+"1827 linalg_vector_norm_35" [id=1827, type=linalg_vector_norm];
+"1828 clamp_min_35" [id=1828, type=clamp_min];
+"1829 expand_as_35" [id=1829, type=expand_as];
+"1830 div_35" [id=1830, type=div];
+"1831 quantize_per_tensor_default_107" [id=1831, type=quantize_per_tensor];
+"1832 dequantize_per_tensor_default_107" [id=1832, type=dequantize_per_tensor];
+"1833 transpose_34" [id=1833, type=transpose];
+"1834 matmul_34" [id=1834, type=matmul];
+"1835 _frozen_param143" [id=1835, type=get_attr];
+"1836 mul_35" [id=1836, type=mul];
+"1837 add_59" [id=1837, type=add];
+"1838 new_zeros_8" [id=1838, type=new_zeros];
+"1839 view_95" [id=1839, type=view];
+"1840 permute_80" [id=1840, type=permute];
+"1841 reshape_78" [id=1841, type=reshape];
+"1842 unsqueeze_50" [id=1842, type=unsqueeze];
+"1843 unsqueeze_51" [id=1843, type=unsqueeze];
+"1844 sub_8" [id=1844, type=sub];
+"1845 ne_8" [id=1845, type=ne];
+"1846 masked_fill_16" [id=1846, type=masked_fill];
+"1847 eq_8" [id=1847, type=eq];
+"1848 masked_fill_17" [id=1848, type=masked_fill];
+"1849 view_96" [id=1849, type=view];
+"1850 unsqueeze_52" [id=1850, type=unsqueeze];
+"1851 unsqueeze_53" [id=1851, type=unsqueeze];
+"1852 add_60" [id=1852, type=add];
+"1853 view_97" [id=1853, type=view];
+"1854 softmax_17" [id=1854, type=softmax];
+"1855 dropout_68" [id=1855, type=dropout];
+"1856 matmul_35" [id=1856, type=matmul];
+"1857 transpose_35" [id=1857, type=transpose];
+"1858 reshape_79" [id=1858, type=reshape];
+"1859 reshape_79_0_0_nncf_smooth_quant_0" [id=1859, type=call_module];
+"1860 quantize_per_tensor_default_108" [id=1860, type=quantize_per_tensor];
+"1861 dequantize_per_tensor_default_108" [id=1861, type=dequantize_per_tensor];
+"1862 linear_107_scale_0" [id=1862, type=get_attr];
+"1863 linear_107_zero_point_0" [id=1863, type=get_attr];
+"1864 compressed_weight_updated_constant72" [id=1864, type=get_attr];
+"1865 quantize_per_channel_default_90" [id=1865, type=quantize_per_channel];
+"1866 dequantize_per_channel_default_90" [id=1866, type=dequantize_per_channel];
+"1867 features_5_13_attn_proj_bias_0_0" [id=1867, type=get_attr];
+"1868 linear_107" [id=1868, type=linear];
+"1869 dropout_69" [id=1869, type=dropout];
+"1870 view_98" [id=1870, type=view];
+"1871 permute_81" [id=1871, type=permute];
+"1872 reshape_80" [id=1872, type=reshape];
+"1873 roll_17" [id=1873, type=roll];
+"1874 slice_269" [id=1874, type=slice];
+"1875 slice_270" [id=1875, type=slice];
+"1876 slice_271" [id=1876, type=slice];
+"1877 slice_272" [id=1877, type=slice];
+"1878 contiguous_33" [id=1878, type=contiguous];
+"1879 layer_norm_37" [id=1879, type=layer_norm];
+"1880 add_61" [id=1880, type=add];
+"1881 add_61_0_0_nncf_smooth_quant_0" [id=1881, type=call_module];
+"1882 quantize_per_tensor_default_109" [id=1882, type=quantize_per_tensor];
+"1883 dequantize_per_tensor_default_109" [id=1883, type=dequantize_per_tensor];
+"1884 linear_108_scale_0" [id=1884, type=get_attr];
+"1885 linear_108_zero_point_0" [id=1885, type=get_attr];
+"1886 compressed_weight_updated_constant73" [id=1886, type=get_attr];
+"1887 quantize_per_channel_default_91" [id=1887, type=quantize_per_channel];
+"1888 dequantize_per_channel_default_91" [id=1888, type=dequantize_per_channel];
+"1889 features_5_13_mlp_0_bias_0_0" [id=1889, type=get_attr];
+"1890 linear_108" [id=1890, type=linear];
+"1891 gelu_17" [id=1891, type=gelu];
+"1892 dropout_70" [id=1892, type=dropout];
+"1893 dropout_70_0_0_nncf_smooth_quant_0" [id=1893, type=call_module];
+"1894 quantize_per_tensor_default_110" [id=1894, type=quantize_per_tensor];
+"1895 dequantize_per_tensor_default_110" [id=1895, type=dequantize_per_tensor];
+"1896 linear_109_scale_0" [id=1896, type=get_attr];
+"1897 linear_109_zero_point_0" [id=1897, type=get_attr];
+"1898 compressed_weight_updated_constant74" [id=1898, type=get_attr];
+"1899 quantize_per_channel_default_92" [id=1899, type=quantize_per_channel];
+"1900 dequantize_per_channel_default_92" [id=1900, type=dequantize_per_channel];
+"1901 features_5_13_mlp_3_bias_0_0" [id=1901, type=get_attr];
+"1902 linear_109" [id=1902, type=linear];
+"1903 dropout_71" [id=1903, type=dropout];
+"1904 layer_norm_38" [id=1904, type=layer_norm];
+"1905 add_62" [id=1905, type=add];
+"1906 _frozen_param153" [id=1906, type=get_attr];
+"1907 pad_20" [id=1907, type=pad];
+"1908 view_101" [id=1908, type=view];
+"1909 permute_83" [id=1909, type=permute];
+"1910 reshape_81" [id=1910, type=reshape];
+"1911 reshape_81_0_0_nncf_smooth_quant_0" [id=1911, type=call_module];
+"1912 quantize_per_tensor_default_111" [id=1912, type=quantize_per_tensor];
+"1913 dequantize_per_tensor_default_111" [id=1913, type=dequantize_per_tensor];
+"1914 linear_112_scale_0" [id=1914, type=get_attr];
+"1915 linear_112_zero_point_0" [id=1915, type=get_attr];
+"1916 compressed_weight_updated_constant75" [id=1916, type=get_attr];
+"1917 quantize_per_channel_default_94" [id=1917, type=quantize_per_channel];
+"1918 dequantize_per_channel_default_94" [id=1918, type=dequantize_per_channel];
+"1919 _frozen_param154_0_0" [id=1919, type=get_attr];
+"1920 linear_112" [id=1920, type=linear];
+"1921 reshape_82" [id=1921, type=reshape];
+"1922 permute_84" [id=1922, type=permute];
+"1923 select_54" [id=1923, type=select];
+"1924 select_55" [id=1924, type=select];
+"1925 select_56" [id=1925, type=select];
+"1926 linalg_vector_norm_36" [id=1926, type=linalg_vector_norm];
+"1927 clamp_min_36" [id=1927, type=clamp_min];
+"1928 expand_as_36" [id=1928, type=expand_as];
+"1929 div_36" [id=1929, type=div];
+"1930 quantize_per_tensor_default_112" [id=1930, type=quantize_per_tensor];
+"1931 dequantize_per_tensor_default_112" [id=1931, type=dequantize_per_tensor];
+"1932 linalg_vector_norm_37" [id=1932, type=linalg_vector_norm];
+"1933 clamp_min_37" [id=1933, type=clamp_min];
+"1934 expand_as_37" [id=1934, type=expand_as];
+"1935 div_37" [id=1935, type=div];
+"1936 quantize_per_tensor_default_113" [id=1936, type=quantize_per_tensor];
+"1937 dequantize_per_tensor_default_113" [id=1937, type=dequantize_per_tensor];
+"1938 transpose_36" [id=1938, type=transpose];
+"1939 matmul_36" [id=1939, type=matmul];
+"1940 _frozen_param156" [id=1940, type=get_attr];
+"1941 mul_37" [id=1941, type=mul];
+"1942 add_63" [id=1942, type=add];
+"1943 softmax_18" [id=1943, type=softmax];
+"1944 dropout_72" [id=1944, type=dropout];
+"1945 matmul_37" [id=1945, type=matmul];
+"1946 transpose_37" [id=1946, type=transpose];
+"1947 reshape_83" [id=1947, type=reshape];
+"1948 reshape_83_0_0_nncf_smooth_quant_0" [id=1948, type=call_module];
+"1949 quantize_per_tensor_default_114" [id=1949, type=quantize_per_tensor];
+"1950 dequantize_per_tensor_default_114" [id=1950, type=dequantize_per_tensor];
+"1951 linear_113_scale_0" [id=1951, type=get_attr];
+"1952 linear_113_zero_point_0" [id=1952, type=get_attr];
+"1953 compressed_weight_updated_constant76" [id=1953, type=get_attr];
+"1954 quantize_per_channel_default_95" [id=1954, type=quantize_per_channel];
+"1955 dequantize_per_channel_default_95" [id=1955, type=dequantize_per_channel];
+"1956 features_5_14_attn_proj_bias_0_0" [id=1956, type=get_attr];
+"1957 linear_113" [id=1957, type=linear];
+"1958 dropout_73" [id=1958, type=dropout];
+"1959 view_102" [id=1959, type=view];
+"1960 permute_85" [id=1960, type=permute];
+"1961 reshape_84" [id=1961, type=reshape];
+"1962 slice_274" [id=1962, type=slice];
+"1963 slice_275" [id=1963, type=slice];
+"1964 slice_276" [id=1964, type=slice];
+"1965 slice_277" [id=1965, type=slice];
+"1966 contiguous_35" [id=1966, type=contiguous];
+"1967 layer_norm_39" [id=1967, type=layer_norm];
+"1968 add_64" [id=1968, type=add];
+"1969 add_64_0_0_nncf_smooth_quant_0" [id=1969, type=call_module];
+"1970 quantize_per_tensor_default_115" [id=1970, type=quantize_per_tensor];
+"1971 dequantize_per_tensor_default_115" [id=1971, type=dequantize_per_tensor];
+"1972 linear_114_scale_0" [id=1972, type=get_attr];
+"1973 linear_114_zero_point_0" [id=1973, type=get_attr];
+"1974 compressed_weight_updated_constant77" [id=1974, type=get_attr];
+"1975 quantize_per_channel_default_96" [id=1975, type=quantize_per_channel];
+"1976 dequantize_per_channel_default_96" [id=1976, type=dequantize_per_channel];
+"1977 features_5_14_mlp_0_bias_0_0" [id=1977, type=get_attr];
+"1978 linear_114" [id=1978, type=linear];
+"1979 gelu_18" [id=1979, type=gelu];
+"1980 dropout_74" [id=1980, type=dropout];
+"1981 dropout_74_0_0_nncf_smooth_quant_0" [id=1981, type=call_module];
+"1982 quantize_per_tensor_default_116" [id=1982, type=quantize_per_tensor];
+"1983 dequantize_per_tensor_default_116" [id=1983, type=dequantize_per_tensor];
+"1984 linear_115_scale_0" [id=1984, type=get_attr];
+"1985 linear_115_zero_point_0" [id=1985, type=get_attr];
+"1986 compressed_weight_updated_constant78" [id=1986, type=get_attr];
+"1987 quantize_per_channel_default_97" [id=1987, type=quantize_per_channel];
+"1988 dequantize_per_channel_default_97" [id=1988, type=dequantize_per_channel];
+"1989 features_5_14_mlp_3_bias_0_0" [id=1989, type=get_attr];
+"1990 linear_115" [id=1990, type=linear];
+"1991 dropout_75" [id=1991, type=dropout];
+"1992 layer_norm_40" [id=1992, type=layer_norm];
+"1993 add_65" [id=1993, type=add];
+"1994 _frozen_param157" [id=1994, type=get_attr];
+"1995 pad_21" [id=1995, type=pad];
+"1996 roll_18" [id=1996, type=roll];
+"1997 view_105" [id=1997, type=view];
+"1998 permute_87" [id=1998, type=permute];
+"1999 reshape_85" [id=1999, type=reshape];
+"2000 reshape_85_0_0_nncf_smooth_quant_0" [id=2000, type=call_module];
+"2001 quantize_per_tensor_default_117" [id=2001, type=quantize_per_tensor];
+"2002 dequantize_per_tensor_default_117" [id=2002, type=dequantize_per_tensor];
+"2003 linear_118_scale_0" [id=2003, type=get_attr];
+"2004 linear_118_zero_point_0" [id=2004, type=get_attr];
+"2005 compressed_weight_updated_constant79" [id=2005, type=get_attr];
+"2006 quantize_per_channel_default_99" [id=2006, type=quantize_per_channel];
+"2007 dequantize_per_channel_default_99" [id=2007, type=dequantize_per_channel];
+"2008 _frozen_param158_0_0" [id=2008, type=get_attr];
+"2009 linear_118" [id=2009, type=linear];
+"2010 reshape_86" [id=2010, type=reshape];
+"2011 permute_88" [id=2011, type=permute];
+"2012 select_57" [id=2012, type=select];
+"2013 select_58" [id=2013, type=select];
+"2014 select_59" [id=2014, type=select];
+"2015 linalg_vector_norm_38" [id=2015, type=linalg_vector_norm];
+"2016 clamp_min_38" [id=2016, type=clamp_min];
+"2017 expand_as_38" [id=2017, type=expand_as];
+"2018 div_38" [id=2018, type=div];
+"2019 quantize_per_tensor_default_118" [id=2019, type=quantize_per_tensor];
+"2020 dequantize_per_tensor_default_118" [id=2020, type=dequantize_per_tensor];
+"2021 linalg_vector_norm_39" [id=2021, type=linalg_vector_norm];
+"2022 clamp_min_39" [id=2022, type=clamp_min];
+"2023 expand_as_39" [id=2023, type=expand_as];
+"2024 div_39" [id=2024, type=div];
+"2025 quantize_per_tensor_default_119" [id=2025, type=quantize_per_tensor];
+"2026 dequantize_per_tensor_default_119" [id=2026, type=dequantize_per_tensor];
+"2027 transpose_38" [id=2027, type=transpose];
+"2028 matmul_38" [id=2028, type=matmul];
+"2029 _frozen_param160" [id=2029, type=get_attr];
+"2030 mul_39" [id=2030, type=mul];
+"2031 add_66" [id=2031, type=add];
+"2032 new_zeros_9" [id=2032, type=new_zeros];
+"2033 view_106" [id=2033, type=view];
+"2034 permute_89" [id=2034, type=permute];
+"2035 reshape_87" [id=2035, type=reshape];
+"2036 unsqueeze_56" [id=2036, type=unsqueeze];
+"2037 unsqueeze_57" [id=2037, type=unsqueeze];
+"2038 sub_9" [id=2038, type=sub];
+"2039 ne_9" [id=2039, type=ne];
+"2040 masked_fill_18" [id=2040, type=masked_fill];
+"2041 eq_9" [id=2041, type=eq];
+"2042 masked_fill_19" [id=2042, type=masked_fill];
+"2043 view_107" [id=2043, type=view];
+"2044 unsqueeze_58" [id=2044, type=unsqueeze];
+"2045 unsqueeze_59" [id=2045, type=unsqueeze];
+"2046 add_67" [id=2046, type=add];
+"2047 view_108" [id=2047, type=view];
+"2048 softmax_19" [id=2048, type=softmax];
+"2049 dropout_76" [id=2049, type=dropout];
+"2050 matmul_39" [id=2050, type=matmul];
+"2051 transpose_39" [id=2051, type=transpose];
+"2052 reshape_88" [id=2052, type=reshape];
+"2053 reshape_88_0_0_nncf_smooth_quant_0" [id=2053, type=call_module];
+"2054 quantize_per_tensor_default_120" [id=2054, type=quantize_per_tensor];
+"2055 dequantize_per_tensor_default_120" [id=2055, type=dequantize_per_tensor];
+"2056 linear_119_scale_0" [id=2056, type=get_attr];
+"2057 linear_119_zero_point_0" [id=2057, type=get_attr];
+"2058 compressed_weight_updated_constant80" [id=2058, type=get_attr];
+"2059 quantize_per_channel_default_100" [id=2059, type=quantize_per_channel];
+"2060 dequantize_per_channel_default_100" [id=2060, type=dequantize_per_channel];
+"2061 features_5_15_attn_proj_bias_0_0" [id=2061, type=get_attr];
+"2062 linear_119" [id=2062, type=linear];
+"2063 dropout_77" [id=2063, type=dropout];
+"2064 view_109" [id=2064, type=view];
+"2065 permute_90" [id=2065, type=permute];
+"2066 reshape_89" [id=2066, type=reshape];
+"2067 roll_19" [id=2067, type=roll];
+"2068 slice_297" [id=2068, type=slice];
+"2069 slice_298" [id=2069, type=slice];
+"2070 slice_299" [id=2070, type=slice];
+"2071 slice_300" [id=2071, type=slice];
+"2072 contiguous_37" [id=2072, type=contiguous];
+"2073 layer_norm_41" [id=2073, type=layer_norm];
+"2074 add_68" [id=2074, type=add];
+"2075 add_68_0_0_nncf_smooth_quant_0" [id=2075, type=call_module];
+"2076 quantize_per_tensor_default_121" [id=2076, type=quantize_per_tensor];
+"2077 dequantize_per_tensor_default_121" [id=2077, type=dequantize_per_tensor];
+"2078 linear_120_scale_0" [id=2078, type=get_attr];
+"2079 linear_120_zero_point_0" [id=2079, type=get_attr];
+"2080 compressed_weight_updated_constant81" [id=2080, type=get_attr];
+"2081 quantize_per_channel_default_101" [id=2081, type=quantize_per_channel];
+"2082 dequantize_per_channel_default_101" [id=2082, type=dequantize_per_channel];
+"2083 features_5_15_mlp_0_bias_0_0" [id=2083, type=get_attr];
+"2084 linear_120" [id=2084, type=linear];
+"2085 gelu_19" [id=2085, type=gelu];
+"2086 dropout_78" [id=2086, type=dropout];
+"2087 dropout_78_0_0_nncf_smooth_quant_0" [id=2087, type=call_module];
+"2088 quantize_per_tensor_default_122" [id=2088, type=quantize_per_tensor];
+"2089 dequantize_per_tensor_default_122" [id=2089, type=dequantize_per_tensor];
+"2090 linear_121_scale_0" [id=2090, type=get_attr];
+"2091 linear_121_zero_point_0" [id=2091, type=get_attr];
+"2092 compressed_weight_updated_constant82" [id=2092, type=get_attr];
+"2093 quantize_per_channel_default_102" [id=2093, type=quantize_per_channel];
+"2094 dequantize_per_channel_default_102" [id=2094, type=dequantize_per_channel];
+"2095 features_5_15_mlp_3_bias_0_0" [id=2095, type=get_attr];
+"2096 linear_121" [id=2096, type=linear];
+"2097 dropout_79" [id=2097, type=dropout];
+"2098 layer_norm_42" [id=2098, type=layer_norm];
+"2099 add_69" [id=2099, type=add];
+"2100 _frozen_param170" [id=2100, type=get_attr];
+"2101 pad_22" [id=2101, type=pad];
+"2102 view_112" [id=2102, type=view];
+"2103 permute_92" [id=2103, type=permute];
+"2104 reshape_90" [id=2104, type=reshape];
+"2105 reshape_90_0_0_nncf_smooth_quant_0" [id=2105, type=call_module];
+"2106 quantize_per_tensor_default_123" [id=2106, type=quantize_per_tensor];
+"2107 dequantize_per_tensor_default_123" [id=2107, type=dequantize_per_tensor];
+"2108 linear_124_scale_0" [id=2108, type=get_attr];
+"2109 linear_124_zero_point_0" [id=2109, type=get_attr];
+"2110 compressed_weight_updated_constant83" [id=2110, type=get_attr];
+"2111 quantize_per_channel_default_104" [id=2111, type=quantize_per_channel];
+"2112 dequantize_per_channel_default_104" [id=2112, type=dequantize_per_channel];
+"2113 _frozen_param171_0_0" [id=2113, type=get_attr];
+"2114 linear_124" [id=2114, type=linear];
+"2115 reshape_91" [id=2115, type=reshape];
+"2116 permute_93" [id=2116, type=permute];
+"2117 select_60" [id=2117, type=select];
+"2118 select_61" [id=2118, type=select];
+"2119 select_62" [id=2119, type=select];
+"2120 linalg_vector_norm_40" [id=2120, type=linalg_vector_norm];
+"2121 clamp_min_40" [id=2121, type=clamp_min];
+"2122 expand_as_40" [id=2122, type=expand_as];
+"2123 div_40" [id=2123, type=div];
+"2124 quantize_per_tensor_default_124" [id=2124, type=quantize_per_tensor];
+"2125 dequantize_per_tensor_default_124" [id=2125, type=dequantize_per_tensor];
+"2126 linalg_vector_norm_41" [id=2126, type=linalg_vector_norm];
+"2127 clamp_min_41" [id=2127, type=clamp_min];
+"2128 expand_as_41" [id=2128, type=expand_as];
+"2129 div_41" [id=2129, type=div];
+"2130 quantize_per_tensor_default_125" [id=2130, type=quantize_per_tensor];
+"2131 dequantize_per_tensor_default_125" [id=2131, type=dequantize_per_tensor];
+"2132 transpose_40" [id=2132, type=transpose];
+"2133 matmul_40" [id=2133, type=matmul];
+"2134 _frozen_param173" [id=2134, type=get_attr];
+"2135 mul_41" [id=2135, type=mul];
+"2136 add_70" [id=2136, type=add];
+"2137 softmax_20" [id=2137, type=softmax];
+"2138 dropout_80" [id=2138, type=dropout];
+"2139 matmul_41" [id=2139, type=matmul];
+"2140 transpose_41" [id=2140, type=transpose];
+"2141 reshape_92" [id=2141, type=reshape];
+"2142 reshape_92_0_0_nncf_smooth_quant_0" [id=2142, type=call_module];
+"2143 quantize_per_tensor_default_126" [id=2143, type=quantize_per_tensor];
+"2144 dequantize_per_tensor_default_126" [id=2144, type=dequantize_per_tensor];
+"2145 linear_125_scale_0" [id=2145, type=get_attr];
+"2146 linear_125_zero_point_0" [id=2146, type=get_attr];
+"2147 compressed_weight_updated_constant84" [id=2147, type=get_attr];
+"2148 quantize_per_channel_default_105" [id=2148, type=quantize_per_channel];
+"2149 dequantize_per_channel_default_105" [id=2149, type=dequantize_per_channel];
+"2150 features_5_16_attn_proj_bias_0_0" [id=2150, type=get_attr];
+"2151 linear_125" [id=2151, type=linear];
+"2152 dropout_81" [id=2152, type=dropout];
+"2153 view_113" [id=2153, type=view];
+"2154 permute_94" [id=2154, type=permute];
+"2155 reshape_93" [id=2155, type=reshape];
+"2156 slice_302" [id=2156, type=slice];
+"2157 slice_303" [id=2157, type=slice];
+"2158 slice_304" [id=2158, type=slice];
+"2159 slice_305" [id=2159, type=slice];
+"2160 contiguous_39" [id=2160, type=contiguous];
+"2161 layer_norm_43" [id=2161, type=layer_norm];
+"2162 add_71" [id=2162, type=add];
+"2163 add_71_0_0_nncf_smooth_quant_0" [id=2163, type=call_module];
+"2164 quantize_per_tensor_default_127" [id=2164, type=quantize_per_tensor];
+"2165 dequantize_per_tensor_default_127" [id=2165, type=dequantize_per_tensor];
+"2166 linear_126_scale_0" [id=2166, type=get_attr];
+"2167 linear_126_zero_point_0" [id=2167, type=get_attr];
+"2168 compressed_weight_updated_constant85" [id=2168, type=get_attr];
+"2169 quantize_per_channel_default_106" [id=2169, type=quantize_per_channel];
+"2170 dequantize_per_channel_default_106" [id=2170, type=dequantize_per_channel];
+"2171 features_5_16_mlp_0_bias_0_0" [id=2171, type=get_attr];
+"2172 linear_126" [id=2172, type=linear];
+"2173 gelu_20" [id=2173, type=gelu];
+"2174 dropout_82" [id=2174, type=dropout];
+"2175 dropout_82_0_0_nncf_smooth_quant_0" [id=2175, type=call_module];
+"2176 quantize_per_tensor_default_128" [id=2176, type=quantize_per_tensor];
+"2177 dequantize_per_tensor_default_128" [id=2177, type=dequantize_per_tensor];
+"2178 linear_127_scale_0" [id=2178, type=get_attr];
+"2179 linear_127_zero_point_0" [id=2179, type=get_attr];
+"2180 compressed_weight_updated_constant86" [id=2180, type=get_attr];
+"2181 quantize_per_channel_default_107" [id=2181, type=quantize_per_channel];
+"2182 dequantize_per_channel_default_107" [id=2182, type=dequantize_per_channel];
+"2183 features_5_16_mlp_3_bias_0_0" [id=2183, type=get_attr];
+"2184 linear_127" [id=2184, type=linear];
+"2185 dropout_83" [id=2185, type=dropout];
+"2186 layer_norm_44" [id=2186, type=layer_norm];
+"2187 add_72" [id=2187, type=add];
+"2188 _frozen_param174" [id=2188, type=get_attr];
+"2189 pad_23" [id=2189, type=pad];
+"2190 roll_20" [id=2190, type=roll];
+"2191 view_116" [id=2191, type=view];
+"2192 permute_96" [id=2192, type=permute];
+"2193 reshape_94" [id=2193, type=reshape];
+"2194 reshape_94_0_0_nncf_smooth_quant_0" [id=2194, type=call_module];
+"2195 quantize_per_tensor_default_129" [id=2195, type=quantize_per_tensor];
+"2196 dequantize_per_tensor_default_129" [id=2196, type=dequantize_per_tensor];
+"2197 linear_130_scale_0" [id=2197, type=get_attr];
+"2198 linear_130_zero_point_0" [id=2198, type=get_attr];
+"2199 compressed_weight_updated_constant87" [id=2199, type=get_attr];
+"2200 quantize_per_channel_default_109" [id=2200, type=quantize_per_channel];
+"2201 dequantize_per_channel_default_109" [id=2201, type=dequantize_per_channel];
+"2202 _frozen_param175_0_0" [id=2202, type=get_attr];
+"2203 linear_130" [id=2203, type=linear];
+"2204 reshape_95" [id=2204, type=reshape];
+"2205 permute_97" [id=2205, type=permute];
+"2206 select_63" [id=2206, type=select];
+"2207 select_64" [id=2207, type=select];
+"2208 select_65" [id=2208, type=select];
+"2209 linalg_vector_norm_42" [id=2209, type=linalg_vector_norm];
+"2210 clamp_min_42" [id=2210, type=clamp_min];
+"2211 expand_as_42" [id=2211, type=expand_as];
+"2212 div_42" [id=2212, type=div];
+"2213 quantize_per_tensor_default_130" [id=2213, type=quantize_per_tensor];
+"2214 dequantize_per_tensor_default_130" [id=2214, type=dequantize_per_tensor];
+"2215 linalg_vector_norm_43" [id=2215, type=linalg_vector_norm];
+"2216 clamp_min_43" [id=2216, type=clamp_min];
+"2217 expand_as_43" [id=2217, type=expand_as];
+"2218 div_43" [id=2218, type=div];
+"2219 quantize_per_tensor_default_131" [id=2219, type=quantize_per_tensor];
+"2220 dequantize_per_tensor_default_131" [id=2220, type=dequantize_per_tensor];
+"2221 transpose_42" [id=2221, type=transpose];
+"2222 matmul_42" [id=2222, type=matmul];
+"2223 _frozen_param177" [id=2223, type=get_attr];
+"2224 mul_43" [id=2224, type=mul];
+"2225 add_73" [id=2225, type=add];
+"2226 new_zeros_10" [id=2226, type=new_zeros];
+"2227 view_117" [id=2227, type=view];
+"2228 permute_98" [id=2228, type=permute];
+"2229 reshape_96" [id=2229, type=reshape];
+"2230 unsqueeze_62" [id=2230, type=unsqueeze];
+"2231 unsqueeze_63" [id=2231, type=unsqueeze];
+"2232 sub_10" [id=2232, type=sub];
+"2233 ne_10" [id=2233, type=ne];
+"2234 masked_fill_20" [id=2234, type=masked_fill];
+"2235 eq_10" [id=2235, type=eq];
+"2236 masked_fill_21" [id=2236, type=masked_fill];
+"2237 view_118" [id=2237, type=view];
+"2238 unsqueeze_64" [id=2238, type=unsqueeze];
+"2239 unsqueeze_65" [id=2239, type=unsqueeze];
+"2240 add_74" [id=2240, type=add];
+"2241 view_119" [id=2241, type=view];
+"2242 softmax_21" [id=2242, type=softmax];
+"2243 dropout_84" [id=2243, type=dropout];
+"2244 matmul_43" [id=2244, type=matmul];
+"2245 transpose_43" [id=2245, type=transpose];
+"2246 reshape_97" [id=2246, type=reshape];
+"2247 reshape_97_0_0_nncf_smooth_quant_0" [id=2247, type=call_module];
+"2248 quantize_per_tensor_default_132" [id=2248, type=quantize_per_tensor];
+"2249 dequantize_per_tensor_default_132" [id=2249, type=dequantize_per_tensor];
+"2250 linear_131_scale_0" [id=2250, type=get_attr];
+"2251 linear_131_zero_point_0" [id=2251, type=get_attr];
+"2252 compressed_weight_updated_constant88" [id=2252, type=get_attr];
+"2253 quantize_per_channel_default_110" [id=2253, type=quantize_per_channel];
+"2254 dequantize_per_channel_default_110" [id=2254, type=dequantize_per_channel];
+"2255 features_5_17_attn_proj_bias_0_0" [id=2255, type=get_attr];
+"2256 linear_131" [id=2256, type=linear];
+"2257 dropout_85" [id=2257, type=dropout];
+"2258 view_120" [id=2258, type=view];
+"2259 permute_99" [id=2259, type=permute];
+"2260 reshape_98" [id=2260, type=reshape];
+"2261 roll_21" [id=2261, type=roll];
+"2262 slice_325" [id=2262, type=slice];
+"2263 slice_326" [id=2263, type=slice];
+"2264 slice_327" [id=2264, type=slice];
+"2265 slice_328" [id=2265, type=slice];
+"2266 contiguous_41" [id=2266, type=contiguous];
+"2267 layer_norm_45" [id=2267, type=layer_norm];
+"2268 add_75" [id=2268, type=add];
+"2269 add_75_0_0_nncf_smooth_quant_0" [id=2269, type=call_module];
+"2270 quantize_per_tensor_default_133" [id=2270, type=quantize_per_tensor];
+"2271 dequantize_per_tensor_default_133" [id=2271, type=dequantize_per_tensor];
+"2272 linear_132_scale_0" [id=2272, type=get_attr];
+"2273 linear_132_zero_point_0" [id=2273, type=get_attr];
+"2274 compressed_weight_updated_constant89" [id=2274, type=get_attr];
+"2275 quantize_per_channel_default_111" [id=2275, type=quantize_per_channel];
+"2276 dequantize_per_channel_default_111" [id=2276, type=dequantize_per_channel];
+"2277 features_5_17_mlp_0_bias_0_0" [id=2277, type=get_attr];
+"2278 linear_132" [id=2278, type=linear];
+"2279 gelu_21" [id=2279, type=gelu];
+"2280 dropout_86" [id=2280, type=dropout];
+"2281 dropout_86_0_0_nncf_smooth_quant_0" [id=2281, type=call_module];
+"2282 quantize_per_tensor_default_134" [id=2282, type=quantize_per_tensor];
+"2283 dequantize_per_tensor_default_134" [id=2283, type=dequantize_per_tensor];
+"2284 linear_133_scale_0" [id=2284, type=get_attr];
+"2285 linear_133_zero_point_0" [id=2285, type=get_attr];
+"2286 compressed_weight_updated_constant90" [id=2286, type=get_attr];
+"2287 quantize_per_channel_default_112" [id=2287, type=quantize_per_channel];
+"2288 dequantize_per_channel_default_112" [id=2288, type=dequantize_per_channel];
+"2289 features_5_17_mlp_3_bias_0_0" [id=2289, type=get_attr];
+"2290 linear_133" [id=2290, type=linear];
+"2291 dropout_87" [id=2291, type=dropout];
+"2292 layer_norm_46" [id=2292, type=layer_norm];
+"2293 add_76" [id=2293, type=add];
+"2294 pad_24" [id=2294, type=pad];
+"2295 slice_329" [id=2295, type=slice];
+"2296 slice_330" [id=2296, type=slice];
+"2297 slice_331" [id=2297, type=slice];
+"2298 slice_332" [id=2298, type=slice];
+"2299 slice_333" [id=2299, type=slice];
+"2300 slice_334" [id=2300, type=slice];
+"2301 slice_335" [id=2301, type=slice];
+"2302 slice_336" [id=2302, type=slice];
+"2303 slice_337" [id=2303, type=slice];
+"2304 slice_338" [id=2304, type=slice];
+"2305 slice_339" [id=2305, type=slice];
+"2306 slice_340" [id=2306, type=slice];
+"2307 cat_2" [id=2307, type=cat];
+"2308 cat_2_0_0_nncf_smooth_quant_0" [id=2308, type=call_module];
+"2309 quantize_per_tensor_default_135" [id=2309, type=quantize_per_tensor];
+"2310 dequantize_per_tensor_default_135" [id=2310, type=dequantize_per_tensor];
+"2311 linear_134_scale_0" [id=2311, type=get_attr];
+"2312 linear_134_zero_point_0" [id=2312, type=get_attr];
+"2313 compressed_weight_updated_constant91" [id=2313, type=get_attr];
+"2314 quantize_per_channel_default_113" [id=2314, type=quantize_per_channel];
+"2315 dequantize_per_channel_default_113" [id=2315, type=dequantize_per_channel];
+"2316 linear_134" [id=2316, type=linear];
+"2317 layer_norm_47" [id=2317, type=layer_norm];
+"2318 _frozen_param187" [id=2318, type=get_attr];
+"2319 pad_25" [id=2319, type=pad];
+"2320 view_123" [id=2320, type=view];
+"2321 permute_101" [id=2321, type=permute];
+"2322 reshape_99" [id=2322, type=reshape];
+"2323 reshape_99_0_0_nncf_smooth_quant_0" [id=2323, type=call_module];
+"2324 quantize_per_tensor_default_136" [id=2324, type=quantize_per_tensor];
+"2325 dequantize_per_tensor_default_136" [id=2325, type=dequantize_per_tensor];
+"2326 linear_137_scale_0" [id=2326, type=get_attr];
+"2327 linear_137_zero_point_0" [id=2327, type=get_attr];
+"2328 compressed_weight_updated_constant92" [id=2328, type=get_attr];
+"2329 quantize_per_channel_default_115" [id=2329, type=quantize_per_channel];
+"2330 dequantize_per_channel_default_115" [id=2330, type=dequantize_per_channel];
+"2331 _frozen_param188_0_0" [id=2331, type=get_attr];
+"2332 linear_137" [id=2332, type=linear];
+"2333 reshape_100" [id=2333, type=reshape];
+"2334 permute_102" [id=2334, type=permute];
+"2335 select_66" [id=2335, type=select];
+"2336 select_67" [id=2336, type=select];
+"2337 select_68" [id=2337, type=select];
+"2338 linalg_vector_norm_44" [id=2338, type=linalg_vector_norm];
+"2339 clamp_min_44" [id=2339, type=clamp_min];
+"2340 expand_as_44" [id=2340, type=expand_as];
+"2341 div_44" [id=2341, type=div];
+"2342 quantize_per_tensor_default_137" [id=2342, type=quantize_per_tensor];
+"2343 dequantize_per_tensor_default_137" [id=2343, type=dequantize_per_tensor];
+"2344 linalg_vector_norm_45" [id=2344, type=linalg_vector_norm];
+"2345 clamp_min_45" [id=2345, type=clamp_min];
+"2346 expand_as_45" [id=2346, type=expand_as];
+"2347 div_45" [id=2347, type=div];
+"2348 quantize_per_tensor_default_138" [id=2348, type=quantize_per_tensor];
+"2349 dequantize_per_tensor_default_138" [id=2349, type=dequantize_per_tensor];
+"2350 transpose_44" [id=2350, type=transpose];
+"2351 matmul_44" [id=2351, type=matmul];
+"2352 _frozen_param190" [id=2352, type=get_attr];
+"2353 mul_45" [id=2353, type=mul];
+"2354 add_77" [id=2354, type=add];
+"2355 softmax_22" [id=2355, type=softmax];
+"2356 dropout_88" [id=2356, type=dropout];
+"2357 matmul_45" [id=2357, type=matmul];
+"2358 transpose_45" [id=2358, type=transpose];
+"2359 reshape_101" [id=2359, type=reshape];
+"2360 reshape_101_0_0_nncf_smooth_quant_0" [id=2360, type=call_module];
+"2361 quantize_per_tensor_default_139" [id=2361, type=quantize_per_tensor];
+"2362 dequantize_per_tensor_default_139" [id=2362, type=dequantize_per_tensor];
+"2363 linear_138_scale_0" [id=2363, type=get_attr];
+"2364 linear_138_zero_point_0" [id=2364, type=get_attr];
+"2365 compressed_weight_updated_constant93" [id=2365, type=get_attr];
+"2366 quantize_per_channel_default_116" [id=2366, type=quantize_per_channel];
+"2367 dequantize_per_channel_default_116" [id=2367, type=dequantize_per_channel];
+"2368 features_7_0_attn_proj_bias_0_0" [id=2368, type=get_attr];
+"2369 linear_138" [id=2369, type=linear];
+"2370 dropout_89" [id=2370, type=dropout];
+"2371 view_124" [id=2371, type=view];
+"2372 permute_103" [id=2372, type=permute];
+"2373 reshape_102" [id=2373, type=reshape];
+"2374 slice_342" [id=2374, type=slice];
+"2375 slice_343" [id=2375, type=slice];
+"2376 slice_344" [id=2376, type=slice];
+"2377 slice_345" [id=2377, type=slice];
+"2378 contiguous_43" [id=2378, type=contiguous];
+"2379 layer_norm_48" [id=2379, type=layer_norm];
+"2380 add_78" [id=2380, type=add];
+"2381 add_78_0_0_nncf_smooth_quant_0" [id=2381, type=call_module];
+"2382 quantize_per_tensor_default_140" [id=2382, type=quantize_per_tensor];
+"2383 dequantize_per_tensor_default_140" [id=2383, type=dequantize_per_tensor];
+"2384 linear_139_scale_0" [id=2384, type=get_attr];
+"2385 linear_139_zero_point_0" [id=2385, type=get_attr];
+"2386 compressed_weight_updated_constant94" [id=2386, type=get_attr];
+"2387 quantize_per_channel_default_117" [id=2387, type=quantize_per_channel];
+"2388 dequantize_per_channel_default_117" [id=2388, type=dequantize_per_channel];
+"2389 features_7_0_mlp_0_bias_0_0" [id=2389, type=get_attr];
+"2390 linear_139" [id=2390, type=linear];
+"2391 gelu_22" [id=2391, type=gelu];
+"2392 dropout_90" [id=2392, type=dropout];
+"2393 dropout_90_0_0_nncf_smooth_quant_0" [id=2393, type=call_module];
+"2394 quantize_per_tensor_default_141" [id=2394, type=quantize_per_tensor];
+"2395 dequantize_per_tensor_default_141" [id=2395, type=dequantize_per_tensor];
+"2396 linear_140_scale_0" [id=2396, type=get_attr];
+"2397 linear_140_zero_point_0" [id=2397, type=get_attr];
+"2398 compressed_weight_updated_constant95" [id=2398, type=get_attr];
+"2399 quantize_per_channel_default_118" [id=2399, type=quantize_per_channel];
+"2400 dequantize_per_channel_default_118" [id=2400, type=dequantize_per_channel];
+"2401 features_7_0_mlp_3_bias_0_0" [id=2401, type=get_attr];
+"2402 linear_140" [id=2402, type=linear];
+"2403 dropout_91" [id=2403, type=dropout];
+"2404 layer_norm_49" [id=2404, type=layer_norm];
+"2405 add_79" [id=2405, type=add];
+"2406 _frozen_param191" [id=2406, type=get_attr];
+"2407 pad_26" [id=2407, type=pad];
+"2408 view_127" [id=2408, type=view];
+"2409 permute_105" [id=2409, type=permute];
+"2410 reshape_103" [id=2410, type=reshape];
+"2411 reshape_103_0_0_nncf_smooth_quant_0" [id=2411, type=call_module];
+"2412 quantize_per_tensor_default_142" [id=2412, type=quantize_per_tensor];
+"2413 dequantize_per_tensor_default_142" [id=2413, type=dequantize_per_tensor];
+"2414 linear_143_scale_0" [id=2414, type=get_attr];
+"2415 linear_143_zero_point_0" [id=2415, type=get_attr];
+"2416 compressed_weight_updated_constant96" [id=2416, type=get_attr];
+"2417 quantize_per_channel_default_120" [id=2417, type=quantize_per_channel];
+"2418 dequantize_per_channel_default_120" [id=2418, type=dequantize_per_channel];
+"2419 _frozen_param192_0_0" [id=2419, type=get_attr];
+"2420 linear_143" [id=2420, type=linear];
+"2421 reshape_104" [id=2421, type=reshape];
+"2422 permute_106" [id=2422, type=permute];
+"2423 select_69" [id=2423, type=select];
+"2424 select_70" [id=2424, type=select];
+"2425 select_71" [id=2425, type=select];
+"2426 linalg_vector_norm_46" [id=2426, type=linalg_vector_norm];
+"2427 clamp_min_46" [id=2427, type=clamp_min];
+"2428 expand_as_46" [id=2428, type=expand_as];
+"2429 div_46" [id=2429, type=div];
+"2430 quantize_per_tensor_default_143" [id=2430, type=quantize_per_tensor];
+"2431 dequantize_per_tensor_default_143" [id=2431, type=dequantize_per_tensor];
+"2432 linalg_vector_norm_47" [id=2432, type=linalg_vector_norm];
+"2433 clamp_min_47" [id=2433, type=clamp_min];
+"2434 expand_as_47" [id=2434, type=expand_as];
+"2435 div_47" [id=2435, type=div];
+"2436 quantize_per_tensor_default_144" [id=2436, type=quantize_per_tensor];
+"2437 dequantize_per_tensor_default_144" [id=2437, type=dequantize_per_tensor];
+"2438 transpose_46" [id=2438, type=transpose];
+"2439 matmul_46" [id=2439, type=matmul];
+"2440 _frozen_param194" [id=2440, type=get_attr];
+"2441 mul_47" [id=2441, type=mul];
+"2442 add_80" [id=2442, type=add];
+"2443 softmax_23" [id=2443, type=softmax];
+"2444 dropout_92" [id=2444, type=dropout];
+"2445 matmul_47" [id=2445, type=matmul];
+"2446 transpose_47" [id=2446, type=transpose];
+"2447 reshape_105" [id=2447, type=reshape];
+"2448 reshape_105_0_0_nncf_smooth_quant_0" [id=2448, type=call_module];
+"2449 quantize_per_tensor_default_145" [id=2449, type=quantize_per_tensor];
+"2450 dequantize_per_tensor_default_145" [id=2450, type=dequantize_per_tensor];
+"2451 linear_144_scale_0" [id=2451, type=get_attr];
+"2452 linear_144_zero_point_0" [id=2452, type=get_attr];
+"2453 compressed_weight_updated_constant97" [id=2453, type=get_attr];
+"2454 quantize_per_channel_default_121" [id=2454, type=quantize_per_channel];
+"2455 dequantize_per_channel_default_121" [id=2455, type=dequantize_per_channel];
+"2456 features_7_1_attn_proj_bias_0_0" [id=2456, type=get_attr];
+"2457 linear_144" [id=2457, type=linear];
+"2458 dropout_93" [id=2458, type=dropout];
+"2459 view_128" [id=2459, type=view];
+"2460 permute_107" [id=2460, type=permute];
+"2461 reshape_106" [id=2461, type=reshape];
+"2462 slice_347" [id=2462, type=slice];
+"2463 slice_348" [id=2463, type=slice];
+"2464 slice_349" [id=2464, type=slice];
+"2465 slice_350" [id=2465, type=slice];
+"2466 contiguous_45" [id=2466, type=contiguous];
+"2467 layer_norm_50" [id=2467, type=layer_norm];
+"2468 add_81" [id=2468, type=add];
+"2469 add_81_0_0_nncf_smooth_quant_0" [id=2469, type=call_module];
+"2470 quantize_per_tensor_default_146" [id=2470, type=quantize_per_tensor];
+"2471 dequantize_per_tensor_default_146" [id=2471, type=dequantize_per_tensor];
+"2472 linear_145_scale_0" [id=2472, type=get_attr];
+"2473 linear_145_zero_point_0" [id=2473, type=get_attr];
+"2474 compressed_weight_updated_constant98" [id=2474, type=get_attr];
+"2475 quantize_per_channel_default_122" [id=2475, type=quantize_per_channel];
+"2476 dequantize_per_channel_default_122" [id=2476, type=dequantize_per_channel];
+"2477 features_7_1_mlp_0_bias_0_0" [id=2477, type=get_attr];
+"2478 linear_145" [id=2478, type=linear];
+"2479 gelu_23" [id=2479, type=gelu];
+"2480 dropout_94" [id=2480, type=dropout];
+"2481 dropout_94_0_0_nncf_smooth_quant_0" [id=2481, type=call_module];
+"2482 quantize_per_tensor_default_147" [id=2482, type=quantize_per_tensor];
+"2483 dequantize_per_tensor_default_147" [id=2483, type=dequantize_per_tensor];
+"2484 linear_146_scale_0" [id=2484, type=get_attr];
+"2485 linear_146_zero_point_0" [id=2485, type=get_attr];
+"2486 compressed_weight_updated_constant99" [id=2486, type=get_attr];
+"2487 quantize_per_channel_default_123" [id=2487, type=quantize_per_channel];
+"2488 dequantize_per_channel_default_123" [id=2488, type=dequantize_per_channel];
+"2489 features_7_1_mlp_3_bias_0_0" [id=2489, type=get_attr];
+"2490 linear_146" [id=2490, type=linear];
+"2491 dropout_95" [id=2491, type=dropout];
+"2492 layer_norm_51" [id=2492, type=layer_norm];
+"2493 add_82" [id=2493, type=add];
 "2494 layer_norm_52" [id=2494, type=layer_norm];
 "2495 permute_108" [id=2495, type=permute];
 "2496 adaptive_avg_pool2d" [id=2496, type=adaptive_avg_pool2d];
@@ -2505,2948 +2505,2948 @@ strict digraph  {
 "2503 compressed_weight_updated_constant100" [id=2503, type=get_attr];
 "2504 quantize_per_channel_default_124" [id=2504, type=quantize_per_channel];
 "2505 dequantize_per_channel_default_124" [id=2505, type=dequantize_per_channel];
-"2506 _param_constant400_0_0" [id=2506, type=get_attr];
+"2506 head_bias_0_0" [id=2506, type=get_attr];
 "2507 linear_147" [id=2507, type=linear];
 "2508 output" [id=2508, type=output];
-"0 arg0_1" -> "1 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"1 quantize_per_tensor_default" -> "2 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"2 dequantize_per_tensor_default" -> "9 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"3 conv2d_scale_0" -> "6 quantize_per_channel_default"  [label="(96,)", style=solid];
-"3 conv2d_scale_0" -> "7 dequantize_per_channel_default"  [label="(96,)", style=solid];
-"4 conv2d_zero_point_0" -> "6 quantize_per_channel_default"  [label="(96,)", style=solid];
-"4 conv2d_zero_point_0" -> "7 dequantize_per_channel_default"  [label="(96,)", style=solid];
-"5 compressed_weight_updated_constant0" -> "6 quantize_per_channel_default"  [label="(96, 3, 4, 4)", style=solid];
-"6 quantize_per_channel_default" -> "7 dequantize_per_channel_default"  [label="(96, 3, 4, 4)", style=solid];
-"7 dequantize_per_channel_default" -> "9 conv2d"  [label="(96, 3, 4, 4)", style=solid];
-"8 _param_constant1_0_0" -> "9 conv2d"  [label="(96,)", style=solid];
-"9 conv2d" -> "10 permute"  [label="(1, 96, 56, 56)", style=solid];
-"10 permute" -> "13 layer_norm"  [label="(1, 56, 56, 96)", style=solid];
-"11 _param_constant2" -> "13 layer_norm"  [label="(96,)", style=solid];
-"12 _param_constant3" -> "13 layer_norm"  [label="(96,)", style=solid];
-"13 layer_norm" -> "15 pad"  [label="(1, 56, 56, 96)", style=solid];
-"13 layer_norm" -> "75 add_1"  [label="(1, 56, 56, 96)", style=solid];
-"14 _frozen_param0" -> "50 add"  [label="(1, 3, 64, 64)", style=solid];
-"15 pad" -> "16 view_2"  [label="(1, 56, 56, 96)", style=solid];
-"16 view_2" -> "17 permute_2"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"17 permute_2" -> "18 reshape"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"18 reshape" -> "19 reshape_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
-"19 reshape_0_0_nncf_smooth_quant_0" -> "20 quantize_per_tensor_default_1"  [label="(49, 64, 96)", style=solid];
-"20 quantize_per_tensor_default_1" -> "21 dequantize_per_tensor_default_1"  [label="(49, 64, 96)", style=solid];
-"21 dequantize_per_tensor_default_1" -> "28 linear_2"  [label="(49, 64, 96)", style=solid];
-"22 linear_2_scale_0" -> "25 quantize_per_channel_default_2"  [label="(288,)", style=solid];
-"22 linear_2_scale_0" -> "26 dequantize_per_channel_default_2"  [label="(288,)", style=solid];
-"23 linear_2_zero_point_0" -> "25 quantize_per_channel_default_2"  [label="(288,)", style=solid];
-"23 linear_2_zero_point_0" -> "26 dequantize_per_channel_default_2"  [label="(288,)", style=solid];
-"24 compressed_weight_updated_constant1" -> "25 quantize_per_channel_default_2"  [label="(288, 96)", style=solid];
-"25 quantize_per_channel_default_2" -> "26 dequantize_per_channel_default_2"  [label="(288, 96)", style=solid];
-"26 dequantize_per_channel_default_2" -> "28 linear_2"  [label="(288, 96)", style=solid];
-"27 _frozen_param1_0_0" -> "28 linear_2"  [label="(288,)", style=solid];
-"28 linear_2" -> "29 reshape_1"  [label="(49, 64, 288)", style=solid];
-"29 reshape_1" -> "30 permute_3"  [label="(49, 64, 3, 3, 32)", style=solid];
-"30 permute_3" -> "31 select"  [label="(3, 49, 3, 64, 32)", style=solid];
-"30 permute_3" -> "32 select_1"  [label="(3, 49, 3, 64, 32)", style=solid];
-"30 permute_3" -> "33 select_2"  [label="(3, 49, 3, 64, 32)", style=solid];
-"31 select" -> "34 linalg_vector_norm"  [label="(49, 3, 64, 32)", style=solid];
-"31 select" -> "36 expand_as"  [label="(49, 3, 64, 32)", style=solid];
-"31 select" -> "37 div"  [label="(49, 3, 64, 32)", style=solid];
-"32 select_1" -> "40 linalg_vector_norm_1"  [label="(49, 3, 64, 32)", style=solid];
-"32 select_1" -> "42 expand_as_1"  [label="(49, 3, 64, 32)", style=solid];
-"32 select_1" -> "43 div_1"  [label="(49, 3, 64, 32)", style=solid];
-"33 select_2" -> "53 matmul_1"  [label="(49, 3, 64, 32)", style=solid];
-"34 linalg_vector_norm" -> "35 clamp_min"  [label="(49, 3, 64, 1)", style=solid];
-"35 clamp_min" -> "36 expand_as"  [label="(49, 3, 64, 1)", style=solid];
-"36 expand_as" -> "37 div"  [label="(49, 3, 64, 32)", style=solid];
-"37 div" -> "38 quantize_per_tensor_default_2"  [label="(49, 3, 64, 32)", style=solid];
-"38 quantize_per_tensor_default_2" -> "39 dequantize_per_tensor_default_2"  [label="(49, 3, 64, 32)", style=solid];
-"39 dequantize_per_tensor_default_2" -> "47 matmul"  [label="(49, 3, 64, 32)", style=solid];
-"40 linalg_vector_norm_1" -> "41 clamp_min_1"  [label="(49, 3, 64, 1)", style=solid];
-"41 clamp_min_1" -> "42 expand_as_1"  [label="(49, 3, 64, 1)", style=solid];
-"42 expand_as_1" -> "43 div_1"  [label="(49, 3, 64, 32)", style=solid];
-"43 div_1" -> "44 quantize_per_tensor_default_3"  [label="(49, 3, 64, 32)", style=solid];
-"44 quantize_per_tensor_default_3" -> "45 dequantize_per_tensor_default_3"  [label="(49, 3, 64, 32)", style=solid];
-"45 dequantize_per_tensor_default_3" -> "46 transpose"  [label="(49, 3, 64, 32)", style=solid];
-"46 transpose" -> "47 matmul"  [label="(49, 3, 32, 64)", style=solid];
-"47 matmul" -> "49 mul_1"  [label="(49, 3, 64, 64)", style=solid];
-"48 _frozen_param3" -> "49 mul_1"  [label="(3, 1, 1)", style=solid];
-"49 mul_1" -> "50 add"  [label="(49, 3, 64, 64)", style=solid];
-"50 add" -> "51 softmax"  [label="(49, 3, 64, 64)", style=solid];
-"51 softmax" -> "52 dropout"  [label="(49, 3, 64, 64)", style=solid];
-"52 dropout" -> "53 matmul_1"  [label="(49, 3, 64, 64)", style=solid];
-"53 matmul_1" -> "54 transpose_1"  [label="(49, 3, 64, 32)", style=solid];
-"54 transpose_1" -> "55 reshape_2"  [label="(49, 64, 3, 32)", style=solid];
-"55 reshape_2" -> "56 reshape_2_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
-"56 reshape_2_0_0_nncf_smooth_quant_0" -> "57 quantize_per_tensor_default_4"  [label="(49, 64, 96)", style=solid];
-"57 quantize_per_tensor_default_4" -> "58 dequantize_per_tensor_default_4"  [label="(49, 64, 96)", style=solid];
-"58 dequantize_per_tensor_default_4" -> "65 linear_3"  [label="(49, 64, 96)", style=solid];
-"59 linear_3_scale_0" -> "62 quantize_per_channel_default_3"  [label="(96,)", style=solid];
-"59 linear_3_scale_0" -> "63 dequantize_per_channel_default_3"  [label="(96,)", style=solid];
-"60 linear_3_zero_point_0" -> "62 quantize_per_channel_default_3"  [label="(96,)", style=solid];
-"60 linear_3_zero_point_0" -> "63 dequantize_per_channel_default_3"  [label="(96,)", style=solid];
-"61 compressed_weight_updated_constant2" -> "62 quantize_per_channel_default_3"  [label="(96, 96)", style=solid];
-"62 quantize_per_channel_default_3" -> "63 dequantize_per_channel_default_3"  [label="(96, 96)", style=solid];
-"63 dequantize_per_channel_default_3" -> "65 linear_3"  [label="(96, 96)", style=solid];
-"64 _param_constant11_0_0" -> "65 linear_3"  [label="(96,)", style=solid];
-"65 linear_3" -> "66 dropout_1"  [label="(49, 64, 96)", style=solid];
-"66 dropout_1" -> "67 view_3"  [label="(49, 64, 96)", style=solid];
-"67 view_3" -> "68 permute_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"68 permute_4" -> "69 reshape_3"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"69 reshape_3" -> "70 slice_2"  [label="(1, 56, 56, 96)", style=solid];
-"70 slice_2" -> "71 slice_3"  [label="(1, 56, 56, 96)", style=solid];
-"71 slice_3" -> "74 layer_norm_1"  [label="(1, 56, 56, 96)", style=solid];
-"72 _param_constant12" -> "74 layer_norm_1"  [label="(96,)", style=solid];
-"73 _param_constant13" -> "74 layer_norm_1"  [label="(96,)", style=solid];
-"74 layer_norm_1" -> "75 add_1"  [label="(1, 56, 56, 96)", style=solid];
-"75 add_1" -> "76 add_1_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 96)", style=solid];
-"75 add_1" -> "102 add_2"  [label="(1, 56, 56, 96)", style=solid];
-"76 add_1_0_0_nncf_smooth_quant_0" -> "77 quantize_per_tensor_default_5"  [label="(1, 56, 56, 96)", style=solid];
-"77 quantize_per_tensor_default_5" -> "78 dequantize_per_tensor_default_5"  [label="(1, 56, 56, 96)", style=solid];
-"78 dequantize_per_tensor_default_5" -> "85 linear_4"  [label="(1, 56, 56, 96)", style=solid];
-"79 linear_4_scale_0" -> "82 quantize_per_channel_default_4"  [label="(384,)", style=solid];
-"79 linear_4_scale_0" -> "83 dequantize_per_channel_default_4"  [label="(384,)", style=solid];
-"80 linear_4_zero_point_0" -> "82 quantize_per_channel_default_4"  [label="(384,)", style=solid];
-"80 linear_4_zero_point_0" -> "83 dequantize_per_channel_default_4"  [label="(384,)", style=solid];
-"81 compressed_weight_updated_constant3" -> "82 quantize_per_channel_default_4"  [label="(384, 96)", style=solid];
-"82 quantize_per_channel_default_4" -> "83 dequantize_per_channel_default_4"  [label="(384, 96)", style=solid];
-"83 dequantize_per_channel_default_4" -> "85 linear_4"  [label="(384, 96)", style=solid];
-"84 _param_constant15_0_0" -> "85 linear_4"  [label="(384,)", style=solid];
-"85 linear_4" -> "86 gelu"  [label="(1, 56, 56, 384)", style=solid];
-"86 gelu" -> "87 dropout_2"  [label="(1, 56, 56, 384)", style=solid];
-"87 dropout_2" -> "88 dropout_2_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 384)", style=solid];
-"88 dropout_2_0_0_nncf_smooth_quant_0" -> "89 quantize_per_tensor_default_6"  [label="(1, 56, 56, 384)", style=solid];
-"89 quantize_per_tensor_default_6" -> "90 dequantize_per_tensor_default_6"  [label="(1, 56, 56, 384)", style=solid];
-"90 dequantize_per_tensor_default_6" -> "97 linear_5"  [label="(1, 56, 56, 384)", style=solid];
-"91 linear_5_scale_0" -> "94 quantize_per_channel_default_5"  [label="(96,)", style=solid];
-"91 linear_5_scale_0" -> "95 dequantize_per_channel_default_5"  [label="(96,)", style=solid];
-"92 linear_5_zero_point_0" -> "94 quantize_per_channel_default_5"  [label="(96,)", style=solid];
-"92 linear_5_zero_point_0" -> "95 dequantize_per_channel_default_5"  [label="(96,)", style=solid];
-"93 compressed_weight_updated_constant4" -> "94 quantize_per_channel_default_5"  [label="(96, 384)", style=solid];
-"94 quantize_per_channel_default_5" -> "95 dequantize_per_channel_default_5"  [label="(96, 384)", style=solid];
-"95 dequantize_per_channel_default_5" -> "97 linear_5"  [label="(96, 384)", style=solid];
-"96 _param_constant17_0_0" -> "97 linear_5"  [label="(96,)", style=solid];
-"97 linear_5" -> "98 dropout_3"  [label="(1, 56, 56, 96)", style=solid];
-"98 dropout_3" -> "101 layer_norm_2"  [label="(1, 56, 56, 96)", style=solid];
-"99 _param_constant18" -> "101 layer_norm_2"  [label="(96,)", style=solid];
-"100 _param_constant19" -> "101 layer_norm_2"  [label="(96,)", style=solid];
-"101 layer_norm_2" -> "102 add_2"  [label="(1, 56, 56, 96)", style=solid];
-"102 add_2" -> "104 pad_1"  [label="(1, 56, 56, 96)", style=solid];
-"102 add_2" -> "182 add_5"  [label="(1, 56, 56, 96)", style=solid];
-"103 _frozen_param4" -> "140 add_3"  [label="(1, 3, 64, 64)", style=solid];
-"104 pad_1" -> "105 roll"  [label="(1, 56, 56, 96)", style=solid];
-"105 roll" -> "106 view_6"  [label="(1, 56, 56, 96)", style=solid];
-"106 view_6" -> "107 permute_6"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"107 permute_6" -> "108 reshape_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"108 reshape_4" -> "109 reshape_4_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
-"108 reshape_4" -> "141 new_zeros"  [label="(49, 64, 96)", style=solid];
-"109 reshape_4_0_0_nncf_smooth_quant_0" -> "110 quantize_per_tensor_default_7"  [label="(49, 64, 96)", style=solid];
-"110 quantize_per_tensor_default_7" -> "111 dequantize_per_tensor_default_7"  [label="(49, 64, 96)", style=solid];
-"111 dequantize_per_tensor_default_7" -> "118 linear_8"  [label="(49, 64, 96)", style=solid];
-"112 linear_8_scale_0" -> "115 quantize_per_channel_default_7"  [label="(288,)", style=solid];
-"112 linear_8_scale_0" -> "116 dequantize_per_channel_default_7"  [label="(288,)", style=solid];
-"113 linear_8_zero_point_0" -> "115 quantize_per_channel_default_7"  [label="(288,)", style=solid];
-"113 linear_8_zero_point_0" -> "116 dequantize_per_channel_default_7"  [label="(288,)", style=solid];
-"114 compressed_weight_updated_constant5" -> "115 quantize_per_channel_default_7"  [label="(288, 96)", style=solid];
-"115 quantize_per_channel_default_7" -> "116 dequantize_per_channel_default_7"  [label="(288, 96)", style=solid];
-"116 dequantize_per_channel_default_7" -> "118 linear_8"  [label="(288, 96)", style=solid];
-"117 _frozen_param5_0_0" -> "118 linear_8"  [label="(288,)", style=solid];
-"118 linear_8" -> "119 reshape_5"  [label="(49, 64, 288)", style=solid];
-"119 reshape_5" -> "120 permute_7"  [label="(49, 64, 3, 3, 32)", style=solid];
-"120 permute_7" -> "121 select_3"  [label="(3, 49, 3, 64, 32)", style=solid];
-"120 permute_7" -> "122 select_4"  [label="(3, 49, 3, 64, 32)", style=solid];
-"120 permute_7" -> "123 select_5"  [label="(3, 49, 3, 64, 32)", style=solid];
-"121 select_3" -> "124 linalg_vector_norm_2"  [label="(49, 3, 64, 32)", style=solid];
-"121 select_3" -> "126 expand_as_2"  [label="(49, 3, 64, 32)", style=solid];
-"121 select_3" -> "127 div_2"  [label="(49, 3, 64, 32)", style=solid];
-"122 select_4" -> "130 linalg_vector_norm_3"  [label="(49, 3, 64, 32)", style=solid];
-"122 select_4" -> "132 expand_as_3"  [label="(49, 3, 64, 32)", style=solid];
-"122 select_4" -> "133 div_3"  [label="(49, 3, 64, 32)", style=solid];
-"123 select_5" -> "159 matmul_3"  [label="(49, 3, 64, 32)", style=solid];
-"124 linalg_vector_norm_2" -> "125 clamp_min_2"  [label="(49, 3, 64, 1)", style=solid];
-"125 clamp_min_2" -> "126 expand_as_2"  [label="(49, 3, 64, 1)", style=solid];
-"126 expand_as_2" -> "127 div_2"  [label="(49, 3, 64, 32)", style=solid];
-"127 div_2" -> "128 quantize_per_tensor_default_8"  [label="(49, 3, 64, 32)", style=solid];
-"128 quantize_per_tensor_default_8" -> "129 dequantize_per_tensor_default_8"  [label="(49, 3, 64, 32)", style=solid];
-"129 dequantize_per_tensor_default_8" -> "137 matmul_2"  [label="(49, 3, 64, 32)", style=solid];
-"130 linalg_vector_norm_3" -> "131 clamp_min_3"  [label="(49, 3, 64, 1)", style=solid];
-"131 clamp_min_3" -> "132 expand_as_3"  [label="(49, 3, 64, 1)", style=solid];
-"132 expand_as_3" -> "133 div_3"  [label="(49, 3, 64, 32)", style=solid];
-"133 div_3" -> "134 quantize_per_tensor_default_9"  [label="(49, 3, 64, 32)", style=solid];
-"134 quantize_per_tensor_default_9" -> "135 dequantize_per_tensor_default_9"  [label="(49, 3, 64, 32)", style=solid];
-"135 dequantize_per_tensor_default_9" -> "136 transpose_2"  [label="(49, 3, 64, 32)", style=solid];
-"136 transpose_2" -> "137 matmul_2"  [label="(49, 3, 32, 64)", style=solid];
-"137 matmul_2" -> "139 mul_3"  [label="(49, 3, 64, 64)", style=solid];
-"138 _frozen_param7" -> "139 mul_3"  [label="(3, 1, 1)", style=solid];
-"139 mul_3" -> "140 add_3"  [label="(49, 3, 64, 64)", style=solid];
-"140 add_3" -> "152 view_8"  [label="(49, 3, 64, 64)", style=solid];
-"141 new_zeros" -> "142 view_7"  [label="(56, 56)", style=solid];
-"142 view_7" -> "143 permute_8"  [label="(7, 8, 7, 8)", style=solid];
-"143 permute_8" -> "144 reshape_6"  [label="(7, 7, 8, 8)", style=solid];
-"144 reshape_6" -> "145 unsqueeze_2"  [label="(49, 64)", style=solid];
-"144 reshape_6" -> "146 unsqueeze_3"  [label="(49, 64)", style=solid];
-"145 unsqueeze_2" -> "147 sub"  [label="(49, 1, 64)", style=solid];
-"146 unsqueeze_3" -> "147 sub"  [label="(49, 64, 1)", style=solid];
-"147 sub" -> "148 ne"  [label="(49, 64, 64)", style=solid];
-"147 sub" -> "149 masked_fill"  [label="(49, 64, 64)", style=solid];
-"147 sub" -> "150 eq"  [label="(49, 64, 64)", style=solid];
-"148 ne" -> "149 masked_fill"  [label="(49, 64, 64)", style=solid];
-"149 masked_fill" -> "151 masked_fill_1"  [label="(49, 64, 64)", style=solid];
-"150 eq" -> "151 masked_fill_1"  [label="(49, 64, 64)", style=solid];
-"151 masked_fill_1" -> "153 unsqueeze_4"  [label="(49, 64, 64)", style=solid];
-"152 view_8" -> "155 add_4"  [label="(1, 49, 3, 64, 64)", style=solid];
-"153 unsqueeze_4" -> "154 unsqueeze_5"  [label="(49, 1, 64, 64)", style=solid];
-"154 unsqueeze_5" -> "155 add_4"  [label="(1, 49, 1, 64, 64)", style=solid];
-"155 add_4" -> "156 view_9"  [label="(1, 49, 3, 64, 64)", style=solid];
-"156 view_9" -> "157 softmax_1"  [label="(49, 3, 64, 64)", style=solid];
-"157 softmax_1" -> "158 dropout_4"  [label="(49, 3, 64, 64)", style=solid];
-"158 dropout_4" -> "159 matmul_3"  [label="(49, 3, 64, 64)", style=solid];
-"159 matmul_3" -> "160 transpose_3"  [label="(49, 3, 64, 32)", style=solid];
-"160 transpose_3" -> "161 reshape_7"  [label="(49, 64, 3, 32)", style=solid];
-"161 reshape_7" -> "162 reshape_7_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
-"162 reshape_7_0_0_nncf_smooth_quant_0" -> "163 quantize_per_tensor_default_10"  [label="(49, 64, 96)", style=solid];
-"163 quantize_per_tensor_default_10" -> "164 dequantize_per_tensor_default_10"  [label="(49, 64, 96)", style=solid];
-"164 dequantize_per_tensor_default_10" -> "171 linear_9"  [label="(49, 64, 96)", style=solid];
-"165 linear_9_scale_0" -> "168 quantize_per_channel_default_8"  [label="(96,)", style=solid];
-"165 linear_9_scale_0" -> "169 dequantize_per_channel_default_8"  [label="(96,)", style=solid];
-"166 linear_9_zero_point_0" -> "168 quantize_per_channel_default_8"  [label="(96,)", style=solid];
-"166 linear_9_zero_point_0" -> "169 dequantize_per_channel_default_8"  [label="(96,)", style=solid];
-"167 compressed_weight_updated_constant6" -> "168 quantize_per_channel_default_8"  [label="(96, 96)", style=solid];
-"168 quantize_per_channel_default_8" -> "169 dequantize_per_channel_default_8"  [label="(96, 96)", style=solid];
-"169 dequantize_per_channel_default_8" -> "171 linear_9"  [label="(96, 96)", style=solid];
-"170 _param_constant27_0_0" -> "171 linear_9"  [label="(96,)", style=solid];
-"171 linear_9" -> "172 dropout_5"  [label="(49, 64, 96)", style=solid];
-"172 dropout_5" -> "173 view_10"  [label="(49, 64, 96)", style=solid];
-"173 view_10" -> "174 permute_9"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"174 permute_9" -> "175 reshape_8"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"175 reshape_8" -> "176 roll_1"  [label="(1, 56, 56, 96)", style=solid];
-"176 roll_1" -> "177 slice_23"  [label="(1, 56, 56, 96)", style=solid];
-"177 slice_23" -> "178 slice_24"  [label="(1, 56, 56, 96)", style=solid];
-"178 slice_24" -> "181 layer_norm_3"  [label="(1, 56, 56, 96)", style=solid];
-"179 _param_constant28" -> "181 layer_norm_3"  [label="(96,)", style=solid];
-"180 _param_constant29" -> "181 layer_norm_3"  [label="(96,)", style=solid];
-"181 layer_norm_3" -> "182 add_5"  [label="(1, 56, 56, 96)", style=solid];
-"182 add_5" -> "183 add_5_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 96)", style=solid];
-"182 add_5" -> "209 add_6"  [label="(1, 56, 56, 96)", style=solid];
-"183 add_5_0_0_nncf_smooth_quant_0" -> "184 quantize_per_tensor_default_11"  [label="(1, 56, 56, 96)", style=solid];
-"184 quantize_per_tensor_default_11" -> "185 dequantize_per_tensor_default_11"  [label="(1, 56, 56, 96)", style=solid];
-"185 dequantize_per_tensor_default_11" -> "192 linear_10"  [label="(1, 56, 56, 96)", style=solid];
-"186 linear_10_scale_0" -> "189 quantize_per_channel_default_9"  [label="(384,)", style=solid];
-"186 linear_10_scale_0" -> "190 dequantize_per_channel_default_9"  [label="(384,)", style=solid];
-"187 linear_10_zero_point_0" -> "189 quantize_per_channel_default_9"  [label="(384,)", style=solid];
-"187 linear_10_zero_point_0" -> "190 dequantize_per_channel_default_9"  [label="(384,)", style=solid];
-"188 compressed_weight_updated_constant7" -> "189 quantize_per_channel_default_9"  [label="(384, 96)", style=solid];
-"189 quantize_per_channel_default_9" -> "190 dequantize_per_channel_default_9"  [label="(384, 96)", style=solid];
-"190 dequantize_per_channel_default_9" -> "192 linear_10"  [label="(384, 96)", style=solid];
-"191 _param_constant31_0_0" -> "192 linear_10"  [label="(384,)", style=solid];
-"192 linear_10" -> "193 gelu_1"  [label="(1, 56, 56, 384)", style=solid];
-"193 gelu_1" -> "194 dropout_6"  [label="(1, 56, 56, 384)", style=solid];
-"194 dropout_6" -> "195 dropout_6_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 384)", style=solid];
-"195 dropout_6_0_0_nncf_smooth_quant_0" -> "196 quantize_per_tensor_default_12"  [label="(1, 56, 56, 384)", style=solid];
-"196 quantize_per_tensor_default_12" -> "197 dequantize_per_tensor_default_12"  [label="(1, 56, 56, 384)", style=solid];
-"197 dequantize_per_tensor_default_12" -> "204 linear_11"  [label="(1, 56, 56, 384)", style=solid];
-"198 linear_11_scale_0" -> "201 quantize_per_channel_default_10"  [label="(96,)", style=solid];
-"198 linear_11_scale_0" -> "202 dequantize_per_channel_default_10"  [label="(96,)", style=solid];
-"199 linear_11_zero_point_0" -> "201 quantize_per_channel_default_10"  [label="(96,)", style=solid];
-"199 linear_11_zero_point_0" -> "202 dequantize_per_channel_default_10"  [label="(96,)", style=solid];
-"200 compressed_weight_updated_constant8" -> "201 quantize_per_channel_default_10"  [label="(96, 384)", style=solid];
-"201 quantize_per_channel_default_10" -> "202 dequantize_per_channel_default_10"  [label="(96, 384)", style=solid];
-"202 dequantize_per_channel_default_10" -> "204 linear_11"  [label="(96, 384)", style=solid];
-"203 _param_constant33_0_0" -> "204 linear_11"  [label="(96,)", style=solid];
-"204 linear_11" -> "205 dropout_7"  [label="(1, 56, 56, 96)", style=solid];
-"205 dropout_7" -> "208 layer_norm_4"  [label="(1, 56, 56, 96)", style=solid];
-"206 _param_constant34" -> "208 layer_norm_4"  [label="(96,)", style=solid];
-"207 _param_constant35" -> "208 layer_norm_4"  [label="(96,)", style=solid];
-"208 layer_norm_4" -> "209 add_6"  [label="(1, 56, 56, 96)", style=solid];
-"209 add_6" -> "210 pad_2"  [label="(1, 56, 56, 96)", style=solid];
-"210 pad_2" -> "211 slice_25"  [label="(1, 56, 56, 96)", style=solid];
-"210 pad_2" -> "214 slice_28"  [label="(1, 56, 56, 96)", style=solid];
-"210 pad_2" -> "217 slice_31"  [label="(1, 56, 56, 96)", style=solid];
-"210 pad_2" -> "220 slice_34"  [label="(1, 56, 56, 96)", style=solid];
-"211 slice_25" -> "212 slice_26"  [label="(1, 28, 56, 96)", style=solid];
-"212 slice_26" -> "213 slice_27"  [label="(1, 28, 28, 96)", style=solid];
-"213 slice_27" -> "223 cat"  [label="(1, 28, 28, 96)", style=solid];
-"214 slice_28" -> "215 slice_29"  [label="(1, 28, 56, 96)", style=solid];
-"215 slice_29" -> "216 slice_30"  [label="(1, 28, 28, 96)", style=solid];
-"216 slice_30" -> "223 cat"  [label="(1, 28, 28, 96)", style=solid];
-"217 slice_31" -> "218 slice_32"  [label="(1, 28, 56, 96)", style=solid];
-"218 slice_32" -> "219 slice_33"  [label="(1, 28, 28, 96)", style=solid];
-"219 slice_33" -> "223 cat"  [label="(1, 28, 28, 96)", style=solid];
-"220 slice_34" -> "221 slice_35"  [label="(1, 28, 56, 96)", style=solid];
-"221 slice_35" -> "222 slice_36"  [label="(1, 28, 28, 96)", style=solid];
-"222 slice_36" -> "223 cat"  [label="(1, 28, 28, 96)", style=solid];
-"223 cat" -> "224 cat_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 384)", style=solid];
-"224 cat_0_0_nncf_smooth_quant_0" -> "225 quantize_per_tensor_default_13"  [label="(1, 28, 28, 384)", style=solid];
-"225 quantize_per_tensor_default_13" -> "226 dequantize_per_tensor_default_13"  [label="(1, 28, 28, 384)", style=solid];
-"226 dequantize_per_tensor_default_13" -> "232 linear_12"  [label="(1, 28, 28, 384)", style=solid];
-"227 linear_12_scale_0" -> "230 quantize_per_channel_default_11"  [label="(192,)", style=solid];
-"227 linear_12_scale_0" -> "231 dequantize_per_channel_default_11"  [label="(192,)", style=solid];
-"228 linear_12_zero_point_0" -> "230 quantize_per_channel_default_11"  [label="(192,)", style=solid];
-"228 linear_12_zero_point_0" -> "231 dequantize_per_channel_default_11"  [label="(192,)", style=solid];
-"229 compressed_weight_updated_constant9" -> "230 quantize_per_channel_default_11"  [label="(192, 384)", style=solid];
-"230 quantize_per_channel_default_11" -> "231 dequantize_per_channel_default_11"  [label="(192, 384)", style=solid];
-"231 dequantize_per_channel_default_11" -> "232 linear_12"  [label="(192, 384)", style=solid];
-"232 linear_12" -> "235 layer_norm_5"  [label="(1, 28, 28, 192)", style=solid];
-"233 _param_constant37" -> "235 layer_norm_5"  [label="(192,)", style=solid];
-"234 _param_constant38" -> "235 layer_norm_5"  [label="(192,)", style=solid];
-"235 layer_norm_5" -> "237 pad_3"  [label="(1, 28, 28, 192)", style=solid];
-"235 layer_norm_5" -> "300 add_8"  [label="(1, 28, 28, 192)", style=solid];
-"236 _frozen_param17" -> "272 add_7"  [label="(1, 6, 64, 64)", style=solid];
-"237 pad_3" -> "238 view_13"  [label="(1, 32, 32, 192)", style=solid];
-"238 view_13" -> "239 permute_11"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"239 permute_11" -> "240 reshape_9"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"240 reshape_9" -> "241 reshape_9_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
-"241 reshape_9_0_0_nncf_smooth_quant_0" -> "242 quantize_per_tensor_default_14"  [label="(16, 64, 192)", style=solid];
-"242 quantize_per_tensor_default_14" -> "243 dequantize_per_tensor_default_14"  [label="(16, 64, 192)", style=solid];
-"243 dequantize_per_tensor_default_14" -> "250 linear_15"  [label="(16, 64, 192)", style=solid];
-"244 linear_15_scale_0" -> "247 quantize_per_channel_default_13"  [label="(576,)", style=solid];
-"244 linear_15_scale_0" -> "248 dequantize_per_channel_default_13"  [label="(576,)", style=solid];
-"245 linear_15_zero_point_0" -> "247 quantize_per_channel_default_13"  [label="(576,)", style=solid];
-"245 linear_15_zero_point_0" -> "248 dequantize_per_channel_default_13"  [label="(576,)", style=solid];
-"246 compressed_weight_updated_constant10" -> "247 quantize_per_channel_default_13"  [label="(576, 192)", style=solid];
-"247 quantize_per_channel_default_13" -> "248 dequantize_per_channel_default_13"  [label="(576, 192)", style=solid];
-"248 dequantize_per_channel_default_13" -> "250 linear_15"  [label="(576, 192)", style=solid];
-"249 _frozen_param18_0_0" -> "250 linear_15"  [label="(576,)", style=solid];
-"250 linear_15" -> "251 reshape_10"  [label="(16, 64, 576)", style=solid];
-"251 reshape_10" -> "252 permute_12"  [label="(16, 64, 3, 6, 32)", style=solid];
-"252 permute_12" -> "253 select_6"  [label="(3, 16, 6, 64, 32)", style=solid];
-"252 permute_12" -> "254 select_7"  [label="(3, 16, 6, 64, 32)", style=solid];
-"252 permute_12" -> "255 select_8"  [label="(3, 16, 6, 64, 32)", style=solid];
-"253 select_6" -> "256 linalg_vector_norm_4"  [label="(16, 6, 64, 32)", style=solid];
-"253 select_6" -> "258 expand_as_4"  [label="(16, 6, 64, 32)", style=solid];
-"253 select_6" -> "259 div_4"  [label="(16, 6, 64, 32)", style=solid];
-"254 select_7" -> "262 linalg_vector_norm_5"  [label="(16, 6, 64, 32)", style=solid];
-"254 select_7" -> "264 expand_as_5"  [label="(16, 6, 64, 32)", style=solid];
-"254 select_7" -> "265 div_5"  [label="(16, 6, 64, 32)", style=solid];
-"255 select_8" -> "275 matmul_5"  [label="(16, 6, 64, 32)", style=solid];
-"256 linalg_vector_norm_4" -> "257 clamp_min_4"  [label="(16, 6, 64, 1)", style=solid];
-"257 clamp_min_4" -> "258 expand_as_4"  [label="(16, 6, 64, 1)", style=solid];
-"258 expand_as_4" -> "259 div_4"  [label="(16, 6, 64, 32)", style=solid];
-"259 div_4" -> "260 quantize_per_tensor_default_15"  [label="(16, 6, 64, 32)", style=solid];
-"260 quantize_per_tensor_default_15" -> "261 dequantize_per_tensor_default_15"  [label="(16, 6, 64, 32)", style=solid];
-"261 dequantize_per_tensor_default_15" -> "269 matmul_4"  [label="(16, 6, 64, 32)", style=solid];
-"262 linalg_vector_norm_5" -> "263 clamp_min_5"  [label="(16, 6, 64, 1)", style=solid];
-"263 clamp_min_5" -> "264 expand_as_5"  [label="(16, 6, 64, 1)", style=solid];
-"264 expand_as_5" -> "265 div_5"  [label="(16, 6, 64, 32)", style=solid];
-"265 div_5" -> "266 quantize_per_tensor_default_16"  [label="(16, 6, 64, 32)", style=solid];
-"266 quantize_per_tensor_default_16" -> "267 dequantize_per_tensor_default_16"  [label="(16, 6, 64, 32)", style=solid];
-"267 dequantize_per_tensor_default_16" -> "268 transpose_4"  [label="(16, 6, 64, 32)", style=solid];
-"268 transpose_4" -> "269 matmul_4"  [label="(16, 6, 32, 64)", style=solid];
-"269 matmul_4" -> "271 mul_5"  [label="(16, 6, 64, 64)", style=solid];
-"270 _frozen_param20" -> "271 mul_5"  [label="(6, 1, 1)", style=solid];
-"271 mul_5" -> "272 add_7"  [label="(16, 6, 64, 64)", style=solid];
-"272 add_7" -> "273 softmax_2"  [label="(16, 6, 64, 64)", style=solid];
-"273 softmax_2" -> "274 dropout_8"  [label="(16, 6, 64, 64)", style=solid];
-"274 dropout_8" -> "275 matmul_5"  [label="(16, 6, 64, 64)", style=solid];
-"275 matmul_5" -> "276 transpose_5"  [label="(16, 6, 64, 32)", style=solid];
-"276 transpose_5" -> "277 reshape_11"  [label="(16, 64, 6, 32)", style=solid];
-"277 reshape_11" -> "278 reshape_11_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
-"278 reshape_11_0_0_nncf_smooth_quant_0" -> "279 quantize_per_tensor_default_17"  [label="(16, 64, 192)", style=solid];
-"279 quantize_per_tensor_default_17" -> "280 dequantize_per_tensor_default_17"  [label="(16, 64, 192)", style=solid];
-"280 dequantize_per_tensor_default_17" -> "287 linear_16"  [label="(16, 64, 192)", style=solid];
-"281 linear_16_scale_0" -> "284 quantize_per_channel_default_14"  [label="(192,)", style=solid];
-"281 linear_16_scale_0" -> "285 dequantize_per_channel_default_14"  [label="(192,)", style=solid];
-"282 linear_16_zero_point_0" -> "284 quantize_per_channel_default_14"  [label="(192,)", style=solid];
-"282 linear_16_zero_point_0" -> "285 dequantize_per_channel_default_14"  [label="(192,)", style=solid];
-"283 compressed_weight_updated_constant11" -> "284 quantize_per_channel_default_14"  [label="(192, 192)", style=solid];
-"284 quantize_per_channel_default_14" -> "285 dequantize_per_channel_default_14"  [label="(192, 192)", style=solid];
-"285 dequantize_per_channel_default_14" -> "287 linear_16"  [label="(192, 192)", style=solid];
-"286 _param_constant46_0_0" -> "287 linear_16"  [label="(192,)", style=solid];
-"287 linear_16" -> "288 dropout_9"  [label="(16, 64, 192)", style=solid];
-"288 dropout_9" -> "289 view_14"  [label="(16, 64, 192)", style=solid];
-"289 view_14" -> "290 permute_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"290 permute_13" -> "291 reshape_12"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"291 reshape_12" -> "292 slice_38"  [label="(1, 32, 32, 192)", style=solid];
-"292 slice_38" -> "293 slice_39"  [label="(1, 32, 32, 192)", style=solid];
-"293 slice_39" -> "294 slice_40"  [label="(1, 28, 32, 192)", style=solid];
-"294 slice_40" -> "295 slice_41"  [label="(1, 28, 28, 192)", style=solid];
-"295 slice_41" -> "296 contiguous_3"  [label="(1, 28, 28, 192)", style=solid];
-"296 contiguous_3" -> "299 layer_norm_6"  [label="(1, 28, 28, 192)", style=solid];
-"297 _param_constant47" -> "299 layer_norm_6"  [label="(192,)", style=solid];
-"298 _param_constant48" -> "299 layer_norm_6"  [label="(192,)", style=solid];
-"299 layer_norm_6" -> "300 add_8"  [label="(1, 28, 28, 192)", style=solid];
-"300 add_8" -> "301 add_8_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 192)", style=solid];
-"300 add_8" -> "327 add_9"  [label="(1, 28, 28, 192)", style=solid];
-"301 add_8_0_0_nncf_smooth_quant_0" -> "302 quantize_per_tensor_default_18"  [label="(1, 28, 28, 192)", style=solid];
-"302 quantize_per_tensor_default_18" -> "303 dequantize_per_tensor_default_18"  [label="(1, 28, 28, 192)", style=solid];
-"303 dequantize_per_tensor_default_18" -> "310 linear_17"  [label="(1, 28, 28, 192)", style=solid];
-"304 linear_17_scale_0" -> "307 quantize_per_channel_default_15"  [label="(768,)", style=solid];
-"304 linear_17_scale_0" -> "308 dequantize_per_channel_default_15"  [label="(768,)", style=solid];
-"305 linear_17_zero_point_0" -> "307 quantize_per_channel_default_15"  [label="(768,)", style=solid];
-"305 linear_17_zero_point_0" -> "308 dequantize_per_channel_default_15"  [label="(768,)", style=solid];
-"306 compressed_weight_updated_constant12" -> "307 quantize_per_channel_default_15"  [label="(768, 192)", style=solid];
-"307 quantize_per_channel_default_15" -> "308 dequantize_per_channel_default_15"  [label="(768, 192)", style=solid];
-"308 dequantize_per_channel_default_15" -> "310 linear_17"  [label="(768, 192)", style=solid];
-"309 _param_constant50_0_0" -> "310 linear_17"  [label="(768,)", style=solid];
-"310 linear_17" -> "311 gelu_2"  [label="(1, 28, 28, 768)", style=solid];
-"311 gelu_2" -> "312 dropout_10"  [label="(1, 28, 28, 768)", style=solid];
-"312 dropout_10" -> "313 dropout_10_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 768)", style=solid];
-"313 dropout_10_0_0_nncf_smooth_quant_0" -> "314 quantize_per_tensor_default_19"  [label="(1, 28, 28, 768)", style=solid];
-"314 quantize_per_tensor_default_19" -> "315 dequantize_per_tensor_default_19"  [label="(1, 28, 28, 768)", style=solid];
-"315 dequantize_per_tensor_default_19" -> "322 linear_18"  [label="(1, 28, 28, 768)", style=solid];
-"316 linear_18_scale_0" -> "319 quantize_per_channel_default_16"  [label="(192,)", style=solid];
-"316 linear_18_scale_0" -> "320 dequantize_per_channel_default_16"  [label="(192,)", style=solid];
-"317 linear_18_zero_point_0" -> "319 quantize_per_channel_default_16"  [label="(192,)", style=solid];
-"317 linear_18_zero_point_0" -> "320 dequantize_per_channel_default_16"  [label="(192,)", style=solid];
-"318 compressed_weight_updated_constant13" -> "319 quantize_per_channel_default_16"  [label="(192, 768)", style=solid];
-"319 quantize_per_channel_default_16" -> "320 dequantize_per_channel_default_16"  [label="(192, 768)", style=solid];
-"320 dequantize_per_channel_default_16" -> "322 linear_18"  [label="(192, 768)", style=solid];
-"321 _param_constant52_0_0" -> "322 linear_18"  [label="(192,)", style=solid];
-"322 linear_18" -> "323 dropout_11"  [label="(1, 28, 28, 192)", style=solid];
-"323 dropout_11" -> "326 layer_norm_7"  [label="(1, 28, 28, 192)", style=solid];
-"324 _param_constant53" -> "326 layer_norm_7"  [label="(192,)", style=solid];
-"325 _param_constant54" -> "326 layer_norm_7"  [label="(192,)", style=solid];
-"326 layer_norm_7" -> "327 add_9"  [label="(1, 28, 28, 192)", style=solid];
-"327 add_9" -> "329 pad_4"  [label="(1, 28, 28, 192)", style=solid];
-"327 add_9" -> "410 add_12"  [label="(1, 28, 28, 192)", style=solid];
-"328 _frozen_param21" -> "365 add_10"  [label="(1, 6, 64, 64)", style=solid];
-"329 pad_4" -> "330 roll_2"  [label="(1, 32, 32, 192)", style=solid];
-"330 roll_2" -> "331 view_17"  [label="(1, 32, 32, 192)", style=solid];
-"331 view_17" -> "332 permute_15"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"332 permute_15" -> "333 reshape_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"333 reshape_13" -> "334 reshape_13_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
-"333 reshape_13" -> "366 new_zeros_1"  [label="(16, 64, 192)", style=solid];
-"334 reshape_13_0_0_nncf_smooth_quant_0" -> "335 quantize_per_tensor_default_20"  [label="(16, 64, 192)", style=solid];
-"335 quantize_per_tensor_default_20" -> "336 dequantize_per_tensor_default_20"  [label="(16, 64, 192)", style=solid];
-"336 dequantize_per_tensor_default_20" -> "343 linear_21"  [label="(16, 64, 192)", style=solid];
-"337 linear_21_scale_0" -> "340 quantize_per_channel_default_18"  [label="(576,)", style=solid];
-"337 linear_21_scale_0" -> "341 dequantize_per_channel_default_18"  [label="(576,)", style=solid];
-"338 linear_21_zero_point_0" -> "340 quantize_per_channel_default_18"  [label="(576,)", style=solid];
-"338 linear_21_zero_point_0" -> "341 dequantize_per_channel_default_18"  [label="(576,)", style=solid];
-"339 compressed_weight_updated_constant14" -> "340 quantize_per_channel_default_18"  [label="(576, 192)", style=solid];
-"340 quantize_per_channel_default_18" -> "341 dequantize_per_channel_default_18"  [label="(576, 192)", style=solid];
-"341 dequantize_per_channel_default_18" -> "343 linear_21"  [label="(576, 192)", style=solid];
-"342 _frozen_param22_0_0" -> "343 linear_21"  [label="(576,)", style=solid];
-"343 linear_21" -> "344 reshape_14"  [label="(16, 64, 576)", style=solid];
-"344 reshape_14" -> "345 permute_16"  [label="(16, 64, 3, 6, 32)", style=solid];
-"345 permute_16" -> "346 select_9"  [label="(3, 16, 6, 64, 32)", style=solid];
-"345 permute_16" -> "347 select_10"  [label="(3, 16, 6, 64, 32)", style=solid];
-"345 permute_16" -> "348 select_11"  [label="(3, 16, 6, 64, 32)", style=solid];
-"346 select_9" -> "349 linalg_vector_norm_6"  [label="(16, 6, 64, 32)", style=solid];
-"346 select_9" -> "351 expand_as_6"  [label="(16, 6, 64, 32)", style=solid];
-"346 select_9" -> "352 div_6"  [label="(16, 6, 64, 32)", style=solid];
-"347 select_10" -> "355 linalg_vector_norm_7"  [label="(16, 6, 64, 32)", style=solid];
-"347 select_10" -> "357 expand_as_7"  [label="(16, 6, 64, 32)", style=solid];
-"347 select_10" -> "358 div_7"  [label="(16, 6, 64, 32)", style=solid];
-"348 select_11" -> "384 matmul_7"  [label="(16, 6, 64, 32)", style=solid];
-"349 linalg_vector_norm_6" -> "350 clamp_min_6"  [label="(16, 6, 64, 1)", style=solid];
-"350 clamp_min_6" -> "351 expand_as_6"  [label="(16, 6, 64, 1)", style=solid];
-"351 expand_as_6" -> "352 div_6"  [label="(16, 6, 64, 32)", style=solid];
-"352 div_6" -> "353 quantize_per_tensor_default_21"  [label="(16, 6, 64, 32)", style=solid];
-"353 quantize_per_tensor_default_21" -> "354 dequantize_per_tensor_default_21"  [label="(16, 6, 64, 32)", style=solid];
-"354 dequantize_per_tensor_default_21" -> "362 matmul_6"  [label="(16, 6, 64, 32)", style=solid];
-"355 linalg_vector_norm_7" -> "356 clamp_min_7"  [label="(16, 6, 64, 1)", style=solid];
-"356 clamp_min_7" -> "357 expand_as_7"  [label="(16, 6, 64, 1)", style=solid];
-"357 expand_as_7" -> "358 div_7"  [label="(16, 6, 64, 32)", style=solid];
-"358 div_7" -> "359 quantize_per_tensor_default_22"  [label="(16, 6, 64, 32)", style=solid];
-"359 quantize_per_tensor_default_22" -> "360 dequantize_per_tensor_default_22"  [label="(16, 6, 64, 32)", style=solid];
-"360 dequantize_per_tensor_default_22" -> "361 transpose_6"  [label="(16, 6, 64, 32)", style=solid];
-"361 transpose_6" -> "362 matmul_6"  [label="(16, 6, 32, 64)", style=solid];
-"362 matmul_6" -> "364 mul_7"  [label="(16, 6, 64, 64)", style=solid];
-"363 _frozen_param24" -> "364 mul_7"  [label="(6, 1, 1)", style=solid];
-"364 mul_7" -> "365 add_10"  [label="(16, 6, 64, 64)", style=solid];
-"365 add_10" -> "377 view_19"  [label="(16, 6, 64, 64)", style=solid];
-"366 new_zeros_1" -> "367 view_18"  [label="(32, 32)", style=solid];
-"367 view_18" -> "368 permute_17"  [label="(4, 8, 4, 8)", style=solid];
-"368 permute_17" -> "369 reshape_15"  [label="(4, 4, 8, 8)", style=solid];
-"369 reshape_15" -> "370 unsqueeze_8"  [label="(16, 64)", style=solid];
-"369 reshape_15" -> "371 unsqueeze_9"  [label="(16, 64)", style=solid];
-"370 unsqueeze_8" -> "372 sub_1"  [label="(16, 1, 64)", style=solid];
-"371 unsqueeze_9" -> "372 sub_1"  [label="(16, 64, 1)", style=solid];
-"372 sub_1" -> "373 ne_1"  [label="(16, 64, 64)", style=solid];
-"372 sub_1" -> "374 masked_fill_2"  [label="(16, 64, 64)", style=solid];
-"372 sub_1" -> "375 eq_1"  [label="(16, 64, 64)", style=solid];
-"373 ne_1" -> "374 masked_fill_2"  [label="(16, 64, 64)", style=solid];
-"374 masked_fill_2" -> "376 masked_fill_3"  [label="(16, 64, 64)", style=solid];
-"375 eq_1" -> "376 masked_fill_3"  [label="(16, 64, 64)", style=solid];
-"376 masked_fill_3" -> "378 unsqueeze_10"  [label="(16, 64, 64)", style=solid];
-"377 view_19" -> "380 add_11"  [label="(1, 16, 6, 64, 64)", style=solid];
-"378 unsqueeze_10" -> "379 unsqueeze_11"  [label="(16, 1, 64, 64)", style=solid];
-"379 unsqueeze_11" -> "380 add_11"  [label="(1, 16, 1, 64, 64)", style=solid];
-"380 add_11" -> "381 view_20"  [label="(1, 16, 6, 64, 64)", style=solid];
-"381 view_20" -> "382 softmax_3"  [label="(16, 6, 64, 64)", style=solid];
-"382 softmax_3" -> "383 dropout_12"  [label="(16, 6, 64, 64)", style=solid];
-"383 dropout_12" -> "384 matmul_7"  [label="(16, 6, 64, 64)", style=solid];
-"384 matmul_7" -> "385 transpose_7"  [label="(16, 6, 64, 32)", style=solid];
-"385 transpose_7" -> "386 reshape_16"  [label="(16, 64, 6, 32)", style=solid];
-"386 reshape_16" -> "387 reshape_16_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
-"387 reshape_16_0_0_nncf_smooth_quant_0" -> "388 quantize_per_tensor_default_23"  [label="(16, 64, 192)", style=solid];
-"388 quantize_per_tensor_default_23" -> "389 dequantize_per_tensor_default_23"  [label="(16, 64, 192)", style=solid];
-"389 dequantize_per_tensor_default_23" -> "396 linear_22"  [label="(16, 64, 192)", style=solid];
-"390 linear_22_scale_0" -> "393 quantize_per_channel_default_19"  [label="(192,)", style=solid];
-"390 linear_22_scale_0" -> "394 dequantize_per_channel_default_19"  [label="(192,)", style=solid];
-"391 linear_22_zero_point_0" -> "393 quantize_per_channel_default_19"  [label="(192,)", style=solid];
-"391 linear_22_zero_point_0" -> "394 dequantize_per_channel_default_19"  [label="(192,)", style=solid];
-"392 compressed_weight_updated_constant15" -> "393 quantize_per_channel_default_19"  [label="(192, 192)", style=solid];
-"393 quantize_per_channel_default_19" -> "394 dequantize_per_channel_default_19"  [label="(192, 192)", style=solid];
-"394 dequantize_per_channel_default_19" -> "396 linear_22"  [label="(192, 192)", style=solid];
-"395 _param_constant62_0_0" -> "396 linear_22"  [label="(192,)", style=solid];
-"396 linear_22" -> "397 dropout_13"  [label="(16, 64, 192)", style=solid];
-"397 dropout_13" -> "398 view_21"  [label="(16, 64, 192)", style=solid];
-"398 view_21" -> "399 permute_18"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"399 permute_18" -> "400 reshape_17"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"400 reshape_17" -> "401 roll_3"  [label="(1, 32, 32, 192)", style=solid];
-"401 roll_3" -> "402 slice_61"  [label="(1, 32, 32, 192)", style=solid];
-"402 slice_61" -> "403 slice_62"  [label="(1, 32, 32, 192)", style=solid];
-"403 slice_62" -> "404 slice_63"  [label="(1, 28, 32, 192)", style=solid];
-"404 slice_63" -> "405 slice_64"  [label="(1, 28, 28, 192)", style=solid];
-"405 slice_64" -> "406 contiguous_5"  [label="(1, 28, 28, 192)", style=solid];
-"406 contiguous_5" -> "409 layer_norm_8"  [label="(1, 28, 28, 192)", style=solid];
-"407 _param_constant63" -> "409 layer_norm_8"  [label="(192,)", style=solid];
-"408 _param_constant64" -> "409 layer_norm_8"  [label="(192,)", style=solid];
-"409 layer_norm_8" -> "410 add_12"  [label="(1, 28, 28, 192)", style=solid];
-"410 add_12" -> "411 add_12_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 192)", style=solid];
-"410 add_12" -> "437 add_13"  [label="(1, 28, 28, 192)", style=solid];
-"411 add_12_0_0_nncf_smooth_quant_0" -> "412 quantize_per_tensor_default_24"  [label="(1, 28, 28, 192)", style=solid];
-"412 quantize_per_tensor_default_24" -> "413 dequantize_per_tensor_default_24"  [label="(1, 28, 28, 192)", style=solid];
-"413 dequantize_per_tensor_default_24" -> "420 linear_23"  [label="(1, 28, 28, 192)", style=solid];
-"414 linear_23_scale_0" -> "417 quantize_per_channel_default_20"  [label="(768,)", style=solid];
-"414 linear_23_scale_0" -> "418 dequantize_per_channel_default_20"  [label="(768,)", style=solid];
-"415 linear_23_zero_point_0" -> "417 quantize_per_channel_default_20"  [label="(768,)", style=solid];
-"415 linear_23_zero_point_0" -> "418 dequantize_per_channel_default_20"  [label="(768,)", style=solid];
-"416 compressed_weight_updated_constant16" -> "417 quantize_per_channel_default_20"  [label="(768, 192)", style=solid];
-"417 quantize_per_channel_default_20" -> "418 dequantize_per_channel_default_20"  [label="(768, 192)", style=solid];
-"418 dequantize_per_channel_default_20" -> "420 linear_23"  [label="(768, 192)", style=solid];
-"419 _param_constant66_0_0" -> "420 linear_23"  [label="(768,)", style=solid];
-"420 linear_23" -> "421 gelu_3"  [label="(1, 28, 28, 768)", style=solid];
-"421 gelu_3" -> "422 dropout_14"  [label="(1, 28, 28, 768)", style=solid];
-"422 dropout_14" -> "423 dropout_14_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 768)", style=solid];
-"423 dropout_14_0_0_nncf_smooth_quant_0" -> "424 quantize_per_tensor_default_25"  [label="(1, 28, 28, 768)", style=solid];
-"424 quantize_per_tensor_default_25" -> "425 dequantize_per_tensor_default_25"  [label="(1, 28, 28, 768)", style=solid];
-"425 dequantize_per_tensor_default_25" -> "432 linear_24"  [label="(1, 28, 28, 768)", style=solid];
-"426 linear_24_scale_0" -> "429 quantize_per_channel_default_21"  [label="(192,)", style=solid];
-"426 linear_24_scale_0" -> "430 dequantize_per_channel_default_21"  [label="(192,)", style=solid];
-"427 linear_24_zero_point_0" -> "429 quantize_per_channel_default_21"  [label="(192,)", style=solid];
-"427 linear_24_zero_point_0" -> "430 dequantize_per_channel_default_21"  [label="(192,)", style=solid];
-"428 compressed_weight_updated_constant17" -> "429 quantize_per_channel_default_21"  [label="(192, 768)", style=solid];
-"429 quantize_per_channel_default_21" -> "430 dequantize_per_channel_default_21"  [label="(192, 768)", style=solid];
-"430 dequantize_per_channel_default_21" -> "432 linear_24"  [label="(192, 768)", style=solid];
-"431 _param_constant68_0_0" -> "432 linear_24"  [label="(192,)", style=solid];
-"432 linear_24" -> "433 dropout_15"  [label="(1, 28, 28, 192)", style=solid];
-"433 dropout_15" -> "436 layer_norm_9"  [label="(1, 28, 28, 192)", style=solid];
-"434 _param_constant69" -> "436 layer_norm_9"  [label="(192,)", style=solid];
-"435 _param_constant70" -> "436 layer_norm_9"  [label="(192,)", style=solid];
-"436 layer_norm_9" -> "437 add_13"  [label="(1, 28, 28, 192)", style=solid];
-"437 add_13" -> "438 pad_5"  [label="(1, 28, 28, 192)", style=solid];
-"438 pad_5" -> "439 slice_65"  [label="(1, 28, 28, 192)", style=solid];
-"438 pad_5" -> "442 slice_68"  [label="(1, 28, 28, 192)", style=solid];
-"438 pad_5" -> "445 slice_71"  [label="(1, 28, 28, 192)", style=solid];
-"438 pad_5" -> "448 slice_74"  [label="(1, 28, 28, 192)", style=solid];
-"439 slice_65" -> "440 slice_66"  [label="(1, 14, 28, 192)", style=solid];
-"440 slice_66" -> "441 slice_67"  [label="(1, 14, 14, 192)", style=solid];
-"441 slice_67" -> "451 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"442 slice_68" -> "443 slice_69"  [label="(1, 14, 28, 192)", style=solid];
-"443 slice_69" -> "444 slice_70"  [label="(1, 14, 14, 192)", style=solid];
-"444 slice_70" -> "451 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"445 slice_71" -> "446 slice_72"  [label="(1, 14, 28, 192)", style=solid];
-"446 slice_72" -> "447 slice_73"  [label="(1, 14, 14, 192)", style=solid];
-"447 slice_73" -> "451 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"448 slice_74" -> "449 slice_75"  [label="(1, 14, 28, 192)", style=solid];
-"449 slice_75" -> "450 slice_76"  [label="(1, 14, 14, 192)", style=solid];
-"450 slice_76" -> "451 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"451 cat_1" -> "452 cat_1_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 768)", style=solid];
-"452 cat_1_0_0_nncf_smooth_quant_0" -> "453 quantize_per_tensor_default_26"  [label="(1, 14, 14, 768)", style=solid];
-"453 quantize_per_tensor_default_26" -> "454 dequantize_per_tensor_default_26"  [label="(1, 14, 14, 768)", style=solid];
-"454 dequantize_per_tensor_default_26" -> "460 linear_25"  [label="(1, 14, 14, 768)", style=solid];
-"455 linear_25_scale_0" -> "458 quantize_per_channel_default_22"  [label="(384,)", style=solid];
-"455 linear_25_scale_0" -> "459 dequantize_per_channel_default_22"  [label="(384,)", style=solid];
-"456 linear_25_zero_point_0" -> "458 quantize_per_channel_default_22"  [label="(384,)", style=solid];
-"456 linear_25_zero_point_0" -> "459 dequantize_per_channel_default_22"  [label="(384,)", style=solid];
-"457 compressed_weight_updated_constant18" -> "458 quantize_per_channel_default_22"  [label="(384, 768)", style=solid];
-"458 quantize_per_channel_default_22" -> "459 dequantize_per_channel_default_22"  [label="(384, 768)", style=solid];
-"459 dequantize_per_channel_default_22" -> "460 linear_25"  [label="(384, 768)", style=solid];
-"460 linear_25" -> "463 layer_norm_10"  [label="(1, 14, 14, 384)", style=solid];
-"461 _param_constant72" -> "463 layer_norm_10"  [label="(384,)", style=solid];
-"462 _param_constant73" -> "463 layer_norm_10"  [label="(384,)", style=solid];
-"463 layer_norm_10" -> "465 pad_6"  [label="(1, 14, 14, 384)", style=solid];
-"463 layer_norm_10" -> "528 add_15"  [label="(1, 14, 14, 384)", style=solid];
-"464 _frozen_param34" -> "500 add_14"  [label="(1, 12, 64, 64)", style=solid];
-"465 pad_6" -> "466 view_24"  [label="(1, 16, 16, 384)", style=solid];
-"466 view_24" -> "467 permute_20"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"467 permute_20" -> "468 reshape_18"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"468 reshape_18" -> "469 reshape_18_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"469 reshape_18_0_0_nncf_smooth_quant_0" -> "470 quantize_per_tensor_default_27"  [label="(4, 64, 384)", style=solid];
-"470 quantize_per_tensor_default_27" -> "471 dequantize_per_tensor_default_27"  [label="(4, 64, 384)", style=solid];
-"471 dequantize_per_tensor_default_27" -> "478 linear_28"  [label="(4, 64, 384)", style=solid];
-"472 linear_28_scale_0" -> "475 quantize_per_channel_default_24"  [label="(1152,)", style=solid];
-"472 linear_28_scale_0" -> "476 dequantize_per_channel_default_24"  [label="(1152,)", style=solid];
-"473 linear_28_zero_point_0" -> "475 quantize_per_channel_default_24"  [label="(1152,)", style=solid];
-"473 linear_28_zero_point_0" -> "476 dequantize_per_channel_default_24"  [label="(1152,)", style=solid];
-"474 compressed_weight_updated_constant19" -> "475 quantize_per_channel_default_24"  [label="(1152, 384)", style=solid];
-"475 quantize_per_channel_default_24" -> "476 dequantize_per_channel_default_24"  [label="(1152, 384)", style=solid];
-"476 dequantize_per_channel_default_24" -> "478 linear_28"  [label="(1152, 384)", style=solid];
-"477 _frozen_param35_0_0" -> "478 linear_28"  [label="(1152,)", style=solid];
-"478 linear_28" -> "479 reshape_19"  [label="(4, 64, 1152)", style=solid];
-"479 reshape_19" -> "480 permute_21"  [label="(4, 64, 3, 12, 32)", style=solid];
-"480 permute_21" -> "481 select_12"  [label="(3, 4, 12, 64, 32)", style=solid];
-"480 permute_21" -> "482 select_13"  [label="(3, 4, 12, 64, 32)", style=solid];
-"480 permute_21" -> "483 select_14"  [label="(3, 4, 12, 64, 32)", style=solid];
-"481 select_12" -> "484 linalg_vector_norm_8"  [label="(4, 12, 64, 32)", style=solid];
-"481 select_12" -> "486 expand_as_8"  [label="(4, 12, 64, 32)", style=solid];
-"481 select_12" -> "487 div_8"  [label="(4, 12, 64, 32)", style=solid];
-"482 select_13" -> "490 linalg_vector_norm_9"  [label="(4, 12, 64, 32)", style=solid];
-"482 select_13" -> "492 expand_as_9"  [label="(4, 12, 64, 32)", style=solid];
-"482 select_13" -> "493 div_9"  [label="(4, 12, 64, 32)", style=solid];
-"483 select_14" -> "503 matmul_9"  [label="(4, 12, 64, 32)", style=solid];
-"484 linalg_vector_norm_8" -> "485 clamp_min_8"  [label="(4, 12, 64, 1)", style=solid];
-"485 clamp_min_8" -> "486 expand_as_8"  [label="(4, 12, 64, 1)", style=solid];
-"486 expand_as_8" -> "487 div_8"  [label="(4, 12, 64, 32)", style=solid];
-"487 div_8" -> "488 quantize_per_tensor_default_28"  [label="(4, 12, 64, 32)", style=solid];
-"488 quantize_per_tensor_default_28" -> "489 dequantize_per_tensor_default_28"  [label="(4, 12, 64, 32)", style=solid];
-"489 dequantize_per_tensor_default_28" -> "497 matmul_8"  [label="(4, 12, 64, 32)", style=solid];
-"490 linalg_vector_norm_9" -> "491 clamp_min_9"  [label="(4, 12, 64, 1)", style=solid];
-"491 clamp_min_9" -> "492 expand_as_9"  [label="(4, 12, 64, 1)", style=solid];
-"492 expand_as_9" -> "493 div_9"  [label="(4, 12, 64, 32)", style=solid];
-"493 div_9" -> "494 quantize_per_tensor_default_29"  [label="(4, 12, 64, 32)", style=solid];
-"494 quantize_per_tensor_default_29" -> "495 dequantize_per_tensor_default_29"  [label="(4, 12, 64, 32)", style=solid];
-"495 dequantize_per_tensor_default_29" -> "496 transpose_8"  [label="(4, 12, 64, 32)", style=solid];
-"496 transpose_8" -> "497 matmul_8"  [label="(4, 12, 32, 64)", style=solid];
-"497 matmul_8" -> "499 mul_9"  [label="(4, 12, 64, 64)", style=solid];
-"498 _frozen_param37" -> "499 mul_9"  [label="(12, 1, 1)", style=solid];
-"499 mul_9" -> "500 add_14"  [label="(4, 12, 64, 64)", style=solid];
-"500 add_14" -> "501 softmax_4"  [label="(4, 12, 64, 64)", style=solid];
-"501 softmax_4" -> "502 dropout_16"  [label="(4, 12, 64, 64)", style=solid];
-"502 dropout_16" -> "503 matmul_9"  [label="(4, 12, 64, 64)", style=solid];
-"503 matmul_9" -> "504 transpose_9"  [label="(4, 12, 64, 32)", style=solid];
-"504 transpose_9" -> "505 reshape_20"  [label="(4, 64, 12, 32)", style=solid];
-"505 reshape_20" -> "506 reshape_20_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"506 reshape_20_0_0_nncf_smooth_quant_0" -> "507 quantize_per_tensor_default_30"  [label="(4, 64, 384)", style=solid];
-"507 quantize_per_tensor_default_30" -> "508 dequantize_per_tensor_default_30"  [label="(4, 64, 384)", style=solid];
-"508 dequantize_per_tensor_default_30" -> "515 linear_29"  [label="(4, 64, 384)", style=solid];
-"509 linear_29_scale_0" -> "512 quantize_per_channel_default_25"  [label="(384,)", style=solid];
-"509 linear_29_scale_0" -> "513 dequantize_per_channel_default_25"  [label="(384,)", style=solid];
-"510 linear_29_zero_point_0" -> "512 quantize_per_channel_default_25"  [label="(384,)", style=solid];
-"510 linear_29_zero_point_0" -> "513 dequantize_per_channel_default_25"  [label="(384,)", style=solid];
-"511 compressed_weight_updated_constant20" -> "512 quantize_per_channel_default_25"  [label="(384, 384)", style=solid];
-"512 quantize_per_channel_default_25" -> "513 dequantize_per_channel_default_25"  [label="(384, 384)", style=solid];
-"513 dequantize_per_channel_default_25" -> "515 linear_29"  [label="(384, 384)", style=solid];
-"514 _param_constant81_0_0" -> "515 linear_29"  [label="(384,)", style=solid];
-"515 linear_29" -> "516 dropout_17"  [label="(4, 64, 384)", style=solid];
-"516 dropout_17" -> "517 view_25"  [label="(4, 64, 384)", style=solid];
-"517 view_25" -> "518 permute_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"518 permute_22" -> "519 reshape_21"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"519 reshape_21" -> "520 slice_78"  [label="(1, 16, 16, 384)", style=solid];
-"520 slice_78" -> "521 slice_79"  [label="(1, 16, 16, 384)", style=solid];
-"521 slice_79" -> "522 slice_80"  [label="(1, 14, 16, 384)", style=solid];
-"522 slice_80" -> "523 slice_81"  [label="(1, 14, 14, 384)", style=solid];
-"523 slice_81" -> "524 contiguous_7"  [label="(1, 14, 14, 384)", style=solid];
-"524 contiguous_7" -> "527 layer_norm_11"  [label="(1, 14, 14, 384)", style=solid];
-"525 _param_constant82" -> "527 layer_norm_11"  [label="(384,)", style=solid];
-"526 _param_constant83" -> "527 layer_norm_11"  [label="(384,)", style=solid];
-"527 layer_norm_11" -> "528 add_15"  [label="(1, 14, 14, 384)", style=solid];
-"528 add_15" -> "529 add_15_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"528 add_15" -> "555 add_16"  [label="(1, 14, 14, 384)", style=solid];
-"529 add_15_0_0_nncf_smooth_quant_0" -> "530 quantize_per_tensor_default_31"  [label="(1, 14, 14, 384)", style=solid];
-"530 quantize_per_tensor_default_31" -> "531 dequantize_per_tensor_default_31"  [label="(1, 14, 14, 384)", style=solid];
-"531 dequantize_per_tensor_default_31" -> "538 linear_30"  [label="(1, 14, 14, 384)", style=solid];
-"532 linear_30_scale_0" -> "535 quantize_per_channel_default_26"  [label="(1536,)", style=solid];
-"532 linear_30_scale_0" -> "536 dequantize_per_channel_default_26"  [label="(1536,)", style=solid];
-"533 linear_30_zero_point_0" -> "535 quantize_per_channel_default_26"  [label="(1536,)", style=solid];
-"533 linear_30_zero_point_0" -> "536 dequantize_per_channel_default_26"  [label="(1536,)", style=solid];
-"534 compressed_weight_updated_constant21" -> "535 quantize_per_channel_default_26"  [label="(1536, 384)", style=solid];
-"535 quantize_per_channel_default_26" -> "536 dequantize_per_channel_default_26"  [label="(1536, 384)", style=solid];
-"536 dequantize_per_channel_default_26" -> "538 linear_30"  [label="(1536, 384)", style=solid];
-"537 _param_constant85_0_0" -> "538 linear_30"  [label="(1536,)", style=solid];
-"538 linear_30" -> "539 gelu_4"  [label="(1, 14, 14, 1536)", style=solid];
-"539 gelu_4" -> "540 dropout_18"  [label="(1, 14, 14, 1536)", style=solid];
-"540 dropout_18" -> "541 dropout_18_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"541 dropout_18_0_0_nncf_smooth_quant_0" -> "542 quantize_per_tensor_default_32"  [label="(1, 14, 14, 1536)", style=solid];
-"542 quantize_per_tensor_default_32" -> "543 dequantize_per_tensor_default_32"  [label="(1, 14, 14, 1536)", style=solid];
-"543 dequantize_per_tensor_default_32" -> "550 linear_31"  [label="(1, 14, 14, 1536)", style=solid];
-"544 linear_31_scale_0" -> "547 quantize_per_channel_default_27"  [label="(384,)", style=solid];
-"544 linear_31_scale_0" -> "548 dequantize_per_channel_default_27"  [label="(384,)", style=solid];
-"545 linear_31_zero_point_0" -> "547 quantize_per_channel_default_27"  [label="(384,)", style=solid];
-"545 linear_31_zero_point_0" -> "548 dequantize_per_channel_default_27"  [label="(384,)", style=solid];
-"546 compressed_weight_updated_constant22" -> "547 quantize_per_channel_default_27"  [label="(384, 1536)", style=solid];
-"547 quantize_per_channel_default_27" -> "548 dequantize_per_channel_default_27"  [label="(384, 1536)", style=solid];
-"548 dequantize_per_channel_default_27" -> "550 linear_31"  [label="(384, 1536)", style=solid];
-"549 _param_constant87_0_0" -> "550 linear_31"  [label="(384,)", style=solid];
-"550 linear_31" -> "551 dropout_19"  [label="(1, 14, 14, 384)", style=solid];
-"551 dropout_19" -> "554 layer_norm_12"  [label="(1, 14, 14, 384)", style=solid];
-"552 _param_constant88" -> "554 layer_norm_12"  [label="(384,)", style=solid];
-"553 _param_constant89" -> "554 layer_norm_12"  [label="(384,)", style=solid];
-"554 layer_norm_12" -> "555 add_16"  [label="(1, 14, 14, 384)", style=solid];
-"555 add_16" -> "557 pad_7"  [label="(1, 14, 14, 384)", style=solid];
-"555 add_16" -> "638 add_19"  [label="(1, 14, 14, 384)", style=solid];
-"556 _frozen_param38" -> "593 add_17"  [label="(1, 12, 64, 64)", style=solid];
-"557 pad_7" -> "558 roll_4"  [label="(1, 16, 16, 384)", style=solid];
-"558 roll_4" -> "559 view_28"  [label="(1, 16, 16, 384)", style=solid];
-"559 view_28" -> "560 permute_24"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"560 permute_24" -> "561 reshape_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"561 reshape_22" -> "562 reshape_22_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"561 reshape_22" -> "594 new_zeros_2"  [label="(4, 64, 384)", style=solid];
-"562 reshape_22_0_0_nncf_smooth_quant_0" -> "563 quantize_per_tensor_default_33"  [label="(4, 64, 384)", style=solid];
-"563 quantize_per_tensor_default_33" -> "564 dequantize_per_tensor_default_33"  [label="(4, 64, 384)", style=solid];
-"564 dequantize_per_tensor_default_33" -> "571 linear_34"  [label="(4, 64, 384)", style=solid];
-"565 linear_34_scale_0" -> "568 quantize_per_channel_default_29"  [label="(1152,)", style=solid];
-"565 linear_34_scale_0" -> "569 dequantize_per_channel_default_29"  [label="(1152,)", style=solid];
-"566 linear_34_zero_point_0" -> "568 quantize_per_channel_default_29"  [label="(1152,)", style=solid];
-"566 linear_34_zero_point_0" -> "569 dequantize_per_channel_default_29"  [label="(1152,)", style=solid];
-"567 compressed_weight_updated_constant23" -> "568 quantize_per_channel_default_29"  [label="(1152, 384)", style=solid];
-"568 quantize_per_channel_default_29" -> "569 dequantize_per_channel_default_29"  [label="(1152, 384)", style=solid];
-"569 dequantize_per_channel_default_29" -> "571 linear_34"  [label="(1152, 384)", style=solid];
-"570 _frozen_param39_0_0" -> "571 linear_34"  [label="(1152,)", style=solid];
-"571 linear_34" -> "572 reshape_23"  [label="(4, 64, 1152)", style=solid];
-"572 reshape_23" -> "573 permute_25"  [label="(4, 64, 3, 12, 32)", style=solid];
-"573 permute_25" -> "574 select_15"  [label="(3, 4, 12, 64, 32)", style=solid];
-"573 permute_25" -> "575 select_16"  [label="(3, 4, 12, 64, 32)", style=solid];
-"573 permute_25" -> "576 select_17"  [label="(3, 4, 12, 64, 32)", style=solid];
-"574 select_15" -> "577 linalg_vector_norm_10"  [label="(4, 12, 64, 32)", style=solid];
-"574 select_15" -> "579 expand_as_10"  [label="(4, 12, 64, 32)", style=solid];
-"574 select_15" -> "580 div_10"  [label="(4, 12, 64, 32)", style=solid];
-"575 select_16" -> "583 linalg_vector_norm_11"  [label="(4, 12, 64, 32)", style=solid];
-"575 select_16" -> "585 expand_as_11"  [label="(4, 12, 64, 32)", style=solid];
-"575 select_16" -> "586 div_11"  [label="(4, 12, 64, 32)", style=solid];
-"576 select_17" -> "612 matmul_11"  [label="(4, 12, 64, 32)", style=solid];
-"577 linalg_vector_norm_10" -> "578 clamp_min_10"  [label="(4, 12, 64, 1)", style=solid];
-"578 clamp_min_10" -> "579 expand_as_10"  [label="(4, 12, 64, 1)", style=solid];
-"579 expand_as_10" -> "580 div_10"  [label="(4, 12, 64, 32)", style=solid];
-"580 div_10" -> "581 quantize_per_tensor_default_34"  [label="(4, 12, 64, 32)", style=solid];
-"581 quantize_per_tensor_default_34" -> "582 dequantize_per_tensor_default_34"  [label="(4, 12, 64, 32)", style=solid];
-"582 dequantize_per_tensor_default_34" -> "590 matmul_10"  [label="(4, 12, 64, 32)", style=solid];
-"583 linalg_vector_norm_11" -> "584 clamp_min_11"  [label="(4, 12, 64, 1)", style=solid];
-"584 clamp_min_11" -> "585 expand_as_11"  [label="(4, 12, 64, 1)", style=solid];
-"585 expand_as_11" -> "586 div_11"  [label="(4, 12, 64, 32)", style=solid];
-"586 div_11" -> "587 quantize_per_tensor_default_35"  [label="(4, 12, 64, 32)", style=solid];
-"587 quantize_per_tensor_default_35" -> "588 dequantize_per_tensor_default_35"  [label="(4, 12, 64, 32)", style=solid];
-"588 dequantize_per_tensor_default_35" -> "589 transpose_10"  [label="(4, 12, 64, 32)", style=solid];
-"589 transpose_10" -> "590 matmul_10"  [label="(4, 12, 32, 64)", style=solid];
-"590 matmul_10" -> "592 mul_11"  [label="(4, 12, 64, 64)", style=solid];
-"591 _frozen_param41" -> "592 mul_11"  [label="(12, 1, 1)", style=solid];
-"592 mul_11" -> "593 add_17"  [label="(4, 12, 64, 64)", style=solid];
-"593 add_17" -> "605 view_30"  [label="(4, 12, 64, 64)", style=solid];
-"594 new_zeros_2" -> "595 view_29"  [label="(16, 16)", style=solid];
-"595 view_29" -> "596 permute_26"  [label="(2, 8, 2, 8)", style=solid];
-"596 permute_26" -> "597 reshape_24"  [label="(2, 2, 8, 8)", style=solid];
-"597 reshape_24" -> "598 unsqueeze_14"  [label="(4, 64)", style=solid];
-"597 reshape_24" -> "599 unsqueeze_15"  [label="(4, 64)", style=solid];
-"598 unsqueeze_14" -> "600 sub_2"  [label="(4, 1, 64)", style=solid];
-"599 unsqueeze_15" -> "600 sub_2"  [label="(4, 64, 1)", style=solid];
-"600 sub_2" -> "601 ne_2"  [label="(4, 64, 64)", style=solid];
-"600 sub_2" -> "602 masked_fill_4"  [label="(4, 64, 64)", style=solid];
-"600 sub_2" -> "603 eq_2"  [label="(4, 64, 64)", style=solid];
-"601 ne_2" -> "602 masked_fill_4"  [label="(4, 64, 64)", style=solid];
-"602 masked_fill_4" -> "604 masked_fill_5"  [label="(4, 64, 64)", style=solid];
-"603 eq_2" -> "604 masked_fill_5"  [label="(4, 64, 64)", style=solid];
-"604 masked_fill_5" -> "606 unsqueeze_16"  [label="(4, 64, 64)", style=solid];
-"605 view_30" -> "608 add_18"  [label="(1, 4, 12, 64, 64)", style=solid];
-"606 unsqueeze_16" -> "607 unsqueeze_17"  [label="(4, 1, 64, 64)", style=solid];
-"607 unsqueeze_17" -> "608 add_18"  [label="(1, 4, 1, 64, 64)", style=solid];
-"608 add_18" -> "609 view_31"  [label="(1, 4, 12, 64, 64)", style=solid];
-"609 view_31" -> "610 softmax_5"  [label="(4, 12, 64, 64)", style=solid];
-"610 softmax_5" -> "611 dropout_20"  [label="(4, 12, 64, 64)", style=solid];
-"611 dropout_20" -> "612 matmul_11"  [label="(4, 12, 64, 64)", style=solid];
-"612 matmul_11" -> "613 transpose_11"  [label="(4, 12, 64, 32)", style=solid];
-"613 transpose_11" -> "614 reshape_25"  [label="(4, 64, 12, 32)", style=solid];
-"614 reshape_25" -> "615 reshape_25_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"615 reshape_25_0_0_nncf_smooth_quant_0" -> "616 quantize_per_tensor_default_36"  [label="(4, 64, 384)", style=solid];
-"616 quantize_per_tensor_default_36" -> "617 dequantize_per_tensor_default_36"  [label="(4, 64, 384)", style=solid];
-"617 dequantize_per_tensor_default_36" -> "624 linear_35"  [label="(4, 64, 384)", style=solid];
-"618 linear_35_scale_0" -> "621 quantize_per_channel_default_30"  [label="(384,)", style=solid];
-"618 linear_35_scale_0" -> "622 dequantize_per_channel_default_30"  [label="(384,)", style=solid];
-"619 linear_35_zero_point_0" -> "621 quantize_per_channel_default_30"  [label="(384,)", style=solid];
-"619 linear_35_zero_point_0" -> "622 dequantize_per_channel_default_30"  [label="(384,)", style=solid];
-"620 compressed_weight_updated_constant24" -> "621 quantize_per_channel_default_30"  [label="(384, 384)", style=solid];
-"621 quantize_per_channel_default_30" -> "622 dequantize_per_channel_default_30"  [label="(384, 384)", style=solid];
-"622 dequantize_per_channel_default_30" -> "624 linear_35"  [label="(384, 384)", style=solid];
-"623 _param_constant97_0_0" -> "624 linear_35"  [label="(384,)", style=solid];
-"624 linear_35" -> "625 dropout_21"  [label="(4, 64, 384)", style=solid];
-"625 dropout_21" -> "626 view_32"  [label="(4, 64, 384)", style=solid];
-"626 view_32" -> "627 permute_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"627 permute_27" -> "628 reshape_26"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"628 reshape_26" -> "629 roll_5"  [label="(1, 16, 16, 384)", style=solid];
-"629 roll_5" -> "630 slice_101"  [label="(1, 16, 16, 384)", style=solid];
-"630 slice_101" -> "631 slice_102"  [label="(1, 16, 16, 384)", style=solid];
-"631 slice_102" -> "632 slice_103"  [label="(1, 14, 16, 384)", style=solid];
-"632 slice_103" -> "633 slice_104"  [label="(1, 14, 14, 384)", style=solid];
-"633 slice_104" -> "634 contiguous_9"  [label="(1, 14, 14, 384)", style=solid];
-"634 contiguous_9" -> "637 layer_norm_13"  [label="(1, 14, 14, 384)", style=solid];
-"635 _param_constant98" -> "637 layer_norm_13"  [label="(384,)", style=solid];
-"636 _param_constant99" -> "637 layer_norm_13"  [label="(384,)", style=solid];
-"637 layer_norm_13" -> "638 add_19"  [label="(1, 14, 14, 384)", style=solid];
-"638 add_19" -> "639 add_19_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"638 add_19" -> "665 add_20"  [label="(1, 14, 14, 384)", style=solid];
-"639 add_19_0_0_nncf_smooth_quant_0" -> "640 quantize_per_tensor_default_37"  [label="(1, 14, 14, 384)", style=solid];
-"640 quantize_per_tensor_default_37" -> "641 dequantize_per_tensor_default_37"  [label="(1, 14, 14, 384)", style=solid];
-"641 dequantize_per_tensor_default_37" -> "648 linear_36"  [label="(1, 14, 14, 384)", style=solid];
-"642 linear_36_scale_0" -> "645 quantize_per_channel_default_31"  [label="(1536,)", style=solid];
-"642 linear_36_scale_0" -> "646 dequantize_per_channel_default_31"  [label="(1536,)", style=solid];
-"643 linear_36_zero_point_0" -> "645 quantize_per_channel_default_31"  [label="(1536,)", style=solid];
-"643 linear_36_zero_point_0" -> "646 dequantize_per_channel_default_31"  [label="(1536,)", style=solid];
-"644 compressed_weight_updated_constant25" -> "645 quantize_per_channel_default_31"  [label="(1536, 384)", style=solid];
-"645 quantize_per_channel_default_31" -> "646 dequantize_per_channel_default_31"  [label="(1536, 384)", style=solid];
-"646 dequantize_per_channel_default_31" -> "648 linear_36"  [label="(1536, 384)", style=solid];
-"647 _param_constant101_0_0" -> "648 linear_36"  [label="(1536,)", style=solid];
-"648 linear_36" -> "649 gelu_5"  [label="(1, 14, 14, 1536)", style=solid];
-"649 gelu_5" -> "650 dropout_22"  [label="(1, 14, 14, 1536)", style=solid];
-"650 dropout_22" -> "651 dropout_22_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"651 dropout_22_0_0_nncf_smooth_quant_0" -> "652 quantize_per_tensor_default_38"  [label="(1, 14, 14, 1536)", style=solid];
-"652 quantize_per_tensor_default_38" -> "653 dequantize_per_tensor_default_38"  [label="(1, 14, 14, 1536)", style=solid];
-"653 dequantize_per_tensor_default_38" -> "660 linear_37"  [label="(1, 14, 14, 1536)", style=solid];
-"654 linear_37_scale_0" -> "657 quantize_per_channel_default_32"  [label="(384,)", style=solid];
-"654 linear_37_scale_0" -> "658 dequantize_per_channel_default_32"  [label="(384,)", style=solid];
-"655 linear_37_zero_point_0" -> "657 quantize_per_channel_default_32"  [label="(384,)", style=solid];
-"655 linear_37_zero_point_0" -> "658 dequantize_per_channel_default_32"  [label="(384,)", style=solid];
-"656 compressed_weight_updated_constant26" -> "657 quantize_per_channel_default_32"  [label="(384, 1536)", style=solid];
-"657 quantize_per_channel_default_32" -> "658 dequantize_per_channel_default_32"  [label="(384, 1536)", style=solid];
-"658 dequantize_per_channel_default_32" -> "660 linear_37"  [label="(384, 1536)", style=solid];
-"659 _param_constant103_0_0" -> "660 linear_37"  [label="(384,)", style=solid];
-"660 linear_37" -> "661 dropout_23"  [label="(1, 14, 14, 384)", style=solid];
-"661 dropout_23" -> "664 layer_norm_14"  [label="(1, 14, 14, 384)", style=solid];
-"662 _param_constant104" -> "664 layer_norm_14"  [label="(384,)", style=solid];
-"663 _param_constant105" -> "664 layer_norm_14"  [label="(384,)", style=solid];
-"664 layer_norm_14" -> "665 add_20"  [label="(1, 14, 14, 384)", style=solid];
-"665 add_20" -> "667 pad_8"  [label="(1, 14, 14, 384)", style=solid];
-"665 add_20" -> "730 add_22"  [label="(1, 14, 14, 384)", style=solid];
-"666 _frozen_param51" -> "702 add_21"  [label="(1, 12, 64, 64)", style=solid];
-"667 pad_8" -> "668 view_35"  [label="(1, 16, 16, 384)", style=solid];
-"668 view_35" -> "669 permute_29"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"669 permute_29" -> "670 reshape_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"670 reshape_27" -> "671 reshape_27_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"671 reshape_27_0_0_nncf_smooth_quant_0" -> "672 quantize_per_tensor_default_39"  [label="(4, 64, 384)", style=solid];
-"672 quantize_per_tensor_default_39" -> "673 dequantize_per_tensor_default_39"  [label="(4, 64, 384)", style=solid];
-"673 dequantize_per_tensor_default_39" -> "680 linear_40"  [label="(4, 64, 384)", style=solid];
-"674 linear_40_scale_0" -> "677 quantize_per_channel_default_34"  [label="(1152,)", style=solid];
-"674 linear_40_scale_0" -> "678 dequantize_per_channel_default_34"  [label="(1152,)", style=solid];
-"675 linear_40_zero_point_0" -> "677 quantize_per_channel_default_34"  [label="(1152,)", style=solid];
-"675 linear_40_zero_point_0" -> "678 dequantize_per_channel_default_34"  [label="(1152,)", style=solid];
-"676 compressed_weight_updated_constant27" -> "677 quantize_per_channel_default_34"  [label="(1152, 384)", style=solid];
-"677 quantize_per_channel_default_34" -> "678 dequantize_per_channel_default_34"  [label="(1152, 384)", style=solid];
-"678 dequantize_per_channel_default_34" -> "680 linear_40"  [label="(1152, 384)", style=solid];
-"679 _frozen_param52_0_0" -> "680 linear_40"  [label="(1152,)", style=solid];
-"680 linear_40" -> "681 reshape_28"  [label="(4, 64, 1152)", style=solid];
-"681 reshape_28" -> "682 permute_30"  [label="(4, 64, 3, 12, 32)", style=solid];
-"682 permute_30" -> "683 select_18"  [label="(3, 4, 12, 64, 32)", style=solid];
-"682 permute_30" -> "684 select_19"  [label="(3, 4, 12, 64, 32)", style=solid];
-"682 permute_30" -> "685 select_20"  [label="(3, 4, 12, 64, 32)", style=solid];
-"683 select_18" -> "686 linalg_vector_norm_12"  [label="(4, 12, 64, 32)", style=solid];
-"683 select_18" -> "688 expand_as_12"  [label="(4, 12, 64, 32)", style=solid];
-"683 select_18" -> "689 div_12"  [label="(4, 12, 64, 32)", style=solid];
-"684 select_19" -> "692 linalg_vector_norm_13"  [label="(4, 12, 64, 32)", style=solid];
-"684 select_19" -> "694 expand_as_13"  [label="(4, 12, 64, 32)", style=solid];
-"684 select_19" -> "695 div_13"  [label="(4, 12, 64, 32)", style=solid];
-"685 select_20" -> "705 matmul_13"  [label="(4, 12, 64, 32)", style=solid];
-"686 linalg_vector_norm_12" -> "687 clamp_min_12"  [label="(4, 12, 64, 1)", style=solid];
-"687 clamp_min_12" -> "688 expand_as_12"  [label="(4, 12, 64, 1)", style=solid];
-"688 expand_as_12" -> "689 div_12"  [label="(4, 12, 64, 32)", style=solid];
-"689 div_12" -> "690 quantize_per_tensor_default_40"  [label="(4, 12, 64, 32)", style=solid];
-"690 quantize_per_tensor_default_40" -> "691 dequantize_per_tensor_default_40"  [label="(4, 12, 64, 32)", style=solid];
-"691 dequantize_per_tensor_default_40" -> "699 matmul_12"  [label="(4, 12, 64, 32)", style=solid];
-"692 linalg_vector_norm_13" -> "693 clamp_min_13"  [label="(4, 12, 64, 1)", style=solid];
-"693 clamp_min_13" -> "694 expand_as_13"  [label="(4, 12, 64, 1)", style=solid];
-"694 expand_as_13" -> "695 div_13"  [label="(4, 12, 64, 32)", style=solid];
-"695 div_13" -> "696 quantize_per_tensor_default_41"  [label="(4, 12, 64, 32)", style=solid];
-"696 quantize_per_tensor_default_41" -> "697 dequantize_per_tensor_default_41"  [label="(4, 12, 64, 32)", style=solid];
-"697 dequantize_per_tensor_default_41" -> "698 transpose_12"  [label="(4, 12, 64, 32)", style=solid];
-"698 transpose_12" -> "699 matmul_12"  [label="(4, 12, 32, 64)", style=solid];
-"699 matmul_12" -> "701 mul_13"  [label="(4, 12, 64, 64)", style=solid];
-"700 _frozen_param54" -> "701 mul_13"  [label="(12, 1, 1)", style=solid];
-"701 mul_13" -> "702 add_21"  [label="(4, 12, 64, 64)", style=solid];
-"702 add_21" -> "703 softmax_6"  [label="(4, 12, 64, 64)", style=solid];
-"703 softmax_6" -> "704 dropout_24"  [label="(4, 12, 64, 64)", style=solid];
-"704 dropout_24" -> "705 matmul_13"  [label="(4, 12, 64, 64)", style=solid];
-"705 matmul_13" -> "706 transpose_13"  [label="(4, 12, 64, 32)", style=solid];
-"706 transpose_13" -> "707 reshape_29"  [label="(4, 64, 12, 32)", style=solid];
-"707 reshape_29" -> "708 reshape_29_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"708 reshape_29_0_0_nncf_smooth_quant_0" -> "709 quantize_per_tensor_default_42"  [label="(4, 64, 384)", style=solid];
-"709 quantize_per_tensor_default_42" -> "710 dequantize_per_tensor_default_42"  [label="(4, 64, 384)", style=solid];
-"710 dequantize_per_tensor_default_42" -> "717 linear_41"  [label="(4, 64, 384)", style=solid];
-"711 linear_41_scale_0" -> "714 quantize_per_channel_default_35"  [label="(384,)", style=solid];
-"711 linear_41_scale_0" -> "715 dequantize_per_channel_default_35"  [label="(384,)", style=solid];
-"712 linear_41_zero_point_0" -> "714 quantize_per_channel_default_35"  [label="(384,)", style=solid];
-"712 linear_41_zero_point_0" -> "715 dequantize_per_channel_default_35"  [label="(384,)", style=solid];
-"713 compressed_weight_updated_constant28" -> "714 quantize_per_channel_default_35"  [label="(384, 384)", style=solid];
-"714 quantize_per_channel_default_35" -> "715 dequantize_per_channel_default_35"  [label="(384, 384)", style=solid];
-"715 dequantize_per_channel_default_35" -> "717 linear_41"  [label="(384, 384)", style=solid];
-"716 _param_constant113_0_0" -> "717 linear_41"  [label="(384,)", style=solid];
-"717 linear_41" -> "718 dropout_25"  [label="(4, 64, 384)", style=solid];
-"718 dropout_25" -> "719 view_36"  [label="(4, 64, 384)", style=solid];
-"719 view_36" -> "720 permute_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"720 permute_31" -> "721 reshape_30"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"721 reshape_30" -> "722 slice_106"  [label="(1, 16, 16, 384)", style=solid];
-"722 slice_106" -> "723 slice_107"  [label="(1, 16, 16, 384)", style=solid];
-"723 slice_107" -> "724 slice_108"  [label="(1, 14, 16, 384)", style=solid];
-"724 slice_108" -> "725 slice_109"  [label="(1, 14, 14, 384)", style=solid];
-"725 slice_109" -> "726 contiguous_11"  [label="(1, 14, 14, 384)", style=solid];
-"726 contiguous_11" -> "729 layer_norm_15"  [label="(1, 14, 14, 384)", style=solid];
-"727 _param_constant114" -> "729 layer_norm_15"  [label="(384,)", style=solid];
-"728 _param_constant115" -> "729 layer_norm_15"  [label="(384,)", style=solid];
-"729 layer_norm_15" -> "730 add_22"  [label="(1, 14, 14, 384)", style=solid];
-"730 add_22" -> "731 add_22_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"730 add_22" -> "757 add_23"  [label="(1, 14, 14, 384)", style=solid];
-"731 add_22_0_0_nncf_smooth_quant_0" -> "732 quantize_per_tensor_default_43"  [label="(1, 14, 14, 384)", style=solid];
-"732 quantize_per_tensor_default_43" -> "733 dequantize_per_tensor_default_43"  [label="(1, 14, 14, 384)", style=solid];
-"733 dequantize_per_tensor_default_43" -> "740 linear_42"  [label="(1, 14, 14, 384)", style=solid];
-"734 linear_42_scale_0" -> "737 quantize_per_channel_default_36"  [label="(1536,)", style=solid];
-"734 linear_42_scale_0" -> "738 dequantize_per_channel_default_36"  [label="(1536,)", style=solid];
-"735 linear_42_zero_point_0" -> "737 quantize_per_channel_default_36"  [label="(1536,)", style=solid];
-"735 linear_42_zero_point_0" -> "738 dequantize_per_channel_default_36"  [label="(1536,)", style=solid];
-"736 compressed_weight_updated_constant29" -> "737 quantize_per_channel_default_36"  [label="(1536, 384)", style=solid];
-"737 quantize_per_channel_default_36" -> "738 dequantize_per_channel_default_36"  [label="(1536, 384)", style=solid];
-"738 dequantize_per_channel_default_36" -> "740 linear_42"  [label="(1536, 384)", style=solid];
-"739 _param_constant117_0_0" -> "740 linear_42"  [label="(1536,)", style=solid];
-"740 linear_42" -> "741 gelu_6"  [label="(1, 14, 14, 1536)", style=solid];
-"741 gelu_6" -> "742 dropout_26"  [label="(1, 14, 14, 1536)", style=solid];
-"742 dropout_26" -> "743 dropout_26_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"743 dropout_26_0_0_nncf_smooth_quant_0" -> "744 quantize_per_tensor_default_44"  [label="(1, 14, 14, 1536)", style=solid];
-"744 quantize_per_tensor_default_44" -> "745 dequantize_per_tensor_default_44"  [label="(1, 14, 14, 1536)", style=solid];
-"745 dequantize_per_tensor_default_44" -> "752 linear_43"  [label="(1, 14, 14, 1536)", style=solid];
-"746 linear_43_scale_0" -> "749 quantize_per_channel_default_37"  [label="(384,)", style=solid];
-"746 linear_43_scale_0" -> "750 dequantize_per_channel_default_37"  [label="(384,)", style=solid];
-"747 linear_43_zero_point_0" -> "749 quantize_per_channel_default_37"  [label="(384,)", style=solid];
-"747 linear_43_zero_point_0" -> "750 dequantize_per_channel_default_37"  [label="(384,)", style=solid];
-"748 compressed_weight_updated_constant30" -> "749 quantize_per_channel_default_37"  [label="(384, 1536)", style=solid];
-"749 quantize_per_channel_default_37" -> "750 dequantize_per_channel_default_37"  [label="(384, 1536)", style=solid];
-"750 dequantize_per_channel_default_37" -> "752 linear_43"  [label="(384, 1536)", style=solid];
-"751 _param_constant119_0_0" -> "752 linear_43"  [label="(384,)", style=solid];
-"752 linear_43" -> "753 dropout_27"  [label="(1, 14, 14, 384)", style=solid];
-"753 dropout_27" -> "756 layer_norm_16"  [label="(1, 14, 14, 384)", style=solid];
-"754 _param_constant120" -> "756 layer_norm_16"  [label="(384,)", style=solid];
-"755 _param_constant121" -> "756 layer_norm_16"  [label="(384,)", style=solid];
-"756 layer_norm_16" -> "757 add_23"  [label="(1, 14, 14, 384)", style=solid];
-"757 add_23" -> "759 pad_9"  [label="(1, 14, 14, 384)", style=solid];
-"757 add_23" -> "840 add_26"  [label="(1, 14, 14, 384)", style=solid];
-"758 _frozen_param55" -> "795 add_24"  [label="(1, 12, 64, 64)", style=solid];
-"759 pad_9" -> "760 roll_6"  [label="(1, 16, 16, 384)", style=solid];
-"760 roll_6" -> "761 view_39"  [label="(1, 16, 16, 384)", style=solid];
-"761 view_39" -> "762 permute_33"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"762 permute_33" -> "763 reshape_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"763 reshape_31" -> "764 reshape_31_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"763 reshape_31" -> "796 new_zeros_3"  [label="(4, 64, 384)", style=solid];
-"764 reshape_31_0_0_nncf_smooth_quant_0" -> "765 quantize_per_tensor_default_45"  [label="(4, 64, 384)", style=solid];
-"765 quantize_per_tensor_default_45" -> "766 dequantize_per_tensor_default_45"  [label="(4, 64, 384)", style=solid];
-"766 dequantize_per_tensor_default_45" -> "773 linear_46"  [label="(4, 64, 384)", style=solid];
-"767 linear_46_scale_0" -> "770 quantize_per_channel_default_39"  [label="(1152,)", style=solid];
-"767 linear_46_scale_0" -> "771 dequantize_per_channel_default_39"  [label="(1152,)", style=solid];
-"768 linear_46_zero_point_0" -> "770 quantize_per_channel_default_39"  [label="(1152,)", style=solid];
-"768 linear_46_zero_point_0" -> "771 dequantize_per_channel_default_39"  [label="(1152,)", style=solid];
-"769 compressed_weight_updated_constant31" -> "770 quantize_per_channel_default_39"  [label="(1152, 384)", style=solid];
-"770 quantize_per_channel_default_39" -> "771 dequantize_per_channel_default_39"  [label="(1152, 384)", style=solid];
-"771 dequantize_per_channel_default_39" -> "773 linear_46"  [label="(1152, 384)", style=solid];
-"772 _frozen_param56_0_0" -> "773 linear_46"  [label="(1152,)", style=solid];
-"773 linear_46" -> "774 reshape_32"  [label="(4, 64, 1152)", style=solid];
-"774 reshape_32" -> "775 permute_34"  [label="(4, 64, 3, 12, 32)", style=solid];
-"775 permute_34" -> "776 select_21"  [label="(3, 4, 12, 64, 32)", style=solid];
-"775 permute_34" -> "777 select_22"  [label="(3, 4, 12, 64, 32)", style=solid];
-"775 permute_34" -> "778 select_23"  [label="(3, 4, 12, 64, 32)", style=solid];
-"776 select_21" -> "779 linalg_vector_norm_14"  [label="(4, 12, 64, 32)", style=solid];
-"776 select_21" -> "781 expand_as_14"  [label="(4, 12, 64, 32)", style=solid];
-"776 select_21" -> "782 div_14"  [label="(4, 12, 64, 32)", style=solid];
-"777 select_22" -> "785 linalg_vector_norm_15"  [label="(4, 12, 64, 32)", style=solid];
-"777 select_22" -> "787 expand_as_15"  [label="(4, 12, 64, 32)", style=solid];
-"777 select_22" -> "788 div_15"  [label="(4, 12, 64, 32)", style=solid];
-"778 select_23" -> "814 matmul_15"  [label="(4, 12, 64, 32)", style=solid];
-"779 linalg_vector_norm_14" -> "780 clamp_min_14"  [label="(4, 12, 64, 1)", style=solid];
-"780 clamp_min_14" -> "781 expand_as_14"  [label="(4, 12, 64, 1)", style=solid];
-"781 expand_as_14" -> "782 div_14"  [label="(4, 12, 64, 32)", style=solid];
-"782 div_14" -> "783 quantize_per_tensor_default_46"  [label="(4, 12, 64, 32)", style=solid];
-"783 quantize_per_tensor_default_46" -> "784 dequantize_per_tensor_default_46"  [label="(4, 12, 64, 32)", style=solid];
-"784 dequantize_per_tensor_default_46" -> "792 matmul_14"  [label="(4, 12, 64, 32)", style=solid];
-"785 linalg_vector_norm_15" -> "786 clamp_min_15"  [label="(4, 12, 64, 1)", style=solid];
-"786 clamp_min_15" -> "787 expand_as_15"  [label="(4, 12, 64, 1)", style=solid];
-"787 expand_as_15" -> "788 div_15"  [label="(4, 12, 64, 32)", style=solid];
-"788 div_15" -> "789 quantize_per_tensor_default_47"  [label="(4, 12, 64, 32)", style=solid];
-"789 quantize_per_tensor_default_47" -> "790 dequantize_per_tensor_default_47"  [label="(4, 12, 64, 32)", style=solid];
-"790 dequantize_per_tensor_default_47" -> "791 transpose_14"  [label="(4, 12, 64, 32)", style=solid];
-"791 transpose_14" -> "792 matmul_14"  [label="(4, 12, 32, 64)", style=solid];
-"792 matmul_14" -> "794 mul_15"  [label="(4, 12, 64, 64)", style=solid];
-"793 _frozen_param58" -> "794 mul_15"  [label="(12, 1, 1)", style=solid];
-"794 mul_15" -> "795 add_24"  [label="(4, 12, 64, 64)", style=solid];
-"795 add_24" -> "807 view_41"  [label="(4, 12, 64, 64)", style=solid];
-"796 new_zeros_3" -> "797 view_40"  [label="(16, 16)", style=solid];
-"797 view_40" -> "798 permute_35"  [label="(2, 8, 2, 8)", style=solid];
-"798 permute_35" -> "799 reshape_33"  [label="(2, 2, 8, 8)", style=solid];
-"799 reshape_33" -> "800 unsqueeze_20"  [label="(4, 64)", style=solid];
-"799 reshape_33" -> "801 unsqueeze_21"  [label="(4, 64)", style=solid];
-"800 unsqueeze_20" -> "802 sub_3"  [label="(4, 1, 64)", style=solid];
-"801 unsqueeze_21" -> "802 sub_3"  [label="(4, 64, 1)", style=solid];
-"802 sub_3" -> "803 ne_3"  [label="(4, 64, 64)", style=solid];
-"802 sub_3" -> "804 masked_fill_6"  [label="(4, 64, 64)", style=solid];
-"802 sub_3" -> "805 eq_3"  [label="(4, 64, 64)", style=solid];
-"803 ne_3" -> "804 masked_fill_6"  [label="(4, 64, 64)", style=solid];
-"804 masked_fill_6" -> "806 masked_fill_7"  [label="(4, 64, 64)", style=solid];
-"805 eq_3" -> "806 masked_fill_7"  [label="(4, 64, 64)", style=solid];
-"806 masked_fill_7" -> "808 unsqueeze_22"  [label="(4, 64, 64)", style=solid];
-"807 view_41" -> "810 add_25"  [label="(1, 4, 12, 64, 64)", style=solid];
-"808 unsqueeze_22" -> "809 unsqueeze_23"  [label="(4, 1, 64, 64)", style=solid];
-"809 unsqueeze_23" -> "810 add_25"  [label="(1, 4, 1, 64, 64)", style=solid];
-"810 add_25" -> "811 view_42"  [label="(1, 4, 12, 64, 64)", style=solid];
-"811 view_42" -> "812 softmax_7"  [label="(4, 12, 64, 64)", style=solid];
-"812 softmax_7" -> "813 dropout_28"  [label="(4, 12, 64, 64)", style=solid];
-"813 dropout_28" -> "814 matmul_15"  [label="(4, 12, 64, 64)", style=solid];
-"814 matmul_15" -> "815 transpose_15"  [label="(4, 12, 64, 32)", style=solid];
-"815 transpose_15" -> "816 reshape_34"  [label="(4, 64, 12, 32)", style=solid];
-"816 reshape_34" -> "817 reshape_34_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"817 reshape_34_0_0_nncf_smooth_quant_0" -> "818 quantize_per_tensor_default_48"  [label="(4, 64, 384)", style=solid];
-"818 quantize_per_tensor_default_48" -> "819 dequantize_per_tensor_default_48"  [label="(4, 64, 384)", style=solid];
-"819 dequantize_per_tensor_default_48" -> "826 linear_47"  [label="(4, 64, 384)", style=solid];
-"820 linear_47_scale_0" -> "823 quantize_per_channel_default_40"  [label="(384,)", style=solid];
-"820 linear_47_scale_0" -> "824 dequantize_per_channel_default_40"  [label="(384,)", style=solid];
-"821 linear_47_zero_point_0" -> "823 quantize_per_channel_default_40"  [label="(384,)", style=solid];
-"821 linear_47_zero_point_0" -> "824 dequantize_per_channel_default_40"  [label="(384,)", style=solid];
-"822 compressed_weight_updated_constant32" -> "823 quantize_per_channel_default_40"  [label="(384, 384)", style=solid];
-"823 quantize_per_channel_default_40" -> "824 dequantize_per_channel_default_40"  [label="(384, 384)", style=solid];
-"824 dequantize_per_channel_default_40" -> "826 linear_47"  [label="(384, 384)", style=solid];
-"825 _param_constant129_0_0" -> "826 linear_47"  [label="(384,)", style=solid];
-"826 linear_47" -> "827 dropout_29"  [label="(4, 64, 384)", style=solid];
-"827 dropout_29" -> "828 view_43"  [label="(4, 64, 384)", style=solid];
-"828 view_43" -> "829 permute_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"829 permute_36" -> "830 reshape_35"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"830 reshape_35" -> "831 roll_7"  [label="(1, 16, 16, 384)", style=solid];
-"831 roll_7" -> "832 slice_129"  [label="(1, 16, 16, 384)", style=solid];
-"832 slice_129" -> "833 slice_130"  [label="(1, 16, 16, 384)", style=solid];
-"833 slice_130" -> "834 slice_131"  [label="(1, 14, 16, 384)", style=solid];
-"834 slice_131" -> "835 slice_132"  [label="(1, 14, 14, 384)", style=solid];
-"835 slice_132" -> "836 contiguous_13"  [label="(1, 14, 14, 384)", style=solid];
-"836 contiguous_13" -> "839 layer_norm_17"  [label="(1, 14, 14, 384)", style=solid];
-"837 _param_constant130" -> "839 layer_norm_17"  [label="(384,)", style=solid];
-"838 _param_constant131" -> "839 layer_norm_17"  [label="(384,)", style=solid];
-"839 layer_norm_17" -> "840 add_26"  [label="(1, 14, 14, 384)", style=solid];
-"840 add_26" -> "841 add_26_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"840 add_26" -> "867 add_27"  [label="(1, 14, 14, 384)", style=solid];
-"841 add_26_0_0_nncf_smooth_quant_0" -> "842 quantize_per_tensor_default_49"  [label="(1, 14, 14, 384)", style=solid];
-"842 quantize_per_tensor_default_49" -> "843 dequantize_per_tensor_default_49"  [label="(1, 14, 14, 384)", style=solid];
-"843 dequantize_per_tensor_default_49" -> "850 linear_48"  [label="(1, 14, 14, 384)", style=solid];
-"844 linear_48_scale_0" -> "847 quantize_per_channel_default_41"  [label="(1536,)", style=solid];
-"844 linear_48_scale_0" -> "848 dequantize_per_channel_default_41"  [label="(1536,)", style=solid];
-"845 linear_48_zero_point_0" -> "847 quantize_per_channel_default_41"  [label="(1536,)", style=solid];
-"845 linear_48_zero_point_0" -> "848 dequantize_per_channel_default_41"  [label="(1536,)", style=solid];
-"846 compressed_weight_updated_constant33" -> "847 quantize_per_channel_default_41"  [label="(1536, 384)", style=solid];
-"847 quantize_per_channel_default_41" -> "848 dequantize_per_channel_default_41"  [label="(1536, 384)", style=solid];
-"848 dequantize_per_channel_default_41" -> "850 linear_48"  [label="(1536, 384)", style=solid];
-"849 _param_constant133_0_0" -> "850 linear_48"  [label="(1536,)", style=solid];
-"850 linear_48" -> "851 gelu_7"  [label="(1, 14, 14, 1536)", style=solid];
-"851 gelu_7" -> "852 dropout_30"  [label="(1, 14, 14, 1536)", style=solid];
-"852 dropout_30" -> "853 dropout_30_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"853 dropout_30_0_0_nncf_smooth_quant_0" -> "854 quantize_per_tensor_default_50"  [label="(1, 14, 14, 1536)", style=solid];
-"854 quantize_per_tensor_default_50" -> "855 dequantize_per_tensor_default_50"  [label="(1, 14, 14, 1536)", style=solid];
-"855 dequantize_per_tensor_default_50" -> "862 linear_49"  [label="(1, 14, 14, 1536)", style=solid];
-"856 linear_49_scale_0" -> "859 quantize_per_channel_default_42"  [label="(384,)", style=solid];
-"856 linear_49_scale_0" -> "860 dequantize_per_channel_default_42"  [label="(384,)", style=solid];
-"857 linear_49_zero_point_0" -> "859 quantize_per_channel_default_42"  [label="(384,)", style=solid];
-"857 linear_49_zero_point_0" -> "860 dequantize_per_channel_default_42"  [label="(384,)", style=solid];
-"858 compressed_weight_updated_constant34" -> "859 quantize_per_channel_default_42"  [label="(384, 1536)", style=solid];
-"859 quantize_per_channel_default_42" -> "860 dequantize_per_channel_default_42"  [label="(384, 1536)", style=solid];
-"860 dequantize_per_channel_default_42" -> "862 linear_49"  [label="(384, 1536)", style=solid];
-"861 _param_constant135_0_0" -> "862 linear_49"  [label="(384,)", style=solid];
-"862 linear_49" -> "863 dropout_31"  [label="(1, 14, 14, 384)", style=solid];
-"863 dropout_31" -> "866 layer_norm_18"  [label="(1, 14, 14, 384)", style=solid];
-"864 _param_constant136" -> "866 layer_norm_18"  [label="(384,)", style=solid];
-"865 _param_constant137" -> "866 layer_norm_18"  [label="(384,)", style=solid];
-"866 layer_norm_18" -> "867 add_27"  [label="(1, 14, 14, 384)", style=solid];
-"867 add_27" -> "869 pad_10"  [label="(1, 14, 14, 384)", style=solid];
-"867 add_27" -> "932 add_29"  [label="(1, 14, 14, 384)", style=solid];
-"868 _frozen_param68" -> "904 add_28"  [label="(1, 12, 64, 64)", style=solid];
-"869 pad_10" -> "870 view_46"  [label="(1, 16, 16, 384)", style=solid];
-"870 view_46" -> "871 permute_38"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"871 permute_38" -> "872 reshape_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"872 reshape_36" -> "873 reshape_36_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"873 reshape_36_0_0_nncf_smooth_quant_0" -> "874 quantize_per_tensor_default_51"  [label="(4, 64, 384)", style=solid];
-"874 quantize_per_tensor_default_51" -> "875 dequantize_per_tensor_default_51"  [label="(4, 64, 384)", style=solid];
-"875 dequantize_per_tensor_default_51" -> "882 linear_52"  [label="(4, 64, 384)", style=solid];
-"876 linear_52_scale_0" -> "879 quantize_per_channel_default_44"  [label="(1152,)", style=solid];
-"876 linear_52_scale_0" -> "880 dequantize_per_channel_default_44"  [label="(1152,)", style=solid];
-"877 linear_52_zero_point_0" -> "879 quantize_per_channel_default_44"  [label="(1152,)", style=solid];
-"877 linear_52_zero_point_0" -> "880 dequantize_per_channel_default_44"  [label="(1152,)", style=solid];
-"878 compressed_weight_updated_constant35" -> "879 quantize_per_channel_default_44"  [label="(1152, 384)", style=solid];
-"879 quantize_per_channel_default_44" -> "880 dequantize_per_channel_default_44"  [label="(1152, 384)", style=solid];
-"880 dequantize_per_channel_default_44" -> "882 linear_52"  [label="(1152, 384)", style=solid];
-"881 _frozen_param69_0_0" -> "882 linear_52"  [label="(1152,)", style=solid];
-"882 linear_52" -> "883 reshape_37"  [label="(4, 64, 1152)", style=solid];
-"883 reshape_37" -> "884 permute_39"  [label="(4, 64, 3, 12, 32)", style=solid];
-"884 permute_39" -> "885 select_24"  [label="(3, 4, 12, 64, 32)", style=solid];
-"884 permute_39" -> "886 select_25"  [label="(3, 4, 12, 64, 32)", style=solid];
-"884 permute_39" -> "887 select_26"  [label="(3, 4, 12, 64, 32)", style=solid];
-"885 select_24" -> "888 linalg_vector_norm_16"  [label="(4, 12, 64, 32)", style=solid];
-"885 select_24" -> "890 expand_as_16"  [label="(4, 12, 64, 32)", style=solid];
-"885 select_24" -> "891 div_16"  [label="(4, 12, 64, 32)", style=solid];
-"886 select_25" -> "894 linalg_vector_norm_17"  [label="(4, 12, 64, 32)", style=solid];
-"886 select_25" -> "896 expand_as_17"  [label="(4, 12, 64, 32)", style=solid];
-"886 select_25" -> "897 div_17"  [label="(4, 12, 64, 32)", style=solid];
-"887 select_26" -> "907 matmul_17"  [label="(4, 12, 64, 32)", style=solid];
-"888 linalg_vector_norm_16" -> "889 clamp_min_16"  [label="(4, 12, 64, 1)", style=solid];
-"889 clamp_min_16" -> "890 expand_as_16"  [label="(4, 12, 64, 1)", style=solid];
-"890 expand_as_16" -> "891 div_16"  [label="(4, 12, 64, 32)", style=solid];
-"891 div_16" -> "892 quantize_per_tensor_default_52"  [label="(4, 12, 64, 32)", style=solid];
-"892 quantize_per_tensor_default_52" -> "893 dequantize_per_tensor_default_52"  [label="(4, 12, 64, 32)", style=solid];
-"893 dequantize_per_tensor_default_52" -> "901 matmul_16"  [label="(4, 12, 64, 32)", style=solid];
-"894 linalg_vector_norm_17" -> "895 clamp_min_17"  [label="(4, 12, 64, 1)", style=solid];
-"895 clamp_min_17" -> "896 expand_as_17"  [label="(4, 12, 64, 1)", style=solid];
-"896 expand_as_17" -> "897 div_17"  [label="(4, 12, 64, 32)", style=solid];
-"897 div_17" -> "898 quantize_per_tensor_default_53"  [label="(4, 12, 64, 32)", style=solid];
-"898 quantize_per_tensor_default_53" -> "899 dequantize_per_tensor_default_53"  [label="(4, 12, 64, 32)", style=solid];
-"899 dequantize_per_tensor_default_53" -> "900 transpose_16"  [label="(4, 12, 64, 32)", style=solid];
-"900 transpose_16" -> "901 matmul_16"  [label="(4, 12, 32, 64)", style=solid];
-"901 matmul_16" -> "903 mul_17"  [label="(4, 12, 64, 64)", style=solid];
-"902 _frozen_param71" -> "903 mul_17"  [label="(12, 1, 1)", style=solid];
-"903 mul_17" -> "904 add_28"  [label="(4, 12, 64, 64)", style=solid];
-"904 add_28" -> "905 softmax_8"  [label="(4, 12, 64, 64)", style=solid];
-"905 softmax_8" -> "906 dropout_32"  [label="(4, 12, 64, 64)", style=solid];
-"906 dropout_32" -> "907 matmul_17"  [label="(4, 12, 64, 64)", style=solid];
-"907 matmul_17" -> "908 transpose_17"  [label="(4, 12, 64, 32)", style=solid];
-"908 transpose_17" -> "909 reshape_38"  [label="(4, 64, 12, 32)", style=solid];
-"909 reshape_38" -> "910 reshape_38_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"910 reshape_38_0_0_nncf_smooth_quant_0" -> "911 quantize_per_tensor_default_54"  [label="(4, 64, 384)", style=solid];
-"911 quantize_per_tensor_default_54" -> "912 dequantize_per_tensor_default_54"  [label="(4, 64, 384)", style=solid];
-"912 dequantize_per_tensor_default_54" -> "919 linear_53"  [label="(4, 64, 384)", style=solid];
-"913 linear_53_scale_0" -> "916 quantize_per_channel_default_45"  [label="(384,)", style=solid];
-"913 linear_53_scale_0" -> "917 dequantize_per_channel_default_45"  [label="(384,)", style=solid];
-"914 linear_53_zero_point_0" -> "916 quantize_per_channel_default_45"  [label="(384,)", style=solid];
-"914 linear_53_zero_point_0" -> "917 dequantize_per_channel_default_45"  [label="(384,)", style=solid];
-"915 compressed_weight_updated_constant36" -> "916 quantize_per_channel_default_45"  [label="(384, 384)", style=solid];
-"916 quantize_per_channel_default_45" -> "917 dequantize_per_channel_default_45"  [label="(384, 384)", style=solid];
-"917 dequantize_per_channel_default_45" -> "919 linear_53"  [label="(384, 384)", style=solid];
-"918 _param_constant145_0_0" -> "919 linear_53"  [label="(384,)", style=solid];
-"919 linear_53" -> "920 dropout_33"  [label="(4, 64, 384)", style=solid];
-"920 dropout_33" -> "921 view_47"  [label="(4, 64, 384)", style=solid];
-"921 view_47" -> "922 permute_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"922 permute_40" -> "923 reshape_39"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"923 reshape_39" -> "924 slice_134"  [label="(1, 16, 16, 384)", style=solid];
-"924 slice_134" -> "925 slice_135"  [label="(1, 16, 16, 384)", style=solid];
-"925 slice_135" -> "926 slice_136"  [label="(1, 14, 16, 384)", style=solid];
-"926 slice_136" -> "927 slice_137"  [label="(1, 14, 14, 384)", style=solid];
-"927 slice_137" -> "928 contiguous_15"  [label="(1, 14, 14, 384)", style=solid];
-"928 contiguous_15" -> "931 layer_norm_19"  [label="(1, 14, 14, 384)", style=solid];
-"929 _param_constant146" -> "931 layer_norm_19"  [label="(384,)", style=solid];
-"930 _param_constant147" -> "931 layer_norm_19"  [label="(384,)", style=solid];
-"931 layer_norm_19" -> "932 add_29"  [label="(1, 14, 14, 384)", style=solid];
-"932 add_29" -> "933 add_29_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"932 add_29" -> "959 add_30"  [label="(1, 14, 14, 384)", style=solid];
-"933 add_29_0_0_nncf_smooth_quant_0" -> "934 quantize_per_tensor_default_55"  [label="(1, 14, 14, 384)", style=solid];
-"934 quantize_per_tensor_default_55" -> "935 dequantize_per_tensor_default_55"  [label="(1, 14, 14, 384)", style=solid];
-"935 dequantize_per_tensor_default_55" -> "942 linear_54"  [label="(1, 14, 14, 384)", style=solid];
-"936 linear_54_scale_0" -> "939 quantize_per_channel_default_46"  [label="(1536,)", style=solid];
-"936 linear_54_scale_0" -> "940 dequantize_per_channel_default_46"  [label="(1536,)", style=solid];
-"937 linear_54_zero_point_0" -> "939 quantize_per_channel_default_46"  [label="(1536,)", style=solid];
-"937 linear_54_zero_point_0" -> "940 dequantize_per_channel_default_46"  [label="(1536,)", style=solid];
-"938 compressed_weight_updated_constant37" -> "939 quantize_per_channel_default_46"  [label="(1536, 384)", style=solid];
-"939 quantize_per_channel_default_46" -> "940 dequantize_per_channel_default_46"  [label="(1536, 384)", style=solid];
-"940 dequantize_per_channel_default_46" -> "942 linear_54"  [label="(1536, 384)", style=solid];
-"941 _param_constant149_0_0" -> "942 linear_54"  [label="(1536,)", style=solid];
-"942 linear_54" -> "943 gelu_8"  [label="(1, 14, 14, 1536)", style=solid];
-"943 gelu_8" -> "944 dropout_34"  [label="(1, 14, 14, 1536)", style=solid];
-"944 dropout_34" -> "945 dropout_34_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"945 dropout_34_0_0_nncf_smooth_quant_0" -> "946 quantize_per_tensor_default_56"  [label="(1, 14, 14, 1536)", style=solid];
-"946 quantize_per_tensor_default_56" -> "947 dequantize_per_tensor_default_56"  [label="(1, 14, 14, 1536)", style=solid];
-"947 dequantize_per_tensor_default_56" -> "954 linear_55"  [label="(1, 14, 14, 1536)", style=solid];
-"948 linear_55_scale_0" -> "951 quantize_per_channel_default_47"  [label="(384,)", style=solid];
-"948 linear_55_scale_0" -> "952 dequantize_per_channel_default_47"  [label="(384,)", style=solid];
-"949 linear_55_zero_point_0" -> "951 quantize_per_channel_default_47"  [label="(384,)", style=solid];
-"949 linear_55_zero_point_0" -> "952 dequantize_per_channel_default_47"  [label="(384,)", style=solid];
-"950 compressed_weight_updated_constant38" -> "951 quantize_per_channel_default_47"  [label="(384, 1536)", style=solid];
-"951 quantize_per_channel_default_47" -> "952 dequantize_per_channel_default_47"  [label="(384, 1536)", style=solid];
-"952 dequantize_per_channel_default_47" -> "954 linear_55"  [label="(384, 1536)", style=solid];
-"953 _param_constant151_0_0" -> "954 linear_55"  [label="(384,)", style=solid];
-"954 linear_55" -> "955 dropout_35"  [label="(1, 14, 14, 384)", style=solid];
-"955 dropout_35" -> "958 layer_norm_20"  [label="(1, 14, 14, 384)", style=solid];
-"956 _param_constant152" -> "958 layer_norm_20"  [label="(384,)", style=solid];
-"957 _param_constant153" -> "958 layer_norm_20"  [label="(384,)", style=solid];
-"958 layer_norm_20" -> "959 add_30"  [label="(1, 14, 14, 384)", style=solid];
-"959 add_30" -> "961 pad_11"  [label="(1, 14, 14, 384)", style=solid];
-"959 add_30" -> "1042 add_33"  [label="(1, 14, 14, 384)", style=solid];
-"960 _frozen_param72" -> "997 add_31"  [label="(1, 12, 64, 64)", style=solid];
-"961 pad_11" -> "962 roll_8"  [label="(1, 16, 16, 384)", style=solid];
-"962 roll_8" -> "963 view_50"  [label="(1, 16, 16, 384)", style=solid];
-"963 view_50" -> "964 permute_42"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"964 permute_42" -> "965 reshape_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"965 reshape_40" -> "966 reshape_40_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"965 reshape_40" -> "998 new_zeros_4"  [label="(4, 64, 384)", style=solid];
-"966 reshape_40_0_0_nncf_smooth_quant_0" -> "967 quantize_per_tensor_default_57"  [label="(4, 64, 384)", style=solid];
-"967 quantize_per_tensor_default_57" -> "968 dequantize_per_tensor_default_57"  [label="(4, 64, 384)", style=solid];
-"968 dequantize_per_tensor_default_57" -> "975 linear_58"  [label="(4, 64, 384)", style=solid];
-"969 linear_58_scale_0" -> "972 quantize_per_channel_default_49"  [label="(1152,)", style=solid];
-"969 linear_58_scale_0" -> "973 dequantize_per_channel_default_49"  [label="(1152,)", style=solid];
-"970 linear_58_zero_point_0" -> "972 quantize_per_channel_default_49"  [label="(1152,)", style=solid];
-"970 linear_58_zero_point_0" -> "973 dequantize_per_channel_default_49"  [label="(1152,)", style=solid];
-"971 compressed_weight_updated_constant39" -> "972 quantize_per_channel_default_49"  [label="(1152, 384)", style=solid];
-"972 quantize_per_channel_default_49" -> "973 dequantize_per_channel_default_49"  [label="(1152, 384)", style=solid];
-"973 dequantize_per_channel_default_49" -> "975 linear_58"  [label="(1152, 384)", style=solid];
-"974 _frozen_param73_0_0" -> "975 linear_58"  [label="(1152,)", style=solid];
-"975 linear_58" -> "976 reshape_41"  [label="(4, 64, 1152)", style=solid];
-"976 reshape_41" -> "977 permute_43"  [label="(4, 64, 3, 12, 32)", style=solid];
-"977 permute_43" -> "978 select_27"  [label="(3, 4, 12, 64, 32)", style=solid];
-"977 permute_43" -> "979 select_28"  [label="(3, 4, 12, 64, 32)", style=solid];
-"977 permute_43" -> "980 select_29"  [label="(3, 4, 12, 64, 32)", style=solid];
-"978 select_27" -> "981 linalg_vector_norm_18"  [label="(4, 12, 64, 32)", style=solid];
-"978 select_27" -> "983 expand_as_18"  [label="(4, 12, 64, 32)", style=solid];
-"978 select_27" -> "984 div_18"  [label="(4, 12, 64, 32)", style=solid];
-"979 select_28" -> "987 linalg_vector_norm_19"  [label="(4, 12, 64, 32)", style=solid];
-"979 select_28" -> "989 expand_as_19"  [label="(4, 12, 64, 32)", style=solid];
-"979 select_28" -> "990 div_19"  [label="(4, 12, 64, 32)", style=solid];
-"980 select_29" -> "1016 matmul_19"  [label="(4, 12, 64, 32)", style=solid];
-"981 linalg_vector_norm_18" -> "982 clamp_min_18"  [label="(4, 12, 64, 1)", style=solid];
-"982 clamp_min_18" -> "983 expand_as_18"  [label="(4, 12, 64, 1)", style=solid];
-"983 expand_as_18" -> "984 div_18"  [label="(4, 12, 64, 32)", style=solid];
-"984 div_18" -> "985 quantize_per_tensor_default_58"  [label="(4, 12, 64, 32)", style=solid];
-"985 quantize_per_tensor_default_58" -> "986 dequantize_per_tensor_default_58"  [label="(4, 12, 64, 32)", style=solid];
-"986 dequantize_per_tensor_default_58" -> "994 matmul_18"  [label="(4, 12, 64, 32)", style=solid];
-"987 linalg_vector_norm_19" -> "988 clamp_min_19"  [label="(4, 12, 64, 1)", style=solid];
-"988 clamp_min_19" -> "989 expand_as_19"  [label="(4, 12, 64, 1)", style=solid];
-"989 expand_as_19" -> "990 div_19"  [label="(4, 12, 64, 32)", style=solid];
-"990 div_19" -> "991 quantize_per_tensor_default_59"  [label="(4, 12, 64, 32)", style=solid];
-"991 quantize_per_tensor_default_59" -> "992 dequantize_per_tensor_default_59"  [label="(4, 12, 64, 32)", style=solid];
-"992 dequantize_per_tensor_default_59" -> "993 transpose_18"  [label="(4, 12, 64, 32)", style=solid];
-"993 transpose_18" -> "994 matmul_18"  [label="(4, 12, 32, 64)", style=solid];
-"994 matmul_18" -> "996 mul_19"  [label="(4, 12, 64, 64)", style=solid];
-"995 _frozen_param75" -> "996 mul_19"  [label="(12, 1, 1)", style=solid];
-"996 mul_19" -> "997 add_31"  [label="(4, 12, 64, 64)", style=solid];
-"997 add_31" -> "1009 view_52"  [label="(4, 12, 64, 64)", style=solid];
-"998 new_zeros_4" -> "999 view_51"  [label="(16, 16)", style=solid];
-"999 view_51" -> "1000 permute_44"  [label="(2, 8, 2, 8)", style=solid];
-"1000 permute_44" -> "1001 reshape_42"  [label="(2, 2, 8, 8)", style=solid];
-"1001 reshape_42" -> "1002 unsqueeze_26"  [label="(4, 64)", style=solid];
-"1001 reshape_42" -> "1003 unsqueeze_27"  [label="(4, 64)", style=solid];
-"1002 unsqueeze_26" -> "1004 sub_4"  [label="(4, 1, 64)", style=solid];
-"1003 unsqueeze_27" -> "1004 sub_4"  [label="(4, 64, 1)", style=solid];
-"1004 sub_4" -> "1005 ne_4"  [label="(4, 64, 64)", style=solid];
-"1004 sub_4" -> "1006 masked_fill_8"  [label="(4, 64, 64)", style=solid];
-"1004 sub_4" -> "1007 eq_4"  [label="(4, 64, 64)", style=solid];
-"1005 ne_4" -> "1006 masked_fill_8"  [label="(4, 64, 64)", style=solid];
-"1006 masked_fill_8" -> "1008 masked_fill_9"  [label="(4, 64, 64)", style=solid];
-"1007 eq_4" -> "1008 masked_fill_9"  [label="(4, 64, 64)", style=solid];
-"1008 masked_fill_9" -> "1010 unsqueeze_28"  [label="(4, 64, 64)", style=solid];
-"1009 view_52" -> "1012 add_32"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1010 unsqueeze_28" -> "1011 unsqueeze_29"  [label="(4, 1, 64, 64)", style=solid];
-"1011 unsqueeze_29" -> "1012 add_32"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1012 add_32" -> "1013 view_53"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1013 view_53" -> "1014 softmax_9"  [label="(4, 12, 64, 64)", style=solid];
-"1014 softmax_9" -> "1015 dropout_36"  [label="(4, 12, 64, 64)", style=solid];
-"1015 dropout_36" -> "1016 matmul_19"  [label="(4, 12, 64, 64)", style=solid];
-"1016 matmul_19" -> "1017 transpose_19"  [label="(4, 12, 64, 32)", style=solid];
-"1017 transpose_19" -> "1018 reshape_43"  [label="(4, 64, 12, 32)", style=solid];
-"1018 reshape_43" -> "1019 reshape_43_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1019 reshape_43_0_0_nncf_smooth_quant_0" -> "1020 quantize_per_tensor_default_60"  [label="(4, 64, 384)", style=solid];
-"1020 quantize_per_tensor_default_60" -> "1021 dequantize_per_tensor_default_60"  [label="(4, 64, 384)", style=solid];
-"1021 dequantize_per_tensor_default_60" -> "1028 linear_59"  [label="(4, 64, 384)", style=solid];
-"1022 linear_59_scale_0" -> "1025 quantize_per_channel_default_50"  [label="(384,)", style=solid];
-"1022 linear_59_scale_0" -> "1026 dequantize_per_channel_default_50"  [label="(384,)", style=solid];
-"1023 linear_59_zero_point_0" -> "1025 quantize_per_channel_default_50"  [label="(384,)", style=solid];
-"1023 linear_59_zero_point_0" -> "1026 dequantize_per_channel_default_50"  [label="(384,)", style=solid];
-"1024 compressed_weight_updated_constant40" -> "1025 quantize_per_channel_default_50"  [label="(384, 384)", style=solid];
-"1025 quantize_per_channel_default_50" -> "1026 dequantize_per_channel_default_50"  [label="(384, 384)", style=solid];
-"1026 dequantize_per_channel_default_50" -> "1028 linear_59"  [label="(384, 384)", style=solid];
-"1027 _param_constant161_0_0" -> "1028 linear_59"  [label="(384,)", style=solid];
-"1028 linear_59" -> "1029 dropout_37"  [label="(4, 64, 384)", style=solid];
-"1029 dropout_37" -> "1030 view_54"  [label="(4, 64, 384)", style=solid];
-"1030 view_54" -> "1031 permute_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1031 permute_45" -> "1032 reshape_44"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1032 reshape_44" -> "1033 roll_9"  [label="(1, 16, 16, 384)", style=solid];
-"1033 roll_9" -> "1034 slice_157"  [label="(1, 16, 16, 384)", style=solid];
-"1034 slice_157" -> "1035 slice_158"  [label="(1, 16, 16, 384)", style=solid];
-"1035 slice_158" -> "1036 slice_159"  [label="(1, 14, 16, 384)", style=solid];
-"1036 slice_159" -> "1037 slice_160"  [label="(1, 14, 14, 384)", style=solid];
-"1037 slice_160" -> "1038 contiguous_17"  [label="(1, 14, 14, 384)", style=solid];
-"1038 contiguous_17" -> "1041 layer_norm_21"  [label="(1, 14, 14, 384)", style=solid];
-"1039 _param_constant162" -> "1041 layer_norm_21"  [label="(384,)", style=solid];
-"1040 _param_constant163" -> "1041 layer_norm_21"  [label="(384,)", style=solid];
-"1041 layer_norm_21" -> "1042 add_33"  [label="(1, 14, 14, 384)", style=solid];
-"1042 add_33" -> "1043 add_33_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1042 add_33" -> "1069 add_34"  [label="(1, 14, 14, 384)", style=solid];
-"1043 add_33_0_0_nncf_smooth_quant_0" -> "1044 quantize_per_tensor_default_61"  [label="(1, 14, 14, 384)", style=solid];
-"1044 quantize_per_tensor_default_61" -> "1045 dequantize_per_tensor_default_61"  [label="(1, 14, 14, 384)", style=solid];
-"1045 dequantize_per_tensor_default_61" -> "1052 linear_60"  [label="(1, 14, 14, 384)", style=solid];
-"1046 linear_60_scale_0" -> "1049 quantize_per_channel_default_51"  [label="(1536,)", style=solid];
-"1046 linear_60_scale_0" -> "1050 dequantize_per_channel_default_51"  [label="(1536,)", style=solid];
-"1047 linear_60_zero_point_0" -> "1049 quantize_per_channel_default_51"  [label="(1536,)", style=solid];
-"1047 linear_60_zero_point_0" -> "1050 dequantize_per_channel_default_51"  [label="(1536,)", style=solid];
-"1048 compressed_weight_updated_constant41" -> "1049 quantize_per_channel_default_51"  [label="(1536, 384)", style=solid];
-"1049 quantize_per_channel_default_51" -> "1050 dequantize_per_channel_default_51"  [label="(1536, 384)", style=solid];
-"1050 dequantize_per_channel_default_51" -> "1052 linear_60"  [label="(1536, 384)", style=solid];
-"1051 _param_constant165_0_0" -> "1052 linear_60"  [label="(1536,)", style=solid];
-"1052 linear_60" -> "1053 gelu_9"  [label="(1, 14, 14, 1536)", style=solid];
-"1053 gelu_9" -> "1054 dropout_38"  [label="(1, 14, 14, 1536)", style=solid];
-"1054 dropout_38" -> "1055 dropout_38_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1055 dropout_38_0_0_nncf_smooth_quant_0" -> "1056 quantize_per_tensor_default_62"  [label="(1, 14, 14, 1536)", style=solid];
-"1056 quantize_per_tensor_default_62" -> "1057 dequantize_per_tensor_default_62"  [label="(1, 14, 14, 1536)", style=solid];
-"1057 dequantize_per_tensor_default_62" -> "1064 linear_61"  [label="(1, 14, 14, 1536)", style=solid];
-"1058 linear_61_scale_0" -> "1061 quantize_per_channel_default_52"  [label="(384,)", style=solid];
-"1058 linear_61_scale_0" -> "1062 dequantize_per_channel_default_52"  [label="(384,)", style=solid];
-"1059 linear_61_zero_point_0" -> "1061 quantize_per_channel_default_52"  [label="(384,)", style=solid];
-"1059 linear_61_zero_point_0" -> "1062 dequantize_per_channel_default_52"  [label="(384,)", style=solid];
-"1060 compressed_weight_updated_constant42" -> "1061 quantize_per_channel_default_52"  [label="(384, 1536)", style=solid];
-"1061 quantize_per_channel_default_52" -> "1062 dequantize_per_channel_default_52"  [label="(384, 1536)", style=solid];
-"1062 dequantize_per_channel_default_52" -> "1064 linear_61"  [label="(384, 1536)", style=solid];
-"1063 _param_constant167_0_0" -> "1064 linear_61"  [label="(384,)", style=solid];
-"1064 linear_61" -> "1065 dropout_39"  [label="(1, 14, 14, 384)", style=solid];
-"1065 dropout_39" -> "1068 layer_norm_22"  [label="(1, 14, 14, 384)", style=solid];
-"1066 _param_constant168" -> "1068 layer_norm_22"  [label="(384,)", style=solid];
-"1067 _param_constant169" -> "1068 layer_norm_22"  [label="(384,)", style=solid];
-"1068 layer_norm_22" -> "1069 add_34"  [label="(1, 14, 14, 384)", style=solid];
-"1069 add_34" -> "1071 pad_12"  [label="(1, 14, 14, 384)", style=solid];
-"1069 add_34" -> "1134 add_36"  [label="(1, 14, 14, 384)", style=solid];
-"1070 _frozen_param85" -> "1106 add_35"  [label="(1, 12, 64, 64)", style=solid];
-"1071 pad_12" -> "1072 view_57"  [label="(1, 16, 16, 384)", style=solid];
-"1072 view_57" -> "1073 permute_47"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1073 permute_47" -> "1074 reshape_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1074 reshape_45" -> "1075 reshape_45_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1075 reshape_45_0_0_nncf_smooth_quant_0" -> "1076 quantize_per_tensor_default_63"  [label="(4, 64, 384)", style=solid];
-"1076 quantize_per_tensor_default_63" -> "1077 dequantize_per_tensor_default_63"  [label="(4, 64, 384)", style=solid];
-"1077 dequantize_per_tensor_default_63" -> "1084 linear_64"  [label="(4, 64, 384)", style=solid];
-"1078 linear_64_scale_0" -> "1081 quantize_per_channel_default_54"  [label="(1152,)", style=solid];
-"1078 linear_64_scale_0" -> "1082 dequantize_per_channel_default_54"  [label="(1152,)", style=solid];
-"1079 linear_64_zero_point_0" -> "1081 quantize_per_channel_default_54"  [label="(1152,)", style=solid];
-"1079 linear_64_zero_point_0" -> "1082 dequantize_per_channel_default_54"  [label="(1152,)", style=solid];
-"1080 compressed_weight_updated_constant43" -> "1081 quantize_per_channel_default_54"  [label="(1152, 384)", style=solid];
-"1081 quantize_per_channel_default_54" -> "1082 dequantize_per_channel_default_54"  [label="(1152, 384)", style=solid];
-"1082 dequantize_per_channel_default_54" -> "1084 linear_64"  [label="(1152, 384)", style=solid];
-"1083 _frozen_param86_0_0" -> "1084 linear_64"  [label="(1152,)", style=solid];
-"1084 linear_64" -> "1085 reshape_46"  [label="(4, 64, 1152)", style=solid];
-"1085 reshape_46" -> "1086 permute_48"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1086 permute_48" -> "1087 select_30"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1086 permute_48" -> "1088 select_31"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1086 permute_48" -> "1089 select_32"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1087 select_30" -> "1090 linalg_vector_norm_20"  [label="(4, 12, 64, 32)", style=solid];
-"1087 select_30" -> "1092 expand_as_20"  [label="(4, 12, 64, 32)", style=solid];
-"1087 select_30" -> "1093 div_20"  [label="(4, 12, 64, 32)", style=solid];
-"1088 select_31" -> "1096 linalg_vector_norm_21"  [label="(4, 12, 64, 32)", style=solid];
-"1088 select_31" -> "1098 expand_as_21"  [label="(4, 12, 64, 32)", style=solid];
-"1088 select_31" -> "1099 div_21"  [label="(4, 12, 64, 32)", style=solid];
-"1089 select_32" -> "1109 matmul_21"  [label="(4, 12, 64, 32)", style=solid];
-"1090 linalg_vector_norm_20" -> "1091 clamp_min_20"  [label="(4, 12, 64, 1)", style=solid];
-"1091 clamp_min_20" -> "1092 expand_as_20"  [label="(4, 12, 64, 1)", style=solid];
-"1092 expand_as_20" -> "1093 div_20"  [label="(4, 12, 64, 32)", style=solid];
-"1093 div_20" -> "1094 quantize_per_tensor_default_64"  [label="(4, 12, 64, 32)", style=solid];
-"1094 quantize_per_tensor_default_64" -> "1095 dequantize_per_tensor_default_64"  [label="(4, 12, 64, 32)", style=solid];
-"1095 dequantize_per_tensor_default_64" -> "1103 matmul_20"  [label="(4, 12, 64, 32)", style=solid];
-"1096 linalg_vector_norm_21" -> "1097 clamp_min_21"  [label="(4, 12, 64, 1)", style=solid];
-"1097 clamp_min_21" -> "1098 expand_as_21"  [label="(4, 12, 64, 1)", style=solid];
-"1098 expand_as_21" -> "1099 div_21"  [label="(4, 12, 64, 32)", style=solid];
-"1099 div_21" -> "1100 quantize_per_tensor_default_65"  [label="(4, 12, 64, 32)", style=solid];
-"1100 quantize_per_tensor_default_65" -> "1101 dequantize_per_tensor_default_65"  [label="(4, 12, 64, 32)", style=solid];
-"1101 dequantize_per_tensor_default_65" -> "1102 transpose_20"  [label="(4, 12, 64, 32)", style=solid];
-"1102 transpose_20" -> "1103 matmul_20"  [label="(4, 12, 32, 64)", style=solid];
-"1103 matmul_20" -> "1105 mul_21"  [label="(4, 12, 64, 64)", style=solid];
-"1104 _frozen_param88" -> "1105 mul_21"  [label="(12, 1, 1)", style=solid];
-"1105 mul_21" -> "1106 add_35"  [label="(4, 12, 64, 64)", style=solid];
-"1106 add_35" -> "1107 softmax_10"  [label="(4, 12, 64, 64)", style=solid];
-"1107 softmax_10" -> "1108 dropout_40"  [label="(4, 12, 64, 64)", style=solid];
-"1108 dropout_40" -> "1109 matmul_21"  [label="(4, 12, 64, 64)", style=solid];
-"1109 matmul_21" -> "1110 transpose_21"  [label="(4, 12, 64, 32)", style=solid];
-"1110 transpose_21" -> "1111 reshape_47"  [label="(4, 64, 12, 32)", style=solid];
-"1111 reshape_47" -> "1112 reshape_47_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1112 reshape_47_0_0_nncf_smooth_quant_0" -> "1113 quantize_per_tensor_default_66"  [label="(4, 64, 384)", style=solid];
-"1113 quantize_per_tensor_default_66" -> "1114 dequantize_per_tensor_default_66"  [label="(4, 64, 384)", style=solid];
-"1114 dequantize_per_tensor_default_66" -> "1121 linear_65"  [label="(4, 64, 384)", style=solid];
-"1115 linear_65_scale_0" -> "1118 quantize_per_channel_default_55"  [label="(384,)", style=solid];
-"1115 linear_65_scale_0" -> "1119 dequantize_per_channel_default_55"  [label="(384,)", style=solid];
-"1116 linear_65_zero_point_0" -> "1118 quantize_per_channel_default_55"  [label="(384,)", style=solid];
-"1116 linear_65_zero_point_0" -> "1119 dequantize_per_channel_default_55"  [label="(384,)", style=solid];
-"1117 compressed_weight_updated_constant44" -> "1118 quantize_per_channel_default_55"  [label="(384, 384)", style=solid];
-"1118 quantize_per_channel_default_55" -> "1119 dequantize_per_channel_default_55"  [label="(384, 384)", style=solid];
-"1119 dequantize_per_channel_default_55" -> "1121 linear_65"  [label="(384, 384)", style=solid];
-"1120 _param_constant177_0_0" -> "1121 linear_65"  [label="(384,)", style=solid];
-"1121 linear_65" -> "1122 dropout_41"  [label="(4, 64, 384)", style=solid];
-"1122 dropout_41" -> "1123 view_58"  [label="(4, 64, 384)", style=solid];
-"1123 view_58" -> "1124 permute_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1124 permute_49" -> "1125 reshape_48"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1125 reshape_48" -> "1126 slice_162"  [label="(1, 16, 16, 384)", style=solid];
-"1126 slice_162" -> "1127 slice_163"  [label="(1, 16, 16, 384)", style=solid];
-"1127 slice_163" -> "1128 slice_164"  [label="(1, 14, 16, 384)", style=solid];
-"1128 slice_164" -> "1129 slice_165"  [label="(1, 14, 14, 384)", style=solid];
-"1129 slice_165" -> "1130 contiguous_19"  [label="(1, 14, 14, 384)", style=solid];
-"1130 contiguous_19" -> "1133 layer_norm_23"  [label="(1, 14, 14, 384)", style=solid];
-"1131 _param_constant178" -> "1133 layer_norm_23"  [label="(384,)", style=solid];
-"1132 _param_constant179" -> "1133 layer_norm_23"  [label="(384,)", style=solid];
-"1133 layer_norm_23" -> "1134 add_36"  [label="(1, 14, 14, 384)", style=solid];
-"1134 add_36" -> "1135 add_36_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1134 add_36" -> "1161 add_37"  [label="(1, 14, 14, 384)", style=solid];
-"1135 add_36_0_0_nncf_smooth_quant_0" -> "1136 quantize_per_tensor_default_67"  [label="(1, 14, 14, 384)", style=solid];
-"1136 quantize_per_tensor_default_67" -> "1137 dequantize_per_tensor_default_67"  [label="(1, 14, 14, 384)", style=solid];
-"1137 dequantize_per_tensor_default_67" -> "1144 linear_66"  [label="(1, 14, 14, 384)", style=solid];
-"1138 linear_66_scale_0" -> "1141 quantize_per_channel_default_56"  [label="(1536,)", style=solid];
-"1138 linear_66_scale_0" -> "1142 dequantize_per_channel_default_56"  [label="(1536,)", style=solid];
-"1139 linear_66_zero_point_0" -> "1141 quantize_per_channel_default_56"  [label="(1536,)", style=solid];
-"1139 linear_66_zero_point_0" -> "1142 dequantize_per_channel_default_56"  [label="(1536,)", style=solid];
-"1140 compressed_weight_updated_constant45" -> "1141 quantize_per_channel_default_56"  [label="(1536, 384)", style=solid];
-"1141 quantize_per_channel_default_56" -> "1142 dequantize_per_channel_default_56"  [label="(1536, 384)", style=solid];
-"1142 dequantize_per_channel_default_56" -> "1144 linear_66"  [label="(1536, 384)", style=solid];
-"1143 _param_constant181_0_0" -> "1144 linear_66"  [label="(1536,)", style=solid];
-"1144 linear_66" -> "1145 gelu_10"  [label="(1, 14, 14, 1536)", style=solid];
-"1145 gelu_10" -> "1146 dropout_42"  [label="(1, 14, 14, 1536)", style=solid];
-"1146 dropout_42" -> "1147 dropout_42_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1147 dropout_42_0_0_nncf_smooth_quant_0" -> "1148 quantize_per_tensor_default_68"  [label="(1, 14, 14, 1536)", style=solid];
-"1148 quantize_per_tensor_default_68" -> "1149 dequantize_per_tensor_default_68"  [label="(1, 14, 14, 1536)", style=solid];
-"1149 dequantize_per_tensor_default_68" -> "1156 linear_67"  [label="(1, 14, 14, 1536)", style=solid];
-"1150 linear_67_scale_0" -> "1153 quantize_per_channel_default_57"  [label="(384,)", style=solid];
-"1150 linear_67_scale_0" -> "1154 dequantize_per_channel_default_57"  [label="(384,)", style=solid];
-"1151 linear_67_zero_point_0" -> "1153 quantize_per_channel_default_57"  [label="(384,)", style=solid];
-"1151 linear_67_zero_point_0" -> "1154 dequantize_per_channel_default_57"  [label="(384,)", style=solid];
-"1152 compressed_weight_updated_constant46" -> "1153 quantize_per_channel_default_57"  [label="(384, 1536)", style=solid];
-"1153 quantize_per_channel_default_57" -> "1154 dequantize_per_channel_default_57"  [label="(384, 1536)", style=solid];
-"1154 dequantize_per_channel_default_57" -> "1156 linear_67"  [label="(384, 1536)", style=solid];
-"1155 _param_constant183_0_0" -> "1156 linear_67"  [label="(384,)", style=solid];
-"1156 linear_67" -> "1157 dropout_43"  [label="(1, 14, 14, 384)", style=solid];
-"1157 dropout_43" -> "1160 layer_norm_24"  [label="(1, 14, 14, 384)", style=solid];
-"1158 _param_constant184" -> "1160 layer_norm_24"  [label="(384,)", style=solid];
-"1159 _param_constant185" -> "1160 layer_norm_24"  [label="(384,)", style=solid];
-"1160 layer_norm_24" -> "1161 add_37"  [label="(1, 14, 14, 384)", style=solid];
-"1161 add_37" -> "1163 pad_13"  [label="(1, 14, 14, 384)", style=solid];
-"1161 add_37" -> "1244 add_40"  [label="(1, 14, 14, 384)", style=solid];
-"1162 _frozen_param89" -> "1199 add_38"  [label="(1, 12, 64, 64)", style=solid];
-"1163 pad_13" -> "1164 roll_10"  [label="(1, 16, 16, 384)", style=solid];
-"1164 roll_10" -> "1165 view_61"  [label="(1, 16, 16, 384)", style=solid];
-"1165 view_61" -> "1166 permute_51"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1166 permute_51" -> "1167 reshape_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1167 reshape_49" -> "1168 reshape_49_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1167 reshape_49" -> "1200 new_zeros_5"  [label="(4, 64, 384)", style=solid];
-"1168 reshape_49_0_0_nncf_smooth_quant_0" -> "1169 quantize_per_tensor_default_69"  [label="(4, 64, 384)", style=solid];
-"1169 quantize_per_tensor_default_69" -> "1170 dequantize_per_tensor_default_69"  [label="(4, 64, 384)", style=solid];
-"1170 dequantize_per_tensor_default_69" -> "1177 linear_70"  [label="(4, 64, 384)", style=solid];
-"1171 linear_70_scale_0" -> "1174 quantize_per_channel_default_59"  [label="(1152,)", style=solid];
-"1171 linear_70_scale_0" -> "1175 dequantize_per_channel_default_59"  [label="(1152,)", style=solid];
-"1172 linear_70_zero_point_0" -> "1174 quantize_per_channel_default_59"  [label="(1152,)", style=solid];
-"1172 linear_70_zero_point_0" -> "1175 dequantize_per_channel_default_59"  [label="(1152,)", style=solid];
-"1173 compressed_weight_updated_constant47" -> "1174 quantize_per_channel_default_59"  [label="(1152, 384)", style=solid];
-"1174 quantize_per_channel_default_59" -> "1175 dequantize_per_channel_default_59"  [label="(1152, 384)", style=solid];
-"1175 dequantize_per_channel_default_59" -> "1177 linear_70"  [label="(1152, 384)", style=solid];
-"1176 _frozen_param90_0_0" -> "1177 linear_70"  [label="(1152,)", style=solid];
-"1177 linear_70" -> "1178 reshape_50"  [label="(4, 64, 1152)", style=solid];
-"1178 reshape_50" -> "1179 permute_52"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1179 permute_52" -> "1180 select_33"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1179 permute_52" -> "1181 select_34"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1179 permute_52" -> "1182 select_35"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1180 select_33" -> "1183 linalg_vector_norm_22"  [label="(4, 12, 64, 32)", style=solid];
-"1180 select_33" -> "1185 expand_as_22"  [label="(4, 12, 64, 32)", style=solid];
-"1180 select_33" -> "1186 div_22"  [label="(4, 12, 64, 32)", style=solid];
-"1181 select_34" -> "1189 linalg_vector_norm_23"  [label="(4, 12, 64, 32)", style=solid];
-"1181 select_34" -> "1191 expand_as_23"  [label="(4, 12, 64, 32)", style=solid];
-"1181 select_34" -> "1192 div_23"  [label="(4, 12, 64, 32)", style=solid];
-"1182 select_35" -> "1218 matmul_23"  [label="(4, 12, 64, 32)", style=solid];
-"1183 linalg_vector_norm_22" -> "1184 clamp_min_22"  [label="(4, 12, 64, 1)", style=solid];
-"1184 clamp_min_22" -> "1185 expand_as_22"  [label="(4, 12, 64, 1)", style=solid];
-"1185 expand_as_22" -> "1186 div_22"  [label="(4, 12, 64, 32)", style=solid];
-"1186 div_22" -> "1187 quantize_per_tensor_default_70"  [label="(4, 12, 64, 32)", style=solid];
-"1187 quantize_per_tensor_default_70" -> "1188 dequantize_per_tensor_default_70"  [label="(4, 12, 64, 32)", style=solid];
-"1188 dequantize_per_tensor_default_70" -> "1196 matmul_22"  [label="(4, 12, 64, 32)", style=solid];
-"1189 linalg_vector_norm_23" -> "1190 clamp_min_23"  [label="(4, 12, 64, 1)", style=solid];
-"1190 clamp_min_23" -> "1191 expand_as_23"  [label="(4, 12, 64, 1)", style=solid];
-"1191 expand_as_23" -> "1192 div_23"  [label="(4, 12, 64, 32)", style=solid];
-"1192 div_23" -> "1193 quantize_per_tensor_default_71"  [label="(4, 12, 64, 32)", style=solid];
-"1193 quantize_per_tensor_default_71" -> "1194 dequantize_per_tensor_default_71"  [label="(4, 12, 64, 32)", style=solid];
-"1194 dequantize_per_tensor_default_71" -> "1195 transpose_22"  [label="(4, 12, 64, 32)", style=solid];
-"1195 transpose_22" -> "1196 matmul_22"  [label="(4, 12, 32, 64)", style=solid];
-"1196 matmul_22" -> "1198 mul_23"  [label="(4, 12, 64, 64)", style=solid];
-"1197 _frozen_param92" -> "1198 mul_23"  [label="(12, 1, 1)", style=solid];
-"1198 mul_23" -> "1199 add_38"  [label="(4, 12, 64, 64)", style=solid];
-"1199 add_38" -> "1211 view_63"  [label="(4, 12, 64, 64)", style=solid];
-"1200 new_zeros_5" -> "1201 view_62"  [label="(16, 16)", style=solid];
-"1201 view_62" -> "1202 permute_53"  [label="(2, 8, 2, 8)", style=solid];
-"1202 permute_53" -> "1203 reshape_51"  [label="(2, 2, 8, 8)", style=solid];
-"1203 reshape_51" -> "1204 unsqueeze_32"  [label="(4, 64)", style=solid];
-"1203 reshape_51" -> "1205 unsqueeze_33"  [label="(4, 64)", style=solid];
-"1204 unsqueeze_32" -> "1206 sub_5"  [label="(4, 1, 64)", style=solid];
-"1205 unsqueeze_33" -> "1206 sub_5"  [label="(4, 64, 1)", style=solid];
-"1206 sub_5" -> "1207 ne_5"  [label="(4, 64, 64)", style=solid];
-"1206 sub_5" -> "1208 masked_fill_10"  [label="(4, 64, 64)", style=solid];
-"1206 sub_5" -> "1209 eq_5"  [label="(4, 64, 64)", style=solid];
-"1207 ne_5" -> "1208 masked_fill_10"  [label="(4, 64, 64)", style=solid];
-"1208 masked_fill_10" -> "1210 masked_fill_11"  [label="(4, 64, 64)", style=solid];
-"1209 eq_5" -> "1210 masked_fill_11"  [label="(4, 64, 64)", style=solid];
-"1210 masked_fill_11" -> "1212 unsqueeze_34"  [label="(4, 64, 64)", style=solid];
-"1211 view_63" -> "1214 add_39"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1212 unsqueeze_34" -> "1213 unsqueeze_35"  [label="(4, 1, 64, 64)", style=solid];
-"1213 unsqueeze_35" -> "1214 add_39"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1214 add_39" -> "1215 view_64"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1215 view_64" -> "1216 softmax_11"  [label="(4, 12, 64, 64)", style=solid];
-"1216 softmax_11" -> "1217 dropout_44"  [label="(4, 12, 64, 64)", style=solid];
-"1217 dropout_44" -> "1218 matmul_23"  [label="(4, 12, 64, 64)", style=solid];
-"1218 matmul_23" -> "1219 transpose_23"  [label="(4, 12, 64, 32)", style=solid];
-"1219 transpose_23" -> "1220 reshape_52"  [label="(4, 64, 12, 32)", style=solid];
-"1220 reshape_52" -> "1221 reshape_52_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1221 reshape_52_0_0_nncf_smooth_quant_0" -> "1222 quantize_per_tensor_default_72"  [label="(4, 64, 384)", style=solid];
-"1222 quantize_per_tensor_default_72" -> "1223 dequantize_per_tensor_default_72"  [label="(4, 64, 384)", style=solid];
-"1223 dequantize_per_tensor_default_72" -> "1230 linear_71"  [label="(4, 64, 384)", style=solid];
-"1224 linear_71_scale_0" -> "1227 quantize_per_channel_default_60"  [label="(384,)", style=solid];
-"1224 linear_71_scale_0" -> "1228 dequantize_per_channel_default_60"  [label="(384,)", style=solid];
-"1225 linear_71_zero_point_0" -> "1227 quantize_per_channel_default_60"  [label="(384,)", style=solid];
-"1225 linear_71_zero_point_0" -> "1228 dequantize_per_channel_default_60"  [label="(384,)", style=solid];
-"1226 compressed_weight_updated_constant48" -> "1227 quantize_per_channel_default_60"  [label="(384, 384)", style=solid];
-"1227 quantize_per_channel_default_60" -> "1228 dequantize_per_channel_default_60"  [label="(384, 384)", style=solid];
-"1228 dequantize_per_channel_default_60" -> "1230 linear_71"  [label="(384, 384)", style=solid];
-"1229 _param_constant193_0_0" -> "1230 linear_71"  [label="(384,)", style=solid];
-"1230 linear_71" -> "1231 dropout_45"  [label="(4, 64, 384)", style=solid];
-"1231 dropout_45" -> "1232 view_65"  [label="(4, 64, 384)", style=solid];
-"1232 view_65" -> "1233 permute_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1233 permute_54" -> "1234 reshape_53"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1234 reshape_53" -> "1235 roll_11"  [label="(1, 16, 16, 384)", style=solid];
-"1235 roll_11" -> "1236 slice_185"  [label="(1, 16, 16, 384)", style=solid];
-"1236 slice_185" -> "1237 slice_186"  [label="(1, 16, 16, 384)", style=solid];
-"1237 slice_186" -> "1238 slice_187"  [label="(1, 14, 16, 384)", style=solid];
-"1238 slice_187" -> "1239 slice_188"  [label="(1, 14, 14, 384)", style=solid];
-"1239 slice_188" -> "1240 contiguous_21"  [label="(1, 14, 14, 384)", style=solid];
-"1240 contiguous_21" -> "1243 layer_norm_25"  [label="(1, 14, 14, 384)", style=solid];
-"1241 _param_constant194" -> "1243 layer_norm_25"  [label="(384,)", style=solid];
-"1242 _param_constant195" -> "1243 layer_norm_25"  [label="(384,)", style=solid];
-"1243 layer_norm_25" -> "1244 add_40"  [label="(1, 14, 14, 384)", style=solid];
-"1244 add_40" -> "1245 add_40_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1244 add_40" -> "1271 add_41"  [label="(1, 14, 14, 384)", style=solid];
-"1245 add_40_0_0_nncf_smooth_quant_0" -> "1246 quantize_per_tensor_default_73"  [label="(1, 14, 14, 384)", style=solid];
-"1246 quantize_per_tensor_default_73" -> "1247 dequantize_per_tensor_default_73"  [label="(1, 14, 14, 384)", style=solid];
-"1247 dequantize_per_tensor_default_73" -> "1254 linear_72"  [label="(1, 14, 14, 384)", style=solid];
-"1248 linear_72_scale_0" -> "1251 quantize_per_channel_default_61"  [label="(1536,)", style=solid];
-"1248 linear_72_scale_0" -> "1252 dequantize_per_channel_default_61"  [label="(1536,)", style=solid];
-"1249 linear_72_zero_point_0" -> "1251 quantize_per_channel_default_61"  [label="(1536,)", style=solid];
-"1249 linear_72_zero_point_0" -> "1252 dequantize_per_channel_default_61"  [label="(1536,)", style=solid];
-"1250 compressed_weight_updated_constant49" -> "1251 quantize_per_channel_default_61"  [label="(1536, 384)", style=solid];
-"1251 quantize_per_channel_default_61" -> "1252 dequantize_per_channel_default_61"  [label="(1536, 384)", style=solid];
-"1252 dequantize_per_channel_default_61" -> "1254 linear_72"  [label="(1536, 384)", style=solid];
-"1253 _param_constant197_0_0" -> "1254 linear_72"  [label="(1536,)", style=solid];
-"1254 linear_72" -> "1255 gelu_11"  [label="(1, 14, 14, 1536)", style=solid];
-"1255 gelu_11" -> "1256 dropout_46"  [label="(1, 14, 14, 1536)", style=solid];
-"1256 dropout_46" -> "1257 dropout_46_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1257 dropout_46_0_0_nncf_smooth_quant_0" -> "1258 quantize_per_tensor_default_74"  [label="(1, 14, 14, 1536)", style=solid];
-"1258 quantize_per_tensor_default_74" -> "1259 dequantize_per_tensor_default_74"  [label="(1, 14, 14, 1536)", style=solid];
-"1259 dequantize_per_tensor_default_74" -> "1266 linear_73"  [label="(1, 14, 14, 1536)", style=solid];
-"1260 linear_73_scale_0" -> "1263 quantize_per_channel_default_62"  [label="(384,)", style=solid];
-"1260 linear_73_scale_0" -> "1264 dequantize_per_channel_default_62"  [label="(384,)", style=solid];
-"1261 linear_73_zero_point_0" -> "1263 quantize_per_channel_default_62"  [label="(384,)", style=solid];
-"1261 linear_73_zero_point_0" -> "1264 dequantize_per_channel_default_62"  [label="(384,)", style=solid];
-"1262 compressed_weight_updated_constant50" -> "1263 quantize_per_channel_default_62"  [label="(384, 1536)", style=solid];
-"1263 quantize_per_channel_default_62" -> "1264 dequantize_per_channel_default_62"  [label="(384, 1536)", style=solid];
-"1264 dequantize_per_channel_default_62" -> "1266 linear_73"  [label="(384, 1536)", style=solid];
-"1265 _param_constant199_0_0" -> "1266 linear_73"  [label="(384,)", style=solid];
-"1266 linear_73" -> "1267 dropout_47"  [label="(1, 14, 14, 384)", style=solid];
-"1267 dropout_47" -> "1270 layer_norm_26"  [label="(1, 14, 14, 384)", style=solid];
-"1268 _param_constant200" -> "1270 layer_norm_26"  [label="(384,)", style=solid];
-"1269 _param_constant201" -> "1270 layer_norm_26"  [label="(384,)", style=solid];
-"1270 layer_norm_26" -> "1271 add_41"  [label="(1, 14, 14, 384)", style=solid];
-"1271 add_41" -> "1273 pad_14"  [label="(1, 14, 14, 384)", style=solid];
-"1271 add_41" -> "1336 add_43"  [label="(1, 14, 14, 384)", style=solid];
-"1272 _frozen_param102" -> "1308 add_42"  [label="(1, 12, 64, 64)", style=solid];
-"1273 pad_14" -> "1274 view_68"  [label="(1, 16, 16, 384)", style=solid];
-"1274 view_68" -> "1275 permute_56"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1275 permute_56" -> "1276 reshape_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1276 reshape_54" -> "1277 reshape_54_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1277 reshape_54_0_0_nncf_smooth_quant_0" -> "1278 quantize_per_tensor_default_75"  [label="(4, 64, 384)", style=solid];
-"1278 quantize_per_tensor_default_75" -> "1279 dequantize_per_tensor_default_75"  [label="(4, 64, 384)", style=solid];
-"1279 dequantize_per_tensor_default_75" -> "1286 linear_76"  [label="(4, 64, 384)", style=solid];
-"1280 linear_76_scale_0" -> "1283 quantize_per_channel_default_64"  [label="(1152,)", style=solid];
-"1280 linear_76_scale_0" -> "1284 dequantize_per_channel_default_64"  [label="(1152,)", style=solid];
-"1281 linear_76_zero_point_0" -> "1283 quantize_per_channel_default_64"  [label="(1152,)", style=solid];
-"1281 linear_76_zero_point_0" -> "1284 dequantize_per_channel_default_64"  [label="(1152,)", style=solid];
-"1282 compressed_weight_updated_constant51" -> "1283 quantize_per_channel_default_64"  [label="(1152, 384)", style=solid];
-"1283 quantize_per_channel_default_64" -> "1284 dequantize_per_channel_default_64"  [label="(1152, 384)", style=solid];
-"1284 dequantize_per_channel_default_64" -> "1286 linear_76"  [label="(1152, 384)", style=solid];
-"1285 _frozen_param103_0_0" -> "1286 linear_76"  [label="(1152,)", style=solid];
-"1286 linear_76" -> "1287 reshape_55"  [label="(4, 64, 1152)", style=solid];
-"1287 reshape_55" -> "1288 permute_57"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1288 permute_57" -> "1289 select_36"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1288 permute_57" -> "1290 select_37"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1288 permute_57" -> "1291 select_38"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1289 select_36" -> "1292 linalg_vector_norm_24"  [label="(4, 12, 64, 32)", style=solid];
-"1289 select_36" -> "1294 expand_as_24"  [label="(4, 12, 64, 32)", style=solid];
-"1289 select_36" -> "1295 div_24"  [label="(4, 12, 64, 32)", style=solid];
-"1290 select_37" -> "1298 linalg_vector_norm_25"  [label="(4, 12, 64, 32)", style=solid];
-"1290 select_37" -> "1300 expand_as_25"  [label="(4, 12, 64, 32)", style=solid];
-"1290 select_37" -> "1301 div_25"  [label="(4, 12, 64, 32)", style=solid];
-"1291 select_38" -> "1311 matmul_25"  [label="(4, 12, 64, 32)", style=solid];
-"1292 linalg_vector_norm_24" -> "1293 clamp_min_24"  [label="(4, 12, 64, 1)", style=solid];
-"1293 clamp_min_24" -> "1294 expand_as_24"  [label="(4, 12, 64, 1)", style=solid];
-"1294 expand_as_24" -> "1295 div_24"  [label="(4, 12, 64, 32)", style=solid];
-"1295 div_24" -> "1296 quantize_per_tensor_default_76"  [label="(4, 12, 64, 32)", style=solid];
-"1296 quantize_per_tensor_default_76" -> "1297 dequantize_per_tensor_default_76"  [label="(4, 12, 64, 32)", style=solid];
-"1297 dequantize_per_tensor_default_76" -> "1305 matmul_24"  [label="(4, 12, 64, 32)", style=solid];
-"1298 linalg_vector_norm_25" -> "1299 clamp_min_25"  [label="(4, 12, 64, 1)", style=solid];
-"1299 clamp_min_25" -> "1300 expand_as_25"  [label="(4, 12, 64, 1)", style=solid];
-"1300 expand_as_25" -> "1301 div_25"  [label="(4, 12, 64, 32)", style=solid];
-"1301 div_25" -> "1302 quantize_per_tensor_default_77"  [label="(4, 12, 64, 32)", style=solid];
-"1302 quantize_per_tensor_default_77" -> "1303 dequantize_per_tensor_default_77"  [label="(4, 12, 64, 32)", style=solid];
-"1303 dequantize_per_tensor_default_77" -> "1304 transpose_24"  [label="(4, 12, 64, 32)", style=solid];
-"1304 transpose_24" -> "1305 matmul_24"  [label="(4, 12, 32, 64)", style=solid];
-"1305 matmul_24" -> "1307 mul_25"  [label="(4, 12, 64, 64)", style=solid];
-"1306 _frozen_param105" -> "1307 mul_25"  [label="(12, 1, 1)", style=solid];
-"1307 mul_25" -> "1308 add_42"  [label="(4, 12, 64, 64)", style=solid];
-"1308 add_42" -> "1309 softmax_12"  [label="(4, 12, 64, 64)", style=solid];
-"1309 softmax_12" -> "1310 dropout_48"  [label="(4, 12, 64, 64)", style=solid];
-"1310 dropout_48" -> "1311 matmul_25"  [label="(4, 12, 64, 64)", style=solid];
-"1311 matmul_25" -> "1312 transpose_25"  [label="(4, 12, 64, 32)", style=solid];
-"1312 transpose_25" -> "1313 reshape_56"  [label="(4, 64, 12, 32)", style=solid];
-"1313 reshape_56" -> "1314 reshape_56_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1314 reshape_56_0_0_nncf_smooth_quant_0" -> "1315 quantize_per_tensor_default_78"  [label="(4, 64, 384)", style=solid];
-"1315 quantize_per_tensor_default_78" -> "1316 dequantize_per_tensor_default_78"  [label="(4, 64, 384)", style=solid];
-"1316 dequantize_per_tensor_default_78" -> "1323 linear_77"  [label="(4, 64, 384)", style=solid];
-"1317 linear_77_scale_0" -> "1320 quantize_per_channel_default_65"  [label="(384,)", style=solid];
-"1317 linear_77_scale_0" -> "1321 dequantize_per_channel_default_65"  [label="(384,)", style=solid];
-"1318 linear_77_zero_point_0" -> "1320 quantize_per_channel_default_65"  [label="(384,)", style=solid];
-"1318 linear_77_zero_point_0" -> "1321 dequantize_per_channel_default_65"  [label="(384,)", style=solid];
-"1319 compressed_weight_updated_constant52" -> "1320 quantize_per_channel_default_65"  [label="(384, 384)", style=solid];
-"1320 quantize_per_channel_default_65" -> "1321 dequantize_per_channel_default_65"  [label="(384, 384)", style=solid];
-"1321 dequantize_per_channel_default_65" -> "1323 linear_77"  [label="(384, 384)", style=solid];
-"1322 _param_constant209_0_0" -> "1323 linear_77"  [label="(384,)", style=solid];
-"1323 linear_77" -> "1324 dropout_49"  [label="(4, 64, 384)", style=solid];
-"1324 dropout_49" -> "1325 view_69"  [label="(4, 64, 384)", style=solid];
-"1325 view_69" -> "1326 permute_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1326 permute_58" -> "1327 reshape_57"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1327 reshape_57" -> "1328 slice_190"  [label="(1, 16, 16, 384)", style=solid];
-"1328 slice_190" -> "1329 slice_191"  [label="(1, 16, 16, 384)", style=solid];
-"1329 slice_191" -> "1330 slice_192"  [label="(1, 14, 16, 384)", style=solid];
-"1330 slice_192" -> "1331 slice_193"  [label="(1, 14, 14, 384)", style=solid];
-"1331 slice_193" -> "1332 contiguous_23"  [label="(1, 14, 14, 384)", style=solid];
-"1332 contiguous_23" -> "1335 layer_norm_27"  [label="(1, 14, 14, 384)", style=solid];
-"1333 _param_constant210" -> "1335 layer_norm_27"  [label="(384,)", style=solid];
-"1334 _param_constant211" -> "1335 layer_norm_27"  [label="(384,)", style=solid];
-"1335 layer_norm_27" -> "1336 add_43"  [label="(1, 14, 14, 384)", style=solid];
-"1336 add_43" -> "1337 add_43_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1336 add_43" -> "1363 add_44"  [label="(1, 14, 14, 384)", style=solid];
-"1337 add_43_0_0_nncf_smooth_quant_0" -> "1338 quantize_per_tensor_default_79"  [label="(1, 14, 14, 384)", style=solid];
-"1338 quantize_per_tensor_default_79" -> "1339 dequantize_per_tensor_default_79"  [label="(1, 14, 14, 384)", style=solid];
-"1339 dequantize_per_tensor_default_79" -> "1346 linear_78"  [label="(1, 14, 14, 384)", style=solid];
-"1340 linear_78_scale_0" -> "1343 quantize_per_channel_default_66"  [label="(1536,)", style=solid];
-"1340 linear_78_scale_0" -> "1344 dequantize_per_channel_default_66"  [label="(1536,)", style=solid];
-"1341 linear_78_zero_point_0" -> "1343 quantize_per_channel_default_66"  [label="(1536,)", style=solid];
-"1341 linear_78_zero_point_0" -> "1344 dequantize_per_channel_default_66"  [label="(1536,)", style=solid];
-"1342 compressed_weight_updated_constant53" -> "1343 quantize_per_channel_default_66"  [label="(1536, 384)", style=solid];
-"1343 quantize_per_channel_default_66" -> "1344 dequantize_per_channel_default_66"  [label="(1536, 384)", style=solid];
-"1344 dequantize_per_channel_default_66" -> "1346 linear_78"  [label="(1536, 384)", style=solid];
-"1345 _param_constant213_0_0" -> "1346 linear_78"  [label="(1536,)", style=solid];
-"1346 linear_78" -> "1347 gelu_12"  [label="(1, 14, 14, 1536)", style=solid];
-"1347 gelu_12" -> "1348 dropout_50"  [label="(1, 14, 14, 1536)", style=solid];
-"1348 dropout_50" -> "1349 dropout_50_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1349 dropout_50_0_0_nncf_smooth_quant_0" -> "1350 quantize_per_tensor_default_80"  [label="(1, 14, 14, 1536)", style=solid];
-"1350 quantize_per_tensor_default_80" -> "1351 dequantize_per_tensor_default_80"  [label="(1, 14, 14, 1536)", style=solid];
-"1351 dequantize_per_tensor_default_80" -> "1358 linear_79"  [label="(1, 14, 14, 1536)", style=solid];
-"1352 linear_79_scale_0" -> "1355 quantize_per_channel_default_67"  [label="(384,)", style=solid];
-"1352 linear_79_scale_0" -> "1356 dequantize_per_channel_default_67"  [label="(384,)", style=solid];
-"1353 linear_79_zero_point_0" -> "1355 quantize_per_channel_default_67"  [label="(384,)", style=solid];
-"1353 linear_79_zero_point_0" -> "1356 dequantize_per_channel_default_67"  [label="(384,)", style=solid];
-"1354 compressed_weight_updated_constant54" -> "1355 quantize_per_channel_default_67"  [label="(384, 1536)", style=solid];
-"1355 quantize_per_channel_default_67" -> "1356 dequantize_per_channel_default_67"  [label="(384, 1536)", style=solid];
-"1356 dequantize_per_channel_default_67" -> "1358 linear_79"  [label="(384, 1536)", style=solid];
-"1357 _param_constant215_0_0" -> "1358 linear_79"  [label="(384,)", style=solid];
-"1358 linear_79" -> "1359 dropout_51"  [label="(1, 14, 14, 384)", style=solid];
-"1359 dropout_51" -> "1362 layer_norm_28"  [label="(1, 14, 14, 384)", style=solid];
-"1360 _param_constant216" -> "1362 layer_norm_28"  [label="(384,)", style=solid];
-"1361 _param_constant217" -> "1362 layer_norm_28"  [label="(384,)", style=solid];
-"1362 layer_norm_28" -> "1363 add_44"  [label="(1, 14, 14, 384)", style=solid];
-"1363 add_44" -> "1365 pad_15"  [label="(1, 14, 14, 384)", style=solid];
-"1363 add_44" -> "1446 add_47"  [label="(1, 14, 14, 384)", style=solid];
-"1364 _frozen_param106" -> "1401 add_45"  [label="(1, 12, 64, 64)", style=solid];
-"1365 pad_15" -> "1366 roll_12"  [label="(1, 16, 16, 384)", style=solid];
-"1366 roll_12" -> "1367 view_72"  [label="(1, 16, 16, 384)", style=solid];
-"1367 view_72" -> "1368 permute_60"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1368 permute_60" -> "1369 reshape_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1369 reshape_58" -> "1370 reshape_58_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1369 reshape_58" -> "1402 new_zeros_6"  [label="(4, 64, 384)", style=solid];
-"1370 reshape_58_0_0_nncf_smooth_quant_0" -> "1371 quantize_per_tensor_default_81"  [label="(4, 64, 384)", style=solid];
-"1371 quantize_per_tensor_default_81" -> "1372 dequantize_per_tensor_default_81"  [label="(4, 64, 384)", style=solid];
-"1372 dequantize_per_tensor_default_81" -> "1379 linear_82"  [label="(4, 64, 384)", style=solid];
-"1373 linear_82_scale_0" -> "1376 quantize_per_channel_default_69"  [label="(1152,)", style=solid];
-"1373 linear_82_scale_0" -> "1377 dequantize_per_channel_default_69"  [label="(1152,)", style=solid];
-"1374 linear_82_zero_point_0" -> "1376 quantize_per_channel_default_69"  [label="(1152,)", style=solid];
-"1374 linear_82_zero_point_0" -> "1377 dequantize_per_channel_default_69"  [label="(1152,)", style=solid];
-"1375 compressed_weight_updated_constant55" -> "1376 quantize_per_channel_default_69"  [label="(1152, 384)", style=solid];
-"1376 quantize_per_channel_default_69" -> "1377 dequantize_per_channel_default_69"  [label="(1152, 384)", style=solid];
-"1377 dequantize_per_channel_default_69" -> "1379 linear_82"  [label="(1152, 384)", style=solid];
-"1378 _frozen_param107_0_0" -> "1379 linear_82"  [label="(1152,)", style=solid];
-"1379 linear_82" -> "1380 reshape_59"  [label="(4, 64, 1152)", style=solid];
-"1380 reshape_59" -> "1381 permute_61"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1381 permute_61" -> "1382 select_39"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1381 permute_61" -> "1383 select_40"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1381 permute_61" -> "1384 select_41"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1382 select_39" -> "1385 linalg_vector_norm_26"  [label="(4, 12, 64, 32)", style=solid];
-"1382 select_39" -> "1387 expand_as_26"  [label="(4, 12, 64, 32)", style=solid];
-"1382 select_39" -> "1388 div_26"  [label="(4, 12, 64, 32)", style=solid];
-"1383 select_40" -> "1391 linalg_vector_norm_27"  [label="(4, 12, 64, 32)", style=solid];
-"1383 select_40" -> "1393 expand_as_27"  [label="(4, 12, 64, 32)", style=solid];
-"1383 select_40" -> "1394 div_27"  [label="(4, 12, 64, 32)", style=solid];
-"1384 select_41" -> "1420 matmul_27"  [label="(4, 12, 64, 32)", style=solid];
-"1385 linalg_vector_norm_26" -> "1386 clamp_min_26"  [label="(4, 12, 64, 1)", style=solid];
-"1386 clamp_min_26" -> "1387 expand_as_26"  [label="(4, 12, 64, 1)", style=solid];
-"1387 expand_as_26" -> "1388 div_26"  [label="(4, 12, 64, 32)", style=solid];
-"1388 div_26" -> "1389 quantize_per_tensor_default_82"  [label="(4, 12, 64, 32)", style=solid];
-"1389 quantize_per_tensor_default_82" -> "1390 dequantize_per_tensor_default_82"  [label="(4, 12, 64, 32)", style=solid];
-"1390 dequantize_per_tensor_default_82" -> "1398 matmul_26"  [label="(4, 12, 64, 32)", style=solid];
-"1391 linalg_vector_norm_27" -> "1392 clamp_min_27"  [label="(4, 12, 64, 1)", style=solid];
-"1392 clamp_min_27" -> "1393 expand_as_27"  [label="(4, 12, 64, 1)", style=solid];
-"1393 expand_as_27" -> "1394 div_27"  [label="(4, 12, 64, 32)", style=solid];
-"1394 div_27" -> "1395 quantize_per_tensor_default_83"  [label="(4, 12, 64, 32)", style=solid];
-"1395 quantize_per_tensor_default_83" -> "1396 dequantize_per_tensor_default_83"  [label="(4, 12, 64, 32)", style=solid];
-"1396 dequantize_per_tensor_default_83" -> "1397 transpose_26"  [label="(4, 12, 64, 32)", style=solid];
-"1397 transpose_26" -> "1398 matmul_26"  [label="(4, 12, 32, 64)", style=solid];
-"1398 matmul_26" -> "1400 mul_27"  [label="(4, 12, 64, 64)", style=solid];
-"1399 _frozen_param109" -> "1400 mul_27"  [label="(12, 1, 1)", style=solid];
-"1400 mul_27" -> "1401 add_45"  [label="(4, 12, 64, 64)", style=solid];
-"1401 add_45" -> "1413 view_74"  [label="(4, 12, 64, 64)", style=solid];
-"1402 new_zeros_6" -> "1403 view_73"  [label="(16, 16)", style=solid];
-"1403 view_73" -> "1404 permute_62"  [label="(2, 8, 2, 8)", style=solid];
-"1404 permute_62" -> "1405 reshape_60"  [label="(2, 2, 8, 8)", style=solid];
-"1405 reshape_60" -> "1406 unsqueeze_38"  [label="(4, 64)", style=solid];
-"1405 reshape_60" -> "1407 unsqueeze_39"  [label="(4, 64)", style=solid];
-"1406 unsqueeze_38" -> "1408 sub_6"  [label="(4, 1, 64)", style=solid];
-"1407 unsqueeze_39" -> "1408 sub_6"  [label="(4, 64, 1)", style=solid];
-"1408 sub_6" -> "1409 ne_6"  [label="(4, 64, 64)", style=solid];
-"1408 sub_6" -> "1410 masked_fill_12"  [label="(4, 64, 64)", style=solid];
-"1408 sub_6" -> "1411 eq_6"  [label="(4, 64, 64)", style=solid];
-"1409 ne_6" -> "1410 masked_fill_12"  [label="(4, 64, 64)", style=solid];
-"1410 masked_fill_12" -> "1412 masked_fill_13"  [label="(4, 64, 64)", style=solid];
-"1411 eq_6" -> "1412 masked_fill_13"  [label="(4, 64, 64)", style=solid];
-"1412 masked_fill_13" -> "1414 unsqueeze_40"  [label="(4, 64, 64)", style=solid];
-"1413 view_74" -> "1416 add_46"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1414 unsqueeze_40" -> "1415 unsqueeze_41"  [label="(4, 1, 64, 64)", style=solid];
-"1415 unsqueeze_41" -> "1416 add_46"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1416 add_46" -> "1417 view_75"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1417 view_75" -> "1418 softmax_13"  [label="(4, 12, 64, 64)", style=solid];
-"1418 softmax_13" -> "1419 dropout_52"  [label="(4, 12, 64, 64)", style=solid];
-"1419 dropout_52" -> "1420 matmul_27"  [label="(4, 12, 64, 64)", style=solid];
-"1420 matmul_27" -> "1421 transpose_27"  [label="(4, 12, 64, 32)", style=solid];
-"1421 transpose_27" -> "1422 reshape_61"  [label="(4, 64, 12, 32)", style=solid];
-"1422 reshape_61" -> "1423 reshape_61_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1423 reshape_61_0_0_nncf_smooth_quant_0" -> "1424 quantize_per_tensor_default_84"  [label="(4, 64, 384)", style=solid];
-"1424 quantize_per_tensor_default_84" -> "1425 dequantize_per_tensor_default_84"  [label="(4, 64, 384)", style=solid];
-"1425 dequantize_per_tensor_default_84" -> "1432 linear_83"  [label="(4, 64, 384)", style=solid];
-"1426 linear_83_scale_0" -> "1429 quantize_per_channel_default_70"  [label="(384,)", style=solid];
-"1426 linear_83_scale_0" -> "1430 dequantize_per_channel_default_70"  [label="(384,)", style=solid];
-"1427 linear_83_zero_point_0" -> "1429 quantize_per_channel_default_70"  [label="(384,)", style=solid];
-"1427 linear_83_zero_point_0" -> "1430 dequantize_per_channel_default_70"  [label="(384,)", style=solid];
-"1428 compressed_weight_updated_constant56" -> "1429 quantize_per_channel_default_70"  [label="(384, 384)", style=solid];
-"1429 quantize_per_channel_default_70" -> "1430 dequantize_per_channel_default_70"  [label="(384, 384)", style=solid];
-"1430 dequantize_per_channel_default_70" -> "1432 linear_83"  [label="(384, 384)", style=solid];
-"1431 _param_constant225_0_0" -> "1432 linear_83"  [label="(384,)", style=solid];
-"1432 linear_83" -> "1433 dropout_53"  [label="(4, 64, 384)", style=solid];
-"1433 dropout_53" -> "1434 view_76"  [label="(4, 64, 384)", style=solid];
-"1434 view_76" -> "1435 permute_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1435 permute_63" -> "1436 reshape_62"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1436 reshape_62" -> "1437 roll_13"  [label="(1, 16, 16, 384)", style=solid];
-"1437 roll_13" -> "1438 slice_213"  [label="(1, 16, 16, 384)", style=solid];
-"1438 slice_213" -> "1439 slice_214"  [label="(1, 16, 16, 384)", style=solid];
-"1439 slice_214" -> "1440 slice_215"  [label="(1, 14, 16, 384)", style=solid];
-"1440 slice_215" -> "1441 slice_216"  [label="(1, 14, 14, 384)", style=solid];
-"1441 slice_216" -> "1442 contiguous_25"  [label="(1, 14, 14, 384)", style=solid];
-"1442 contiguous_25" -> "1445 layer_norm_29"  [label="(1, 14, 14, 384)", style=solid];
-"1443 _param_constant226" -> "1445 layer_norm_29"  [label="(384,)", style=solid];
-"1444 _param_constant227" -> "1445 layer_norm_29"  [label="(384,)", style=solid];
-"1445 layer_norm_29" -> "1446 add_47"  [label="(1, 14, 14, 384)", style=solid];
-"1446 add_47" -> "1447 add_47_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1446 add_47" -> "1473 add_48"  [label="(1, 14, 14, 384)", style=solid];
-"1447 add_47_0_0_nncf_smooth_quant_0" -> "1448 quantize_per_tensor_default_85"  [label="(1, 14, 14, 384)", style=solid];
-"1448 quantize_per_tensor_default_85" -> "1449 dequantize_per_tensor_default_85"  [label="(1, 14, 14, 384)", style=solid];
-"1449 dequantize_per_tensor_default_85" -> "1456 linear_84"  [label="(1, 14, 14, 384)", style=solid];
-"1450 linear_84_scale_0" -> "1453 quantize_per_channel_default_71"  [label="(1536,)", style=solid];
-"1450 linear_84_scale_0" -> "1454 dequantize_per_channel_default_71"  [label="(1536,)", style=solid];
-"1451 linear_84_zero_point_0" -> "1453 quantize_per_channel_default_71"  [label="(1536,)", style=solid];
-"1451 linear_84_zero_point_0" -> "1454 dequantize_per_channel_default_71"  [label="(1536,)", style=solid];
-"1452 compressed_weight_updated_constant57" -> "1453 quantize_per_channel_default_71"  [label="(1536, 384)", style=solid];
-"1453 quantize_per_channel_default_71" -> "1454 dequantize_per_channel_default_71"  [label="(1536, 384)", style=solid];
-"1454 dequantize_per_channel_default_71" -> "1456 linear_84"  [label="(1536, 384)", style=solid];
-"1455 _param_constant229_0_0" -> "1456 linear_84"  [label="(1536,)", style=solid];
-"1456 linear_84" -> "1457 gelu_13"  [label="(1, 14, 14, 1536)", style=solid];
-"1457 gelu_13" -> "1458 dropout_54"  [label="(1, 14, 14, 1536)", style=solid];
-"1458 dropout_54" -> "1459 dropout_54_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1459 dropout_54_0_0_nncf_smooth_quant_0" -> "1460 quantize_per_tensor_default_86"  [label="(1, 14, 14, 1536)", style=solid];
-"1460 quantize_per_tensor_default_86" -> "1461 dequantize_per_tensor_default_86"  [label="(1, 14, 14, 1536)", style=solid];
-"1461 dequantize_per_tensor_default_86" -> "1468 linear_85"  [label="(1, 14, 14, 1536)", style=solid];
-"1462 linear_85_scale_0" -> "1465 quantize_per_channel_default_72"  [label="(384,)", style=solid];
-"1462 linear_85_scale_0" -> "1466 dequantize_per_channel_default_72"  [label="(384,)", style=solid];
-"1463 linear_85_zero_point_0" -> "1465 quantize_per_channel_default_72"  [label="(384,)", style=solid];
-"1463 linear_85_zero_point_0" -> "1466 dequantize_per_channel_default_72"  [label="(384,)", style=solid];
-"1464 compressed_weight_updated_constant58" -> "1465 quantize_per_channel_default_72"  [label="(384, 1536)", style=solid];
-"1465 quantize_per_channel_default_72" -> "1466 dequantize_per_channel_default_72"  [label="(384, 1536)", style=solid];
-"1466 dequantize_per_channel_default_72" -> "1468 linear_85"  [label="(384, 1536)", style=solid];
-"1467 _param_constant231_0_0" -> "1468 linear_85"  [label="(384,)", style=solid];
-"1468 linear_85" -> "1469 dropout_55"  [label="(1, 14, 14, 384)", style=solid];
-"1469 dropout_55" -> "1472 layer_norm_30"  [label="(1, 14, 14, 384)", style=solid];
-"1470 _param_constant232" -> "1472 layer_norm_30"  [label="(384,)", style=solid];
-"1471 _param_constant233" -> "1472 layer_norm_30"  [label="(384,)", style=solid];
-"1472 layer_norm_30" -> "1473 add_48"  [label="(1, 14, 14, 384)", style=solid];
-"1473 add_48" -> "1475 pad_16"  [label="(1, 14, 14, 384)", style=solid];
-"1473 add_48" -> "1538 add_50"  [label="(1, 14, 14, 384)", style=solid];
-"1474 _frozen_param119" -> "1510 add_49"  [label="(1, 12, 64, 64)", style=solid];
-"1475 pad_16" -> "1476 view_79"  [label="(1, 16, 16, 384)", style=solid];
-"1476 view_79" -> "1477 permute_65"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1477 permute_65" -> "1478 reshape_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1478 reshape_63" -> "1479 reshape_63_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1479 reshape_63_0_0_nncf_smooth_quant_0" -> "1480 quantize_per_tensor_default_87"  [label="(4, 64, 384)", style=solid];
-"1480 quantize_per_tensor_default_87" -> "1481 dequantize_per_tensor_default_87"  [label="(4, 64, 384)", style=solid];
-"1481 dequantize_per_tensor_default_87" -> "1488 linear_88"  [label="(4, 64, 384)", style=solid];
-"1482 linear_88_scale_0" -> "1485 quantize_per_channel_default_74"  [label="(1152,)", style=solid];
-"1482 linear_88_scale_0" -> "1486 dequantize_per_channel_default_74"  [label="(1152,)", style=solid];
-"1483 linear_88_zero_point_0" -> "1485 quantize_per_channel_default_74"  [label="(1152,)", style=solid];
-"1483 linear_88_zero_point_0" -> "1486 dequantize_per_channel_default_74"  [label="(1152,)", style=solid];
-"1484 compressed_weight_updated_constant59" -> "1485 quantize_per_channel_default_74"  [label="(1152, 384)", style=solid];
-"1485 quantize_per_channel_default_74" -> "1486 dequantize_per_channel_default_74"  [label="(1152, 384)", style=solid];
-"1486 dequantize_per_channel_default_74" -> "1488 linear_88"  [label="(1152, 384)", style=solid];
-"1487 _frozen_param120_0_0" -> "1488 linear_88"  [label="(1152,)", style=solid];
-"1488 linear_88" -> "1489 reshape_64"  [label="(4, 64, 1152)", style=solid];
-"1489 reshape_64" -> "1490 permute_66"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1490 permute_66" -> "1491 select_42"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1490 permute_66" -> "1492 select_43"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1490 permute_66" -> "1493 select_44"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1491 select_42" -> "1494 linalg_vector_norm_28"  [label="(4, 12, 64, 32)", style=solid];
-"1491 select_42" -> "1496 expand_as_28"  [label="(4, 12, 64, 32)", style=solid];
-"1491 select_42" -> "1497 div_28"  [label="(4, 12, 64, 32)", style=solid];
-"1492 select_43" -> "1500 linalg_vector_norm_29"  [label="(4, 12, 64, 32)", style=solid];
-"1492 select_43" -> "1502 expand_as_29"  [label="(4, 12, 64, 32)", style=solid];
-"1492 select_43" -> "1503 div_29"  [label="(4, 12, 64, 32)", style=solid];
-"1493 select_44" -> "1513 matmul_29"  [label="(4, 12, 64, 32)", style=solid];
-"1494 linalg_vector_norm_28" -> "1495 clamp_min_28"  [label="(4, 12, 64, 1)", style=solid];
-"1495 clamp_min_28" -> "1496 expand_as_28"  [label="(4, 12, 64, 1)", style=solid];
-"1496 expand_as_28" -> "1497 div_28"  [label="(4, 12, 64, 32)", style=solid];
-"1497 div_28" -> "1498 quantize_per_tensor_default_88"  [label="(4, 12, 64, 32)", style=solid];
-"1498 quantize_per_tensor_default_88" -> "1499 dequantize_per_tensor_default_88"  [label="(4, 12, 64, 32)", style=solid];
-"1499 dequantize_per_tensor_default_88" -> "1507 matmul_28"  [label="(4, 12, 64, 32)", style=solid];
-"1500 linalg_vector_norm_29" -> "1501 clamp_min_29"  [label="(4, 12, 64, 1)", style=solid];
-"1501 clamp_min_29" -> "1502 expand_as_29"  [label="(4, 12, 64, 1)", style=solid];
-"1502 expand_as_29" -> "1503 div_29"  [label="(4, 12, 64, 32)", style=solid];
-"1503 div_29" -> "1504 quantize_per_tensor_default_89"  [label="(4, 12, 64, 32)", style=solid];
-"1504 quantize_per_tensor_default_89" -> "1505 dequantize_per_tensor_default_89"  [label="(4, 12, 64, 32)", style=solid];
-"1505 dequantize_per_tensor_default_89" -> "1506 transpose_28"  [label="(4, 12, 64, 32)", style=solid];
-"1506 transpose_28" -> "1507 matmul_28"  [label="(4, 12, 32, 64)", style=solid];
-"1507 matmul_28" -> "1509 mul_29"  [label="(4, 12, 64, 64)", style=solid];
-"1508 _frozen_param122" -> "1509 mul_29"  [label="(12, 1, 1)", style=solid];
-"1509 mul_29" -> "1510 add_49"  [label="(4, 12, 64, 64)", style=solid];
-"1510 add_49" -> "1511 softmax_14"  [label="(4, 12, 64, 64)", style=solid];
-"1511 softmax_14" -> "1512 dropout_56"  [label="(4, 12, 64, 64)", style=solid];
-"1512 dropout_56" -> "1513 matmul_29"  [label="(4, 12, 64, 64)", style=solid];
-"1513 matmul_29" -> "1514 transpose_29"  [label="(4, 12, 64, 32)", style=solid];
-"1514 transpose_29" -> "1515 reshape_65"  [label="(4, 64, 12, 32)", style=solid];
-"1515 reshape_65" -> "1516 reshape_65_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1516 reshape_65_0_0_nncf_smooth_quant_0" -> "1517 quantize_per_tensor_default_90"  [label="(4, 64, 384)", style=solid];
-"1517 quantize_per_tensor_default_90" -> "1518 dequantize_per_tensor_default_90"  [label="(4, 64, 384)", style=solid];
-"1518 dequantize_per_tensor_default_90" -> "1525 linear_89"  [label="(4, 64, 384)", style=solid];
-"1519 linear_89_scale_0" -> "1522 quantize_per_channel_default_75"  [label="(384,)", style=solid];
-"1519 linear_89_scale_0" -> "1523 dequantize_per_channel_default_75"  [label="(384,)", style=solid];
-"1520 linear_89_zero_point_0" -> "1522 quantize_per_channel_default_75"  [label="(384,)", style=solid];
-"1520 linear_89_zero_point_0" -> "1523 dequantize_per_channel_default_75"  [label="(384,)", style=solid];
-"1521 compressed_weight_updated_constant60" -> "1522 quantize_per_channel_default_75"  [label="(384, 384)", style=solid];
-"1522 quantize_per_channel_default_75" -> "1523 dequantize_per_channel_default_75"  [label="(384, 384)", style=solid];
-"1523 dequantize_per_channel_default_75" -> "1525 linear_89"  [label="(384, 384)", style=solid];
-"1524 _param_constant241_0_0" -> "1525 linear_89"  [label="(384,)", style=solid];
-"1525 linear_89" -> "1526 dropout_57"  [label="(4, 64, 384)", style=solid];
-"1526 dropout_57" -> "1527 view_80"  [label="(4, 64, 384)", style=solid];
-"1527 view_80" -> "1528 permute_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1528 permute_67" -> "1529 reshape_66"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1529 reshape_66" -> "1530 slice_218"  [label="(1, 16, 16, 384)", style=solid];
-"1530 slice_218" -> "1531 slice_219"  [label="(1, 16, 16, 384)", style=solid];
-"1531 slice_219" -> "1532 slice_220"  [label="(1, 14, 16, 384)", style=solid];
-"1532 slice_220" -> "1533 slice_221"  [label="(1, 14, 14, 384)", style=solid];
-"1533 slice_221" -> "1534 contiguous_27"  [label="(1, 14, 14, 384)", style=solid];
-"1534 contiguous_27" -> "1537 layer_norm_31"  [label="(1, 14, 14, 384)", style=solid];
-"1535 _param_constant242" -> "1537 layer_norm_31"  [label="(384,)", style=solid];
-"1536 _param_constant243" -> "1537 layer_norm_31"  [label="(384,)", style=solid];
-"1537 layer_norm_31" -> "1538 add_50"  [label="(1, 14, 14, 384)", style=solid];
-"1538 add_50" -> "1539 add_50_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1538 add_50" -> "1565 add_51"  [label="(1, 14, 14, 384)", style=solid];
-"1539 add_50_0_0_nncf_smooth_quant_0" -> "1540 quantize_per_tensor_default_91"  [label="(1, 14, 14, 384)", style=solid];
-"1540 quantize_per_tensor_default_91" -> "1541 dequantize_per_tensor_default_91"  [label="(1, 14, 14, 384)", style=solid];
-"1541 dequantize_per_tensor_default_91" -> "1548 linear_90"  [label="(1, 14, 14, 384)", style=solid];
-"1542 linear_90_scale_0" -> "1545 quantize_per_channel_default_76"  [label="(1536,)", style=solid];
-"1542 linear_90_scale_0" -> "1546 dequantize_per_channel_default_76"  [label="(1536,)", style=solid];
-"1543 linear_90_zero_point_0" -> "1545 quantize_per_channel_default_76"  [label="(1536,)", style=solid];
-"1543 linear_90_zero_point_0" -> "1546 dequantize_per_channel_default_76"  [label="(1536,)", style=solid];
-"1544 compressed_weight_updated_constant61" -> "1545 quantize_per_channel_default_76"  [label="(1536, 384)", style=solid];
-"1545 quantize_per_channel_default_76" -> "1546 dequantize_per_channel_default_76"  [label="(1536, 384)", style=solid];
-"1546 dequantize_per_channel_default_76" -> "1548 linear_90"  [label="(1536, 384)", style=solid];
-"1547 _param_constant245_0_0" -> "1548 linear_90"  [label="(1536,)", style=solid];
-"1548 linear_90" -> "1549 gelu_14"  [label="(1, 14, 14, 1536)", style=solid];
-"1549 gelu_14" -> "1550 dropout_58"  [label="(1, 14, 14, 1536)", style=solid];
-"1550 dropout_58" -> "1551 dropout_58_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1551 dropout_58_0_0_nncf_smooth_quant_0" -> "1552 quantize_per_tensor_default_92"  [label="(1, 14, 14, 1536)", style=solid];
-"1552 quantize_per_tensor_default_92" -> "1553 dequantize_per_tensor_default_92"  [label="(1, 14, 14, 1536)", style=solid];
-"1553 dequantize_per_tensor_default_92" -> "1560 linear_91"  [label="(1, 14, 14, 1536)", style=solid];
-"1554 linear_91_scale_0" -> "1557 quantize_per_channel_default_77"  [label="(384,)", style=solid];
-"1554 linear_91_scale_0" -> "1558 dequantize_per_channel_default_77"  [label="(384,)", style=solid];
-"1555 linear_91_zero_point_0" -> "1557 quantize_per_channel_default_77"  [label="(384,)", style=solid];
-"1555 linear_91_zero_point_0" -> "1558 dequantize_per_channel_default_77"  [label="(384,)", style=solid];
-"1556 compressed_weight_updated_constant62" -> "1557 quantize_per_channel_default_77"  [label="(384, 1536)", style=solid];
-"1557 quantize_per_channel_default_77" -> "1558 dequantize_per_channel_default_77"  [label="(384, 1536)", style=solid];
-"1558 dequantize_per_channel_default_77" -> "1560 linear_91"  [label="(384, 1536)", style=solid];
-"1559 _param_constant247_0_0" -> "1560 linear_91"  [label="(384,)", style=solid];
-"1560 linear_91" -> "1561 dropout_59"  [label="(1, 14, 14, 384)", style=solid];
-"1561 dropout_59" -> "1564 layer_norm_32"  [label="(1, 14, 14, 384)", style=solid];
-"1562 _param_constant248" -> "1564 layer_norm_32"  [label="(384,)", style=solid];
-"1563 _param_constant249" -> "1564 layer_norm_32"  [label="(384,)", style=solid];
-"1564 layer_norm_32" -> "1565 add_51"  [label="(1, 14, 14, 384)", style=solid];
-"1565 add_51" -> "1567 pad_17"  [label="(1, 14, 14, 384)", style=solid];
-"1565 add_51" -> "1648 add_54"  [label="(1, 14, 14, 384)", style=solid];
-"1566 _frozen_param123" -> "1603 add_52"  [label="(1, 12, 64, 64)", style=solid];
-"1567 pad_17" -> "1568 roll_14"  [label="(1, 16, 16, 384)", style=solid];
-"1568 roll_14" -> "1569 view_83"  [label="(1, 16, 16, 384)", style=solid];
-"1569 view_83" -> "1570 permute_69"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1570 permute_69" -> "1571 reshape_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1571 reshape_67" -> "1572 reshape_67_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1571 reshape_67" -> "1604 new_zeros_7"  [label="(4, 64, 384)", style=solid];
-"1572 reshape_67_0_0_nncf_smooth_quant_0" -> "1573 quantize_per_tensor_default_93"  [label="(4, 64, 384)", style=solid];
-"1573 quantize_per_tensor_default_93" -> "1574 dequantize_per_tensor_default_93"  [label="(4, 64, 384)", style=solid];
-"1574 dequantize_per_tensor_default_93" -> "1581 linear_94"  [label="(4, 64, 384)", style=solid];
-"1575 linear_94_scale_0" -> "1578 quantize_per_channel_default_79"  [label="(1152,)", style=solid];
-"1575 linear_94_scale_0" -> "1579 dequantize_per_channel_default_79"  [label="(1152,)", style=solid];
-"1576 linear_94_zero_point_0" -> "1578 quantize_per_channel_default_79"  [label="(1152,)", style=solid];
-"1576 linear_94_zero_point_0" -> "1579 dequantize_per_channel_default_79"  [label="(1152,)", style=solid];
-"1577 compressed_weight_updated_constant63" -> "1578 quantize_per_channel_default_79"  [label="(1152, 384)", style=solid];
-"1578 quantize_per_channel_default_79" -> "1579 dequantize_per_channel_default_79"  [label="(1152, 384)", style=solid];
-"1579 dequantize_per_channel_default_79" -> "1581 linear_94"  [label="(1152, 384)", style=solid];
-"1580 _frozen_param124_0_0" -> "1581 linear_94"  [label="(1152,)", style=solid];
-"1581 linear_94" -> "1582 reshape_68"  [label="(4, 64, 1152)", style=solid];
-"1582 reshape_68" -> "1583 permute_70"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1583 permute_70" -> "1584 select_45"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1583 permute_70" -> "1585 select_46"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1583 permute_70" -> "1586 select_47"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1584 select_45" -> "1587 linalg_vector_norm_30"  [label="(4, 12, 64, 32)", style=solid];
-"1584 select_45" -> "1589 expand_as_30"  [label="(4, 12, 64, 32)", style=solid];
-"1584 select_45" -> "1590 div_30"  [label="(4, 12, 64, 32)", style=solid];
-"1585 select_46" -> "1593 linalg_vector_norm_31"  [label="(4, 12, 64, 32)", style=solid];
-"1585 select_46" -> "1595 expand_as_31"  [label="(4, 12, 64, 32)", style=solid];
-"1585 select_46" -> "1596 div_31"  [label="(4, 12, 64, 32)", style=solid];
-"1586 select_47" -> "1622 matmul_31"  [label="(4, 12, 64, 32)", style=solid];
-"1587 linalg_vector_norm_30" -> "1588 clamp_min_30"  [label="(4, 12, 64, 1)", style=solid];
-"1588 clamp_min_30" -> "1589 expand_as_30"  [label="(4, 12, 64, 1)", style=solid];
-"1589 expand_as_30" -> "1590 div_30"  [label="(4, 12, 64, 32)", style=solid];
-"1590 div_30" -> "1591 quantize_per_tensor_default_94"  [label="(4, 12, 64, 32)", style=solid];
-"1591 quantize_per_tensor_default_94" -> "1592 dequantize_per_tensor_default_94"  [label="(4, 12, 64, 32)", style=solid];
-"1592 dequantize_per_tensor_default_94" -> "1600 matmul_30"  [label="(4, 12, 64, 32)", style=solid];
-"1593 linalg_vector_norm_31" -> "1594 clamp_min_31"  [label="(4, 12, 64, 1)", style=solid];
-"1594 clamp_min_31" -> "1595 expand_as_31"  [label="(4, 12, 64, 1)", style=solid];
-"1595 expand_as_31" -> "1596 div_31"  [label="(4, 12, 64, 32)", style=solid];
-"1596 div_31" -> "1597 quantize_per_tensor_default_95"  [label="(4, 12, 64, 32)", style=solid];
-"1597 quantize_per_tensor_default_95" -> "1598 dequantize_per_tensor_default_95"  [label="(4, 12, 64, 32)", style=solid];
-"1598 dequantize_per_tensor_default_95" -> "1599 transpose_30"  [label="(4, 12, 64, 32)", style=solid];
-"1599 transpose_30" -> "1600 matmul_30"  [label="(4, 12, 32, 64)", style=solid];
-"1600 matmul_30" -> "1602 mul_31"  [label="(4, 12, 64, 64)", style=solid];
-"1601 _frozen_param126" -> "1602 mul_31"  [label="(12, 1, 1)", style=solid];
-"1602 mul_31" -> "1603 add_52"  [label="(4, 12, 64, 64)", style=solid];
-"1603 add_52" -> "1615 view_85"  [label="(4, 12, 64, 64)", style=solid];
-"1604 new_zeros_7" -> "1605 view_84"  [label="(16, 16)", style=solid];
-"1605 view_84" -> "1606 permute_71"  [label="(2, 8, 2, 8)", style=solid];
-"1606 permute_71" -> "1607 reshape_69"  [label="(2, 2, 8, 8)", style=solid];
-"1607 reshape_69" -> "1608 unsqueeze_44"  [label="(4, 64)", style=solid];
-"1607 reshape_69" -> "1609 unsqueeze_45"  [label="(4, 64)", style=solid];
-"1608 unsqueeze_44" -> "1610 sub_7"  [label="(4, 1, 64)", style=solid];
-"1609 unsqueeze_45" -> "1610 sub_7"  [label="(4, 64, 1)", style=solid];
-"1610 sub_7" -> "1611 ne_7"  [label="(4, 64, 64)", style=solid];
-"1610 sub_7" -> "1612 masked_fill_14"  [label="(4, 64, 64)", style=solid];
-"1610 sub_7" -> "1613 eq_7"  [label="(4, 64, 64)", style=solid];
-"1611 ne_7" -> "1612 masked_fill_14"  [label="(4, 64, 64)", style=solid];
-"1612 masked_fill_14" -> "1614 masked_fill_15"  [label="(4, 64, 64)", style=solid];
-"1613 eq_7" -> "1614 masked_fill_15"  [label="(4, 64, 64)", style=solid];
-"1614 masked_fill_15" -> "1616 unsqueeze_46"  [label="(4, 64, 64)", style=solid];
-"1615 view_85" -> "1618 add_53"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1616 unsqueeze_46" -> "1617 unsqueeze_47"  [label="(4, 1, 64, 64)", style=solid];
-"1617 unsqueeze_47" -> "1618 add_53"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1618 add_53" -> "1619 view_86"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1619 view_86" -> "1620 softmax_15"  [label="(4, 12, 64, 64)", style=solid];
-"1620 softmax_15" -> "1621 dropout_60"  [label="(4, 12, 64, 64)", style=solid];
-"1621 dropout_60" -> "1622 matmul_31"  [label="(4, 12, 64, 64)", style=solid];
-"1622 matmul_31" -> "1623 transpose_31"  [label="(4, 12, 64, 32)", style=solid];
-"1623 transpose_31" -> "1624 reshape_70"  [label="(4, 64, 12, 32)", style=solid];
-"1624 reshape_70" -> "1625 reshape_70_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1625 reshape_70_0_0_nncf_smooth_quant_0" -> "1626 quantize_per_tensor_default_96"  [label="(4, 64, 384)", style=solid];
-"1626 quantize_per_tensor_default_96" -> "1627 dequantize_per_tensor_default_96"  [label="(4, 64, 384)", style=solid];
-"1627 dequantize_per_tensor_default_96" -> "1634 linear_95"  [label="(4, 64, 384)", style=solid];
-"1628 linear_95_scale_0" -> "1631 quantize_per_channel_default_80"  [label="(384,)", style=solid];
-"1628 linear_95_scale_0" -> "1632 dequantize_per_channel_default_80"  [label="(384,)", style=solid];
-"1629 linear_95_zero_point_0" -> "1631 quantize_per_channel_default_80"  [label="(384,)", style=solid];
-"1629 linear_95_zero_point_0" -> "1632 dequantize_per_channel_default_80"  [label="(384,)", style=solid];
-"1630 compressed_weight_updated_constant64" -> "1631 quantize_per_channel_default_80"  [label="(384, 384)", style=solid];
-"1631 quantize_per_channel_default_80" -> "1632 dequantize_per_channel_default_80"  [label="(384, 384)", style=solid];
-"1632 dequantize_per_channel_default_80" -> "1634 linear_95"  [label="(384, 384)", style=solid];
-"1633 _param_constant257_0_0" -> "1634 linear_95"  [label="(384,)", style=solid];
-"1634 linear_95" -> "1635 dropout_61"  [label="(4, 64, 384)", style=solid];
-"1635 dropout_61" -> "1636 view_87"  [label="(4, 64, 384)", style=solid];
-"1636 view_87" -> "1637 permute_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1637 permute_72" -> "1638 reshape_71"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1638 reshape_71" -> "1639 roll_15"  [label="(1, 16, 16, 384)", style=solid];
-"1639 roll_15" -> "1640 slice_241"  [label="(1, 16, 16, 384)", style=solid];
-"1640 slice_241" -> "1641 slice_242"  [label="(1, 16, 16, 384)", style=solid];
-"1641 slice_242" -> "1642 slice_243"  [label="(1, 14, 16, 384)", style=solid];
-"1642 slice_243" -> "1643 slice_244"  [label="(1, 14, 14, 384)", style=solid];
-"1643 slice_244" -> "1644 contiguous_29"  [label="(1, 14, 14, 384)", style=solid];
-"1644 contiguous_29" -> "1647 layer_norm_33"  [label="(1, 14, 14, 384)", style=solid];
-"1645 _param_constant258" -> "1647 layer_norm_33"  [label="(384,)", style=solid];
-"1646 _param_constant259" -> "1647 layer_norm_33"  [label="(384,)", style=solid];
-"1647 layer_norm_33" -> "1648 add_54"  [label="(1, 14, 14, 384)", style=solid];
-"1648 add_54" -> "1649 add_54_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1648 add_54" -> "1675 add_55"  [label="(1, 14, 14, 384)", style=solid];
-"1649 add_54_0_0_nncf_smooth_quant_0" -> "1650 quantize_per_tensor_default_97"  [label="(1, 14, 14, 384)", style=solid];
-"1650 quantize_per_tensor_default_97" -> "1651 dequantize_per_tensor_default_97"  [label="(1, 14, 14, 384)", style=solid];
-"1651 dequantize_per_tensor_default_97" -> "1658 linear_96"  [label="(1, 14, 14, 384)", style=solid];
-"1652 linear_96_scale_0" -> "1655 quantize_per_channel_default_81"  [label="(1536,)", style=solid];
-"1652 linear_96_scale_0" -> "1656 dequantize_per_channel_default_81"  [label="(1536,)", style=solid];
-"1653 linear_96_zero_point_0" -> "1655 quantize_per_channel_default_81"  [label="(1536,)", style=solid];
-"1653 linear_96_zero_point_0" -> "1656 dequantize_per_channel_default_81"  [label="(1536,)", style=solid];
-"1654 compressed_weight_updated_constant65" -> "1655 quantize_per_channel_default_81"  [label="(1536, 384)", style=solid];
-"1655 quantize_per_channel_default_81" -> "1656 dequantize_per_channel_default_81"  [label="(1536, 384)", style=solid];
-"1656 dequantize_per_channel_default_81" -> "1658 linear_96"  [label="(1536, 384)", style=solid];
-"1657 _param_constant261_0_0" -> "1658 linear_96"  [label="(1536,)", style=solid];
-"1658 linear_96" -> "1659 gelu_15"  [label="(1, 14, 14, 1536)", style=solid];
-"1659 gelu_15" -> "1660 dropout_62"  [label="(1, 14, 14, 1536)", style=solid];
-"1660 dropout_62" -> "1661 dropout_62_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1661 dropout_62_0_0_nncf_smooth_quant_0" -> "1662 quantize_per_tensor_default_98"  [label="(1, 14, 14, 1536)", style=solid];
-"1662 quantize_per_tensor_default_98" -> "1663 dequantize_per_tensor_default_98"  [label="(1, 14, 14, 1536)", style=solid];
-"1663 dequantize_per_tensor_default_98" -> "1670 linear_97"  [label="(1, 14, 14, 1536)", style=solid];
-"1664 linear_97_scale_0" -> "1667 quantize_per_channel_default_82"  [label="(384,)", style=solid];
-"1664 linear_97_scale_0" -> "1668 dequantize_per_channel_default_82"  [label="(384,)", style=solid];
-"1665 linear_97_zero_point_0" -> "1667 quantize_per_channel_default_82"  [label="(384,)", style=solid];
-"1665 linear_97_zero_point_0" -> "1668 dequantize_per_channel_default_82"  [label="(384,)", style=solid];
-"1666 compressed_weight_updated_constant66" -> "1667 quantize_per_channel_default_82"  [label="(384, 1536)", style=solid];
-"1667 quantize_per_channel_default_82" -> "1668 dequantize_per_channel_default_82"  [label="(384, 1536)", style=solid];
-"1668 dequantize_per_channel_default_82" -> "1670 linear_97"  [label="(384, 1536)", style=solid];
-"1669 _param_constant263_0_0" -> "1670 linear_97"  [label="(384,)", style=solid];
-"1670 linear_97" -> "1671 dropout_63"  [label="(1, 14, 14, 384)", style=solid];
-"1671 dropout_63" -> "1674 layer_norm_34"  [label="(1, 14, 14, 384)", style=solid];
-"1672 _param_constant264" -> "1674 layer_norm_34"  [label="(384,)", style=solid];
-"1673 _param_constant265" -> "1674 layer_norm_34"  [label="(384,)", style=solid];
-"1674 layer_norm_34" -> "1675 add_55"  [label="(1, 14, 14, 384)", style=solid];
-"1675 add_55" -> "1677 pad_18"  [label="(1, 14, 14, 384)", style=solid];
-"1675 add_55" -> "1740 add_57"  [label="(1, 14, 14, 384)", style=solid];
-"1676 _frozen_param136" -> "1712 add_56"  [label="(1, 12, 64, 64)", style=solid];
-"1677 pad_18" -> "1678 view_90"  [label="(1, 16, 16, 384)", style=solid];
-"1678 view_90" -> "1679 permute_74"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1679 permute_74" -> "1680 reshape_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1680 reshape_72" -> "1681 reshape_72_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1681 reshape_72_0_0_nncf_smooth_quant_0" -> "1682 quantize_per_tensor_default_99"  [label="(4, 64, 384)", style=solid];
-"1682 quantize_per_tensor_default_99" -> "1683 dequantize_per_tensor_default_99"  [label="(4, 64, 384)", style=solid];
-"1683 dequantize_per_tensor_default_99" -> "1690 linear_100"  [label="(4, 64, 384)", style=solid];
-"1684 linear_100_scale_0" -> "1687 quantize_per_channel_default_84"  [label="(1152,)", style=solid];
-"1684 linear_100_scale_0" -> "1688 dequantize_per_channel_default_84"  [label="(1152,)", style=solid];
-"1685 linear_100_zero_point_0" -> "1687 quantize_per_channel_default_84"  [label="(1152,)", style=solid];
-"1685 linear_100_zero_point_0" -> "1688 dequantize_per_channel_default_84"  [label="(1152,)", style=solid];
-"1686 compressed_weight_updated_constant67" -> "1687 quantize_per_channel_default_84"  [label="(1152, 384)", style=solid];
-"1687 quantize_per_channel_default_84" -> "1688 dequantize_per_channel_default_84"  [label="(1152, 384)", style=solid];
-"1688 dequantize_per_channel_default_84" -> "1690 linear_100"  [label="(1152, 384)", style=solid];
-"1689 _frozen_param137_0_0" -> "1690 linear_100"  [label="(1152,)", style=solid];
-"1690 linear_100" -> "1691 reshape_73"  [label="(4, 64, 1152)", style=solid];
-"1691 reshape_73" -> "1692 permute_75"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1692 permute_75" -> "1693 select_48"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1692 permute_75" -> "1694 select_49"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1692 permute_75" -> "1695 select_50"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1693 select_48" -> "1696 linalg_vector_norm_32"  [label="(4, 12, 64, 32)", style=solid];
-"1693 select_48" -> "1698 expand_as_32"  [label="(4, 12, 64, 32)", style=solid];
-"1693 select_48" -> "1699 div_32"  [label="(4, 12, 64, 32)", style=solid];
-"1694 select_49" -> "1702 linalg_vector_norm_33"  [label="(4, 12, 64, 32)", style=solid];
-"1694 select_49" -> "1704 expand_as_33"  [label="(4, 12, 64, 32)", style=solid];
-"1694 select_49" -> "1705 div_33"  [label="(4, 12, 64, 32)", style=solid];
-"1695 select_50" -> "1715 matmul_33"  [label="(4, 12, 64, 32)", style=solid];
-"1696 linalg_vector_norm_32" -> "1697 clamp_min_32"  [label="(4, 12, 64, 1)", style=solid];
-"1697 clamp_min_32" -> "1698 expand_as_32"  [label="(4, 12, 64, 1)", style=solid];
-"1698 expand_as_32" -> "1699 div_32"  [label="(4, 12, 64, 32)", style=solid];
-"1699 div_32" -> "1700 quantize_per_tensor_default_100"  [label="(4, 12, 64, 32)", style=solid];
-"1700 quantize_per_tensor_default_100" -> "1701 dequantize_per_tensor_default_100"  [label="(4, 12, 64, 32)", style=solid];
-"1701 dequantize_per_tensor_default_100" -> "1709 matmul_32"  [label="(4, 12, 64, 32)", style=solid];
-"1702 linalg_vector_norm_33" -> "1703 clamp_min_33"  [label="(4, 12, 64, 1)", style=solid];
-"1703 clamp_min_33" -> "1704 expand_as_33"  [label="(4, 12, 64, 1)", style=solid];
-"1704 expand_as_33" -> "1705 div_33"  [label="(4, 12, 64, 32)", style=solid];
-"1705 div_33" -> "1706 quantize_per_tensor_default_101"  [label="(4, 12, 64, 32)", style=solid];
-"1706 quantize_per_tensor_default_101" -> "1707 dequantize_per_tensor_default_101"  [label="(4, 12, 64, 32)", style=solid];
-"1707 dequantize_per_tensor_default_101" -> "1708 transpose_32"  [label="(4, 12, 64, 32)", style=solid];
-"1708 transpose_32" -> "1709 matmul_32"  [label="(4, 12, 32, 64)", style=solid];
-"1709 matmul_32" -> "1711 mul_33"  [label="(4, 12, 64, 64)", style=solid];
-"1710 _frozen_param139" -> "1711 mul_33"  [label="(12, 1, 1)", style=solid];
-"1711 mul_33" -> "1712 add_56"  [label="(4, 12, 64, 64)", style=solid];
-"1712 add_56" -> "1713 softmax_16"  [label="(4, 12, 64, 64)", style=solid];
-"1713 softmax_16" -> "1714 dropout_64"  [label="(4, 12, 64, 64)", style=solid];
-"1714 dropout_64" -> "1715 matmul_33"  [label="(4, 12, 64, 64)", style=solid];
-"1715 matmul_33" -> "1716 transpose_33"  [label="(4, 12, 64, 32)", style=solid];
-"1716 transpose_33" -> "1717 reshape_74"  [label="(4, 64, 12, 32)", style=solid];
-"1717 reshape_74" -> "1718 reshape_74_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1718 reshape_74_0_0_nncf_smooth_quant_0" -> "1719 quantize_per_tensor_default_102"  [label="(4, 64, 384)", style=solid];
-"1719 quantize_per_tensor_default_102" -> "1720 dequantize_per_tensor_default_102"  [label="(4, 64, 384)", style=solid];
-"1720 dequantize_per_tensor_default_102" -> "1727 linear_101"  [label="(4, 64, 384)", style=solid];
-"1721 linear_101_scale_0" -> "1724 quantize_per_channel_default_85"  [label="(384,)", style=solid];
-"1721 linear_101_scale_0" -> "1725 dequantize_per_channel_default_85"  [label="(384,)", style=solid];
-"1722 linear_101_zero_point_0" -> "1724 quantize_per_channel_default_85"  [label="(384,)", style=solid];
-"1722 linear_101_zero_point_0" -> "1725 dequantize_per_channel_default_85"  [label="(384,)", style=solid];
-"1723 compressed_weight_updated_constant68" -> "1724 quantize_per_channel_default_85"  [label="(384, 384)", style=solid];
-"1724 quantize_per_channel_default_85" -> "1725 dequantize_per_channel_default_85"  [label="(384, 384)", style=solid];
-"1725 dequantize_per_channel_default_85" -> "1727 linear_101"  [label="(384, 384)", style=solid];
-"1726 _param_constant273_0_0" -> "1727 linear_101"  [label="(384,)", style=solid];
-"1727 linear_101" -> "1728 dropout_65"  [label="(4, 64, 384)", style=solid];
-"1728 dropout_65" -> "1729 view_91"  [label="(4, 64, 384)", style=solid];
-"1729 view_91" -> "1730 permute_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1730 permute_76" -> "1731 reshape_75"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1731 reshape_75" -> "1732 slice_246"  [label="(1, 16, 16, 384)", style=solid];
-"1732 slice_246" -> "1733 slice_247"  [label="(1, 16, 16, 384)", style=solid];
-"1733 slice_247" -> "1734 slice_248"  [label="(1, 14, 16, 384)", style=solid];
-"1734 slice_248" -> "1735 slice_249"  [label="(1, 14, 14, 384)", style=solid];
-"1735 slice_249" -> "1736 contiguous_31"  [label="(1, 14, 14, 384)", style=solid];
-"1736 contiguous_31" -> "1739 layer_norm_35"  [label="(1, 14, 14, 384)", style=solid];
-"1737 _param_constant274" -> "1739 layer_norm_35"  [label="(384,)", style=solid];
-"1738 _param_constant275" -> "1739 layer_norm_35"  [label="(384,)", style=solid];
-"1739 layer_norm_35" -> "1740 add_57"  [label="(1, 14, 14, 384)", style=solid];
-"1740 add_57" -> "1741 add_57_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1740 add_57" -> "1767 add_58"  [label="(1, 14, 14, 384)", style=solid];
-"1741 add_57_0_0_nncf_smooth_quant_0" -> "1742 quantize_per_tensor_default_103"  [label="(1, 14, 14, 384)", style=solid];
-"1742 quantize_per_tensor_default_103" -> "1743 dequantize_per_tensor_default_103"  [label="(1, 14, 14, 384)", style=solid];
-"1743 dequantize_per_tensor_default_103" -> "1750 linear_102"  [label="(1, 14, 14, 384)", style=solid];
-"1744 linear_102_scale_0" -> "1747 quantize_per_channel_default_86"  [label="(1536,)", style=solid];
-"1744 linear_102_scale_0" -> "1748 dequantize_per_channel_default_86"  [label="(1536,)", style=solid];
-"1745 linear_102_zero_point_0" -> "1747 quantize_per_channel_default_86"  [label="(1536,)", style=solid];
-"1745 linear_102_zero_point_0" -> "1748 dequantize_per_channel_default_86"  [label="(1536,)", style=solid];
-"1746 compressed_weight_updated_constant69" -> "1747 quantize_per_channel_default_86"  [label="(1536, 384)", style=solid];
-"1747 quantize_per_channel_default_86" -> "1748 dequantize_per_channel_default_86"  [label="(1536, 384)", style=solid];
-"1748 dequantize_per_channel_default_86" -> "1750 linear_102"  [label="(1536, 384)", style=solid];
-"1749 _param_constant277_0_0" -> "1750 linear_102"  [label="(1536,)", style=solid];
-"1750 linear_102" -> "1751 gelu_16"  [label="(1, 14, 14, 1536)", style=solid];
-"1751 gelu_16" -> "1752 dropout_66"  [label="(1, 14, 14, 1536)", style=solid];
-"1752 dropout_66" -> "1753 dropout_66_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1753 dropout_66_0_0_nncf_smooth_quant_0" -> "1754 quantize_per_tensor_default_104"  [label="(1, 14, 14, 1536)", style=solid];
-"1754 quantize_per_tensor_default_104" -> "1755 dequantize_per_tensor_default_104"  [label="(1, 14, 14, 1536)", style=solid];
-"1755 dequantize_per_tensor_default_104" -> "1762 linear_103"  [label="(1, 14, 14, 1536)", style=solid];
-"1756 linear_103_scale_0" -> "1759 quantize_per_channel_default_87"  [label="(384,)", style=solid];
-"1756 linear_103_scale_0" -> "1760 dequantize_per_channel_default_87"  [label="(384,)", style=solid];
-"1757 linear_103_zero_point_0" -> "1759 quantize_per_channel_default_87"  [label="(384,)", style=solid];
-"1757 linear_103_zero_point_0" -> "1760 dequantize_per_channel_default_87"  [label="(384,)", style=solid];
-"1758 compressed_weight_updated_constant70" -> "1759 quantize_per_channel_default_87"  [label="(384, 1536)", style=solid];
-"1759 quantize_per_channel_default_87" -> "1760 dequantize_per_channel_default_87"  [label="(384, 1536)", style=solid];
-"1760 dequantize_per_channel_default_87" -> "1762 linear_103"  [label="(384, 1536)", style=solid];
-"1761 _param_constant279_0_0" -> "1762 linear_103"  [label="(384,)", style=solid];
-"1762 linear_103" -> "1763 dropout_67"  [label="(1, 14, 14, 384)", style=solid];
-"1763 dropout_67" -> "1766 layer_norm_36"  [label="(1, 14, 14, 384)", style=solid];
-"1764 _param_constant280" -> "1766 layer_norm_36"  [label="(384,)", style=solid];
-"1765 _param_constant281" -> "1766 layer_norm_36"  [label="(384,)", style=solid];
-"1766 layer_norm_36" -> "1767 add_58"  [label="(1, 14, 14, 384)", style=solid];
-"1767 add_58" -> "1769 pad_19"  [label="(1, 14, 14, 384)", style=solid];
-"1767 add_58" -> "1850 add_61"  [label="(1, 14, 14, 384)", style=solid];
-"1768 _frozen_param140" -> "1805 add_59"  [label="(1, 12, 64, 64)", style=solid];
-"1769 pad_19" -> "1770 roll_16"  [label="(1, 16, 16, 384)", style=solid];
-"1770 roll_16" -> "1771 view_94"  [label="(1, 16, 16, 384)", style=solid];
-"1771 view_94" -> "1772 permute_78"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1772 permute_78" -> "1773 reshape_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1773 reshape_76" -> "1774 reshape_76_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1773 reshape_76" -> "1806 new_zeros_8"  [label="(4, 64, 384)", style=solid];
-"1774 reshape_76_0_0_nncf_smooth_quant_0" -> "1775 quantize_per_tensor_default_105"  [label="(4, 64, 384)", style=solid];
-"1775 quantize_per_tensor_default_105" -> "1776 dequantize_per_tensor_default_105"  [label="(4, 64, 384)", style=solid];
-"1776 dequantize_per_tensor_default_105" -> "1783 linear_106"  [label="(4, 64, 384)", style=solid];
-"1777 linear_106_scale_0" -> "1780 quantize_per_channel_default_89"  [label="(1152,)", style=solid];
-"1777 linear_106_scale_0" -> "1781 dequantize_per_channel_default_89"  [label="(1152,)", style=solid];
-"1778 linear_106_zero_point_0" -> "1780 quantize_per_channel_default_89"  [label="(1152,)", style=solid];
-"1778 linear_106_zero_point_0" -> "1781 dequantize_per_channel_default_89"  [label="(1152,)", style=solid];
-"1779 compressed_weight_updated_constant71" -> "1780 quantize_per_channel_default_89"  [label="(1152, 384)", style=solid];
-"1780 quantize_per_channel_default_89" -> "1781 dequantize_per_channel_default_89"  [label="(1152, 384)", style=solid];
-"1781 dequantize_per_channel_default_89" -> "1783 linear_106"  [label="(1152, 384)", style=solid];
-"1782 _frozen_param141_0_0" -> "1783 linear_106"  [label="(1152,)", style=solid];
-"1783 linear_106" -> "1784 reshape_77"  [label="(4, 64, 1152)", style=solid];
-"1784 reshape_77" -> "1785 permute_79"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1785 permute_79" -> "1786 select_51"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1785 permute_79" -> "1787 select_52"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1785 permute_79" -> "1788 select_53"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1786 select_51" -> "1789 linalg_vector_norm_34"  [label="(4, 12, 64, 32)", style=solid];
-"1786 select_51" -> "1791 expand_as_34"  [label="(4, 12, 64, 32)", style=solid];
-"1786 select_51" -> "1792 div_34"  [label="(4, 12, 64, 32)", style=solid];
-"1787 select_52" -> "1795 linalg_vector_norm_35"  [label="(4, 12, 64, 32)", style=solid];
-"1787 select_52" -> "1797 expand_as_35"  [label="(4, 12, 64, 32)", style=solid];
-"1787 select_52" -> "1798 div_35"  [label="(4, 12, 64, 32)", style=solid];
-"1788 select_53" -> "1824 matmul_35"  [label="(4, 12, 64, 32)", style=solid];
-"1789 linalg_vector_norm_34" -> "1790 clamp_min_34"  [label="(4, 12, 64, 1)", style=solid];
-"1790 clamp_min_34" -> "1791 expand_as_34"  [label="(4, 12, 64, 1)", style=solid];
-"1791 expand_as_34" -> "1792 div_34"  [label="(4, 12, 64, 32)", style=solid];
-"1792 div_34" -> "1793 quantize_per_tensor_default_106"  [label="(4, 12, 64, 32)", style=solid];
-"1793 quantize_per_tensor_default_106" -> "1794 dequantize_per_tensor_default_106"  [label="(4, 12, 64, 32)", style=solid];
-"1794 dequantize_per_tensor_default_106" -> "1802 matmul_34"  [label="(4, 12, 64, 32)", style=solid];
-"1795 linalg_vector_norm_35" -> "1796 clamp_min_35"  [label="(4, 12, 64, 1)", style=solid];
-"1796 clamp_min_35" -> "1797 expand_as_35"  [label="(4, 12, 64, 1)", style=solid];
-"1797 expand_as_35" -> "1798 div_35"  [label="(4, 12, 64, 32)", style=solid];
-"1798 div_35" -> "1799 quantize_per_tensor_default_107"  [label="(4, 12, 64, 32)", style=solid];
-"1799 quantize_per_tensor_default_107" -> "1800 dequantize_per_tensor_default_107"  [label="(4, 12, 64, 32)", style=solid];
-"1800 dequantize_per_tensor_default_107" -> "1801 transpose_34"  [label="(4, 12, 64, 32)", style=solid];
-"1801 transpose_34" -> "1802 matmul_34"  [label="(4, 12, 32, 64)", style=solid];
-"1802 matmul_34" -> "1804 mul_35"  [label="(4, 12, 64, 64)", style=solid];
-"1803 _frozen_param143" -> "1804 mul_35"  [label="(12, 1, 1)", style=solid];
-"1804 mul_35" -> "1805 add_59"  [label="(4, 12, 64, 64)", style=solid];
-"1805 add_59" -> "1817 view_96"  [label="(4, 12, 64, 64)", style=solid];
-"1806 new_zeros_8" -> "1807 view_95"  [label="(16, 16)", style=solid];
-"1807 view_95" -> "1808 permute_80"  [label="(2, 8, 2, 8)", style=solid];
-"1808 permute_80" -> "1809 reshape_78"  [label="(2, 2, 8, 8)", style=solid];
-"1809 reshape_78" -> "1810 unsqueeze_50"  [label="(4, 64)", style=solid];
-"1809 reshape_78" -> "1811 unsqueeze_51"  [label="(4, 64)", style=solid];
-"1810 unsqueeze_50" -> "1812 sub_8"  [label="(4, 1, 64)", style=solid];
-"1811 unsqueeze_51" -> "1812 sub_8"  [label="(4, 64, 1)", style=solid];
-"1812 sub_8" -> "1813 ne_8"  [label="(4, 64, 64)", style=solid];
-"1812 sub_8" -> "1814 masked_fill_16"  [label="(4, 64, 64)", style=solid];
-"1812 sub_8" -> "1815 eq_8"  [label="(4, 64, 64)", style=solid];
-"1813 ne_8" -> "1814 masked_fill_16"  [label="(4, 64, 64)", style=solid];
-"1814 masked_fill_16" -> "1816 masked_fill_17"  [label="(4, 64, 64)", style=solid];
-"1815 eq_8" -> "1816 masked_fill_17"  [label="(4, 64, 64)", style=solid];
-"1816 masked_fill_17" -> "1818 unsqueeze_52"  [label="(4, 64, 64)", style=solid];
-"1817 view_96" -> "1820 add_60"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1818 unsqueeze_52" -> "1819 unsqueeze_53"  [label="(4, 1, 64, 64)", style=solid];
-"1819 unsqueeze_53" -> "1820 add_60"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1820 add_60" -> "1821 view_97"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1821 view_97" -> "1822 softmax_17"  [label="(4, 12, 64, 64)", style=solid];
-"1822 softmax_17" -> "1823 dropout_68"  [label="(4, 12, 64, 64)", style=solid];
-"1823 dropout_68" -> "1824 matmul_35"  [label="(4, 12, 64, 64)", style=solid];
-"1824 matmul_35" -> "1825 transpose_35"  [label="(4, 12, 64, 32)", style=solid];
-"1825 transpose_35" -> "1826 reshape_79"  [label="(4, 64, 12, 32)", style=solid];
-"1826 reshape_79" -> "1827 reshape_79_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1827 reshape_79_0_0_nncf_smooth_quant_0" -> "1828 quantize_per_tensor_default_108"  [label="(4, 64, 384)", style=solid];
-"1828 quantize_per_tensor_default_108" -> "1829 dequantize_per_tensor_default_108"  [label="(4, 64, 384)", style=solid];
-"1829 dequantize_per_tensor_default_108" -> "1836 linear_107"  [label="(4, 64, 384)", style=solid];
-"1830 linear_107_scale_0" -> "1833 quantize_per_channel_default_90"  [label="(384,)", style=solid];
-"1830 linear_107_scale_0" -> "1834 dequantize_per_channel_default_90"  [label="(384,)", style=solid];
-"1831 linear_107_zero_point_0" -> "1833 quantize_per_channel_default_90"  [label="(384,)", style=solid];
-"1831 linear_107_zero_point_0" -> "1834 dequantize_per_channel_default_90"  [label="(384,)", style=solid];
-"1832 compressed_weight_updated_constant72" -> "1833 quantize_per_channel_default_90"  [label="(384, 384)", style=solid];
-"1833 quantize_per_channel_default_90" -> "1834 dequantize_per_channel_default_90"  [label="(384, 384)", style=solid];
-"1834 dequantize_per_channel_default_90" -> "1836 linear_107"  [label="(384, 384)", style=solid];
-"1835 _param_constant289_0_0" -> "1836 linear_107"  [label="(384,)", style=solid];
-"1836 linear_107" -> "1837 dropout_69"  [label="(4, 64, 384)", style=solid];
-"1837 dropout_69" -> "1838 view_98"  [label="(4, 64, 384)", style=solid];
-"1838 view_98" -> "1839 permute_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1839 permute_81" -> "1840 reshape_80"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1840 reshape_80" -> "1841 roll_17"  [label="(1, 16, 16, 384)", style=solid];
-"1841 roll_17" -> "1842 slice_269"  [label="(1, 16, 16, 384)", style=solid];
-"1842 slice_269" -> "1843 slice_270"  [label="(1, 16, 16, 384)", style=solid];
-"1843 slice_270" -> "1844 slice_271"  [label="(1, 14, 16, 384)", style=solid];
-"1844 slice_271" -> "1845 slice_272"  [label="(1, 14, 14, 384)", style=solid];
-"1845 slice_272" -> "1846 contiguous_33"  [label="(1, 14, 14, 384)", style=solid];
-"1846 contiguous_33" -> "1849 layer_norm_37"  [label="(1, 14, 14, 384)", style=solid];
-"1847 _param_constant290" -> "1849 layer_norm_37"  [label="(384,)", style=solid];
-"1848 _param_constant291" -> "1849 layer_norm_37"  [label="(384,)", style=solid];
-"1849 layer_norm_37" -> "1850 add_61"  [label="(1, 14, 14, 384)", style=solid];
-"1850 add_61" -> "1851 add_61_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1850 add_61" -> "1877 add_62"  [label="(1, 14, 14, 384)", style=solid];
-"1851 add_61_0_0_nncf_smooth_quant_0" -> "1852 quantize_per_tensor_default_109"  [label="(1, 14, 14, 384)", style=solid];
-"1852 quantize_per_tensor_default_109" -> "1853 dequantize_per_tensor_default_109"  [label="(1, 14, 14, 384)", style=solid];
-"1853 dequantize_per_tensor_default_109" -> "1860 linear_108"  [label="(1, 14, 14, 384)", style=solid];
-"1854 linear_108_scale_0" -> "1857 quantize_per_channel_default_91"  [label="(1536,)", style=solid];
-"1854 linear_108_scale_0" -> "1858 dequantize_per_channel_default_91"  [label="(1536,)", style=solid];
-"1855 linear_108_zero_point_0" -> "1857 quantize_per_channel_default_91"  [label="(1536,)", style=solid];
-"1855 linear_108_zero_point_0" -> "1858 dequantize_per_channel_default_91"  [label="(1536,)", style=solid];
-"1856 compressed_weight_updated_constant73" -> "1857 quantize_per_channel_default_91"  [label="(1536, 384)", style=solid];
-"1857 quantize_per_channel_default_91" -> "1858 dequantize_per_channel_default_91"  [label="(1536, 384)", style=solid];
-"1858 dequantize_per_channel_default_91" -> "1860 linear_108"  [label="(1536, 384)", style=solid];
-"1859 _param_constant293_0_0" -> "1860 linear_108"  [label="(1536,)", style=solid];
-"1860 linear_108" -> "1861 gelu_17"  [label="(1, 14, 14, 1536)", style=solid];
-"1861 gelu_17" -> "1862 dropout_70"  [label="(1, 14, 14, 1536)", style=solid];
-"1862 dropout_70" -> "1863 dropout_70_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1863 dropout_70_0_0_nncf_smooth_quant_0" -> "1864 quantize_per_tensor_default_110"  [label="(1, 14, 14, 1536)", style=solid];
-"1864 quantize_per_tensor_default_110" -> "1865 dequantize_per_tensor_default_110"  [label="(1, 14, 14, 1536)", style=solid];
-"1865 dequantize_per_tensor_default_110" -> "1872 linear_109"  [label="(1, 14, 14, 1536)", style=solid];
-"1866 linear_109_scale_0" -> "1869 quantize_per_channel_default_92"  [label="(384,)", style=solid];
-"1866 linear_109_scale_0" -> "1870 dequantize_per_channel_default_92"  [label="(384,)", style=solid];
-"1867 linear_109_zero_point_0" -> "1869 quantize_per_channel_default_92"  [label="(384,)", style=solid];
-"1867 linear_109_zero_point_0" -> "1870 dequantize_per_channel_default_92"  [label="(384,)", style=solid];
-"1868 compressed_weight_updated_constant74" -> "1869 quantize_per_channel_default_92"  [label="(384, 1536)", style=solid];
-"1869 quantize_per_channel_default_92" -> "1870 dequantize_per_channel_default_92"  [label="(384, 1536)", style=solid];
-"1870 dequantize_per_channel_default_92" -> "1872 linear_109"  [label="(384, 1536)", style=solid];
-"1871 _param_constant295_0_0" -> "1872 linear_109"  [label="(384,)", style=solid];
-"1872 linear_109" -> "1873 dropout_71"  [label="(1, 14, 14, 384)", style=solid];
-"1873 dropout_71" -> "1876 layer_norm_38"  [label="(1, 14, 14, 384)", style=solid];
-"1874 _param_constant296" -> "1876 layer_norm_38"  [label="(384,)", style=solid];
-"1875 _param_constant297" -> "1876 layer_norm_38"  [label="(384,)", style=solid];
-"1876 layer_norm_38" -> "1877 add_62"  [label="(1, 14, 14, 384)", style=solid];
-"1877 add_62" -> "1879 pad_20"  [label="(1, 14, 14, 384)", style=solid];
-"1877 add_62" -> "1942 add_64"  [label="(1, 14, 14, 384)", style=solid];
-"1878 _frozen_param153" -> "1914 add_63"  [label="(1, 12, 64, 64)", style=solid];
-"1879 pad_20" -> "1880 view_101"  [label="(1, 16, 16, 384)", style=solid];
-"1880 view_101" -> "1881 permute_83"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1881 permute_83" -> "1882 reshape_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1882 reshape_81" -> "1883 reshape_81_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1883 reshape_81_0_0_nncf_smooth_quant_0" -> "1884 quantize_per_tensor_default_111"  [label="(4, 64, 384)", style=solid];
-"1884 quantize_per_tensor_default_111" -> "1885 dequantize_per_tensor_default_111"  [label="(4, 64, 384)", style=solid];
-"1885 dequantize_per_tensor_default_111" -> "1892 linear_112"  [label="(4, 64, 384)", style=solid];
-"1886 linear_112_scale_0" -> "1889 quantize_per_channel_default_94"  [label="(1152,)", style=solid];
-"1886 linear_112_scale_0" -> "1890 dequantize_per_channel_default_94"  [label="(1152,)", style=solid];
-"1887 linear_112_zero_point_0" -> "1889 quantize_per_channel_default_94"  [label="(1152,)", style=solid];
-"1887 linear_112_zero_point_0" -> "1890 dequantize_per_channel_default_94"  [label="(1152,)", style=solid];
-"1888 compressed_weight_updated_constant75" -> "1889 quantize_per_channel_default_94"  [label="(1152, 384)", style=solid];
-"1889 quantize_per_channel_default_94" -> "1890 dequantize_per_channel_default_94"  [label="(1152, 384)", style=solid];
-"1890 dequantize_per_channel_default_94" -> "1892 linear_112"  [label="(1152, 384)", style=solid];
-"1891 _frozen_param154_0_0" -> "1892 linear_112"  [label="(1152,)", style=solid];
-"1892 linear_112" -> "1893 reshape_82"  [label="(4, 64, 1152)", style=solid];
-"1893 reshape_82" -> "1894 permute_84"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1894 permute_84" -> "1895 select_54"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1894 permute_84" -> "1896 select_55"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1894 permute_84" -> "1897 select_56"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1895 select_54" -> "1898 linalg_vector_norm_36"  [label="(4, 12, 64, 32)", style=solid];
-"1895 select_54" -> "1900 expand_as_36"  [label="(4, 12, 64, 32)", style=solid];
-"1895 select_54" -> "1901 div_36"  [label="(4, 12, 64, 32)", style=solid];
-"1896 select_55" -> "1904 linalg_vector_norm_37"  [label="(4, 12, 64, 32)", style=solid];
-"1896 select_55" -> "1906 expand_as_37"  [label="(4, 12, 64, 32)", style=solid];
-"1896 select_55" -> "1907 div_37"  [label="(4, 12, 64, 32)", style=solid];
-"1897 select_56" -> "1917 matmul_37"  [label="(4, 12, 64, 32)", style=solid];
-"1898 linalg_vector_norm_36" -> "1899 clamp_min_36"  [label="(4, 12, 64, 1)", style=solid];
-"1899 clamp_min_36" -> "1900 expand_as_36"  [label="(4, 12, 64, 1)", style=solid];
-"1900 expand_as_36" -> "1901 div_36"  [label="(4, 12, 64, 32)", style=solid];
-"1901 div_36" -> "1902 quantize_per_tensor_default_112"  [label="(4, 12, 64, 32)", style=solid];
-"1902 quantize_per_tensor_default_112" -> "1903 dequantize_per_tensor_default_112"  [label="(4, 12, 64, 32)", style=solid];
-"1903 dequantize_per_tensor_default_112" -> "1911 matmul_36"  [label="(4, 12, 64, 32)", style=solid];
-"1904 linalg_vector_norm_37" -> "1905 clamp_min_37"  [label="(4, 12, 64, 1)", style=solid];
-"1905 clamp_min_37" -> "1906 expand_as_37"  [label="(4, 12, 64, 1)", style=solid];
-"1906 expand_as_37" -> "1907 div_37"  [label="(4, 12, 64, 32)", style=solid];
-"1907 div_37" -> "1908 quantize_per_tensor_default_113"  [label="(4, 12, 64, 32)", style=solid];
-"1908 quantize_per_tensor_default_113" -> "1909 dequantize_per_tensor_default_113"  [label="(4, 12, 64, 32)", style=solid];
-"1909 dequantize_per_tensor_default_113" -> "1910 transpose_36"  [label="(4, 12, 64, 32)", style=solid];
-"1910 transpose_36" -> "1911 matmul_36"  [label="(4, 12, 32, 64)", style=solid];
-"1911 matmul_36" -> "1913 mul_37"  [label="(4, 12, 64, 64)", style=solid];
-"1912 _frozen_param156" -> "1913 mul_37"  [label="(12, 1, 1)", style=solid];
-"1913 mul_37" -> "1914 add_63"  [label="(4, 12, 64, 64)", style=solid];
-"1914 add_63" -> "1915 softmax_18"  [label="(4, 12, 64, 64)", style=solid];
-"1915 softmax_18" -> "1916 dropout_72"  [label="(4, 12, 64, 64)", style=solid];
-"1916 dropout_72" -> "1917 matmul_37"  [label="(4, 12, 64, 64)", style=solid];
-"1917 matmul_37" -> "1918 transpose_37"  [label="(4, 12, 64, 32)", style=solid];
-"1918 transpose_37" -> "1919 reshape_83"  [label="(4, 64, 12, 32)", style=solid];
-"1919 reshape_83" -> "1920 reshape_83_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1920 reshape_83_0_0_nncf_smooth_quant_0" -> "1921 quantize_per_tensor_default_114"  [label="(4, 64, 384)", style=solid];
-"1921 quantize_per_tensor_default_114" -> "1922 dequantize_per_tensor_default_114"  [label="(4, 64, 384)", style=solid];
-"1922 dequantize_per_tensor_default_114" -> "1929 linear_113"  [label="(4, 64, 384)", style=solid];
-"1923 linear_113_scale_0" -> "1926 quantize_per_channel_default_95"  [label="(384,)", style=solid];
-"1923 linear_113_scale_0" -> "1927 dequantize_per_channel_default_95"  [label="(384,)", style=solid];
-"1924 linear_113_zero_point_0" -> "1926 quantize_per_channel_default_95"  [label="(384,)", style=solid];
-"1924 linear_113_zero_point_0" -> "1927 dequantize_per_channel_default_95"  [label="(384,)", style=solid];
-"1925 compressed_weight_updated_constant76" -> "1926 quantize_per_channel_default_95"  [label="(384, 384)", style=solid];
-"1926 quantize_per_channel_default_95" -> "1927 dequantize_per_channel_default_95"  [label="(384, 384)", style=solid];
-"1927 dequantize_per_channel_default_95" -> "1929 linear_113"  [label="(384, 384)", style=solid];
-"1928 _param_constant305_0_0" -> "1929 linear_113"  [label="(384,)", style=solid];
-"1929 linear_113" -> "1930 dropout_73"  [label="(4, 64, 384)", style=solid];
-"1930 dropout_73" -> "1931 view_102"  [label="(4, 64, 384)", style=solid];
-"1931 view_102" -> "1932 permute_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1932 permute_85" -> "1933 reshape_84"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1933 reshape_84" -> "1934 slice_274"  [label="(1, 16, 16, 384)", style=solid];
-"1934 slice_274" -> "1935 slice_275"  [label="(1, 16, 16, 384)", style=solid];
-"1935 slice_275" -> "1936 slice_276"  [label="(1, 14, 16, 384)", style=solid];
-"1936 slice_276" -> "1937 slice_277"  [label="(1, 14, 14, 384)", style=solid];
-"1937 slice_277" -> "1938 contiguous_35"  [label="(1, 14, 14, 384)", style=solid];
-"1938 contiguous_35" -> "1941 layer_norm_39"  [label="(1, 14, 14, 384)", style=solid];
-"1939 _param_constant306" -> "1941 layer_norm_39"  [label="(384,)", style=solid];
-"1940 _param_constant307" -> "1941 layer_norm_39"  [label="(384,)", style=solid];
-"1941 layer_norm_39" -> "1942 add_64"  [label="(1, 14, 14, 384)", style=solid];
-"1942 add_64" -> "1943 add_64_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"1942 add_64" -> "1969 add_65"  [label="(1, 14, 14, 384)", style=solid];
-"1943 add_64_0_0_nncf_smooth_quant_0" -> "1944 quantize_per_tensor_default_115"  [label="(1, 14, 14, 384)", style=solid];
-"1944 quantize_per_tensor_default_115" -> "1945 dequantize_per_tensor_default_115"  [label="(1, 14, 14, 384)", style=solid];
-"1945 dequantize_per_tensor_default_115" -> "1952 linear_114"  [label="(1, 14, 14, 384)", style=solid];
-"1946 linear_114_scale_0" -> "1949 quantize_per_channel_default_96"  [label="(1536,)", style=solid];
-"1946 linear_114_scale_0" -> "1950 dequantize_per_channel_default_96"  [label="(1536,)", style=solid];
-"1947 linear_114_zero_point_0" -> "1949 quantize_per_channel_default_96"  [label="(1536,)", style=solid];
-"1947 linear_114_zero_point_0" -> "1950 dequantize_per_channel_default_96"  [label="(1536,)", style=solid];
-"1948 compressed_weight_updated_constant77" -> "1949 quantize_per_channel_default_96"  [label="(1536, 384)", style=solid];
-"1949 quantize_per_channel_default_96" -> "1950 dequantize_per_channel_default_96"  [label="(1536, 384)", style=solid];
-"1950 dequantize_per_channel_default_96" -> "1952 linear_114"  [label="(1536, 384)", style=solid];
-"1951 _param_constant309_0_0" -> "1952 linear_114"  [label="(1536,)", style=solid];
-"1952 linear_114" -> "1953 gelu_18"  [label="(1, 14, 14, 1536)", style=solid];
-"1953 gelu_18" -> "1954 dropout_74"  [label="(1, 14, 14, 1536)", style=solid];
-"1954 dropout_74" -> "1955 dropout_74_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"1955 dropout_74_0_0_nncf_smooth_quant_0" -> "1956 quantize_per_tensor_default_116"  [label="(1, 14, 14, 1536)", style=solid];
-"1956 quantize_per_tensor_default_116" -> "1957 dequantize_per_tensor_default_116"  [label="(1, 14, 14, 1536)", style=solid];
-"1957 dequantize_per_tensor_default_116" -> "1964 linear_115"  [label="(1, 14, 14, 1536)", style=solid];
-"1958 linear_115_scale_0" -> "1961 quantize_per_channel_default_97"  [label="(384,)", style=solid];
-"1958 linear_115_scale_0" -> "1962 dequantize_per_channel_default_97"  [label="(384,)", style=solid];
-"1959 linear_115_zero_point_0" -> "1961 quantize_per_channel_default_97"  [label="(384,)", style=solid];
-"1959 linear_115_zero_point_0" -> "1962 dequantize_per_channel_default_97"  [label="(384,)", style=solid];
-"1960 compressed_weight_updated_constant78" -> "1961 quantize_per_channel_default_97"  [label="(384, 1536)", style=solid];
-"1961 quantize_per_channel_default_97" -> "1962 dequantize_per_channel_default_97"  [label="(384, 1536)", style=solid];
-"1962 dequantize_per_channel_default_97" -> "1964 linear_115"  [label="(384, 1536)", style=solid];
-"1963 _param_constant311_0_0" -> "1964 linear_115"  [label="(384,)", style=solid];
-"1964 linear_115" -> "1965 dropout_75"  [label="(1, 14, 14, 384)", style=solid];
-"1965 dropout_75" -> "1968 layer_norm_40"  [label="(1, 14, 14, 384)", style=solid];
-"1966 _param_constant312" -> "1968 layer_norm_40"  [label="(384,)", style=solid];
-"1967 _param_constant313" -> "1968 layer_norm_40"  [label="(384,)", style=solid];
-"1968 layer_norm_40" -> "1969 add_65"  [label="(1, 14, 14, 384)", style=solid];
-"1969 add_65" -> "1971 pad_21"  [label="(1, 14, 14, 384)", style=solid];
-"1969 add_65" -> "2052 add_68"  [label="(1, 14, 14, 384)", style=solid];
-"1970 _frozen_param157" -> "2007 add_66"  [label="(1, 12, 64, 64)", style=solid];
-"1971 pad_21" -> "1972 roll_18"  [label="(1, 16, 16, 384)", style=solid];
-"1972 roll_18" -> "1973 view_105"  [label="(1, 16, 16, 384)", style=solid];
-"1973 view_105" -> "1974 permute_87"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1974 permute_87" -> "1975 reshape_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1975 reshape_85" -> "1976 reshape_85_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"1975 reshape_85" -> "2008 new_zeros_9"  [label="(4, 64, 384)", style=solid];
-"1976 reshape_85_0_0_nncf_smooth_quant_0" -> "1977 quantize_per_tensor_default_117"  [label="(4, 64, 384)", style=solid];
-"1977 quantize_per_tensor_default_117" -> "1978 dequantize_per_tensor_default_117"  [label="(4, 64, 384)", style=solid];
-"1978 dequantize_per_tensor_default_117" -> "1985 linear_118"  [label="(4, 64, 384)", style=solid];
-"1979 linear_118_scale_0" -> "1982 quantize_per_channel_default_99"  [label="(1152,)", style=solid];
-"1979 linear_118_scale_0" -> "1983 dequantize_per_channel_default_99"  [label="(1152,)", style=solid];
-"1980 linear_118_zero_point_0" -> "1982 quantize_per_channel_default_99"  [label="(1152,)", style=solid];
-"1980 linear_118_zero_point_0" -> "1983 dequantize_per_channel_default_99"  [label="(1152,)", style=solid];
-"1981 compressed_weight_updated_constant79" -> "1982 quantize_per_channel_default_99"  [label="(1152, 384)", style=solid];
-"1982 quantize_per_channel_default_99" -> "1983 dequantize_per_channel_default_99"  [label="(1152, 384)", style=solid];
-"1983 dequantize_per_channel_default_99" -> "1985 linear_118"  [label="(1152, 384)", style=solid];
-"1984 _frozen_param158_0_0" -> "1985 linear_118"  [label="(1152,)", style=solid];
-"1985 linear_118" -> "1986 reshape_86"  [label="(4, 64, 1152)", style=solid];
-"1986 reshape_86" -> "1987 permute_88"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1987 permute_88" -> "1988 select_57"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1987 permute_88" -> "1989 select_58"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1987 permute_88" -> "1990 select_59"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1988 select_57" -> "1991 linalg_vector_norm_38"  [label="(4, 12, 64, 32)", style=solid];
-"1988 select_57" -> "1993 expand_as_38"  [label="(4, 12, 64, 32)", style=solid];
-"1988 select_57" -> "1994 div_38"  [label="(4, 12, 64, 32)", style=solid];
-"1989 select_58" -> "1997 linalg_vector_norm_39"  [label="(4, 12, 64, 32)", style=solid];
-"1989 select_58" -> "1999 expand_as_39"  [label="(4, 12, 64, 32)", style=solid];
-"1989 select_58" -> "2000 div_39"  [label="(4, 12, 64, 32)", style=solid];
-"1990 select_59" -> "2026 matmul_39"  [label="(4, 12, 64, 32)", style=solid];
-"1991 linalg_vector_norm_38" -> "1992 clamp_min_38"  [label="(4, 12, 64, 1)", style=solid];
-"1992 clamp_min_38" -> "1993 expand_as_38"  [label="(4, 12, 64, 1)", style=solid];
-"1993 expand_as_38" -> "1994 div_38"  [label="(4, 12, 64, 32)", style=solid];
-"1994 div_38" -> "1995 quantize_per_tensor_default_118"  [label="(4, 12, 64, 32)", style=solid];
-"1995 quantize_per_tensor_default_118" -> "1996 dequantize_per_tensor_default_118"  [label="(4, 12, 64, 32)", style=solid];
-"1996 dequantize_per_tensor_default_118" -> "2004 matmul_38"  [label="(4, 12, 64, 32)", style=solid];
-"1997 linalg_vector_norm_39" -> "1998 clamp_min_39"  [label="(4, 12, 64, 1)", style=solid];
-"1998 clamp_min_39" -> "1999 expand_as_39"  [label="(4, 12, 64, 1)", style=solid];
-"1999 expand_as_39" -> "2000 div_39"  [label="(4, 12, 64, 32)", style=solid];
-"2000 div_39" -> "2001 quantize_per_tensor_default_119"  [label="(4, 12, 64, 32)", style=solid];
-"2001 quantize_per_tensor_default_119" -> "2002 dequantize_per_tensor_default_119"  [label="(4, 12, 64, 32)", style=solid];
-"2002 dequantize_per_tensor_default_119" -> "2003 transpose_38"  [label="(4, 12, 64, 32)", style=solid];
-"2003 transpose_38" -> "2004 matmul_38"  [label="(4, 12, 32, 64)", style=solid];
-"2004 matmul_38" -> "2006 mul_39"  [label="(4, 12, 64, 64)", style=solid];
-"2005 _frozen_param160" -> "2006 mul_39"  [label="(12, 1, 1)", style=solid];
-"2006 mul_39" -> "2007 add_66"  [label="(4, 12, 64, 64)", style=solid];
-"2007 add_66" -> "2019 view_107"  [label="(4, 12, 64, 64)", style=solid];
-"2008 new_zeros_9" -> "2009 view_106"  [label="(16, 16)", style=solid];
-"2009 view_106" -> "2010 permute_89"  [label="(2, 8, 2, 8)", style=solid];
-"2010 permute_89" -> "2011 reshape_87"  [label="(2, 2, 8, 8)", style=solid];
-"2011 reshape_87" -> "2012 unsqueeze_56"  [label="(4, 64)", style=solid];
-"2011 reshape_87" -> "2013 unsqueeze_57"  [label="(4, 64)", style=solid];
-"2012 unsqueeze_56" -> "2014 sub_9"  [label="(4, 1, 64)", style=solid];
-"2013 unsqueeze_57" -> "2014 sub_9"  [label="(4, 64, 1)", style=solid];
-"2014 sub_9" -> "2015 ne_9"  [label="(4, 64, 64)", style=solid];
-"2014 sub_9" -> "2016 masked_fill_18"  [label="(4, 64, 64)", style=solid];
-"2014 sub_9" -> "2017 eq_9"  [label="(4, 64, 64)", style=solid];
-"2015 ne_9" -> "2016 masked_fill_18"  [label="(4, 64, 64)", style=solid];
-"2016 masked_fill_18" -> "2018 masked_fill_19"  [label="(4, 64, 64)", style=solid];
-"2017 eq_9" -> "2018 masked_fill_19"  [label="(4, 64, 64)", style=solid];
-"2018 masked_fill_19" -> "2020 unsqueeze_58"  [label="(4, 64, 64)", style=solid];
-"2019 view_107" -> "2022 add_67"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2020 unsqueeze_58" -> "2021 unsqueeze_59"  [label="(4, 1, 64, 64)", style=solid];
-"2021 unsqueeze_59" -> "2022 add_67"  [label="(1, 4, 1, 64, 64)", style=solid];
-"2022 add_67" -> "2023 view_108"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2023 view_108" -> "2024 softmax_19"  [label="(4, 12, 64, 64)", style=solid];
-"2024 softmax_19" -> "2025 dropout_76"  [label="(4, 12, 64, 64)", style=solid];
-"2025 dropout_76" -> "2026 matmul_39"  [label="(4, 12, 64, 64)", style=solid];
-"2026 matmul_39" -> "2027 transpose_39"  [label="(4, 12, 64, 32)", style=solid];
-"2027 transpose_39" -> "2028 reshape_88"  [label="(4, 64, 12, 32)", style=solid];
-"2028 reshape_88" -> "2029 reshape_88_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"2029 reshape_88_0_0_nncf_smooth_quant_0" -> "2030 quantize_per_tensor_default_120"  [label="(4, 64, 384)", style=solid];
-"2030 quantize_per_tensor_default_120" -> "2031 dequantize_per_tensor_default_120"  [label="(4, 64, 384)", style=solid];
-"2031 dequantize_per_tensor_default_120" -> "2038 linear_119"  [label="(4, 64, 384)", style=solid];
-"2032 linear_119_scale_0" -> "2035 quantize_per_channel_default_100"  [label="(384,)", style=solid];
-"2032 linear_119_scale_0" -> "2036 dequantize_per_channel_default_100"  [label="(384,)", style=solid];
-"2033 linear_119_zero_point_0" -> "2035 quantize_per_channel_default_100"  [label="(384,)", style=solid];
-"2033 linear_119_zero_point_0" -> "2036 dequantize_per_channel_default_100"  [label="(384,)", style=solid];
-"2034 compressed_weight_updated_constant80" -> "2035 quantize_per_channel_default_100"  [label="(384, 384)", style=solid];
-"2035 quantize_per_channel_default_100" -> "2036 dequantize_per_channel_default_100"  [label="(384, 384)", style=solid];
-"2036 dequantize_per_channel_default_100" -> "2038 linear_119"  [label="(384, 384)", style=solid];
-"2037 _param_constant321_0_0" -> "2038 linear_119"  [label="(384,)", style=solid];
-"2038 linear_119" -> "2039 dropout_77"  [label="(4, 64, 384)", style=solid];
-"2039 dropout_77" -> "2040 view_109"  [label="(4, 64, 384)", style=solid];
-"2040 view_109" -> "2041 permute_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2041 permute_90" -> "2042 reshape_89"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2042 reshape_89" -> "2043 roll_19"  [label="(1, 16, 16, 384)", style=solid];
-"2043 roll_19" -> "2044 slice_297"  [label="(1, 16, 16, 384)", style=solid];
-"2044 slice_297" -> "2045 slice_298"  [label="(1, 16, 16, 384)", style=solid];
-"2045 slice_298" -> "2046 slice_299"  [label="(1, 14, 16, 384)", style=solid];
-"2046 slice_299" -> "2047 slice_300"  [label="(1, 14, 14, 384)", style=solid];
-"2047 slice_300" -> "2048 contiguous_37"  [label="(1, 14, 14, 384)", style=solid];
-"2048 contiguous_37" -> "2051 layer_norm_41"  [label="(1, 14, 14, 384)", style=solid];
-"2049 _param_constant322" -> "2051 layer_norm_41"  [label="(384,)", style=solid];
-"2050 _param_constant323" -> "2051 layer_norm_41"  [label="(384,)", style=solid];
-"2051 layer_norm_41" -> "2052 add_68"  [label="(1, 14, 14, 384)", style=solid];
-"2052 add_68" -> "2053 add_68_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"2052 add_68" -> "2079 add_69"  [label="(1, 14, 14, 384)", style=solid];
-"2053 add_68_0_0_nncf_smooth_quant_0" -> "2054 quantize_per_tensor_default_121"  [label="(1, 14, 14, 384)", style=solid];
-"2054 quantize_per_tensor_default_121" -> "2055 dequantize_per_tensor_default_121"  [label="(1, 14, 14, 384)", style=solid];
-"2055 dequantize_per_tensor_default_121" -> "2062 linear_120"  [label="(1, 14, 14, 384)", style=solid];
-"2056 linear_120_scale_0" -> "2059 quantize_per_channel_default_101"  [label="(1536,)", style=solid];
-"2056 linear_120_scale_0" -> "2060 dequantize_per_channel_default_101"  [label="(1536,)", style=solid];
-"2057 linear_120_zero_point_0" -> "2059 quantize_per_channel_default_101"  [label="(1536,)", style=solid];
-"2057 linear_120_zero_point_0" -> "2060 dequantize_per_channel_default_101"  [label="(1536,)", style=solid];
-"2058 compressed_weight_updated_constant81" -> "2059 quantize_per_channel_default_101"  [label="(1536, 384)", style=solid];
-"2059 quantize_per_channel_default_101" -> "2060 dequantize_per_channel_default_101"  [label="(1536, 384)", style=solid];
-"2060 dequantize_per_channel_default_101" -> "2062 linear_120"  [label="(1536, 384)", style=solid];
-"2061 _param_constant325_0_0" -> "2062 linear_120"  [label="(1536,)", style=solid];
-"2062 linear_120" -> "2063 gelu_19"  [label="(1, 14, 14, 1536)", style=solid];
-"2063 gelu_19" -> "2064 dropout_78"  [label="(1, 14, 14, 1536)", style=solid];
-"2064 dropout_78" -> "2065 dropout_78_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"2065 dropout_78_0_0_nncf_smooth_quant_0" -> "2066 quantize_per_tensor_default_122"  [label="(1, 14, 14, 1536)", style=solid];
-"2066 quantize_per_tensor_default_122" -> "2067 dequantize_per_tensor_default_122"  [label="(1, 14, 14, 1536)", style=solid];
-"2067 dequantize_per_tensor_default_122" -> "2074 linear_121"  [label="(1, 14, 14, 1536)", style=solid];
-"2068 linear_121_scale_0" -> "2071 quantize_per_channel_default_102"  [label="(384,)", style=solid];
-"2068 linear_121_scale_0" -> "2072 dequantize_per_channel_default_102"  [label="(384,)", style=solid];
-"2069 linear_121_zero_point_0" -> "2071 quantize_per_channel_default_102"  [label="(384,)", style=solid];
-"2069 linear_121_zero_point_0" -> "2072 dequantize_per_channel_default_102"  [label="(384,)", style=solid];
-"2070 compressed_weight_updated_constant82" -> "2071 quantize_per_channel_default_102"  [label="(384, 1536)", style=solid];
-"2071 quantize_per_channel_default_102" -> "2072 dequantize_per_channel_default_102"  [label="(384, 1536)", style=solid];
-"2072 dequantize_per_channel_default_102" -> "2074 linear_121"  [label="(384, 1536)", style=solid];
-"2073 _param_constant327_0_0" -> "2074 linear_121"  [label="(384,)", style=solid];
-"2074 linear_121" -> "2075 dropout_79"  [label="(1, 14, 14, 384)", style=solid];
-"2075 dropout_79" -> "2078 layer_norm_42"  [label="(1, 14, 14, 384)", style=solid];
-"2076 _param_constant328" -> "2078 layer_norm_42"  [label="(384,)", style=solid];
-"2077 _param_constant329" -> "2078 layer_norm_42"  [label="(384,)", style=solid];
-"2078 layer_norm_42" -> "2079 add_69"  [label="(1, 14, 14, 384)", style=solid];
-"2079 add_69" -> "2081 pad_22"  [label="(1, 14, 14, 384)", style=solid];
-"2079 add_69" -> "2144 add_71"  [label="(1, 14, 14, 384)", style=solid];
-"2080 _frozen_param170" -> "2116 add_70"  [label="(1, 12, 64, 64)", style=solid];
-"2081 pad_22" -> "2082 view_112"  [label="(1, 16, 16, 384)", style=solid];
-"2082 view_112" -> "2083 permute_92"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2083 permute_92" -> "2084 reshape_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2084 reshape_90" -> "2085 reshape_90_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"2085 reshape_90_0_0_nncf_smooth_quant_0" -> "2086 quantize_per_tensor_default_123"  [label="(4, 64, 384)", style=solid];
-"2086 quantize_per_tensor_default_123" -> "2087 dequantize_per_tensor_default_123"  [label="(4, 64, 384)", style=solid];
-"2087 dequantize_per_tensor_default_123" -> "2094 linear_124"  [label="(4, 64, 384)", style=solid];
-"2088 linear_124_scale_0" -> "2091 quantize_per_channel_default_104"  [label="(1152,)", style=solid];
-"2088 linear_124_scale_0" -> "2092 dequantize_per_channel_default_104"  [label="(1152,)", style=solid];
-"2089 linear_124_zero_point_0" -> "2091 quantize_per_channel_default_104"  [label="(1152,)", style=solid];
-"2089 linear_124_zero_point_0" -> "2092 dequantize_per_channel_default_104"  [label="(1152,)", style=solid];
-"2090 compressed_weight_updated_constant83" -> "2091 quantize_per_channel_default_104"  [label="(1152, 384)", style=solid];
-"2091 quantize_per_channel_default_104" -> "2092 dequantize_per_channel_default_104"  [label="(1152, 384)", style=solid];
-"2092 dequantize_per_channel_default_104" -> "2094 linear_124"  [label="(1152, 384)", style=solid];
-"2093 _frozen_param171_0_0" -> "2094 linear_124"  [label="(1152,)", style=solid];
-"2094 linear_124" -> "2095 reshape_91"  [label="(4, 64, 1152)", style=solid];
-"2095 reshape_91" -> "2096 permute_93"  [label="(4, 64, 3, 12, 32)", style=solid];
-"2096 permute_93" -> "2097 select_60"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2096 permute_93" -> "2098 select_61"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2096 permute_93" -> "2099 select_62"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2097 select_60" -> "2100 linalg_vector_norm_40"  [label="(4, 12, 64, 32)", style=solid];
-"2097 select_60" -> "2102 expand_as_40"  [label="(4, 12, 64, 32)", style=solid];
-"2097 select_60" -> "2103 div_40"  [label="(4, 12, 64, 32)", style=solid];
-"2098 select_61" -> "2106 linalg_vector_norm_41"  [label="(4, 12, 64, 32)", style=solid];
-"2098 select_61" -> "2108 expand_as_41"  [label="(4, 12, 64, 32)", style=solid];
-"2098 select_61" -> "2109 div_41"  [label="(4, 12, 64, 32)", style=solid];
-"2099 select_62" -> "2119 matmul_41"  [label="(4, 12, 64, 32)", style=solid];
-"2100 linalg_vector_norm_40" -> "2101 clamp_min_40"  [label="(4, 12, 64, 1)", style=solid];
-"2101 clamp_min_40" -> "2102 expand_as_40"  [label="(4, 12, 64, 1)", style=solid];
-"2102 expand_as_40" -> "2103 div_40"  [label="(4, 12, 64, 32)", style=solid];
-"2103 div_40" -> "2104 quantize_per_tensor_default_124"  [label="(4, 12, 64, 32)", style=solid];
-"2104 quantize_per_tensor_default_124" -> "2105 dequantize_per_tensor_default_124"  [label="(4, 12, 64, 32)", style=solid];
-"2105 dequantize_per_tensor_default_124" -> "2113 matmul_40"  [label="(4, 12, 64, 32)", style=solid];
-"2106 linalg_vector_norm_41" -> "2107 clamp_min_41"  [label="(4, 12, 64, 1)", style=solid];
-"2107 clamp_min_41" -> "2108 expand_as_41"  [label="(4, 12, 64, 1)", style=solid];
-"2108 expand_as_41" -> "2109 div_41"  [label="(4, 12, 64, 32)", style=solid];
-"2109 div_41" -> "2110 quantize_per_tensor_default_125"  [label="(4, 12, 64, 32)", style=solid];
-"2110 quantize_per_tensor_default_125" -> "2111 dequantize_per_tensor_default_125"  [label="(4, 12, 64, 32)", style=solid];
-"2111 dequantize_per_tensor_default_125" -> "2112 transpose_40"  [label="(4, 12, 64, 32)", style=solid];
-"2112 transpose_40" -> "2113 matmul_40"  [label="(4, 12, 32, 64)", style=solid];
-"2113 matmul_40" -> "2115 mul_41"  [label="(4, 12, 64, 64)", style=solid];
-"2114 _frozen_param173" -> "2115 mul_41"  [label="(12, 1, 1)", style=solid];
-"2115 mul_41" -> "2116 add_70"  [label="(4, 12, 64, 64)", style=solid];
-"2116 add_70" -> "2117 softmax_20"  [label="(4, 12, 64, 64)", style=solid];
-"2117 softmax_20" -> "2118 dropout_80"  [label="(4, 12, 64, 64)", style=solid];
-"2118 dropout_80" -> "2119 matmul_41"  [label="(4, 12, 64, 64)", style=solid];
-"2119 matmul_41" -> "2120 transpose_41"  [label="(4, 12, 64, 32)", style=solid];
-"2120 transpose_41" -> "2121 reshape_92"  [label="(4, 64, 12, 32)", style=solid];
-"2121 reshape_92" -> "2122 reshape_92_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"2122 reshape_92_0_0_nncf_smooth_quant_0" -> "2123 quantize_per_tensor_default_126"  [label="(4, 64, 384)", style=solid];
-"2123 quantize_per_tensor_default_126" -> "2124 dequantize_per_tensor_default_126"  [label="(4, 64, 384)", style=solid];
-"2124 dequantize_per_tensor_default_126" -> "2131 linear_125"  [label="(4, 64, 384)", style=solid];
-"2125 linear_125_scale_0" -> "2128 quantize_per_channel_default_105"  [label="(384,)", style=solid];
-"2125 linear_125_scale_0" -> "2129 dequantize_per_channel_default_105"  [label="(384,)", style=solid];
-"2126 linear_125_zero_point_0" -> "2128 quantize_per_channel_default_105"  [label="(384,)", style=solid];
-"2126 linear_125_zero_point_0" -> "2129 dequantize_per_channel_default_105"  [label="(384,)", style=solid];
-"2127 compressed_weight_updated_constant84" -> "2128 quantize_per_channel_default_105"  [label="(384, 384)", style=solid];
-"2128 quantize_per_channel_default_105" -> "2129 dequantize_per_channel_default_105"  [label="(384, 384)", style=solid];
-"2129 dequantize_per_channel_default_105" -> "2131 linear_125"  [label="(384, 384)", style=solid];
-"2130 _param_constant337_0_0" -> "2131 linear_125"  [label="(384,)", style=solid];
-"2131 linear_125" -> "2132 dropout_81"  [label="(4, 64, 384)", style=solid];
-"2132 dropout_81" -> "2133 view_113"  [label="(4, 64, 384)", style=solid];
-"2133 view_113" -> "2134 permute_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2134 permute_94" -> "2135 reshape_93"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2135 reshape_93" -> "2136 slice_302"  [label="(1, 16, 16, 384)", style=solid];
-"2136 slice_302" -> "2137 slice_303"  [label="(1, 16, 16, 384)", style=solid];
-"2137 slice_303" -> "2138 slice_304"  [label="(1, 14, 16, 384)", style=solid];
-"2138 slice_304" -> "2139 slice_305"  [label="(1, 14, 14, 384)", style=solid];
-"2139 slice_305" -> "2140 contiguous_39"  [label="(1, 14, 14, 384)", style=solid];
-"2140 contiguous_39" -> "2143 layer_norm_43"  [label="(1, 14, 14, 384)", style=solid];
-"2141 _param_constant338" -> "2143 layer_norm_43"  [label="(384,)", style=solid];
-"2142 _param_constant339" -> "2143 layer_norm_43"  [label="(384,)", style=solid];
-"2143 layer_norm_43" -> "2144 add_71"  [label="(1, 14, 14, 384)", style=solid];
-"2144 add_71" -> "2145 add_71_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"2144 add_71" -> "2171 add_72"  [label="(1, 14, 14, 384)", style=solid];
-"2145 add_71_0_0_nncf_smooth_quant_0" -> "2146 quantize_per_tensor_default_127"  [label="(1, 14, 14, 384)", style=solid];
-"2146 quantize_per_tensor_default_127" -> "2147 dequantize_per_tensor_default_127"  [label="(1, 14, 14, 384)", style=solid];
-"2147 dequantize_per_tensor_default_127" -> "2154 linear_126"  [label="(1, 14, 14, 384)", style=solid];
-"2148 linear_126_scale_0" -> "2151 quantize_per_channel_default_106"  [label="(1536,)", style=solid];
-"2148 linear_126_scale_0" -> "2152 dequantize_per_channel_default_106"  [label="(1536,)", style=solid];
-"2149 linear_126_zero_point_0" -> "2151 quantize_per_channel_default_106"  [label="(1536,)", style=solid];
-"2149 linear_126_zero_point_0" -> "2152 dequantize_per_channel_default_106"  [label="(1536,)", style=solid];
-"2150 compressed_weight_updated_constant85" -> "2151 quantize_per_channel_default_106"  [label="(1536, 384)", style=solid];
-"2151 quantize_per_channel_default_106" -> "2152 dequantize_per_channel_default_106"  [label="(1536, 384)", style=solid];
-"2152 dequantize_per_channel_default_106" -> "2154 linear_126"  [label="(1536, 384)", style=solid];
-"2153 _param_constant341_0_0" -> "2154 linear_126"  [label="(1536,)", style=solid];
-"2154 linear_126" -> "2155 gelu_20"  [label="(1, 14, 14, 1536)", style=solid];
-"2155 gelu_20" -> "2156 dropout_82"  [label="(1, 14, 14, 1536)", style=solid];
-"2156 dropout_82" -> "2157 dropout_82_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"2157 dropout_82_0_0_nncf_smooth_quant_0" -> "2158 quantize_per_tensor_default_128"  [label="(1, 14, 14, 1536)", style=solid];
-"2158 quantize_per_tensor_default_128" -> "2159 dequantize_per_tensor_default_128"  [label="(1, 14, 14, 1536)", style=solid];
-"2159 dequantize_per_tensor_default_128" -> "2166 linear_127"  [label="(1, 14, 14, 1536)", style=solid];
-"2160 linear_127_scale_0" -> "2163 quantize_per_channel_default_107"  [label="(384,)", style=solid];
-"2160 linear_127_scale_0" -> "2164 dequantize_per_channel_default_107"  [label="(384,)", style=solid];
-"2161 linear_127_zero_point_0" -> "2163 quantize_per_channel_default_107"  [label="(384,)", style=solid];
-"2161 linear_127_zero_point_0" -> "2164 dequantize_per_channel_default_107"  [label="(384,)", style=solid];
-"2162 compressed_weight_updated_constant86" -> "2163 quantize_per_channel_default_107"  [label="(384, 1536)", style=solid];
-"2163 quantize_per_channel_default_107" -> "2164 dequantize_per_channel_default_107"  [label="(384, 1536)", style=solid];
-"2164 dequantize_per_channel_default_107" -> "2166 linear_127"  [label="(384, 1536)", style=solid];
-"2165 _param_constant343_0_0" -> "2166 linear_127"  [label="(384,)", style=solid];
-"2166 linear_127" -> "2167 dropout_83"  [label="(1, 14, 14, 384)", style=solid];
-"2167 dropout_83" -> "2170 layer_norm_44"  [label="(1, 14, 14, 384)", style=solid];
-"2168 _param_constant344" -> "2170 layer_norm_44"  [label="(384,)", style=solid];
-"2169 _param_constant345" -> "2170 layer_norm_44"  [label="(384,)", style=solid];
-"2170 layer_norm_44" -> "2171 add_72"  [label="(1, 14, 14, 384)", style=solid];
-"2171 add_72" -> "2173 pad_23"  [label="(1, 14, 14, 384)", style=solid];
-"2171 add_72" -> "2254 add_75"  [label="(1, 14, 14, 384)", style=solid];
-"2172 _frozen_param174" -> "2209 add_73"  [label="(1, 12, 64, 64)", style=solid];
-"2173 pad_23" -> "2174 roll_20"  [label="(1, 16, 16, 384)", style=solid];
-"2174 roll_20" -> "2175 view_116"  [label="(1, 16, 16, 384)", style=solid];
-"2175 view_116" -> "2176 permute_96"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2176 permute_96" -> "2177 reshape_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2177 reshape_94" -> "2178 reshape_94_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"2177 reshape_94" -> "2210 new_zeros_10"  [label="(4, 64, 384)", style=solid];
-"2178 reshape_94_0_0_nncf_smooth_quant_0" -> "2179 quantize_per_tensor_default_129"  [label="(4, 64, 384)", style=solid];
-"2179 quantize_per_tensor_default_129" -> "2180 dequantize_per_tensor_default_129"  [label="(4, 64, 384)", style=solid];
-"2180 dequantize_per_tensor_default_129" -> "2187 linear_130"  [label="(4, 64, 384)", style=solid];
-"2181 linear_130_scale_0" -> "2184 quantize_per_channel_default_109"  [label="(1152,)", style=solid];
-"2181 linear_130_scale_0" -> "2185 dequantize_per_channel_default_109"  [label="(1152,)", style=solid];
-"2182 linear_130_zero_point_0" -> "2184 quantize_per_channel_default_109"  [label="(1152,)", style=solid];
-"2182 linear_130_zero_point_0" -> "2185 dequantize_per_channel_default_109"  [label="(1152,)", style=solid];
-"2183 compressed_weight_updated_constant87" -> "2184 quantize_per_channel_default_109"  [label="(1152, 384)", style=solid];
-"2184 quantize_per_channel_default_109" -> "2185 dequantize_per_channel_default_109"  [label="(1152, 384)", style=solid];
-"2185 dequantize_per_channel_default_109" -> "2187 linear_130"  [label="(1152, 384)", style=solid];
-"2186 _frozen_param175_0_0" -> "2187 linear_130"  [label="(1152,)", style=solid];
-"2187 linear_130" -> "2188 reshape_95"  [label="(4, 64, 1152)", style=solid];
-"2188 reshape_95" -> "2189 permute_97"  [label="(4, 64, 3, 12, 32)", style=solid];
-"2189 permute_97" -> "2190 select_63"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2189 permute_97" -> "2191 select_64"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2189 permute_97" -> "2192 select_65"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2190 select_63" -> "2193 linalg_vector_norm_42"  [label="(4, 12, 64, 32)", style=solid];
-"2190 select_63" -> "2195 expand_as_42"  [label="(4, 12, 64, 32)", style=solid];
-"2190 select_63" -> "2196 div_42"  [label="(4, 12, 64, 32)", style=solid];
-"2191 select_64" -> "2199 linalg_vector_norm_43"  [label="(4, 12, 64, 32)", style=solid];
-"2191 select_64" -> "2201 expand_as_43"  [label="(4, 12, 64, 32)", style=solid];
-"2191 select_64" -> "2202 div_43"  [label="(4, 12, 64, 32)", style=solid];
-"2192 select_65" -> "2228 matmul_43"  [label="(4, 12, 64, 32)", style=solid];
-"2193 linalg_vector_norm_42" -> "2194 clamp_min_42"  [label="(4, 12, 64, 1)", style=solid];
-"2194 clamp_min_42" -> "2195 expand_as_42"  [label="(4, 12, 64, 1)", style=solid];
-"2195 expand_as_42" -> "2196 div_42"  [label="(4, 12, 64, 32)", style=solid];
-"2196 div_42" -> "2197 quantize_per_tensor_default_130"  [label="(4, 12, 64, 32)", style=solid];
-"2197 quantize_per_tensor_default_130" -> "2198 dequantize_per_tensor_default_130"  [label="(4, 12, 64, 32)", style=solid];
-"2198 dequantize_per_tensor_default_130" -> "2206 matmul_42"  [label="(4, 12, 64, 32)", style=solid];
-"2199 linalg_vector_norm_43" -> "2200 clamp_min_43"  [label="(4, 12, 64, 1)", style=solid];
-"2200 clamp_min_43" -> "2201 expand_as_43"  [label="(4, 12, 64, 1)", style=solid];
-"2201 expand_as_43" -> "2202 div_43"  [label="(4, 12, 64, 32)", style=solid];
-"2202 div_43" -> "2203 quantize_per_tensor_default_131"  [label="(4, 12, 64, 32)", style=solid];
-"2203 quantize_per_tensor_default_131" -> "2204 dequantize_per_tensor_default_131"  [label="(4, 12, 64, 32)", style=solid];
-"2204 dequantize_per_tensor_default_131" -> "2205 transpose_42"  [label="(4, 12, 64, 32)", style=solid];
-"2205 transpose_42" -> "2206 matmul_42"  [label="(4, 12, 32, 64)", style=solid];
-"2206 matmul_42" -> "2208 mul_43"  [label="(4, 12, 64, 64)", style=solid];
-"2207 _frozen_param177" -> "2208 mul_43"  [label="(12, 1, 1)", style=solid];
-"2208 mul_43" -> "2209 add_73"  [label="(4, 12, 64, 64)", style=solid];
-"2209 add_73" -> "2221 view_118"  [label="(4, 12, 64, 64)", style=solid];
-"2210 new_zeros_10" -> "2211 view_117"  [label="(16, 16)", style=solid];
-"2211 view_117" -> "2212 permute_98"  [label="(2, 8, 2, 8)", style=solid];
-"2212 permute_98" -> "2213 reshape_96"  [label="(2, 2, 8, 8)", style=solid];
-"2213 reshape_96" -> "2214 unsqueeze_62"  [label="(4, 64)", style=solid];
-"2213 reshape_96" -> "2215 unsqueeze_63"  [label="(4, 64)", style=solid];
-"2214 unsqueeze_62" -> "2216 sub_10"  [label="(4, 1, 64)", style=solid];
-"2215 unsqueeze_63" -> "2216 sub_10"  [label="(4, 64, 1)", style=solid];
-"2216 sub_10" -> "2217 ne_10"  [label="(4, 64, 64)", style=solid];
-"2216 sub_10" -> "2218 masked_fill_20"  [label="(4, 64, 64)", style=solid];
-"2216 sub_10" -> "2219 eq_10"  [label="(4, 64, 64)", style=solid];
-"2217 ne_10" -> "2218 masked_fill_20"  [label="(4, 64, 64)", style=solid];
-"2218 masked_fill_20" -> "2220 masked_fill_21"  [label="(4, 64, 64)", style=solid];
-"2219 eq_10" -> "2220 masked_fill_21"  [label="(4, 64, 64)", style=solid];
-"2220 masked_fill_21" -> "2222 unsqueeze_64"  [label="(4, 64, 64)", style=solid];
-"2221 view_118" -> "2224 add_74"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2222 unsqueeze_64" -> "2223 unsqueeze_65"  [label="(4, 1, 64, 64)", style=solid];
-"2223 unsqueeze_65" -> "2224 add_74"  [label="(1, 4, 1, 64, 64)", style=solid];
-"2224 add_74" -> "2225 view_119"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2225 view_119" -> "2226 softmax_21"  [label="(4, 12, 64, 64)", style=solid];
-"2226 softmax_21" -> "2227 dropout_84"  [label="(4, 12, 64, 64)", style=solid];
-"2227 dropout_84" -> "2228 matmul_43"  [label="(4, 12, 64, 64)", style=solid];
-"2228 matmul_43" -> "2229 transpose_43"  [label="(4, 12, 64, 32)", style=solid];
-"2229 transpose_43" -> "2230 reshape_97"  [label="(4, 64, 12, 32)", style=solid];
-"2230 reshape_97" -> "2231 reshape_97_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
-"2231 reshape_97_0_0_nncf_smooth_quant_0" -> "2232 quantize_per_tensor_default_132"  [label="(4, 64, 384)", style=solid];
-"2232 quantize_per_tensor_default_132" -> "2233 dequantize_per_tensor_default_132"  [label="(4, 64, 384)", style=solid];
-"2233 dequantize_per_tensor_default_132" -> "2240 linear_131"  [label="(4, 64, 384)", style=solid];
-"2234 linear_131_scale_0" -> "2237 quantize_per_channel_default_110"  [label="(384,)", style=solid];
-"2234 linear_131_scale_0" -> "2238 dequantize_per_channel_default_110"  [label="(384,)", style=solid];
-"2235 linear_131_zero_point_0" -> "2237 quantize_per_channel_default_110"  [label="(384,)", style=solid];
-"2235 linear_131_zero_point_0" -> "2238 dequantize_per_channel_default_110"  [label="(384,)", style=solid];
-"2236 compressed_weight_updated_constant88" -> "2237 quantize_per_channel_default_110"  [label="(384, 384)", style=solid];
-"2237 quantize_per_channel_default_110" -> "2238 dequantize_per_channel_default_110"  [label="(384, 384)", style=solid];
-"2238 dequantize_per_channel_default_110" -> "2240 linear_131"  [label="(384, 384)", style=solid];
-"2239 _param_constant353_0_0" -> "2240 linear_131"  [label="(384,)", style=solid];
-"2240 linear_131" -> "2241 dropout_85"  [label="(4, 64, 384)", style=solid];
-"2241 dropout_85" -> "2242 view_120"  [label="(4, 64, 384)", style=solid];
-"2242 view_120" -> "2243 permute_99"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2243 permute_99" -> "2244 reshape_98"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2244 reshape_98" -> "2245 roll_21"  [label="(1, 16, 16, 384)", style=solid];
-"2245 roll_21" -> "2246 slice_325"  [label="(1, 16, 16, 384)", style=solid];
-"2246 slice_325" -> "2247 slice_326"  [label="(1, 16, 16, 384)", style=solid];
-"2247 slice_326" -> "2248 slice_327"  [label="(1, 14, 16, 384)", style=solid];
-"2248 slice_327" -> "2249 slice_328"  [label="(1, 14, 14, 384)", style=solid];
-"2249 slice_328" -> "2250 contiguous_41"  [label="(1, 14, 14, 384)", style=solid];
-"2250 contiguous_41" -> "2253 layer_norm_45"  [label="(1, 14, 14, 384)", style=solid];
-"2251 _param_constant354" -> "2253 layer_norm_45"  [label="(384,)", style=solid];
-"2252 _param_constant355" -> "2253 layer_norm_45"  [label="(384,)", style=solid];
-"2253 layer_norm_45" -> "2254 add_75"  [label="(1, 14, 14, 384)", style=solid];
-"2254 add_75" -> "2255 add_75_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
-"2254 add_75" -> "2281 add_76"  [label="(1, 14, 14, 384)", style=solid];
-"2255 add_75_0_0_nncf_smooth_quant_0" -> "2256 quantize_per_tensor_default_133"  [label="(1, 14, 14, 384)", style=solid];
-"2256 quantize_per_tensor_default_133" -> "2257 dequantize_per_tensor_default_133"  [label="(1, 14, 14, 384)", style=solid];
-"2257 dequantize_per_tensor_default_133" -> "2264 linear_132"  [label="(1, 14, 14, 384)", style=solid];
-"2258 linear_132_scale_0" -> "2261 quantize_per_channel_default_111"  [label="(1536,)", style=solid];
-"2258 linear_132_scale_0" -> "2262 dequantize_per_channel_default_111"  [label="(1536,)", style=solid];
-"2259 linear_132_zero_point_0" -> "2261 quantize_per_channel_default_111"  [label="(1536,)", style=solid];
-"2259 linear_132_zero_point_0" -> "2262 dequantize_per_channel_default_111"  [label="(1536,)", style=solid];
-"2260 compressed_weight_updated_constant89" -> "2261 quantize_per_channel_default_111"  [label="(1536, 384)", style=solid];
-"2261 quantize_per_channel_default_111" -> "2262 dequantize_per_channel_default_111"  [label="(1536, 384)", style=solid];
-"2262 dequantize_per_channel_default_111" -> "2264 linear_132"  [label="(1536, 384)", style=solid];
-"2263 _param_constant357_0_0" -> "2264 linear_132"  [label="(1536,)", style=solid];
-"2264 linear_132" -> "2265 gelu_21"  [label="(1, 14, 14, 1536)", style=solid];
-"2265 gelu_21" -> "2266 dropout_86"  [label="(1, 14, 14, 1536)", style=solid];
-"2266 dropout_86" -> "2267 dropout_86_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
-"2267 dropout_86_0_0_nncf_smooth_quant_0" -> "2268 quantize_per_tensor_default_134"  [label="(1, 14, 14, 1536)", style=solid];
-"2268 quantize_per_tensor_default_134" -> "2269 dequantize_per_tensor_default_134"  [label="(1, 14, 14, 1536)", style=solid];
-"2269 dequantize_per_tensor_default_134" -> "2276 linear_133"  [label="(1, 14, 14, 1536)", style=solid];
-"2270 linear_133_scale_0" -> "2273 quantize_per_channel_default_112"  [label="(384,)", style=solid];
-"2270 linear_133_scale_0" -> "2274 dequantize_per_channel_default_112"  [label="(384,)", style=solid];
-"2271 linear_133_zero_point_0" -> "2273 quantize_per_channel_default_112"  [label="(384,)", style=solid];
-"2271 linear_133_zero_point_0" -> "2274 dequantize_per_channel_default_112"  [label="(384,)", style=solid];
-"2272 compressed_weight_updated_constant90" -> "2273 quantize_per_channel_default_112"  [label="(384, 1536)", style=solid];
-"2273 quantize_per_channel_default_112" -> "2274 dequantize_per_channel_default_112"  [label="(384, 1536)", style=solid];
-"2274 dequantize_per_channel_default_112" -> "2276 linear_133"  [label="(384, 1536)", style=solid];
-"2275 _param_constant359_0_0" -> "2276 linear_133"  [label="(384,)", style=solid];
-"2276 linear_133" -> "2277 dropout_87"  [label="(1, 14, 14, 384)", style=solid];
-"2277 dropout_87" -> "2280 layer_norm_46"  [label="(1, 14, 14, 384)", style=solid];
-"2278 _param_constant360" -> "2280 layer_norm_46"  [label="(384,)", style=solid];
-"2279 _param_constant361" -> "2280 layer_norm_46"  [label="(384,)", style=solid];
-"2280 layer_norm_46" -> "2281 add_76"  [label="(1, 14, 14, 384)", style=solid];
-"2281 add_76" -> "2282 pad_24"  [label="(1, 14, 14, 384)", style=solid];
-"2282 pad_24" -> "2283 slice_329"  [label="(1, 14, 14, 384)", style=solid];
-"2282 pad_24" -> "2286 slice_332"  [label="(1, 14, 14, 384)", style=solid];
-"2282 pad_24" -> "2289 slice_335"  [label="(1, 14, 14, 384)", style=solid];
-"2282 pad_24" -> "2292 slice_338"  [label="(1, 14, 14, 384)", style=solid];
-"2283 slice_329" -> "2284 slice_330"  [label="(1, 7, 14, 384)", style=solid];
-"2284 slice_330" -> "2285 slice_331"  [label="(1, 7, 7, 384)", style=solid];
-"2285 slice_331" -> "2295 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2286 slice_332" -> "2287 slice_333"  [label="(1, 7, 14, 384)", style=solid];
-"2287 slice_333" -> "2288 slice_334"  [label="(1, 7, 7, 384)", style=solid];
-"2288 slice_334" -> "2295 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2289 slice_335" -> "2290 slice_336"  [label="(1, 7, 14, 384)", style=solid];
-"2290 slice_336" -> "2291 slice_337"  [label="(1, 7, 7, 384)", style=solid];
-"2291 slice_337" -> "2295 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2292 slice_338" -> "2293 slice_339"  [label="(1, 7, 14, 384)", style=solid];
-"2293 slice_339" -> "2294 slice_340"  [label="(1, 7, 7, 384)", style=solid];
-"2294 slice_340" -> "2295 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2295 cat_2" -> "2296 cat_2_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 1536)", style=solid];
-"2296 cat_2_0_0_nncf_smooth_quant_0" -> "2297 quantize_per_tensor_default_135"  [label="(1, 7, 7, 1536)", style=solid];
-"2297 quantize_per_tensor_default_135" -> "2298 dequantize_per_tensor_default_135"  [label="(1, 7, 7, 1536)", style=solid];
-"2298 dequantize_per_tensor_default_135" -> "2304 linear_134"  [label="(1, 7, 7, 1536)", style=solid];
-"2299 linear_134_scale_0" -> "2302 quantize_per_channel_default_113"  [label="(768,)", style=solid];
-"2299 linear_134_scale_0" -> "2303 dequantize_per_channel_default_113"  [label="(768,)", style=solid];
-"2300 linear_134_zero_point_0" -> "2302 quantize_per_channel_default_113"  [label="(768,)", style=solid];
-"2300 linear_134_zero_point_0" -> "2303 dequantize_per_channel_default_113"  [label="(768,)", style=solid];
-"2301 compressed_weight_updated_constant91" -> "2302 quantize_per_channel_default_113"  [label="(768, 1536)", style=solid];
-"2302 quantize_per_channel_default_113" -> "2303 dequantize_per_channel_default_113"  [label="(768, 1536)", style=solid];
-"2303 dequantize_per_channel_default_113" -> "2304 linear_134"  [label="(768, 1536)", style=solid];
-"2304 linear_134" -> "2307 layer_norm_47"  [label="(1, 7, 7, 768)", style=solid];
-"2305 _param_constant363" -> "2307 layer_norm_47"  [label="(768,)", style=solid];
-"2306 _param_constant364" -> "2307 layer_norm_47"  [label="(768,)", style=solid];
-"2307 layer_norm_47" -> "2309 pad_25"  [label="(1, 7, 7, 768)", style=solid];
-"2307 layer_norm_47" -> "2372 add_78"  [label="(1, 7, 7, 768)", style=solid];
-"2308 _frozen_param187" -> "2344 add_77"  [label="(1, 24, 64, 64)", style=solid];
-"2309 pad_25" -> "2310 view_123"  [label="(1, 8, 8, 768)", style=solid];
-"2310 view_123" -> "2311 permute_101"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2311 permute_101" -> "2312 reshape_99"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2312 reshape_99" -> "2313 reshape_99_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
-"2313 reshape_99_0_0_nncf_smooth_quant_0" -> "2314 quantize_per_tensor_default_136"  [label="(1, 64, 768)", style=solid];
-"2314 quantize_per_tensor_default_136" -> "2315 dequantize_per_tensor_default_136"  [label="(1, 64, 768)", style=solid];
-"2315 dequantize_per_tensor_default_136" -> "2322 linear_137"  [label="(1, 64, 768)", style=solid];
-"2316 linear_137_scale_0" -> "2319 quantize_per_channel_default_115"  [label="(2304,)", style=solid];
-"2316 linear_137_scale_0" -> "2320 dequantize_per_channel_default_115"  [label="(2304,)", style=solid];
-"2317 linear_137_zero_point_0" -> "2319 quantize_per_channel_default_115"  [label="(2304,)", style=solid];
-"2317 linear_137_zero_point_0" -> "2320 dequantize_per_channel_default_115"  [label="(2304,)", style=solid];
-"2318 compressed_weight_updated_constant92" -> "2319 quantize_per_channel_default_115"  [label="(2304, 768)", style=solid];
-"2319 quantize_per_channel_default_115" -> "2320 dequantize_per_channel_default_115"  [label="(2304, 768)", style=solid];
-"2320 dequantize_per_channel_default_115" -> "2322 linear_137"  [label="(2304, 768)", style=solid];
-"2321 _frozen_param188_0_0" -> "2322 linear_137"  [label="(2304,)", style=solid];
-"2322 linear_137" -> "2323 reshape_100"  [label="(1, 64, 2304)", style=solid];
-"2323 reshape_100" -> "2324 permute_102"  [label="(1, 64, 3, 24, 32)", style=solid];
-"2324 permute_102" -> "2325 select_66"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2324 permute_102" -> "2326 select_67"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2324 permute_102" -> "2327 select_68"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2325 select_66" -> "2328 linalg_vector_norm_44"  [label="(1, 24, 64, 32)", style=solid];
-"2325 select_66" -> "2330 expand_as_44"  [label="(1, 24, 64, 32)", style=solid];
-"2325 select_66" -> "2331 div_44"  [label="(1, 24, 64, 32)", style=solid];
-"2326 select_67" -> "2334 linalg_vector_norm_45"  [label="(1, 24, 64, 32)", style=solid];
-"2326 select_67" -> "2336 expand_as_45"  [label="(1, 24, 64, 32)", style=solid];
-"2326 select_67" -> "2337 div_45"  [label="(1, 24, 64, 32)", style=solid];
-"2327 select_68" -> "2347 matmul_45"  [label="(1, 24, 64, 32)", style=solid];
-"2328 linalg_vector_norm_44" -> "2329 clamp_min_44"  [label="(1, 24, 64, 1)", style=solid];
-"2329 clamp_min_44" -> "2330 expand_as_44"  [label="(1, 24, 64, 1)", style=solid];
-"2330 expand_as_44" -> "2331 div_44"  [label="(1, 24, 64, 32)", style=solid];
-"2331 div_44" -> "2332 quantize_per_tensor_default_137"  [label="(1, 24, 64, 32)", style=solid];
-"2332 quantize_per_tensor_default_137" -> "2333 dequantize_per_tensor_default_137"  [label="(1, 24, 64, 32)", style=solid];
-"2333 dequantize_per_tensor_default_137" -> "2341 matmul_44"  [label="(1, 24, 64, 32)", style=solid];
-"2334 linalg_vector_norm_45" -> "2335 clamp_min_45"  [label="(1, 24, 64, 1)", style=solid];
-"2335 clamp_min_45" -> "2336 expand_as_45"  [label="(1, 24, 64, 1)", style=solid];
-"2336 expand_as_45" -> "2337 div_45"  [label="(1, 24, 64, 32)", style=solid];
-"2337 div_45" -> "2338 quantize_per_tensor_default_138"  [label="(1, 24, 64, 32)", style=solid];
-"2338 quantize_per_tensor_default_138" -> "2339 dequantize_per_tensor_default_138"  [label="(1, 24, 64, 32)", style=solid];
-"2339 dequantize_per_tensor_default_138" -> "2340 transpose_44"  [label="(1, 24, 64, 32)", style=solid];
-"2340 transpose_44" -> "2341 matmul_44"  [label="(1, 24, 32, 64)", style=solid];
-"2341 matmul_44" -> "2343 mul_45"  [label="(1, 24, 64, 64)", style=solid];
-"2342 _frozen_param190" -> "2343 mul_45"  [label="(24, 1, 1)", style=solid];
-"2343 mul_45" -> "2344 add_77"  [label="(1, 24, 64, 64)", style=solid];
-"2344 add_77" -> "2345 softmax_22"  [label="(1, 24, 64, 64)", style=solid];
-"2345 softmax_22" -> "2346 dropout_88"  [label="(1, 24, 64, 64)", style=solid];
-"2346 dropout_88" -> "2347 matmul_45"  [label="(1, 24, 64, 64)", style=solid];
-"2347 matmul_45" -> "2348 transpose_45"  [label="(1, 24, 64, 32)", style=solid];
-"2348 transpose_45" -> "2349 reshape_101"  [label="(1, 64, 24, 32)", style=solid];
-"2349 reshape_101" -> "2350 reshape_101_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
-"2350 reshape_101_0_0_nncf_smooth_quant_0" -> "2351 quantize_per_tensor_default_139"  [label="(1, 64, 768)", style=solid];
-"2351 quantize_per_tensor_default_139" -> "2352 dequantize_per_tensor_default_139"  [label="(1, 64, 768)", style=solid];
-"2352 dequantize_per_tensor_default_139" -> "2359 linear_138"  [label="(1, 64, 768)", style=solid];
-"2353 linear_138_scale_0" -> "2356 quantize_per_channel_default_116"  [label="(768,)", style=solid];
-"2353 linear_138_scale_0" -> "2357 dequantize_per_channel_default_116"  [label="(768,)", style=solid];
-"2354 linear_138_zero_point_0" -> "2356 quantize_per_channel_default_116"  [label="(768,)", style=solid];
-"2354 linear_138_zero_point_0" -> "2357 dequantize_per_channel_default_116"  [label="(768,)", style=solid];
-"2355 compressed_weight_updated_constant93" -> "2356 quantize_per_channel_default_116"  [label="(768, 768)", style=solid];
-"2356 quantize_per_channel_default_116" -> "2357 dequantize_per_channel_default_116"  [label="(768, 768)", style=solid];
-"2357 dequantize_per_channel_default_116" -> "2359 linear_138"  [label="(768, 768)", style=solid];
-"2358 _param_constant372_0_0" -> "2359 linear_138"  [label="(768,)", style=solid];
-"2359 linear_138" -> "2360 dropout_89"  [label="(1, 64, 768)", style=solid];
-"2360 dropout_89" -> "2361 view_124"  [label="(1, 64, 768)", style=solid];
-"2361 view_124" -> "2362 permute_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2362 permute_103" -> "2363 reshape_102"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2363 reshape_102" -> "2364 slice_342"  [label="(1, 8, 8, 768)", style=solid];
-"2364 slice_342" -> "2365 slice_343"  [label="(1, 8, 8, 768)", style=solid];
-"2365 slice_343" -> "2366 slice_344"  [label="(1, 7, 8, 768)", style=solid];
-"2366 slice_344" -> "2367 slice_345"  [label="(1, 7, 7, 768)", style=solid];
-"2367 slice_345" -> "2368 contiguous_43"  [label="(1, 7, 7, 768)", style=solid];
-"2368 contiguous_43" -> "2371 layer_norm_48"  [label="(1, 7, 7, 768)", style=solid];
-"2369 _param_constant373" -> "2371 layer_norm_48"  [label="(768,)", style=solid];
-"2370 _param_constant374" -> "2371 layer_norm_48"  [label="(768,)", style=solid];
-"2371 layer_norm_48" -> "2372 add_78"  [label="(1, 7, 7, 768)", style=solid];
-"2372 add_78" -> "2373 add_78_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 768)", style=solid];
-"2372 add_78" -> "2399 add_79"  [label="(1, 7, 7, 768)", style=solid];
-"2373 add_78_0_0_nncf_smooth_quant_0" -> "2374 quantize_per_tensor_default_140"  [label="(1, 7, 7, 768)", style=solid];
-"2374 quantize_per_tensor_default_140" -> "2375 dequantize_per_tensor_default_140"  [label="(1, 7, 7, 768)", style=solid];
-"2375 dequantize_per_tensor_default_140" -> "2382 linear_139"  [label="(1, 7, 7, 768)", style=solid];
-"2376 linear_139_scale_0" -> "2379 quantize_per_channel_default_117"  [label="(3072,)", style=solid];
-"2376 linear_139_scale_0" -> "2380 dequantize_per_channel_default_117"  [label="(3072,)", style=solid];
-"2377 linear_139_zero_point_0" -> "2379 quantize_per_channel_default_117"  [label="(3072,)", style=solid];
-"2377 linear_139_zero_point_0" -> "2380 dequantize_per_channel_default_117"  [label="(3072,)", style=solid];
-"2378 compressed_weight_updated_constant94" -> "2379 quantize_per_channel_default_117"  [label="(3072, 768)", style=solid];
-"2379 quantize_per_channel_default_117" -> "2380 dequantize_per_channel_default_117"  [label="(3072, 768)", style=solid];
-"2380 dequantize_per_channel_default_117" -> "2382 linear_139"  [label="(3072, 768)", style=solid];
-"2381 _param_constant376_0_0" -> "2382 linear_139"  [label="(3072,)", style=solid];
-"2382 linear_139" -> "2383 gelu_22"  [label="(1, 7, 7, 3072)", style=solid];
-"2383 gelu_22" -> "2384 dropout_90"  [label="(1, 7, 7, 3072)", style=solid];
-"2384 dropout_90" -> "2385 dropout_90_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 3072)", style=solid];
-"2385 dropout_90_0_0_nncf_smooth_quant_0" -> "2386 quantize_per_tensor_default_141"  [label="(1, 7, 7, 3072)", style=solid];
-"2386 quantize_per_tensor_default_141" -> "2387 dequantize_per_tensor_default_141"  [label="(1, 7, 7, 3072)", style=solid];
-"2387 dequantize_per_tensor_default_141" -> "2394 linear_140"  [label="(1, 7, 7, 3072)", style=solid];
-"2388 linear_140_scale_0" -> "2391 quantize_per_channel_default_118"  [label="(768,)", style=solid];
-"2388 linear_140_scale_0" -> "2392 dequantize_per_channel_default_118"  [label="(768,)", style=solid];
-"2389 linear_140_zero_point_0" -> "2391 quantize_per_channel_default_118"  [label="(768,)", style=solid];
-"2389 linear_140_zero_point_0" -> "2392 dequantize_per_channel_default_118"  [label="(768,)", style=solid];
-"2390 compressed_weight_updated_constant95" -> "2391 quantize_per_channel_default_118"  [label="(768, 3072)", style=solid];
-"2391 quantize_per_channel_default_118" -> "2392 dequantize_per_channel_default_118"  [label="(768, 3072)", style=solid];
-"2392 dequantize_per_channel_default_118" -> "2394 linear_140"  [label="(768, 3072)", style=solid];
-"2393 _param_constant378_0_0" -> "2394 linear_140"  [label="(768,)", style=solid];
-"2394 linear_140" -> "2395 dropout_91"  [label="(1, 7, 7, 768)", style=solid];
-"2395 dropout_91" -> "2398 layer_norm_49"  [label="(1, 7, 7, 768)", style=solid];
-"2396 _param_constant379" -> "2398 layer_norm_49"  [label="(768,)", style=solid];
-"2397 _param_constant380" -> "2398 layer_norm_49"  [label="(768,)", style=solid];
-"2398 layer_norm_49" -> "2399 add_79"  [label="(1, 7, 7, 768)", style=solid];
-"2399 add_79" -> "2401 pad_26"  [label="(1, 7, 7, 768)", style=solid];
-"2399 add_79" -> "2464 add_81"  [label="(1, 7, 7, 768)", style=solid];
-"2400 _frozen_param191" -> "2436 add_80"  [label="(1, 24, 64, 64)", style=solid];
-"2401 pad_26" -> "2402 view_127"  [label="(1, 8, 8, 768)", style=solid];
-"2402 view_127" -> "2403 permute_105"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2403 permute_105" -> "2404 reshape_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2404 reshape_103" -> "2405 reshape_103_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
-"2405 reshape_103_0_0_nncf_smooth_quant_0" -> "2406 quantize_per_tensor_default_142"  [label="(1, 64, 768)", style=solid];
-"2406 quantize_per_tensor_default_142" -> "2407 dequantize_per_tensor_default_142"  [label="(1, 64, 768)", style=solid];
-"2407 dequantize_per_tensor_default_142" -> "2414 linear_143"  [label="(1, 64, 768)", style=solid];
-"2408 linear_143_scale_0" -> "2411 quantize_per_channel_default_120"  [label="(2304,)", style=solid];
-"2408 linear_143_scale_0" -> "2412 dequantize_per_channel_default_120"  [label="(2304,)", style=solid];
-"2409 linear_143_zero_point_0" -> "2411 quantize_per_channel_default_120"  [label="(2304,)", style=solid];
-"2409 linear_143_zero_point_0" -> "2412 dequantize_per_channel_default_120"  [label="(2304,)", style=solid];
-"2410 compressed_weight_updated_constant96" -> "2411 quantize_per_channel_default_120"  [label="(2304, 768)", style=solid];
-"2411 quantize_per_channel_default_120" -> "2412 dequantize_per_channel_default_120"  [label="(2304, 768)", style=solid];
-"2412 dequantize_per_channel_default_120" -> "2414 linear_143"  [label="(2304, 768)", style=solid];
-"2413 _frozen_param192_0_0" -> "2414 linear_143"  [label="(2304,)", style=solid];
-"2414 linear_143" -> "2415 reshape_104"  [label="(1, 64, 2304)", style=solid];
-"2415 reshape_104" -> "2416 permute_106"  [label="(1, 64, 3, 24, 32)", style=solid];
-"2416 permute_106" -> "2417 select_69"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2416 permute_106" -> "2418 select_70"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2416 permute_106" -> "2419 select_71"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2417 select_69" -> "2420 linalg_vector_norm_46"  [label="(1, 24, 64, 32)", style=solid];
-"2417 select_69" -> "2422 expand_as_46"  [label="(1, 24, 64, 32)", style=solid];
-"2417 select_69" -> "2423 div_46"  [label="(1, 24, 64, 32)", style=solid];
-"2418 select_70" -> "2426 linalg_vector_norm_47"  [label="(1, 24, 64, 32)", style=solid];
-"2418 select_70" -> "2428 expand_as_47"  [label="(1, 24, 64, 32)", style=solid];
-"2418 select_70" -> "2429 div_47"  [label="(1, 24, 64, 32)", style=solid];
-"2419 select_71" -> "2439 matmul_47"  [label="(1, 24, 64, 32)", style=solid];
-"2420 linalg_vector_norm_46" -> "2421 clamp_min_46"  [label="(1, 24, 64, 1)", style=solid];
-"2421 clamp_min_46" -> "2422 expand_as_46"  [label="(1, 24, 64, 1)", style=solid];
-"2422 expand_as_46" -> "2423 div_46"  [label="(1, 24, 64, 32)", style=solid];
-"2423 div_46" -> "2424 quantize_per_tensor_default_143"  [label="(1, 24, 64, 32)", style=solid];
-"2424 quantize_per_tensor_default_143" -> "2425 dequantize_per_tensor_default_143"  [label="(1, 24, 64, 32)", style=solid];
-"2425 dequantize_per_tensor_default_143" -> "2433 matmul_46"  [label="(1, 24, 64, 32)", style=solid];
-"2426 linalg_vector_norm_47" -> "2427 clamp_min_47"  [label="(1, 24, 64, 1)", style=solid];
-"2427 clamp_min_47" -> "2428 expand_as_47"  [label="(1, 24, 64, 1)", style=solid];
-"2428 expand_as_47" -> "2429 div_47"  [label="(1, 24, 64, 32)", style=solid];
-"2429 div_47" -> "2430 quantize_per_tensor_default_144"  [label="(1, 24, 64, 32)", style=solid];
-"2430 quantize_per_tensor_default_144" -> "2431 dequantize_per_tensor_default_144"  [label="(1, 24, 64, 32)", style=solid];
-"2431 dequantize_per_tensor_default_144" -> "2432 transpose_46"  [label="(1, 24, 64, 32)", style=solid];
-"2432 transpose_46" -> "2433 matmul_46"  [label="(1, 24, 32, 64)", style=solid];
-"2433 matmul_46" -> "2435 mul_47"  [label="(1, 24, 64, 64)", style=solid];
-"2434 _frozen_param194" -> "2435 mul_47"  [label="(24, 1, 1)", style=solid];
-"2435 mul_47" -> "2436 add_80"  [label="(1, 24, 64, 64)", style=solid];
-"2436 add_80" -> "2437 softmax_23"  [label="(1, 24, 64, 64)", style=solid];
-"2437 softmax_23" -> "2438 dropout_92"  [label="(1, 24, 64, 64)", style=solid];
-"2438 dropout_92" -> "2439 matmul_47"  [label="(1, 24, 64, 64)", style=solid];
-"2439 matmul_47" -> "2440 transpose_47"  [label="(1, 24, 64, 32)", style=solid];
-"2440 transpose_47" -> "2441 reshape_105"  [label="(1, 64, 24, 32)", style=solid];
-"2441 reshape_105" -> "2442 reshape_105_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
-"2442 reshape_105_0_0_nncf_smooth_quant_0" -> "2443 quantize_per_tensor_default_145"  [label="(1, 64, 768)", style=solid];
-"2443 quantize_per_tensor_default_145" -> "2444 dequantize_per_tensor_default_145"  [label="(1, 64, 768)", style=solid];
-"2444 dequantize_per_tensor_default_145" -> "2451 linear_144"  [label="(1, 64, 768)", style=solid];
-"2445 linear_144_scale_0" -> "2448 quantize_per_channel_default_121"  [label="(768,)", style=solid];
-"2445 linear_144_scale_0" -> "2449 dequantize_per_channel_default_121"  [label="(768,)", style=solid];
-"2446 linear_144_zero_point_0" -> "2448 quantize_per_channel_default_121"  [label="(768,)", style=solid];
-"2446 linear_144_zero_point_0" -> "2449 dequantize_per_channel_default_121"  [label="(768,)", style=solid];
-"2447 compressed_weight_updated_constant97" -> "2448 quantize_per_channel_default_121"  [label="(768, 768)", style=solid];
-"2448 quantize_per_channel_default_121" -> "2449 dequantize_per_channel_default_121"  [label="(768, 768)", style=solid];
-"2449 dequantize_per_channel_default_121" -> "2451 linear_144"  [label="(768, 768)", style=solid];
-"2450 _param_constant388_0_0" -> "2451 linear_144"  [label="(768,)", style=solid];
-"2451 linear_144" -> "2452 dropout_93"  [label="(1, 64, 768)", style=solid];
-"2452 dropout_93" -> "2453 view_128"  [label="(1, 64, 768)", style=solid];
-"2453 view_128" -> "2454 permute_107"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2454 permute_107" -> "2455 reshape_106"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2455 reshape_106" -> "2456 slice_347"  [label="(1, 8, 8, 768)", style=solid];
-"2456 slice_347" -> "2457 slice_348"  [label="(1, 8, 8, 768)", style=solid];
-"2457 slice_348" -> "2458 slice_349"  [label="(1, 7, 8, 768)", style=solid];
-"2458 slice_349" -> "2459 slice_350"  [label="(1, 7, 7, 768)", style=solid];
-"2459 slice_350" -> "2460 contiguous_45"  [label="(1, 7, 7, 768)", style=solid];
-"2460 contiguous_45" -> "2463 layer_norm_50"  [label="(1, 7, 7, 768)", style=solid];
-"2461 _param_constant389" -> "2463 layer_norm_50"  [label="(768,)", style=solid];
-"2462 _param_constant390" -> "2463 layer_norm_50"  [label="(768,)", style=solid];
-"2463 layer_norm_50" -> "2464 add_81"  [label="(1, 7, 7, 768)", style=solid];
-"2464 add_81" -> "2465 add_81_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 768)", style=solid];
-"2464 add_81" -> "2491 add_82"  [label="(1, 7, 7, 768)", style=solid];
-"2465 add_81_0_0_nncf_smooth_quant_0" -> "2466 quantize_per_tensor_default_146"  [label="(1, 7, 7, 768)", style=solid];
-"2466 quantize_per_tensor_default_146" -> "2467 dequantize_per_tensor_default_146"  [label="(1, 7, 7, 768)", style=solid];
-"2467 dequantize_per_tensor_default_146" -> "2474 linear_145"  [label="(1, 7, 7, 768)", style=solid];
-"2468 linear_145_scale_0" -> "2471 quantize_per_channel_default_122"  [label="(3072,)", style=solid];
-"2468 linear_145_scale_0" -> "2472 dequantize_per_channel_default_122"  [label="(3072,)", style=solid];
-"2469 linear_145_zero_point_0" -> "2471 quantize_per_channel_default_122"  [label="(3072,)", style=solid];
-"2469 linear_145_zero_point_0" -> "2472 dequantize_per_channel_default_122"  [label="(3072,)", style=solid];
-"2470 compressed_weight_updated_constant98" -> "2471 quantize_per_channel_default_122"  [label="(3072, 768)", style=solid];
-"2471 quantize_per_channel_default_122" -> "2472 dequantize_per_channel_default_122"  [label="(3072, 768)", style=solid];
-"2472 dequantize_per_channel_default_122" -> "2474 linear_145"  [label="(3072, 768)", style=solid];
-"2473 _param_constant392_0_0" -> "2474 linear_145"  [label="(3072,)", style=solid];
-"2474 linear_145" -> "2475 gelu_23"  [label="(1, 7, 7, 3072)", style=solid];
-"2475 gelu_23" -> "2476 dropout_94"  [label="(1, 7, 7, 3072)", style=solid];
-"2476 dropout_94" -> "2477 dropout_94_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 3072)", style=solid];
-"2477 dropout_94_0_0_nncf_smooth_quant_0" -> "2478 quantize_per_tensor_default_147"  [label="(1, 7, 7, 3072)", style=solid];
-"2478 quantize_per_tensor_default_147" -> "2479 dequantize_per_tensor_default_147"  [label="(1, 7, 7, 3072)", style=solid];
-"2479 dequantize_per_tensor_default_147" -> "2486 linear_146"  [label="(1, 7, 7, 3072)", style=solid];
-"2480 linear_146_scale_0" -> "2483 quantize_per_channel_default_123"  [label="(768,)", style=solid];
-"2480 linear_146_scale_0" -> "2484 dequantize_per_channel_default_123"  [label="(768,)", style=solid];
-"2481 linear_146_zero_point_0" -> "2483 quantize_per_channel_default_123"  [label="(768,)", style=solid];
-"2481 linear_146_zero_point_0" -> "2484 dequantize_per_channel_default_123"  [label="(768,)", style=solid];
-"2482 compressed_weight_updated_constant99" -> "2483 quantize_per_channel_default_123"  [label="(768, 3072)", style=solid];
-"2483 quantize_per_channel_default_123" -> "2484 dequantize_per_channel_default_123"  [label="(768, 3072)", style=solid];
-"2484 dequantize_per_channel_default_123" -> "2486 linear_146"  [label="(768, 3072)", style=solid];
-"2485 _param_constant394_0_0" -> "2486 linear_146"  [label="(768,)", style=solid];
-"2486 linear_146" -> "2487 dropout_95"  [label="(1, 7, 7, 768)", style=solid];
-"2487 dropout_95" -> "2490 layer_norm_51"  [label="(1, 7, 7, 768)", style=solid];
-"2488 _param_constant395" -> "2490 layer_norm_51"  [label="(768,)", style=solid];
-"2489 _param_constant396" -> "2490 layer_norm_51"  [label="(768,)", style=solid];
-"2490 layer_norm_51" -> "2491 add_82"  [label="(1, 7, 7, 768)", style=solid];
-"2491 add_82" -> "2494 layer_norm_52"  [label="(1, 7, 7, 768)", style=solid];
-"2492 _param_constant397" -> "2494 layer_norm_52"  [label="(768,)", style=solid];
-"2493 _param_constant398" -> "2494 layer_norm_52"  [label="(768,)", style=solid];
+"0 features_0_2_weight" -> "117 layer_norm"  [label="(96,)", style=solid];
+"1 features_0_2_bias" -> "117 layer_norm"  [label="(96,)", style=solid];
+"2 features_1_0_norm1_weight" -> "176 layer_norm_1"  [label="(96,)", style=solid];
+"3 features_1_0_norm1_bias" -> "176 layer_norm_1"  [label="(96,)", style=solid];
+"4 features_1_0_norm2_weight" -> "201 layer_norm_2"  [label="(96,)", style=solid];
+"5 features_1_0_norm2_bias" -> "201 layer_norm_2"  [label="(96,)", style=solid];
+"6 features_1_1_norm1_weight" -> "279 layer_norm_3"  [label="(96,)", style=solid];
+"7 features_1_1_norm1_bias" -> "279 layer_norm_3"  [label="(96,)", style=solid];
+"8 features_1_1_norm2_weight" -> "304 layer_norm_4"  [label="(96,)", style=solid];
+"9 features_1_1_norm2_bias" -> "304 layer_norm_4"  [label="(96,)", style=solid];
+"10 features_2_norm_weight" -> "329 layer_norm_5"  [label="(192,)", style=solid];
+"11 features_2_norm_bias" -> "329 layer_norm_5"  [label="(192,)", style=solid];
+"12 features_3_0_norm1_weight" -> "391 layer_norm_6"  [label="(192,)", style=solid];
+"13 features_3_0_norm1_bias" -> "391 layer_norm_6"  [label="(192,)", style=solid];
+"14 features_3_0_norm2_weight" -> "416 layer_norm_7"  [label="(192,)", style=solid];
+"15 features_3_0_norm2_bias" -> "416 layer_norm_7"  [label="(192,)", style=solid];
+"16 features_3_1_norm1_weight" -> "497 layer_norm_8"  [label="(192,)", style=solid];
+"17 features_3_1_norm1_bias" -> "497 layer_norm_8"  [label="(192,)", style=solid];
+"18 features_3_1_norm2_weight" -> "522 layer_norm_9"  [label="(192,)", style=solid];
+"19 features_3_1_norm2_bias" -> "522 layer_norm_9"  [label="(192,)", style=solid];
+"20 features_4_norm_weight" -> "547 layer_norm_10"  [label="(384,)", style=solid];
+"21 features_4_norm_bias" -> "547 layer_norm_10"  [label="(384,)", style=solid];
+"22 features_5_0_norm1_weight" -> "609 layer_norm_11"  [label="(384,)", style=solid];
+"23 features_5_0_norm1_bias" -> "609 layer_norm_11"  [label="(384,)", style=solid];
+"24 features_5_0_norm2_weight" -> "634 layer_norm_12"  [label="(384,)", style=solid];
+"25 features_5_0_norm2_bias" -> "634 layer_norm_12"  [label="(384,)", style=solid];
+"26 features_5_1_norm1_weight" -> "715 layer_norm_13"  [label="(384,)", style=solid];
+"27 features_5_1_norm1_bias" -> "715 layer_norm_13"  [label="(384,)", style=solid];
+"28 features_5_1_norm2_weight" -> "740 layer_norm_14"  [label="(384,)", style=solid];
+"29 features_5_1_norm2_bias" -> "740 layer_norm_14"  [label="(384,)", style=solid];
+"30 features_5_2_norm1_weight" -> "803 layer_norm_15"  [label="(384,)", style=solid];
+"31 features_5_2_norm1_bias" -> "803 layer_norm_15"  [label="(384,)", style=solid];
+"32 features_5_2_norm2_weight" -> "828 layer_norm_16"  [label="(384,)", style=solid];
+"33 features_5_2_norm2_bias" -> "828 layer_norm_16"  [label="(384,)", style=solid];
+"34 features_5_3_norm1_weight" -> "909 layer_norm_17"  [label="(384,)", style=solid];
+"35 features_5_3_norm1_bias" -> "909 layer_norm_17"  [label="(384,)", style=solid];
+"36 features_5_3_norm2_weight" -> "934 layer_norm_18"  [label="(384,)", style=solid];
+"37 features_5_3_norm2_bias" -> "934 layer_norm_18"  [label="(384,)", style=solid];
+"38 features_5_4_norm1_weight" -> "997 layer_norm_19"  [label="(384,)", style=solid];
+"39 features_5_4_norm1_bias" -> "997 layer_norm_19"  [label="(384,)", style=solid];
+"40 features_5_4_norm2_weight" -> "1022 layer_norm_20"  [label="(384,)", style=solid];
+"41 features_5_4_norm2_bias" -> "1022 layer_norm_20"  [label="(384,)", style=solid];
+"42 features_5_5_norm1_weight" -> "1103 layer_norm_21"  [label="(384,)", style=solid];
+"43 features_5_5_norm1_bias" -> "1103 layer_norm_21"  [label="(384,)", style=solid];
+"44 features_5_5_norm2_weight" -> "1128 layer_norm_22"  [label="(384,)", style=solid];
+"45 features_5_5_norm2_bias" -> "1128 layer_norm_22"  [label="(384,)", style=solid];
+"46 features_5_6_norm1_weight" -> "1191 layer_norm_23"  [label="(384,)", style=solid];
+"47 features_5_6_norm1_bias" -> "1191 layer_norm_23"  [label="(384,)", style=solid];
+"48 features_5_6_norm2_weight" -> "1216 layer_norm_24"  [label="(384,)", style=solid];
+"49 features_5_6_norm2_bias" -> "1216 layer_norm_24"  [label="(384,)", style=solid];
+"50 features_5_7_norm1_weight" -> "1297 layer_norm_25"  [label="(384,)", style=solid];
+"51 features_5_7_norm1_bias" -> "1297 layer_norm_25"  [label="(384,)", style=solid];
+"52 features_5_7_norm2_weight" -> "1322 layer_norm_26"  [label="(384,)", style=solid];
+"53 features_5_7_norm2_bias" -> "1322 layer_norm_26"  [label="(384,)", style=solid];
+"54 features_5_8_norm1_weight" -> "1385 layer_norm_27"  [label="(384,)", style=solid];
+"55 features_5_8_norm1_bias" -> "1385 layer_norm_27"  [label="(384,)", style=solid];
+"56 features_5_8_norm2_weight" -> "1410 layer_norm_28"  [label="(384,)", style=solid];
+"57 features_5_8_norm2_bias" -> "1410 layer_norm_28"  [label="(384,)", style=solid];
+"58 features_5_9_norm1_weight" -> "1491 layer_norm_29"  [label="(384,)", style=solid];
+"59 features_5_9_norm1_bias" -> "1491 layer_norm_29"  [label="(384,)", style=solid];
+"60 features_5_9_norm2_weight" -> "1516 layer_norm_30"  [label="(384,)", style=solid];
+"61 features_5_9_norm2_bias" -> "1516 layer_norm_30"  [label="(384,)", style=solid];
+"62 features_5_10_norm1_weight" -> "1579 layer_norm_31"  [label="(384,)", style=solid];
+"63 features_5_10_norm1_bias" -> "1579 layer_norm_31"  [label="(384,)", style=solid];
+"64 features_5_10_norm2_weight" -> "1604 layer_norm_32"  [label="(384,)", style=solid];
+"65 features_5_10_norm2_bias" -> "1604 layer_norm_32"  [label="(384,)", style=solid];
+"66 features_5_11_norm1_weight" -> "1685 layer_norm_33"  [label="(384,)", style=solid];
+"67 features_5_11_norm1_bias" -> "1685 layer_norm_33"  [label="(384,)", style=solid];
+"68 features_5_11_norm2_weight" -> "1710 layer_norm_34"  [label="(384,)", style=solid];
+"69 features_5_11_norm2_bias" -> "1710 layer_norm_34"  [label="(384,)", style=solid];
+"70 features_5_12_norm1_weight" -> "1773 layer_norm_35"  [label="(384,)", style=solid];
+"71 features_5_12_norm1_bias" -> "1773 layer_norm_35"  [label="(384,)", style=solid];
+"72 features_5_12_norm2_weight" -> "1798 layer_norm_36"  [label="(384,)", style=solid];
+"73 features_5_12_norm2_bias" -> "1798 layer_norm_36"  [label="(384,)", style=solid];
+"74 features_5_13_norm1_weight" -> "1879 layer_norm_37"  [label="(384,)", style=solid];
+"75 features_5_13_norm1_bias" -> "1879 layer_norm_37"  [label="(384,)", style=solid];
+"76 features_5_13_norm2_weight" -> "1904 layer_norm_38"  [label="(384,)", style=solid];
+"77 features_5_13_norm2_bias" -> "1904 layer_norm_38"  [label="(384,)", style=solid];
+"78 features_5_14_norm1_weight" -> "1967 layer_norm_39"  [label="(384,)", style=solid];
+"79 features_5_14_norm1_bias" -> "1967 layer_norm_39"  [label="(384,)", style=solid];
+"80 features_5_14_norm2_weight" -> "1992 layer_norm_40"  [label="(384,)", style=solid];
+"81 features_5_14_norm2_bias" -> "1992 layer_norm_40"  [label="(384,)", style=solid];
+"82 features_5_15_norm1_weight" -> "2073 layer_norm_41"  [label="(384,)", style=solid];
+"83 features_5_15_norm1_bias" -> "2073 layer_norm_41"  [label="(384,)", style=solid];
+"84 features_5_15_norm2_weight" -> "2098 layer_norm_42"  [label="(384,)", style=solid];
+"85 features_5_15_norm2_bias" -> "2098 layer_norm_42"  [label="(384,)", style=solid];
+"86 features_5_16_norm1_weight" -> "2161 layer_norm_43"  [label="(384,)", style=solid];
+"87 features_5_16_norm1_bias" -> "2161 layer_norm_43"  [label="(384,)", style=solid];
+"88 features_5_16_norm2_weight" -> "2186 layer_norm_44"  [label="(384,)", style=solid];
+"89 features_5_16_norm2_bias" -> "2186 layer_norm_44"  [label="(384,)", style=solid];
+"90 features_5_17_norm1_weight" -> "2267 layer_norm_45"  [label="(384,)", style=solid];
+"91 features_5_17_norm1_bias" -> "2267 layer_norm_45"  [label="(384,)", style=solid];
+"92 features_5_17_norm2_weight" -> "2292 layer_norm_46"  [label="(384,)", style=solid];
+"93 features_5_17_norm2_bias" -> "2292 layer_norm_46"  [label="(384,)", style=solid];
+"94 features_6_norm_weight" -> "2317 layer_norm_47"  [label="(768,)", style=solid];
+"95 features_6_norm_bias" -> "2317 layer_norm_47"  [label="(768,)", style=solid];
+"96 features_7_0_norm1_weight" -> "2379 layer_norm_48"  [label="(768,)", style=solid];
+"97 features_7_0_norm1_bias" -> "2379 layer_norm_48"  [label="(768,)", style=solid];
+"98 features_7_0_norm2_weight" -> "2404 layer_norm_49"  [label="(768,)", style=solid];
+"99 features_7_0_norm2_bias" -> "2404 layer_norm_49"  [label="(768,)", style=solid];
+"100 features_7_1_norm1_weight" -> "2467 layer_norm_50"  [label="(768,)", style=solid];
+"101 features_7_1_norm1_bias" -> "2467 layer_norm_50"  [label="(768,)", style=solid];
+"102 features_7_1_norm2_weight" -> "2492 layer_norm_51"  [label="(768,)", style=solid];
+"103 features_7_1_norm2_bias" -> "2492 layer_norm_51"  [label="(768,)", style=solid];
+"104 norm_weight" -> "2494 layer_norm_52"  [label="(768,)", style=solid];
+"105 norm_bias" -> "2494 layer_norm_52"  [label="(768,)", style=solid];
+"106 x" -> "107 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"107 quantize_per_tensor_default" -> "108 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"108 dequantize_per_tensor_default" -> "115 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"109 conv2d_scale_0" -> "112 quantize_per_channel_default"  [label="(96,)", style=solid];
+"109 conv2d_scale_0" -> "113 dequantize_per_channel_default"  [label="(96,)", style=solid];
+"110 conv2d_zero_point_0" -> "112 quantize_per_channel_default"  [label="(96,)", style=solid];
+"110 conv2d_zero_point_0" -> "113 dequantize_per_channel_default"  [label="(96,)", style=solid];
+"111 compressed_weight_updated_constant0" -> "112 quantize_per_channel_default"  [label="(96, 3, 4, 4)", style=solid];
+"112 quantize_per_channel_default" -> "113 dequantize_per_channel_default"  [label="(96, 3, 4, 4)", style=solid];
+"113 dequantize_per_channel_default" -> "115 conv2d"  [label="(96, 3, 4, 4)", style=solid];
+"114 features_0_0_bias_0_0" -> "115 conv2d"  [label="(96,)", style=solid];
+"115 conv2d" -> "116 permute"  [label="(1, 96, 56, 56)", style=solid];
+"116 permute" -> "117 layer_norm"  [label="(1, 56, 56, 96)", style=solid];
+"117 layer_norm" -> "119 pad"  [label="(1, 56, 56, 96)", style=solid];
+"117 layer_norm" -> "177 add_1"  [label="(1, 56, 56, 96)", style=solid];
+"118 _frozen_param0" -> "154 add"  [label="(1, 3, 64, 64)", style=solid];
+"119 pad" -> "120 view_2"  [label="(1, 56, 56, 96)", style=solid];
+"120 view_2" -> "121 permute_2"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"121 permute_2" -> "122 reshape"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"122 reshape" -> "123 reshape_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
+"123 reshape_0_0_nncf_smooth_quant_0" -> "124 quantize_per_tensor_default_1"  [label="(49, 64, 96)", style=solid];
+"124 quantize_per_tensor_default_1" -> "125 dequantize_per_tensor_default_1"  [label="(49, 64, 96)", style=solid];
+"125 dequantize_per_tensor_default_1" -> "132 linear_2"  [label="(49, 64, 96)", style=solid];
+"126 linear_2_scale_0" -> "129 quantize_per_channel_default_2"  [label="(288,)", style=solid];
+"126 linear_2_scale_0" -> "130 dequantize_per_channel_default_2"  [label="(288,)", style=solid];
+"127 linear_2_zero_point_0" -> "129 quantize_per_channel_default_2"  [label="(288,)", style=solid];
+"127 linear_2_zero_point_0" -> "130 dequantize_per_channel_default_2"  [label="(288,)", style=solid];
+"128 compressed_weight_updated_constant1" -> "129 quantize_per_channel_default_2"  [label="(288, 96)", style=solid];
+"129 quantize_per_channel_default_2" -> "130 dequantize_per_channel_default_2"  [label="(288, 96)", style=solid];
+"130 dequantize_per_channel_default_2" -> "132 linear_2"  [label="(288, 96)", style=solid];
+"131 _frozen_param1_0_0" -> "132 linear_2"  [label="(288,)", style=solid];
+"132 linear_2" -> "133 reshape_1"  [label="(49, 64, 288)", style=solid];
+"133 reshape_1" -> "134 permute_3"  [label="(49, 64, 3, 3, 32)", style=solid];
+"134 permute_3" -> "135 select"  [label="(3, 49, 3, 64, 32)", style=solid];
+"134 permute_3" -> "136 select_1"  [label="(3, 49, 3, 64, 32)", style=solid];
+"134 permute_3" -> "137 select_2"  [label="(3, 49, 3, 64, 32)", style=solid];
+"135 select" -> "138 linalg_vector_norm"  [label="(49, 3, 64, 32)", style=solid];
+"135 select" -> "140 expand_as"  [label="(49, 3, 64, 32)", style=solid];
+"135 select" -> "141 div"  [label="(49, 3, 64, 32)", style=solid];
+"136 select_1" -> "144 linalg_vector_norm_1"  [label="(49, 3, 64, 32)", style=solid];
+"136 select_1" -> "146 expand_as_1"  [label="(49, 3, 64, 32)", style=solid];
+"136 select_1" -> "147 div_1"  [label="(49, 3, 64, 32)", style=solid];
+"137 select_2" -> "157 matmul_1"  [label="(49, 3, 64, 32)", style=solid];
+"138 linalg_vector_norm" -> "139 clamp_min"  [label="(49, 3, 64, 1)", style=solid];
+"139 clamp_min" -> "140 expand_as"  [label="(49, 3, 64, 1)", style=solid];
+"140 expand_as" -> "141 div"  [label="(49, 3, 64, 32)", style=solid];
+"141 div" -> "142 quantize_per_tensor_default_2"  [label="(49, 3, 64, 32)", style=solid];
+"142 quantize_per_tensor_default_2" -> "143 dequantize_per_tensor_default_2"  [label="(49, 3, 64, 32)", style=solid];
+"143 dequantize_per_tensor_default_2" -> "151 matmul"  [label="(49, 3, 64, 32)", style=solid];
+"144 linalg_vector_norm_1" -> "145 clamp_min_1"  [label="(49, 3, 64, 1)", style=solid];
+"145 clamp_min_1" -> "146 expand_as_1"  [label="(49, 3, 64, 1)", style=solid];
+"146 expand_as_1" -> "147 div_1"  [label="(49, 3, 64, 32)", style=solid];
+"147 div_1" -> "148 quantize_per_tensor_default_3"  [label="(49, 3, 64, 32)", style=solid];
+"148 quantize_per_tensor_default_3" -> "149 dequantize_per_tensor_default_3"  [label="(49, 3, 64, 32)", style=solid];
+"149 dequantize_per_tensor_default_3" -> "150 transpose"  [label="(49, 3, 64, 32)", style=solid];
+"150 transpose" -> "151 matmul"  [label="(49, 3, 32, 64)", style=solid];
+"151 matmul" -> "153 mul_1"  [label="(49, 3, 64, 64)", style=solid];
+"152 _frozen_param3" -> "153 mul_1"  [label="(3, 1, 1)", style=solid];
+"153 mul_1" -> "154 add"  [label="(49, 3, 64, 64)", style=solid];
+"154 add" -> "155 softmax"  [label="(49, 3, 64, 64)", style=solid];
+"155 softmax" -> "156 dropout"  [label="(49, 3, 64, 64)", style=solid];
+"156 dropout" -> "157 matmul_1"  [label="(49, 3, 64, 64)", style=solid];
+"157 matmul_1" -> "158 transpose_1"  [label="(49, 3, 64, 32)", style=solid];
+"158 transpose_1" -> "159 reshape_2"  [label="(49, 64, 3, 32)", style=solid];
+"159 reshape_2" -> "160 reshape_2_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
+"160 reshape_2_0_0_nncf_smooth_quant_0" -> "161 quantize_per_tensor_default_4"  [label="(49, 64, 96)", style=solid];
+"161 quantize_per_tensor_default_4" -> "162 dequantize_per_tensor_default_4"  [label="(49, 64, 96)", style=solid];
+"162 dequantize_per_tensor_default_4" -> "169 linear_3"  [label="(49, 64, 96)", style=solid];
+"163 linear_3_scale_0" -> "166 quantize_per_channel_default_3"  [label="(96,)", style=solid];
+"163 linear_3_scale_0" -> "167 dequantize_per_channel_default_3"  [label="(96,)", style=solid];
+"164 linear_3_zero_point_0" -> "166 quantize_per_channel_default_3"  [label="(96,)", style=solid];
+"164 linear_3_zero_point_0" -> "167 dequantize_per_channel_default_3"  [label="(96,)", style=solid];
+"165 compressed_weight_updated_constant2" -> "166 quantize_per_channel_default_3"  [label="(96, 96)", style=solid];
+"166 quantize_per_channel_default_3" -> "167 dequantize_per_channel_default_3"  [label="(96, 96)", style=solid];
+"167 dequantize_per_channel_default_3" -> "169 linear_3"  [label="(96, 96)", style=solid];
+"168 features_1_0_attn_proj_bias_0_0" -> "169 linear_3"  [label="(96,)", style=solid];
+"169 linear_3" -> "170 dropout_1"  [label="(49, 64, 96)", style=solid];
+"170 dropout_1" -> "171 view_3"  [label="(49, 64, 96)", style=solid];
+"171 view_3" -> "172 permute_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"172 permute_4" -> "173 reshape_3"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"173 reshape_3" -> "174 slice_2"  [label="(1, 56, 56, 96)", style=solid];
+"174 slice_2" -> "175 slice_3"  [label="(1, 56, 56, 96)", style=solid];
+"175 slice_3" -> "176 layer_norm_1"  [label="(1, 56, 56, 96)", style=solid];
+"176 layer_norm_1" -> "177 add_1"  [label="(1, 56, 56, 96)", style=solid];
+"177 add_1" -> "178 add_1_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 96)", style=solid];
+"177 add_1" -> "202 add_2"  [label="(1, 56, 56, 96)", style=solid];
+"178 add_1_0_0_nncf_smooth_quant_0" -> "179 quantize_per_tensor_default_5"  [label="(1, 56, 56, 96)", style=solid];
+"179 quantize_per_tensor_default_5" -> "180 dequantize_per_tensor_default_5"  [label="(1, 56, 56, 96)", style=solid];
+"180 dequantize_per_tensor_default_5" -> "187 linear_4"  [label="(1, 56, 56, 96)", style=solid];
+"181 linear_4_scale_0" -> "184 quantize_per_channel_default_4"  [label="(384,)", style=solid];
+"181 linear_4_scale_0" -> "185 dequantize_per_channel_default_4"  [label="(384,)", style=solid];
+"182 linear_4_zero_point_0" -> "184 quantize_per_channel_default_4"  [label="(384,)", style=solid];
+"182 linear_4_zero_point_0" -> "185 dequantize_per_channel_default_4"  [label="(384,)", style=solid];
+"183 compressed_weight_updated_constant3" -> "184 quantize_per_channel_default_4"  [label="(384, 96)", style=solid];
+"184 quantize_per_channel_default_4" -> "185 dequantize_per_channel_default_4"  [label="(384, 96)", style=solid];
+"185 dequantize_per_channel_default_4" -> "187 linear_4"  [label="(384, 96)", style=solid];
+"186 features_1_0_mlp_0_bias_0_0" -> "187 linear_4"  [label="(384,)", style=solid];
+"187 linear_4" -> "188 gelu"  [label="(1, 56, 56, 384)", style=solid];
+"188 gelu" -> "189 dropout_2"  [label="(1, 56, 56, 384)", style=solid];
+"189 dropout_2" -> "190 dropout_2_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 384)", style=solid];
+"190 dropout_2_0_0_nncf_smooth_quant_0" -> "191 quantize_per_tensor_default_6"  [label="(1, 56, 56, 384)", style=solid];
+"191 quantize_per_tensor_default_6" -> "192 dequantize_per_tensor_default_6"  [label="(1, 56, 56, 384)", style=solid];
+"192 dequantize_per_tensor_default_6" -> "199 linear_5"  [label="(1, 56, 56, 384)", style=solid];
+"193 linear_5_scale_0" -> "196 quantize_per_channel_default_5"  [label="(96,)", style=solid];
+"193 linear_5_scale_0" -> "197 dequantize_per_channel_default_5"  [label="(96,)", style=solid];
+"194 linear_5_zero_point_0" -> "196 quantize_per_channel_default_5"  [label="(96,)", style=solid];
+"194 linear_5_zero_point_0" -> "197 dequantize_per_channel_default_5"  [label="(96,)", style=solid];
+"195 compressed_weight_updated_constant4" -> "196 quantize_per_channel_default_5"  [label="(96, 384)", style=solid];
+"196 quantize_per_channel_default_5" -> "197 dequantize_per_channel_default_5"  [label="(96, 384)", style=solid];
+"197 dequantize_per_channel_default_5" -> "199 linear_5"  [label="(96, 384)", style=solid];
+"198 features_1_0_mlp_3_bias_0_0" -> "199 linear_5"  [label="(96,)", style=solid];
+"199 linear_5" -> "200 dropout_3"  [label="(1, 56, 56, 96)", style=solid];
+"200 dropout_3" -> "201 layer_norm_2"  [label="(1, 56, 56, 96)", style=solid];
+"201 layer_norm_2" -> "202 add_2"  [label="(1, 56, 56, 96)", style=solid];
+"202 add_2" -> "204 pad_1"  [label="(1, 56, 56, 96)", style=solid];
+"202 add_2" -> "280 add_5"  [label="(1, 56, 56, 96)", style=solid];
+"203 _frozen_param4" -> "240 add_3"  [label="(1, 3, 64, 64)", style=solid];
+"204 pad_1" -> "205 roll"  [label="(1, 56, 56, 96)", style=solid];
+"205 roll" -> "206 view_6"  [label="(1, 56, 56, 96)", style=solid];
+"206 view_6" -> "207 permute_6"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"207 permute_6" -> "208 reshape_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"208 reshape_4" -> "209 reshape_4_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
+"208 reshape_4" -> "241 new_zeros"  [label="(49, 64, 96)", style=solid];
+"209 reshape_4_0_0_nncf_smooth_quant_0" -> "210 quantize_per_tensor_default_7"  [label="(49, 64, 96)", style=solid];
+"210 quantize_per_tensor_default_7" -> "211 dequantize_per_tensor_default_7"  [label="(49, 64, 96)", style=solid];
+"211 dequantize_per_tensor_default_7" -> "218 linear_8"  [label="(49, 64, 96)", style=solid];
+"212 linear_8_scale_0" -> "215 quantize_per_channel_default_7"  [label="(288,)", style=solid];
+"212 linear_8_scale_0" -> "216 dequantize_per_channel_default_7"  [label="(288,)", style=solid];
+"213 linear_8_zero_point_0" -> "215 quantize_per_channel_default_7"  [label="(288,)", style=solid];
+"213 linear_8_zero_point_0" -> "216 dequantize_per_channel_default_7"  [label="(288,)", style=solid];
+"214 compressed_weight_updated_constant5" -> "215 quantize_per_channel_default_7"  [label="(288, 96)", style=solid];
+"215 quantize_per_channel_default_7" -> "216 dequantize_per_channel_default_7"  [label="(288, 96)", style=solid];
+"216 dequantize_per_channel_default_7" -> "218 linear_8"  [label="(288, 96)", style=solid];
+"217 _frozen_param5_0_0" -> "218 linear_8"  [label="(288,)", style=solid];
+"218 linear_8" -> "219 reshape_5"  [label="(49, 64, 288)", style=solid];
+"219 reshape_5" -> "220 permute_7"  [label="(49, 64, 3, 3, 32)", style=solid];
+"220 permute_7" -> "221 select_3"  [label="(3, 49, 3, 64, 32)", style=solid];
+"220 permute_7" -> "222 select_4"  [label="(3, 49, 3, 64, 32)", style=solid];
+"220 permute_7" -> "223 select_5"  [label="(3, 49, 3, 64, 32)", style=solid];
+"221 select_3" -> "224 linalg_vector_norm_2"  [label="(49, 3, 64, 32)", style=solid];
+"221 select_3" -> "226 expand_as_2"  [label="(49, 3, 64, 32)", style=solid];
+"221 select_3" -> "227 div_2"  [label="(49, 3, 64, 32)", style=solid];
+"222 select_4" -> "230 linalg_vector_norm_3"  [label="(49, 3, 64, 32)", style=solid];
+"222 select_4" -> "232 expand_as_3"  [label="(49, 3, 64, 32)", style=solid];
+"222 select_4" -> "233 div_3"  [label="(49, 3, 64, 32)", style=solid];
+"223 select_5" -> "259 matmul_3"  [label="(49, 3, 64, 32)", style=solid];
+"224 linalg_vector_norm_2" -> "225 clamp_min_2"  [label="(49, 3, 64, 1)", style=solid];
+"225 clamp_min_2" -> "226 expand_as_2"  [label="(49, 3, 64, 1)", style=solid];
+"226 expand_as_2" -> "227 div_2"  [label="(49, 3, 64, 32)", style=solid];
+"227 div_2" -> "228 quantize_per_tensor_default_8"  [label="(49, 3, 64, 32)", style=solid];
+"228 quantize_per_tensor_default_8" -> "229 dequantize_per_tensor_default_8"  [label="(49, 3, 64, 32)", style=solid];
+"229 dequantize_per_tensor_default_8" -> "237 matmul_2"  [label="(49, 3, 64, 32)", style=solid];
+"230 linalg_vector_norm_3" -> "231 clamp_min_3"  [label="(49, 3, 64, 1)", style=solid];
+"231 clamp_min_3" -> "232 expand_as_3"  [label="(49, 3, 64, 1)", style=solid];
+"232 expand_as_3" -> "233 div_3"  [label="(49, 3, 64, 32)", style=solid];
+"233 div_3" -> "234 quantize_per_tensor_default_9"  [label="(49, 3, 64, 32)", style=solid];
+"234 quantize_per_tensor_default_9" -> "235 dequantize_per_tensor_default_9"  [label="(49, 3, 64, 32)", style=solid];
+"235 dequantize_per_tensor_default_9" -> "236 transpose_2"  [label="(49, 3, 64, 32)", style=solid];
+"236 transpose_2" -> "237 matmul_2"  [label="(49, 3, 32, 64)", style=solid];
+"237 matmul_2" -> "239 mul_3"  [label="(49, 3, 64, 64)", style=solid];
+"238 _frozen_param7" -> "239 mul_3"  [label="(3, 1, 1)", style=solid];
+"239 mul_3" -> "240 add_3"  [label="(49, 3, 64, 64)", style=solid];
+"240 add_3" -> "252 view_8"  [label="(49, 3, 64, 64)", style=solid];
+"241 new_zeros" -> "242 view_7"  [label="(56, 56)", style=solid];
+"242 view_7" -> "243 permute_8"  [label="(7, 8, 7, 8)", style=solid];
+"243 permute_8" -> "244 reshape_6"  [label="(7, 7, 8, 8)", style=solid];
+"244 reshape_6" -> "245 unsqueeze_2"  [label="(49, 64)", style=solid];
+"244 reshape_6" -> "246 unsqueeze_3"  [label="(49, 64)", style=solid];
+"245 unsqueeze_2" -> "247 sub"  [label="(49, 1, 64)", style=solid];
+"246 unsqueeze_3" -> "247 sub"  [label="(49, 64, 1)", style=solid];
+"247 sub" -> "248 ne"  [label="(49, 64, 64)", style=solid];
+"247 sub" -> "249 masked_fill"  [label="(49, 64, 64)", style=solid];
+"247 sub" -> "250 eq"  [label="(49, 64, 64)", style=solid];
+"248 ne" -> "249 masked_fill"  [label="(49, 64, 64)", style=solid];
+"249 masked_fill" -> "251 masked_fill_1"  [label="(49, 64, 64)", style=solid];
+"250 eq" -> "251 masked_fill_1"  [label="(49, 64, 64)", style=solid];
+"251 masked_fill_1" -> "253 unsqueeze_4"  [label="(49, 64, 64)", style=solid];
+"252 view_8" -> "255 add_4"  [label="(1, 49, 3, 64, 64)", style=solid];
+"253 unsqueeze_4" -> "254 unsqueeze_5"  [label="(49, 1, 64, 64)", style=solid];
+"254 unsqueeze_5" -> "255 add_4"  [label="(1, 49, 1, 64, 64)", style=solid];
+"255 add_4" -> "256 view_9"  [label="(1, 49, 3, 64, 64)", style=solid];
+"256 view_9" -> "257 softmax_1"  [label="(49, 3, 64, 64)", style=solid];
+"257 softmax_1" -> "258 dropout_4"  [label="(49, 3, 64, 64)", style=solid];
+"258 dropout_4" -> "259 matmul_3"  [label="(49, 3, 64, 64)", style=solid];
+"259 matmul_3" -> "260 transpose_3"  [label="(49, 3, 64, 32)", style=solid];
+"260 transpose_3" -> "261 reshape_7"  [label="(49, 64, 3, 32)", style=solid];
+"261 reshape_7" -> "262 reshape_7_0_0_nncf_smooth_quant_0"  [label="(49, 64, 96)", style=solid];
+"262 reshape_7_0_0_nncf_smooth_quant_0" -> "263 quantize_per_tensor_default_10"  [label="(49, 64, 96)", style=solid];
+"263 quantize_per_tensor_default_10" -> "264 dequantize_per_tensor_default_10"  [label="(49, 64, 96)", style=solid];
+"264 dequantize_per_tensor_default_10" -> "271 linear_9"  [label="(49, 64, 96)", style=solid];
+"265 linear_9_scale_0" -> "268 quantize_per_channel_default_8"  [label="(96,)", style=solid];
+"265 linear_9_scale_0" -> "269 dequantize_per_channel_default_8"  [label="(96,)", style=solid];
+"266 linear_9_zero_point_0" -> "268 quantize_per_channel_default_8"  [label="(96,)", style=solid];
+"266 linear_9_zero_point_0" -> "269 dequantize_per_channel_default_8"  [label="(96,)", style=solid];
+"267 compressed_weight_updated_constant6" -> "268 quantize_per_channel_default_8"  [label="(96, 96)", style=solid];
+"268 quantize_per_channel_default_8" -> "269 dequantize_per_channel_default_8"  [label="(96, 96)", style=solid];
+"269 dequantize_per_channel_default_8" -> "271 linear_9"  [label="(96, 96)", style=solid];
+"270 features_1_1_attn_proj_bias_0_0" -> "271 linear_9"  [label="(96,)", style=solid];
+"271 linear_9" -> "272 dropout_5"  [label="(49, 64, 96)", style=solid];
+"272 dropout_5" -> "273 view_10"  [label="(49, 64, 96)", style=solid];
+"273 view_10" -> "274 permute_9"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"274 permute_9" -> "275 reshape_8"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"275 reshape_8" -> "276 roll_1"  [label="(1, 56, 56, 96)", style=solid];
+"276 roll_1" -> "277 slice_23"  [label="(1, 56, 56, 96)", style=solid];
+"277 slice_23" -> "278 slice_24"  [label="(1, 56, 56, 96)", style=solid];
+"278 slice_24" -> "279 layer_norm_3"  [label="(1, 56, 56, 96)", style=solid];
+"279 layer_norm_3" -> "280 add_5"  [label="(1, 56, 56, 96)", style=solid];
+"280 add_5" -> "281 add_5_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 96)", style=solid];
+"280 add_5" -> "305 add_6"  [label="(1, 56, 56, 96)", style=solid];
+"281 add_5_0_0_nncf_smooth_quant_0" -> "282 quantize_per_tensor_default_11"  [label="(1, 56, 56, 96)", style=solid];
+"282 quantize_per_tensor_default_11" -> "283 dequantize_per_tensor_default_11"  [label="(1, 56, 56, 96)", style=solid];
+"283 dequantize_per_tensor_default_11" -> "290 linear_10"  [label="(1, 56, 56, 96)", style=solid];
+"284 linear_10_scale_0" -> "287 quantize_per_channel_default_9"  [label="(384,)", style=solid];
+"284 linear_10_scale_0" -> "288 dequantize_per_channel_default_9"  [label="(384,)", style=solid];
+"285 linear_10_zero_point_0" -> "287 quantize_per_channel_default_9"  [label="(384,)", style=solid];
+"285 linear_10_zero_point_0" -> "288 dequantize_per_channel_default_9"  [label="(384,)", style=solid];
+"286 compressed_weight_updated_constant7" -> "287 quantize_per_channel_default_9"  [label="(384, 96)", style=solid];
+"287 quantize_per_channel_default_9" -> "288 dequantize_per_channel_default_9"  [label="(384, 96)", style=solid];
+"288 dequantize_per_channel_default_9" -> "290 linear_10"  [label="(384, 96)", style=solid];
+"289 features_1_1_mlp_0_bias_0_0" -> "290 linear_10"  [label="(384,)", style=solid];
+"290 linear_10" -> "291 gelu_1"  [label="(1, 56, 56, 384)", style=solid];
+"291 gelu_1" -> "292 dropout_6"  [label="(1, 56, 56, 384)", style=solid];
+"292 dropout_6" -> "293 dropout_6_0_0_nncf_smooth_quant_0"  [label="(1, 56, 56, 384)", style=solid];
+"293 dropout_6_0_0_nncf_smooth_quant_0" -> "294 quantize_per_tensor_default_12"  [label="(1, 56, 56, 384)", style=solid];
+"294 quantize_per_tensor_default_12" -> "295 dequantize_per_tensor_default_12"  [label="(1, 56, 56, 384)", style=solid];
+"295 dequantize_per_tensor_default_12" -> "302 linear_11"  [label="(1, 56, 56, 384)", style=solid];
+"296 linear_11_scale_0" -> "299 quantize_per_channel_default_10"  [label="(96,)", style=solid];
+"296 linear_11_scale_0" -> "300 dequantize_per_channel_default_10"  [label="(96,)", style=solid];
+"297 linear_11_zero_point_0" -> "299 quantize_per_channel_default_10"  [label="(96,)", style=solid];
+"297 linear_11_zero_point_0" -> "300 dequantize_per_channel_default_10"  [label="(96,)", style=solid];
+"298 compressed_weight_updated_constant8" -> "299 quantize_per_channel_default_10"  [label="(96, 384)", style=solid];
+"299 quantize_per_channel_default_10" -> "300 dequantize_per_channel_default_10"  [label="(96, 384)", style=solid];
+"300 dequantize_per_channel_default_10" -> "302 linear_11"  [label="(96, 384)", style=solid];
+"301 features_1_1_mlp_3_bias_0_0" -> "302 linear_11"  [label="(96,)", style=solid];
+"302 linear_11" -> "303 dropout_7"  [label="(1, 56, 56, 96)", style=solid];
+"303 dropout_7" -> "304 layer_norm_4"  [label="(1, 56, 56, 96)", style=solid];
+"304 layer_norm_4" -> "305 add_6"  [label="(1, 56, 56, 96)", style=solid];
+"305 add_6" -> "306 pad_2"  [label="(1, 56, 56, 96)", style=solid];
+"306 pad_2" -> "307 slice_25"  [label="(1, 56, 56, 96)", style=solid];
+"306 pad_2" -> "310 slice_28"  [label="(1, 56, 56, 96)", style=solid];
+"306 pad_2" -> "313 slice_31"  [label="(1, 56, 56, 96)", style=solid];
+"306 pad_2" -> "316 slice_34"  [label="(1, 56, 56, 96)", style=solid];
+"307 slice_25" -> "308 slice_26"  [label="(1, 28, 56, 96)", style=solid];
+"308 slice_26" -> "309 slice_27"  [label="(1, 28, 28, 96)", style=solid];
+"309 slice_27" -> "319 cat"  [label="(1, 28, 28, 96)", style=solid];
+"310 slice_28" -> "311 slice_29"  [label="(1, 28, 56, 96)", style=solid];
+"311 slice_29" -> "312 slice_30"  [label="(1, 28, 28, 96)", style=solid];
+"312 slice_30" -> "319 cat"  [label="(1, 28, 28, 96)", style=solid];
+"313 slice_31" -> "314 slice_32"  [label="(1, 28, 56, 96)", style=solid];
+"314 slice_32" -> "315 slice_33"  [label="(1, 28, 28, 96)", style=solid];
+"315 slice_33" -> "319 cat"  [label="(1, 28, 28, 96)", style=solid];
+"316 slice_34" -> "317 slice_35"  [label="(1, 28, 56, 96)", style=solid];
+"317 slice_35" -> "318 slice_36"  [label="(1, 28, 28, 96)", style=solid];
+"318 slice_36" -> "319 cat"  [label="(1, 28, 28, 96)", style=solid];
+"319 cat" -> "320 cat_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 384)", style=solid];
+"320 cat_0_0_nncf_smooth_quant_0" -> "321 quantize_per_tensor_default_13"  [label="(1, 28, 28, 384)", style=solid];
+"321 quantize_per_tensor_default_13" -> "322 dequantize_per_tensor_default_13"  [label="(1, 28, 28, 384)", style=solid];
+"322 dequantize_per_tensor_default_13" -> "328 linear_12"  [label="(1, 28, 28, 384)", style=solid];
+"323 linear_12_scale_0" -> "326 quantize_per_channel_default_11"  [label="(192,)", style=solid];
+"323 linear_12_scale_0" -> "327 dequantize_per_channel_default_11"  [label="(192,)", style=solid];
+"324 linear_12_zero_point_0" -> "326 quantize_per_channel_default_11"  [label="(192,)", style=solid];
+"324 linear_12_zero_point_0" -> "327 dequantize_per_channel_default_11"  [label="(192,)", style=solid];
+"325 compressed_weight_updated_constant9" -> "326 quantize_per_channel_default_11"  [label="(192, 384)", style=solid];
+"326 quantize_per_channel_default_11" -> "327 dequantize_per_channel_default_11"  [label="(192, 384)", style=solid];
+"327 dequantize_per_channel_default_11" -> "328 linear_12"  [label="(192, 384)", style=solid];
+"328 linear_12" -> "329 layer_norm_5"  [label="(1, 28, 28, 192)", style=solid];
+"329 layer_norm_5" -> "331 pad_3"  [label="(1, 28, 28, 192)", style=solid];
+"329 layer_norm_5" -> "392 add_8"  [label="(1, 28, 28, 192)", style=solid];
+"330 _frozen_param17" -> "366 add_7"  [label="(1, 6, 64, 64)", style=solid];
+"331 pad_3" -> "332 view_13"  [label="(1, 32, 32, 192)", style=solid];
+"332 view_13" -> "333 permute_11"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"333 permute_11" -> "334 reshape_9"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"334 reshape_9" -> "335 reshape_9_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
+"335 reshape_9_0_0_nncf_smooth_quant_0" -> "336 quantize_per_tensor_default_14"  [label="(16, 64, 192)", style=solid];
+"336 quantize_per_tensor_default_14" -> "337 dequantize_per_tensor_default_14"  [label="(16, 64, 192)", style=solid];
+"337 dequantize_per_tensor_default_14" -> "344 linear_15"  [label="(16, 64, 192)", style=solid];
+"338 linear_15_scale_0" -> "341 quantize_per_channel_default_13"  [label="(576,)", style=solid];
+"338 linear_15_scale_0" -> "342 dequantize_per_channel_default_13"  [label="(576,)", style=solid];
+"339 linear_15_zero_point_0" -> "341 quantize_per_channel_default_13"  [label="(576,)", style=solid];
+"339 linear_15_zero_point_0" -> "342 dequantize_per_channel_default_13"  [label="(576,)", style=solid];
+"340 compressed_weight_updated_constant10" -> "341 quantize_per_channel_default_13"  [label="(576, 192)", style=solid];
+"341 quantize_per_channel_default_13" -> "342 dequantize_per_channel_default_13"  [label="(576, 192)", style=solid];
+"342 dequantize_per_channel_default_13" -> "344 linear_15"  [label="(576, 192)", style=solid];
+"343 _frozen_param18_0_0" -> "344 linear_15"  [label="(576,)", style=solid];
+"344 linear_15" -> "345 reshape_10"  [label="(16, 64, 576)", style=solid];
+"345 reshape_10" -> "346 permute_12"  [label="(16, 64, 3, 6, 32)", style=solid];
+"346 permute_12" -> "347 select_6"  [label="(3, 16, 6, 64, 32)", style=solid];
+"346 permute_12" -> "348 select_7"  [label="(3, 16, 6, 64, 32)", style=solid];
+"346 permute_12" -> "349 select_8"  [label="(3, 16, 6, 64, 32)", style=solid];
+"347 select_6" -> "350 linalg_vector_norm_4"  [label="(16, 6, 64, 32)", style=solid];
+"347 select_6" -> "352 expand_as_4"  [label="(16, 6, 64, 32)", style=solid];
+"347 select_6" -> "353 div_4"  [label="(16, 6, 64, 32)", style=solid];
+"348 select_7" -> "356 linalg_vector_norm_5"  [label="(16, 6, 64, 32)", style=solid];
+"348 select_7" -> "358 expand_as_5"  [label="(16, 6, 64, 32)", style=solid];
+"348 select_7" -> "359 div_5"  [label="(16, 6, 64, 32)", style=solid];
+"349 select_8" -> "369 matmul_5"  [label="(16, 6, 64, 32)", style=solid];
+"350 linalg_vector_norm_4" -> "351 clamp_min_4"  [label="(16, 6, 64, 1)", style=solid];
+"351 clamp_min_4" -> "352 expand_as_4"  [label="(16, 6, 64, 1)", style=solid];
+"352 expand_as_4" -> "353 div_4"  [label="(16, 6, 64, 32)", style=solid];
+"353 div_4" -> "354 quantize_per_tensor_default_15"  [label="(16, 6, 64, 32)", style=solid];
+"354 quantize_per_tensor_default_15" -> "355 dequantize_per_tensor_default_15"  [label="(16, 6, 64, 32)", style=solid];
+"355 dequantize_per_tensor_default_15" -> "363 matmul_4"  [label="(16, 6, 64, 32)", style=solid];
+"356 linalg_vector_norm_5" -> "357 clamp_min_5"  [label="(16, 6, 64, 1)", style=solid];
+"357 clamp_min_5" -> "358 expand_as_5"  [label="(16, 6, 64, 1)", style=solid];
+"358 expand_as_5" -> "359 div_5"  [label="(16, 6, 64, 32)", style=solid];
+"359 div_5" -> "360 quantize_per_tensor_default_16"  [label="(16, 6, 64, 32)", style=solid];
+"360 quantize_per_tensor_default_16" -> "361 dequantize_per_tensor_default_16"  [label="(16, 6, 64, 32)", style=solid];
+"361 dequantize_per_tensor_default_16" -> "362 transpose_4"  [label="(16, 6, 64, 32)", style=solid];
+"362 transpose_4" -> "363 matmul_4"  [label="(16, 6, 32, 64)", style=solid];
+"363 matmul_4" -> "365 mul_5"  [label="(16, 6, 64, 64)", style=solid];
+"364 _frozen_param20" -> "365 mul_5"  [label="(6, 1, 1)", style=solid];
+"365 mul_5" -> "366 add_7"  [label="(16, 6, 64, 64)", style=solid];
+"366 add_7" -> "367 softmax_2"  [label="(16, 6, 64, 64)", style=solid];
+"367 softmax_2" -> "368 dropout_8"  [label="(16, 6, 64, 64)", style=solid];
+"368 dropout_8" -> "369 matmul_5"  [label="(16, 6, 64, 64)", style=solid];
+"369 matmul_5" -> "370 transpose_5"  [label="(16, 6, 64, 32)", style=solid];
+"370 transpose_5" -> "371 reshape_11"  [label="(16, 64, 6, 32)", style=solid];
+"371 reshape_11" -> "372 reshape_11_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
+"372 reshape_11_0_0_nncf_smooth_quant_0" -> "373 quantize_per_tensor_default_17"  [label="(16, 64, 192)", style=solid];
+"373 quantize_per_tensor_default_17" -> "374 dequantize_per_tensor_default_17"  [label="(16, 64, 192)", style=solid];
+"374 dequantize_per_tensor_default_17" -> "381 linear_16"  [label="(16, 64, 192)", style=solid];
+"375 linear_16_scale_0" -> "378 quantize_per_channel_default_14"  [label="(192,)", style=solid];
+"375 linear_16_scale_0" -> "379 dequantize_per_channel_default_14"  [label="(192,)", style=solid];
+"376 linear_16_zero_point_0" -> "378 quantize_per_channel_default_14"  [label="(192,)", style=solid];
+"376 linear_16_zero_point_0" -> "379 dequantize_per_channel_default_14"  [label="(192,)", style=solid];
+"377 compressed_weight_updated_constant11" -> "378 quantize_per_channel_default_14"  [label="(192, 192)", style=solid];
+"378 quantize_per_channel_default_14" -> "379 dequantize_per_channel_default_14"  [label="(192, 192)", style=solid];
+"379 dequantize_per_channel_default_14" -> "381 linear_16"  [label="(192, 192)", style=solid];
+"380 features_3_0_attn_proj_bias_0_0" -> "381 linear_16"  [label="(192,)", style=solid];
+"381 linear_16" -> "382 dropout_9"  [label="(16, 64, 192)", style=solid];
+"382 dropout_9" -> "383 view_14"  [label="(16, 64, 192)", style=solid];
+"383 view_14" -> "384 permute_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"384 permute_13" -> "385 reshape_12"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"385 reshape_12" -> "386 slice_38"  [label="(1, 32, 32, 192)", style=solid];
+"386 slice_38" -> "387 slice_39"  [label="(1, 32, 32, 192)", style=solid];
+"387 slice_39" -> "388 slice_40"  [label="(1, 28, 32, 192)", style=solid];
+"388 slice_40" -> "389 slice_41"  [label="(1, 28, 28, 192)", style=solid];
+"389 slice_41" -> "390 contiguous_3"  [label="(1, 28, 28, 192)", style=solid];
+"390 contiguous_3" -> "391 layer_norm_6"  [label="(1, 28, 28, 192)", style=solid];
+"391 layer_norm_6" -> "392 add_8"  [label="(1, 28, 28, 192)", style=solid];
+"392 add_8" -> "393 add_8_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 192)", style=solid];
+"392 add_8" -> "417 add_9"  [label="(1, 28, 28, 192)", style=solid];
+"393 add_8_0_0_nncf_smooth_quant_0" -> "394 quantize_per_tensor_default_18"  [label="(1, 28, 28, 192)", style=solid];
+"394 quantize_per_tensor_default_18" -> "395 dequantize_per_tensor_default_18"  [label="(1, 28, 28, 192)", style=solid];
+"395 dequantize_per_tensor_default_18" -> "402 linear_17"  [label="(1, 28, 28, 192)", style=solid];
+"396 linear_17_scale_0" -> "399 quantize_per_channel_default_15"  [label="(768,)", style=solid];
+"396 linear_17_scale_0" -> "400 dequantize_per_channel_default_15"  [label="(768,)", style=solid];
+"397 linear_17_zero_point_0" -> "399 quantize_per_channel_default_15"  [label="(768,)", style=solid];
+"397 linear_17_zero_point_0" -> "400 dequantize_per_channel_default_15"  [label="(768,)", style=solid];
+"398 compressed_weight_updated_constant12" -> "399 quantize_per_channel_default_15"  [label="(768, 192)", style=solid];
+"399 quantize_per_channel_default_15" -> "400 dequantize_per_channel_default_15"  [label="(768, 192)", style=solid];
+"400 dequantize_per_channel_default_15" -> "402 linear_17"  [label="(768, 192)", style=solid];
+"401 features_3_0_mlp_0_bias_0_0" -> "402 linear_17"  [label="(768,)", style=solid];
+"402 linear_17" -> "403 gelu_2"  [label="(1, 28, 28, 768)", style=solid];
+"403 gelu_2" -> "404 dropout_10"  [label="(1, 28, 28, 768)", style=solid];
+"404 dropout_10" -> "405 dropout_10_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 768)", style=solid];
+"405 dropout_10_0_0_nncf_smooth_quant_0" -> "406 quantize_per_tensor_default_19"  [label="(1, 28, 28, 768)", style=solid];
+"406 quantize_per_tensor_default_19" -> "407 dequantize_per_tensor_default_19"  [label="(1, 28, 28, 768)", style=solid];
+"407 dequantize_per_tensor_default_19" -> "414 linear_18"  [label="(1, 28, 28, 768)", style=solid];
+"408 linear_18_scale_0" -> "411 quantize_per_channel_default_16"  [label="(192,)", style=solid];
+"408 linear_18_scale_0" -> "412 dequantize_per_channel_default_16"  [label="(192,)", style=solid];
+"409 linear_18_zero_point_0" -> "411 quantize_per_channel_default_16"  [label="(192,)", style=solid];
+"409 linear_18_zero_point_0" -> "412 dequantize_per_channel_default_16"  [label="(192,)", style=solid];
+"410 compressed_weight_updated_constant13" -> "411 quantize_per_channel_default_16"  [label="(192, 768)", style=solid];
+"411 quantize_per_channel_default_16" -> "412 dequantize_per_channel_default_16"  [label="(192, 768)", style=solid];
+"412 dequantize_per_channel_default_16" -> "414 linear_18"  [label="(192, 768)", style=solid];
+"413 features_3_0_mlp_3_bias_0_0" -> "414 linear_18"  [label="(192,)", style=solid];
+"414 linear_18" -> "415 dropout_11"  [label="(1, 28, 28, 192)", style=solid];
+"415 dropout_11" -> "416 layer_norm_7"  [label="(1, 28, 28, 192)", style=solid];
+"416 layer_norm_7" -> "417 add_9"  [label="(1, 28, 28, 192)", style=solid];
+"417 add_9" -> "419 pad_4"  [label="(1, 28, 28, 192)", style=solid];
+"417 add_9" -> "498 add_12"  [label="(1, 28, 28, 192)", style=solid];
+"418 _frozen_param21" -> "455 add_10"  [label="(1, 6, 64, 64)", style=solid];
+"419 pad_4" -> "420 roll_2"  [label="(1, 32, 32, 192)", style=solid];
+"420 roll_2" -> "421 view_17"  [label="(1, 32, 32, 192)", style=solid];
+"421 view_17" -> "422 permute_15"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"422 permute_15" -> "423 reshape_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"423 reshape_13" -> "424 reshape_13_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
+"423 reshape_13" -> "456 new_zeros_1"  [label="(16, 64, 192)", style=solid];
+"424 reshape_13_0_0_nncf_smooth_quant_0" -> "425 quantize_per_tensor_default_20"  [label="(16, 64, 192)", style=solid];
+"425 quantize_per_tensor_default_20" -> "426 dequantize_per_tensor_default_20"  [label="(16, 64, 192)", style=solid];
+"426 dequantize_per_tensor_default_20" -> "433 linear_21"  [label="(16, 64, 192)", style=solid];
+"427 linear_21_scale_0" -> "430 quantize_per_channel_default_18"  [label="(576,)", style=solid];
+"427 linear_21_scale_0" -> "431 dequantize_per_channel_default_18"  [label="(576,)", style=solid];
+"428 linear_21_zero_point_0" -> "430 quantize_per_channel_default_18"  [label="(576,)", style=solid];
+"428 linear_21_zero_point_0" -> "431 dequantize_per_channel_default_18"  [label="(576,)", style=solid];
+"429 compressed_weight_updated_constant14" -> "430 quantize_per_channel_default_18"  [label="(576, 192)", style=solid];
+"430 quantize_per_channel_default_18" -> "431 dequantize_per_channel_default_18"  [label="(576, 192)", style=solid];
+"431 dequantize_per_channel_default_18" -> "433 linear_21"  [label="(576, 192)", style=solid];
+"432 _frozen_param22_0_0" -> "433 linear_21"  [label="(576,)", style=solid];
+"433 linear_21" -> "434 reshape_14"  [label="(16, 64, 576)", style=solid];
+"434 reshape_14" -> "435 permute_16"  [label="(16, 64, 3, 6, 32)", style=solid];
+"435 permute_16" -> "436 select_9"  [label="(3, 16, 6, 64, 32)", style=solid];
+"435 permute_16" -> "437 select_10"  [label="(3, 16, 6, 64, 32)", style=solid];
+"435 permute_16" -> "438 select_11"  [label="(3, 16, 6, 64, 32)", style=solid];
+"436 select_9" -> "439 linalg_vector_norm_6"  [label="(16, 6, 64, 32)", style=solid];
+"436 select_9" -> "441 expand_as_6"  [label="(16, 6, 64, 32)", style=solid];
+"436 select_9" -> "442 div_6"  [label="(16, 6, 64, 32)", style=solid];
+"437 select_10" -> "445 linalg_vector_norm_7"  [label="(16, 6, 64, 32)", style=solid];
+"437 select_10" -> "447 expand_as_7"  [label="(16, 6, 64, 32)", style=solid];
+"437 select_10" -> "448 div_7"  [label="(16, 6, 64, 32)", style=solid];
+"438 select_11" -> "474 matmul_7"  [label="(16, 6, 64, 32)", style=solid];
+"439 linalg_vector_norm_6" -> "440 clamp_min_6"  [label="(16, 6, 64, 1)", style=solid];
+"440 clamp_min_6" -> "441 expand_as_6"  [label="(16, 6, 64, 1)", style=solid];
+"441 expand_as_6" -> "442 div_6"  [label="(16, 6, 64, 32)", style=solid];
+"442 div_6" -> "443 quantize_per_tensor_default_21"  [label="(16, 6, 64, 32)", style=solid];
+"443 quantize_per_tensor_default_21" -> "444 dequantize_per_tensor_default_21"  [label="(16, 6, 64, 32)", style=solid];
+"444 dequantize_per_tensor_default_21" -> "452 matmul_6"  [label="(16, 6, 64, 32)", style=solid];
+"445 linalg_vector_norm_7" -> "446 clamp_min_7"  [label="(16, 6, 64, 1)", style=solid];
+"446 clamp_min_7" -> "447 expand_as_7"  [label="(16, 6, 64, 1)", style=solid];
+"447 expand_as_7" -> "448 div_7"  [label="(16, 6, 64, 32)", style=solid];
+"448 div_7" -> "449 quantize_per_tensor_default_22"  [label="(16, 6, 64, 32)", style=solid];
+"449 quantize_per_tensor_default_22" -> "450 dequantize_per_tensor_default_22"  [label="(16, 6, 64, 32)", style=solid];
+"450 dequantize_per_tensor_default_22" -> "451 transpose_6"  [label="(16, 6, 64, 32)", style=solid];
+"451 transpose_6" -> "452 matmul_6"  [label="(16, 6, 32, 64)", style=solid];
+"452 matmul_6" -> "454 mul_7"  [label="(16, 6, 64, 64)", style=solid];
+"453 _frozen_param24" -> "454 mul_7"  [label="(6, 1, 1)", style=solid];
+"454 mul_7" -> "455 add_10"  [label="(16, 6, 64, 64)", style=solid];
+"455 add_10" -> "467 view_19"  [label="(16, 6, 64, 64)", style=solid];
+"456 new_zeros_1" -> "457 view_18"  [label="(32, 32)", style=solid];
+"457 view_18" -> "458 permute_17"  [label="(4, 8, 4, 8)", style=solid];
+"458 permute_17" -> "459 reshape_15"  [label="(4, 4, 8, 8)", style=solid];
+"459 reshape_15" -> "460 unsqueeze_8"  [label="(16, 64)", style=solid];
+"459 reshape_15" -> "461 unsqueeze_9"  [label="(16, 64)", style=solid];
+"460 unsqueeze_8" -> "462 sub_1"  [label="(16, 1, 64)", style=solid];
+"461 unsqueeze_9" -> "462 sub_1"  [label="(16, 64, 1)", style=solid];
+"462 sub_1" -> "463 ne_1"  [label="(16, 64, 64)", style=solid];
+"462 sub_1" -> "464 masked_fill_2"  [label="(16, 64, 64)", style=solid];
+"462 sub_1" -> "465 eq_1"  [label="(16, 64, 64)", style=solid];
+"463 ne_1" -> "464 masked_fill_2"  [label="(16, 64, 64)", style=solid];
+"464 masked_fill_2" -> "466 masked_fill_3"  [label="(16, 64, 64)", style=solid];
+"465 eq_1" -> "466 masked_fill_3"  [label="(16, 64, 64)", style=solid];
+"466 masked_fill_3" -> "468 unsqueeze_10"  [label="(16, 64, 64)", style=solid];
+"467 view_19" -> "470 add_11"  [label="(1, 16, 6, 64, 64)", style=solid];
+"468 unsqueeze_10" -> "469 unsqueeze_11"  [label="(16, 1, 64, 64)", style=solid];
+"469 unsqueeze_11" -> "470 add_11"  [label="(1, 16, 1, 64, 64)", style=solid];
+"470 add_11" -> "471 view_20"  [label="(1, 16, 6, 64, 64)", style=solid];
+"471 view_20" -> "472 softmax_3"  [label="(16, 6, 64, 64)", style=solid];
+"472 softmax_3" -> "473 dropout_12"  [label="(16, 6, 64, 64)", style=solid];
+"473 dropout_12" -> "474 matmul_7"  [label="(16, 6, 64, 64)", style=solid];
+"474 matmul_7" -> "475 transpose_7"  [label="(16, 6, 64, 32)", style=solid];
+"475 transpose_7" -> "476 reshape_16"  [label="(16, 64, 6, 32)", style=solid];
+"476 reshape_16" -> "477 reshape_16_0_0_nncf_smooth_quant_0"  [label="(16, 64, 192)", style=solid];
+"477 reshape_16_0_0_nncf_smooth_quant_0" -> "478 quantize_per_tensor_default_23"  [label="(16, 64, 192)", style=solid];
+"478 quantize_per_tensor_default_23" -> "479 dequantize_per_tensor_default_23"  [label="(16, 64, 192)", style=solid];
+"479 dequantize_per_tensor_default_23" -> "486 linear_22"  [label="(16, 64, 192)", style=solid];
+"480 linear_22_scale_0" -> "483 quantize_per_channel_default_19"  [label="(192,)", style=solid];
+"480 linear_22_scale_0" -> "484 dequantize_per_channel_default_19"  [label="(192,)", style=solid];
+"481 linear_22_zero_point_0" -> "483 quantize_per_channel_default_19"  [label="(192,)", style=solid];
+"481 linear_22_zero_point_0" -> "484 dequantize_per_channel_default_19"  [label="(192,)", style=solid];
+"482 compressed_weight_updated_constant15" -> "483 quantize_per_channel_default_19"  [label="(192, 192)", style=solid];
+"483 quantize_per_channel_default_19" -> "484 dequantize_per_channel_default_19"  [label="(192, 192)", style=solid];
+"484 dequantize_per_channel_default_19" -> "486 linear_22"  [label="(192, 192)", style=solid];
+"485 features_3_1_attn_proj_bias_0_0" -> "486 linear_22"  [label="(192,)", style=solid];
+"486 linear_22" -> "487 dropout_13"  [label="(16, 64, 192)", style=solid];
+"487 dropout_13" -> "488 view_21"  [label="(16, 64, 192)", style=solid];
+"488 view_21" -> "489 permute_18"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"489 permute_18" -> "490 reshape_17"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"490 reshape_17" -> "491 roll_3"  [label="(1, 32, 32, 192)", style=solid];
+"491 roll_3" -> "492 slice_61"  [label="(1, 32, 32, 192)", style=solid];
+"492 slice_61" -> "493 slice_62"  [label="(1, 32, 32, 192)", style=solid];
+"493 slice_62" -> "494 slice_63"  [label="(1, 28, 32, 192)", style=solid];
+"494 slice_63" -> "495 slice_64"  [label="(1, 28, 28, 192)", style=solid];
+"495 slice_64" -> "496 contiguous_5"  [label="(1, 28, 28, 192)", style=solid];
+"496 contiguous_5" -> "497 layer_norm_8"  [label="(1, 28, 28, 192)", style=solid];
+"497 layer_norm_8" -> "498 add_12"  [label="(1, 28, 28, 192)", style=solid];
+"498 add_12" -> "499 add_12_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 192)", style=solid];
+"498 add_12" -> "523 add_13"  [label="(1, 28, 28, 192)", style=solid];
+"499 add_12_0_0_nncf_smooth_quant_0" -> "500 quantize_per_tensor_default_24"  [label="(1, 28, 28, 192)", style=solid];
+"500 quantize_per_tensor_default_24" -> "501 dequantize_per_tensor_default_24"  [label="(1, 28, 28, 192)", style=solid];
+"501 dequantize_per_tensor_default_24" -> "508 linear_23"  [label="(1, 28, 28, 192)", style=solid];
+"502 linear_23_scale_0" -> "505 quantize_per_channel_default_20"  [label="(768,)", style=solid];
+"502 linear_23_scale_0" -> "506 dequantize_per_channel_default_20"  [label="(768,)", style=solid];
+"503 linear_23_zero_point_0" -> "505 quantize_per_channel_default_20"  [label="(768,)", style=solid];
+"503 linear_23_zero_point_0" -> "506 dequantize_per_channel_default_20"  [label="(768,)", style=solid];
+"504 compressed_weight_updated_constant16" -> "505 quantize_per_channel_default_20"  [label="(768, 192)", style=solid];
+"505 quantize_per_channel_default_20" -> "506 dequantize_per_channel_default_20"  [label="(768, 192)", style=solid];
+"506 dequantize_per_channel_default_20" -> "508 linear_23"  [label="(768, 192)", style=solid];
+"507 features_3_1_mlp_0_bias_0_0" -> "508 linear_23"  [label="(768,)", style=solid];
+"508 linear_23" -> "509 gelu_3"  [label="(1, 28, 28, 768)", style=solid];
+"509 gelu_3" -> "510 dropout_14"  [label="(1, 28, 28, 768)", style=solid];
+"510 dropout_14" -> "511 dropout_14_0_0_nncf_smooth_quant_0"  [label="(1, 28, 28, 768)", style=solid];
+"511 dropout_14_0_0_nncf_smooth_quant_0" -> "512 quantize_per_tensor_default_25"  [label="(1, 28, 28, 768)", style=solid];
+"512 quantize_per_tensor_default_25" -> "513 dequantize_per_tensor_default_25"  [label="(1, 28, 28, 768)", style=solid];
+"513 dequantize_per_tensor_default_25" -> "520 linear_24"  [label="(1, 28, 28, 768)", style=solid];
+"514 linear_24_scale_0" -> "517 quantize_per_channel_default_21"  [label="(192,)", style=solid];
+"514 linear_24_scale_0" -> "518 dequantize_per_channel_default_21"  [label="(192,)", style=solid];
+"515 linear_24_zero_point_0" -> "517 quantize_per_channel_default_21"  [label="(192,)", style=solid];
+"515 linear_24_zero_point_0" -> "518 dequantize_per_channel_default_21"  [label="(192,)", style=solid];
+"516 compressed_weight_updated_constant17" -> "517 quantize_per_channel_default_21"  [label="(192, 768)", style=solid];
+"517 quantize_per_channel_default_21" -> "518 dequantize_per_channel_default_21"  [label="(192, 768)", style=solid];
+"518 dequantize_per_channel_default_21" -> "520 linear_24"  [label="(192, 768)", style=solid];
+"519 features_3_1_mlp_3_bias_0_0" -> "520 linear_24"  [label="(192,)", style=solid];
+"520 linear_24" -> "521 dropout_15"  [label="(1, 28, 28, 192)", style=solid];
+"521 dropout_15" -> "522 layer_norm_9"  [label="(1, 28, 28, 192)", style=solid];
+"522 layer_norm_9" -> "523 add_13"  [label="(1, 28, 28, 192)", style=solid];
+"523 add_13" -> "524 pad_5"  [label="(1, 28, 28, 192)", style=solid];
+"524 pad_5" -> "525 slice_65"  [label="(1, 28, 28, 192)", style=solid];
+"524 pad_5" -> "528 slice_68"  [label="(1, 28, 28, 192)", style=solid];
+"524 pad_5" -> "531 slice_71"  [label="(1, 28, 28, 192)", style=solid];
+"524 pad_5" -> "534 slice_74"  [label="(1, 28, 28, 192)", style=solid];
+"525 slice_65" -> "526 slice_66"  [label="(1, 14, 28, 192)", style=solid];
+"526 slice_66" -> "527 slice_67"  [label="(1, 14, 14, 192)", style=solid];
+"527 slice_67" -> "537 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"528 slice_68" -> "529 slice_69"  [label="(1, 14, 28, 192)", style=solid];
+"529 slice_69" -> "530 slice_70"  [label="(1, 14, 14, 192)", style=solid];
+"530 slice_70" -> "537 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"531 slice_71" -> "532 slice_72"  [label="(1, 14, 28, 192)", style=solid];
+"532 slice_72" -> "533 slice_73"  [label="(1, 14, 14, 192)", style=solid];
+"533 slice_73" -> "537 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"534 slice_74" -> "535 slice_75"  [label="(1, 14, 28, 192)", style=solid];
+"535 slice_75" -> "536 slice_76"  [label="(1, 14, 14, 192)", style=solid];
+"536 slice_76" -> "537 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"537 cat_1" -> "538 cat_1_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 768)", style=solid];
+"538 cat_1_0_0_nncf_smooth_quant_0" -> "539 quantize_per_tensor_default_26"  [label="(1, 14, 14, 768)", style=solid];
+"539 quantize_per_tensor_default_26" -> "540 dequantize_per_tensor_default_26"  [label="(1, 14, 14, 768)", style=solid];
+"540 dequantize_per_tensor_default_26" -> "546 linear_25"  [label="(1, 14, 14, 768)", style=solid];
+"541 linear_25_scale_0" -> "544 quantize_per_channel_default_22"  [label="(384,)", style=solid];
+"541 linear_25_scale_0" -> "545 dequantize_per_channel_default_22"  [label="(384,)", style=solid];
+"542 linear_25_zero_point_0" -> "544 quantize_per_channel_default_22"  [label="(384,)", style=solid];
+"542 linear_25_zero_point_0" -> "545 dequantize_per_channel_default_22"  [label="(384,)", style=solid];
+"543 compressed_weight_updated_constant18" -> "544 quantize_per_channel_default_22"  [label="(384, 768)", style=solid];
+"544 quantize_per_channel_default_22" -> "545 dequantize_per_channel_default_22"  [label="(384, 768)", style=solid];
+"545 dequantize_per_channel_default_22" -> "546 linear_25"  [label="(384, 768)", style=solid];
+"546 linear_25" -> "547 layer_norm_10"  [label="(1, 14, 14, 384)", style=solid];
+"547 layer_norm_10" -> "549 pad_6"  [label="(1, 14, 14, 384)", style=solid];
+"547 layer_norm_10" -> "610 add_15"  [label="(1, 14, 14, 384)", style=solid];
+"548 _frozen_param34" -> "584 add_14"  [label="(1, 12, 64, 64)", style=solid];
+"549 pad_6" -> "550 view_24"  [label="(1, 16, 16, 384)", style=solid];
+"550 view_24" -> "551 permute_20"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"551 permute_20" -> "552 reshape_18"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"552 reshape_18" -> "553 reshape_18_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"553 reshape_18_0_0_nncf_smooth_quant_0" -> "554 quantize_per_tensor_default_27"  [label="(4, 64, 384)", style=solid];
+"554 quantize_per_tensor_default_27" -> "555 dequantize_per_tensor_default_27"  [label="(4, 64, 384)", style=solid];
+"555 dequantize_per_tensor_default_27" -> "562 linear_28"  [label="(4, 64, 384)", style=solid];
+"556 linear_28_scale_0" -> "559 quantize_per_channel_default_24"  [label="(1152,)", style=solid];
+"556 linear_28_scale_0" -> "560 dequantize_per_channel_default_24"  [label="(1152,)", style=solid];
+"557 linear_28_zero_point_0" -> "559 quantize_per_channel_default_24"  [label="(1152,)", style=solid];
+"557 linear_28_zero_point_0" -> "560 dequantize_per_channel_default_24"  [label="(1152,)", style=solid];
+"558 compressed_weight_updated_constant19" -> "559 quantize_per_channel_default_24"  [label="(1152, 384)", style=solid];
+"559 quantize_per_channel_default_24" -> "560 dequantize_per_channel_default_24"  [label="(1152, 384)", style=solid];
+"560 dequantize_per_channel_default_24" -> "562 linear_28"  [label="(1152, 384)", style=solid];
+"561 _frozen_param35_0_0" -> "562 linear_28"  [label="(1152,)", style=solid];
+"562 linear_28" -> "563 reshape_19"  [label="(4, 64, 1152)", style=solid];
+"563 reshape_19" -> "564 permute_21"  [label="(4, 64, 3, 12, 32)", style=solid];
+"564 permute_21" -> "565 select_12"  [label="(3, 4, 12, 64, 32)", style=solid];
+"564 permute_21" -> "566 select_13"  [label="(3, 4, 12, 64, 32)", style=solid];
+"564 permute_21" -> "567 select_14"  [label="(3, 4, 12, 64, 32)", style=solid];
+"565 select_12" -> "568 linalg_vector_norm_8"  [label="(4, 12, 64, 32)", style=solid];
+"565 select_12" -> "570 expand_as_8"  [label="(4, 12, 64, 32)", style=solid];
+"565 select_12" -> "571 div_8"  [label="(4, 12, 64, 32)", style=solid];
+"566 select_13" -> "574 linalg_vector_norm_9"  [label="(4, 12, 64, 32)", style=solid];
+"566 select_13" -> "576 expand_as_9"  [label="(4, 12, 64, 32)", style=solid];
+"566 select_13" -> "577 div_9"  [label="(4, 12, 64, 32)", style=solid];
+"567 select_14" -> "587 matmul_9"  [label="(4, 12, 64, 32)", style=solid];
+"568 linalg_vector_norm_8" -> "569 clamp_min_8"  [label="(4, 12, 64, 1)", style=solid];
+"569 clamp_min_8" -> "570 expand_as_8"  [label="(4, 12, 64, 1)", style=solid];
+"570 expand_as_8" -> "571 div_8"  [label="(4, 12, 64, 32)", style=solid];
+"571 div_8" -> "572 quantize_per_tensor_default_28"  [label="(4, 12, 64, 32)", style=solid];
+"572 quantize_per_tensor_default_28" -> "573 dequantize_per_tensor_default_28"  [label="(4, 12, 64, 32)", style=solid];
+"573 dequantize_per_tensor_default_28" -> "581 matmul_8"  [label="(4, 12, 64, 32)", style=solid];
+"574 linalg_vector_norm_9" -> "575 clamp_min_9"  [label="(4, 12, 64, 1)", style=solid];
+"575 clamp_min_9" -> "576 expand_as_9"  [label="(4, 12, 64, 1)", style=solid];
+"576 expand_as_9" -> "577 div_9"  [label="(4, 12, 64, 32)", style=solid];
+"577 div_9" -> "578 quantize_per_tensor_default_29"  [label="(4, 12, 64, 32)", style=solid];
+"578 quantize_per_tensor_default_29" -> "579 dequantize_per_tensor_default_29"  [label="(4, 12, 64, 32)", style=solid];
+"579 dequantize_per_tensor_default_29" -> "580 transpose_8"  [label="(4, 12, 64, 32)", style=solid];
+"580 transpose_8" -> "581 matmul_8"  [label="(4, 12, 32, 64)", style=solid];
+"581 matmul_8" -> "583 mul_9"  [label="(4, 12, 64, 64)", style=solid];
+"582 _frozen_param37" -> "583 mul_9"  [label="(12, 1, 1)", style=solid];
+"583 mul_9" -> "584 add_14"  [label="(4, 12, 64, 64)", style=solid];
+"584 add_14" -> "585 softmax_4"  [label="(4, 12, 64, 64)", style=solid];
+"585 softmax_4" -> "586 dropout_16"  [label="(4, 12, 64, 64)", style=solid];
+"586 dropout_16" -> "587 matmul_9"  [label="(4, 12, 64, 64)", style=solid];
+"587 matmul_9" -> "588 transpose_9"  [label="(4, 12, 64, 32)", style=solid];
+"588 transpose_9" -> "589 reshape_20"  [label="(4, 64, 12, 32)", style=solid];
+"589 reshape_20" -> "590 reshape_20_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"590 reshape_20_0_0_nncf_smooth_quant_0" -> "591 quantize_per_tensor_default_30"  [label="(4, 64, 384)", style=solid];
+"591 quantize_per_tensor_default_30" -> "592 dequantize_per_tensor_default_30"  [label="(4, 64, 384)", style=solid];
+"592 dequantize_per_tensor_default_30" -> "599 linear_29"  [label="(4, 64, 384)", style=solid];
+"593 linear_29_scale_0" -> "596 quantize_per_channel_default_25"  [label="(384,)", style=solid];
+"593 linear_29_scale_0" -> "597 dequantize_per_channel_default_25"  [label="(384,)", style=solid];
+"594 linear_29_zero_point_0" -> "596 quantize_per_channel_default_25"  [label="(384,)", style=solid];
+"594 linear_29_zero_point_0" -> "597 dequantize_per_channel_default_25"  [label="(384,)", style=solid];
+"595 compressed_weight_updated_constant20" -> "596 quantize_per_channel_default_25"  [label="(384, 384)", style=solid];
+"596 quantize_per_channel_default_25" -> "597 dequantize_per_channel_default_25"  [label="(384, 384)", style=solid];
+"597 dequantize_per_channel_default_25" -> "599 linear_29"  [label="(384, 384)", style=solid];
+"598 features_5_0_attn_proj_bias_0_0" -> "599 linear_29"  [label="(384,)", style=solid];
+"599 linear_29" -> "600 dropout_17"  [label="(4, 64, 384)", style=solid];
+"600 dropout_17" -> "601 view_25"  [label="(4, 64, 384)", style=solid];
+"601 view_25" -> "602 permute_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"602 permute_22" -> "603 reshape_21"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"603 reshape_21" -> "604 slice_78"  [label="(1, 16, 16, 384)", style=solid];
+"604 slice_78" -> "605 slice_79"  [label="(1, 16, 16, 384)", style=solid];
+"605 slice_79" -> "606 slice_80"  [label="(1, 14, 16, 384)", style=solid];
+"606 slice_80" -> "607 slice_81"  [label="(1, 14, 14, 384)", style=solid];
+"607 slice_81" -> "608 contiguous_7"  [label="(1, 14, 14, 384)", style=solid];
+"608 contiguous_7" -> "609 layer_norm_11"  [label="(1, 14, 14, 384)", style=solid];
+"609 layer_norm_11" -> "610 add_15"  [label="(1, 14, 14, 384)", style=solid];
+"610 add_15" -> "611 add_15_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"610 add_15" -> "635 add_16"  [label="(1, 14, 14, 384)", style=solid];
+"611 add_15_0_0_nncf_smooth_quant_0" -> "612 quantize_per_tensor_default_31"  [label="(1, 14, 14, 384)", style=solid];
+"612 quantize_per_tensor_default_31" -> "613 dequantize_per_tensor_default_31"  [label="(1, 14, 14, 384)", style=solid];
+"613 dequantize_per_tensor_default_31" -> "620 linear_30"  [label="(1, 14, 14, 384)", style=solid];
+"614 linear_30_scale_0" -> "617 quantize_per_channel_default_26"  [label="(1536,)", style=solid];
+"614 linear_30_scale_0" -> "618 dequantize_per_channel_default_26"  [label="(1536,)", style=solid];
+"615 linear_30_zero_point_0" -> "617 quantize_per_channel_default_26"  [label="(1536,)", style=solid];
+"615 linear_30_zero_point_0" -> "618 dequantize_per_channel_default_26"  [label="(1536,)", style=solid];
+"616 compressed_weight_updated_constant21" -> "617 quantize_per_channel_default_26"  [label="(1536, 384)", style=solid];
+"617 quantize_per_channel_default_26" -> "618 dequantize_per_channel_default_26"  [label="(1536, 384)", style=solid];
+"618 dequantize_per_channel_default_26" -> "620 linear_30"  [label="(1536, 384)", style=solid];
+"619 features_5_0_mlp_0_bias_0_0" -> "620 linear_30"  [label="(1536,)", style=solid];
+"620 linear_30" -> "621 gelu_4"  [label="(1, 14, 14, 1536)", style=solid];
+"621 gelu_4" -> "622 dropout_18"  [label="(1, 14, 14, 1536)", style=solid];
+"622 dropout_18" -> "623 dropout_18_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"623 dropout_18_0_0_nncf_smooth_quant_0" -> "624 quantize_per_tensor_default_32"  [label="(1, 14, 14, 1536)", style=solid];
+"624 quantize_per_tensor_default_32" -> "625 dequantize_per_tensor_default_32"  [label="(1, 14, 14, 1536)", style=solid];
+"625 dequantize_per_tensor_default_32" -> "632 linear_31"  [label="(1, 14, 14, 1536)", style=solid];
+"626 linear_31_scale_0" -> "629 quantize_per_channel_default_27"  [label="(384,)", style=solid];
+"626 linear_31_scale_0" -> "630 dequantize_per_channel_default_27"  [label="(384,)", style=solid];
+"627 linear_31_zero_point_0" -> "629 quantize_per_channel_default_27"  [label="(384,)", style=solid];
+"627 linear_31_zero_point_0" -> "630 dequantize_per_channel_default_27"  [label="(384,)", style=solid];
+"628 compressed_weight_updated_constant22" -> "629 quantize_per_channel_default_27"  [label="(384, 1536)", style=solid];
+"629 quantize_per_channel_default_27" -> "630 dequantize_per_channel_default_27"  [label="(384, 1536)", style=solid];
+"630 dequantize_per_channel_default_27" -> "632 linear_31"  [label="(384, 1536)", style=solid];
+"631 features_5_0_mlp_3_bias_0_0" -> "632 linear_31"  [label="(384,)", style=solid];
+"632 linear_31" -> "633 dropout_19"  [label="(1, 14, 14, 384)", style=solid];
+"633 dropout_19" -> "634 layer_norm_12"  [label="(1, 14, 14, 384)", style=solid];
+"634 layer_norm_12" -> "635 add_16"  [label="(1, 14, 14, 384)", style=solid];
+"635 add_16" -> "637 pad_7"  [label="(1, 14, 14, 384)", style=solid];
+"635 add_16" -> "716 add_19"  [label="(1, 14, 14, 384)", style=solid];
+"636 _frozen_param38" -> "673 add_17"  [label="(1, 12, 64, 64)", style=solid];
+"637 pad_7" -> "638 roll_4"  [label="(1, 16, 16, 384)", style=solid];
+"638 roll_4" -> "639 view_28"  [label="(1, 16, 16, 384)", style=solid];
+"639 view_28" -> "640 permute_24"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"640 permute_24" -> "641 reshape_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"641 reshape_22" -> "642 reshape_22_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"641 reshape_22" -> "674 new_zeros_2"  [label="(4, 64, 384)", style=solid];
+"642 reshape_22_0_0_nncf_smooth_quant_0" -> "643 quantize_per_tensor_default_33"  [label="(4, 64, 384)", style=solid];
+"643 quantize_per_tensor_default_33" -> "644 dequantize_per_tensor_default_33"  [label="(4, 64, 384)", style=solid];
+"644 dequantize_per_tensor_default_33" -> "651 linear_34"  [label="(4, 64, 384)", style=solid];
+"645 linear_34_scale_0" -> "648 quantize_per_channel_default_29"  [label="(1152,)", style=solid];
+"645 linear_34_scale_0" -> "649 dequantize_per_channel_default_29"  [label="(1152,)", style=solid];
+"646 linear_34_zero_point_0" -> "648 quantize_per_channel_default_29"  [label="(1152,)", style=solid];
+"646 linear_34_zero_point_0" -> "649 dequantize_per_channel_default_29"  [label="(1152,)", style=solid];
+"647 compressed_weight_updated_constant23" -> "648 quantize_per_channel_default_29"  [label="(1152, 384)", style=solid];
+"648 quantize_per_channel_default_29" -> "649 dequantize_per_channel_default_29"  [label="(1152, 384)", style=solid];
+"649 dequantize_per_channel_default_29" -> "651 linear_34"  [label="(1152, 384)", style=solid];
+"650 _frozen_param39_0_0" -> "651 linear_34"  [label="(1152,)", style=solid];
+"651 linear_34" -> "652 reshape_23"  [label="(4, 64, 1152)", style=solid];
+"652 reshape_23" -> "653 permute_25"  [label="(4, 64, 3, 12, 32)", style=solid];
+"653 permute_25" -> "654 select_15"  [label="(3, 4, 12, 64, 32)", style=solid];
+"653 permute_25" -> "655 select_16"  [label="(3, 4, 12, 64, 32)", style=solid];
+"653 permute_25" -> "656 select_17"  [label="(3, 4, 12, 64, 32)", style=solid];
+"654 select_15" -> "657 linalg_vector_norm_10"  [label="(4, 12, 64, 32)", style=solid];
+"654 select_15" -> "659 expand_as_10"  [label="(4, 12, 64, 32)", style=solid];
+"654 select_15" -> "660 div_10"  [label="(4, 12, 64, 32)", style=solid];
+"655 select_16" -> "663 linalg_vector_norm_11"  [label="(4, 12, 64, 32)", style=solid];
+"655 select_16" -> "665 expand_as_11"  [label="(4, 12, 64, 32)", style=solid];
+"655 select_16" -> "666 div_11"  [label="(4, 12, 64, 32)", style=solid];
+"656 select_17" -> "692 matmul_11"  [label="(4, 12, 64, 32)", style=solid];
+"657 linalg_vector_norm_10" -> "658 clamp_min_10"  [label="(4, 12, 64, 1)", style=solid];
+"658 clamp_min_10" -> "659 expand_as_10"  [label="(4, 12, 64, 1)", style=solid];
+"659 expand_as_10" -> "660 div_10"  [label="(4, 12, 64, 32)", style=solid];
+"660 div_10" -> "661 quantize_per_tensor_default_34"  [label="(4, 12, 64, 32)", style=solid];
+"661 quantize_per_tensor_default_34" -> "662 dequantize_per_tensor_default_34"  [label="(4, 12, 64, 32)", style=solid];
+"662 dequantize_per_tensor_default_34" -> "670 matmul_10"  [label="(4, 12, 64, 32)", style=solid];
+"663 linalg_vector_norm_11" -> "664 clamp_min_11"  [label="(4, 12, 64, 1)", style=solid];
+"664 clamp_min_11" -> "665 expand_as_11"  [label="(4, 12, 64, 1)", style=solid];
+"665 expand_as_11" -> "666 div_11"  [label="(4, 12, 64, 32)", style=solid];
+"666 div_11" -> "667 quantize_per_tensor_default_35"  [label="(4, 12, 64, 32)", style=solid];
+"667 quantize_per_tensor_default_35" -> "668 dequantize_per_tensor_default_35"  [label="(4, 12, 64, 32)", style=solid];
+"668 dequantize_per_tensor_default_35" -> "669 transpose_10"  [label="(4, 12, 64, 32)", style=solid];
+"669 transpose_10" -> "670 matmul_10"  [label="(4, 12, 32, 64)", style=solid];
+"670 matmul_10" -> "672 mul_11"  [label="(4, 12, 64, 64)", style=solid];
+"671 _frozen_param41" -> "672 mul_11"  [label="(12, 1, 1)", style=solid];
+"672 mul_11" -> "673 add_17"  [label="(4, 12, 64, 64)", style=solid];
+"673 add_17" -> "685 view_30"  [label="(4, 12, 64, 64)", style=solid];
+"674 new_zeros_2" -> "675 view_29"  [label="(16, 16)", style=solid];
+"675 view_29" -> "676 permute_26"  [label="(2, 8, 2, 8)", style=solid];
+"676 permute_26" -> "677 reshape_24"  [label="(2, 2, 8, 8)", style=solid];
+"677 reshape_24" -> "678 unsqueeze_14"  [label="(4, 64)", style=solid];
+"677 reshape_24" -> "679 unsqueeze_15"  [label="(4, 64)", style=solid];
+"678 unsqueeze_14" -> "680 sub_2"  [label="(4, 1, 64)", style=solid];
+"679 unsqueeze_15" -> "680 sub_2"  [label="(4, 64, 1)", style=solid];
+"680 sub_2" -> "681 ne_2"  [label="(4, 64, 64)", style=solid];
+"680 sub_2" -> "682 masked_fill_4"  [label="(4, 64, 64)", style=solid];
+"680 sub_2" -> "683 eq_2"  [label="(4, 64, 64)", style=solid];
+"681 ne_2" -> "682 masked_fill_4"  [label="(4, 64, 64)", style=solid];
+"682 masked_fill_4" -> "684 masked_fill_5"  [label="(4, 64, 64)", style=solid];
+"683 eq_2" -> "684 masked_fill_5"  [label="(4, 64, 64)", style=solid];
+"684 masked_fill_5" -> "686 unsqueeze_16"  [label="(4, 64, 64)", style=solid];
+"685 view_30" -> "688 add_18"  [label="(1, 4, 12, 64, 64)", style=solid];
+"686 unsqueeze_16" -> "687 unsqueeze_17"  [label="(4, 1, 64, 64)", style=solid];
+"687 unsqueeze_17" -> "688 add_18"  [label="(1, 4, 1, 64, 64)", style=solid];
+"688 add_18" -> "689 view_31"  [label="(1, 4, 12, 64, 64)", style=solid];
+"689 view_31" -> "690 softmax_5"  [label="(4, 12, 64, 64)", style=solid];
+"690 softmax_5" -> "691 dropout_20"  [label="(4, 12, 64, 64)", style=solid];
+"691 dropout_20" -> "692 matmul_11"  [label="(4, 12, 64, 64)", style=solid];
+"692 matmul_11" -> "693 transpose_11"  [label="(4, 12, 64, 32)", style=solid];
+"693 transpose_11" -> "694 reshape_25"  [label="(4, 64, 12, 32)", style=solid];
+"694 reshape_25" -> "695 reshape_25_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"695 reshape_25_0_0_nncf_smooth_quant_0" -> "696 quantize_per_tensor_default_36"  [label="(4, 64, 384)", style=solid];
+"696 quantize_per_tensor_default_36" -> "697 dequantize_per_tensor_default_36"  [label="(4, 64, 384)", style=solid];
+"697 dequantize_per_tensor_default_36" -> "704 linear_35"  [label="(4, 64, 384)", style=solid];
+"698 linear_35_scale_0" -> "701 quantize_per_channel_default_30"  [label="(384,)", style=solid];
+"698 linear_35_scale_0" -> "702 dequantize_per_channel_default_30"  [label="(384,)", style=solid];
+"699 linear_35_zero_point_0" -> "701 quantize_per_channel_default_30"  [label="(384,)", style=solid];
+"699 linear_35_zero_point_0" -> "702 dequantize_per_channel_default_30"  [label="(384,)", style=solid];
+"700 compressed_weight_updated_constant24" -> "701 quantize_per_channel_default_30"  [label="(384, 384)", style=solid];
+"701 quantize_per_channel_default_30" -> "702 dequantize_per_channel_default_30"  [label="(384, 384)", style=solid];
+"702 dequantize_per_channel_default_30" -> "704 linear_35"  [label="(384, 384)", style=solid];
+"703 features_5_1_attn_proj_bias_0_0" -> "704 linear_35"  [label="(384,)", style=solid];
+"704 linear_35" -> "705 dropout_21"  [label="(4, 64, 384)", style=solid];
+"705 dropout_21" -> "706 view_32"  [label="(4, 64, 384)", style=solid];
+"706 view_32" -> "707 permute_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"707 permute_27" -> "708 reshape_26"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"708 reshape_26" -> "709 roll_5"  [label="(1, 16, 16, 384)", style=solid];
+"709 roll_5" -> "710 slice_101"  [label="(1, 16, 16, 384)", style=solid];
+"710 slice_101" -> "711 slice_102"  [label="(1, 16, 16, 384)", style=solid];
+"711 slice_102" -> "712 slice_103"  [label="(1, 14, 16, 384)", style=solid];
+"712 slice_103" -> "713 slice_104"  [label="(1, 14, 14, 384)", style=solid];
+"713 slice_104" -> "714 contiguous_9"  [label="(1, 14, 14, 384)", style=solid];
+"714 contiguous_9" -> "715 layer_norm_13"  [label="(1, 14, 14, 384)", style=solid];
+"715 layer_norm_13" -> "716 add_19"  [label="(1, 14, 14, 384)", style=solid];
+"716 add_19" -> "717 add_19_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"716 add_19" -> "741 add_20"  [label="(1, 14, 14, 384)", style=solid];
+"717 add_19_0_0_nncf_smooth_quant_0" -> "718 quantize_per_tensor_default_37"  [label="(1, 14, 14, 384)", style=solid];
+"718 quantize_per_tensor_default_37" -> "719 dequantize_per_tensor_default_37"  [label="(1, 14, 14, 384)", style=solid];
+"719 dequantize_per_tensor_default_37" -> "726 linear_36"  [label="(1, 14, 14, 384)", style=solid];
+"720 linear_36_scale_0" -> "723 quantize_per_channel_default_31"  [label="(1536,)", style=solid];
+"720 linear_36_scale_0" -> "724 dequantize_per_channel_default_31"  [label="(1536,)", style=solid];
+"721 linear_36_zero_point_0" -> "723 quantize_per_channel_default_31"  [label="(1536,)", style=solid];
+"721 linear_36_zero_point_0" -> "724 dequantize_per_channel_default_31"  [label="(1536,)", style=solid];
+"722 compressed_weight_updated_constant25" -> "723 quantize_per_channel_default_31"  [label="(1536, 384)", style=solid];
+"723 quantize_per_channel_default_31" -> "724 dequantize_per_channel_default_31"  [label="(1536, 384)", style=solid];
+"724 dequantize_per_channel_default_31" -> "726 linear_36"  [label="(1536, 384)", style=solid];
+"725 features_5_1_mlp_0_bias_0_0" -> "726 linear_36"  [label="(1536,)", style=solid];
+"726 linear_36" -> "727 gelu_5"  [label="(1, 14, 14, 1536)", style=solid];
+"727 gelu_5" -> "728 dropout_22"  [label="(1, 14, 14, 1536)", style=solid];
+"728 dropout_22" -> "729 dropout_22_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"729 dropout_22_0_0_nncf_smooth_quant_0" -> "730 quantize_per_tensor_default_38"  [label="(1, 14, 14, 1536)", style=solid];
+"730 quantize_per_tensor_default_38" -> "731 dequantize_per_tensor_default_38"  [label="(1, 14, 14, 1536)", style=solid];
+"731 dequantize_per_tensor_default_38" -> "738 linear_37"  [label="(1, 14, 14, 1536)", style=solid];
+"732 linear_37_scale_0" -> "735 quantize_per_channel_default_32"  [label="(384,)", style=solid];
+"732 linear_37_scale_0" -> "736 dequantize_per_channel_default_32"  [label="(384,)", style=solid];
+"733 linear_37_zero_point_0" -> "735 quantize_per_channel_default_32"  [label="(384,)", style=solid];
+"733 linear_37_zero_point_0" -> "736 dequantize_per_channel_default_32"  [label="(384,)", style=solid];
+"734 compressed_weight_updated_constant26" -> "735 quantize_per_channel_default_32"  [label="(384, 1536)", style=solid];
+"735 quantize_per_channel_default_32" -> "736 dequantize_per_channel_default_32"  [label="(384, 1536)", style=solid];
+"736 dequantize_per_channel_default_32" -> "738 linear_37"  [label="(384, 1536)", style=solid];
+"737 features_5_1_mlp_3_bias_0_0" -> "738 linear_37"  [label="(384,)", style=solid];
+"738 linear_37" -> "739 dropout_23"  [label="(1, 14, 14, 384)", style=solid];
+"739 dropout_23" -> "740 layer_norm_14"  [label="(1, 14, 14, 384)", style=solid];
+"740 layer_norm_14" -> "741 add_20"  [label="(1, 14, 14, 384)", style=solid];
+"741 add_20" -> "743 pad_8"  [label="(1, 14, 14, 384)", style=solid];
+"741 add_20" -> "804 add_22"  [label="(1, 14, 14, 384)", style=solid];
+"742 _frozen_param51" -> "778 add_21"  [label="(1, 12, 64, 64)", style=solid];
+"743 pad_8" -> "744 view_35"  [label="(1, 16, 16, 384)", style=solid];
+"744 view_35" -> "745 permute_29"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"745 permute_29" -> "746 reshape_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"746 reshape_27" -> "747 reshape_27_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"747 reshape_27_0_0_nncf_smooth_quant_0" -> "748 quantize_per_tensor_default_39"  [label="(4, 64, 384)", style=solid];
+"748 quantize_per_tensor_default_39" -> "749 dequantize_per_tensor_default_39"  [label="(4, 64, 384)", style=solid];
+"749 dequantize_per_tensor_default_39" -> "756 linear_40"  [label="(4, 64, 384)", style=solid];
+"750 linear_40_scale_0" -> "753 quantize_per_channel_default_34"  [label="(1152,)", style=solid];
+"750 linear_40_scale_0" -> "754 dequantize_per_channel_default_34"  [label="(1152,)", style=solid];
+"751 linear_40_zero_point_0" -> "753 quantize_per_channel_default_34"  [label="(1152,)", style=solid];
+"751 linear_40_zero_point_0" -> "754 dequantize_per_channel_default_34"  [label="(1152,)", style=solid];
+"752 compressed_weight_updated_constant27" -> "753 quantize_per_channel_default_34"  [label="(1152, 384)", style=solid];
+"753 quantize_per_channel_default_34" -> "754 dequantize_per_channel_default_34"  [label="(1152, 384)", style=solid];
+"754 dequantize_per_channel_default_34" -> "756 linear_40"  [label="(1152, 384)", style=solid];
+"755 _frozen_param52_0_0" -> "756 linear_40"  [label="(1152,)", style=solid];
+"756 linear_40" -> "757 reshape_28"  [label="(4, 64, 1152)", style=solid];
+"757 reshape_28" -> "758 permute_30"  [label="(4, 64, 3, 12, 32)", style=solid];
+"758 permute_30" -> "759 select_18"  [label="(3, 4, 12, 64, 32)", style=solid];
+"758 permute_30" -> "760 select_19"  [label="(3, 4, 12, 64, 32)", style=solid];
+"758 permute_30" -> "761 select_20"  [label="(3, 4, 12, 64, 32)", style=solid];
+"759 select_18" -> "762 linalg_vector_norm_12"  [label="(4, 12, 64, 32)", style=solid];
+"759 select_18" -> "764 expand_as_12"  [label="(4, 12, 64, 32)", style=solid];
+"759 select_18" -> "765 div_12"  [label="(4, 12, 64, 32)", style=solid];
+"760 select_19" -> "768 linalg_vector_norm_13"  [label="(4, 12, 64, 32)", style=solid];
+"760 select_19" -> "770 expand_as_13"  [label="(4, 12, 64, 32)", style=solid];
+"760 select_19" -> "771 div_13"  [label="(4, 12, 64, 32)", style=solid];
+"761 select_20" -> "781 matmul_13"  [label="(4, 12, 64, 32)", style=solid];
+"762 linalg_vector_norm_12" -> "763 clamp_min_12"  [label="(4, 12, 64, 1)", style=solid];
+"763 clamp_min_12" -> "764 expand_as_12"  [label="(4, 12, 64, 1)", style=solid];
+"764 expand_as_12" -> "765 div_12"  [label="(4, 12, 64, 32)", style=solid];
+"765 div_12" -> "766 quantize_per_tensor_default_40"  [label="(4, 12, 64, 32)", style=solid];
+"766 quantize_per_tensor_default_40" -> "767 dequantize_per_tensor_default_40"  [label="(4, 12, 64, 32)", style=solid];
+"767 dequantize_per_tensor_default_40" -> "775 matmul_12"  [label="(4, 12, 64, 32)", style=solid];
+"768 linalg_vector_norm_13" -> "769 clamp_min_13"  [label="(4, 12, 64, 1)", style=solid];
+"769 clamp_min_13" -> "770 expand_as_13"  [label="(4, 12, 64, 1)", style=solid];
+"770 expand_as_13" -> "771 div_13"  [label="(4, 12, 64, 32)", style=solid];
+"771 div_13" -> "772 quantize_per_tensor_default_41"  [label="(4, 12, 64, 32)", style=solid];
+"772 quantize_per_tensor_default_41" -> "773 dequantize_per_tensor_default_41"  [label="(4, 12, 64, 32)", style=solid];
+"773 dequantize_per_tensor_default_41" -> "774 transpose_12"  [label="(4, 12, 64, 32)", style=solid];
+"774 transpose_12" -> "775 matmul_12"  [label="(4, 12, 32, 64)", style=solid];
+"775 matmul_12" -> "777 mul_13"  [label="(4, 12, 64, 64)", style=solid];
+"776 _frozen_param54" -> "777 mul_13"  [label="(12, 1, 1)", style=solid];
+"777 mul_13" -> "778 add_21"  [label="(4, 12, 64, 64)", style=solid];
+"778 add_21" -> "779 softmax_6"  [label="(4, 12, 64, 64)", style=solid];
+"779 softmax_6" -> "780 dropout_24"  [label="(4, 12, 64, 64)", style=solid];
+"780 dropout_24" -> "781 matmul_13"  [label="(4, 12, 64, 64)", style=solid];
+"781 matmul_13" -> "782 transpose_13"  [label="(4, 12, 64, 32)", style=solid];
+"782 transpose_13" -> "783 reshape_29"  [label="(4, 64, 12, 32)", style=solid];
+"783 reshape_29" -> "784 reshape_29_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"784 reshape_29_0_0_nncf_smooth_quant_0" -> "785 quantize_per_tensor_default_42"  [label="(4, 64, 384)", style=solid];
+"785 quantize_per_tensor_default_42" -> "786 dequantize_per_tensor_default_42"  [label="(4, 64, 384)", style=solid];
+"786 dequantize_per_tensor_default_42" -> "793 linear_41"  [label="(4, 64, 384)", style=solid];
+"787 linear_41_scale_0" -> "790 quantize_per_channel_default_35"  [label="(384,)", style=solid];
+"787 linear_41_scale_0" -> "791 dequantize_per_channel_default_35"  [label="(384,)", style=solid];
+"788 linear_41_zero_point_0" -> "790 quantize_per_channel_default_35"  [label="(384,)", style=solid];
+"788 linear_41_zero_point_0" -> "791 dequantize_per_channel_default_35"  [label="(384,)", style=solid];
+"789 compressed_weight_updated_constant28" -> "790 quantize_per_channel_default_35"  [label="(384, 384)", style=solid];
+"790 quantize_per_channel_default_35" -> "791 dequantize_per_channel_default_35"  [label="(384, 384)", style=solid];
+"791 dequantize_per_channel_default_35" -> "793 linear_41"  [label="(384, 384)", style=solid];
+"792 features_5_2_attn_proj_bias_0_0" -> "793 linear_41"  [label="(384,)", style=solid];
+"793 linear_41" -> "794 dropout_25"  [label="(4, 64, 384)", style=solid];
+"794 dropout_25" -> "795 view_36"  [label="(4, 64, 384)", style=solid];
+"795 view_36" -> "796 permute_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"796 permute_31" -> "797 reshape_30"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"797 reshape_30" -> "798 slice_106"  [label="(1, 16, 16, 384)", style=solid];
+"798 slice_106" -> "799 slice_107"  [label="(1, 16, 16, 384)", style=solid];
+"799 slice_107" -> "800 slice_108"  [label="(1, 14, 16, 384)", style=solid];
+"800 slice_108" -> "801 slice_109"  [label="(1, 14, 14, 384)", style=solid];
+"801 slice_109" -> "802 contiguous_11"  [label="(1, 14, 14, 384)", style=solid];
+"802 contiguous_11" -> "803 layer_norm_15"  [label="(1, 14, 14, 384)", style=solid];
+"803 layer_norm_15" -> "804 add_22"  [label="(1, 14, 14, 384)", style=solid];
+"804 add_22" -> "805 add_22_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"804 add_22" -> "829 add_23"  [label="(1, 14, 14, 384)", style=solid];
+"805 add_22_0_0_nncf_smooth_quant_0" -> "806 quantize_per_tensor_default_43"  [label="(1, 14, 14, 384)", style=solid];
+"806 quantize_per_tensor_default_43" -> "807 dequantize_per_tensor_default_43"  [label="(1, 14, 14, 384)", style=solid];
+"807 dequantize_per_tensor_default_43" -> "814 linear_42"  [label="(1, 14, 14, 384)", style=solid];
+"808 linear_42_scale_0" -> "811 quantize_per_channel_default_36"  [label="(1536,)", style=solid];
+"808 linear_42_scale_0" -> "812 dequantize_per_channel_default_36"  [label="(1536,)", style=solid];
+"809 linear_42_zero_point_0" -> "811 quantize_per_channel_default_36"  [label="(1536,)", style=solid];
+"809 linear_42_zero_point_0" -> "812 dequantize_per_channel_default_36"  [label="(1536,)", style=solid];
+"810 compressed_weight_updated_constant29" -> "811 quantize_per_channel_default_36"  [label="(1536, 384)", style=solid];
+"811 quantize_per_channel_default_36" -> "812 dequantize_per_channel_default_36"  [label="(1536, 384)", style=solid];
+"812 dequantize_per_channel_default_36" -> "814 linear_42"  [label="(1536, 384)", style=solid];
+"813 features_5_2_mlp_0_bias_0_0" -> "814 linear_42"  [label="(1536,)", style=solid];
+"814 linear_42" -> "815 gelu_6"  [label="(1, 14, 14, 1536)", style=solid];
+"815 gelu_6" -> "816 dropout_26"  [label="(1, 14, 14, 1536)", style=solid];
+"816 dropout_26" -> "817 dropout_26_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"817 dropout_26_0_0_nncf_smooth_quant_0" -> "818 quantize_per_tensor_default_44"  [label="(1, 14, 14, 1536)", style=solid];
+"818 quantize_per_tensor_default_44" -> "819 dequantize_per_tensor_default_44"  [label="(1, 14, 14, 1536)", style=solid];
+"819 dequantize_per_tensor_default_44" -> "826 linear_43"  [label="(1, 14, 14, 1536)", style=solid];
+"820 linear_43_scale_0" -> "823 quantize_per_channel_default_37"  [label="(384,)", style=solid];
+"820 linear_43_scale_0" -> "824 dequantize_per_channel_default_37"  [label="(384,)", style=solid];
+"821 linear_43_zero_point_0" -> "823 quantize_per_channel_default_37"  [label="(384,)", style=solid];
+"821 linear_43_zero_point_0" -> "824 dequantize_per_channel_default_37"  [label="(384,)", style=solid];
+"822 compressed_weight_updated_constant30" -> "823 quantize_per_channel_default_37"  [label="(384, 1536)", style=solid];
+"823 quantize_per_channel_default_37" -> "824 dequantize_per_channel_default_37"  [label="(384, 1536)", style=solid];
+"824 dequantize_per_channel_default_37" -> "826 linear_43"  [label="(384, 1536)", style=solid];
+"825 features_5_2_mlp_3_bias_0_0" -> "826 linear_43"  [label="(384,)", style=solid];
+"826 linear_43" -> "827 dropout_27"  [label="(1, 14, 14, 384)", style=solid];
+"827 dropout_27" -> "828 layer_norm_16"  [label="(1, 14, 14, 384)", style=solid];
+"828 layer_norm_16" -> "829 add_23"  [label="(1, 14, 14, 384)", style=solid];
+"829 add_23" -> "831 pad_9"  [label="(1, 14, 14, 384)", style=solid];
+"829 add_23" -> "910 add_26"  [label="(1, 14, 14, 384)", style=solid];
+"830 _frozen_param55" -> "867 add_24"  [label="(1, 12, 64, 64)", style=solid];
+"831 pad_9" -> "832 roll_6"  [label="(1, 16, 16, 384)", style=solid];
+"832 roll_6" -> "833 view_39"  [label="(1, 16, 16, 384)", style=solid];
+"833 view_39" -> "834 permute_33"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"834 permute_33" -> "835 reshape_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"835 reshape_31" -> "836 reshape_31_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"835 reshape_31" -> "868 new_zeros_3"  [label="(4, 64, 384)", style=solid];
+"836 reshape_31_0_0_nncf_smooth_quant_0" -> "837 quantize_per_tensor_default_45"  [label="(4, 64, 384)", style=solid];
+"837 quantize_per_tensor_default_45" -> "838 dequantize_per_tensor_default_45"  [label="(4, 64, 384)", style=solid];
+"838 dequantize_per_tensor_default_45" -> "845 linear_46"  [label="(4, 64, 384)", style=solid];
+"839 linear_46_scale_0" -> "842 quantize_per_channel_default_39"  [label="(1152,)", style=solid];
+"839 linear_46_scale_0" -> "843 dequantize_per_channel_default_39"  [label="(1152,)", style=solid];
+"840 linear_46_zero_point_0" -> "842 quantize_per_channel_default_39"  [label="(1152,)", style=solid];
+"840 linear_46_zero_point_0" -> "843 dequantize_per_channel_default_39"  [label="(1152,)", style=solid];
+"841 compressed_weight_updated_constant31" -> "842 quantize_per_channel_default_39"  [label="(1152, 384)", style=solid];
+"842 quantize_per_channel_default_39" -> "843 dequantize_per_channel_default_39"  [label="(1152, 384)", style=solid];
+"843 dequantize_per_channel_default_39" -> "845 linear_46"  [label="(1152, 384)", style=solid];
+"844 _frozen_param56_0_0" -> "845 linear_46"  [label="(1152,)", style=solid];
+"845 linear_46" -> "846 reshape_32"  [label="(4, 64, 1152)", style=solid];
+"846 reshape_32" -> "847 permute_34"  [label="(4, 64, 3, 12, 32)", style=solid];
+"847 permute_34" -> "848 select_21"  [label="(3, 4, 12, 64, 32)", style=solid];
+"847 permute_34" -> "849 select_22"  [label="(3, 4, 12, 64, 32)", style=solid];
+"847 permute_34" -> "850 select_23"  [label="(3, 4, 12, 64, 32)", style=solid];
+"848 select_21" -> "851 linalg_vector_norm_14"  [label="(4, 12, 64, 32)", style=solid];
+"848 select_21" -> "853 expand_as_14"  [label="(4, 12, 64, 32)", style=solid];
+"848 select_21" -> "854 div_14"  [label="(4, 12, 64, 32)", style=solid];
+"849 select_22" -> "857 linalg_vector_norm_15"  [label="(4, 12, 64, 32)", style=solid];
+"849 select_22" -> "859 expand_as_15"  [label="(4, 12, 64, 32)", style=solid];
+"849 select_22" -> "860 div_15"  [label="(4, 12, 64, 32)", style=solid];
+"850 select_23" -> "886 matmul_15"  [label="(4, 12, 64, 32)", style=solid];
+"851 linalg_vector_norm_14" -> "852 clamp_min_14"  [label="(4, 12, 64, 1)", style=solid];
+"852 clamp_min_14" -> "853 expand_as_14"  [label="(4, 12, 64, 1)", style=solid];
+"853 expand_as_14" -> "854 div_14"  [label="(4, 12, 64, 32)", style=solid];
+"854 div_14" -> "855 quantize_per_tensor_default_46"  [label="(4, 12, 64, 32)", style=solid];
+"855 quantize_per_tensor_default_46" -> "856 dequantize_per_tensor_default_46"  [label="(4, 12, 64, 32)", style=solid];
+"856 dequantize_per_tensor_default_46" -> "864 matmul_14"  [label="(4, 12, 64, 32)", style=solid];
+"857 linalg_vector_norm_15" -> "858 clamp_min_15"  [label="(4, 12, 64, 1)", style=solid];
+"858 clamp_min_15" -> "859 expand_as_15"  [label="(4, 12, 64, 1)", style=solid];
+"859 expand_as_15" -> "860 div_15"  [label="(4, 12, 64, 32)", style=solid];
+"860 div_15" -> "861 quantize_per_tensor_default_47"  [label="(4, 12, 64, 32)", style=solid];
+"861 quantize_per_tensor_default_47" -> "862 dequantize_per_tensor_default_47"  [label="(4, 12, 64, 32)", style=solid];
+"862 dequantize_per_tensor_default_47" -> "863 transpose_14"  [label="(4, 12, 64, 32)", style=solid];
+"863 transpose_14" -> "864 matmul_14"  [label="(4, 12, 32, 64)", style=solid];
+"864 matmul_14" -> "866 mul_15"  [label="(4, 12, 64, 64)", style=solid];
+"865 _frozen_param58" -> "866 mul_15"  [label="(12, 1, 1)", style=solid];
+"866 mul_15" -> "867 add_24"  [label="(4, 12, 64, 64)", style=solid];
+"867 add_24" -> "879 view_41"  [label="(4, 12, 64, 64)", style=solid];
+"868 new_zeros_3" -> "869 view_40"  [label="(16, 16)", style=solid];
+"869 view_40" -> "870 permute_35"  [label="(2, 8, 2, 8)", style=solid];
+"870 permute_35" -> "871 reshape_33"  [label="(2, 2, 8, 8)", style=solid];
+"871 reshape_33" -> "872 unsqueeze_20"  [label="(4, 64)", style=solid];
+"871 reshape_33" -> "873 unsqueeze_21"  [label="(4, 64)", style=solid];
+"872 unsqueeze_20" -> "874 sub_3"  [label="(4, 1, 64)", style=solid];
+"873 unsqueeze_21" -> "874 sub_3"  [label="(4, 64, 1)", style=solid];
+"874 sub_3" -> "875 ne_3"  [label="(4, 64, 64)", style=solid];
+"874 sub_3" -> "876 masked_fill_6"  [label="(4, 64, 64)", style=solid];
+"874 sub_3" -> "877 eq_3"  [label="(4, 64, 64)", style=solid];
+"875 ne_3" -> "876 masked_fill_6"  [label="(4, 64, 64)", style=solid];
+"876 masked_fill_6" -> "878 masked_fill_7"  [label="(4, 64, 64)", style=solid];
+"877 eq_3" -> "878 masked_fill_7"  [label="(4, 64, 64)", style=solid];
+"878 masked_fill_7" -> "880 unsqueeze_22"  [label="(4, 64, 64)", style=solid];
+"879 view_41" -> "882 add_25"  [label="(1, 4, 12, 64, 64)", style=solid];
+"880 unsqueeze_22" -> "881 unsqueeze_23"  [label="(4, 1, 64, 64)", style=solid];
+"881 unsqueeze_23" -> "882 add_25"  [label="(1, 4, 1, 64, 64)", style=solid];
+"882 add_25" -> "883 view_42"  [label="(1, 4, 12, 64, 64)", style=solid];
+"883 view_42" -> "884 softmax_7"  [label="(4, 12, 64, 64)", style=solid];
+"884 softmax_7" -> "885 dropout_28"  [label="(4, 12, 64, 64)", style=solid];
+"885 dropout_28" -> "886 matmul_15"  [label="(4, 12, 64, 64)", style=solid];
+"886 matmul_15" -> "887 transpose_15"  [label="(4, 12, 64, 32)", style=solid];
+"887 transpose_15" -> "888 reshape_34"  [label="(4, 64, 12, 32)", style=solid];
+"888 reshape_34" -> "889 reshape_34_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"889 reshape_34_0_0_nncf_smooth_quant_0" -> "890 quantize_per_tensor_default_48"  [label="(4, 64, 384)", style=solid];
+"890 quantize_per_tensor_default_48" -> "891 dequantize_per_tensor_default_48"  [label="(4, 64, 384)", style=solid];
+"891 dequantize_per_tensor_default_48" -> "898 linear_47"  [label="(4, 64, 384)", style=solid];
+"892 linear_47_scale_0" -> "895 quantize_per_channel_default_40"  [label="(384,)", style=solid];
+"892 linear_47_scale_0" -> "896 dequantize_per_channel_default_40"  [label="(384,)", style=solid];
+"893 linear_47_zero_point_0" -> "895 quantize_per_channel_default_40"  [label="(384,)", style=solid];
+"893 linear_47_zero_point_0" -> "896 dequantize_per_channel_default_40"  [label="(384,)", style=solid];
+"894 compressed_weight_updated_constant32" -> "895 quantize_per_channel_default_40"  [label="(384, 384)", style=solid];
+"895 quantize_per_channel_default_40" -> "896 dequantize_per_channel_default_40"  [label="(384, 384)", style=solid];
+"896 dequantize_per_channel_default_40" -> "898 linear_47"  [label="(384, 384)", style=solid];
+"897 features_5_3_attn_proj_bias_0_0" -> "898 linear_47"  [label="(384,)", style=solid];
+"898 linear_47" -> "899 dropout_29"  [label="(4, 64, 384)", style=solid];
+"899 dropout_29" -> "900 view_43"  [label="(4, 64, 384)", style=solid];
+"900 view_43" -> "901 permute_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"901 permute_36" -> "902 reshape_35"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"902 reshape_35" -> "903 roll_7"  [label="(1, 16, 16, 384)", style=solid];
+"903 roll_7" -> "904 slice_129"  [label="(1, 16, 16, 384)", style=solid];
+"904 slice_129" -> "905 slice_130"  [label="(1, 16, 16, 384)", style=solid];
+"905 slice_130" -> "906 slice_131"  [label="(1, 14, 16, 384)", style=solid];
+"906 slice_131" -> "907 slice_132"  [label="(1, 14, 14, 384)", style=solid];
+"907 slice_132" -> "908 contiguous_13"  [label="(1, 14, 14, 384)", style=solid];
+"908 contiguous_13" -> "909 layer_norm_17"  [label="(1, 14, 14, 384)", style=solid];
+"909 layer_norm_17" -> "910 add_26"  [label="(1, 14, 14, 384)", style=solid];
+"910 add_26" -> "911 add_26_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"910 add_26" -> "935 add_27"  [label="(1, 14, 14, 384)", style=solid];
+"911 add_26_0_0_nncf_smooth_quant_0" -> "912 quantize_per_tensor_default_49"  [label="(1, 14, 14, 384)", style=solid];
+"912 quantize_per_tensor_default_49" -> "913 dequantize_per_tensor_default_49"  [label="(1, 14, 14, 384)", style=solid];
+"913 dequantize_per_tensor_default_49" -> "920 linear_48"  [label="(1, 14, 14, 384)", style=solid];
+"914 linear_48_scale_0" -> "917 quantize_per_channel_default_41"  [label="(1536,)", style=solid];
+"914 linear_48_scale_0" -> "918 dequantize_per_channel_default_41"  [label="(1536,)", style=solid];
+"915 linear_48_zero_point_0" -> "917 quantize_per_channel_default_41"  [label="(1536,)", style=solid];
+"915 linear_48_zero_point_0" -> "918 dequantize_per_channel_default_41"  [label="(1536,)", style=solid];
+"916 compressed_weight_updated_constant33" -> "917 quantize_per_channel_default_41"  [label="(1536, 384)", style=solid];
+"917 quantize_per_channel_default_41" -> "918 dequantize_per_channel_default_41"  [label="(1536, 384)", style=solid];
+"918 dequantize_per_channel_default_41" -> "920 linear_48"  [label="(1536, 384)", style=solid];
+"919 features_5_3_mlp_0_bias_0_0" -> "920 linear_48"  [label="(1536,)", style=solid];
+"920 linear_48" -> "921 gelu_7"  [label="(1, 14, 14, 1536)", style=solid];
+"921 gelu_7" -> "922 dropout_30"  [label="(1, 14, 14, 1536)", style=solid];
+"922 dropout_30" -> "923 dropout_30_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"923 dropout_30_0_0_nncf_smooth_quant_0" -> "924 quantize_per_tensor_default_50"  [label="(1, 14, 14, 1536)", style=solid];
+"924 quantize_per_tensor_default_50" -> "925 dequantize_per_tensor_default_50"  [label="(1, 14, 14, 1536)", style=solid];
+"925 dequantize_per_tensor_default_50" -> "932 linear_49"  [label="(1, 14, 14, 1536)", style=solid];
+"926 linear_49_scale_0" -> "929 quantize_per_channel_default_42"  [label="(384,)", style=solid];
+"926 linear_49_scale_0" -> "930 dequantize_per_channel_default_42"  [label="(384,)", style=solid];
+"927 linear_49_zero_point_0" -> "929 quantize_per_channel_default_42"  [label="(384,)", style=solid];
+"927 linear_49_zero_point_0" -> "930 dequantize_per_channel_default_42"  [label="(384,)", style=solid];
+"928 compressed_weight_updated_constant34" -> "929 quantize_per_channel_default_42"  [label="(384, 1536)", style=solid];
+"929 quantize_per_channel_default_42" -> "930 dequantize_per_channel_default_42"  [label="(384, 1536)", style=solid];
+"930 dequantize_per_channel_default_42" -> "932 linear_49"  [label="(384, 1536)", style=solid];
+"931 features_5_3_mlp_3_bias_0_0" -> "932 linear_49"  [label="(384,)", style=solid];
+"932 linear_49" -> "933 dropout_31"  [label="(1, 14, 14, 384)", style=solid];
+"933 dropout_31" -> "934 layer_norm_18"  [label="(1, 14, 14, 384)", style=solid];
+"934 layer_norm_18" -> "935 add_27"  [label="(1, 14, 14, 384)", style=solid];
+"935 add_27" -> "937 pad_10"  [label="(1, 14, 14, 384)", style=solid];
+"935 add_27" -> "998 add_29"  [label="(1, 14, 14, 384)", style=solid];
+"936 _frozen_param68" -> "972 add_28"  [label="(1, 12, 64, 64)", style=solid];
+"937 pad_10" -> "938 view_46"  [label="(1, 16, 16, 384)", style=solid];
+"938 view_46" -> "939 permute_38"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"939 permute_38" -> "940 reshape_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"940 reshape_36" -> "941 reshape_36_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"941 reshape_36_0_0_nncf_smooth_quant_0" -> "942 quantize_per_tensor_default_51"  [label="(4, 64, 384)", style=solid];
+"942 quantize_per_tensor_default_51" -> "943 dequantize_per_tensor_default_51"  [label="(4, 64, 384)", style=solid];
+"943 dequantize_per_tensor_default_51" -> "950 linear_52"  [label="(4, 64, 384)", style=solid];
+"944 linear_52_scale_0" -> "947 quantize_per_channel_default_44"  [label="(1152,)", style=solid];
+"944 linear_52_scale_0" -> "948 dequantize_per_channel_default_44"  [label="(1152,)", style=solid];
+"945 linear_52_zero_point_0" -> "947 quantize_per_channel_default_44"  [label="(1152,)", style=solid];
+"945 linear_52_zero_point_0" -> "948 dequantize_per_channel_default_44"  [label="(1152,)", style=solid];
+"946 compressed_weight_updated_constant35" -> "947 quantize_per_channel_default_44"  [label="(1152, 384)", style=solid];
+"947 quantize_per_channel_default_44" -> "948 dequantize_per_channel_default_44"  [label="(1152, 384)", style=solid];
+"948 dequantize_per_channel_default_44" -> "950 linear_52"  [label="(1152, 384)", style=solid];
+"949 _frozen_param69_0_0" -> "950 linear_52"  [label="(1152,)", style=solid];
+"950 linear_52" -> "951 reshape_37"  [label="(4, 64, 1152)", style=solid];
+"951 reshape_37" -> "952 permute_39"  [label="(4, 64, 3, 12, 32)", style=solid];
+"952 permute_39" -> "953 select_24"  [label="(3, 4, 12, 64, 32)", style=solid];
+"952 permute_39" -> "954 select_25"  [label="(3, 4, 12, 64, 32)", style=solid];
+"952 permute_39" -> "955 select_26"  [label="(3, 4, 12, 64, 32)", style=solid];
+"953 select_24" -> "956 linalg_vector_norm_16"  [label="(4, 12, 64, 32)", style=solid];
+"953 select_24" -> "958 expand_as_16"  [label="(4, 12, 64, 32)", style=solid];
+"953 select_24" -> "959 div_16"  [label="(4, 12, 64, 32)", style=solid];
+"954 select_25" -> "962 linalg_vector_norm_17"  [label="(4, 12, 64, 32)", style=solid];
+"954 select_25" -> "964 expand_as_17"  [label="(4, 12, 64, 32)", style=solid];
+"954 select_25" -> "965 div_17"  [label="(4, 12, 64, 32)", style=solid];
+"955 select_26" -> "975 matmul_17"  [label="(4, 12, 64, 32)", style=solid];
+"956 linalg_vector_norm_16" -> "957 clamp_min_16"  [label="(4, 12, 64, 1)", style=solid];
+"957 clamp_min_16" -> "958 expand_as_16"  [label="(4, 12, 64, 1)", style=solid];
+"958 expand_as_16" -> "959 div_16"  [label="(4, 12, 64, 32)", style=solid];
+"959 div_16" -> "960 quantize_per_tensor_default_52"  [label="(4, 12, 64, 32)", style=solid];
+"960 quantize_per_tensor_default_52" -> "961 dequantize_per_tensor_default_52"  [label="(4, 12, 64, 32)", style=solid];
+"961 dequantize_per_tensor_default_52" -> "969 matmul_16"  [label="(4, 12, 64, 32)", style=solid];
+"962 linalg_vector_norm_17" -> "963 clamp_min_17"  [label="(4, 12, 64, 1)", style=solid];
+"963 clamp_min_17" -> "964 expand_as_17"  [label="(4, 12, 64, 1)", style=solid];
+"964 expand_as_17" -> "965 div_17"  [label="(4, 12, 64, 32)", style=solid];
+"965 div_17" -> "966 quantize_per_tensor_default_53"  [label="(4, 12, 64, 32)", style=solid];
+"966 quantize_per_tensor_default_53" -> "967 dequantize_per_tensor_default_53"  [label="(4, 12, 64, 32)", style=solid];
+"967 dequantize_per_tensor_default_53" -> "968 transpose_16"  [label="(4, 12, 64, 32)", style=solid];
+"968 transpose_16" -> "969 matmul_16"  [label="(4, 12, 32, 64)", style=solid];
+"969 matmul_16" -> "971 mul_17"  [label="(4, 12, 64, 64)", style=solid];
+"970 _frozen_param71" -> "971 mul_17"  [label="(12, 1, 1)", style=solid];
+"971 mul_17" -> "972 add_28"  [label="(4, 12, 64, 64)", style=solid];
+"972 add_28" -> "973 softmax_8"  [label="(4, 12, 64, 64)", style=solid];
+"973 softmax_8" -> "974 dropout_32"  [label="(4, 12, 64, 64)", style=solid];
+"974 dropout_32" -> "975 matmul_17"  [label="(4, 12, 64, 64)", style=solid];
+"975 matmul_17" -> "976 transpose_17"  [label="(4, 12, 64, 32)", style=solid];
+"976 transpose_17" -> "977 reshape_38"  [label="(4, 64, 12, 32)", style=solid];
+"977 reshape_38" -> "978 reshape_38_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"978 reshape_38_0_0_nncf_smooth_quant_0" -> "979 quantize_per_tensor_default_54"  [label="(4, 64, 384)", style=solid];
+"979 quantize_per_tensor_default_54" -> "980 dequantize_per_tensor_default_54"  [label="(4, 64, 384)", style=solid];
+"980 dequantize_per_tensor_default_54" -> "987 linear_53"  [label="(4, 64, 384)", style=solid];
+"981 linear_53_scale_0" -> "984 quantize_per_channel_default_45"  [label="(384,)", style=solid];
+"981 linear_53_scale_0" -> "985 dequantize_per_channel_default_45"  [label="(384,)", style=solid];
+"982 linear_53_zero_point_0" -> "984 quantize_per_channel_default_45"  [label="(384,)", style=solid];
+"982 linear_53_zero_point_0" -> "985 dequantize_per_channel_default_45"  [label="(384,)", style=solid];
+"983 compressed_weight_updated_constant36" -> "984 quantize_per_channel_default_45"  [label="(384, 384)", style=solid];
+"984 quantize_per_channel_default_45" -> "985 dequantize_per_channel_default_45"  [label="(384, 384)", style=solid];
+"985 dequantize_per_channel_default_45" -> "987 linear_53"  [label="(384, 384)", style=solid];
+"986 features_5_4_attn_proj_bias_0_0" -> "987 linear_53"  [label="(384,)", style=solid];
+"987 linear_53" -> "988 dropout_33"  [label="(4, 64, 384)", style=solid];
+"988 dropout_33" -> "989 view_47"  [label="(4, 64, 384)", style=solid];
+"989 view_47" -> "990 permute_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"990 permute_40" -> "991 reshape_39"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"991 reshape_39" -> "992 slice_134"  [label="(1, 16, 16, 384)", style=solid];
+"992 slice_134" -> "993 slice_135"  [label="(1, 16, 16, 384)", style=solid];
+"993 slice_135" -> "994 slice_136"  [label="(1, 14, 16, 384)", style=solid];
+"994 slice_136" -> "995 slice_137"  [label="(1, 14, 14, 384)", style=solid];
+"995 slice_137" -> "996 contiguous_15"  [label="(1, 14, 14, 384)", style=solid];
+"996 contiguous_15" -> "997 layer_norm_19"  [label="(1, 14, 14, 384)", style=solid];
+"997 layer_norm_19" -> "998 add_29"  [label="(1, 14, 14, 384)", style=solid];
+"998 add_29" -> "999 add_29_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"998 add_29" -> "1023 add_30"  [label="(1, 14, 14, 384)", style=solid];
+"999 add_29_0_0_nncf_smooth_quant_0" -> "1000 quantize_per_tensor_default_55"  [label="(1, 14, 14, 384)", style=solid];
+"1000 quantize_per_tensor_default_55" -> "1001 dequantize_per_tensor_default_55"  [label="(1, 14, 14, 384)", style=solid];
+"1001 dequantize_per_tensor_default_55" -> "1008 linear_54"  [label="(1, 14, 14, 384)", style=solid];
+"1002 linear_54_scale_0" -> "1005 quantize_per_channel_default_46"  [label="(1536,)", style=solid];
+"1002 linear_54_scale_0" -> "1006 dequantize_per_channel_default_46"  [label="(1536,)", style=solid];
+"1003 linear_54_zero_point_0" -> "1005 quantize_per_channel_default_46"  [label="(1536,)", style=solid];
+"1003 linear_54_zero_point_0" -> "1006 dequantize_per_channel_default_46"  [label="(1536,)", style=solid];
+"1004 compressed_weight_updated_constant37" -> "1005 quantize_per_channel_default_46"  [label="(1536, 384)", style=solid];
+"1005 quantize_per_channel_default_46" -> "1006 dequantize_per_channel_default_46"  [label="(1536, 384)", style=solid];
+"1006 dequantize_per_channel_default_46" -> "1008 linear_54"  [label="(1536, 384)", style=solid];
+"1007 features_5_4_mlp_0_bias_0_0" -> "1008 linear_54"  [label="(1536,)", style=solid];
+"1008 linear_54" -> "1009 gelu_8"  [label="(1, 14, 14, 1536)", style=solid];
+"1009 gelu_8" -> "1010 dropout_34"  [label="(1, 14, 14, 1536)", style=solid];
+"1010 dropout_34" -> "1011 dropout_34_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1011 dropout_34_0_0_nncf_smooth_quant_0" -> "1012 quantize_per_tensor_default_56"  [label="(1, 14, 14, 1536)", style=solid];
+"1012 quantize_per_tensor_default_56" -> "1013 dequantize_per_tensor_default_56"  [label="(1, 14, 14, 1536)", style=solid];
+"1013 dequantize_per_tensor_default_56" -> "1020 linear_55"  [label="(1, 14, 14, 1536)", style=solid];
+"1014 linear_55_scale_0" -> "1017 quantize_per_channel_default_47"  [label="(384,)", style=solid];
+"1014 linear_55_scale_0" -> "1018 dequantize_per_channel_default_47"  [label="(384,)", style=solid];
+"1015 linear_55_zero_point_0" -> "1017 quantize_per_channel_default_47"  [label="(384,)", style=solid];
+"1015 linear_55_zero_point_0" -> "1018 dequantize_per_channel_default_47"  [label="(384,)", style=solid];
+"1016 compressed_weight_updated_constant38" -> "1017 quantize_per_channel_default_47"  [label="(384, 1536)", style=solid];
+"1017 quantize_per_channel_default_47" -> "1018 dequantize_per_channel_default_47"  [label="(384, 1536)", style=solid];
+"1018 dequantize_per_channel_default_47" -> "1020 linear_55"  [label="(384, 1536)", style=solid];
+"1019 features_5_4_mlp_3_bias_0_0" -> "1020 linear_55"  [label="(384,)", style=solid];
+"1020 linear_55" -> "1021 dropout_35"  [label="(1, 14, 14, 384)", style=solid];
+"1021 dropout_35" -> "1022 layer_norm_20"  [label="(1, 14, 14, 384)", style=solid];
+"1022 layer_norm_20" -> "1023 add_30"  [label="(1, 14, 14, 384)", style=solid];
+"1023 add_30" -> "1025 pad_11"  [label="(1, 14, 14, 384)", style=solid];
+"1023 add_30" -> "1104 add_33"  [label="(1, 14, 14, 384)", style=solid];
+"1024 _frozen_param72" -> "1061 add_31"  [label="(1, 12, 64, 64)", style=solid];
+"1025 pad_11" -> "1026 roll_8"  [label="(1, 16, 16, 384)", style=solid];
+"1026 roll_8" -> "1027 view_50"  [label="(1, 16, 16, 384)", style=solid];
+"1027 view_50" -> "1028 permute_42"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1028 permute_42" -> "1029 reshape_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1029 reshape_40" -> "1030 reshape_40_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1029 reshape_40" -> "1062 new_zeros_4"  [label="(4, 64, 384)", style=solid];
+"1030 reshape_40_0_0_nncf_smooth_quant_0" -> "1031 quantize_per_tensor_default_57"  [label="(4, 64, 384)", style=solid];
+"1031 quantize_per_tensor_default_57" -> "1032 dequantize_per_tensor_default_57"  [label="(4, 64, 384)", style=solid];
+"1032 dequantize_per_tensor_default_57" -> "1039 linear_58"  [label="(4, 64, 384)", style=solid];
+"1033 linear_58_scale_0" -> "1036 quantize_per_channel_default_49"  [label="(1152,)", style=solid];
+"1033 linear_58_scale_0" -> "1037 dequantize_per_channel_default_49"  [label="(1152,)", style=solid];
+"1034 linear_58_zero_point_0" -> "1036 quantize_per_channel_default_49"  [label="(1152,)", style=solid];
+"1034 linear_58_zero_point_0" -> "1037 dequantize_per_channel_default_49"  [label="(1152,)", style=solid];
+"1035 compressed_weight_updated_constant39" -> "1036 quantize_per_channel_default_49"  [label="(1152, 384)", style=solid];
+"1036 quantize_per_channel_default_49" -> "1037 dequantize_per_channel_default_49"  [label="(1152, 384)", style=solid];
+"1037 dequantize_per_channel_default_49" -> "1039 linear_58"  [label="(1152, 384)", style=solid];
+"1038 _frozen_param73_0_0" -> "1039 linear_58"  [label="(1152,)", style=solid];
+"1039 linear_58" -> "1040 reshape_41"  [label="(4, 64, 1152)", style=solid];
+"1040 reshape_41" -> "1041 permute_43"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1041 permute_43" -> "1042 select_27"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1041 permute_43" -> "1043 select_28"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1041 permute_43" -> "1044 select_29"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1042 select_27" -> "1045 linalg_vector_norm_18"  [label="(4, 12, 64, 32)", style=solid];
+"1042 select_27" -> "1047 expand_as_18"  [label="(4, 12, 64, 32)", style=solid];
+"1042 select_27" -> "1048 div_18"  [label="(4, 12, 64, 32)", style=solid];
+"1043 select_28" -> "1051 linalg_vector_norm_19"  [label="(4, 12, 64, 32)", style=solid];
+"1043 select_28" -> "1053 expand_as_19"  [label="(4, 12, 64, 32)", style=solid];
+"1043 select_28" -> "1054 div_19"  [label="(4, 12, 64, 32)", style=solid];
+"1044 select_29" -> "1080 matmul_19"  [label="(4, 12, 64, 32)", style=solid];
+"1045 linalg_vector_norm_18" -> "1046 clamp_min_18"  [label="(4, 12, 64, 1)", style=solid];
+"1046 clamp_min_18" -> "1047 expand_as_18"  [label="(4, 12, 64, 1)", style=solid];
+"1047 expand_as_18" -> "1048 div_18"  [label="(4, 12, 64, 32)", style=solid];
+"1048 div_18" -> "1049 quantize_per_tensor_default_58"  [label="(4, 12, 64, 32)", style=solid];
+"1049 quantize_per_tensor_default_58" -> "1050 dequantize_per_tensor_default_58"  [label="(4, 12, 64, 32)", style=solid];
+"1050 dequantize_per_tensor_default_58" -> "1058 matmul_18"  [label="(4, 12, 64, 32)", style=solid];
+"1051 linalg_vector_norm_19" -> "1052 clamp_min_19"  [label="(4, 12, 64, 1)", style=solid];
+"1052 clamp_min_19" -> "1053 expand_as_19"  [label="(4, 12, 64, 1)", style=solid];
+"1053 expand_as_19" -> "1054 div_19"  [label="(4, 12, 64, 32)", style=solid];
+"1054 div_19" -> "1055 quantize_per_tensor_default_59"  [label="(4, 12, 64, 32)", style=solid];
+"1055 quantize_per_tensor_default_59" -> "1056 dequantize_per_tensor_default_59"  [label="(4, 12, 64, 32)", style=solid];
+"1056 dequantize_per_tensor_default_59" -> "1057 transpose_18"  [label="(4, 12, 64, 32)", style=solid];
+"1057 transpose_18" -> "1058 matmul_18"  [label="(4, 12, 32, 64)", style=solid];
+"1058 matmul_18" -> "1060 mul_19"  [label="(4, 12, 64, 64)", style=solid];
+"1059 _frozen_param75" -> "1060 mul_19"  [label="(12, 1, 1)", style=solid];
+"1060 mul_19" -> "1061 add_31"  [label="(4, 12, 64, 64)", style=solid];
+"1061 add_31" -> "1073 view_52"  [label="(4, 12, 64, 64)", style=solid];
+"1062 new_zeros_4" -> "1063 view_51"  [label="(16, 16)", style=solid];
+"1063 view_51" -> "1064 permute_44"  [label="(2, 8, 2, 8)", style=solid];
+"1064 permute_44" -> "1065 reshape_42"  [label="(2, 2, 8, 8)", style=solid];
+"1065 reshape_42" -> "1066 unsqueeze_26"  [label="(4, 64)", style=solid];
+"1065 reshape_42" -> "1067 unsqueeze_27"  [label="(4, 64)", style=solid];
+"1066 unsqueeze_26" -> "1068 sub_4"  [label="(4, 1, 64)", style=solid];
+"1067 unsqueeze_27" -> "1068 sub_4"  [label="(4, 64, 1)", style=solid];
+"1068 sub_4" -> "1069 ne_4"  [label="(4, 64, 64)", style=solid];
+"1068 sub_4" -> "1070 masked_fill_8"  [label="(4, 64, 64)", style=solid];
+"1068 sub_4" -> "1071 eq_4"  [label="(4, 64, 64)", style=solid];
+"1069 ne_4" -> "1070 masked_fill_8"  [label="(4, 64, 64)", style=solid];
+"1070 masked_fill_8" -> "1072 masked_fill_9"  [label="(4, 64, 64)", style=solid];
+"1071 eq_4" -> "1072 masked_fill_9"  [label="(4, 64, 64)", style=solid];
+"1072 masked_fill_9" -> "1074 unsqueeze_28"  [label="(4, 64, 64)", style=solid];
+"1073 view_52" -> "1076 add_32"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1074 unsqueeze_28" -> "1075 unsqueeze_29"  [label="(4, 1, 64, 64)", style=solid];
+"1075 unsqueeze_29" -> "1076 add_32"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1076 add_32" -> "1077 view_53"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1077 view_53" -> "1078 softmax_9"  [label="(4, 12, 64, 64)", style=solid];
+"1078 softmax_9" -> "1079 dropout_36"  [label="(4, 12, 64, 64)", style=solid];
+"1079 dropout_36" -> "1080 matmul_19"  [label="(4, 12, 64, 64)", style=solid];
+"1080 matmul_19" -> "1081 transpose_19"  [label="(4, 12, 64, 32)", style=solid];
+"1081 transpose_19" -> "1082 reshape_43"  [label="(4, 64, 12, 32)", style=solid];
+"1082 reshape_43" -> "1083 reshape_43_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1083 reshape_43_0_0_nncf_smooth_quant_0" -> "1084 quantize_per_tensor_default_60"  [label="(4, 64, 384)", style=solid];
+"1084 quantize_per_tensor_default_60" -> "1085 dequantize_per_tensor_default_60"  [label="(4, 64, 384)", style=solid];
+"1085 dequantize_per_tensor_default_60" -> "1092 linear_59"  [label="(4, 64, 384)", style=solid];
+"1086 linear_59_scale_0" -> "1089 quantize_per_channel_default_50"  [label="(384,)", style=solid];
+"1086 linear_59_scale_0" -> "1090 dequantize_per_channel_default_50"  [label="(384,)", style=solid];
+"1087 linear_59_zero_point_0" -> "1089 quantize_per_channel_default_50"  [label="(384,)", style=solid];
+"1087 linear_59_zero_point_0" -> "1090 dequantize_per_channel_default_50"  [label="(384,)", style=solid];
+"1088 compressed_weight_updated_constant40" -> "1089 quantize_per_channel_default_50"  [label="(384, 384)", style=solid];
+"1089 quantize_per_channel_default_50" -> "1090 dequantize_per_channel_default_50"  [label="(384, 384)", style=solid];
+"1090 dequantize_per_channel_default_50" -> "1092 linear_59"  [label="(384, 384)", style=solid];
+"1091 features_5_5_attn_proj_bias_0_0" -> "1092 linear_59"  [label="(384,)", style=solid];
+"1092 linear_59" -> "1093 dropout_37"  [label="(4, 64, 384)", style=solid];
+"1093 dropout_37" -> "1094 view_54"  [label="(4, 64, 384)", style=solid];
+"1094 view_54" -> "1095 permute_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1095 permute_45" -> "1096 reshape_44"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1096 reshape_44" -> "1097 roll_9"  [label="(1, 16, 16, 384)", style=solid];
+"1097 roll_9" -> "1098 slice_157"  [label="(1, 16, 16, 384)", style=solid];
+"1098 slice_157" -> "1099 slice_158"  [label="(1, 16, 16, 384)", style=solid];
+"1099 slice_158" -> "1100 slice_159"  [label="(1, 14, 16, 384)", style=solid];
+"1100 slice_159" -> "1101 slice_160"  [label="(1, 14, 14, 384)", style=solid];
+"1101 slice_160" -> "1102 contiguous_17"  [label="(1, 14, 14, 384)", style=solid];
+"1102 contiguous_17" -> "1103 layer_norm_21"  [label="(1, 14, 14, 384)", style=solid];
+"1103 layer_norm_21" -> "1104 add_33"  [label="(1, 14, 14, 384)", style=solid];
+"1104 add_33" -> "1105 add_33_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1104 add_33" -> "1129 add_34"  [label="(1, 14, 14, 384)", style=solid];
+"1105 add_33_0_0_nncf_smooth_quant_0" -> "1106 quantize_per_tensor_default_61"  [label="(1, 14, 14, 384)", style=solid];
+"1106 quantize_per_tensor_default_61" -> "1107 dequantize_per_tensor_default_61"  [label="(1, 14, 14, 384)", style=solid];
+"1107 dequantize_per_tensor_default_61" -> "1114 linear_60"  [label="(1, 14, 14, 384)", style=solid];
+"1108 linear_60_scale_0" -> "1111 quantize_per_channel_default_51"  [label="(1536,)", style=solid];
+"1108 linear_60_scale_0" -> "1112 dequantize_per_channel_default_51"  [label="(1536,)", style=solid];
+"1109 linear_60_zero_point_0" -> "1111 quantize_per_channel_default_51"  [label="(1536,)", style=solid];
+"1109 linear_60_zero_point_0" -> "1112 dequantize_per_channel_default_51"  [label="(1536,)", style=solid];
+"1110 compressed_weight_updated_constant41" -> "1111 quantize_per_channel_default_51"  [label="(1536, 384)", style=solid];
+"1111 quantize_per_channel_default_51" -> "1112 dequantize_per_channel_default_51"  [label="(1536, 384)", style=solid];
+"1112 dequantize_per_channel_default_51" -> "1114 linear_60"  [label="(1536, 384)", style=solid];
+"1113 features_5_5_mlp_0_bias_0_0" -> "1114 linear_60"  [label="(1536,)", style=solid];
+"1114 linear_60" -> "1115 gelu_9"  [label="(1, 14, 14, 1536)", style=solid];
+"1115 gelu_9" -> "1116 dropout_38"  [label="(1, 14, 14, 1536)", style=solid];
+"1116 dropout_38" -> "1117 dropout_38_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1117 dropout_38_0_0_nncf_smooth_quant_0" -> "1118 quantize_per_tensor_default_62"  [label="(1, 14, 14, 1536)", style=solid];
+"1118 quantize_per_tensor_default_62" -> "1119 dequantize_per_tensor_default_62"  [label="(1, 14, 14, 1536)", style=solid];
+"1119 dequantize_per_tensor_default_62" -> "1126 linear_61"  [label="(1, 14, 14, 1536)", style=solid];
+"1120 linear_61_scale_0" -> "1123 quantize_per_channel_default_52"  [label="(384,)", style=solid];
+"1120 linear_61_scale_0" -> "1124 dequantize_per_channel_default_52"  [label="(384,)", style=solid];
+"1121 linear_61_zero_point_0" -> "1123 quantize_per_channel_default_52"  [label="(384,)", style=solid];
+"1121 linear_61_zero_point_0" -> "1124 dequantize_per_channel_default_52"  [label="(384,)", style=solid];
+"1122 compressed_weight_updated_constant42" -> "1123 quantize_per_channel_default_52"  [label="(384, 1536)", style=solid];
+"1123 quantize_per_channel_default_52" -> "1124 dequantize_per_channel_default_52"  [label="(384, 1536)", style=solid];
+"1124 dequantize_per_channel_default_52" -> "1126 linear_61"  [label="(384, 1536)", style=solid];
+"1125 features_5_5_mlp_3_bias_0_0" -> "1126 linear_61"  [label="(384,)", style=solid];
+"1126 linear_61" -> "1127 dropout_39"  [label="(1, 14, 14, 384)", style=solid];
+"1127 dropout_39" -> "1128 layer_norm_22"  [label="(1, 14, 14, 384)", style=solid];
+"1128 layer_norm_22" -> "1129 add_34"  [label="(1, 14, 14, 384)", style=solid];
+"1129 add_34" -> "1131 pad_12"  [label="(1, 14, 14, 384)", style=solid];
+"1129 add_34" -> "1192 add_36"  [label="(1, 14, 14, 384)", style=solid];
+"1130 _frozen_param85" -> "1166 add_35"  [label="(1, 12, 64, 64)", style=solid];
+"1131 pad_12" -> "1132 view_57"  [label="(1, 16, 16, 384)", style=solid];
+"1132 view_57" -> "1133 permute_47"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1133 permute_47" -> "1134 reshape_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1134 reshape_45" -> "1135 reshape_45_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1135 reshape_45_0_0_nncf_smooth_quant_0" -> "1136 quantize_per_tensor_default_63"  [label="(4, 64, 384)", style=solid];
+"1136 quantize_per_tensor_default_63" -> "1137 dequantize_per_tensor_default_63"  [label="(4, 64, 384)", style=solid];
+"1137 dequantize_per_tensor_default_63" -> "1144 linear_64"  [label="(4, 64, 384)", style=solid];
+"1138 linear_64_scale_0" -> "1141 quantize_per_channel_default_54"  [label="(1152,)", style=solid];
+"1138 linear_64_scale_0" -> "1142 dequantize_per_channel_default_54"  [label="(1152,)", style=solid];
+"1139 linear_64_zero_point_0" -> "1141 quantize_per_channel_default_54"  [label="(1152,)", style=solid];
+"1139 linear_64_zero_point_0" -> "1142 dequantize_per_channel_default_54"  [label="(1152,)", style=solid];
+"1140 compressed_weight_updated_constant43" -> "1141 quantize_per_channel_default_54"  [label="(1152, 384)", style=solid];
+"1141 quantize_per_channel_default_54" -> "1142 dequantize_per_channel_default_54"  [label="(1152, 384)", style=solid];
+"1142 dequantize_per_channel_default_54" -> "1144 linear_64"  [label="(1152, 384)", style=solid];
+"1143 _frozen_param86_0_0" -> "1144 linear_64"  [label="(1152,)", style=solid];
+"1144 linear_64" -> "1145 reshape_46"  [label="(4, 64, 1152)", style=solid];
+"1145 reshape_46" -> "1146 permute_48"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1146 permute_48" -> "1147 select_30"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1146 permute_48" -> "1148 select_31"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1146 permute_48" -> "1149 select_32"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1147 select_30" -> "1150 linalg_vector_norm_20"  [label="(4, 12, 64, 32)", style=solid];
+"1147 select_30" -> "1152 expand_as_20"  [label="(4, 12, 64, 32)", style=solid];
+"1147 select_30" -> "1153 div_20"  [label="(4, 12, 64, 32)", style=solid];
+"1148 select_31" -> "1156 linalg_vector_norm_21"  [label="(4, 12, 64, 32)", style=solid];
+"1148 select_31" -> "1158 expand_as_21"  [label="(4, 12, 64, 32)", style=solid];
+"1148 select_31" -> "1159 div_21"  [label="(4, 12, 64, 32)", style=solid];
+"1149 select_32" -> "1169 matmul_21"  [label="(4, 12, 64, 32)", style=solid];
+"1150 linalg_vector_norm_20" -> "1151 clamp_min_20"  [label="(4, 12, 64, 1)", style=solid];
+"1151 clamp_min_20" -> "1152 expand_as_20"  [label="(4, 12, 64, 1)", style=solid];
+"1152 expand_as_20" -> "1153 div_20"  [label="(4, 12, 64, 32)", style=solid];
+"1153 div_20" -> "1154 quantize_per_tensor_default_64"  [label="(4, 12, 64, 32)", style=solid];
+"1154 quantize_per_tensor_default_64" -> "1155 dequantize_per_tensor_default_64"  [label="(4, 12, 64, 32)", style=solid];
+"1155 dequantize_per_tensor_default_64" -> "1163 matmul_20"  [label="(4, 12, 64, 32)", style=solid];
+"1156 linalg_vector_norm_21" -> "1157 clamp_min_21"  [label="(4, 12, 64, 1)", style=solid];
+"1157 clamp_min_21" -> "1158 expand_as_21"  [label="(4, 12, 64, 1)", style=solid];
+"1158 expand_as_21" -> "1159 div_21"  [label="(4, 12, 64, 32)", style=solid];
+"1159 div_21" -> "1160 quantize_per_tensor_default_65"  [label="(4, 12, 64, 32)", style=solid];
+"1160 quantize_per_tensor_default_65" -> "1161 dequantize_per_tensor_default_65"  [label="(4, 12, 64, 32)", style=solid];
+"1161 dequantize_per_tensor_default_65" -> "1162 transpose_20"  [label="(4, 12, 64, 32)", style=solid];
+"1162 transpose_20" -> "1163 matmul_20"  [label="(4, 12, 32, 64)", style=solid];
+"1163 matmul_20" -> "1165 mul_21"  [label="(4, 12, 64, 64)", style=solid];
+"1164 _frozen_param88" -> "1165 mul_21"  [label="(12, 1, 1)", style=solid];
+"1165 mul_21" -> "1166 add_35"  [label="(4, 12, 64, 64)", style=solid];
+"1166 add_35" -> "1167 softmax_10"  [label="(4, 12, 64, 64)", style=solid];
+"1167 softmax_10" -> "1168 dropout_40"  [label="(4, 12, 64, 64)", style=solid];
+"1168 dropout_40" -> "1169 matmul_21"  [label="(4, 12, 64, 64)", style=solid];
+"1169 matmul_21" -> "1170 transpose_21"  [label="(4, 12, 64, 32)", style=solid];
+"1170 transpose_21" -> "1171 reshape_47"  [label="(4, 64, 12, 32)", style=solid];
+"1171 reshape_47" -> "1172 reshape_47_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1172 reshape_47_0_0_nncf_smooth_quant_0" -> "1173 quantize_per_tensor_default_66"  [label="(4, 64, 384)", style=solid];
+"1173 quantize_per_tensor_default_66" -> "1174 dequantize_per_tensor_default_66"  [label="(4, 64, 384)", style=solid];
+"1174 dequantize_per_tensor_default_66" -> "1181 linear_65"  [label="(4, 64, 384)", style=solid];
+"1175 linear_65_scale_0" -> "1178 quantize_per_channel_default_55"  [label="(384,)", style=solid];
+"1175 linear_65_scale_0" -> "1179 dequantize_per_channel_default_55"  [label="(384,)", style=solid];
+"1176 linear_65_zero_point_0" -> "1178 quantize_per_channel_default_55"  [label="(384,)", style=solid];
+"1176 linear_65_zero_point_0" -> "1179 dequantize_per_channel_default_55"  [label="(384,)", style=solid];
+"1177 compressed_weight_updated_constant44" -> "1178 quantize_per_channel_default_55"  [label="(384, 384)", style=solid];
+"1178 quantize_per_channel_default_55" -> "1179 dequantize_per_channel_default_55"  [label="(384, 384)", style=solid];
+"1179 dequantize_per_channel_default_55" -> "1181 linear_65"  [label="(384, 384)", style=solid];
+"1180 features_5_6_attn_proj_bias_0_0" -> "1181 linear_65"  [label="(384,)", style=solid];
+"1181 linear_65" -> "1182 dropout_41"  [label="(4, 64, 384)", style=solid];
+"1182 dropout_41" -> "1183 view_58"  [label="(4, 64, 384)", style=solid];
+"1183 view_58" -> "1184 permute_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1184 permute_49" -> "1185 reshape_48"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1185 reshape_48" -> "1186 slice_162"  [label="(1, 16, 16, 384)", style=solid];
+"1186 slice_162" -> "1187 slice_163"  [label="(1, 16, 16, 384)", style=solid];
+"1187 slice_163" -> "1188 slice_164"  [label="(1, 14, 16, 384)", style=solid];
+"1188 slice_164" -> "1189 slice_165"  [label="(1, 14, 14, 384)", style=solid];
+"1189 slice_165" -> "1190 contiguous_19"  [label="(1, 14, 14, 384)", style=solid];
+"1190 contiguous_19" -> "1191 layer_norm_23"  [label="(1, 14, 14, 384)", style=solid];
+"1191 layer_norm_23" -> "1192 add_36"  [label="(1, 14, 14, 384)", style=solid];
+"1192 add_36" -> "1193 add_36_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1192 add_36" -> "1217 add_37"  [label="(1, 14, 14, 384)", style=solid];
+"1193 add_36_0_0_nncf_smooth_quant_0" -> "1194 quantize_per_tensor_default_67"  [label="(1, 14, 14, 384)", style=solid];
+"1194 quantize_per_tensor_default_67" -> "1195 dequantize_per_tensor_default_67"  [label="(1, 14, 14, 384)", style=solid];
+"1195 dequantize_per_tensor_default_67" -> "1202 linear_66"  [label="(1, 14, 14, 384)", style=solid];
+"1196 linear_66_scale_0" -> "1199 quantize_per_channel_default_56"  [label="(1536,)", style=solid];
+"1196 linear_66_scale_0" -> "1200 dequantize_per_channel_default_56"  [label="(1536,)", style=solid];
+"1197 linear_66_zero_point_0" -> "1199 quantize_per_channel_default_56"  [label="(1536,)", style=solid];
+"1197 linear_66_zero_point_0" -> "1200 dequantize_per_channel_default_56"  [label="(1536,)", style=solid];
+"1198 compressed_weight_updated_constant45" -> "1199 quantize_per_channel_default_56"  [label="(1536, 384)", style=solid];
+"1199 quantize_per_channel_default_56" -> "1200 dequantize_per_channel_default_56"  [label="(1536, 384)", style=solid];
+"1200 dequantize_per_channel_default_56" -> "1202 linear_66"  [label="(1536, 384)", style=solid];
+"1201 features_5_6_mlp_0_bias_0_0" -> "1202 linear_66"  [label="(1536,)", style=solid];
+"1202 linear_66" -> "1203 gelu_10"  [label="(1, 14, 14, 1536)", style=solid];
+"1203 gelu_10" -> "1204 dropout_42"  [label="(1, 14, 14, 1536)", style=solid];
+"1204 dropout_42" -> "1205 dropout_42_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1205 dropout_42_0_0_nncf_smooth_quant_0" -> "1206 quantize_per_tensor_default_68"  [label="(1, 14, 14, 1536)", style=solid];
+"1206 quantize_per_tensor_default_68" -> "1207 dequantize_per_tensor_default_68"  [label="(1, 14, 14, 1536)", style=solid];
+"1207 dequantize_per_tensor_default_68" -> "1214 linear_67"  [label="(1, 14, 14, 1536)", style=solid];
+"1208 linear_67_scale_0" -> "1211 quantize_per_channel_default_57"  [label="(384,)", style=solid];
+"1208 linear_67_scale_0" -> "1212 dequantize_per_channel_default_57"  [label="(384,)", style=solid];
+"1209 linear_67_zero_point_0" -> "1211 quantize_per_channel_default_57"  [label="(384,)", style=solid];
+"1209 linear_67_zero_point_0" -> "1212 dequantize_per_channel_default_57"  [label="(384,)", style=solid];
+"1210 compressed_weight_updated_constant46" -> "1211 quantize_per_channel_default_57"  [label="(384, 1536)", style=solid];
+"1211 quantize_per_channel_default_57" -> "1212 dequantize_per_channel_default_57"  [label="(384, 1536)", style=solid];
+"1212 dequantize_per_channel_default_57" -> "1214 linear_67"  [label="(384, 1536)", style=solid];
+"1213 features_5_6_mlp_3_bias_0_0" -> "1214 linear_67"  [label="(384,)", style=solid];
+"1214 linear_67" -> "1215 dropout_43"  [label="(1, 14, 14, 384)", style=solid];
+"1215 dropout_43" -> "1216 layer_norm_24"  [label="(1, 14, 14, 384)", style=solid];
+"1216 layer_norm_24" -> "1217 add_37"  [label="(1, 14, 14, 384)", style=solid];
+"1217 add_37" -> "1219 pad_13"  [label="(1, 14, 14, 384)", style=solid];
+"1217 add_37" -> "1298 add_40"  [label="(1, 14, 14, 384)", style=solid];
+"1218 _frozen_param89" -> "1255 add_38"  [label="(1, 12, 64, 64)", style=solid];
+"1219 pad_13" -> "1220 roll_10"  [label="(1, 16, 16, 384)", style=solid];
+"1220 roll_10" -> "1221 view_61"  [label="(1, 16, 16, 384)", style=solid];
+"1221 view_61" -> "1222 permute_51"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1222 permute_51" -> "1223 reshape_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1223 reshape_49" -> "1224 reshape_49_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1223 reshape_49" -> "1256 new_zeros_5"  [label="(4, 64, 384)", style=solid];
+"1224 reshape_49_0_0_nncf_smooth_quant_0" -> "1225 quantize_per_tensor_default_69"  [label="(4, 64, 384)", style=solid];
+"1225 quantize_per_tensor_default_69" -> "1226 dequantize_per_tensor_default_69"  [label="(4, 64, 384)", style=solid];
+"1226 dequantize_per_tensor_default_69" -> "1233 linear_70"  [label="(4, 64, 384)", style=solid];
+"1227 linear_70_scale_0" -> "1230 quantize_per_channel_default_59"  [label="(1152,)", style=solid];
+"1227 linear_70_scale_0" -> "1231 dequantize_per_channel_default_59"  [label="(1152,)", style=solid];
+"1228 linear_70_zero_point_0" -> "1230 quantize_per_channel_default_59"  [label="(1152,)", style=solid];
+"1228 linear_70_zero_point_0" -> "1231 dequantize_per_channel_default_59"  [label="(1152,)", style=solid];
+"1229 compressed_weight_updated_constant47" -> "1230 quantize_per_channel_default_59"  [label="(1152, 384)", style=solid];
+"1230 quantize_per_channel_default_59" -> "1231 dequantize_per_channel_default_59"  [label="(1152, 384)", style=solid];
+"1231 dequantize_per_channel_default_59" -> "1233 linear_70"  [label="(1152, 384)", style=solid];
+"1232 _frozen_param90_0_0" -> "1233 linear_70"  [label="(1152,)", style=solid];
+"1233 linear_70" -> "1234 reshape_50"  [label="(4, 64, 1152)", style=solid];
+"1234 reshape_50" -> "1235 permute_52"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1235 permute_52" -> "1236 select_33"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1235 permute_52" -> "1237 select_34"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1235 permute_52" -> "1238 select_35"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1236 select_33" -> "1239 linalg_vector_norm_22"  [label="(4, 12, 64, 32)", style=solid];
+"1236 select_33" -> "1241 expand_as_22"  [label="(4, 12, 64, 32)", style=solid];
+"1236 select_33" -> "1242 div_22"  [label="(4, 12, 64, 32)", style=solid];
+"1237 select_34" -> "1245 linalg_vector_norm_23"  [label="(4, 12, 64, 32)", style=solid];
+"1237 select_34" -> "1247 expand_as_23"  [label="(4, 12, 64, 32)", style=solid];
+"1237 select_34" -> "1248 div_23"  [label="(4, 12, 64, 32)", style=solid];
+"1238 select_35" -> "1274 matmul_23"  [label="(4, 12, 64, 32)", style=solid];
+"1239 linalg_vector_norm_22" -> "1240 clamp_min_22"  [label="(4, 12, 64, 1)", style=solid];
+"1240 clamp_min_22" -> "1241 expand_as_22"  [label="(4, 12, 64, 1)", style=solid];
+"1241 expand_as_22" -> "1242 div_22"  [label="(4, 12, 64, 32)", style=solid];
+"1242 div_22" -> "1243 quantize_per_tensor_default_70"  [label="(4, 12, 64, 32)", style=solid];
+"1243 quantize_per_tensor_default_70" -> "1244 dequantize_per_tensor_default_70"  [label="(4, 12, 64, 32)", style=solid];
+"1244 dequantize_per_tensor_default_70" -> "1252 matmul_22"  [label="(4, 12, 64, 32)", style=solid];
+"1245 linalg_vector_norm_23" -> "1246 clamp_min_23"  [label="(4, 12, 64, 1)", style=solid];
+"1246 clamp_min_23" -> "1247 expand_as_23"  [label="(4, 12, 64, 1)", style=solid];
+"1247 expand_as_23" -> "1248 div_23"  [label="(4, 12, 64, 32)", style=solid];
+"1248 div_23" -> "1249 quantize_per_tensor_default_71"  [label="(4, 12, 64, 32)", style=solid];
+"1249 quantize_per_tensor_default_71" -> "1250 dequantize_per_tensor_default_71"  [label="(4, 12, 64, 32)", style=solid];
+"1250 dequantize_per_tensor_default_71" -> "1251 transpose_22"  [label="(4, 12, 64, 32)", style=solid];
+"1251 transpose_22" -> "1252 matmul_22"  [label="(4, 12, 32, 64)", style=solid];
+"1252 matmul_22" -> "1254 mul_23"  [label="(4, 12, 64, 64)", style=solid];
+"1253 _frozen_param92" -> "1254 mul_23"  [label="(12, 1, 1)", style=solid];
+"1254 mul_23" -> "1255 add_38"  [label="(4, 12, 64, 64)", style=solid];
+"1255 add_38" -> "1267 view_63"  [label="(4, 12, 64, 64)", style=solid];
+"1256 new_zeros_5" -> "1257 view_62"  [label="(16, 16)", style=solid];
+"1257 view_62" -> "1258 permute_53"  [label="(2, 8, 2, 8)", style=solid];
+"1258 permute_53" -> "1259 reshape_51"  [label="(2, 2, 8, 8)", style=solid];
+"1259 reshape_51" -> "1260 unsqueeze_32"  [label="(4, 64)", style=solid];
+"1259 reshape_51" -> "1261 unsqueeze_33"  [label="(4, 64)", style=solid];
+"1260 unsqueeze_32" -> "1262 sub_5"  [label="(4, 1, 64)", style=solid];
+"1261 unsqueeze_33" -> "1262 sub_5"  [label="(4, 64, 1)", style=solid];
+"1262 sub_5" -> "1263 ne_5"  [label="(4, 64, 64)", style=solid];
+"1262 sub_5" -> "1264 masked_fill_10"  [label="(4, 64, 64)", style=solid];
+"1262 sub_5" -> "1265 eq_5"  [label="(4, 64, 64)", style=solid];
+"1263 ne_5" -> "1264 masked_fill_10"  [label="(4, 64, 64)", style=solid];
+"1264 masked_fill_10" -> "1266 masked_fill_11"  [label="(4, 64, 64)", style=solid];
+"1265 eq_5" -> "1266 masked_fill_11"  [label="(4, 64, 64)", style=solid];
+"1266 masked_fill_11" -> "1268 unsqueeze_34"  [label="(4, 64, 64)", style=solid];
+"1267 view_63" -> "1270 add_39"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1268 unsqueeze_34" -> "1269 unsqueeze_35"  [label="(4, 1, 64, 64)", style=solid];
+"1269 unsqueeze_35" -> "1270 add_39"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1270 add_39" -> "1271 view_64"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1271 view_64" -> "1272 softmax_11"  [label="(4, 12, 64, 64)", style=solid];
+"1272 softmax_11" -> "1273 dropout_44"  [label="(4, 12, 64, 64)", style=solid];
+"1273 dropout_44" -> "1274 matmul_23"  [label="(4, 12, 64, 64)", style=solid];
+"1274 matmul_23" -> "1275 transpose_23"  [label="(4, 12, 64, 32)", style=solid];
+"1275 transpose_23" -> "1276 reshape_52"  [label="(4, 64, 12, 32)", style=solid];
+"1276 reshape_52" -> "1277 reshape_52_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1277 reshape_52_0_0_nncf_smooth_quant_0" -> "1278 quantize_per_tensor_default_72"  [label="(4, 64, 384)", style=solid];
+"1278 quantize_per_tensor_default_72" -> "1279 dequantize_per_tensor_default_72"  [label="(4, 64, 384)", style=solid];
+"1279 dequantize_per_tensor_default_72" -> "1286 linear_71"  [label="(4, 64, 384)", style=solid];
+"1280 linear_71_scale_0" -> "1283 quantize_per_channel_default_60"  [label="(384,)", style=solid];
+"1280 linear_71_scale_0" -> "1284 dequantize_per_channel_default_60"  [label="(384,)", style=solid];
+"1281 linear_71_zero_point_0" -> "1283 quantize_per_channel_default_60"  [label="(384,)", style=solid];
+"1281 linear_71_zero_point_0" -> "1284 dequantize_per_channel_default_60"  [label="(384,)", style=solid];
+"1282 compressed_weight_updated_constant48" -> "1283 quantize_per_channel_default_60"  [label="(384, 384)", style=solid];
+"1283 quantize_per_channel_default_60" -> "1284 dequantize_per_channel_default_60"  [label="(384, 384)", style=solid];
+"1284 dequantize_per_channel_default_60" -> "1286 linear_71"  [label="(384, 384)", style=solid];
+"1285 features_5_7_attn_proj_bias_0_0" -> "1286 linear_71"  [label="(384,)", style=solid];
+"1286 linear_71" -> "1287 dropout_45"  [label="(4, 64, 384)", style=solid];
+"1287 dropout_45" -> "1288 view_65"  [label="(4, 64, 384)", style=solid];
+"1288 view_65" -> "1289 permute_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1289 permute_54" -> "1290 reshape_53"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1290 reshape_53" -> "1291 roll_11"  [label="(1, 16, 16, 384)", style=solid];
+"1291 roll_11" -> "1292 slice_185"  [label="(1, 16, 16, 384)", style=solid];
+"1292 slice_185" -> "1293 slice_186"  [label="(1, 16, 16, 384)", style=solid];
+"1293 slice_186" -> "1294 slice_187"  [label="(1, 14, 16, 384)", style=solid];
+"1294 slice_187" -> "1295 slice_188"  [label="(1, 14, 14, 384)", style=solid];
+"1295 slice_188" -> "1296 contiguous_21"  [label="(1, 14, 14, 384)", style=solid];
+"1296 contiguous_21" -> "1297 layer_norm_25"  [label="(1, 14, 14, 384)", style=solid];
+"1297 layer_norm_25" -> "1298 add_40"  [label="(1, 14, 14, 384)", style=solid];
+"1298 add_40" -> "1299 add_40_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1298 add_40" -> "1323 add_41"  [label="(1, 14, 14, 384)", style=solid];
+"1299 add_40_0_0_nncf_smooth_quant_0" -> "1300 quantize_per_tensor_default_73"  [label="(1, 14, 14, 384)", style=solid];
+"1300 quantize_per_tensor_default_73" -> "1301 dequantize_per_tensor_default_73"  [label="(1, 14, 14, 384)", style=solid];
+"1301 dequantize_per_tensor_default_73" -> "1308 linear_72"  [label="(1, 14, 14, 384)", style=solid];
+"1302 linear_72_scale_0" -> "1305 quantize_per_channel_default_61"  [label="(1536,)", style=solid];
+"1302 linear_72_scale_0" -> "1306 dequantize_per_channel_default_61"  [label="(1536,)", style=solid];
+"1303 linear_72_zero_point_0" -> "1305 quantize_per_channel_default_61"  [label="(1536,)", style=solid];
+"1303 linear_72_zero_point_0" -> "1306 dequantize_per_channel_default_61"  [label="(1536,)", style=solid];
+"1304 compressed_weight_updated_constant49" -> "1305 quantize_per_channel_default_61"  [label="(1536, 384)", style=solid];
+"1305 quantize_per_channel_default_61" -> "1306 dequantize_per_channel_default_61"  [label="(1536, 384)", style=solid];
+"1306 dequantize_per_channel_default_61" -> "1308 linear_72"  [label="(1536, 384)", style=solid];
+"1307 features_5_7_mlp_0_bias_0_0" -> "1308 linear_72"  [label="(1536,)", style=solid];
+"1308 linear_72" -> "1309 gelu_11"  [label="(1, 14, 14, 1536)", style=solid];
+"1309 gelu_11" -> "1310 dropout_46"  [label="(1, 14, 14, 1536)", style=solid];
+"1310 dropout_46" -> "1311 dropout_46_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1311 dropout_46_0_0_nncf_smooth_quant_0" -> "1312 quantize_per_tensor_default_74"  [label="(1, 14, 14, 1536)", style=solid];
+"1312 quantize_per_tensor_default_74" -> "1313 dequantize_per_tensor_default_74"  [label="(1, 14, 14, 1536)", style=solid];
+"1313 dequantize_per_tensor_default_74" -> "1320 linear_73"  [label="(1, 14, 14, 1536)", style=solid];
+"1314 linear_73_scale_0" -> "1317 quantize_per_channel_default_62"  [label="(384,)", style=solid];
+"1314 linear_73_scale_0" -> "1318 dequantize_per_channel_default_62"  [label="(384,)", style=solid];
+"1315 linear_73_zero_point_0" -> "1317 quantize_per_channel_default_62"  [label="(384,)", style=solid];
+"1315 linear_73_zero_point_0" -> "1318 dequantize_per_channel_default_62"  [label="(384,)", style=solid];
+"1316 compressed_weight_updated_constant50" -> "1317 quantize_per_channel_default_62"  [label="(384, 1536)", style=solid];
+"1317 quantize_per_channel_default_62" -> "1318 dequantize_per_channel_default_62"  [label="(384, 1536)", style=solid];
+"1318 dequantize_per_channel_default_62" -> "1320 linear_73"  [label="(384, 1536)", style=solid];
+"1319 features_5_7_mlp_3_bias_0_0" -> "1320 linear_73"  [label="(384,)", style=solid];
+"1320 linear_73" -> "1321 dropout_47"  [label="(1, 14, 14, 384)", style=solid];
+"1321 dropout_47" -> "1322 layer_norm_26"  [label="(1, 14, 14, 384)", style=solid];
+"1322 layer_norm_26" -> "1323 add_41"  [label="(1, 14, 14, 384)", style=solid];
+"1323 add_41" -> "1325 pad_14"  [label="(1, 14, 14, 384)", style=solid];
+"1323 add_41" -> "1386 add_43"  [label="(1, 14, 14, 384)", style=solid];
+"1324 _frozen_param102" -> "1360 add_42"  [label="(1, 12, 64, 64)", style=solid];
+"1325 pad_14" -> "1326 view_68"  [label="(1, 16, 16, 384)", style=solid];
+"1326 view_68" -> "1327 permute_56"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1327 permute_56" -> "1328 reshape_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1328 reshape_54" -> "1329 reshape_54_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1329 reshape_54_0_0_nncf_smooth_quant_0" -> "1330 quantize_per_tensor_default_75"  [label="(4, 64, 384)", style=solid];
+"1330 quantize_per_tensor_default_75" -> "1331 dequantize_per_tensor_default_75"  [label="(4, 64, 384)", style=solid];
+"1331 dequantize_per_tensor_default_75" -> "1338 linear_76"  [label="(4, 64, 384)", style=solid];
+"1332 linear_76_scale_0" -> "1335 quantize_per_channel_default_64"  [label="(1152,)", style=solid];
+"1332 linear_76_scale_0" -> "1336 dequantize_per_channel_default_64"  [label="(1152,)", style=solid];
+"1333 linear_76_zero_point_0" -> "1335 quantize_per_channel_default_64"  [label="(1152,)", style=solid];
+"1333 linear_76_zero_point_0" -> "1336 dequantize_per_channel_default_64"  [label="(1152,)", style=solid];
+"1334 compressed_weight_updated_constant51" -> "1335 quantize_per_channel_default_64"  [label="(1152, 384)", style=solid];
+"1335 quantize_per_channel_default_64" -> "1336 dequantize_per_channel_default_64"  [label="(1152, 384)", style=solid];
+"1336 dequantize_per_channel_default_64" -> "1338 linear_76"  [label="(1152, 384)", style=solid];
+"1337 _frozen_param103_0_0" -> "1338 linear_76"  [label="(1152,)", style=solid];
+"1338 linear_76" -> "1339 reshape_55"  [label="(4, 64, 1152)", style=solid];
+"1339 reshape_55" -> "1340 permute_57"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1340 permute_57" -> "1341 select_36"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1340 permute_57" -> "1342 select_37"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1340 permute_57" -> "1343 select_38"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1341 select_36" -> "1344 linalg_vector_norm_24"  [label="(4, 12, 64, 32)", style=solid];
+"1341 select_36" -> "1346 expand_as_24"  [label="(4, 12, 64, 32)", style=solid];
+"1341 select_36" -> "1347 div_24"  [label="(4, 12, 64, 32)", style=solid];
+"1342 select_37" -> "1350 linalg_vector_norm_25"  [label="(4, 12, 64, 32)", style=solid];
+"1342 select_37" -> "1352 expand_as_25"  [label="(4, 12, 64, 32)", style=solid];
+"1342 select_37" -> "1353 div_25"  [label="(4, 12, 64, 32)", style=solid];
+"1343 select_38" -> "1363 matmul_25"  [label="(4, 12, 64, 32)", style=solid];
+"1344 linalg_vector_norm_24" -> "1345 clamp_min_24"  [label="(4, 12, 64, 1)", style=solid];
+"1345 clamp_min_24" -> "1346 expand_as_24"  [label="(4, 12, 64, 1)", style=solid];
+"1346 expand_as_24" -> "1347 div_24"  [label="(4, 12, 64, 32)", style=solid];
+"1347 div_24" -> "1348 quantize_per_tensor_default_76"  [label="(4, 12, 64, 32)", style=solid];
+"1348 quantize_per_tensor_default_76" -> "1349 dequantize_per_tensor_default_76"  [label="(4, 12, 64, 32)", style=solid];
+"1349 dequantize_per_tensor_default_76" -> "1357 matmul_24"  [label="(4, 12, 64, 32)", style=solid];
+"1350 linalg_vector_norm_25" -> "1351 clamp_min_25"  [label="(4, 12, 64, 1)", style=solid];
+"1351 clamp_min_25" -> "1352 expand_as_25"  [label="(4, 12, 64, 1)", style=solid];
+"1352 expand_as_25" -> "1353 div_25"  [label="(4, 12, 64, 32)", style=solid];
+"1353 div_25" -> "1354 quantize_per_tensor_default_77"  [label="(4, 12, 64, 32)", style=solid];
+"1354 quantize_per_tensor_default_77" -> "1355 dequantize_per_tensor_default_77"  [label="(4, 12, 64, 32)", style=solid];
+"1355 dequantize_per_tensor_default_77" -> "1356 transpose_24"  [label="(4, 12, 64, 32)", style=solid];
+"1356 transpose_24" -> "1357 matmul_24"  [label="(4, 12, 32, 64)", style=solid];
+"1357 matmul_24" -> "1359 mul_25"  [label="(4, 12, 64, 64)", style=solid];
+"1358 _frozen_param105" -> "1359 mul_25"  [label="(12, 1, 1)", style=solid];
+"1359 mul_25" -> "1360 add_42"  [label="(4, 12, 64, 64)", style=solid];
+"1360 add_42" -> "1361 softmax_12"  [label="(4, 12, 64, 64)", style=solid];
+"1361 softmax_12" -> "1362 dropout_48"  [label="(4, 12, 64, 64)", style=solid];
+"1362 dropout_48" -> "1363 matmul_25"  [label="(4, 12, 64, 64)", style=solid];
+"1363 matmul_25" -> "1364 transpose_25"  [label="(4, 12, 64, 32)", style=solid];
+"1364 transpose_25" -> "1365 reshape_56"  [label="(4, 64, 12, 32)", style=solid];
+"1365 reshape_56" -> "1366 reshape_56_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1366 reshape_56_0_0_nncf_smooth_quant_0" -> "1367 quantize_per_tensor_default_78"  [label="(4, 64, 384)", style=solid];
+"1367 quantize_per_tensor_default_78" -> "1368 dequantize_per_tensor_default_78"  [label="(4, 64, 384)", style=solid];
+"1368 dequantize_per_tensor_default_78" -> "1375 linear_77"  [label="(4, 64, 384)", style=solid];
+"1369 linear_77_scale_0" -> "1372 quantize_per_channel_default_65"  [label="(384,)", style=solid];
+"1369 linear_77_scale_0" -> "1373 dequantize_per_channel_default_65"  [label="(384,)", style=solid];
+"1370 linear_77_zero_point_0" -> "1372 quantize_per_channel_default_65"  [label="(384,)", style=solid];
+"1370 linear_77_zero_point_0" -> "1373 dequantize_per_channel_default_65"  [label="(384,)", style=solid];
+"1371 compressed_weight_updated_constant52" -> "1372 quantize_per_channel_default_65"  [label="(384, 384)", style=solid];
+"1372 quantize_per_channel_default_65" -> "1373 dequantize_per_channel_default_65"  [label="(384, 384)", style=solid];
+"1373 dequantize_per_channel_default_65" -> "1375 linear_77"  [label="(384, 384)", style=solid];
+"1374 features_5_8_attn_proj_bias_0_0" -> "1375 linear_77"  [label="(384,)", style=solid];
+"1375 linear_77" -> "1376 dropout_49"  [label="(4, 64, 384)", style=solid];
+"1376 dropout_49" -> "1377 view_69"  [label="(4, 64, 384)", style=solid];
+"1377 view_69" -> "1378 permute_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1378 permute_58" -> "1379 reshape_57"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1379 reshape_57" -> "1380 slice_190"  [label="(1, 16, 16, 384)", style=solid];
+"1380 slice_190" -> "1381 slice_191"  [label="(1, 16, 16, 384)", style=solid];
+"1381 slice_191" -> "1382 slice_192"  [label="(1, 14, 16, 384)", style=solid];
+"1382 slice_192" -> "1383 slice_193"  [label="(1, 14, 14, 384)", style=solid];
+"1383 slice_193" -> "1384 contiguous_23"  [label="(1, 14, 14, 384)", style=solid];
+"1384 contiguous_23" -> "1385 layer_norm_27"  [label="(1, 14, 14, 384)", style=solid];
+"1385 layer_norm_27" -> "1386 add_43"  [label="(1, 14, 14, 384)", style=solid];
+"1386 add_43" -> "1387 add_43_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1386 add_43" -> "1411 add_44"  [label="(1, 14, 14, 384)", style=solid];
+"1387 add_43_0_0_nncf_smooth_quant_0" -> "1388 quantize_per_tensor_default_79"  [label="(1, 14, 14, 384)", style=solid];
+"1388 quantize_per_tensor_default_79" -> "1389 dequantize_per_tensor_default_79"  [label="(1, 14, 14, 384)", style=solid];
+"1389 dequantize_per_tensor_default_79" -> "1396 linear_78"  [label="(1, 14, 14, 384)", style=solid];
+"1390 linear_78_scale_0" -> "1393 quantize_per_channel_default_66"  [label="(1536,)", style=solid];
+"1390 linear_78_scale_0" -> "1394 dequantize_per_channel_default_66"  [label="(1536,)", style=solid];
+"1391 linear_78_zero_point_0" -> "1393 quantize_per_channel_default_66"  [label="(1536,)", style=solid];
+"1391 linear_78_zero_point_0" -> "1394 dequantize_per_channel_default_66"  [label="(1536,)", style=solid];
+"1392 compressed_weight_updated_constant53" -> "1393 quantize_per_channel_default_66"  [label="(1536, 384)", style=solid];
+"1393 quantize_per_channel_default_66" -> "1394 dequantize_per_channel_default_66"  [label="(1536, 384)", style=solid];
+"1394 dequantize_per_channel_default_66" -> "1396 linear_78"  [label="(1536, 384)", style=solid];
+"1395 features_5_8_mlp_0_bias_0_0" -> "1396 linear_78"  [label="(1536,)", style=solid];
+"1396 linear_78" -> "1397 gelu_12"  [label="(1, 14, 14, 1536)", style=solid];
+"1397 gelu_12" -> "1398 dropout_50"  [label="(1, 14, 14, 1536)", style=solid];
+"1398 dropout_50" -> "1399 dropout_50_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1399 dropout_50_0_0_nncf_smooth_quant_0" -> "1400 quantize_per_tensor_default_80"  [label="(1, 14, 14, 1536)", style=solid];
+"1400 quantize_per_tensor_default_80" -> "1401 dequantize_per_tensor_default_80"  [label="(1, 14, 14, 1536)", style=solid];
+"1401 dequantize_per_tensor_default_80" -> "1408 linear_79"  [label="(1, 14, 14, 1536)", style=solid];
+"1402 linear_79_scale_0" -> "1405 quantize_per_channel_default_67"  [label="(384,)", style=solid];
+"1402 linear_79_scale_0" -> "1406 dequantize_per_channel_default_67"  [label="(384,)", style=solid];
+"1403 linear_79_zero_point_0" -> "1405 quantize_per_channel_default_67"  [label="(384,)", style=solid];
+"1403 linear_79_zero_point_0" -> "1406 dequantize_per_channel_default_67"  [label="(384,)", style=solid];
+"1404 compressed_weight_updated_constant54" -> "1405 quantize_per_channel_default_67"  [label="(384, 1536)", style=solid];
+"1405 quantize_per_channel_default_67" -> "1406 dequantize_per_channel_default_67"  [label="(384, 1536)", style=solid];
+"1406 dequantize_per_channel_default_67" -> "1408 linear_79"  [label="(384, 1536)", style=solid];
+"1407 features_5_8_mlp_3_bias_0_0" -> "1408 linear_79"  [label="(384,)", style=solid];
+"1408 linear_79" -> "1409 dropout_51"  [label="(1, 14, 14, 384)", style=solid];
+"1409 dropout_51" -> "1410 layer_norm_28"  [label="(1, 14, 14, 384)", style=solid];
+"1410 layer_norm_28" -> "1411 add_44"  [label="(1, 14, 14, 384)", style=solid];
+"1411 add_44" -> "1413 pad_15"  [label="(1, 14, 14, 384)", style=solid];
+"1411 add_44" -> "1492 add_47"  [label="(1, 14, 14, 384)", style=solid];
+"1412 _frozen_param106" -> "1449 add_45"  [label="(1, 12, 64, 64)", style=solid];
+"1413 pad_15" -> "1414 roll_12"  [label="(1, 16, 16, 384)", style=solid];
+"1414 roll_12" -> "1415 view_72"  [label="(1, 16, 16, 384)", style=solid];
+"1415 view_72" -> "1416 permute_60"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1416 permute_60" -> "1417 reshape_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1417 reshape_58" -> "1418 reshape_58_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1417 reshape_58" -> "1450 new_zeros_6"  [label="(4, 64, 384)", style=solid];
+"1418 reshape_58_0_0_nncf_smooth_quant_0" -> "1419 quantize_per_tensor_default_81"  [label="(4, 64, 384)", style=solid];
+"1419 quantize_per_tensor_default_81" -> "1420 dequantize_per_tensor_default_81"  [label="(4, 64, 384)", style=solid];
+"1420 dequantize_per_tensor_default_81" -> "1427 linear_82"  [label="(4, 64, 384)", style=solid];
+"1421 linear_82_scale_0" -> "1424 quantize_per_channel_default_69"  [label="(1152,)", style=solid];
+"1421 linear_82_scale_0" -> "1425 dequantize_per_channel_default_69"  [label="(1152,)", style=solid];
+"1422 linear_82_zero_point_0" -> "1424 quantize_per_channel_default_69"  [label="(1152,)", style=solid];
+"1422 linear_82_zero_point_0" -> "1425 dequantize_per_channel_default_69"  [label="(1152,)", style=solid];
+"1423 compressed_weight_updated_constant55" -> "1424 quantize_per_channel_default_69"  [label="(1152, 384)", style=solid];
+"1424 quantize_per_channel_default_69" -> "1425 dequantize_per_channel_default_69"  [label="(1152, 384)", style=solid];
+"1425 dequantize_per_channel_default_69" -> "1427 linear_82"  [label="(1152, 384)", style=solid];
+"1426 _frozen_param107_0_0" -> "1427 linear_82"  [label="(1152,)", style=solid];
+"1427 linear_82" -> "1428 reshape_59"  [label="(4, 64, 1152)", style=solid];
+"1428 reshape_59" -> "1429 permute_61"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1429 permute_61" -> "1430 select_39"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1429 permute_61" -> "1431 select_40"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1429 permute_61" -> "1432 select_41"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1430 select_39" -> "1433 linalg_vector_norm_26"  [label="(4, 12, 64, 32)", style=solid];
+"1430 select_39" -> "1435 expand_as_26"  [label="(4, 12, 64, 32)", style=solid];
+"1430 select_39" -> "1436 div_26"  [label="(4, 12, 64, 32)", style=solid];
+"1431 select_40" -> "1439 linalg_vector_norm_27"  [label="(4, 12, 64, 32)", style=solid];
+"1431 select_40" -> "1441 expand_as_27"  [label="(4, 12, 64, 32)", style=solid];
+"1431 select_40" -> "1442 div_27"  [label="(4, 12, 64, 32)", style=solid];
+"1432 select_41" -> "1468 matmul_27"  [label="(4, 12, 64, 32)", style=solid];
+"1433 linalg_vector_norm_26" -> "1434 clamp_min_26"  [label="(4, 12, 64, 1)", style=solid];
+"1434 clamp_min_26" -> "1435 expand_as_26"  [label="(4, 12, 64, 1)", style=solid];
+"1435 expand_as_26" -> "1436 div_26"  [label="(4, 12, 64, 32)", style=solid];
+"1436 div_26" -> "1437 quantize_per_tensor_default_82"  [label="(4, 12, 64, 32)", style=solid];
+"1437 quantize_per_tensor_default_82" -> "1438 dequantize_per_tensor_default_82"  [label="(4, 12, 64, 32)", style=solid];
+"1438 dequantize_per_tensor_default_82" -> "1446 matmul_26"  [label="(4, 12, 64, 32)", style=solid];
+"1439 linalg_vector_norm_27" -> "1440 clamp_min_27"  [label="(4, 12, 64, 1)", style=solid];
+"1440 clamp_min_27" -> "1441 expand_as_27"  [label="(4, 12, 64, 1)", style=solid];
+"1441 expand_as_27" -> "1442 div_27"  [label="(4, 12, 64, 32)", style=solid];
+"1442 div_27" -> "1443 quantize_per_tensor_default_83"  [label="(4, 12, 64, 32)", style=solid];
+"1443 quantize_per_tensor_default_83" -> "1444 dequantize_per_tensor_default_83"  [label="(4, 12, 64, 32)", style=solid];
+"1444 dequantize_per_tensor_default_83" -> "1445 transpose_26"  [label="(4, 12, 64, 32)", style=solid];
+"1445 transpose_26" -> "1446 matmul_26"  [label="(4, 12, 32, 64)", style=solid];
+"1446 matmul_26" -> "1448 mul_27"  [label="(4, 12, 64, 64)", style=solid];
+"1447 _frozen_param109" -> "1448 mul_27"  [label="(12, 1, 1)", style=solid];
+"1448 mul_27" -> "1449 add_45"  [label="(4, 12, 64, 64)", style=solid];
+"1449 add_45" -> "1461 view_74"  [label="(4, 12, 64, 64)", style=solid];
+"1450 new_zeros_6" -> "1451 view_73"  [label="(16, 16)", style=solid];
+"1451 view_73" -> "1452 permute_62"  [label="(2, 8, 2, 8)", style=solid];
+"1452 permute_62" -> "1453 reshape_60"  [label="(2, 2, 8, 8)", style=solid];
+"1453 reshape_60" -> "1454 unsqueeze_38"  [label="(4, 64)", style=solid];
+"1453 reshape_60" -> "1455 unsqueeze_39"  [label="(4, 64)", style=solid];
+"1454 unsqueeze_38" -> "1456 sub_6"  [label="(4, 1, 64)", style=solid];
+"1455 unsqueeze_39" -> "1456 sub_6"  [label="(4, 64, 1)", style=solid];
+"1456 sub_6" -> "1457 ne_6"  [label="(4, 64, 64)", style=solid];
+"1456 sub_6" -> "1458 masked_fill_12"  [label="(4, 64, 64)", style=solid];
+"1456 sub_6" -> "1459 eq_6"  [label="(4, 64, 64)", style=solid];
+"1457 ne_6" -> "1458 masked_fill_12"  [label="(4, 64, 64)", style=solid];
+"1458 masked_fill_12" -> "1460 masked_fill_13"  [label="(4, 64, 64)", style=solid];
+"1459 eq_6" -> "1460 masked_fill_13"  [label="(4, 64, 64)", style=solid];
+"1460 masked_fill_13" -> "1462 unsqueeze_40"  [label="(4, 64, 64)", style=solid];
+"1461 view_74" -> "1464 add_46"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1462 unsqueeze_40" -> "1463 unsqueeze_41"  [label="(4, 1, 64, 64)", style=solid];
+"1463 unsqueeze_41" -> "1464 add_46"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1464 add_46" -> "1465 view_75"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1465 view_75" -> "1466 softmax_13"  [label="(4, 12, 64, 64)", style=solid];
+"1466 softmax_13" -> "1467 dropout_52"  [label="(4, 12, 64, 64)", style=solid];
+"1467 dropout_52" -> "1468 matmul_27"  [label="(4, 12, 64, 64)", style=solid];
+"1468 matmul_27" -> "1469 transpose_27"  [label="(4, 12, 64, 32)", style=solid];
+"1469 transpose_27" -> "1470 reshape_61"  [label="(4, 64, 12, 32)", style=solid];
+"1470 reshape_61" -> "1471 reshape_61_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1471 reshape_61_0_0_nncf_smooth_quant_0" -> "1472 quantize_per_tensor_default_84"  [label="(4, 64, 384)", style=solid];
+"1472 quantize_per_tensor_default_84" -> "1473 dequantize_per_tensor_default_84"  [label="(4, 64, 384)", style=solid];
+"1473 dequantize_per_tensor_default_84" -> "1480 linear_83"  [label="(4, 64, 384)", style=solid];
+"1474 linear_83_scale_0" -> "1477 quantize_per_channel_default_70"  [label="(384,)", style=solid];
+"1474 linear_83_scale_0" -> "1478 dequantize_per_channel_default_70"  [label="(384,)", style=solid];
+"1475 linear_83_zero_point_0" -> "1477 quantize_per_channel_default_70"  [label="(384,)", style=solid];
+"1475 linear_83_zero_point_0" -> "1478 dequantize_per_channel_default_70"  [label="(384,)", style=solid];
+"1476 compressed_weight_updated_constant56" -> "1477 quantize_per_channel_default_70"  [label="(384, 384)", style=solid];
+"1477 quantize_per_channel_default_70" -> "1478 dequantize_per_channel_default_70"  [label="(384, 384)", style=solid];
+"1478 dequantize_per_channel_default_70" -> "1480 linear_83"  [label="(384, 384)", style=solid];
+"1479 features_5_9_attn_proj_bias_0_0" -> "1480 linear_83"  [label="(384,)", style=solid];
+"1480 linear_83" -> "1481 dropout_53"  [label="(4, 64, 384)", style=solid];
+"1481 dropout_53" -> "1482 view_76"  [label="(4, 64, 384)", style=solid];
+"1482 view_76" -> "1483 permute_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1483 permute_63" -> "1484 reshape_62"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1484 reshape_62" -> "1485 roll_13"  [label="(1, 16, 16, 384)", style=solid];
+"1485 roll_13" -> "1486 slice_213"  [label="(1, 16, 16, 384)", style=solid];
+"1486 slice_213" -> "1487 slice_214"  [label="(1, 16, 16, 384)", style=solid];
+"1487 slice_214" -> "1488 slice_215"  [label="(1, 14, 16, 384)", style=solid];
+"1488 slice_215" -> "1489 slice_216"  [label="(1, 14, 14, 384)", style=solid];
+"1489 slice_216" -> "1490 contiguous_25"  [label="(1, 14, 14, 384)", style=solid];
+"1490 contiguous_25" -> "1491 layer_norm_29"  [label="(1, 14, 14, 384)", style=solid];
+"1491 layer_norm_29" -> "1492 add_47"  [label="(1, 14, 14, 384)", style=solid];
+"1492 add_47" -> "1493 add_47_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1492 add_47" -> "1517 add_48"  [label="(1, 14, 14, 384)", style=solid];
+"1493 add_47_0_0_nncf_smooth_quant_0" -> "1494 quantize_per_tensor_default_85"  [label="(1, 14, 14, 384)", style=solid];
+"1494 quantize_per_tensor_default_85" -> "1495 dequantize_per_tensor_default_85"  [label="(1, 14, 14, 384)", style=solid];
+"1495 dequantize_per_tensor_default_85" -> "1502 linear_84"  [label="(1, 14, 14, 384)", style=solid];
+"1496 linear_84_scale_0" -> "1499 quantize_per_channel_default_71"  [label="(1536,)", style=solid];
+"1496 linear_84_scale_0" -> "1500 dequantize_per_channel_default_71"  [label="(1536,)", style=solid];
+"1497 linear_84_zero_point_0" -> "1499 quantize_per_channel_default_71"  [label="(1536,)", style=solid];
+"1497 linear_84_zero_point_0" -> "1500 dequantize_per_channel_default_71"  [label="(1536,)", style=solid];
+"1498 compressed_weight_updated_constant57" -> "1499 quantize_per_channel_default_71"  [label="(1536, 384)", style=solid];
+"1499 quantize_per_channel_default_71" -> "1500 dequantize_per_channel_default_71"  [label="(1536, 384)", style=solid];
+"1500 dequantize_per_channel_default_71" -> "1502 linear_84"  [label="(1536, 384)", style=solid];
+"1501 features_5_9_mlp_0_bias_0_0" -> "1502 linear_84"  [label="(1536,)", style=solid];
+"1502 linear_84" -> "1503 gelu_13"  [label="(1, 14, 14, 1536)", style=solid];
+"1503 gelu_13" -> "1504 dropout_54"  [label="(1, 14, 14, 1536)", style=solid];
+"1504 dropout_54" -> "1505 dropout_54_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1505 dropout_54_0_0_nncf_smooth_quant_0" -> "1506 quantize_per_tensor_default_86"  [label="(1, 14, 14, 1536)", style=solid];
+"1506 quantize_per_tensor_default_86" -> "1507 dequantize_per_tensor_default_86"  [label="(1, 14, 14, 1536)", style=solid];
+"1507 dequantize_per_tensor_default_86" -> "1514 linear_85"  [label="(1, 14, 14, 1536)", style=solid];
+"1508 linear_85_scale_0" -> "1511 quantize_per_channel_default_72"  [label="(384,)", style=solid];
+"1508 linear_85_scale_0" -> "1512 dequantize_per_channel_default_72"  [label="(384,)", style=solid];
+"1509 linear_85_zero_point_0" -> "1511 quantize_per_channel_default_72"  [label="(384,)", style=solid];
+"1509 linear_85_zero_point_0" -> "1512 dequantize_per_channel_default_72"  [label="(384,)", style=solid];
+"1510 compressed_weight_updated_constant58" -> "1511 quantize_per_channel_default_72"  [label="(384, 1536)", style=solid];
+"1511 quantize_per_channel_default_72" -> "1512 dequantize_per_channel_default_72"  [label="(384, 1536)", style=solid];
+"1512 dequantize_per_channel_default_72" -> "1514 linear_85"  [label="(384, 1536)", style=solid];
+"1513 features_5_9_mlp_3_bias_0_0" -> "1514 linear_85"  [label="(384,)", style=solid];
+"1514 linear_85" -> "1515 dropout_55"  [label="(1, 14, 14, 384)", style=solid];
+"1515 dropout_55" -> "1516 layer_norm_30"  [label="(1, 14, 14, 384)", style=solid];
+"1516 layer_norm_30" -> "1517 add_48"  [label="(1, 14, 14, 384)", style=solid];
+"1517 add_48" -> "1519 pad_16"  [label="(1, 14, 14, 384)", style=solid];
+"1517 add_48" -> "1580 add_50"  [label="(1, 14, 14, 384)", style=solid];
+"1518 _frozen_param119" -> "1554 add_49"  [label="(1, 12, 64, 64)", style=solid];
+"1519 pad_16" -> "1520 view_79"  [label="(1, 16, 16, 384)", style=solid];
+"1520 view_79" -> "1521 permute_65"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1521 permute_65" -> "1522 reshape_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1522 reshape_63" -> "1523 reshape_63_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1523 reshape_63_0_0_nncf_smooth_quant_0" -> "1524 quantize_per_tensor_default_87"  [label="(4, 64, 384)", style=solid];
+"1524 quantize_per_tensor_default_87" -> "1525 dequantize_per_tensor_default_87"  [label="(4, 64, 384)", style=solid];
+"1525 dequantize_per_tensor_default_87" -> "1532 linear_88"  [label="(4, 64, 384)", style=solid];
+"1526 linear_88_scale_0" -> "1529 quantize_per_channel_default_74"  [label="(1152,)", style=solid];
+"1526 linear_88_scale_0" -> "1530 dequantize_per_channel_default_74"  [label="(1152,)", style=solid];
+"1527 linear_88_zero_point_0" -> "1529 quantize_per_channel_default_74"  [label="(1152,)", style=solid];
+"1527 linear_88_zero_point_0" -> "1530 dequantize_per_channel_default_74"  [label="(1152,)", style=solid];
+"1528 compressed_weight_updated_constant59" -> "1529 quantize_per_channel_default_74"  [label="(1152, 384)", style=solid];
+"1529 quantize_per_channel_default_74" -> "1530 dequantize_per_channel_default_74"  [label="(1152, 384)", style=solid];
+"1530 dequantize_per_channel_default_74" -> "1532 linear_88"  [label="(1152, 384)", style=solid];
+"1531 _frozen_param120_0_0" -> "1532 linear_88"  [label="(1152,)", style=solid];
+"1532 linear_88" -> "1533 reshape_64"  [label="(4, 64, 1152)", style=solid];
+"1533 reshape_64" -> "1534 permute_66"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1534 permute_66" -> "1535 select_42"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1534 permute_66" -> "1536 select_43"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1534 permute_66" -> "1537 select_44"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1535 select_42" -> "1538 linalg_vector_norm_28"  [label="(4, 12, 64, 32)", style=solid];
+"1535 select_42" -> "1540 expand_as_28"  [label="(4, 12, 64, 32)", style=solid];
+"1535 select_42" -> "1541 div_28"  [label="(4, 12, 64, 32)", style=solid];
+"1536 select_43" -> "1544 linalg_vector_norm_29"  [label="(4, 12, 64, 32)", style=solid];
+"1536 select_43" -> "1546 expand_as_29"  [label="(4, 12, 64, 32)", style=solid];
+"1536 select_43" -> "1547 div_29"  [label="(4, 12, 64, 32)", style=solid];
+"1537 select_44" -> "1557 matmul_29"  [label="(4, 12, 64, 32)", style=solid];
+"1538 linalg_vector_norm_28" -> "1539 clamp_min_28"  [label="(4, 12, 64, 1)", style=solid];
+"1539 clamp_min_28" -> "1540 expand_as_28"  [label="(4, 12, 64, 1)", style=solid];
+"1540 expand_as_28" -> "1541 div_28"  [label="(4, 12, 64, 32)", style=solid];
+"1541 div_28" -> "1542 quantize_per_tensor_default_88"  [label="(4, 12, 64, 32)", style=solid];
+"1542 quantize_per_tensor_default_88" -> "1543 dequantize_per_tensor_default_88"  [label="(4, 12, 64, 32)", style=solid];
+"1543 dequantize_per_tensor_default_88" -> "1551 matmul_28"  [label="(4, 12, 64, 32)", style=solid];
+"1544 linalg_vector_norm_29" -> "1545 clamp_min_29"  [label="(4, 12, 64, 1)", style=solid];
+"1545 clamp_min_29" -> "1546 expand_as_29"  [label="(4, 12, 64, 1)", style=solid];
+"1546 expand_as_29" -> "1547 div_29"  [label="(4, 12, 64, 32)", style=solid];
+"1547 div_29" -> "1548 quantize_per_tensor_default_89"  [label="(4, 12, 64, 32)", style=solid];
+"1548 quantize_per_tensor_default_89" -> "1549 dequantize_per_tensor_default_89"  [label="(4, 12, 64, 32)", style=solid];
+"1549 dequantize_per_tensor_default_89" -> "1550 transpose_28"  [label="(4, 12, 64, 32)", style=solid];
+"1550 transpose_28" -> "1551 matmul_28"  [label="(4, 12, 32, 64)", style=solid];
+"1551 matmul_28" -> "1553 mul_29"  [label="(4, 12, 64, 64)", style=solid];
+"1552 _frozen_param122" -> "1553 mul_29"  [label="(12, 1, 1)", style=solid];
+"1553 mul_29" -> "1554 add_49"  [label="(4, 12, 64, 64)", style=solid];
+"1554 add_49" -> "1555 softmax_14"  [label="(4, 12, 64, 64)", style=solid];
+"1555 softmax_14" -> "1556 dropout_56"  [label="(4, 12, 64, 64)", style=solid];
+"1556 dropout_56" -> "1557 matmul_29"  [label="(4, 12, 64, 64)", style=solid];
+"1557 matmul_29" -> "1558 transpose_29"  [label="(4, 12, 64, 32)", style=solid];
+"1558 transpose_29" -> "1559 reshape_65"  [label="(4, 64, 12, 32)", style=solid];
+"1559 reshape_65" -> "1560 reshape_65_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1560 reshape_65_0_0_nncf_smooth_quant_0" -> "1561 quantize_per_tensor_default_90"  [label="(4, 64, 384)", style=solid];
+"1561 quantize_per_tensor_default_90" -> "1562 dequantize_per_tensor_default_90"  [label="(4, 64, 384)", style=solid];
+"1562 dequantize_per_tensor_default_90" -> "1569 linear_89"  [label="(4, 64, 384)", style=solid];
+"1563 linear_89_scale_0" -> "1566 quantize_per_channel_default_75"  [label="(384,)", style=solid];
+"1563 linear_89_scale_0" -> "1567 dequantize_per_channel_default_75"  [label="(384,)", style=solid];
+"1564 linear_89_zero_point_0" -> "1566 quantize_per_channel_default_75"  [label="(384,)", style=solid];
+"1564 linear_89_zero_point_0" -> "1567 dequantize_per_channel_default_75"  [label="(384,)", style=solid];
+"1565 compressed_weight_updated_constant60" -> "1566 quantize_per_channel_default_75"  [label="(384, 384)", style=solid];
+"1566 quantize_per_channel_default_75" -> "1567 dequantize_per_channel_default_75"  [label="(384, 384)", style=solid];
+"1567 dequantize_per_channel_default_75" -> "1569 linear_89"  [label="(384, 384)", style=solid];
+"1568 features_5_10_attn_proj_bias_0_0" -> "1569 linear_89"  [label="(384,)", style=solid];
+"1569 linear_89" -> "1570 dropout_57"  [label="(4, 64, 384)", style=solid];
+"1570 dropout_57" -> "1571 view_80"  [label="(4, 64, 384)", style=solid];
+"1571 view_80" -> "1572 permute_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1572 permute_67" -> "1573 reshape_66"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1573 reshape_66" -> "1574 slice_218"  [label="(1, 16, 16, 384)", style=solid];
+"1574 slice_218" -> "1575 slice_219"  [label="(1, 16, 16, 384)", style=solid];
+"1575 slice_219" -> "1576 slice_220"  [label="(1, 14, 16, 384)", style=solid];
+"1576 slice_220" -> "1577 slice_221"  [label="(1, 14, 14, 384)", style=solid];
+"1577 slice_221" -> "1578 contiguous_27"  [label="(1, 14, 14, 384)", style=solid];
+"1578 contiguous_27" -> "1579 layer_norm_31"  [label="(1, 14, 14, 384)", style=solid];
+"1579 layer_norm_31" -> "1580 add_50"  [label="(1, 14, 14, 384)", style=solid];
+"1580 add_50" -> "1581 add_50_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1580 add_50" -> "1605 add_51"  [label="(1, 14, 14, 384)", style=solid];
+"1581 add_50_0_0_nncf_smooth_quant_0" -> "1582 quantize_per_tensor_default_91"  [label="(1, 14, 14, 384)", style=solid];
+"1582 quantize_per_tensor_default_91" -> "1583 dequantize_per_tensor_default_91"  [label="(1, 14, 14, 384)", style=solid];
+"1583 dequantize_per_tensor_default_91" -> "1590 linear_90"  [label="(1, 14, 14, 384)", style=solid];
+"1584 linear_90_scale_0" -> "1587 quantize_per_channel_default_76"  [label="(1536,)", style=solid];
+"1584 linear_90_scale_0" -> "1588 dequantize_per_channel_default_76"  [label="(1536,)", style=solid];
+"1585 linear_90_zero_point_0" -> "1587 quantize_per_channel_default_76"  [label="(1536,)", style=solid];
+"1585 linear_90_zero_point_0" -> "1588 dequantize_per_channel_default_76"  [label="(1536,)", style=solid];
+"1586 compressed_weight_updated_constant61" -> "1587 quantize_per_channel_default_76"  [label="(1536, 384)", style=solid];
+"1587 quantize_per_channel_default_76" -> "1588 dequantize_per_channel_default_76"  [label="(1536, 384)", style=solid];
+"1588 dequantize_per_channel_default_76" -> "1590 linear_90"  [label="(1536, 384)", style=solid];
+"1589 features_5_10_mlp_0_bias_0_0" -> "1590 linear_90"  [label="(1536,)", style=solid];
+"1590 linear_90" -> "1591 gelu_14"  [label="(1, 14, 14, 1536)", style=solid];
+"1591 gelu_14" -> "1592 dropout_58"  [label="(1, 14, 14, 1536)", style=solid];
+"1592 dropout_58" -> "1593 dropout_58_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1593 dropout_58_0_0_nncf_smooth_quant_0" -> "1594 quantize_per_tensor_default_92"  [label="(1, 14, 14, 1536)", style=solid];
+"1594 quantize_per_tensor_default_92" -> "1595 dequantize_per_tensor_default_92"  [label="(1, 14, 14, 1536)", style=solid];
+"1595 dequantize_per_tensor_default_92" -> "1602 linear_91"  [label="(1, 14, 14, 1536)", style=solid];
+"1596 linear_91_scale_0" -> "1599 quantize_per_channel_default_77"  [label="(384,)", style=solid];
+"1596 linear_91_scale_0" -> "1600 dequantize_per_channel_default_77"  [label="(384,)", style=solid];
+"1597 linear_91_zero_point_0" -> "1599 quantize_per_channel_default_77"  [label="(384,)", style=solid];
+"1597 linear_91_zero_point_0" -> "1600 dequantize_per_channel_default_77"  [label="(384,)", style=solid];
+"1598 compressed_weight_updated_constant62" -> "1599 quantize_per_channel_default_77"  [label="(384, 1536)", style=solid];
+"1599 quantize_per_channel_default_77" -> "1600 dequantize_per_channel_default_77"  [label="(384, 1536)", style=solid];
+"1600 dequantize_per_channel_default_77" -> "1602 linear_91"  [label="(384, 1536)", style=solid];
+"1601 features_5_10_mlp_3_bias_0_0" -> "1602 linear_91"  [label="(384,)", style=solid];
+"1602 linear_91" -> "1603 dropout_59"  [label="(1, 14, 14, 384)", style=solid];
+"1603 dropout_59" -> "1604 layer_norm_32"  [label="(1, 14, 14, 384)", style=solid];
+"1604 layer_norm_32" -> "1605 add_51"  [label="(1, 14, 14, 384)", style=solid];
+"1605 add_51" -> "1607 pad_17"  [label="(1, 14, 14, 384)", style=solid];
+"1605 add_51" -> "1686 add_54"  [label="(1, 14, 14, 384)", style=solid];
+"1606 _frozen_param123" -> "1643 add_52"  [label="(1, 12, 64, 64)", style=solid];
+"1607 pad_17" -> "1608 roll_14"  [label="(1, 16, 16, 384)", style=solid];
+"1608 roll_14" -> "1609 view_83"  [label="(1, 16, 16, 384)", style=solid];
+"1609 view_83" -> "1610 permute_69"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1610 permute_69" -> "1611 reshape_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1611 reshape_67" -> "1612 reshape_67_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1611 reshape_67" -> "1644 new_zeros_7"  [label="(4, 64, 384)", style=solid];
+"1612 reshape_67_0_0_nncf_smooth_quant_0" -> "1613 quantize_per_tensor_default_93"  [label="(4, 64, 384)", style=solid];
+"1613 quantize_per_tensor_default_93" -> "1614 dequantize_per_tensor_default_93"  [label="(4, 64, 384)", style=solid];
+"1614 dequantize_per_tensor_default_93" -> "1621 linear_94"  [label="(4, 64, 384)", style=solid];
+"1615 linear_94_scale_0" -> "1618 quantize_per_channel_default_79"  [label="(1152,)", style=solid];
+"1615 linear_94_scale_0" -> "1619 dequantize_per_channel_default_79"  [label="(1152,)", style=solid];
+"1616 linear_94_zero_point_0" -> "1618 quantize_per_channel_default_79"  [label="(1152,)", style=solid];
+"1616 linear_94_zero_point_0" -> "1619 dequantize_per_channel_default_79"  [label="(1152,)", style=solid];
+"1617 compressed_weight_updated_constant63" -> "1618 quantize_per_channel_default_79"  [label="(1152, 384)", style=solid];
+"1618 quantize_per_channel_default_79" -> "1619 dequantize_per_channel_default_79"  [label="(1152, 384)", style=solid];
+"1619 dequantize_per_channel_default_79" -> "1621 linear_94"  [label="(1152, 384)", style=solid];
+"1620 _frozen_param124_0_0" -> "1621 linear_94"  [label="(1152,)", style=solid];
+"1621 linear_94" -> "1622 reshape_68"  [label="(4, 64, 1152)", style=solid];
+"1622 reshape_68" -> "1623 permute_70"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1623 permute_70" -> "1624 select_45"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1623 permute_70" -> "1625 select_46"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1623 permute_70" -> "1626 select_47"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1624 select_45" -> "1627 linalg_vector_norm_30"  [label="(4, 12, 64, 32)", style=solid];
+"1624 select_45" -> "1629 expand_as_30"  [label="(4, 12, 64, 32)", style=solid];
+"1624 select_45" -> "1630 div_30"  [label="(4, 12, 64, 32)", style=solid];
+"1625 select_46" -> "1633 linalg_vector_norm_31"  [label="(4, 12, 64, 32)", style=solid];
+"1625 select_46" -> "1635 expand_as_31"  [label="(4, 12, 64, 32)", style=solid];
+"1625 select_46" -> "1636 div_31"  [label="(4, 12, 64, 32)", style=solid];
+"1626 select_47" -> "1662 matmul_31"  [label="(4, 12, 64, 32)", style=solid];
+"1627 linalg_vector_norm_30" -> "1628 clamp_min_30"  [label="(4, 12, 64, 1)", style=solid];
+"1628 clamp_min_30" -> "1629 expand_as_30"  [label="(4, 12, 64, 1)", style=solid];
+"1629 expand_as_30" -> "1630 div_30"  [label="(4, 12, 64, 32)", style=solid];
+"1630 div_30" -> "1631 quantize_per_tensor_default_94"  [label="(4, 12, 64, 32)", style=solid];
+"1631 quantize_per_tensor_default_94" -> "1632 dequantize_per_tensor_default_94"  [label="(4, 12, 64, 32)", style=solid];
+"1632 dequantize_per_tensor_default_94" -> "1640 matmul_30"  [label="(4, 12, 64, 32)", style=solid];
+"1633 linalg_vector_norm_31" -> "1634 clamp_min_31"  [label="(4, 12, 64, 1)", style=solid];
+"1634 clamp_min_31" -> "1635 expand_as_31"  [label="(4, 12, 64, 1)", style=solid];
+"1635 expand_as_31" -> "1636 div_31"  [label="(4, 12, 64, 32)", style=solid];
+"1636 div_31" -> "1637 quantize_per_tensor_default_95"  [label="(4, 12, 64, 32)", style=solid];
+"1637 quantize_per_tensor_default_95" -> "1638 dequantize_per_tensor_default_95"  [label="(4, 12, 64, 32)", style=solid];
+"1638 dequantize_per_tensor_default_95" -> "1639 transpose_30"  [label="(4, 12, 64, 32)", style=solid];
+"1639 transpose_30" -> "1640 matmul_30"  [label="(4, 12, 32, 64)", style=solid];
+"1640 matmul_30" -> "1642 mul_31"  [label="(4, 12, 64, 64)", style=solid];
+"1641 _frozen_param126" -> "1642 mul_31"  [label="(12, 1, 1)", style=solid];
+"1642 mul_31" -> "1643 add_52"  [label="(4, 12, 64, 64)", style=solid];
+"1643 add_52" -> "1655 view_85"  [label="(4, 12, 64, 64)", style=solid];
+"1644 new_zeros_7" -> "1645 view_84"  [label="(16, 16)", style=solid];
+"1645 view_84" -> "1646 permute_71"  [label="(2, 8, 2, 8)", style=solid];
+"1646 permute_71" -> "1647 reshape_69"  [label="(2, 2, 8, 8)", style=solid];
+"1647 reshape_69" -> "1648 unsqueeze_44"  [label="(4, 64)", style=solid];
+"1647 reshape_69" -> "1649 unsqueeze_45"  [label="(4, 64)", style=solid];
+"1648 unsqueeze_44" -> "1650 sub_7"  [label="(4, 1, 64)", style=solid];
+"1649 unsqueeze_45" -> "1650 sub_7"  [label="(4, 64, 1)", style=solid];
+"1650 sub_7" -> "1651 ne_7"  [label="(4, 64, 64)", style=solid];
+"1650 sub_7" -> "1652 masked_fill_14"  [label="(4, 64, 64)", style=solid];
+"1650 sub_7" -> "1653 eq_7"  [label="(4, 64, 64)", style=solid];
+"1651 ne_7" -> "1652 masked_fill_14"  [label="(4, 64, 64)", style=solid];
+"1652 masked_fill_14" -> "1654 masked_fill_15"  [label="(4, 64, 64)", style=solid];
+"1653 eq_7" -> "1654 masked_fill_15"  [label="(4, 64, 64)", style=solid];
+"1654 masked_fill_15" -> "1656 unsqueeze_46"  [label="(4, 64, 64)", style=solid];
+"1655 view_85" -> "1658 add_53"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1656 unsqueeze_46" -> "1657 unsqueeze_47"  [label="(4, 1, 64, 64)", style=solid];
+"1657 unsqueeze_47" -> "1658 add_53"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1658 add_53" -> "1659 view_86"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1659 view_86" -> "1660 softmax_15"  [label="(4, 12, 64, 64)", style=solid];
+"1660 softmax_15" -> "1661 dropout_60"  [label="(4, 12, 64, 64)", style=solid];
+"1661 dropout_60" -> "1662 matmul_31"  [label="(4, 12, 64, 64)", style=solid];
+"1662 matmul_31" -> "1663 transpose_31"  [label="(4, 12, 64, 32)", style=solid];
+"1663 transpose_31" -> "1664 reshape_70"  [label="(4, 64, 12, 32)", style=solid];
+"1664 reshape_70" -> "1665 reshape_70_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1665 reshape_70_0_0_nncf_smooth_quant_0" -> "1666 quantize_per_tensor_default_96"  [label="(4, 64, 384)", style=solid];
+"1666 quantize_per_tensor_default_96" -> "1667 dequantize_per_tensor_default_96"  [label="(4, 64, 384)", style=solid];
+"1667 dequantize_per_tensor_default_96" -> "1674 linear_95"  [label="(4, 64, 384)", style=solid];
+"1668 linear_95_scale_0" -> "1671 quantize_per_channel_default_80"  [label="(384,)", style=solid];
+"1668 linear_95_scale_0" -> "1672 dequantize_per_channel_default_80"  [label="(384,)", style=solid];
+"1669 linear_95_zero_point_0" -> "1671 quantize_per_channel_default_80"  [label="(384,)", style=solid];
+"1669 linear_95_zero_point_0" -> "1672 dequantize_per_channel_default_80"  [label="(384,)", style=solid];
+"1670 compressed_weight_updated_constant64" -> "1671 quantize_per_channel_default_80"  [label="(384, 384)", style=solid];
+"1671 quantize_per_channel_default_80" -> "1672 dequantize_per_channel_default_80"  [label="(384, 384)", style=solid];
+"1672 dequantize_per_channel_default_80" -> "1674 linear_95"  [label="(384, 384)", style=solid];
+"1673 features_5_11_attn_proj_bias_0_0" -> "1674 linear_95"  [label="(384,)", style=solid];
+"1674 linear_95" -> "1675 dropout_61"  [label="(4, 64, 384)", style=solid];
+"1675 dropout_61" -> "1676 view_87"  [label="(4, 64, 384)", style=solid];
+"1676 view_87" -> "1677 permute_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1677 permute_72" -> "1678 reshape_71"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1678 reshape_71" -> "1679 roll_15"  [label="(1, 16, 16, 384)", style=solid];
+"1679 roll_15" -> "1680 slice_241"  [label="(1, 16, 16, 384)", style=solid];
+"1680 slice_241" -> "1681 slice_242"  [label="(1, 16, 16, 384)", style=solid];
+"1681 slice_242" -> "1682 slice_243"  [label="(1, 14, 16, 384)", style=solid];
+"1682 slice_243" -> "1683 slice_244"  [label="(1, 14, 14, 384)", style=solid];
+"1683 slice_244" -> "1684 contiguous_29"  [label="(1, 14, 14, 384)", style=solid];
+"1684 contiguous_29" -> "1685 layer_norm_33"  [label="(1, 14, 14, 384)", style=solid];
+"1685 layer_norm_33" -> "1686 add_54"  [label="(1, 14, 14, 384)", style=solid];
+"1686 add_54" -> "1687 add_54_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1686 add_54" -> "1711 add_55"  [label="(1, 14, 14, 384)", style=solid];
+"1687 add_54_0_0_nncf_smooth_quant_0" -> "1688 quantize_per_tensor_default_97"  [label="(1, 14, 14, 384)", style=solid];
+"1688 quantize_per_tensor_default_97" -> "1689 dequantize_per_tensor_default_97"  [label="(1, 14, 14, 384)", style=solid];
+"1689 dequantize_per_tensor_default_97" -> "1696 linear_96"  [label="(1, 14, 14, 384)", style=solid];
+"1690 linear_96_scale_0" -> "1693 quantize_per_channel_default_81"  [label="(1536,)", style=solid];
+"1690 linear_96_scale_0" -> "1694 dequantize_per_channel_default_81"  [label="(1536,)", style=solid];
+"1691 linear_96_zero_point_0" -> "1693 quantize_per_channel_default_81"  [label="(1536,)", style=solid];
+"1691 linear_96_zero_point_0" -> "1694 dequantize_per_channel_default_81"  [label="(1536,)", style=solid];
+"1692 compressed_weight_updated_constant65" -> "1693 quantize_per_channel_default_81"  [label="(1536, 384)", style=solid];
+"1693 quantize_per_channel_default_81" -> "1694 dequantize_per_channel_default_81"  [label="(1536, 384)", style=solid];
+"1694 dequantize_per_channel_default_81" -> "1696 linear_96"  [label="(1536, 384)", style=solid];
+"1695 features_5_11_mlp_0_bias_0_0" -> "1696 linear_96"  [label="(1536,)", style=solid];
+"1696 linear_96" -> "1697 gelu_15"  [label="(1, 14, 14, 1536)", style=solid];
+"1697 gelu_15" -> "1698 dropout_62"  [label="(1, 14, 14, 1536)", style=solid];
+"1698 dropout_62" -> "1699 dropout_62_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1699 dropout_62_0_0_nncf_smooth_quant_0" -> "1700 quantize_per_tensor_default_98"  [label="(1, 14, 14, 1536)", style=solid];
+"1700 quantize_per_tensor_default_98" -> "1701 dequantize_per_tensor_default_98"  [label="(1, 14, 14, 1536)", style=solid];
+"1701 dequantize_per_tensor_default_98" -> "1708 linear_97"  [label="(1, 14, 14, 1536)", style=solid];
+"1702 linear_97_scale_0" -> "1705 quantize_per_channel_default_82"  [label="(384,)", style=solid];
+"1702 linear_97_scale_0" -> "1706 dequantize_per_channel_default_82"  [label="(384,)", style=solid];
+"1703 linear_97_zero_point_0" -> "1705 quantize_per_channel_default_82"  [label="(384,)", style=solid];
+"1703 linear_97_zero_point_0" -> "1706 dequantize_per_channel_default_82"  [label="(384,)", style=solid];
+"1704 compressed_weight_updated_constant66" -> "1705 quantize_per_channel_default_82"  [label="(384, 1536)", style=solid];
+"1705 quantize_per_channel_default_82" -> "1706 dequantize_per_channel_default_82"  [label="(384, 1536)", style=solid];
+"1706 dequantize_per_channel_default_82" -> "1708 linear_97"  [label="(384, 1536)", style=solid];
+"1707 features_5_11_mlp_3_bias_0_0" -> "1708 linear_97"  [label="(384,)", style=solid];
+"1708 linear_97" -> "1709 dropout_63"  [label="(1, 14, 14, 384)", style=solid];
+"1709 dropout_63" -> "1710 layer_norm_34"  [label="(1, 14, 14, 384)", style=solid];
+"1710 layer_norm_34" -> "1711 add_55"  [label="(1, 14, 14, 384)", style=solid];
+"1711 add_55" -> "1713 pad_18"  [label="(1, 14, 14, 384)", style=solid];
+"1711 add_55" -> "1774 add_57"  [label="(1, 14, 14, 384)", style=solid];
+"1712 _frozen_param136" -> "1748 add_56"  [label="(1, 12, 64, 64)", style=solid];
+"1713 pad_18" -> "1714 view_90"  [label="(1, 16, 16, 384)", style=solid];
+"1714 view_90" -> "1715 permute_74"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1715 permute_74" -> "1716 reshape_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1716 reshape_72" -> "1717 reshape_72_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1717 reshape_72_0_0_nncf_smooth_quant_0" -> "1718 quantize_per_tensor_default_99"  [label="(4, 64, 384)", style=solid];
+"1718 quantize_per_tensor_default_99" -> "1719 dequantize_per_tensor_default_99"  [label="(4, 64, 384)", style=solid];
+"1719 dequantize_per_tensor_default_99" -> "1726 linear_100"  [label="(4, 64, 384)", style=solid];
+"1720 linear_100_scale_0" -> "1723 quantize_per_channel_default_84"  [label="(1152,)", style=solid];
+"1720 linear_100_scale_0" -> "1724 dequantize_per_channel_default_84"  [label="(1152,)", style=solid];
+"1721 linear_100_zero_point_0" -> "1723 quantize_per_channel_default_84"  [label="(1152,)", style=solid];
+"1721 linear_100_zero_point_0" -> "1724 dequantize_per_channel_default_84"  [label="(1152,)", style=solid];
+"1722 compressed_weight_updated_constant67" -> "1723 quantize_per_channel_default_84"  [label="(1152, 384)", style=solid];
+"1723 quantize_per_channel_default_84" -> "1724 dequantize_per_channel_default_84"  [label="(1152, 384)", style=solid];
+"1724 dequantize_per_channel_default_84" -> "1726 linear_100"  [label="(1152, 384)", style=solid];
+"1725 _frozen_param137_0_0" -> "1726 linear_100"  [label="(1152,)", style=solid];
+"1726 linear_100" -> "1727 reshape_73"  [label="(4, 64, 1152)", style=solid];
+"1727 reshape_73" -> "1728 permute_75"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1728 permute_75" -> "1729 select_48"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1728 permute_75" -> "1730 select_49"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1728 permute_75" -> "1731 select_50"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1729 select_48" -> "1732 linalg_vector_norm_32"  [label="(4, 12, 64, 32)", style=solid];
+"1729 select_48" -> "1734 expand_as_32"  [label="(4, 12, 64, 32)", style=solid];
+"1729 select_48" -> "1735 div_32"  [label="(4, 12, 64, 32)", style=solid];
+"1730 select_49" -> "1738 linalg_vector_norm_33"  [label="(4, 12, 64, 32)", style=solid];
+"1730 select_49" -> "1740 expand_as_33"  [label="(4, 12, 64, 32)", style=solid];
+"1730 select_49" -> "1741 div_33"  [label="(4, 12, 64, 32)", style=solid];
+"1731 select_50" -> "1751 matmul_33"  [label="(4, 12, 64, 32)", style=solid];
+"1732 linalg_vector_norm_32" -> "1733 clamp_min_32"  [label="(4, 12, 64, 1)", style=solid];
+"1733 clamp_min_32" -> "1734 expand_as_32"  [label="(4, 12, 64, 1)", style=solid];
+"1734 expand_as_32" -> "1735 div_32"  [label="(4, 12, 64, 32)", style=solid];
+"1735 div_32" -> "1736 quantize_per_tensor_default_100"  [label="(4, 12, 64, 32)", style=solid];
+"1736 quantize_per_tensor_default_100" -> "1737 dequantize_per_tensor_default_100"  [label="(4, 12, 64, 32)", style=solid];
+"1737 dequantize_per_tensor_default_100" -> "1745 matmul_32"  [label="(4, 12, 64, 32)", style=solid];
+"1738 linalg_vector_norm_33" -> "1739 clamp_min_33"  [label="(4, 12, 64, 1)", style=solid];
+"1739 clamp_min_33" -> "1740 expand_as_33"  [label="(4, 12, 64, 1)", style=solid];
+"1740 expand_as_33" -> "1741 div_33"  [label="(4, 12, 64, 32)", style=solid];
+"1741 div_33" -> "1742 quantize_per_tensor_default_101"  [label="(4, 12, 64, 32)", style=solid];
+"1742 quantize_per_tensor_default_101" -> "1743 dequantize_per_tensor_default_101"  [label="(4, 12, 64, 32)", style=solid];
+"1743 dequantize_per_tensor_default_101" -> "1744 transpose_32"  [label="(4, 12, 64, 32)", style=solid];
+"1744 transpose_32" -> "1745 matmul_32"  [label="(4, 12, 32, 64)", style=solid];
+"1745 matmul_32" -> "1747 mul_33"  [label="(4, 12, 64, 64)", style=solid];
+"1746 _frozen_param139" -> "1747 mul_33"  [label="(12, 1, 1)", style=solid];
+"1747 mul_33" -> "1748 add_56"  [label="(4, 12, 64, 64)", style=solid];
+"1748 add_56" -> "1749 softmax_16"  [label="(4, 12, 64, 64)", style=solid];
+"1749 softmax_16" -> "1750 dropout_64"  [label="(4, 12, 64, 64)", style=solid];
+"1750 dropout_64" -> "1751 matmul_33"  [label="(4, 12, 64, 64)", style=solid];
+"1751 matmul_33" -> "1752 transpose_33"  [label="(4, 12, 64, 32)", style=solid];
+"1752 transpose_33" -> "1753 reshape_74"  [label="(4, 64, 12, 32)", style=solid];
+"1753 reshape_74" -> "1754 reshape_74_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1754 reshape_74_0_0_nncf_smooth_quant_0" -> "1755 quantize_per_tensor_default_102"  [label="(4, 64, 384)", style=solid];
+"1755 quantize_per_tensor_default_102" -> "1756 dequantize_per_tensor_default_102"  [label="(4, 64, 384)", style=solid];
+"1756 dequantize_per_tensor_default_102" -> "1763 linear_101"  [label="(4, 64, 384)", style=solid];
+"1757 linear_101_scale_0" -> "1760 quantize_per_channel_default_85"  [label="(384,)", style=solid];
+"1757 linear_101_scale_0" -> "1761 dequantize_per_channel_default_85"  [label="(384,)", style=solid];
+"1758 linear_101_zero_point_0" -> "1760 quantize_per_channel_default_85"  [label="(384,)", style=solid];
+"1758 linear_101_zero_point_0" -> "1761 dequantize_per_channel_default_85"  [label="(384,)", style=solid];
+"1759 compressed_weight_updated_constant68" -> "1760 quantize_per_channel_default_85"  [label="(384, 384)", style=solid];
+"1760 quantize_per_channel_default_85" -> "1761 dequantize_per_channel_default_85"  [label="(384, 384)", style=solid];
+"1761 dequantize_per_channel_default_85" -> "1763 linear_101"  [label="(384, 384)", style=solid];
+"1762 features_5_12_attn_proj_bias_0_0" -> "1763 linear_101"  [label="(384,)", style=solid];
+"1763 linear_101" -> "1764 dropout_65"  [label="(4, 64, 384)", style=solid];
+"1764 dropout_65" -> "1765 view_91"  [label="(4, 64, 384)", style=solid];
+"1765 view_91" -> "1766 permute_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1766 permute_76" -> "1767 reshape_75"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1767 reshape_75" -> "1768 slice_246"  [label="(1, 16, 16, 384)", style=solid];
+"1768 slice_246" -> "1769 slice_247"  [label="(1, 16, 16, 384)", style=solid];
+"1769 slice_247" -> "1770 slice_248"  [label="(1, 14, 16, 384)", style=solid];
+"1770 slice_248" -> "1771 slice_249"  [label="(1, 14, 14, 384)", style=solid];
+"1771 slice_249" -> "1772 contiguous_31"  [label="(1, 14, 14, 384)", style=solid];
+"1772 contiguous_31" -> "1773 layer_norm_35"  [label="(1, 14, 14, 384)", style=solid];
+"1773 layer_norm_35" -> "1774 add_57"  [label="(1, 14, 14, 384)", style=solid];
+"1774 add_57" -> "1775 add_57_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1774 add_57" -> "1799 add_58"  [label="(1, 14, 14, 384)", style=solid];
+"1775 add_57_0_0_nncf_smooth_quant_0" -> "1776 quantize_per_tensor_default_103"  [label="(1, 14, 14, 384)", style=solid];
+"1776 quantize_per_tensor_default_103" -> "1777 dequantize_per_tensor_default_103"  [label="(1, 14, 14, 384)", style=solid];
+"1777 dequantize_per_tensor_default_103" -> "1784 linear_102"  [label="(1, 14, 14, 384)", style=solid];
+"1778 linear_102_scale_0" -> "1781 quantize_per_channel_default_86"  [label="(1536,)", style=solid];
+"1778 linear_102_scale_0" -> "1782 dequantize_per_channel_default_86"  [label="(1536,)", style=solid];
+"1779 linear_102_zero_point_0" -> "1781 quantize_per_channel_default_86"  [label="(1536,)", style=solid];
+"1779 linear_102_zero_point_0" -> "1782 dequantize_per_channel_default_86"  [label="(1536,)", style=solid];
+"1780 compressed_weight_updated_constant69" -> "1781 quantize_per_channel_default_86"  [label="(1536, 384)", style=solid];
+"1781 quantize_per_channel_default_86" -> "1782 dequantize_per_channel_default_86"  [label="(1536, 384)", style=solid];
+"1782 dequantize_per_channel_default_86" -> "1784 linear_102"  [label="(1536, 384)", style=solid];
+"1783 features_5_12_mlp_0_bias_0_0" -> "1784 linear_102"  [label="(1536,)", style=solid];
+"1784 linear_102" -> "1785 gelu_16"  [label="(1, 14, 14, 1536)", style=solid];
+"1785 gelu_16" -> "1786 dropout_66"  [label="(1, 14, 14, 1536)", style=solid];
+"1786 dropout_66" -> "1787 dropout_66_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1787 dropout_66_0_0_nncf_smooth_quant_0" -> "1788 quantize_per_tensor_default_104"  [label="(1, 14, 14, 1536)", style=solid];
+"1788 quantize_per_tensor_default_104" -> "1789 dequantize_per_tensor_default_104"  [label="(1, 14, 14, 1536)", style=solid];
+"1789 dequantize_per_tensor_default_104" -> "1796 linear_103"  [label="(1, 14, 14, 1536)", style=solid];
+"1790 linear_103_scale_0" -> "1793 quantize_per_channel_default_87"  [label="(384,)", style=solid];
+"1790 linear_103_scale_0" -> "1794 dequantize_per_channel_default_87"  [label="(384,)", style=solid];
+"1791 linear_103_zero_point_0" -> "1793 quantize_per_channel_default_87"  [label="(384,)", style=solid];
+"1791 linear_103_zero_point_0" -> "1794 dequantize_per_channel_default_87"  [label="(384,)", style=solid];
+"1792 compressed_weight_updated_constant70" -> "1793 quantize_per_channel_default_87"  [label="(384, 1536)", style=solid];
+"1793 quantize_per_channel_default_87" -> "1794 dequantize_per_channel_default_87"  [label="(384, 1536)", style=solid];
+"1794 dequantize_per_channel_default_87" -> "1796 linear_103"  [label="(384, 1536)", style=solid];
+"1795 features_5_12_mlp_3_bias_0_0" -> "1796 linear_103"  [label="(384,)", style=solid];
+"1796 linear_103" -> "1797 dropout_67"  [label="(1, 14, 14, 384)", style=solid];
+"1797 dropout_67" -> "1798 layer_norm_36"  [label="(1, 14, 14, 384)", style=solid];
+"1798 layer_norm_36" -> "1799 add_58"  [label="(1, 14, 14, 384)", style=solid];
+"1799 add_58" -> "1801 pad_19"  [label="(1, 14, 14, 384)", style=solid];
+"1799 add_58" -> "1880 add_61"  [label="(1, 14, 14, 384)", style=solid];
+"1800 _frozen_param140" -> "1837 add_59"  [label="(1, 12, 64, 64)", style=solid];
+"1801 pad_19" -> "1802 roll_16"  [label="(1, 16, 16, 384)", style=solid];
+"1802 roll_16" -> "1803 view_94"  [label="(1, 16, 16, 384)", style=solid];
+"1803 view_94" -> "1804 permute_78"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1804 permute_78" -> "1805 reshape_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1805 reshape_76" -> "1806 reshape_76_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1805 reshape_76" -> "1838 new_zeros_8"  [label="(4, 64, 384)", style=solid];
+"1806 reshape_76_0_0_nncf_smooth_quant_0" -> "1807 quantize_per_tensor_default_105"  [label="(4, 64, 384)", style=solid];
+"1807 quantize_per_tensor_default_105" -> "1808 dequantize_per_tensor_default_105"  [label="(4, 64, 384)", style=solid];
+"1808 dequantize_per_tensor_default_105" -> "1815 linear_106"  [label="(4, 64, 384)", style=solid];
+"1809 linear_106_scale_0" -> "1812 quantize_per_channel_default_89"  [label="(1152,)", style=solid];
+"1809 linear_106_scale_0" -> "1813 dequantize_per_channel_default_89"  [label="(1152,)", style=solid];
+"1810 linear_106_zero_point_0" -> "1812 quantize_per_channel_default_89"  [label="(1152,)", style=solid];
+"1810 linear_106_zero_point_0" -> "1813 dequantize_per_channel_default_89"  [label="(1152,)", style=solid];
+"1811 compressed_weight_updated_constant71" -> "1812 quantize_per_channel_default_89"  [label="(1152, 384)", style=solid];
+"1812 quantize_per_channel_default_89" -> "1813 dequantize_per_channel_default_89"  [label="(1152, 384)", style=solid];
+"1813 dequantize_per_channel_default_89" -> "1815 linear_106"  [label="(1152, 384)", style=solid];
+"1814 _frozen_param141_0_0" -> "1815 linear_106"  [label="(1152,)", style=solid];
+"1815 linear_106" -> "1816 reshape_77"  [label="(4, 64, 1152)", style=solid];
+"1816 reshape_77" -> "1817 permute_79"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1817 permute_79" -> "1818 select_51"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1817 permute_79" -> "1819 select_52"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1817 permute_79" -> "1820 select_53"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1818 select_51" -> "1821 linalg_vector_norm_34"  [label="(4, 12, 64, 32)", style=solid];
+"1818 select_51" -> "1823 expand_as_34"  [label="(4, 12, 64, 32)", style=solid];
+"1818 select_51" -> "1824 div_34"  [label="(4, 12, 64, 32)", style=solid];
+"1819 select_52" -> "1827 linalg_vector_norm_35"  [label="(4, 12, 64, 32)", style=solid];
+"1819 select_52" -> "1829 expand_as_35"  [label="(4, 12, 64, 32)", style=solid];
+"1819 select_52" -> "1830 div_35"  [label="(4, 12, 64, 32)", style=solid];
+"1820 select_53" -> "1856 matmul_35"  [label="(4, 12, 64, 32)", style=solid];
+"1821 linalg_vector_norm_34" -> "1822 clamp_min_34"  [label="(4, 12, 64, 1)", style=solid];
+"1822 clamp_min_34" -> "1823 expand_as_34"  [label="(4, 12, 64, 1)", style=solid];
+"1823 expand_as_34" -> "1824 div_34"  [label="(4, 12, 64, 32)", style=solid];
+"1824 div_34" -> "1825 quantize_per_tensor_default_106"  [label="(4, 12, 64, 32)", style=solid];
+"1825 quantize_per_tensor_default_106" -> "1826 dequantize_per_tensor_default_106"  [label="(4, 12, 64, 32)", style=solid];
+"1826 dequantize_per_tensor_default_106" -> "1834 matmul_34"  [label="(4, 12, 64, 32)", style=solid];
+"1827 linalg_vector_norm_35" -> "1828 clamp_min_35"  [label="(4, 12, 64, 1)", style=solid];
+"1828 clamp_min_35" -> "1829 expand_as_35"  [label="(4, 12, 64, 1)", style=solid];
+"1829 expand_as_35" -> "1830 div_35"  [label="(4, 12, 64, 32)", style=solid];
+"1830 div_35" -> "1831 quantize_per_tensor_default_107"  [label="(4, 12, 64, 32)", style=solid];
+"1831 quantize_per_tensor_default_107" -> "1832 dequantize_per_tensor_default_107"  [label="(4, 12, 64, 32)", style=solid];
+"1832 dequantize_per_tensor_default_107" -> "1833 transpose_34"  [label="(4, 12, 64, 32)", style=solid];
+"1833 transpose_34" -> "1834 matmul_34"  [label="(4, 12, 32, 64)", style=solid];
+"1834 matmul_34" -> "1836 mul_35"  [label="(4, 12, 64, 64)", style=solid];
+"1835 _frozen_param143" -> "1836 mul_35"  [label="(12, 1, 1)", style=solid];
+"1836 mul_35" -> "1837 add_59"  [label="(4, 12, 64, 64)", style=solid];
+"1837 add_59" -> "1849 view_96"  [label="(4, 12, 64, 64)", style=solid];
+"1838 new_zeros_8" -> "1839 view_95"  [label="(16, 16)", style=solid];
+"1839 view_95" -> "1840 permute_80"  [label="(2, 8, 2, 8)", style=solid];
+"1840 permute_80" -> "1841 reshape_78"  [label="(2, 2, 8, 8)", style=solid];
+"1841 reshape_78" -> "1842 unsqueeze_50"  [label="(4, 64)", style=solid];
+"1841 reshape_78" -> "1843 unsqueeze_51"  [label="(4, 64)", style=solid];
+"1842 unsqueeze_50" -> "1844 sub_8"  [label="(4, 1, 64)", style=solid];
+"1843 unsqueeze_51" -> "1844 sub_8"  [label="(4, 64, 1)", style=solid];
+"1844 sub_8" -> "1845 ne_8"  [label="(4, 64, 64)", style=solid];
+"1844 sub_8" -> "1846 masked_fill_16"  [label="(4, 64, 64)", style=solid];
+"1844 sub_8" -> "1847 eq_8"  [label="(4, 64, 64)", style=solid];
+"1845 ne_8" -> "1846 masked_fill_16"  [label="(4, 64, 64)", style=solid];
+"1846 masked_fill_16" -> "1848 masked_fill_17"  [label="(4, 64, 64)", style=solid];
+"1847 eq_8" -> "1848 masked_fill_17"  [label="(4, 64, 64)", style=solid];
+"1848 masked_fill_17" -> "1850 unsqueeze_52"  [label="(4, 64, 64)", style=solid];
+"1849 view_96" -> "1852 add_60"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1850 unsqueeze_52" -> "1851 unsqueeze_53"  [label="(4, 1, 64, 64)", style=solid];
+"1851 unsqueeze_53" -> "1852 add_60"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1852 add_60" -> "1853 view_97"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1853 view_97" -> "1854 softmax_17"  [label="(4, 12, 64, 64)", style=solid];
+"1854 softmax_17" -> "1855 dropout_68"  [label="(4, 12, 64, 64)", style=solid];
+"1855 dropout_68" -> "1856 matmul_35"  [label="(4, 12, 64, 64)", style=solid];
+"1856 matmul_35" -> "1857 transpose_35"  [label="(4, 12, 64, 32)", style=solid];
+"1857 transpose_35" -> "1858 reshape_79"  [label="(4, 64, 12, 32)", style=solid];
+"1858 reshape_79" -> "1859 reshape_79_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1859 reshape_79_0_0_nncf_smooth_quant_0" -> "1860 quantize_per_tensor_default_108"  [label="(4, 64, 384)", style=solid];
+"1860 quantize_per_tensor_default_108" -> "1861 dequantize_per_tensor_default_108"  [label="(4, 64, 384)", style=solid];
+"1861 dequantize_per_tensor_default_108" -> "1868 linear_107"  [label="(4, 64, 384)", style=solid];
+"1862 linear_107_scale_0" -> "1865 quantize_per_channel_default_90"  [label="(384,)", style=solid];
+"1862 linear_107_scale_0" -> "1866 dequantize_per_channel_default_90"  [label="(384,)", style=solid];
+"1863 linear_107_zero_point_0" -> "1865 quantize_per_channel_default_90"  [label="(384,)", style=solid];
+"1863 linear_107_zero_point_0" -> "1866 dequantize_per_channel_default_90"  [label="(384,)", style=solid];
+"1864 compressed_weight_updated_constant72" -> "1865 quantize_per_channel_default_90"  [label="(384, 384)", style=solid];
+"1865 quantize_per_channel_default_90" -> "1866 dequantize_per_channel_default_90"  [label="(384, 384)", style=solid];
+"1866 dequantize_per_channel_default_90" -> "1868 linear_107"  [label="(384, 384)", style=solid];
+"1867 features_5_13_attn_proj_bias_0_0" -> "1868 linear_107"  [label="(384,)", style=solid];
+"1868 linear_107" -> "1869 dropout_69"  [label="(4, 64, 384)", style=solid];
+"1869 dropout_69" -> "1870 view_98"  [label="(4, 64, 384)", style=solid];
+"1870 view_98" -> "1871 permute_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1871 permute_81" -> "1872 reshape_80"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1872 reshape_80" -> "1873 roll_17"  [label="(1, 16, 16, 384)", style=solid];
+"1873 roll_17" -> "1874 slice_269"  [label="(1, 16, 16, 384)", style=solid];
+"1874 slice_269" -> "1875 slice_270"  [label="(1, 16, 16, 384)", style=solid];
+"1875 slice_270" -> "1876 slice_271"  [label="(1, 14, 16, 384)", style=solid];
+"1876 slice_271" -> "1877 slice_272"  [label="(1, 14, 14, 384)", style=solid];
+"1877 slice_272" -> "1878 contiguous_33"  [label="(1, 14, 14, 384)", style=solid];
+"1878 contiguous_33" -> "1879 layer_norm_37"  [label="(1, 14, 14, 384)", style=solid];
+"1879 layer_norm_37" -> "1880 add_61"  [label="(1, 14, 14, 384)", style=solid];
+"1880 add_61" -> "1881 add_61_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1880 add_61" -> "1905 add_62"  [label="(1, 14, 14, 384)", style=solid];
+"1881 add_61_0_0_nncf_smooth_quant_0" -> "1882 quantize_per_tensor_default_109"  [label="(1, 14, 14, 384)", style=solid];
+"1882 quantize_per_tensor_default_109" -> "1883 dequantize_per_tensor_default_109"  [label="(1, 14, 14, 384)", style=solid];
+"1883 dequantize_per_tensor_default_109" -> "1890 linear_108"  [label="(1, 14, 14, 384)", style=solid];
+"1884 linear_108_scale_0" -> "1887 quantize_per_channel_default_91"  [label="(1536,)", style=solid];
+"1884 linear_108_scale_0" -> "1888 dequantize_per_channel_default_91"  [label="(1536,)", style=solid];
+"1885 linear_108_zero_point_0" -> "1887 quantize_per_channel_default_91"  [label="(1536,)", style=solid];
+"1885 linear_108_zero_point_0" -> "1888 dequantize_per_channel_default_91"  [label="(1536,)", style=solid];
+"1886 compressed_weight_updated_constant73" -> "1887 quantize_per_channel_default_91"  [label="(1536, 384)", style=solid];
+"1887 quantize_per_channel_default_91" -> "1888 dequantize_per_channel_default_91"  [label="(1536, 384)", style=solid];
+"1888 dequantize_per_channel_default_91" -> "1890 linear_108"  [label="(1536, 384)", style=solid];
+"1889 features_5_13_mlp_0_bias_0_0" -> "1890 linear_108"  [label="(1536,)", style=solid];
+"1890 linear_108" -> "1891 gelu_17"  [label="(1, 14, 14, 1536)", style=solid];
+"1891 gelu_17" -> "1892 dropout_70"  [label="(1, 14, 14, 1536)", style=solid];
+"1892 dropout_70" -> "1893 dropout_70_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1893 dropout_70_0_0_nncf_smooth_quant_0" -> "1894 quantize_per_tensor_default_110"  [label="(1, 14, 14, 1536)", style=solid];
+"1894 quantize_per_tensor_default_110" -> "1895 dequantize_per_tensor_default_110"  [label="(1, 14, 14, 1536)", style=solid];
+"1895 dequantize_per_tensor_default_110" -> "1902 linear_109"  [label="(1, 14, 14, 1536)", style=solid];
+"1896 linear_109_scale_0" -> "1899 quantize_per_channel_default_92"  [label="(384,)", style=solid];
+"1896 linear_109_scale_0" -> "1900 dequantize_per_channel_default_92"  [label="(384,)", style=solid];
+"1897 linear_109_zero_point_0" -> "1899 quantize_per_channel_default_92"  [label="(384,)", style=solid];
+"1897 linear_109_zero_point_0" -> "1900 dequantize_per_channel_default_92"  [label="(384,)", style=solid];
+"1898 compressed_weight_updated_constant74" -> "1899 quantize_per_channel_default_92"  [label="(384, 1536)", style=solid];
+"1899 quantize_per_channel_default_92" -> "1900 dequantize_per_channel_default_92"  [label="(384, 1536)", style=solid];
+"1900 dequantize_per_channel_default_92" -> "1902 linear_109"  [label="(384, 1536)", style=solid];
+"1901 features_5_13_mlp_3_bias_0_0" -> "1902 linear_109"  [label="(384,)", style=solid];
+"1902 linear_109" -> "1903 dropout_71"  [label="(1, 14, 14, 384)", style=solid];
+"1903 dropout_71" -> "1904 layer_norm_38"  [label="(1, 14, 14, 384)", style=solid];
+"1904 layer_norm_38" -> "1905 add_62"  [label="(1, 14, 14, 384)", style=solid];
+"1905 add_62" -> "1907 pad_20"  [label="(1, 14, 14, 384)", style=solid];
+"1905 add_62" -> "1968 add_64"  [label="(1, 14, 14, 384)", style=solid];
+"1906 _frozen_param153" -> "1942 add_63"  [label="(1, 12, 64, 64)", style=solid];
+"1907 pad_20" -> "1908 view_101"  [label="(1, 16, 16, 384)", style=solid];
+"1908 view_101" -> "1909 permute_83"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1909 permute_83" -> "1910 reshape_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1910 reshape_81" -> "1911 reshape_81_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1911 reshape_81_0_0_nncf_smooth_quant_0" -> "1912 quantize_per_tensor_default_111"  [label="(4, 64, 384)", style=solid];
+"1912 quantize_per_tensor_default_111" -> "1913 dequantize_per_tensor_default_111"  [label="(4, 64, 384)", style=solid];
+"1913 dequantize_per_tensor_default_111" -> "1920 linear_112"  [label="(4, 64, 384)", style=solid];
+"1914 linear_112_scale_0" -> "1917 quantize_per_channel_default_94"  [label="(1152,)", style=solid];
+"1914 linear_112_scale_0" -> "1918 dequantize_per_channel_default_94"  [label="(1152,)", style=solid];
+"1915 linear_112_zero_point_0" -> "1917 quantize_per_channel_default_94"  [label="(1152,)", style=solid];
+"1915 linear_112_zero_point_0" -> "1918 dequantize_per_channel_default_94"  [label="(1152,)", style=solid];
+"1916 compressed_weight_updated_constant75" -> "1917 quantize_per_channel_default_94"  [label="(1152, 384)", style=solid];
+"1917 quantize_per_channel_default_94" -> "1918 dequantize_per_channel_default_94"  [label="(1152, 384)", style=solid];
+"1918 dequantize_per_channel_default_94" -> "1920 linear_112"  [label="(1152, 384)", style=solid];
+"1919 _frozen_param154_0_0" -> "1920 linear_112"  [label="(1152,)", style=solid];
+"1920 linear_112" -> "1921 reshape_82"  [label="(4, 64, 1152)", style=solid];
+"1921 reshape_82" -> "1922 permute_84"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1922 permute_84" -> "1923 select_54"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1922 permute_84" -> "1924 select_55"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1922 permute_84" -> "1925 select_56"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1923 select_54" -> "1926 linalg_vector_norm_36"  [label="(4, 12, 64, 32)", style=solid];
+"1923 select_54" -> "1928 expand_as_36"  [label="(4, 12, 64, 32)", style=solid];
+"1923 select_54" -> "1929 div_36"  [label="(4, 12, 64, 32)", style=solid];
+"1924 select_55" -> "1932 linalg_vector_norm_37"  [label="(4, 12, 64, 32)", style=solid];
+"1924 select_55" -> "1934 expand_as_37"  [label="(4, 12, 64, 32)", style=solid];
+"1924 select_55" -> "1935 div_37"  [label="(4, 12, 64, 32)", style=solid];
+"1925 select_56" -> "1945 matmul_37"  [label="(4, 12, 64, 32)", style=solid];
+"1926 linalg_vector_norm_36" -> "1927 clamp_min_36"  [label="(4, 12, 64, 1)", style=solid];
+"1927 clamp_min_36" -> "1928 expand_as_36"  [label="(4, 12, 64, 1)", style=solid];
+"1928 expand_as_36" -> "1929 div_36"  [label="(4, 12, 64, 32)", style=solid];
+"1929 div_36" -> "1930 quantize_per_tensor_default_112"  [label="(4, 12, 64, 32)", style=solid];
+"1930 quantize_per_tensor_default_112" -> "1931 dequantize_per_tensor_default_112"  [label="(4, 12, 64, 32)", style=solid];
+"1931 dequantize_per_tensor_default_112" -> "1939 matmul_36"  [label="(4, 12, 64, 32)", style=solid];
+"1932 linalg_vector_norm_37" -> "1933 clamp_min_37"  [label="(4, 12, 64, 1)", style=solid];
+"1933 clamp_min_37" -> "1934 expand_as_37"  [label="(4, 12, 64, 1)", style=solid];
+"1934 expand_as_37" -> "1935 div_37"  [label="(4, 12, 64, 32)", style=solid];
+"1935 div_37" -> "1936 quantize_per_tensor_default_113"  [label="(4, 12, 64, 32)", style=solid];
+"1936 quantize_per_tensor_default_113" -> "1937 dequantize_per_tensor_default_113"  [label="(4, 12, 64, 32)", style=solid];
+"1937 dequantize_per_tensor_default_113" -> "1938 transpose_36"  [label="(4, 12, 64, 32)", style=solid];
+"1938 transpose_36" -> "1939 matmul_36"  [label="(4, 12, 32, 64)", style=solid];
+"1939 matmul_36" -> "1941 mul_37"  [label="(4, 12, 64, 64)", style=solid];
+"1940 _frozen_param156" -> "1941 mul_37"  [label="(12, 1, 1)", style=solid];
+"1941 mul_37" -> "1942 add_63"  [label="(4, 12, 64, 64)", style=solid];
+"1942 add_63" -> "1943 softmax_18"  [label="(4, 12, 64, 64)", style=solid];
+"1943 softmax_18" -> "1944 dropout_72"  [label="(4, 12, 64, 64)", style=solid];
+"1944 dropout_72" -> "1945 matmul_37"  [label="(4, 12, 64, 64)", style=solid];
+"1945 matmul_37" -> "1946 transpose_37"  [label="(4, 12, 64, 32)", style=solid];
+"1946 transpose_37" -> "1947 reshape_83"  [label="(4, 64, 12, 32)", style=solid];
+"1947 reshape_83" -> "1948 reshape_83_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1948 reshape_83_0_0_nncf_smooth_quant_0" -> "1949 quantize_per_tensor_default_114"  [label="(4, 64, 384)", style=solid];
+"1949 quantize_per_tensor_default_114" -> "1950 dequantize_per_tensor_default_114"  [label="(4, 64, 384)", style=solid];
+"1950 dequantize_per_tensor_default_114" -> "1957 linear_113"  [label="(4, 64, 384)", style=solid];
+"1951 linear_113_scale_0" -> "1954 quantize_per_channel_default_95"  [label="(384,)", style=solid];
+"1951 linear_113_scale_0" -> "1955 dequantize_per_channel_default_95"  [label="(384,)", style=solid];
+"1952 linear_113_zero_point_0" -> "1954 quantize_per_channel_default_95"  [label="(384,)", style=solid];
+"1952 linear_113_zero_point_0" -> "1955 dequantize_per_channel_default_95"  [label="(384,)", style=solid];
+"1953 compressed_weight_updated_constant76" -> "1954 quantize_per_channel_default_95"  [label="(384, 384)", style=solid];
+"1954 quantize_per_channel_default_95" -> "1955 dequantize_per_channel_default_95"  [label="(384, 384)", style=solid];
+"1955 dequantize_per_channel_default_95" -> "1957 linear_113"  [label="(384, 384)", style=solid];
+"1956 features_5_14_attn_proj_bias_0_0" -> "1957 linear_113"  [label="(384,)", style=solid];
+"1957 linear_113" -> "1958 dropout_73"  [label="(4, 64, 384)", style=solid];
+"1958 dropout_73" -> "1959 view_102"  [label="(4, 64, 384)", style=solid];
+"1959 view_102" -> "1960 permute_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1960 permute_85" -> "1961 reshape_84"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1961 reshape_84" -> "1962 slice_274"  [label="(1, 16, 16, 384)", style=solid];
+"1962 slice_274" -> "1963 slice_275"  [label="(1, 16, 16, 384)", style=solid];
+"1963 slice_275" -> "1964 slice_276"  [label="(1, 14, 16, 384)", style=solid];
+"1964 slice_276" -> "1965 slice_277"  [label="(1, 14, 14, 384)", style=solid];
+"1965 slice_277" -> "1966 contiguous_35"  [label="(1, 14, 14, 384)", style=solid];
+"1966 contiguous_35" -> "1967 layer_norm_39"  [label="(1, 14, 14, 384)", style=solid];
+"1967 layer_norm_39" -> "1968 add_64"  [label="(1, 14, 14, 384)", style=solid];
+"1968 add_64" -> "1969 add_64_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"1968 add_64" -> "1993 add_65"  [label="(1, 14, 14, 384)", style=solid];
+"1969 add_64_0_0_nncf_smooth_quant_0" -> "1970 quantize_per_tensor_default_115"  [label="(1, 14, 14, 384)", style=solid];
+"1970 quantize_per_tensor_default_115" -> "1971 dequantize_per_tensor_default_115"  [label="(1, 14, 14, 384)", style=solid];
+"1971 dequantize_per_tensor_default_115" -> "1978 linear_114"  [label="(1, 14, 14, 384)", style=solid];
+"1972 linear_114_scale_0" -> "1975 quantize_per_channel_default_96"  [label="(1536,)", style=solid];
+"1972 linear_114_scale_0" -> "1976 dequantize_per_channel_default_96"  [label="(1536,)", style=solid];
+"1973 linear_114_zero_point_0" -> "1975 quantize_per_channel_default_96"  [label="(1536,)", style=solid];
+"1973 linear_114_zero_point_0" -> "1976 dequantize_per_channel_default_96"  [label="(1536,)", style=solid];
+"1974 compressed_weight_updated_constant77" -> "1975 quantize_per_channel_default_96"  [label="(1536, 384)", style=solid];
+"1975 quantize_per_channel_default_96" -> "1976 dequantize_per_channel_default_96"  [label="(1536, 384)", style=solid];
+"1976 dequantize_per_channel_default_96" -> "1978 linear_114"  [label="(1536, 384)", style=solid];
+"1977 features_5_14_mlp_0_bias_0_0" -> "1978 linear_114"  [label="(1536,)", style=solid];
+"1978 linear_114" -> "1979 gelu_18"  [label="(1, 14, 14, 1536)", style=solid];
+"1979 gelu_18" -> "1980 dropout_74"  [label="(1, 14, 14, 1536)", style=solid];
+"1980 dropout_74" -> "1981 dropout_74_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"1981 dropout_74_0_0_nncf_smooth_quant_0" -> "1982 quantize_per_tensor_default_116"  [label="(1, 14, 14, 1536)", style=solid];
+"1982 quantize_per_tensor_default_116" -> "1983 dequantize_per_tensor_default_116"  [label="(1, 14, 14, 1536)", style=solid];
+"1983 dequantize_per_tensor_default_116" -> "1990 linear_115"  [label="(1, 14, 14, 1536)", style=solid];
+"1984 linear_115_scale_0" -> "1987 quantize_per_channel_default_97"  [label="(384,)", style=solid];
+"1984 linear_115_scale_0" -> "1988 dequantize_per_channel_default_97"  [label="(384,)", style=solid];
+"1985 linear_115_zero_point_0" -> "1987 quantize_per_channel_default_97"  [label="(384,)", style=solid];
+"1985 linear_115_zero_point_0" -> "1988 dequantize_per_channel_default_97"  [label="(384,)", style=solid];
+"1986 compressed_weight_updated_constant78" -> "1987 quantize_per_channel_default_97"  [label="(384, 1536)", style=solid];
+"1987 quantize_per_channel_default_97" -> "1988 dequantize_per_channel_default_97"  [label="(384, 1536)", style=solid];
+"1988 dequantize_per_channel_default_97" -> "1990 linear_115"  [label="(384, 1536)", style=solid];
+"1989 features_5_14_mlp_3_bias_0_0" -> "1990 linear_115"  [label="(384,)", style=solid];
+"1990 linear_115" -> "1991 dropout_75"  [label="(1, 14, 14, 384)", style=solid];
+"1991 dropout_75" -> "1992 layer_norm_40"  [label="(1, 14, 14, 384)", style=solid];
+"1992 layer_norm_40" -> "1993 add_65"  [label="(1, 14, 14, 384)", style=solid];
+"1993 add_65" -> "1995 pad_21"  [label="(1, 14, 14, 384)", style=solid];
+"1993 add_65" -> "2074 add_68"  [label="(1, 14, 14, 384)", style=solid];
+"1994 _frozen_param157" -> "2031 add_66"  [label="(1, 12, 64, 64)", style=solid];
+"1995 pad_21" -> "1996 roll_18"  [label="(1, 16, 16, 384)", style=solid];
+"1996 roll_18" -> "1997 view_105"  [label="(1, 16, 16, 384)", style=solid];
+"1997 view_105" -> "1998 permute_87"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1998 permute_87" -> "1999 reshape_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1999 reshape_85" -> "2000 reshape_85_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"1999 reshape_85" -> "2032 new_zeros_9"  [label="(4, 64, 384)", style=solid];
+"2000 reshape_85_0_0_nncf_smooth_quant_0" -> "2001 quantize_per_tensor_default_117"  [label="(4, 64, 384)", style=solid];
+"2001 quantize_per_tensor_default_117" -> "2002 dequantize_per_tensor_default_117"  [label="(4, 64, 384)", style=solid];
+"2002 dequantize_per_tensor_default_117" -> "2009 linear_118"  [label="(4, 64, 384)", style=solid];
+"2003 linear_118_scale_0" -> "2006 quantize_per_channel_default_99"  [label="(1152,)", style=solid];
+"2003 linear_118_scale_0" -> "2007 dequantize_per_channel_default_99"  [label="(1152,)", style=solid];
+"2004 linear_118_zero_point_0" -> "2006 quantize_per_channel_default_99"  [label="(1152,)", style=solid];
+"2004 linear_118_zero_point_0" -> "2007 dequantize_per_channel_default_99"  [label="(1152,)", style=solid];
+"2005 compressed_weight_updated_constant79" -> "2006 quantize_per_channel_default_99"  [label="(1152, 384)", style=solid];
+"2006 quantize_per_channel_default_99" -> "2007 dequantize_per_channel_default_99"  [label="(1152, 384)", style=solid];
+"2007 dequantize_per_channel_default_99" -> "2009 linear_118"  [label="(1152, 384)", style=solid];
+"2008 _frozen_param158_0_0" -> "2009 linear_118"  [label="(1152,)", style=solid];
+"2009 linear_118" -> "2010 reshape_86"  [label="(4, 64, 1152)", style=solid];
+"2010 reshape_86" -> "2011 permute_88"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2011 permute_88" -> "2012 select_57"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2011 permute_88" -> "2013 select_58"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2011 permute_88" -> "2014 select_59"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2012 select_57" -> "2015 linalg_vector_norm_38"  [label="(4, 12, 64, 32)", style=solid];
+"2012 select_57" -> "2017 expand_as_38"  [label="(4, 12, 64, 32)", style=solid];
+"2012 select_57" -> "2018 div_38"  [label="(4, 12, 64, 32)", style=solid];
+"2013 select_58" -> "2021 linalg_vector_norm_39"  [label="(4, 12, 64, 32)", style=solid];
+"2013 select_58" -> "2023 expand_as_39"  [label="(4, 12, 64, 32)", style=solid];
+"2013 select_58" -> "2024 div_39"  [label="(4, 12, 64, 32)", style=solid];
+"2014 select_59" -> "2050 matmul_39"  [label="(4, 12, 64, 32)", style=solid];
+"2015 linalg_vector_norm_38" -> "2016 clamp_min_38"  [label="(4, 12, 64, 1)", style=solid];
+"2016 clamp_min_38" -> "2017 expand_as_38"  [label="(4, 12, 64, 1)", style=solid];
+"2017 expand_as_38" -> "2018 div_38"  [label="(4, 12, 64, 32)", style=solid];
+"2018 div_38" -> "2019 quantize_per_tensor_default_118"  [label="(4, 12, 64, 32)", style=solid];
+"2019 quantize_per_tensor_default_118" -> "2020 dequantize_per_tensor_default_118"  [label="(4, 12, 64, 32)", style=solid];
+"2020 dequantize_per_tensor_default_118" -> "2028 matmul_38"  [label="(4, 12, 64, 32)", style=solid];
+"2021 linalg_vector_norm_39" -> "2022 clamp_min_39"  [label="(4, 12, 64, 1)", style=solid];
+"2022 clamp_min_39" -> "2023 expand_as_39"  [label="(4, 12, 64, 1)", style=solid];
+"2023 expand_as_39" -> "2024 div_39"  [label="(4, 12, 64, 32)", style=solid];
+"2024 div_39" -> "2025 quantize_per_tensor_default_119"  [label="(4, 12, 64, 32)", style=solid];
+"2025 quantize_per_tensor_default_119" -> "2026 dequantize_per_tensor_default_119"  [label="(4, 12, 64, 32)", style=solid];
+"2026 dequantize_per_tensor_default_119" -> "2027 transpose_38"  [label="(4, 12, 64, 32)", style=solid];
+"2027 transpose_38" -> "2028 matmul_38"  [label="(4, 12, 32, 64)", style=solid];
+"2028 matmul_38" -> "2030 mul_39"  [label="(4, 12, 64, 64)", style=solid];
+"2029 _frozen_param160" -> "2030 mul_39"  [label="(12, 1, 1)", style=solid];
+"2030 mul_39" -> "2031 add_66"  [label="(4, 12, 64, 64)", style=solid];
+"2031 add_66" -> "2043 view_107"  [label="(4, 12, 64, 64)", style=solid];
+"2032 new_zeros_9" -> "2033 view_106"  [label="(16, 16)", style=solid];
+"2033 view_106" -> "2034 permute_89"  [label="(2, 8, 2, 8)", style=solid];
+"2034 permute_89" -> "2035 reshape_87"  [label="(2, 2, 8, 8)", style=solid];
+"2035 reshape_87" -> "2036 unsqueeze_56"  [label="(4, 64)", style=solid];
+"2035 reshape_87" -> "2037 unsqueeze_57"  [label="(4, 64)", style=solid];
+"2036 unsqueeze_56" -> "2038 sub_9"  [label="(4, 1, 64)", style=solid];
+"2037 unsqueeze_57" -> "2038 sub_9"  [label="(4, 64, 1)", style=solid];
+"2038 sub_9" -> "2039 ne_9"  [label="(4, 64, 64)", style=solid];
+"2038 sub_9" -> "2040 masked_fill_18"  [label="(4, 64, 64)", style=solid];
+"2038 sub_9" -> "2041 eq_9"  [label="(4, 64, 64)", style=solid];
+"2039 ne_9" -> "2040 masked_fill_18"  [label="(4, 64, 64)", style=solid];
+"2040 masked_fill_18" -> "2042 masked_fill_19"  [label="(4, 64, 64)", style=solid];
+"2041 eq_9" -> "2042 masked_fill_19"  [label="(4, 64, 64)", style=solid];
+"2042 masked_fill_19" -> "2044 unsqueeze_58"  [label="(4, 64, 64)", style=solid];
+"2043 view_107" -> "2046 add_67"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2044 unsqueeze_58" -> "2045 unsqueeze_59"  [label="(4, 1, 64, 64)", style=solid];
+"2045 unsqueeze_59" -> "2046 add_67"  [label="(1, 4, 1, 64, 64)", style=solid];
+"2046 add_67" -> "2047 view_108"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2047 view_108" -> "2048 softmax_19"  [label="(4, 12, 64, 64)", style=solid];
+"2048 softmax_19" -> "2049 dropout_76"  [label="(4, 12, 64, 64)", style=solid];
+"2049 dropout_76" -> "2050 matmul_39"  [label="(4, 12, 64, 64)", style=solid];
+"2050 matmul_39" -> "2051 transpose_39"  [label="(4, 12, 64, 32)", style=solid];
+"2051 transpose_39" -> "2052 reshape_88"  [label="(4, 64, 12, 32)", style=solid];
+"2052 reshape_88" -> "2053 reshape_88_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"2053 reshape_88_0_0_nncf_smooth_quant_0" -> "2054 quantize_per_tensor_default_120"  [label="(4, 64, 384)", style=solid];
+"2054 quantize_per_tensor_default_120" -> "2055 dequantize_per_tensor_default_120"  [label="(4, 64, 384)", style=solid];
+"2055 dequantize_per_tensor_default_120" -> "2062 linear_119"  [label="(4, 64, 384)", style=solid];
+"2056 linear_119_scale_0" -> "2059 quantize_per_channel_default_100"  [label="(384,)", style=solid];
+"2056 linear_119_scale_0" -> "2060 dequantize_per_channel_default_100"  [label="(384,)", style=solid];
+"2057 linear_119_zero_point_0" -> "2059 quantize_per_channel_default_100"  [label="(384,)", style=solid];
+"2057 linear_119_zero_point_0" -> "2060 dequantize_per_channel_default_100"  [label="(384,)", style=solid];
+"2058 compressed_weight_updated_constant80" -> "2059 quantize_per_channel_default_100"  [label="(384, 384)", style=solid];
+"2059 quantize_per_channel_default_100" -> "2060 dequantize_per_channel_default_100"  [label="(384, 384)", style=solid];
+"2060 dequantize_per_channel_default_100" -> "2062 linear_119"  [label="(384, 384)", style=solid];
+"2061 features_5_15_attn_proj_bias_0_0" -> "2062 linear_119"  [label="(384,)", style=solid];
+"2062 linear_119" -> "2063 dropout_77"  [label="(4, 64, 384)", style=solid];
+"2063 dropout_77" -> "2064 view_109"  [label="(4, 64, 384)", style=solid];
+"2064 view_109" -> "2065 permute_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2065 permute_90" -> "2066 reshape_89"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2066 reshape_89" -> "2067 roll_19"  [label="(1, 16, 16, 384)", style=solid];
+"2067 roll_19" -> "2068 slice_297"  [label="(1, 16, 16, 384)", style=solid];
+"2068 slice_297" -> "2069 slice_298"  [label="(1, 16, 16, 384)", style=solid];
+"2069 slice_298" -> "2070 slice_299"  [label="(1, 14, 16, 384)", style=solid];
+"2070 slice_299" -> "2071 slice_300"  [label="(1, 14, 14, 384)", style=solid];
+"2071 slice_300" -> "2072 contiguous_37"  [label="(1, 14, 14, 384)", style=solid];
+"2072 contiguous_37" -> "2073 layer_norm_41"  [label="(1, 14, 14, 384)", style=solid];
+"2073 layer_norm_41" -> "2074 add_68"  [label="(1, 14, 14, 384)", style=solid];
+"2074 add_68" -> "2075 add_68_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"2074 add_68" -> "2099 add_69"  [label="(1, 14, 14, 384)", style=solid];
+"2075 add_68_0_0_nncf_smooth_quant_0" -> "2076 quantize_per_tensor_default_121"  [label="(1, 14, 14, 384)", style=solid];
+"2076 quantize_per_tensor_default_121" -> "2077 dequantize_per_tensor_default_121"  [label="(1, 14, 14, 384)", style=solid];
+"2077 dequantize_per_tensor_default_121" -> "2084 linear_120"  [label="(1, 14, 14, 384)", style=solid];
+"2078 linear_120_scale_0" -> "2081 quantize_per_channel_default_101"  [label="(1536,)", style=solid];
+"2078 linear_120_scale_0" -> "2082 dequantize_per_channel_default_101"  [label="(1536,)", style=solid];
+"2079 linear_120_zero_point_0" -> "2081 quantize_per_channel_default_101"  [label="(1536,)", style=solid];
+"2079 linear_120_zero_point_0" -> "2082 dequantize_per_channel_default_101"  [label="(1536,)", style=solid];
+"2080 compressed_weight_updated_constant81" -> "2081 quantize_per_channel_default_101"  [label="(1536, 384)", style=solid];
+"2081 quantize_per_channel_default_101" -> "2082 dequantize_per_channel_default_101"  [label="(1536, 384)", style=solid];
+"2082 dequantize_per_channel_default_101" -> "2084 linear_120"  [label="(1536, 384)", style=solid];
+"2083 features_5_15_mlp_0_bias_0_0" -> "2084 linear_120"  [label="(1536,)", style=solid];
+"2084 linear_120" -> "2085 gelu_19"  [label="(1, 14, 14, 1536)", style=solid];
+"2085 gelu_19" -> "2086 dropout_78"  [label="(1, 14, 14, 1536)", style=solid];
+"2086 dropout_78" -> "2087 dropout_78_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"2087 dropout_78_0_0_nncf_smooth_quant_0" -> "2088 quantize_per_tensor_default_122"  [label="(1, 14, 14, 1536)", style=solid];
+"2088 quantize_per_tensor_default_122" -> "2089 dequantize_per_tensor_default_122"  [label="(1, 14, 14, 1536)", style=solid];
+"2089 dequantize_per_tensor_default_122" -> "2096 linear_121"  [label="(1, 14, 14, 1536)", style=solid];
+"2090 linear_121_scale_0" -> "2093 quantize_per_channel_default_102"  [label="(384,)", style=solid];
+"2090 linear_121_scale_0" -> "2094 dequantize_per_channel_default_102"  [label="(384,)", style=solid];
+"2091 linear_121_zero_point_0" -> "2093 quantize_per_channel_default_102"  [label="(384,)", style=solid];
+"2091 linear_121_zero_point_0" -> "2094 dequantize_per_channel_default_102"  [label="(384,)", style=solid];
+"2092 compressed_weight_updated_constant82" -> "2093 quantize_per_channel_default_102"  [label="(384, 1536)", style=solid];
+"2093 quantize_per_channel_default_102" -> "2094 dequantize_per_channel_default_102"  [label="(384, 1536)", style=solid];
+"2094 dequantize_per_channel_default_102" -> "2096 linear_121"  [label="(384, 1536)", style=solid];
+"2095 features_5_15_mlp_3_bias_0_0" -> "2096 linear_121"  [label="(384,)", style=solid];
+"2096 linear_121" -> "2097 dropout_79"  [label="(1, 14, 14, 384)", style=solid];
+"2097 dropout_79" -> "2098 layer_norm_42"  [label="(1, 14, 14, 384)", style=solid];
+"2098 layer_norm_42" -> "2099 add_69"  [label="(1, 14, 14, 384)", style=solid];
+"2099 add_69" -> "2101 pad_22"  [label="(1, 14, 14, 384)", style=solid];
+"2099 add_69" -> "2162 add_71"  [label="(1, 14, 14, 384)", style=solid];
+"2100 _frozen_param170" -> "2136 add_70"  [label="(1, 12, 64, 64)", style=solid];
+"2101 pad_22" -> "2102 view_112"  [label="(1, 16, 16, 384)", style=solid];
+"2102 view_112" -> "2103 permute_92"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2103 permute_92" -> "2104 reshape_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2104 reshape_90" -> "2105 reshape_90_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"2105 reshape_90_0_0_nncf_smooth_quant_0" -> "2106 quantize_per_tensor_default_123"  [label="(4, 64, 384)", style=solid];
+"2106 quantize_per_tensor_default_123" -> "2107 dequantize_per_tensor_default_123"  [label="(4, 64, 384)", style=solid];
+"2107 dequantize_per_tensor_default_123" -> "2114 linear_124"  [label="(4, 64, 384)", style=solid];
+"2108 linear_124_scale_0" -> "2111 quantize_per_channel_default_104"  [label="(1152,)", style=solid];
+"2108 linear_124_scale_0" -> "2112 dequantize_per_channel_default_104"  [label="(1152,)", style=solid];
+"2109 linear_124_zero_point_0" -> "2111 quantize_per_channel_default_104"  [label="(1152,)", style=solid];
+"2109 linear_124_zero_point_0" -> "2112 dequantize_per_channel_default_104"  [label="(1152,)", style=solid];
+"2110 compressed_weight_updated_constant83" -> "2111 quantize_per_channel_default_104"  [label="(1152, 384)", style=solid];
+"2111 quantize_per_channel_default_104" -> "2112 dequantize_per_channel_default_104"  [label="(1152, 384)", style=solid];
+"2112 dequantize_per_channel_default_104" -> "2114 linear_124"  [label="(1152, 384)", style=solid];
+"2113 _frozen_param171_0_0" -> "2114 linear_124"  [label="(1152,)", style=solid];
+"2114 linear_124" -> "2115 reshape_91"  [label="(4, 64, 1152)", style=solid];
+"2115 reshape_91" -> "2116 permute_93"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2116 permute_93" -> "2117 select_60"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2116 permute_93" -> "2118 select_61"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2116 permute_93" -> "2119 select_62"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2117 select_60" -> "2120 linalg_vector_norm_40"  [label="(4, 12, 64, 32)", style=solid];
+"2117 select_60" -> "2122 expand_as_40"  [label="(4, 12, 64, 32)", style=solid];
+"2117 select_60" -> "2123 div_40"  [label="(4, 12, 64, 32)", style=solid];
+"2118 select_61" -> "2126 linalg_vector_norm_41"  [label="(4, 12, 64, 32)", style=solid];
+"2118 select_61" -> "2128 expand_as_41"  [label="(4, 12, 64, 32)", style=solid];
+"2118 select_61" -> "2129 div_41"  [label="(4, 12, 64, 32)", style=solid];
+"2119 select_62" -> "2139 matmul_41"  [label="(4, 12, 64, 32)", style=solid];
+"2120 linalg_vector_norm_40" -> "2121 clamp_min_40"  [label="(4, 12, 64, 1)", style=solid];
+"2121 clamp_min_40" -> "2122 expand_as_40"  [label="(4, 12, 64, 1)", style=solid];
+"2122 expand_as_40" -> "2123 div_40"  [label="(4, 12, 64, 32)", style=solid];
+"2123 div_40" -> "2124 quantize_per_tensor_default_124"  [label="(4, 12, 64, 32)", style=solid];
+"2124 quantize_per_tensor_default_124" -> "2125 dequantize_per_tensor_default_124"  [label="(4, 12, 64, 32)", style=solid];
+"2125 dequantize_per_tensor_default_124" -> "2133 matmul_40"  [label="(4, 12, 64, 32)", style=solid];
+"2126 linalg_vector_norm_41" -> "2127 clamp_min_41"  [label="(4, 12, 64, 1)", style=solid];
+"2127 clamp_min_41" -> "2128 expand_as_41"  [label="(4, 12, 64, 1)", style=solid];
+"2128 expand_as_41" -> "2129 div_41"  [label="(4, 12, 64, 32)", style=solid];
+"2129 div_41" -> "2130 quantize_per_tensor_default_125"  [label="(4, 12, 64, 32)", style=solid];
+"2130 quantize_per_tensor_default_125" -> "2131 dequantize_per_tensor_default_125"  [label="(4, 12, 64, 32)", style=solid];
+"2131 dequantize_per_tensor_default_125" -> "2132 transpose_40"  [label="(4, 12, 64, 32)", style=solid];
+"2132 transpose_40" -> "2133 matmul_40"  [label="(4, 12, 32, 64)", style=solid];
+"2133 matmul_40" -> "2135 mul_41"  [label="(4, 12, 64, 64)", style=solid];
+"2134 _frozen_param173" -> "2135 mul_41"  [label="(12, 1, 1)", style=solid];
+"2135 mul_41" -> "2136 add_70"  [label="(4, 12, 64, 64)", style=solid];
+"2136 add_70" -> "2137 softmax_20"  [label="(4, 12, 64, 64)", style=solid];
+"2137 softmax_20" -> "2138 dropout_80"  [label="(4, 12, 64, 64)", style=solid];
+"2138 dropout_80" -> "2139 matmul_41"  [label="(4, 12, 64, 64)", style=solid];
+"2139 matmul_41" -> "2140 transpose_41"  [label="(4, 12, 64, 32)", style=solid];
+"2140 transpose_41" -> "2141 reshape_92"  [label="(4, 64, 12, 32)", style=solid];
+"2141 reshape_92" -> "2142 reshape_92_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"2142 reshape_92_0_0_nncf_smooth_quant_0" -> "2143 quantize_per_tensor_default_126"  [label="(4, 64, 384)", style=solid];
+"2143 quantize_per_tensor_default_126" -> "2144 dequantize_per_tensor_default_126"  [label="(4, 64, 384)", style=solid];
+"2144 dequantize_per_tensor_default_126" -> "2151 linear_125"  [label="(4, 64, 384)", style=solid];
+"2145 linear_125_scale_0" -> "2148 quantize_per_channel_default_105"  [label="(384,)", style=solid];
+"2145 linear_125_scale_0" -> "2149 dequantize_per_channel_default_105"  [label="(384,)", style=solid];
+"2146 linear_125_zero_point_0" -> "2148 quantize_per_channel_default_105"  [label="(384,)", style=solid];
+"2146 linear_125_zero_point_0" -> "2149 dequantize_per_channel_default_105"  [label="(384,)", style=solid];
+"2147 compressed_weight_updated_constant84" -> "2148 quantize_per_channel_default_105"  [label="(384, 384)", style=solid];
+"2148 quantize_per_channel_default_105" -> "2149 dequantize_per_channel_default_105"  [label="(384, 384)", style=solid];
+"2149 dequantize_per_channel_default_105" -> "2151 linear_125"  [label="(384, 384)", style=solid];
+"2150 features_5_16_attn_proj_bias_0_0" -> "2151 linear_125"  [label="(384,)", style=solid];
+"2151 linear_125" -> "2152 dropout_81"  [label="(4, 64, 384)", style=solid];
+"2152 dropout_81" -> "2153 view_113"  [label="(4, 64, 384)", style=solid];
+"2153 view_113" -> "2154 permute_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2154 permute_94" -> "2155 reshape_93"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2155 reshape_93" -> "2156 slice_302"  [label="(1, 16, 16, 384)", style=solid];
+"2156 slice_302" -> "2157 slice_303"  [label="(1, 16, 16, 384)", style=solid];
+"2157 slice_303" -> "2158 slice_304"  [label="(1, 14, 16, 384)", style=solid];
+"2158 slice_304" -> "2159 slice_305"  [label="(1, 14, 14, 384)", style=solid];
+"2159 slice_305" -> "2160 contiguous_39"  [label="(1, 14, 14, 384)", style=solid];
+"2160 contiguous_39" -> "2161 layer_norm_43"  [label="(1, 14, 14, 384)", style=solid];
+"2161 layer_norm_43" -> "2162 add_71"  [label="(1, 14, 14, 384)", style=solid];
+"2162 add_71" -> "2163 add_71_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"2162 add_71" -> "2187 add_72"  [label="(1, 14, 14, 384)", style=solid];
+"2163 add_71_0_0_nncf_smooth_quant_0" -> "2164 quantize_per_tensor_default_127"  [label="(1, 14, 14, 384)", style=solid];
+"2164 quantize_per_tensor_default_127" -> "2165 dequantize_per_tensor_default_127"  [label="(1, 14, 14, 384)", style=solid];
+"2165 dequantize_per_tensor_default_127" -> "2172 linear_126"  [label="(1, 14, 14, 384)", style=solid];
+"2166 linear_126_scale_0" -> "2169 quantize_per_channel_default_106"  [label="(1536,)", style=solid];
+"2166 linear_126_scale_0" -> "2170 dequantize_per_channel_default_106"  [label="(1536,)", style=solid];
+"2167 linear_126_zero_point_0" -> "2169 quantize_per_channel_default_106"  [label="(1536,)", style=solid];
+"2167 linear_126_zero_point_0" -> "2170 dequantize_per_channel_default_106"  [label="(1536,)", style=solid];
+"2168 compressed_weight_updated_constant85" -> "2169 quantize_per_channel_default_106"  [label="(1536, 384)", style=solid];
+"2169 quantize_per_channel_default_106" -> "2170 dequantize_per_channel_default_106"  [label="(1536, 384)", style=solid];
+"2170 dequantize_per_channel_default_106" -> "2172 linear_126"  [label="(1536, 384)", style=solid];
+"2171 features_5_16_mlp_0_bias_0_0" -> "2172 linear_126"  [label="(1536,)", style=solid];
+"2172 linear_126" -> "2173 gelu_20"  [label="(1, 14, 14, 1536)", style=solid];
+"2173 gelu_20" -> "2174 dropout_82"  [label="(1, 14, 14, 1536)", style=solid];
+"2174 dropout_82" -> "2175 dropout_82_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"2175 dropout_82_0_0_nncf_smooth_quant_0" -> "2176 quantize_per_tensor_default_128"  [label="(1, 14, 14, 1536)", style=solid];
+"2176 quantize_per_tensor_default_128" -> "2177 dequantize_per_tensor_default_128"  [label="(1, 14, 14, 1536)", style=solid];
+"2177 dequantize_per_tensor_default_128" -> "2184 linear_127"  [label="(1, 14, 14, 1536)", style=solid];
+"2178 linear_127_scale_0" -> "2181 quantize_per_channel_default_107"  [label="(384,)", style=solid];
+"2178 linear_127_scale_0" -> "2182 dequantize_per_channel_default_107"  [label="(384,)", style=solid];
+"2179 linear_127_zero_point_0" -> "2181 quantize_per_channel_default_107"  [label="(384,)", style=solid];
+"2179 linear_127_zero_point_0" -> "2182 dequantize_per_channel_default_107"  [label="(384,)", style=solid];
+"2180 compressed_weight_updated_constant86" -> "2181 quantize_per_channel_default_107"  [label="(384, 1536)", style=solid];
+"2181 quantize_per_channel_default_107" -> "2182 dequantize_per_channel_default_107"  [label="(384, 1536)", style=solid];
+"2182 dequantize_per_channel_default_107" -> "2184 linear_127"  [label="(384, 1536)", style=solid];
+"2183 features_5_16_mlp_3_bias_0_0" -> "2184 linear_127"  [label="(384,)", style=solid];
+"2184 linear_127" -> "2185 dropout_83"  [label="(1, 14, 14, 384)", style=solid];
+"2185 dropout_83" -> "2186 layer_norm_44"  [label="(1, 14, 14, 384)", style=solid];
+"2186 layer_norm_44" -> "2187 add_72"  [label="(1, 14, 14, 384)", style=solid];
+"2187 add_72" -> "2189 pad_23"  [label="(1, 14, 14, 384)", style=solid];
+"2187 add_72" -> "2268 add_75"  [label="(1, 14, 14, 384)", style=solid];
+"2188 _frozen_param174" -> "2225 add_73"  [label="(1, 12, 64, 64)", style=solid];
+"2189 pad_23" -> "2190 roll_20"  [label="(1, 16, 16, 384)", style=solid];
+"2190 roll_20" -> "2191 view_116"  [label="(1, 16, 16, 384)", style=solid];
+"2191 view_116" -> "2192 permute_96"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2192 permute_96" -> "2193 reshape_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2193 reshape_94" -> "2194 reshape_94_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"2193 reshape_94" -> "2226 new_zeros_10"  [label="(4, 64, 384)", style=solid];
+"2194 reshape_94_0_0_nncf_smooth_quant_0" -> "2195 quantize_per_tensor_default_129"  [label="(4, 64, 384)", style=solid];
+"2195 quantize_per_tensor_default_129" -> "2196 dequantize_per_tensor_default_129"  [label="(4, 64, 384)", style=solid];
+"2196 dequantize_per_tensor_default_129" -> "2203 linear_130"  [label="(4, 64, 384)", style=solid];
+"2197 linear_130_scale_0" -> "2200 quantize_per_channel_default_109"  [label="(1152,)", style=solid];
+"2197 linear_130_scale_0" -> "2201 dequantize_per_channel_default_109"  [label="(1152,)", style=solid];
+"2198 linear_130_zero_point_0" -> "2200 quantize_per_channel_default_109"  [label="(1152,)", style=solid];
+"2198 linear_130_zero_point_0" -> "2201 dequantize_per_channel_default_109"  [label="(1152,)", style=solid];
+"2199 compressed_weight_updated_constant87" -> "2200 quantize_per_channel_default_109"  [label="(1152, 384)", style=solid];
+"2200 quantize_per_channel_default_109" -> "2201 dequantize_per_channel_default_109"  [label="(1152, 384)", style=solid];
+"2201 dequantize_per_channel_default_109" -> "2203 linear_130"  [label="(1152, 384)", style=solid];
+"2202 _frozen_param175_0_0" -> "2203 linear_130"  [label="(1152,)", style=solid];
+"2203 linear_130" -> "2204 reshape_95"  [label="(4, 64, 1152)", style=solid];
+"2204 reshape_95" -> "2205 permute_97"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2205 permute_97" -> "2206 select_63"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2205 permute_97" -> "2207 select_64"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2205 permute_97" -> "2208 select_65"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2206 select_63" -> "2209 linalg_vector_norm_42"  [label="(4, 12, 64, 32)", style=solid];
+"2206 select_63" -> "2211 expand_as_42"  [label="(4, 12, 64, 32)", style=solid];
+"2206 select_63" -> "2212 div_42"  [label="(4, 12, 64, 32)", style=solid];
+"2207 select_64" -> "2215 linalg_vector_norm_43"  [label="(4, 12, 64, 32)", style=solid];
+"2207 select_64" -> "2217 expand_as_43"  [label="(4, 12, 64, 32)", style=solid];
+"2207 select_64" -> "2218 div_43"  [label="(4, 12, 64, 32)", style=solid];
+"2208 select_65" -> "2244 matmul_43"  [label="(4, 12, 64, 32)", style=solid];
+"2209 linalg_vector_norm_42" -> "2210 clamp_min_42"  [label="(4, 12, 64, 1)", style=solid];
+"2210 clamp_min_42" -> "2211 expand_as_42"  [label="(4, 12, 64, 1)", style=solid];
+"2211 expand_as_42" -> "2212 div_42"  [label="(4, 12, 64, 32)", style=solid];
+"2212 div_42" -> "2213 quantize_per_tensor_default_130"  [label="(4, 12, 64, 32)", style=solid];
+"2213 quantize_per_tensor_default_130" -> "2214 dequantize_per_tensor_default_130"  [label="(4, 12, 64, 32)", style=solid];
+"2214 dequantize_per_tensor_default_130" -> "2222 matmul_42"  [label="(4, 12, 64, 32)", style=solid];
+"2215 linalg_vector_norm_43" -> "2216 clamp_min_43"  [label="(4, 12, 64, 1)", style=solid];
+"2216 clamp_min_43" -> "2217 expand_as_43"  [label="(4, 12, 64, 1)", style=solid];
+"2217 expand_as_43" -> "2218 div_43"  [label="(4, 12, 64, 32)", style=solid];
+"2218 div_43" -> "2219 quantize_per_tensor_default_131"  [label="(4, 12, 64, 32)", style=solid];
+"2219 quantize_per_tensor_default_131" -> "2220 dequantize_per_tensor_default_131"  [label="(4, 12, 64, 32)", style=solid];
+"2220 dequantize_per_tensor_default_131" -> "2221 transpose_42"  [label="(4, 12, 64, 32)", style=solid];
+"2221 transpose_42" -> "2222 matmul_42"  [label="(4, 12, 32, 64)", style=solid];
+"2222 matmul_42" -> "2224 mul_43"  [label="(4, 12, 64, 64)", style=solid];
+"2223 _frozen_param177" -> "2224 mul_43"  [label="(12, 1, 1)", style=solid];
+"2224 mul_43" -> "2225 add_73"  [label="(4, 12, 64, 64)", style=solid];
+"2225 add_73" -> "2237 view_118"  [label="(4, 12, 64, 64)", style=solid];
+"2226 new_zeros_10" -> "2227 view_117"  [label="(16, 16)", style=solid];
+"2227 view_117" -> "2228 permute_98"  [label="(2, 8, 2, 8)", style=solid];
+"2228 permute_98" -> "2229 reshape_96"  [label="(2, 2, 8, 8)", style=solid];
+"2229 reshape_96" -> "2230 unsqueeze_62"  [label="(4, 64)", style=solid];
+"2229 reshape_96" -> "2231 unsqueeze_63"  [label="(4, 64)", style=solid];
+"2230 unsqueeze_62" -> "2232 sub_10"  [label="(4, 1, 64)", style=solid];
+"2231 unsqueeze_63" -> "2232 sub_10"  [label="(4, 64, 1)", style=solid];
+"2232 sub_10" -> "2233 ne_10"  [label="(4, 64, 64)", style=solid];
+"2232 sub_10" -> "2234 masked_fill_20"  [label="(4, 64, 64)", style=solid];
+"2232 sub_10" -> "2235 eq_10"  [label="(4, 64, 64)", style=solid];
+"2233 ne_10" -> "2234 masked_fill_20"  [label="(4, 64, 64)", style=solid];
+"2234 masked_fill_20" -> "2236 masked_fill_21"  [label="(4, 64, 64)", style=solid];
+"2235 eq_10" -> "2236 masked_fill_21"  [label="(4, 64, 64)", style=solid];
+"2236 masked_fill_21" -> "2238 unsqueeze_64"  [label="(4, 64, 64)", style=solid];
+"2237 view_118" -> "2240 add_74"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2238 unsqueeze_64" -> "2239 unsqueeze_65"  [label="(4, 1, 64, 64)", style=solid];
+"2239 unsqueeze_65" -> "2240 add_74"  [label="(1, 4, 1, 64, 64)", style=solid];
+"2240 add_74" -> "2241 view_119"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2241 view_119" -> "2242 softmax_21"  [label="(4, 12, 64, 64)", style=solid];
+"2242 softmax_21" -> "2243 dropout_84"  [label="(4, 12, 64, 64)", style=solid];
+"2243 dropout_84" -> "2244 matmul_43"  [label="(4, 12, 64, 64)", style=solid];
+"2244 matmul_43" -> "2245 transpose_43"  [label="(4, 12, 64, 32)", style=solid];
+"2245 transpose_43" -> "2246 reshape_97"  [label="(4, 64, 12, 32)", style=solid];
+"2246 reshape_97" -> "2247 reshape_97_0_0_nncf_smooth_quant_0"  [label="(4, 64, 384)", style=solid];
+"2247 reshape_97_0_0_nncf_smooth_quant_0" -> "2248 quantize_per_tensor_default_132"  [label="(4, 64, 384)", style=solid];
+"2248 quantize_per_tensor_default_132" -> "2249 dequantize_per_tensor_default_132"  [label="(4, 64, 384)", style=solid];
+"2249 dequantize_per_tensor_default_132" -> "2256 linear_131"  [label="(4, 64, 384)", style=solid];
+"2250 linear_131_scale_0" -> "2253 quantize_per_channel_default_110"  [label="(384,)", style=solid];
+"2250 linear_131_scale_0" -> "2254 dequantize_per_channel_default_110"  [label="(384,)", style=solid];
+"2251 linear_131_zero_point_0" -> "2253 quantize_per_channel_default_110"  [label="(384,)", style=solid];
+"2251 linear_131_zero_point_0" -> "2254 dequantize_per_channel_default_110"  [label="(384,)", style=solid];
+"2252 compressed_weight_updated_constant88" -> "2253 quantize_per_channel_default_110"  [label="(384, 384)", style=solid];
+"2253 quantize_per_channel_default_110" -> "2254 dequantize_per_channel_default_110"  [label="(384, 384)", style=solid];
+"2254 dequantize_per_channel_default_110" -> "2256 linear_131"  [label="(384, 384)", style=solid];
+"2255 features_5_17_attn_proj_bias_0_0" -> "2256 linear_131"  [label="(384,)", style=solid];
+"2256 linear_131" -> "2257 dropout_85"  [label="(4, 64, 384)", style=solid];
+"2257 dropout_85" -> "2258 view_120"  [label="(4, 64, 384)", style=solid];
+"2258 view_120" -> "2259 permute_99"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2259 permute_99" -> "2260 reshape_98"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2260 reshape_98" -> "2261 roll_21"  [label="(1, 16, 16, 384)", style=solid];
+"2261 roll_21" -> "2262 slice_325"  [label="(1, 16, 16, 384)", style=solid];
+"2262 slice_325" -> "2263 slice_326"  [label="(1, 16, 16, 384)", style=solid];
+"2263 slice_326" -> "2264 slice_327"  [label="(1, 14, 16, 384)", style=solid];
+"2264 slice_327" -> "2265 slice_328"  [label="(1, 14, 14, 384)", style=solid];
+"2265 slice_328" -> "2266 contiguous_41"  [label="(1, 14, 14, 384)", style=solid];
+"2266 contiguous_41" -> "2267 layer_norm_45"  [label="(1, 14, 14, 384)", style=solid];
+"2267 layer_norm_45" -> "2268 add_75"  [label="(1, 14, 14, 384)", style=solid];
+"2268 add_75" -> "2269 add_75_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 384)", style=solid];
+"2268 add_75" -> "2293 add_76"  [label="(1, 14, 14, 384)", style=solid];
+"2269 add_75_0_0_nncf_smooth_quant_0" -> "2270 quantize_per_tensor_default_133"  [label="(1, 14, 14, 384)", style=solid];
+"2270 quantize_per_tensor_default_133" -> "2271 dequantize_per_tensor_default_133"  [label="(1, 14, 14, 384)", style=solid];
+"2271 dequantize_per_tensor_default_133" -> "2278 linear_132"  [label="(1, 14, 14, 384)", style=solid];
+"2272 linear_132_scale_0" -> "2275 quantize_per_channel_default_111"  [label="(1536,)", style=solid];
+"2272 linear_132_scale_0" -> "2276 dequantize_per_channel_default_111"  [label="(1536,)", style=solid];
+"2273 linear_132_zero_point_0" -> "2275 quantize_per_channel_default_111"  [label="(1536,)", style=solid];
+"2273 linear_132_zero_point_0" -> "2276 dequantize_per_channel_default_111"  [label="(1536,)", style=solid];
+"2274 compressed_weight_updated_constant89" -> "2275 quantize_per_channel_default_111"  [label="(1536, 384)", style=solid];
+"2275 quantize_per_channel_default_111" -> "2276 dequantize_per_channel_default_111"  [label="(1536, 384)", style=solid];
+"2276 dequantize_per_channel_default_111" -> "2278 linear_132"  [label="(1536, 384)", style=solid];
+"2277 features_5_17_mlp_0_bias_0_0" -> "2278 linear_132"  [label="(1536,)", style=solid];
+"2278 linear_132" -> "2279 gelu_21"  [label="(1, 14, 14, 1536)", style=solid];
+"2279 gelu_21" -> "2280 dropout_86"  [label="(1, 14, 14, 1536)", style=solid];
+"2280 dropout_86" -> "2281 dropout_86_0_0_nncf_smooth_quant_0"  [label="(1, 14, 14, 1536)", style=solid];
+"2281 dropout_86_0_0_nncf_smooth_quant_0" -> "2282 quantize_per_tensor_default_134"  [label="(1, 14, 14, 1536)", style=solid];
+"2282 quantize_per_tensor_default_134" -> "2283 dequantize_per_tensor_default_134"  [label="(1, 14, 14, 1536)", style=solid];
+"2283 dequantize_per_tensor_default_134" -> "2290 linear_133"  [label="(1, 14, 14, 1536)", style=solid];
+"2284 linear_133_scale_0" -> "2287 quantize_per_channel_default_112"  [label="(384,)", style=solid];
+"2284 linear_133_scale_0" -> "2288 dequantize_per_channel_default_112"  [label="(384,)", style=solid];
+"2285 linear_133_zero_point_0" -> "2287 quantize_per_channel_default_112"  [label="(384,)", style=solid];
+"2285 linear_133_zero_point_0" -> "2288 dequantize_per_channel_default_112"  [label="(384,)", style=solid];
+"2286 compressed_weight_updated_constant90" -> "2287 quantize_per_channel_default_112"  [label="(384, 1536)", style=solid];
+"2287 quantize_per_channel_default_112" -> "2288 dequantize_per_channel_default_112"  [label="(384, 1536)", style=solid];
+"2288 dequantize_per_channel_default_112" -> "2290 linear_133"  [label="(384, 1536)", style=solid];
+"2289 features_5_17_mlp_3_bias_0_0" -> "2290 linear_133"  [label="(384,)", style=solid];
+"2290 linear_133" -> "2291 dropout_87"  [label="(1, 14, 14, 384)", style=solid];
+"2291 dropout_87" -> "2292 layer_norm_46"  [label="(1, 14, 14, 384)", style=solid];
+"2292 layer_norm_46" -> "2293 add_76"  [label="(1, 14, 14, 384)", style=solid];
+"2293 add_76" -> "2294 pad_24"  [label="(1, 14, 14, 384)", style=solid];
+"2294 pad_24" -> "2295 slice_329"  [label="(1, 14, 14, 384)", style=solid];
+"2294 pad_24" -> "2298 slice_332"  [label="(1, 14, 14, 384)", style=solid];
+"2294 pad_24" -> "2301 slice_335"  [label="(1, 14, 14, 384)", style=solid];
+"2294 pad_24" -> "2304 slice_338"  [label="(1, 14, 14, 384)", style=solid];
+"2295 slice_329" -> "2296 slice_330"  [label="(1, 7, 14, 384)", style=solid];
+"2296 slice_330" -> "2297 slice_331"  [label="(1, 7, 7, 384)", style=solid];
+"2297 slice_331" -> "2307 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2298 slice_332" -> "2299 slice_333"  [label="(1, 7, 14, 384)", style=solid];
+"2299 slice_333" -> "2300 slice_334"  [label="(1, 7, 7, 384)", style=solid];
+"2300 slice_334" -> "2307 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2301 slice_335" -> "2302 slice_336"  [label="(1, 7, 14, 384)", style=solid];
+"2302 slice_336" -> "2303 slice_337"  [label="(1, 7, 7, 384)", style=solid];
+"2303 slice_337" -> "2307 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2304 slice_338" -> "2305 slice_339"  [label="(1, 7, 14, 384)", style=solid];
+"2305 slice_339" -> "2306 slice_340"  [label="(1, 7, 7, 384)", style=solid];
+"2306 slice_340" -> "2307 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2307 cat_2" -> "2308 cat_2_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 1536)", style=solid];
+"2308 cat_2_0_0_nncf_smooth_quant_0" -> "2309 quantize_per_tensor_default_135"  [label="(1, 7, 7, 1536)", style=solid];
+"2309 quantize_per_tensor_default_135" -> "2310 dequantize_per_tensor_default_135"  [label="(1, 7, 7, 1536)", style=solid];
+"2310 dequantize_per_tensor_default_135" -> "2316 linear_134"  [label="(1, 7, 7, 1536)", style=solid];
+"2311 linear_134_scale_0" -> "2314 quantize_per_channel_default_113"  [label="(768,)", style=solid];
+"2311 linear_134_scale_0" -> "2315 dequantize_per_channel_default_113"  [label="(768,)", style=solid];
+"2312 linear_134_zero_point_0" -> "2314 quantize_per_channel_default_113"  [label="(768,)", style=solid];
+"2312 linear_134_zero_point_0" -> "2315 dequantize_per_channel_default_113"  [label="(768,)", style=solid];
+"2313 compressed_weight_updated_constant91" -> "2314 quantize_per_channel_default_113"  [label="(768, 1536)", style=solid];
+"2314 quantize_per_channel_default_113" -> "2315 dequantize_per_channel_default_113"  [label="(768, 1536)", style=solid];
+"2315 dequantize_per_channel_default_113" -> "2316 linear_134"  [label="(768, 1536)", style=solid];
+"2316 linear_134" -> "2317 layer_norm_47"  [label="(1, 7, 7, 768)", style=solid];
+"2317 layer_norm_47" -> "2319 pad_25"  [label="(1, 7, 7, 768)", style=solid];
+"2317 layer_norm_47" -> "2380 add_78"  [label="(1, 7, 7, 768)", style=solid];
+"2318 _frozen_param187" -> "2354 add_77"  [label="(1, 24, 64, 64)", style=solid];
+"2319 pad_25" -> "2320 view_123"  [label="(1, 8, 8, 768)", style=solid];
+"2320 view_123" -> "2321 permute_101"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2321 permute_101" -> "2322 reshape_99"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2322 reshape_99" -> "2323 reshape_99_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
+"2323 reshape_99_0_0_nncf_smooth_quant_0" -> "2324 quantize_per_tensor_default_136"  [label="(1, 64, 768)", style=solid];
+"2324 quantize_per_tensor_default_136" -> "2325 dequantize_per_tensor_default_136"  [label="(1, 64, 768)", style=solid];
+"2325 dequantize_per_tensor_default_136" -> "2332 linear_137"  [label="(1, 64, 768)", style=solid];
+"2326 linear_137_scale_0" -> "2329 quantize_per_channel_default_115"  [label="(2304,)", style=solid];
+"2326 linear_137_scale_0" -> "2330 dequantize_per_channel_default_115"  [label="(2304,)", style=solid];
+"2327 linear_137_zero_point_0" -> "2329 quantize_per_channel_default_115"  [label="(2304,)", style=solid];
+"2327 linear_137_zero_point_0" -> "2330 dequantize_per_channel_default_115"  [label="(2304,)", style=solid];
+"2328 compressed_weight_updated_constant92" -> "2329 quantize_per_channel_default_115"  [label="(2304, 768)", style=solid];
+"2329 quantize_per_channel_default_115" -> "2330 dequantize_per_channel_default_115"  [label="(2304, 768)", style=solid];
+"2330 dequantize_per_channel_default_115" -> "2332 linear_137"  [label="(2304, 768)", style=solid];
+"2331 _frozen_param188_0_0" -> "2332 linear_137"  [label="(2304,)", style=solid];
+"2332 linear_137" -> "2333 reshape_100"  [label="(1, 64, 2304)", style=solid];
+"2333 reshape_100" -> "2334 permute_102"  [label="(1, 64, 3, 24, 32)", style=solid];
+"2334 permute_102" -> "2335 select_66"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2334 permute_102" -> "2336 select_67"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2334 permute_102" -> "2337 select_68"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2335 select_66" -> "2338 linalg_vector_norm_44"  [label="(1, 24, 64, 32)", style=solid];
+"2335 select_66" -> "2340 expand_as_44"  [label="(1, 24, 64, 32)", style=solid];
+"2335 select_66" -> "2341 div_44"  [label="(1, 24, 64, 32)", style=solid];
+"2336 select_67" -> "2344 linalg_vector_norm_45"  [label="(1, 24, 64, 32)", style=solid];
+"2336 select_67" -> "2346 expand_as_45"  [label="(1, 24, 64, 32)", style=solid];
+"2336 select_67" -> "2347 div_45"  [label="(1, 24, 64, 32)", style=solid];
+"2337 select_68" -> "2357 matmul_45"  [label="(1, 24, 64, 32)", style=solid];
+"2338 linalg_vector_norm_44" -> "2339 clamp_min_44"  [label="(1, 24, 64, 1)", style=solid];
+"2339 clamp_min_44" -> "2340 expand_as_44"  [label="(1, 24, 64, 1)", style=solid];
+"2340 expand_as_44" -> "2341 div_44"  [label="(1, 24, 64, 32)", style=solid];
+"2341 div_44" -> "2342 quantize_per_tensor_default_137"  [label="(1, 24, 64, 32)", style=solid];
+"2342 quantize_per_tensor_default_137" -> "2343 dequantize_per_tensor_default_137"  [label="(1, 24, 64, 32)", style=solid];
+"2343 dequantize_per_tensor_default_137" -> "2351 matmul_44"  [label="(1, 24, 64, 32)", style=solid];
+"2344 linalg_vector_norm_45" -> "2345 clamp_min_45"  [label="(1, 24, 64, 1)", style=solid];
+"2345 clamp_min_45" -> "2346 expand_as_45"  [label="(1, 24, 64, 1)", style=solid];
+"2346 expand_as_45" -> "2347 div_45"  [label="(1, 24, 64, 32)", style=solid];
+"2347 div_45" -> "2348 quantize_per_tensor_default_138"  [label="(1, 24, 64, 32)", style=solid];
+"2348 quantize_per_tensor_default_138" -> "2349 dequantize_per_tensor_default_138"  [label="(1, 24, 64, 32)", style=solid];
+"2349 dequantize_per_tensor_default_138" -> "2350 transpose_44"  [label="(1, 24, 64, 32)", style=solid];
+"2350 transpose_44" -> "2351 matmul_44"  [label="(1, 24, 32, 64)", style=solid];
+"2351 matmul_44" -> "2353 mul_45"  [label="(1, 24, 64, 64)", style=solid];
+"2352 _frozen_param190" -> "2353 mul_45"  [label="(24, 1, 1)", style=solid];
+"2353 mul_45" -> "2354 add_77"  [label="(1, 24, 64, 64)", style=solid];
+"2354 add_77" -> "2355 softmax_22"  [label="(1, 24, 64, 64)", style=solid];
+"2355 softmax_22" -> "2356 dropout_88"  [label="(1, 24, 64, 64)", style=solid];
+"2356 dropout_88" -> "2357 matmul_45"  [label="(1, 24, 64, 64)", style=solid];
+"2357 matmul_45" -> "2358 transpose_45"  [label="(1, 24, 64, 32)", style=solid];
+"2358 transpose_45" -> "2359 reshape_101"  [label="(1, 64, 24, 32)", style=solid];
+"2359 reshape_101" -> "2360 reshape_101_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
+"2360 reshape_101_0_0_nncf_smooth_quant_0" -> "2361 quantize_per_tensor_default_139"  [label="(1, 64, 768)", style=solid];
+"2361 quantize_per_tensor_default_139" -> "2362 dequantize_per_tensor_default_139"  [label="(1, 64, 768)", style=solid];
+"2362 dequantize_per_tensor_default_139" -> "2369 linear_138"  [label="(1, 64, 768)", style=solid];
+"2363 linear_138_scale_0" -> "2366 quantize_per_channel_default_116"  [label="(768,)", style=solid];
+"2363 linear_138_scale_0" -> "2367 dequantize_per_channel_default_116"  [label="(768,)", style=solid];
+"2364 linear_138_zero_point_0" -> "2366 quantize_per_channel_default_116"  [label="(768,)", style=solid];
+"2364 linear_138_zero_point_0" -> "2367 dequantize_per_channel_default_116"  [label="(768,)", style=solid];
+"2365 compressed_weight_updated_constant93" -> "2366 quantize_per_channel_default_116"  [label="(768, 768)", style=solid];
+"2366 quantize_per_channel_default_116" -> "2367 dequantize_per_channel_default_116"  [label="(768, 768)", style=solid];
+"2367 dequantize_per_channel_default_116" -> "2369 linear_138"  [label="(768, 768)", style=solid];
+"2368 features_7_0_attn_proj_bias_0_0" -> "2369 linear_138"  [label="(768,)", style=solid];
+"2369 linear_138" -> "2370 dropout_89"  [label="(1, 64, 768)", style=solid];
+"2370 dropout_89" -> "2371 view_124"  [label="(1, 64, 768)", style=solid];
+"2371 view_124" -> "2372 permute_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2372 permute_103" -> "2373 reshape_102"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2373 reshape_102" -> "2374 slice_342"  [label="(1, 8, 8, 768)", style=solid];
+"2374 slice_342" -> "2375 slice_343"  [label="(1, 8, 8, 768)", style=solid];
+"2375 slice_343" -> "2376 slice_344"  [label="(1, 7, 8, 768)", style=solid];
+"2376 slice_344" -> "2377 slice_345"  [label="(1, 7, 7, 768)", style=solid];
+"2377 slice_345" -> "2378 contiguous_43"  [label="(1, 7, 7, 768)", style=solid];
+"2378 contiguous_43" -> "2379 layer_norm_48"  [label="(1, 7, 7, 768)", style=solid];
+"2379 layer_norm_48" -> "2380 add_78"  [label="(1, 7, 7, 768)", style=solid];
+"2380 add_78" -> "2381 add_78_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 768)", style=solid];
+"2380 add_78" -> "2405 add_79"  [label="(1, 7, 7, 768)", style=solid];
+"2381 add_78_0_0_nncf_smooth_quant_0" -> "2382 quantize_per_tensor_default_140"  [label="(1, 7, 7, 768)", style=solid];
+"2382 quantize_per_tensor_default_140" -> "2383 dequantize_per_tensor_default_140"  [label="(1, 7, 7, 768)", style=solid];
+"2383 dequantize_per_tensor_default_140" -> "2390 linear_139"  [label="(1, 7, 7, 768)", style=solid];
+"2384 linear_139_scale_0" -> "2387 quantize_per_channel_default_117"  [label="(3072,)", style=solid];
+"2384 linear_139_scale_0" -> "2388 dequantize_per_channel_default_117"  [label="(3072,)", style=solid];
+"2385 linear_139_zero_point_0" -> "2387 quantize_per_channel_default_117"  [label="(3072,)", style=solid];
+"2385 linear_139_zero_point_0" -> "2388 dequantize_per_channel_default_117"  [label="(3072,)", style=solid];
+"2386 compressed_weight_updated_constant94" -> "2387 quantize_per_channel_default_117"  [label="(3072, 768)", style=solid];
+"2387 quantize_per_channel_default_117" -> "2388 dequantize_per_channel_default_117"  [label="(3072, 768)", style=solid];
+"2388 dequantize_per_channel_default_117" -> "2390 linear_139"  [label="(3072, 768)", style=solid];
+"2389 features_7_0_mlp_0_bias_0_0" -> "2390 linear_139"  [label="(3072,)", style=solid];
+"2390 linear_139" -> "2391 gelu_22"  [label="(1, 7, 7, 3072)", style=solid];
+"2391 gelu_22" -> "2392 dropout_90"  [label="(1, 7, 7, 3072)", style=solid];
+"2392 dropout_90" -> "2393 dropout_90_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 3072)", style=solid];
+"2393 dropout_90_0_0_nncf_smooth_quant_0" -> "2394 quantize_per_tensor_default_141"  [label="(1, 7, 7, 3072)", style=solid];
+"2394 quantize_per_tensor_default_141" -> "2395 dequantize_per_tensor_default_141"  [label="(1, 7, 7, 3072)", style=solid];
+"2395 dequantize_per_tensor_default_141" -> "2402 linear_140"  [label="(1, 7, 7, 3072)", style=solid];
+"2396 linear_140_scale_0" -> "2399 quantize_per_channel_default_118"  [label="(768,)", style=solid];
+"2396 linear_140_scale_0" -> "2400 dequantize_per_channel_default_118"  [label="(768,)", style=solid];
+"2397 linear_140_zero_point_0" -> "2399 quantize_per_channel_default_118"  [label="(768,)", style=solid];
+"2397 linear_140_zero_point_0" -> "2400 dequantize_per_channel_default_118"  [label="(768,)", style=solid];
+"2398 compressed_weight_updated_constant95" -> "2399 quantize_per_channel_default_118"  [label="(768, 3072)", style=solid];
+"2399 quantize_per_channel_default_118" -> "2400 dequantize_per_channel_default_118"  [label="(768, 3072)", style=solid];
+"2400 dequantize_per_channel_default_118" -> "2402 linear_140"  [label="(768, 3072)", style=solid];
+"2401 features_7_0_mlp_3_bias_0_0" -> "2402 linear_140"  [label="(768,)", style=solid];
+"2402 linear_140" -> "2403 dropout_91"  [label="(1, 7, 7, 768)", style=solid];
+"2403 dropout_91" -> "2404 layer_norm_49"  [label="(1, 7, 7, 768)", style=solid];
+"2404 layer_norm_49" -> "2405 add_79"  [label="(1, 7, 7, 768)", style=solid];
+"2405 add_79" -> "2407 pad_26"  [label="(1, 7, 7, 768)", style=solid];
+"2405 add_79" -> "2468 add_81"  [label="(1, 7, 7, 768)", style=solid];
+"2406 _frozen_param191" -> "2442 add_80"  [label="(1, 24, 64, 64)", style=solid];
+"2407 pad_26" -> "2408 view_127"  [label="(1, 8, 8, 768)", style=solid];
+"2408 view_127" -> "2409 permute_105"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2409 permute_105" -> "2410 reshape_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2410 reshape_103" -> "2411 reshape_103_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
+"2411 reshape_103_0_0_nncf_smooth_quant_0" -> "2412 quantize_per_tensor_default_142"  [label="(1, 64, 768)", style=solid];
+"2412 quantize_per_tensor_default_142" -> "2413 dequantize_per_tensor_default_142"  [label="(1, 64, 768)", style=solid];
+"2413 dequantize_per_tensor_default_142" -> "2420 linear_143"  [label="(1, 64, 768)", style=solid];
+"2414 linear_143_scale_0" -> "2417 quantize_per_channel_default_120"  [label="(2304,)", style=solid];
+"2414 linear_143_scale_0" -> "2418 dequantize_per_channel_default_120"  [label="(2304,)", style=solid];
+"2415 linear_143_zero_point_0" -> "2417 quantize_per_channel_default_120"  [label="(2304,)", style=solid];
+"2415 linear_143_zero_point_0" -> "2418 dequantize_per_channel_default_120"  [label="(2304,)", style=solid];
+"2416 compressed_weight_updated_constant96" -> "2417 quantize_per_channel_default_120"  [label="(2304, 768)", style=solid];
+"2417 quantize_per_channel_default_120" -> "2418 dequantize_per_channel_default_120"  [label="(2304, 768)", style=solid];
+"2418 dequantize_per_channel_default_120" -> "2420 linear_143"  [label="(2304, 768)", style=solid];
+"2419 _frozen_param192_0_0" -> "2420 linear_143"  [label="(2304,)", style=solid];
+"2420 linear_143" -> "2421 reshape_104"  [label="(1, 64, 2304)", style=solid];
+"2421 reshape_104" -> "2422 permute_106"  [label="(1, 64, 3, 24, 32)", style=solid];
+"2422 permute_106" -> "2423 select_69"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2422 permute_106" -> "2424 select_70"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2422 permute_106" -> "2425 select_71"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2423 select_69" -> "2426 linalg_vector_norm_46"  [label="(1, 24, 64, 32)", style=solid];
+"2423 select_69" -> "2428 expand_as_46"  [label="(1, 24, 64, 32)", style=solid];
+"2423 select_69" -> "2429 div_46"  [label="(1, 24, 64, 32)", style=solid];
+"2424 select_70" -> "2432 linalg_vector_norm_47"  [label="(1, 24, 64, 32)", style=solid];
+"2424 select_70" -> "2434 expand_as_47"  [label="(1, 24, 64, 32)", style=solid];
+"2424 select_70" -> "2435 div_47"  [label="(1, 24, 64, 32)", style=solid];
+"2425 select_71" -> "2445 matmul_47"  [label="(1, 24, 64, 32)", style=solid];
+"2426 linalg_vector_norm_46" -> "2427 clamp_min_46"  [label="(1, 24, 64, 1)", style=solid];
+"2427 clamp_min_46" -> "2428 expand_as_46"  [label="(1, 24, 64, 1)", style=solid];
+"2428 expand_as_46" -> "2429 div_46"  [label="(1, 24, 64, 32)", style=solid];
+"2429 div_46" -> "2430 quantize_per_tensor_default_143"  [label="(1, 24, 64, 32)", style=solid];
+"2430 quantize_per_tensor_default_143" -> "2431 dequantize_per_tensor_default_143"  [label="(1, 24, 64, 32)", style=solid];
+"2431 dequantize_per_tensor_default_143" -> "2439 matmul_46"  [label="(1, 24, 64, 32)", style=solid];
+"2432 linalg_vector_norm_47" -> "2433 clamp_min_47"  [label="(1, 24, 64, 1)", style=solid];
+"2433 clamp_min_47" -> "2434 expand_as_47"  [label="(1, 24, 64, 1)", style=solid];
+"2434 expand_as_47" -> "2435 div_47"  [label="(1, 24, 64, 32)", style=solid];
+"2435 div_47" -> "2436 quantize_per_tensor_default_144"  [label="(1, 24, 64, 32)", style=solid];
+"2436 quantize_per_tensor_default_144" -> "2437 dequantize_per_tensor_default_144"  [label="(1, 24, 64, 32)", style=solid];
+"2437 dequantize_per_tensor_default_144" -> "2438 transpose_46"  [label="(1, 24, 64, 32)", style=solid];
+"2438 transpose_46" -> "2439 matmul_46"  [label="(1, 24, 32, 64)", style=solid];
+"2439 matmul_46" -> "2441 mul_47"  [label="(1, 24, 64, 64)", style=solid];
+"2440 _frozen_param194" -> "2441 mul_47"  [label="(24, 1, 1)", style=solid];
+"2441 mul_47" -> "2442 add_80"  [label="(1, 24, 64, 64)", style=solid];
+"2442 add_80" -> "2443 softmax_23"  [label="(1, 24, 64, 64)", style=solid];
+"2443 softmax_23" -> "2444 dropout_92"  [label="(1, 24, 64, 64)", style=solid];
+"2444 dropout_92" -> "2445 matmul_47"  [label="(1, 24, 64, 64)", style=solid];
+"2445 matmul_47" -> "2446 transpose_47"  [label="(1, 24, 64, 32)", style=solid];
+"2446 transpose_47" -> "2447 reshape_105"  [label="(1, 64, 24, 32)", style=solid];
+"2447 reshape_105" -> "2448 reshape_105_0_0_nncf_smooth_quant_0"  [label="(1, 64, 768)", style=solid];
+"2448 reshape_105_0_0_nncf_smooth_quant_0" -> "2449 quantize_per_tensor_default_145"  [label="(1, 64, 768)", style=solid];
+"2449 quantize_per_tensor_default_145" -> "2450 dequantize_per_tensor_default_145"  [label="(1, 64, 768)", style=solid];
+"2450 dequantize_per_tensor_default_145" -> "2457 linear_144"  [label="(1, 64, 768)", style=solid];
+"2451 linear_144_scale_0" -> "2454 quantize_per_channel_default_121"  [label="(768,)", style=solid];
+"2451 linear_144_scale_0" -> "2455 dequantize_per_channel_default_121"  [label="(768,)", style=solid];
+"2452 linear_144_zero_point_0" -> "2454 quantize_per_channel_default_121"  [label="(768,)", style=solid];
+"2452 linear_144_zero_point_0" -> "2455 dequantize_per_channel_default_121"  [label="(768,)", style=solid];
+"2453 compressed_weight_updated_constant97" -> "2454 quantize_per_channel_default_121"  [label="(768, 768)", style=solid];
+"2454 quantize_per_channel_default_121" -> "2455 dequantize_per_channel_default_121"  [label="(768, 768)", style=solid];
+"2455 dequantize_per_channel_default_121" -> "2457 linear_144"  [label="(768, 768)", style=solid];
+"2456 features_7_1_attn_proj_bias_0_0" -> "2457 linear_144"  [label="(768,)", style=solid];
+"2457 linear_144" -> "2458 dropout_93"  [label="(1, 64, 768)", style=solid];
+"2458 dropout_93" -> "2459 view_128"  [label="(1, 64, 768)", style=solid];
+"2459 view_128" -> "2460 permute_107"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2460 permute_107" -> "2461 reshape_106"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2461 reshape_106" -> "2462 slice_347"  [label="(1, 8, 8, 768)", style=solid];
+"2462 slice_347" -> "2463 slice_348"  [label="(1, 8, 8, 768)", style=solid];
+"2463 slice_348" -> "2464 slice_349"  [label="(1, 7, 8, 768)", style=solid];
+"2464 slice_349" -> "2465 slice_350"  [label="(1, 7, 7, 768)", style=solid];
+"2465 slice_350" -> "2466 contiguous_45"  [label="(1, 7, 7, 768)", style=solid];
+"2466 contiguous_45" -> "2467 layer_norm_50"  [label="(1, 7, 7, 768)", style=solid];
+"2467 layer_norm_50" -> "2468 add_81"  [label="(1, 7, 7, 768)", style=solid];
+"2468 add_81" -> "2469 add_81_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 768)", style=solid];
+"2468 add_81" -> "2493 add_82"  [label="(1, 7, 7, 768)", style=solid];
+"2469 add_81_0_0_nncf_smooth_quant_0" -> "2470 quantize_per_tensor_default_146"  [label="(1, 7, 7, 768)", style=solid];
+"2470 quantize_per_tensor_default_146" -> "2471 dequantize_per_tensor_default_146"  [label="(1, 7, 7, 768)", style=solid];
+"2471 dequantize_per_tensor_default_146" -> "2478 linear_145"  [label="(1, 7, 7, 768)", style=solid];
+"2472 linear_145_scale_0" -> "2475 quantize_per_channel_default_122"  [label="(3072,)", style=solid];
+"2472 linear_145_scale_0" -> "2476 dequantize_per_channel_default_122"  [label="(3072,)", style=solid];
+"2473 linear_145_zero_point_0" -> "2475 quantize_per_channel_default_122"  [label="(3072,)", style=solid];
+"2473 linear_145_zero_point_0" -> "2476 dequantize_per_channel_default_122"  [label="(3072,)", style=solid];
+"2474 compressed_weight_updated_constant98" -> "2475 quantize_per_channel_default_122"  [label="(3072, 768)", style=solid];
+"2475 quantize_per_channel_default_122" -> "2476 dequantize_per_channel_default_122"  [label="(3072, 768)", style=solid];
+"2476 dequantize_per_channel_default_122" -> "2478 linear_145"  [label="(3072, 768)", style=solid];
+"2477 features_7_1_mlp_0_bias_0_0" -> "2478 linear_145"  [label="(3072,)", style=solid];
+"2478 linear_145" -> "2479 gelu_23"  [label="(1, 7, 7, 3072)", style=solid];
+"2479 gelu_23" -> "2480 dropout_94"  [label="(1, 7, 7, 3072)", style=solid];
+"2480 dropout_94" -> "2481 dropout_94_0_0_nncf_smooth_quant_0"  [label="(1, 7, 7, 3072)", style=solid];
+"2481 dropout_94_0_0_nncf_smooth_quant_0" -> "2482 quantize_per_tensor_default_147"  [label="(1, 7, 7, 3072)", style=solid];
+"2482 quantize_per_tensor_default_147" -> "2483 dequantize_per_tensor_default_147"  [label="(1, 7, 7, 3072)", style=solid];
+"2483 dequantize_per_tensor_default_147" -> "2490 linear_146"  [label="(1, 7, 7, 3072)", style=solid];
+"2484 linear_146_scale_0" -> "2487 quantize_per_channel_default_123"  [label="(768,)", style=solid];
+"2484 linear_146_scale_0" -> "2488 dequantize_per_channel_default_123"  [label="(768,)", style=solid];
+"2485 linear_146_zero_point_0" -> "2487 quantize_per_channel_default_123"  [label="(768,)", style=solid];
+"2485 linear_146_zero_point_0" -> "2488 dequantize_per_channel_default_123"  [label="(768,)", style=solid];
+"2486 compressed_weight_updated_constant99" -> "2487 quantize_per_channel_default_123"  [label="(768, 3072)", style=solid];
+"2487 quantize_per_channel_default_123" -> "2488 dequantize_per_channel_default_123"  [label="(768, 3072)", style=solid];
+"2488 dequantize_per_channel_default_123" -> "2490 linear_146"  [label="(768, 3072)", style=solid];
+"2489 features_7_1_mlp_3_bias_0_0" -> "2490 linear_146"  [label="(768,)", style=solid];
+"2490 linear_146" -> "2491 dropout_95"  [label="(1, 7, 7, 768)", style=solid];
+"2491 dropout_95" -> "2492 layer_norm_51"  [label="(1, 7, 7, 768)", style=solid];
+"2492 layer_norm_51" -> "2493 add_82"  [label="(1, 7, 7, 768)", style=solid];
+"2493 add_82" -> "2494 layer_norm_52"  [label="(1, 7, 7, 768)", style=solid];
 "2494 layer_norm_52" -> "2495 permute_108"  [label="(1, 7, 7, 768)", style=solid];
 "2495 permute_108" -> "2496 adaptive_avg_pool2d"  [label="(1, 768, 7, 7)", style=solid];
 "2496 adaptive_avg_pool2d" -> "2497 flatten"  [label="(1, 768, 1, 1)", style=solid];
@@ -5461,6 +5461,6 @@ strict digraph  {
 "2503 compressed_weight_updated_constant100" -> "2504 quantize_per_channel_default_124"  [label="(1000, 768)", style=solid];
 "2504 quantize_per_channel_default_124" -> "2505 dequantize_per_channel_default_124"  [label="(1000, 768)", style=solid];
 "2505 dequantize_per_channel_default_124" -> "2507 linear_147"  [label="(1000, 768)", style=solid];
-"2506 _param_constant400_0_0" -> "2507 linear_147"  [label="(1000,)", style=solid];
+"2506 head_bias_0_0" -> "2507 linear_147"  [label="(1000,)", style=solid];
 "2507 linear_147" -> "2508 output"  [label="(1, 1000)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/quantized/synthetic_transformer.dot b/tests/torch/data/reference_graphs/fx/quantized/synthetic_transformer.dot
index a3a6542ed3c..3fd5285222d 100644
--- a/tests/torch/data/reference_graphs/fx/quantized/synthetic_transformer.dot
+++ b/tests/torch/data/reference_graphs/fx/quantized/synthetic_transformer.dot
@@ -1,6 +1,6 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
+"0 wte_weight" [id=0, type=get_attr];
+"1 input_ids" [id=1, type=input];
 "2 embedding" [id=2, type=embedding];
 "3 embedding_0_0_nncf_smooth_quant_0" [id=3, type=call_module];
 "4 quantize_per_tensor_default" [id=4, type=quantize_per_tensor];
@@ -10,7 +10,7 @@ strict digraph  {
 "8 compressed_weight_updated_constant0" [id=8, type=get_attr];
 "9 quantize_per_channel_default" [id=9, type=quantize_per_channel];
 "10 dequantize_per_channel_default" [id=10, type=dequantize_per_channel];
-"11 _param_constant2_0_0" [id=11, type=get_attr];
+"11 linear_bias_0_0" [id=11, type=get_attr];
 "12 linear" [id=12, type=linear];
 "13 add_tensor_0_0_nncf_smooth_quant_0" [id=13, type=call_module];
 "14 quantize_per_tensor_default_1" [id=14, type=quantize_per_tensor];
@@ -20,11 +20,11 @@ strict digraph  {
 "18 compressed_weight_updated_constant1" [id=18, type=get_attr];
 "19 quantize_per_channel_default_1" [id=19, type=quantize_per_channel];
 "20 dequantize_per_channel_default_1" [id=20, type=dequantize_per_channel];
-"21 _param_constant4_0_0" [id=21, type=get_attr];
+"21 lm_head_bias_0_0" [id=21, type=get_attr];
 "22 linear_1" [id=22, type=linear];
 "23 output" [id=23, type=output];
-"0 arg0_1" -> "2 embedding"  [label="(5,)", style=solid];
-"1 _param_constant0" -> "2 embedding"  [label="(10, 5)", style=solid];
+"0 wte_weight" -> "2 embedding"  [label="(10, 5)", style=solid];
+"1 input_ids" -> "2 embedding"  [label="(5,)", style=solid];
 "2 embedding" -> "3 embedding_0_0_nncf_smooth_quant_0"  [label="(5, 5)", style=solid];
 "3 embedding_0_0_nncf_smooth_quant_0" -> "4 quantize_per_tensor_default"  [label="(5, 5)", style=solid];
 "4 quantize_per_tensor_default" -> "5 dequantize_per_tensor_default"  [label="(5, 5)", style=solid];
@@ -36,7 +36,7 @@ strict digraph  {
 "8 compressed_weight_updated_constant0" -> "9 quantize_per_channel_default"  [label="(5, 5)", style=solid];
 "9 quantize_per_channel_default" -> "10 dequantize_per_channel_default"  [label="(5, 5)", style=solid];
 "10 dequantize_per_channel_default" -> "12 linear"  [label="(5, 5)", style=solid];
-"11 _param_constant2_0_0" -> "12 linear"  [label="(5,)", style=solid];
+"11 linear_bias_0_0" -> "12 linear"  [label="(5,)", style=solid];
 "12 linear" -> "13 add_tensor_0_0_nncf_smooth_quant_0"  [label="(5, 5)", style=solid];
 "13 add_tensor_0_0_nncf_smooth_quant_0" -> "14 quantize_per_tensor_default_1"  [label="(5, 5)", style=solid];
 "14 quantize_per_tensor_default_1" -> "15 dequantize_per_tensor_default_1"  [label="(5, 5)", style=solid];
@@ -48,6 +48,6 @@ strict digraph  {
 "18 compressed_weight_updated_constant1" -> "19 quantize_per_channel_default_1"  [label="(10, 5)", style=solid];
 "19 quantize_per_channel_default_1" -> "20 dequantize_per_channel_default_1"  [label="(10, 5)", style=solid];
 "20 dequantize_per_channel_default_1" -> "22 linear_1"  [label="(10, 5)", style=solid];
-"21 _param_constant4_0_0" -> "22 linear_1"  [label="(10,)", style=solid];
+"21 lm_head_bias_0_0" -> "22 linear_1"  [label="(10,)", style=solid];
 "22 linear_1" -> "23 output"  [label="(5, 10)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/quantized/unet.dot b/tests/torch/data/reference_graphs/fx/quantized/unet.dot
index 22972f24a32..b299be70340 100644
--- a/tests/torch/data/reference_graphs/fx/quantized/unet.dot
+++ b/tests/torch/data/reference_graphs/fx/quantized/unet.dot
@@ -1,5 +1,5 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
+"0 x" [id=0, type=input];
 "1 quantize_per_tensor_default_8" [id=1, type=quantize_per_tensor];
 "2 dequantize_per_tensor_default_12" [id=2, type=dequantize_per_tensor];
 "3 conv2d_scale_0" [id=3, type=get_attr];
@@ -7,7 +7,7 @@ strict digraph  {
 "5 compressed_weight_updated_constant0" [id=5, type=get_attr];
 "6 quantize_per_channel_default" [id=6, type=quantize_per_channel];
 "7 dequantize_per_channel_default" [id=7, type=dequantize_per_channel];
-"8 _param_constant1_0_0" [id=8, type=get_attr];
+"8 down_path_0_block_0_bias_0_0" [id=8, type=get_attr];
 "9 conv2d" [id=9, type=conv2d];
 "10 relu" [id=10, type=relu];
 "11 quantize_per_tensor_default_9" [id=11, type=quantize_per_tensor];
@@ -17,7 +17,7 @@ strict digraph  {
 "15 compressed_weight_updated_constant1" [id=15, type=get_attr];
 "16 quantize_per_channel_default_1" [id=16, type=quantize_per_channel];
 "17 dequantize_per_channel_default_1" [id=17, type=dequantize_per_channel];
-"18 _param_constant5_0_0" [id=18, type=get_attr];
+"18 down_path_0_block_3_bias_0_0" [id=18, type=get_attr];
 "19 conv2d_1" [id=19, type=conv2d];
 "20 relu_1" [id=20, type=relu];
 "21 quantize_per_tensor_default" [id=21, type=quantize_per_tensor];
@@ -29,7 +29,7 @@ strict digraph  {
 "27 compressed_weight_updated_constant2" [id=27, type=get_attr];
 "28 quantize_per_channel_default_2" [id=28, type=quantize_per_channel];
 "29 dequantize_per_channel_default_2" [id=29, type=dequantize_per_channel];
-"30 _param_constant9_0_0" [id=30, type=get_attr];
+"30 down_path_1_block_0_bias_0_0" [id=30, type=get_attr];
 "31 conv2d_2" [id=31, type=conv2d];
 "32 relu_2" [id=32, type=relu];
 "33 quantize_per_tensor_default_10" [id=33, type=quantize_per_tensor];
@@ -39,7 +39,7 @@ strict digraph  {
 "37 compressed_weight_updated_constant3" [id=37, type=get_attr];
 "38 quantize_per_channel_default_3" [id=38, type=quantize_per_channel];
 "39 dequantize_per_channel_default_3" [id=39, type=dequantize_per_channel];
-"40 _param_constant13_0_0" [id=40, type=get_attr];
+"40 down_path_1_block_3_bias_0_0" [id=40, type=get_attr];
 "41 conv2d_3" [id=41, type=conv2d];
 "42 relu_3" [id=42, type=relu];
 "43 quantize_per_tensor_default_3" [id=43, type=quantize_per_tensor];
@@ -51,7 +51,7 @@ strict digraph  {
 "49 compressed_weight_updated_constant4" [id=49, type=get_attr];
 "50 quantize_per_channel_default_4" [id=50, type=quantize_per_channel];
 "51 dequantize_per_channel_default_4" [id=51, type=dequantize_per_channel];
-"52 _param_constant17_0_0" [id=52, type=get_attr];
+"52 down_path_2_block_0_bias_0_0" [id=52, type=get_attr];
 "53 conv2d_4" [id=53, type=conv2d];
 "54 relu_4" [id=54, type=relu];
 "55 quantize_per_tensor_default_11" [id=55, type=quantize_per_tensor];
@@ -61,7 +61,7 @@ strict digraph  {
 "59 compressed_weight_updated_constant5" [id=59, type=get_attr];
 "60 quantize_per_channel_default_5" [id=60, type=quantize_per_channel];
 "61 dequantize_per_channel_default_5" [id=61, type=dequantize_per_channel];
-"62 _param_constant21_0_0" [id=62, type=get_attr];
+"62 down_path_2_block_3_bias_0_0" [id=62, type=get_attr];
 "63 conv2d_5" [id=63, type=conv2d];
 "64 relu_5" [id=64, type=relu];
 "65 quantize_per_tensor_default_4" [id=65, type=quantize_per_tensor];
@@ -73,7 +73,7 @@ strict digraph  {
 "71 compressed_weight_updated_constant6" [id=71, type=get_attr];
 "72 quantize_per_channel_default_6" [id=72, type=quantize_per_channel];
 "73 dequantize_per_channel_default_6" [id=73, type=dequantize_per_channel];
-"74 _param_constant25_0_0" [id=74, type=get_attr];
+"74 down_path_3_block_0_bias_0_0" [id=74, type=get_attr];
 "75 conv2d_6" [id=75, type=conv2d];
 "76 relu_6" [id=76, type=relu];
 "77 quantize_per_tensor_default_12" [id=77, type=quantize_per_tensor];
@@ -83,7 +83,7 @@ strict digraph  {
 "81 compressed_weight_updated_constant7" [id=81, type=get_attr];
 "82 quantize_per_channel_default_7" [id=82, type=quantize_per_channel];
 "83 dequantize_per_channel_default_7" [id=83, type=dequantize_per_channel];
-"84 _param_constant29_0_0" [id=84, type=get_attr];
+"84 down_path_3_block_3_bias_0_0" [id=84, type=get_attr];
 "85 conv2d_7" [id=85, type=conv2d];
 "86 relu_7" [id=86, type=relu];
 "87 quantize_per_tensor_default_7" [id=87, type=quantize_per_tensor];
@@ -95,7 +95,7 @@ strict digraph  {
 "93 compressed_weight_updated_constant8" [id=93, type=get_attr];
 "94 quantize_per_channel_default_8" [id=94, type=quantize_per_channel];
 "95 dequantize_per_channel_default_8" [id=95, type=dequantize_per_channel];
-"96 _param_constant33_0_0" [id=96, type=get_attr];
+"96 down_path_4_block_0_bias_0_0" [id=96, type=get_attr];
 "97 conv2d_8" [id=97, type=conv2d];
 "98 relu_8" [id=98, type=relu];
 "99 quantize_per_tensor_default_13" [id=99, type=quantize_per_tensor];
@@ -105,7 +105,7 @@ strict digraph  {
 "103 compressed_weight_updated_constant9" [id=103, type=get_attr];
 "104 quantize_per_channel_default_9" [id=104, type=quantize_per_channel];
 "105 dequantize_per_channel_default_9" [id=105, type=dequantize_per_channel];
-"106 _param_constant37_0_0" [id=106, type=get_attr];
+"106 down_path_4_block_3_bias_0_0" [id=106, type=get_attr];
 "107 conv2d_9" [id=107, type=conv2d];
 "108 relu_9" [id=108, type=relu];
 "109 quantize_per_tensor_default_14" [id=109, type=quantize_per_tensor];
@@ -115,7 +115,7 @@ strict digraph  {
 "113 compressed_weight_updated_constant10" [id=113, type=get_attr];
 "114 quantize_per_channel_default_10" [id=114, type=quantize_per_channel];
 "115 dequantize_per_channel_default_10" [id=115, type=dequantize_per_channel];
-"116 _param_constant41_0_0" [id=116, type=get_attr];
+"116 up_path_0_up_bias_0_0" [id=116, type=get_attr];
 "117 conv_transpose2d" [id=117, type=conv_transpose2d];
 "118 quantize_per_tensor_default_6" [id=118, type=quantize_per_tensor];
 "119 dequantize_per_tensor_default_9" [id=119, type=dequantize_per_tensor];
@@ -129,7 +129,7 @@ strict digraph  {
 "127 compressed_weight_updated_constant11" [id=127, type=get_attr];
 "128 quantize_per_channel_default_11" [id=128, type=quantize_per_channel];
 "129 dequantize_per_channel_default_11" [id=129, type=dequantize_per_channel];
-"130 _param_constant43_0_0" [id=130, type=get_attr];
+"130 up_path_0_conv_block_block_0_bias_0_0" [id=130, type=get_attr];
 "131 conv2d_10" [id=131, type=conv2d];
 "132 relu_10" [id=132, type=relu];
 "133 quantize_per_tensor_default_15" [id=133, type=quantize_per_tensor];
@@ -139,7 +139,7 @@ strict digraph  {
 "137 compressed_weight_updated_constant12" [id=137, type=get_attr];
 "138 quantize_per_channel_default_12" [id=138, type=quantize_per_channel];
 "139 dequantize_per_channel_default_12" [id=139, type=dequantize_per_channel];
-"140 _param_constant47_0_0" [id=140, type=get_attr];
+"140 up_path_0_conv_block_block_3_bias_0_0" [id=140, type=get_attr];
 "141 conv2d_11" [id=141, type=conv2d];
 "142 relu_11" [id=142, type=relu];
 "143 quantize_per_tensor_default_16" [id=143, type=quantize_per_tensor];
@@ -149,7 +149,7 @@ strict digraph  {
 "147 compressed_weight_updated_constant13" [id=147, type=get_attr];
 "148 quantize_per_channel_default_13" [id=148, type=quantize_per_channel];
 "149 dequantize_per_channel_default_13" [id=149, type=dequantize_per_channel];
-"150 _param_constant51_0_0" [id=150, type=get_attr];
+"150 up_path_1_up_bias_0_0" [id=150, type=get_attr];
 "151 conv_transpose2d_1" [id=151, type=conv_transpose2d];
 "152 quantize_per_tensor_default_5" [id=152, type=quantize_per_tensor];
 "153 dequantize_per_tensor_default_8" [id=153, type=dequantize_per_tensor];
@@ -163,7 +163,7 @@ strict digraph  {
 "161 compressed_weight_updated_constant14" [id=161, type=get_attr];
 "162 quantize_per_channel_default_14" [id=162, type=quantize_per_channel];
 "163 dequantize_per_channel_default_14" [id=163, type=dequantize_per_channel];
-"164 _param_constant53_0_0" [id=164, type=get_attr];
+"164 up_path_1_conv_block_block_0_bias_0_0" [id=164, type=get_attr];
 "165 conv2d_12" [id=165, type=conv2d];
 "166 relu_12" [id=166, type=relu];
 "167 quantize_per_tensor_default_17" [id=167, type=quantize_per_tensor];
@@ -173,7 +173,7 @@ strict digraph  {
 "171 compressed_weight_updated_constant15" [id=171, type=get_attr];
 "172 quantize_per_channel_default_15" [id=172, type=quantize_per_channel];
 "173 dequantize_per_channel_default_15" [id=173, type=dequantize_per_channel];
-"174 _param_constant57_0_0" [id=174, type=get_attr];
+"174 up_path_1_conv_block_block_3_bias_0_0" [id=174, type=get_attr];
 "175 conv2d_13" [id=175, type=conv2d];
 "176 relu_13" [id=176, type=relu];
 "177 quantize_per_tensor_default_18" [id=177, type=quantize_per_tensor];
@@ -183,7 +183,7 @@ strict digraph  {
 "181 compressed_weight_updated_constant16" [id=181, type=get_attr];
 "182 quantize_per_channel_default_16" [id=182, type=quantize_per_channel];
 "183 dequantize_per_channel_default_16" [id=183, type=dequantize_per_channel];
-"184 _param_constant61_0_0" [id=184, type=get_attr];
+"184 up_path_2_up_bias_0_0" [id=184, type=get_attr];
 "185 conv_transpose2d_2" [id=185, type=conv_transpose2d];
 "186 quantize_per_tensor_default_2" [id=186, type=quantize_per_tensor];
 "187 dequantize_per_tensor_default_3" [id=187, type=dequantize_per_tensor];
@@ -197,7 +197,7 @@ strict digraph  {
 "195 compressed_weight_updated_constant17" [id=195, type=get_attr];
 "196 quantize_per_channel_default_17" [id=196, type=quantize_per_channel];
 "197 dequantize_per_channel_default_17" [id=197, type=dequantize_per_channel];
-"198 _param_constant63_0_0" [id=198, type=get_attr];
+"198 up_path_2_conv_block_block_0_bias_0_0" [id=198, type=get_attr];
 "199 conv2d_14" [id=199, type=conv2d];
 "200 relu_14" [id=200, type=relu];
 "201 quantize_per_tensor_default_19" [id=201, type=quantize_per_tensor];
@@ -207,7 +207,7 @@ strict digraph  {
 "205 compressed_weight_updated_constant18" [id=205, type=get_attr];
 "206 quantize_per_channel_default_18" [id=206, type=quantize_per_channel];
 "207 dequantize_per_channel_default_18" [id=207, type=dequantize_per_channel];
-"208 _param_constant67_0_0" [id=208, type=get_attr];
+"208 up_path_2_conv_block_block_3_bias_0_0" [id=208, type=get_attr];
 "209 conv2d_15" [id=209, type=conv2d];
 "210 relu_15" [id=210, type=relu];
 "211 quantize_per_tensor_default_20" [id=211, type=quantize_per_tensor];
@@ -217,7 +217,7 @@ strict digraph  {
 "215 compressed_weight_updated_constant19" [id=215, type=get_attr];
 "216 quantize_per_channel_default_19" [id=216, type=quantize_per_channel];
 "217 dequantize_per_channel_default_19" [id=217, type=dequantize_per_channel];
-"218 _param_constant71_0_0" [id=218, type=get_attr];
+"218 up_path_3_up_bias_0_0" [id=218, type=get_attr];
 "219 conv_transpose2d_3" [id=219, type=conv_transpose2d];
 "220 quantize_per_tensor_default_1" [id=220, type=quantize_per_tensor];
 "221 dequantize_per_tensor_default_2" [id=221, type=dequantize_per_tensor];
@@ -231,7 +231,7 @@ strict digraph  {
 "229 compressed_weight_updated_constant20" [id=229, type=get_attr];
 "230 quantize_per_channel_default_20" [id=230, type=quantize_per_channel];
 "231 dequantize_per_channel_default_20" [id=231, type=dequantize_per_channel];
-"232 _param_constant73_0_0" [id=232, type=get_attr];
+"232 up_path_3_conv_block_block_0_bias_0_0" [id=232, type=get_attr];
 "233 conv2d_16" [id=233, type=conv2d];
 "234 relu_16" [id=234, type=relu];
 "235 quantize_per_tensor_default_21" [id=235, type=quantize_per_tensor];
@@ -241,7 +241,7 @@ strict digraph  {
 "239 compressed_weight_updated_constant21" [id=239, type=get_attr];
 "240 quantize_per_channel_default_21" [id=240, type=quantize_per_channel];
 "241 dequantize_per_channel_default_21" [id=241, type=dequantize_per_channel];
-"242 _param_constant77_0_0" [id=242, type=get_attr];
+"242 up_path_3_conv_block_block_3_bias_0_0" [id=242, type=get_attr];
 "243 conv2d_17" [id=243, type=conv2d];
 "244 relu_17" [id=244, type=relu];
 "245 quantize_per_tensor_default_22" [id=245, type=quantize_per_tensor];
@@ -251,10 +251,10 @@ strict digraph  {
 "249 compressed_weight_updated_constant22" [id=249, type=get_attr];
 "250 quantize_per_channel_default_22" [id=250, type=quantize_per_channel];
 "251 dequantize_per_channel_default_22" [id=251, type=dequantize_per_channel];
-"252 _param_constant81_0_0" [id=252, type=get_attr];
+"252 last_bias_0_0" [id=252, type=get_attr];
 "253 conv2d_18" [id=253, type=conv2d];
 "254 output" [id=254, type=output];
-"0 arg0_1" -> "1 quantize_per_tensor_default_8"  [label="(1, 3, 224, 224)", style=solid];
+"0 x" -> "1 quantize_per_tensor_default_8"  [label="(1, 3, 224, 224)", style=solid];
 "1 quantize_per_tensor_default_8" -> "2 dequantize_per_tensor_default_12"  [label="(1, 3, 224, 224)", style=solid];
 "2 dequantize_per_tensor_default_12" -> "9 conv2d"  [label="(1, 3, 224, 224)", style=solid];
 "3 conv2d_scale_0" -> "6 quantize_per_channel_default"  [label="(64,)", style=solid];
@@ -264,7 +264,7 @@ strict digraph  {
 "5 compressed_weight_updated_constant0" -> "6 quantize_per_channel_default"  [label="(64, 3, 3, 3)", style=solid];
 "6 quantize_per_channel_default" -> "7 dequantize_per_channel_default"  [label="(64, 3, 3, 3)", style=solid];
 "7 dequantize_per_channel_default" -> "9 conv2d"  [label="(64, 3, 3, 3)", style=solid];
-"8 _param_constant1_0_0" -> "9 conv2d"  [label="(64,)", style=solid];
+"8 down_path_0_block_0_bias_0_0" -> "9 conv2d"  [label="(64,)", style=solid];
 "9 conv2d" -> "10 relu"  [label="(1, 64, 222, 222)", style=solid];
 "10 relu" -> "11 quantize_per_tensor_default_9"  [label="(1, 64, 222, 222)", style=solid];
 "11 quantize_per_tensor_default_9" -> "12 dequantize_per_tensor_default_13"  [label="(1, 64, 222, 222)", style=solid];
@@ -276,7 +276,7 @@ strict digraph  {
 "15 compressed_weight_updated_constant1" -> "16 quantize_per_channel_default_1"  [label="(64, 64, 3, 3)", style=solid];
 "16 quantize_per_channel_default_1" -> "17 dequantize_per_channel_default_1"  [label="(64, 64, 3, 3)", style=solid];
 "17 dequantize_per_channel_default_1" -> "19 conv2d_1"  [label="(64, 64, 3, 3)", style=solid];
-"18 _param_constant5_0_0" -> "19 conv2d_1"  [label="(64,)", style=solid];
+"18 down_path_0_block_3_bias_0_0" -> "19 conv2d_1"  [label="(64,)", style=solid];
 "19 conv2d_1" -> "20 relu_1"  [label="(1, 64, 220, 220)", style=solid];
 "20 relu_1" -> "21 quantize_per_tensor_default"  [label="(1, 64, 220, 220)", style=solid];
 "21 quantize_per_tensor_default" -> "22 dequantize_per_tensor_default_1"  [label="(1, 64, 220, 220)", style=solid];
@@ -291,7 +291,7 @@ strict digraph  {
 "27 compressed_weight_updated_constant2" -> "28 quantize_per_channel_default_2"  [label="(128, 64, 3, 3)", style=solid];
 "28 quantize_per_channel_default_2" -> "29 dequantize_per_channel_default_2"  [label="(128, 64, 3, 3)", style=solid];
 "29 dequantize_per_channel_default_2" -> "31 conv2d_2"  [label="(128, 64, 3, 3)", style=solid];
-"30 _param_constant9_0_0" -> "31 conv2d_2"  [label="(128,)", style=solid];
+"30 down_path_1_block_0_bias_0_0" -> "31 conv2d_2"  [label="(128,)", style=solid];
 "31 conv2d_2" -> "32 relu_2"  [label="(1, 128, 108, 108)", style=solid];
 "32 relu_2" -> "33 quantize_per_tensor_default_10"  [label="(1, 128, 108, 108)", style=solid];
 "33 quantize_per_tensor_default_10" -> "34 dequantize_per_tensor_default_14"  [label="(1, 128, 108, 108)", style=solid];
@@ -303,7 +303,7 @@ strict digraph  {
 "37 compressed_weight_updated_constant3" -> "38 quantize_per_channel_default_3"  [label="(128, 128, 3, 3)", style=solid];
 "38 quantize_per_channel_default_3" -> "39 dequantize_per_channel_default_3"  [label="(128, 128, 3, 3)", style=solid];
 "39 dequantize_per_channel_default_3" -> "41 conv2d_3"  [label="(128, 128, 3, 3)", style=solid];
-"40 _param_constant13_0_0" -> "41 conv2d_3"  [label="(128,)", style=solid];
+"40 down_path_1_block_3_bias_0_0" -> "41 conv2d_3"  [label="(128,)", style=solid];
 "41 conv2d_3" -> "42 relu_3"  [label="(1, 128, 106, 106)", style=solid];
 "42 relu_3" -> "43 quantize_per_tensor_default_3"  [label="(1, 128, 106, 106)", style=solid];
 "43 quantize_per_tensor_default_3" -> "44 dequantize_per_tensor_default_5"  [label="(1, 128, 106, 106)", style=solid];
@@ -318,7 +318,7 @@ strict digraph  {
 "49 compressed_weight_updated_constant4" -> "50 quantize_per_channel_default_4"  [label="(256, 128, 3, 3)", style=solid];
 "50 quantize_per_channel_default_4" -> "51 dequantize_per_channel_default_4"  [label="(256, 128, 3, 3)", style=solid];
 "51 dequantize_per_channel_default_4" -> "53 conv2d_4"  [label="(256, 128, 3, 3)", style=solid];
-"52 _param_constant17_0_0" -> "53 conv2d_4"  [label="(256,)", style=solid];
+"52 down_path_2_block_0_bias_0_0" -> "53 conv2d_4"  [label="(256,)", style=solid];
 "53 conv2d_4" -> "54 relu_4"  [label="(1, 256, 51, 51)", style=solid];
 "54 relu_4" -> "55 quantize_per_tensor_default_11"  [label="(1, 256, 51, 51)", style=solid];
 "55 quantize_per_tensor_default_11" -> "56 dequantize_per_tensor_default_15"  [label="(1, 256, 51, 51)", style=solid];
@@ -330,7 +330,7 @@ strict digraph  {
 "59 compressed_weight_updated_constant5" -> "60 quantize_per_channel_default_5"  [label="(256, 256, 3, 3)", style=solid];
 "60 quantize_per_channel_default_5" -> "61 dequantize_per_channel_default_5"  [label="(256, 256, 3, 3)", style=solid];
 "61 dequantize_per_channel_default_5" -> "63 conv2d_5"  [label="(256, 256, 3, 3)", style=solid];
-"62 _param_constant21_0_0" -> "63 conv2d_5"  [label="(256,)", style=solid];
+"62 down_path_2_block_3_bias_0_0" -> "63 conv2d_5"  [label="(256,)", style=solid];
 "63 conv2d_5" -> "64 relu_5"  [label="(1, 256, 49, 49)", style=solid];
 "64 relu_5" -> "65 quantize_per_tensor_default_4"  [label="(1, 256, 49, 49)", style=solid];
 "65 quantize_per_tensor_default_4" -> "66 dequantize_per_tensor_default_7"  [label="(1, 256, 49, 49)", style=solid];
@@ -345,7 +345,7 @@ strict digraph  {
 "71 compressed_weight_updated_constant6" -> "72 quantize_per_channel_default_6"  [label="(512, 256, 3, 3)", style=solid];
 "72 quantize_per_channel_default_6" -> "73 dequantize_per_channel_default_6"  [label="(512, 256, 3, 3)", style=solid];
 "73 dequantize_per_channel_default_6" -> "75 conv2d_6"  [label="(512, 256, 3, 3)", style=solid];
-"74 _param_constant25_0_0" -> "75 conv2d_6"  [label="(512,)", style=solid];
+"74 down_path_3_block_0_bias_0_0" -> "75 conv2d_6"  [label="(512,)", style=solid];
 "75 conv2d_6" -> "76 relu_6"  [label="(1, 512, 22, 22)", style=solid];
 "76 relu_6" -> "77 quantize_per_tensor_default_12"  [label="(1, 512, 22, 22)", style=solid];
 "77 quantize_per_tensor_default_12" -> "78 dequantize_per_tensor_default_16"  [label="(1, 512, 22, 22)", style=solid];
@@ -357,7 +357,7 @@ strict digraph  {
 "81 compressed_weight_updated_constant7" -> "82 quantize_per_channel_default_7"  [label="(512, 512, 3, 3)", style=solid];
 "82 quantize_per_channel_default_7" -> "83 dequantize_per_channel_default_7"  [label="(512, 512, 3, 3)", style=solid];
 "83 dequantize_per_channel_default_7" -> "85 conv2d_7"  [label="(512, 512, 3, 3)", style=solid];
-"84 _param_constant29_0_0" -> "85 conv2d_7"  [label="(512,)", style=solid];
+"84 down_path_3_block_3_bias_0_0" -> "85 conv2d_7"  [label="(512,)", style=solid];
 "85 conv2d_7" -> "86 relu_7"  [label="(1, 512, 20, 20)", style=solid];
 "86 relu_7" -> "87 quantize_per_tensor_default_7"  [label="(1, 512, 20, 20)", style=solid];
 "87 quantize_per_tensor_default_7" -> "88 dequantize_per_tensor_default_11"  [label="(1, 512, 20, 20)", style=solid];
@@ -372,7 +372,7 @@ strict digraph  {
 "93 compressed_weight_updated_constant8" -> "94 quantize_per_channel_default_8"  [label="(1024, 512, 3, 3)", style=solid];
 "94 quantize_per_channel_default_8" -> "95 dequantize_per_channel_default_8"  [label="(1024, 512, 3, 3)", style=solid];
 "95 dequantize_per_channel_default_8" -> "97 conv2d_8"  [label="(1024, 512, 3, 3)", style=solid];
-"96 _param_constant33_0_0" -> "97 conv2d_8"  [label="(1024,)", style=solid];
+"96 down_path_4_block_0_bias_0_0" -> "97 conv2d_8"  [label="(1024,)", style=solid];
 "97 conv2d_8" -> "98 relu_8"  [label="(1, 1024, 8, 8)", style=solid];
 "98 relu_8" -> "99 quantize_per_tensor_default_13"  [label="(1, 1024, 8, 8)", style=solid];
 "99 quantize_per_tensor_default_13" -> "100 dequantize_per_tensor_default_17"  [label="(1, 1024, 8, 8)", style=solid];
@@ -384,7 +384,7 @@ strict digraph  {
 "103 compressed_weight_updated_constant9" -> "104 quantize_per_channel_default_9"  [label="(1024, 1024, 3, 3)", style=solid];
 "104 quantize_per_channel_default_9" -> "105 dequantize_per_channel_default_9"  [label="(1024, 1024, 3, 3)", style=solid];
 "105 dequantize_per_channel_default_9" -> "107 conv2d_9"  [label="(1024, 1024, 3, 3)", style=solid];
-"106 _param_constant37_0_0" -> "107 conv2d_9"  [label="(1024,)", style=solid];
+"106 down_path_4_block_3_bias_0_0" -> "107 conv2d_9"  [label="(1024,)", style=solid];
 "107 conv2d_9" -> "108 relu_9"  [label="(1, 1024, 6, 6)", style=solid];
 "108 relu_9" -> "109 quantize_per_tensor_default_14"  [label="(1, 1024, 6, 6)", style=solid];
 "109 quantize_per_tensor_default_14" -> "110 dequantize_per_tensor_default_18"  [label="(1, 1024, 6, 6)", style=solid];
@@ -396,7 +396,7 @@ strict digraph  {
 "113 compressed_weight_updated_constant10" -> "114 quantize_per_channel_default_10"  [label="(1024, 512, 2, 2)", style=solid];
 "114 quantize_per_channel_default_10" -> "115 dequantize_per_channel_default_10"  [label="(1024, 512, 2, 2)", style=solid];
 "115 dequantize_per_channel_default_10" -> "117 conv_transpose2d"  [label="(1024, 512, 2, 2)", style=solid];
-"116 _param_constant41_0_0" -> "117 conv_transpose2d"  [label="(512,)", style=solid];
+"116 up_path_0_up_bias_0_0" -> "117 conv_transpose2d"  [label="(512,)", style=solid];
 "117 conv_transpose2d" -> "118 quantize_per_tensor_default_6"  [label="(1, 512, 12, 12)", style=solid];
 "118 quantize_per_tensor_default_6" -> "119 dequantize_per_tensor_default_9"  [label="(1, 512, 12, 12)", style=solid];
 "119 dequantize_per_tensor_default_9" -> "124 cat"  [label="(1, 512, 12, 12)", style=solid];
@@ -412,7 +412,7 @@ strict digraph  {
 "127 compressed_weight_updated_constant11" -> "128 quantize_per_channel_default_11"  [label="(512, 1024, 3, 3)", style=solid];
 "128 quantize_per_channel_default_11" -> "129 dequantize_per_channel_default_11"  [label="(512, 1024, 3, 3)", style=solid];
 "129 dequantize_per_channel_default_11" -> "131 conv2d_10"  [label="(512, 1024, 3, 3)", style=solid];
-"130 _param_constant43_0_0" -> "131 conv2d_10"  [label="(512,)", style=solid];
+"130 up_path_0_conv_block_block_0_bias_0_0" -> "131 conv2d_10"  [label="(512,)", style=solid];
 "131 conv2d_10" -> "132 relu_10"  [label="(1, 512, 10, 10)", style=solid];
 "132 relu_10" -> "133 quantize_per_tensor_default_15"  [label="(1, 512, 10, 10)", style=solid];
 "133 quantize_per_tensor_default_15" -> "134 dequantize_per_tensor_default_19"  [label="(1, 512, 10, 10)", style=solid];
@@ -424,7 +424,7 @@ strict digraph  {
 "137 compressed_weight_updated_constant12" -> "138 quantize_per_channel_default_12"  [label="(512, 512, 3, 3)", style=solid];
 "138 quantize_per_channel_default_12" -> "139 dequantize_per_channel_default_12"  [label="(512, 512, 3, 3)", style=solid];
 "139 dequantize_per_channel_default_12" -> "141 conv2d_11"  [label="(512, 512, 3, 3)", style=solid];
-"140 _param_constant47_0_0" -> "141 conv2d_11"  [label="(512,)", style=solid];
+"140 up_path_0_conv_block_block_3_bias_0_0" -> "141 conv2d_11"  [label="(512,)", style=solid];
 "141 conv2d_11" -> "142 relu_11"  [label="(1, 512, 8, 8)", style=solid];
 "142 relu_11" -> "143 quantize_per_tensor_default_16"  [label="(1, 512, 8, 8)", style=solid];
 "143 quantize_per_tensor_default_16" -> "144 dequantize_per_tensor_default_20"  [label="(1, 512, 8, 8)", style=solid];
@@ -436,7 +436,7 @@ strict digraph  {
 "147 compressed_weight_updated_constant13" -> "148 quantize_per_channel_default_13"  [label="(512, 256, 2, 2)", style=solid];
 "148 quantize_per_channel_default_13" -> "149 dequantize_per_channel_default_13"  [label="(512, 256, 2, 2)", style=solid];
 "149 dequantize_per_channel_default_13" -> "151 conv_transpose2d_1"  [label="(512, 256, 2, 2)", style=solid];
-"150 _param_constant51_0_0" -> "151 conv_transpose2d_1"  [label="(256,)", style=solid];
+"150 up_path_1_up_bias_0_0" -> "151 conv_transpose2d_1"  [label="(256,)", style=solid];
 "151 conv_transpose2d_1" -> "152 quantize_per_tensor_default_5"  [label="(1, 256, 16, 16)", style=solid];
 "152 quantize_per_tensor_default_5" -> "153 dequantize_per_tensor_default_8"  [label="(1, 256, 16, 16)", style=solid];
 "153 dequantize_per_tensor_default_8" -> "158 cat_1"  [label="(1, 256, 16, 16)", style=solid];
@@ -452,7 +452,7 @@ strict digraph  {
 "161 compressed_weight_updated_constant14" -> "162 quantize_per_channel_default_14"  [label="(256, 512, 3, 3)", style=solid];
 "162 quantize_per_channel_default_14" -> "163 dequantize_per_channel_default_14"  [label="(256, 512, 3, 3)", style=solid];
 "163 dequantize_per_channel_default_14" -> "165 conv2d_12"  [label="(256, 512, 3, 3)", style=solid];
-"164 _param_constant53_0_0" -> "165 conv2d_12"  [label="(256,)", style=solid];
+"164 up_path_1_conv_block_block_0_bias_0_0" -> "165 conv2d_12"  [label="(256,)", style=solid];
 "165 conv2d_12" -> "166 relu_12"  [label="(1, 256, 14, 14)", style=solid];
 "166 relu_12" -> "167 quantize_per_tensor_default_17"  [label="(1, 256, 14, 14)", style=solid];
 "167 quantize_per_tensor_default_17" -> "168 dequantize_per_tensor_default_21"  [label="(1, 256, 14, 14)", style=solid];
@@ -464,7 +464,7 @@ strict digraph  {
 "171 compressed_weight_updated_constant15" -> "172 quantize_per_channel_default_15"  [label="(256, 256, 3, 3)", style=solid];
 "172 quantize_per_channel_default_15" -> "173 dequantize_per_channel_default_15"  [label="(256, 256, 3, 3)", style=solid];
 "173 dequantize_per_channel_default_15" -> "175 conv2d_13"  [label="(256, 256, 3, 3)", style=solid];
-"174 _param_constant57_0_0" -> "175 conv2d_13"  [label="(256,)", style=solid];
+"174 up_path_1_conv_block_block_3_bias_0_0" -> "175 conv2d_13"  [label="(256,)", style=solid];
 "175 conv2d_13" -> "176 relu_13"  [label="(1, 256, 12, 12)", style=solid];
 "176 relu_13" -> "177 quantize_per_tensor_default_18"  [label="(1, 256, 12, 12)", style=solid];
 "177 quantize_per_tensor_default_18" -> "178 dequantize_per_tensor_default_22"  [label="(1, 256, 12, 12)", style=solid];
@@ -476,7 +476,7 @@ strict digraph  {
 "181 compressed_weight_updated_constant16" -> "182 quantize_per_channel_default_16"  [label="(256, 128, 2, 2)", style=solid];
 "182 quantize_per_channel_default_16" -> "183 dequantize_per_channel_default_16"  [label="(256, 128, 2, 2)", style=solid];
 "183 dequantize_per_channel_default_16" -> "185 conv_transpose2d_2"  [label="(256, 128, 2, 2)", style=solid];
-"184 _param_constant61_0_0" -> "185 conv_transpose2d_2"  [label="(128,)", style=solid];
+"184 up_path_2_up_bias_0_0" -> "185 conv_transpose2d_2"  [label="(128,)", style=solid];
 "185 conv_transpose2d_2" -> "186 quantize_per_tensor_default_2"  [label="(1, 128, 24, 24)", style=solid];
 "186 quantize_per_tensor_default_2" -> "187 dequantize_per_tensor_default_3"  [label="(1, 128, 24, 24)", style=solid];
 "187 dequantize_per_tensor_default_3" -> "192 cat_2"  [label="(1, 128, 24, 24)", style=solid];
@@ -492,7 +492,7 @@ strict digraph  {
 "195 compressed_weight_updated_constant17" -> "196 quantize_per_channel_default_17"  [label="(128, 256, 3, 3)", style=solid];
 "196 quantize_per_channel_default_17" -> "197 dequantize_per_channel_default_17"  [label="(128, 256, 3, 3)", style=solid];
 "197 dequantize_per_channel_default_17" -> "199 conv2d_14"  [label="(128, 256, 3, 3)", style=solid];
-"198 _param_constant63_0_0" -> "199 conv2d_14"  [label="(128,)", style=solid];
+"198 up_path_2_conv_block_block_0_bias_0_0" -> "199 conv2d_14"  [label="(128,)", style=solid];
 "199 conv2d_14" -> "200 relu_14"  [label="(1, 128, 22, 22)", style=solid];
 "200 relu_14" -> "201 quantize_per_tensor_default_19"  [label="(1, 128, 22, 22)", style=solid];
 "201 quantize_per_tensor_default_19" -> "202 dequantize_per_tensor_default_23"  [label="(1, 128, 22, 22)", style=solid];
@@ -504,7 +504,7 @@ strict digraph  {
 "205 compressed_weight_updated_constant18" -> "206 quantize_per_channel_default_18"  [label="(128, 128, 3, 3)", style=solid];
 "206 quantize_per_channel_default_18" -> "207 dequantize_per_channel_default_18"  [label="(128, 128, 3, 3)", style=solid];
 "207 dequantize_per_channel_default_18" -> "209 conv2d_15"  [label="(128, 128, 3, 3)", style=solid];
-"208 _param_constant67_0_0" -> "209 conv2d_15"  [label="(128,)", style=solid];
+"208 up_path_2_conv_block_block_3_bias_0_0" -> "209 conv2d_15"  [label="(128,)", style=solid];
 "209 conv2d_15" -> "210 relu_15"  [label="(1, 128, 20, 20)", style=solid];
 "210 relu_15" -> "211 quantize_per_tensor_default_20"  [label="(1, 128, 20, 20)", style=solid];
 "211 quantize_per_tensor_default_20" -> "212 dequantize_per_tensor_default_24"  [label="(1, 128, 20, 20)", style=solid];
@@ -516,7 +516,7 @@ strict digraph  {
 "215 compressed_weight_updated_constant19" -> "216 quantize_per_channel_default_19"  [label="(128, 64, 2, 2)", style=solid];
 "216 quantize_per_channel_default_19" -> "217 dequantize_per_channel_default_19"  [label="(128, 64, 2, 2)", style=solid];
 "217 dequantize_per_channel_default_19" -> "219 conv_transpose2d_3"  [label="(128, 64, 2, 2)", style=solid];
-"218 _param_constant71_0_0" -> "219 conv_transpose2d_3"  [label="(64,)", style=solid];
+"218 up_path_3_up_bias_0_0" -> "219 conv_transpose2d_3"  [label="(64,)", style=solid];
 "219 conv_transpose2d_3" -> "220 quantize_per_tensor_default_1"  [label="(1, 64, 40, 40)", style=solid];
 "220 quantize_per_tensor_default_1" -> "221 dequantize_per_tensor_default_2"  [label="(1, 64, 40, 40)", style=solid];
 "221 dequantize_per_tensor_default_2" -> "226 cat_3"  [label="(1, 64, 40, 40)", style=solid];
@@ -532,7 +532,7 @@ strict digraph  {
 "229 compressed_weight_updated_constant20" -> "230 quantize_per_channel_default_20"  [label="(64, 128, 3, 3)", style=solid];
 "230 quantize_per_channel_default_20" -> "231 dequantize_per_channel_default_20"  [label="(64, 128, 3, 3)", style=solid];
 "231 dequantize_per_channel_default_20" -> "233 conv2d_16"  [label="(64, 128, 3, 3)", style=solid];
-"232 _param_constant73_0_0" -> "233 conv2d_16"  [label="(64,)", style=solid];
+"232 up_path_3_conv_block_block_0_bias_0_0" -> "233 conv2d_16"  [label="(64,)", style=solid];
 "233 conv2d_16" -> "234 relu_16"  [label="(1, 64, 38, 38)", style=solid];
 "234 relu_16" -> "235 quantize_per_tensor_default_21"  [label="(1, 64, 38, 38)", style=solid];
 "235 quantize_per_tensor_default_21" -> "236 dequantize_per_tensor_default_25"  [label="(1, 64, 38, 38)", style=solid];
@@ -544,7 +544,7 @@ strict digraph  {
 "239 compressed_weight_updated_constant21" -> "240 quantize_per_channel_default_21"  [label="(64, 64, 3, 3)", style=solid];
 "240 quantize_per_channel_default_21" -> "241 dequantize_per_channel_default_21"  [label="(64, 64, 3, 3)", style=solid];
 "241 dequantize_per_channel_default_21" -> "243 conv2d_17"  [label="(64, 64, 3, 3)", style=solid];
-"242 _param_constant77_0_0" -> "243 conv2d_17"  [label="(64,)", style=solid];
+"242 up_path_3_conv_block_block_3_bias_0_0" -> "243 conv2d_17"  [label="(64,)", style=solid];
 "243 conv2d_17" -> "244 relu_17"  [label="(1, 64, 36, 36)", style=solid];
 "244 relu_17" -> "245 quantize_per_tensor_default_22"  [label="(1, 64, 36, 36)", style=solid];
 "245 quantize_per_tensor_default_22" -> "246 dequantize_per_tensor_default_26"  [label="(1, 64, 36, 36)", style=solid];
@@ -556,6 +556,6 @@ strict digraph  {
 "249 compressed_weight_updated_constant22" -> "250 quantize_per_channel_default_22"  [label="(12, 64, 1, 1)", style=solid];
 "250 quantize_per_channel_default_22" -> "251 dequantize_per_channel_default_22"  [label="(12, 64, 1, 1)", style=solid];
 "251 dequantize_per_channel_default_22" -> "253 conv2d_18"  [label="(12, 64, 1, 1)", style=solid];
-"252 _param_constant81_0_0" -> "253 conv2d_18"  [label="(12,)", style=solid];
+"252 last_bias_0_0" -> "253 conv2d_18"  [label="(12,)", style=solid];
 "253 conv2d_18" -> "254 output"  [label="(1, 12, 36, 36)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/quantized/vit_b_16.dot b/tests/torch/data/reference_graphs/fx/quantized/vit_b_16.dot
index 501d066ef82..51f70602ba7 100644
--- a/tests/torch/data/reference_graphs/fx/quantized/vit_b_16.dot
+++ b/tests/torch/data/reference_graphs/fx/quantized/vit_b_16.dot
@@ -1,971 +1,971 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 quantize_per_tensor_default" [id=1, type=quantize_per_tensor];
-"2 dequantize_per_tensor_default" [id=2, type=dequantize_per_tensor];
-"3 conv2d_scale_0" [id=3, type=get_attr];
-"4 conv2d_zero_point_0" [id=4, type=get_attr];
-"5 compressed_weight_updated_constant0" [id=5, type=get_attr];
-"6 quantize_per_channel_default" [id=6, type=quantize_per_channel];
-"7 dequantize_per_channel_default" [id=7, type=dequantize_per_channel];
-"8 _param_constant1_0_0" [id=8, type=get_attr];
-"9 conv2d" [id=9, type=conv2d];
-"10 reshape" [id=10, type=reshape];
-"11 permute" [id=11, type=permute];
-"12 _frozen_param0" [id=12, type=get_attr];
-"13 cat" [id=13, type=cat];
-"14 _param_constant3" [id=14, type=get_attr];
-"15 add" [id=15, type=add];
-"16 dropout" [id=16, type=dropout];
-"17 _param_constant4" [id=17, type=get_attr];
-"18 _param_constant5" [id=18, type=get_attr];
-"19 layer_norm" [id=19, type=layer_norm];
-"20 transpose" [id=20, type=transpose];
-"21 transpose_0_0_nncf_smooth_quant_0" [id=21, type=call_module];
-"22 quantize_per_tensor_default_1" [id=22, type=quantize_per_tensor];
-"23 dequantize_per_tensor_default_1" [id=23, type=dequantize_per_tensor];
-"24 linear_scale_0" [id=24, type=get_attr];
-"25 linear_zero_point_0" [id=25, type=get_attr];
-"26 compressed_weight_updated_constant1" [id=26, type=get_attr];
-"27 quantize_per_channel_default_1" [id=27, type=quantize_per_channel];
-"28 dequantize_per_channel_default_1" [id=28, type=dequantize_per_channel];
-"29 _param_constant7_0_0" [id=29, type=get_attr];
-"30 linear" [id=30, type=linear];
-"31 unflatten" [id=31, type=unflatten];
-"32 unsqueeze" [id=32, type=unsqueeze];
-"33 transpose_1" [id=33, type=transpose];
-"34 squeeze" [id=34, type=squeeze];
-"35 contiguous" [id=35, type=contiguous];
-"36 quantize_per_tensor_default_2" [id=36, type=quantize_per_tensor];
-"37 dequantize_per_tensor_default_2" [id=37, type=dequantize_per_tensor];
-"38 select" [id=38, type=select];
-"39 quantize_per_tensor_default_3" [id=39, type=quantize_per_tensor];
-"40 dequantize_per_tensor_default_3" [id=40, type=dequantize_per_tensor];
-"41 select_1" [id=41, type=select];
-"42 select_2" [id=42, type=select];
-"43 view" [id=43, type=view];
-"44 transpose_2" [id=44, type=transpose];
-"45 view_1" [id=45, type=view];
-"46 transpose_3" [id=46, type=transpose];
-"47 view_2" [id=47, type=view];
-"48 transpose_4" [id=48, type=transpose];
-"49 view_3" [id=49, type=view];
-"50 view_4" [id=50, type=view];
-"51 view_5" [id=51, type=view];
-"52 scaled_dot_product_attention" [id=52, type=scaled_dot_product_attention];
-"53 permute_1" [id=53, type=permute];
-"54 view_6" [id=54, type=view];
-"55 view_6_0_0_nncf_smooth_quant_0" [id=55, type=call_module];
-"56 quantize_per_tensor_default_4" [id=56, type=quantize_per_tensor];
-"57 dequantize_per_tensor_default_4" [id=57, type=dequantize_per_tensor];
-"58 linear_1_scale_0" [id=58, type=get_attr];
-"59 linear_1_zero_point_0" [id=59, type=get_attr];
-"60 compressed_weight_updated_constant2" [id=60, type=get_attr];
-"61 quantize_per_channel_default_2" [id=61, type=quantize_per_channel];
-"62 dequantize_per_channel_default_2" [id=62, type=dequantize_per_channel];
-"63 _param_constant9_0_0" [id=63, type=get_attr];
-"64 linear_1" [id=64, type=linear];
-"65 view_7" [id=65, type=view];
-"66 transpose_5" [id=66, type=transpose];
-"67 dropout_1" [id=67, type=dropout];
-"68 add_1" [id=68, type=add];
-"69 _param_constant10" [id=69, type=get_attr];
-"70 _param_constant11" [id=70, type=get_attr];
-"71 layer_norm_1" [id=71, type=layer_norm];
-"72 layer_norm_1_0_0_nncf_smooth_quant_0" [id=72, type=call_module];
-"73 quantize_per_tensor_default_5" [id=73, type=quantize_per_tensor];
-"74 dequantize_per_tensor_default_5" [id=74, type=dequantize_per_tensor];
-"75 linear_2_scale_0" [id=75, type=get_attr];
-"76 linear_2_zero_point_0" [id=76, type=get_attr];
-"77 compressed_weight_updated_constant3" [id=77, type=get_attr];
-"78 quantize_per_channel_default_3" [id=78, type=quantize_per_channel];
-"79 dequantize_per_channel_default_3" [id=79, type=dequantize_per_channel];
-"80 _param_constant13_0_0" [id=80, type=get_attr];
-"81 linear_2" [id=81, type=linear];
-"82 gelu" [id=82, type=gelu];
-"83 dropout_2" [id=83, type=dropout];
-"84 dropout_2_0_0_nncf_smooth_quant_0" [id=84, type=call_module];
-"85 quantize_per_tensor_default_6" [id=85, type=quantize_per_tensor];
-"86 dequantize_per_tensor_default_6" [id=86, type=dequantize_per_tensor];
-"87 linear_3_scale_0" [id=87, type=get_attr];
-"88 linear_3_zero_point_0" [id=88, type=get_attr];
-"89 compressed_weight_updated_constant4" [id=89, type=get_attr];
-"90 quantize_per_channel_default_4" [id=90, type=quantize_per_channel];
-"91 dequantize_per_channel_default_4" [id=91, type=dequantize_per_channel];
-"92 _param_constant15_0_0" [id=92, type=get_attr];
-"93 linear_3" [id=93, type=linear];
-"94 dropout_3" [id=94, type=dropout];
-"95 add_2" [id=95, type=add];
-"96 _param_constant16" [id=96, type=get_attr];
-"97 _param_constant17" [id=97, type=get_attr];
-"98 layer_norm_2" [id=98, type=layer_norm];
-"99 transpose_6" [id=99, type=transpose];
-"100 transpose_6_0_0_nncf_smooth_quant_0" [id=100, type=call_module];
-"101 quantize_per_tensor_default_7" [id=101, type=quantize_per_tensor];
-"102 dequantize_per_tensor_default_7" [id=102, type=dequantize_per_tensor];
-"103 linear_4_scale_0" [id=103, type=get_attr];
-"104 linear_4_zero_point_0" [id=104, type=get_attr];
-"105 compressed_weight_updated_constant5" [id=105, type=get_attr];
-"106 quantize_per_channel_default_5" [id=106, type=quantize_per_channel];
-"107 dequantize_per_channel_default_5" [id=107, type=dequantize_per_channel];
-"108 _param_constant19_0_0" [id=108, type=get_attr];
-"109 linear_4" [id=109, type=linear];
-"110 unflatten_1" [id=110, type=unflatten];
-"111 unsqueeze_1" [id=111, type=unsqueeze];
-"112 transpose_7" [id=112, type=transpose];
-"113 squeeze_1" [id=113, type=squeeze];
-"114 contiguous_1" [id=114, type=contiguous];
-"115 quantize_per_tensor_default_8" [id=115, type=quantize_per_tensor];
-"116 dequantize_per_tensor_default_8" [id=116, type=dequantize_per_tensor];
-"117 select_3" [id=117, type=select];
-"118 quantize_per_tensor_default_9" [id=118, type=quantize_per_tensor];
-"119 dequantize_per_tensor_default_9" [id=119, type=dequantize_per_tensor];
-"120 select_4" [id=120, type=select];
-"121 select_5" [id=121, type=select];
-"122 view_8" [id=122, type=view];
-"123 transpose_8" [id=123, type=transpose];
-"124 view_9" [id=124, type=view];
-"125 transpose_9" [id=125, type=transpose];
-"126 view_10" [id=126, type=view];
-"127 transpose_10" [id=127, type=transpose];
-"128 view_11" [id=128, type=view];
-"129 view_12" [id=129, type=view];
-"130 view_13" [id=130, type=view];
-"131 scaled_dot_product_attention_1" [id=131, type=scaled_dot_product_attention];
-"132 permute_2" [id=132, type=permute];
-"133 view_14" [id=133, type=view];
-"134 view_14_0_0_nncf_smooth_quant_0" [id=134, type=call_module];
-"135 quantize_per_tensor_default_10" [id=135, type=quantize_per_tensor];
-"136 dequantize_per_tensor_default_10" [id=136, type=dequantize_per_tensor];
-"137 linear_5_scale_0" [id=137, type=get_attr];
-"138 linear_5_zero_point_0" [id=138, type=get_attr];
-"139 compressed_weight_updated_constant6" [id=139, type=get_attr];
-"140 quantize_per_channel_default_6" [id=140, type=quantize_per_channel];
-"141 dequantize_per_channel_default_6" [id=141, type=dequantize_per_channel];
-"142 _param_constant21_0_0" [id=142, type=get_attr];
-"143 linear_5" [id=143, type=linear];
-"144 view_15" [id=144, type=view];
-"145 transpose_11" [id=145, type=transpose];
-"146 dropout_4" [id=146, type=dropout];
-"147 add_3" [id=147, type=add];
-"148 _param_constant22" [id=148, type=get_attr];
-"149 _param_constant23" [id=149, type=get_attr];
-"150 layer_norm_3" [id=150, type=layer_norm];
-"151 layer_norm_3_0_0_nncf_smooth_quant_0" [id=151, type=call_module];
-"152 quantize_per_tensor_default_11" [id=152, type=quantize_per_tensor];
-"153 dequantize_per_tensor_default_11" [id=153, type=dequantize_per_tensor];
-"154 linear_6_scale_0" [id=154, type=get_attr];
-"155 linear_6_zero_point_0" [id=155, type=get_attr];
-"156 compressed_weight_updated_constant7" [id=156, type=get_attr];
-"157 quantize_per_channel_default_7" [id=157, type=quantize_per_channel];
-"158 dequantize_per_channel_default_7" [id=158, type=dequantize_per_channel];
-"159 _param_constant25_0_0" [id=159, type=get_attr];
-"160 linear_6" [id=160, type=linear];
-"161 gelu_1" [id=161, type=gelu];
-"162 dropout_5" [id=162, type=dropout];
-"163 dropout_5_0_0_nncf_smooth_quant_0" [id=163, type=call_module];
-"164 quantize_per_tensor_default_12" [id=164, type=quantize_per_tensor];
-"165 dequantize_per_tensor_default_12" [id=165, type=dequantize_per_tensor];
-"166 linear_7_scale_0" [id=166, type=get_attr];
-"167 linear_7_zero_point_0" [id=167, type=get_attr];
-"168 compressed_weight_updated_constant8" [id=168, type=get_attr];
-"169 quantize_per_channel_default_8" [id=169, type=quantize_per_channel];
-"170 dequantize_per_channel_default_8" [id=170, type=dequantize_per_channel];
-"171 _param_constant27_0_0" [id=171, type=get_attr];
-"172 linear_7" [id=172, type=linear];
-"173 dropout_6" [id=173, type=dropout];
-"174 add_4" [id=174, type=add];
-"175 _param_constant28" [id=175, type=get_attr];
-"176 _param_constant29" [id=176, type=get_attr];
-"177 layer_norm_4" [id=177, type=layer_norm];
-"178 transpose_12" [id=178, type=transpose];
-"179 transpose_12_0_0_nncf_smooth_quant_0" [id=179, type=call_module];
-"180 quantize_per_tensor_default_13" [id=180, type=quantize_per_tensor];
-"181 dequantize_per_tensor_default_13" [id=181, type=dequantize_per_tensor];
-"182 linear_8_scale_0" [id=182, type=get_attr];
-"183 linear_8_zero_point_0" [id=183, type=get_attr];
-"184 compressed_weight_updated_constant9" [id=184, type=get_attr];
-"185 quantize_per_channel_default_9" [id=185, type=quantize_per_channel];
-"186 dequantize_per_channel_default_9" [id=186, type=dequantize_per_channel];
-"187 _param_constant31_0_0" [id=187, type=get_attr];
-"188 linear_8" [id=188, type=linear];
-"189 unflatten_2" [id=189, type=unflatten];
-"190 unsqueeze_2" [id=190, type=unsqueeze];
-"191 transpose_13" [id=191, type=transpose];
-"192 squeeze_2" [id=192, type=squeeze];
-"193 contiguous_2" [id=193, type=contiguous];
-"194 quantize_per_tensor_default_14" [id=194, type=quantize_per_tensor];
-"195 dequantize_per_tensor_default_14" [id=195, type=dequantize_per_tensor];
-"196 select_6" [id=196, type=select];
-"197 quantize_per_tensor_default_15" [id=197, type=quantize_per_tensor];
-"198 dequantize_per_tensor_default_15" [id=198, type=dequantize_per_tensor];
-"199 select_7" [id=199, type=select];
-"200 select_8" [id=200, type=select];
-"201 view_16" [id=201, type=view];
-"202 transpose_14" [id=202, type=transpose];
-"203 view_17" [id=203, type=view];
-"204 transpose_15" [id=204, type=transpose];
-"205 view_18" [id=205, type=view];
-"206 transpose_16" [id=206, type=transpose];
-"207 view_19" [id=207, type=view];
-"208 view_20" [id=208, type=view];
-"209 view_21" [id=209, type=view];
-"210 scaled_dot_product_attention_2" [id=210, type=scaled_dot_product_attention];
-"211 permute_3" [id=211, type=permute];
-"212 view_22" [id=212, type=view];
-"213 view_22_0_0_nncf_smooth_quant_0" [id=213, type=call_module];
-"214 quantize_per_tensor_default_16" [id=214, type=quantize_per_tensor];
-"215 dequantize_per_tensor_default_16" [id=215, type=dequantize_per_tensor];
-"216 linear_9_scale_0" [id=216, type=get_attr];
-"217 linear_9_zero_point_0" [id=217, type=get_attr];
-"218 compressed_weight_updated_constant10" [id=218, type=get_attr];
-"219 quantize_per_channel_default_10" [id=219, type=quantize_per_channel];
-"220 dequantize_per_channel_default_10" [id=220, type=dequantize_per_channel];
-"221 _param_constant33_0_0" [id=221, type=get_attr];
-"222 linear_9" [id=222, type=linear];
-"223 view_23" [id=223, type=view];
-"224 transpose_17" [id=224, type=transpose];
-"225 dropout_7" [id=225, type=dropout];
-"226 add_5" [id=226, type=add];
-"227 _param_constant34" [id=227, type=get_attr];
-"228 _param_constant35" [id=228, type=get_attr];
-"229 layer_norm_5" [id=229, type=layer_norm];
-"230 layer_norm_5_0_0_nncf_smooth_quant_0" [id=230, type=call_module];
-"231 quantize_per_tensor_default_17" [id=231, type=quantize_per_tensor];
-"232 dequantize_per_tensor_default_17" [id=232, type=dequantize_per_tensor];
-"233 linear_10_scale_0" [id=233, type=get_attr];
-"234 linear_10_zero_point_0" [id=234, type=get_attr];
-"235 compressed_weight_updated_constant11" [id=235, type=get_attr];
-"236 quantize_per_channel_default_11" [id=236, type=quantize_per_channel];
-"237 dequantize_per_channel_default_11" [id=237, type=dequantize_per_channel];
-"238 _param_constant37_0_0" [id=238, type=get_attr];
-"239 linear_10" [id=239, type=linear];
-"240 gelu_2" [id=240, type=gelu];
-"241 dropout_8" [id=241, type=dropout];
-"242 dropout_8_0_0_nncf_smooth_quant_0" [id=242, type=call_module];
-"243 quantize_per_tensor_default_18" [id=243, type=quantize_per_tensor];
-"244 dequantize_per_tensor_default_18" [id=244, type=dequantize_per_tensor];
-"245 linear_11_scale_0" [id=245, type=get_attr];
-"246 linear_11_zero_point_0" [id=246, type=get_attr];
-"247 compressed_weight_updated_constant12" [id=247, type=get_attr];
-"248 quantize_per_channel_default_12" [id=248, type=quantize_per_channel];
-"249 dequantize_per_channel_default_12" [id=249, type=dequantize_per_channel];
-"250 _param_constant39_0_0" [id=250, type=get_attr];
-"251 linear_11" [id=251, type=linear];
-"252 dropout_9" [id=252, type=dropout];
-"253 add_6" [id=253, type=add];
-"254 _param_constant40" [id=254, type=get_attr];
-"255 _param_constant41" [id=255, type=get_attr];
-"256 layer_norm_6" [id=256, type=layer_norm];
-"257 transpose_18" [id=257, type=transpose];
-"258 transpose_18_0_0_nncf_smooth_quant_0" [id=258, type=call_module];
-"259 quantize_per_tensor_default_19" [id=259, type=quantize_per_tensor];
-"260 dequantize_per_tensor_default_19" [id=260, type=dequantize_per_tensor];
-"261 linear_12_scale_0" [id=261, type=get_attr];
-"262 linear_12_zero_point_0" [id=262, type=get_attr];
-"263 compressed_weight_updated_constant13" [id=263, type=get_attr];
-"264 quantize_per_channel_default_13" [id=264, type=quantize_per_channel];
-"265 dequantize_per_channel_default_13" [id=265, type=dequantize_per_channel];
-"266 _param_constant43_0_0" [id=266, type=get_attr];
-"267 linear_12" [id=267, type=linear];
-"268 unflatten_3" [id=268, type=unflatten];
-"269 unsqueeze_3" [id=269, type=unsqueeze];
-"270 transpose_19" [id=270, type=transpose];
-"271 squeeze_3" [id=271, type=squeeze];
-"272 contiguous_3" [id=272, type=contiguous];
-"273 quantize_per_tensor_default_20" [id=273, type=quantize_per_tensor];
-"274 dequantize_per_tensor_default_20" [id=274, type=dequantize_per_tensor];
-"275 select_9" [id=275, type=select];
-"276 quantize_per_tensor_default_21" [id=276, type=quantize_per_tensor];
-"277 dequantize_per_tensor_default_21" [id=277, type=dequantize_per_tensor];
-"278 select_10" [id=278, type=select];
-"279 select_11" [id=279, type=select];
-"280 view_24" [id=280, type=view];
-"281 transpose_20" [id=281, type=transpose];
-"282 view_25" [id=282, type=view];
-"283 transpose_21" [id=283, type=transpose];
-"284 view_26" [id=284, type=view];
-"285 transpose_22" [id=285, type=transpose];
-"286 view_27" [id=286, type=view];
-"287 view_28" [id=287, type=view];
-"288 view_29" [id=288, type=view];
-"289 scaled_dot_product_attention_3" [id=289, type=scaled_dot_product_attention];
-"290 permute_4" [id=290, type=permute];
-"291 view_30" [id=291, type=view];
-"292 view_30_0_0_nncf_smooth_quant_0" [id=292, type=call_module];
-"293 quantize_per_tensor_default_22" [id=293, type=quantize_per_tensor];
-"294 dequantize_per_tensor_default_22" [id=294, type=dequantize_per_tensor];
-"295 linear_13_scale_0" [id=295, type=get_attr];
-"296 linear_13_zero_point_0" [id=296, type=get_attr];
-"297 compressed_weight_updated_constant14" [id=297, type=get_attr];
-"298 quantize_per_channel_default_14" [id=298, type=quantize_per_channel];
-"299 dequantize_per_channel_default_14" [id=299, type=dequantize_per_channel];
-"300 _param_constant45_0_0" [id=300, type=get_attr];
-"301 linear_13" [id=301, type=linear];
-"302 view_31" [id=302, type=view];
-"303 transpose_23" [id=303, type=transpose];
-"304 dropout_10" [id=304, type=dropout];
-"305 add_7" [id=305, type=add];
-"306 _param_constant46" [id=306, type=get_attr];
-"307 _param_constant47" [id=307, type=get_attr];
-"308 layer_norm_7" [id=308, type=layer_norm];
-"309 layer_norm_7_0_0_nncf_smooth_quant_0" [id=309, type=call_module];
-"310 quantize_per_tensor_default_23" [id=310, type=quantize_per_tensor];
-"311 dequantize_per_tensor_default_23" [id=311, type=dequantize_per_tensor];
-"312 linear_14_scale_0" [id=312, type=get_attr];
-"313 linear_14_zero_point_0" [id=313, type=get_attr];
-"314 compressed_weight_updated_constant15" [id=314, type=get_attr];
-"315 quantize_per_channel_default_15" [id=315, type=quantize_per_channel];
-"316 dequantize_per_channel_default_15" [id=316, type=dequantize_per_channel];
-"317 _param_constant49_0_0" [id=317, type=get_attr];
-"318 linear_14" [id=318, type=linear];
-"319 gelu_3" [id=319, type=gelu];
-"320 dropout_11" [id=320, type=dropout];
-"321 dropout_11_0_0_nncf_smooth_quant_0" [id=321, type=call_module];
-"322 quantize_per_tensor_default_24" [id=322, type=quantize_per_tensor];
-"323 dequantize_per_tensor_default_24" [id=323, type=dequantize_per_tensor];
-"324 linear_15_scale_0" [id=324, type=get_attr];
-"325 linear_15_zero_point_0" [id=325, type=get_attr];
-"326 compressed_weight_updated_constant16" [id=326, type=get_attr];
-"327 quantize_per_channel_default_16" [id=327, type=quantize_per_channel];
-"328 dequantize_per_channel_default_16" [id=328, type=dequantize_per_channel];
-"329 _param_constant51_0_0" [id=329, type=get_attr];
-"330 linear_15" [id=330, type=linear];
-"331 dropout_12" [id=331, type=dropout];
-"332 add_8" [id=332, type=add];
-"333 _param_constant52" [id=333, type=get_attr];
-"334 _param_constant53" [id=334, type=get_attr];
-"335 layer_norm_8" [id=335, type=layer_norm];
-"336 transpose_24" [id=336, type=transpose];
-"337 transpose_24_0_0_nncf_smooth_quant_0" [id=337, type=call_module];
-"338 quantize_per_tensor_default_25" [id=338, type=quantize_per_tensor];
-"339 dequantize_per_tensor_default_25" [id=339, type=dequantize_per_tensor];
-"340 linear_16_scale_0" [id=340, type=get_attr];
-"341 linear_16_zero_point_0" [id=341, type=get_attr];
-"342 compressed_weight_updated_constant17" [id=342, type=get_attr];
-"343 quantize_per_channel_default_17" [id=343, type=quantize_per_channel];
-"344 dequantize_per_channel_default_17" [id=344, type=dequantize_per_channel];
-"345 _param_constant55_0_0" [id=345, type=get_attr];
-"346 linear_16" [id=346, type=linear];
-"347 unflatten_4" [id=347, type=unflatten];
-"348 unsqueeze_4" [id=348, type=unsqueeze];
-"349 transpose_25" [id=349, type=transpose];
-"350 squeeze_4" [id=350, type=squeeze];
-"351 contiguous_4" [id=351, type=contiguous];
-"352 quantize_per_tensor_default_26" [id=352, type=quantize_per_tensor];
-"353 dequantize_per_tensor_default_26" [id=353, type=dequantize_per_tensor];
-"354 select_12" [id=354, type=select];
-"355 quantize_per_tensor_default_27" [id=355, type=quantize_per_tensor];
-"356 dequantize_per_tensor_default_27" [id=356, type=dequantize_per_tensor];
-"357 select_13" [id=357, type=select];
-"358 select_14" [id=358, type=select];
-"359 view_32" [id=359, type=view];
-"360 transpose_26" [id=360, type=transpose];
-"361 view_33" [id=361, type=view];
-"362 transpose_27" [id=362, type=transpose];
-"363 view_34" [id=363, type=view];
-"364 transpose_28" [id=364, type=transpose];
-"365 view_35" [id=365, type=view];
-"366 view_36" [id=366, type=view];
-"367 view_37" [id=367, type=view];
-"368 scaled_dot_product_attention_4" [id=368, type=scaled_dot_product_attention];
-"369 permute_5" [id=369, type=permute];
-"370 view_38" [id=370, type=view];
-"371 view_38_0_0_nncf_smooth_quant_0" [id=371, type=call_module];
-"372 quantize_per_tensor_default_28" [id=372, type=quantize_per_tensor];
-"373 dequantize_per_tensor_default_28" [id=373, type=dequantize_per_tensor];
-"374 linear_17_scale_0" [id=374, type=get_attr];
-"375 linear_17_zero_point_0" [id=375, type=get_attr];
-"376 compressed_weight_updated_constant18" [id=376, type=get_attr];
-"377 quantize_per_channel_default_18" [id=377, type=quantize_per_channel];
-"378 dequantize_per_channel_default_18" [id=378, type=dequantize_per_channel];
-"379 _param_constant57_0_0" [id=379, type=get_attr];
-"380 linear_17" [id=380, type=linear];
-"381 view_39" [id=381, type=view];
-"382 transpose_29" [id=382, type=transpose];
-"383 dropout_13" [id=383, type=dropout];
-"384 add_9" [id=384, type=add];
-"385 _param_constant58" [id=385, type=get_attr];
-"386 _param_constant59" [id=386, type=get_attr];
-"387 layer_norm_9" [id=387, type=layer_norm];
-"388 layer_norm_9_0_0_nncf_smooth_quant_0" [id=388, type=call_module];
-"389 quantize_per_tensor_default_29" [id=389, type=quantize_per_tensor];
-"390 dequantize_per_tensor_default_29" [id=390, type=dequantize_per_tensor];
-"391 linear_18_scale_0" [id=391, type=get_attr];
-"392 linear_18_zero_point_0" [id=392, type=get_attr];
-"393 compressed_weight_updated_constant19" [id=393, type=get_attr];
-"394 quantize_per_channel_default_19" [id=394, type=quantize_per_channel];
-"395 dequantize_per_channel_default_19" [id=395, type=dequantize_per_channel];
-"396 _param_constant61_0_0" [id=396, type=get_attr];
-"397 linear_18" [id=397, type=linear];
-"398 gelu_4" [id=398, type=gelu];
-"399 dropout_14" [id=399, type=dropout];
-"400 dropout_14_0_0_nncf_smooth_quant_0" [id=400, type=call_module];
-"401 quantize_per_tensor_default_30" [id=401, type=quantize_per_tensor];
-"402 dequantize_per_tensor_default_30" [id=402, type=dequantize_per_tensor];
-"403 linear_19_scale_0" [id=403, type=get_attr];
-"404 linear_19_zero_point_0" [id=404, type=get_attr];
-"405 compressed_weight_updated_constant20" [id=405, type=get_attr];
-"406 quantize_per_channel_default_20" [id=406, type=quantize_per_channel];
-"407 dequantize_per_channel_default_20" [id=407, type=dequantize_per_channel];
-"408 _param_constant63_0_0" [id=408, type=get_attr];
-"409 linear_19" [id=409, type=linear];
-"410 dropout_15" [id=410, type=dropout];
-"411 add_10" [id=411, type=add];
-"412 _param_constant64" [id=412, type=get_attr];
-"413 _param_constant65" [id=413, type=get_attr];
-"414 layer_norm_10" [id=414, type=layer_norm];
-"415 transpose_30" [id=415, type=transpose];
-"416 transpose_30_0_0_nncf_smooth_quant_0" [id=416, type=call_module];
-"417 quantize_per_tensor_default_31" [id=417, type=quantize_per_tensor];
-"418 dequantize_per_tensor_default_31" [id=418, type=dequantize_per_tensor];
-"419 linear_20_scale_0" [id=419, type=get_attr];
-"420 linear_20_zero_point_0" [id=420, type=get_attr];
-"421 compressed_weight_updated_constant21" [id=421, type=get_attr];
-"422 quantize_per_channel_default_21" [id=422, type=quantize_per_channel];
-"423 dequantize_per_channel_default_21" [id=423, type=dequantize_per_channel];
-"424 _param_constant67_0_0" [id=424, type=get_attr];
-"425 linear_20" [id=425, type=linear];
-"426 unflatten_5" [id=426, type=unflatten];
-"427 unsqueeze_5" [id=427, type=unsqueeze];
-"428 transpose_31" [id=428, type=transpose];
-"429 squeeze_5" [id=429, type=squeeze];
-"430 contiguous_5" [id=430, type=contiguous];
-"431 quantize_per_tensor_default_32" [id=431, type=quantize_per_tensor];
-"432 dequantize_per_tensor_default_32" [id=432, type=dequantize_per_tensor];
-"433 select_15" [id=433, type=select];
-"434 quantize_per_tensor_default_33" [id=434, type=quantize_per_tensor];
-"435 dequantize_per_tensor_default_33" [id=435, type=dequantize_per_tensor];
-"436 select_16" [id=436, type=select];
-"437 select_17" [id=437, type=select];
-"438 view_40" [id=438, type=view];
-"439 transpose_32" [id=439, type=transpose];
-"440 view_41" [id=440, type=view];
-"441 transpose_33" [id=441, type=transpose];
-"442 view_42" [id=442, type=view];
-"443 transpose_34" [id=443, type=transpose];
-"444 view_43" [id=444, type=view];
-"445 view_44" [id=445, type=view];
-"446 view_45" [id=446, type=view];
-"447 scaled_dot_product_attention_5" [id=447, type=scaled_dot_product_attention];
-"448 permute_6" [id=448, type=permute];
-"449 view_46" [id=449, type=view];
-"450 view_46_0_0_nncf_smooth_quant_0" [id=450, type=call_module];
-"451 quantize_per_tensor_default_34" [id=451, type=quantize_per_tensor];
-"452 dequantize_per_tensor_default_34" [id=452, type=dequantize_per_tensor];
-"453 linear_21_scale_0" [id=453, type=get_attr];
-"454 linear_21_zero_point_0" [id=454, type=get_attr];
-"455 compressed_weight_updated_constant22" [id=455, type=get_attr];
-"456 quantize_per_channel_default_22" [id=456, type=quantize_per_channel];
-"457 dequantize_per_channel_default_22" [id=457, type=dequantize_per_channel];
-"458 _param_constant69_0_0" [id=458, type=get_attr];
-"459 linear_21" [id=459, type=linear];
-"460 view_47" [id=460, type=view];
-"461 transpose_35" [id=461, type=transpose];
-"462 dropout_16" [id=462, type=dropout];
-"463 add_11" [id=463, type=add];
-"464 _param_constant70" [id=464, type=get_attr];
-"465 _param_constant71" [id=465, type=get_attr];
-"466 layer_norm_11" [id=466, type=layer_norm];
-"467 layer_norm_11_0_0_nncf_smooth_quant_0" [id=467, type=call_module];
-"468 quantize_per_tensor_default_35" [id=468, type=quantize_per_tensor];
-"469 dequantize_per_tensor_default_35" [id=469, type=dequantize_per_tensor];
-"470 linear_22_scale_0" [id=470, type=get_attr];
-"471 linear_22_zero_point_0" [id=471, type=get_attr];
-"472 compressed_weight_updated_constant23" [id=472, type=get_attr];
-"473 quantize_per_channel_default_23" [id=473, type=quantize_per_channel];
-"474 dequantize_per_channel_default_23" [id=474, type=dequantize_per_channel];
-"475 _param_constant73_0_0" [id=475, type=get_attr];
-"476 linear_22" [id=476, type=linear];
-"477 gelu_5" [id=477, type=gelu];
-"478 dropout_17" [id=478, type=dropout];
-"479 dropout_17_0_0_nncf_smooth_quant_0" [id=479, type=call_module];
-"480 quantize_per_tensor_default_36" [id=480, type=quantize_per_tensor];
-"481 dequantize_per_tensor_default_36" [id=481, type=dequantize_per_tensor];
-"482 linear_23_scale_0" [id=482, type=get_attr];
-"483 linear_23_zero_point_0" [id=483, type=get_attr];
-"484 compressed_weight_updated_constant24" [id=484, type=get_attr];
-"485 quantize_per_channel_default_24" [id=485, type=quantize_per_channel];
-"486 dequantize_per_channel_default_24" [id=486, type=dequantize_per_channel];
-"487 _param_constant75_0_0" [id=487, type=get_attr];
-"488 linear_23" [id=488, type=linear];
-"489 dropout_18" [id=489, type=dropout];
-"490 add_12" [id=490, type=add];
-"491 _param_constant76" [id=491, type=get_attr];
-"492 _param_constant77" [id=492, type=get_attr];
-"493 layer_norm_12" [id=493, type=layer_norm];
-"494 transpose_36" [id=494, type=transpose];
-"495 transpose_36_0_0_nncf_smooth_quant_0" [id=495, type=call_module];
-"496 quantize_per_tensor_default_37" [id=496, type=quantize_per_tensor];
-"497 dequantize_per_tensor_default_37" [id=497, type=dequantize_per_tensor];
-"498 linear_24_scale_0" [id=498, type=get_attr];
-"499 linear_24_zero_point_0" [id=499, type=get_attr];
-"500 compressed_weight_updated_constant25" [id=500, type=get_attr];
-"501 quantize_per_channel_default_25" [id=501, type=quantize_per_channel];
-"502 dequantize_per_channel_default_25" [id=502, type=dequantize_per_channel];
-"503 _param_constant79_0_0" [id=503, type=get_attr];
-"504 linear_24" [id=504, type=linear];
-"505 unflatten_6" [id=505, type=unflatten];
-"506 unsqueeze_6" [id=506, type=unsqueeze];
-"507 transpose_37" [id=507, type=transpose];
-"508 squeeze_6" [id=508, type=squeeze];
-"509 contiguous_6" [id=509, type=contiguous];
-"510 quantize_per_tensor_default_38" [id=510, type=quantize_per_tensor];
-"511 dequantize_per_tensor_default_38" [id=511, type=dequantize_per_tensor];
-"512 select_18" [id=512, type=select];
-"513 quantize_per_tensor_default_39" [id=513, type=quantize_per_tensor];
-"514 dequantize_per_tensor_default_39" [id=514, type=dequantize_per_tensor];
-"515 select_19" [id=515, type=select];
-"516 select_20" [id=516, type=select];
-"517 view_48" [id=517, type=view];
-"518 transpose_38" [id=518, type=transpose];
-"519 view_49" [id=519, type=view];
-"520 transpose_39" [id=520, type=transpose];
-"521 view_50" [id=521, type=view];
-"522 transpose_40" [id=522, type=transpose];
-"523 view_51" [id=523, type=view];
-"524 view_52" [id=524, type=view];
-"525 view_53" [id=525, type=view];
-"526 scaled_dot_product_attention_6" [id=526, type=scaled_dot_product_attention];
-"527 permute_7" [id=527, type=permute];
-"528 view_54" [id=528, type=view];
-"529 view_54_0_0_nncf_smooth_quant_0" [id=529, type=call_module];
-"530 quantize_per_tensor_default_40" [id=530, type=quantize_per_tensor];
-"531 dequantize_per_tensor_default_40" [id=531, type=dequantize_per_tensor];
-"532 linear_25_scale_0" [id=532, type=get_attr];
-"533 linear_25_zero_point_0" [id=533, type=get_attr];
-"534 compressed_weight_updated_constant26" [id=534, type=get_attr];
-"535 quantize_per_channel_default_26" [id=535, type=quantize_per_channel];
-"536 dequantize_per_channel_default_26" [id=536, type=dequantize_per_channel];
-"537 _param_constant81_0_0" [id=537, type=get_attr];
-"538 linear_25" [id=538, type=linear];
-"539 view_55" [id=539, type=view];
-"540 transpose_41" [id=540, type=transpose];
-"541 dropout_19" [id=541, type=dropout];
-"542 add_13" [id=542, type=add];
-"543 _param_constant82" [id=543, type=get_attr];
-"544 _param_constant83" [id=544, type=get_attr];
-"545 layer_norm_13" [id=545, type=layer_norm];
-"546 layer_norm_13_0_0_nncf_smooth_quant_0" [id=546, type=call_module];
-"547 quantize_per_tensor_default_41" [id=547, type=quantize_per_tensor];
-"548 dequantize_per_tensor_default_41" [id=548, type=dequantize_per_tensor];
-"549 linear_26_scale_0" [id=549, type=get_attr];
-"550 linear_26_zero_point_0" [id=550, type=get_attr];
-"551 compressed_weight_updated_constant27" [id=551, type=get_attr];
-"552 quantize_per_channel_default_27" [id=552, type=quantize_per_channel];
-"553 dequantize_per_channel_default_27" [id=553, type=dequantize_per_channel];
-"554 _param_constant85_0_0" [id=554, type=get_attr];
-"555 linear_26" [id=555, type=linear];
-"556 gelu_6" [id=556, type=gelu];
-"557 dropout_20" [id=557, type=dropout];
-"558 dropout_20_0_0_nncf_smooth_quant_0" [id=558, type=call_module];
-"559 quantize_per_tensor_default_42" [id=559, type=quantize_per_tensor];
-"560 dequantize_per_tensor_default_42" [id=560, type=dequantize_per_tensor];
-"561 linear_27_scale_0" [id=561, type=get_attr];
-"562 linear_27_zero_point_0" [id=562, type=get_attr];
-"563 compressed_weight_updated_constant28" [id=563, type=get_attr];
-"564 quantize_per_channel_default_28" [id=564, type=quantize_per_channel];
-"565 dequantize_per_channel_default_28" [id=565, type=dequantize_per_channel];
-"566 _param_constant87_0_0" [id=566, type=get_attr];
-"567 linear_27" [id=567, type=linear];
-"568 dropout_21" [id=568, type=dropout];
-"569 add_14" [id=569, type=add];
-"570 _param_constant88" [id=570, type=get_attr];
-"571 _param_constant89" [id=571, type=get_attr];
-"572 layer_norm_14" [id=572, type=layer_norm];
-"573 transpose_42" [id=573, type=transpose];
-"574 transpose_42_0_0_nncf_smooth_quant_0" [id=574, type=call_module];
-"575 quantize_per_tensor_default_43" [id=575, type=quantize_per_tensor];
-"576 dequantize_per_tensor_default_43" [id=576, type=dequantize_per_tensor];
-"577 linear_28_scale_0" [id=577, type=get_attr];
-"578 linear_28_zero_point_0" [id=578, type=get_attr];
-"579 compressed_weight_updated_constant29" [id=579, type=get_attr];
-"580 quantize_per_channel_default_29" [id=580, type=quantize_per_channel];
-"581 dequantize_per_channel_default_29" [id=581, type=dequantize_per_channel];
-"582 _param_constant91_0_0" [id=582, type=get_attr];
-"583 linear_28" [id=583, type=linear];
-"584 unflatten_7" [id=584, type=unflatten];
-"585 unsqueeze_7" [id=585, type=unsqueeze];
-"586 transpose_43" [id=586, type=transpose];
-"587 squeeze_7" [id=587, type=squeeze];
-"588 contiguous_7" [id=588, type=contiguous];
-"589 quantize_per_tensor_default_44" [id=589, type=quantize_per_tensor];
-"590 dequantize_per_tensor_default_44" [id=590, type=dequantize_per_tensor];
-"591 select_21" [id=591, type=select];
-"592 quantize_per_tensor_default_45" [id=592, type=quantize_per_tensor];
-"593 dequantize_per_tensor_default_45" [id=593, type=dequantize_per_tensor];
-"594 select_22" [id=594, type=select];
-"595 select_23" [id=595, type=select];
-"596 view_56" [id=596, type=view];
-"597 transpose_44" [id=597, type=transpose];
-"598 view_57" [id=598, type=view];
-"599 transpose_45" [id=599, type=transpose];
-"600 view_58" [id=600, type=view];
-"601 transpose_46" [id=601, type=transpose];
-"602 view_59" [id=602, type=view];
-"603 view_60" [id=603, type=view];
-"604 view_61" [id=604, type=view];
-"605 scaled_dot_product_attention_7" [id=605, type=scaled_dot_product_attention];
-"606 permute_8" [id=606, type=permute];
-"607 view_62" [id=607, type=view];
-"608 view_62_0_0_nncf_smooth_quant_0" [id=608, type=call_module];
-"609 quantize_per_tensor_default_46" [id=609, type=quantize_per_tensor];
-"610 dequantize_per_tensor_default_46" [id=610, type=dequantize_per_tensor];
-"611 linear_29_scale_0" [id=611, type=get_attr];
-"612 linear_29_zero_point_0" [id=612, type=get_attr];
-"613 compressed_weight_updated_constant30" [id=613, type=get_attr];
-"614 quantize_per_channel_default_30" [id=614, type=quantize_per_channel];
-"615 dequantize_per_channel_default_30" [id=615, type=dequantize_per_channel];
-"616 _param_constant93_0_0" [id=616, type=get_attr];
-"617 linear_29" [id=617, type=linear];
-"618 view_63" [id=618, type=view];
-"619 transpose_47" [id=619, type=transpose];
-"620 dropout_22" [id=620, type=dropout];
-"621 add_15" [id=621, type=add];
-"622 _param_constant94" [id=622, type=get_attr];
-"623 _param_constant95" [id=623, type=get_attr];
-"624 layer_norm_15" [id=624, type=layer_norm];
-"625 layer_norm_15_0_0_nncf_smooth_quant_0" [id=625, type=call_module];
-"626 quantize_per_tensor_default_47" [id=626, type=quantize_per_tensor];
-"627 dequantize_per_tensor_default_47" [id=627, type=dequantize_per_tensor];
-"628 linear_30_scale_0" [id=628, type=get_attr];
-"629 linear_30_zero_point_0" [id=629, type=get_attr];
-"630 compressed_weight_updated_constant31" [id=630, type=get_attr];
-"631 quantize_per_channel_default_31" [id=631, type=quantize_per_channel];
-"632 dequantize_per_channel_default_31" [id=632, type=dequantize_per_channel];
-"633 _param_constant97_0_0" [id=633, type=get_attr];
-"634 linear_30" [id=634, type=linear];
-"635 gelu_7" [id=635, type=gelu];
-"636 dropout_23" [id=636, type=dropout];
-"637 dropout_23_0_0_nncf_smooth_quant_0" [id=637, type=call_module];
-"638 quantize_per_tensor_default_48" [id=638, type=quantize_per_tensor];
-"639 dequantize_per_tensor_default_48" [id=639, type=dequantize_per_tensor];
-"640 linear_31_scale_0" [id=640, type=get_attr];
-"641 linear_31_zero_point_0" [id=641, type=get_attr];
-"642 compressed_weight_updated_constant32" [id=642, type=get_attr];
-"643 quantize_per_channel_default_32" [id=643, type=quantize_per_channel];
-"644 dequantize_per_channel_default_32" [id=644, type=dequantize_per_channel];
-"645 _param_constant99_0_0" [id=645, type=get_attr];
-"646 linear_31" [id=646, type=linear];
-"647 dropout_24" [id=647, type=dropout];
-"648 add_16" [id=648, type=add];
-"649 _param_constant100" [id=649, type=get_attr];
-"650 _param_constant101" [id=650, type=get_attr];
-"651 layer_norm_16" [id=651, type=layer_norm];
-"652 transpose_48" [id=652, type=transpose];
-"653 transpose_48_0_0_nncf_smooth_quant_0" [id=653, type=call_module];
-"654 quantize_per_tensor_default_49" [id=654, type=quantize_per_tensor];
-"655 dequantize_per_tensor_default_49" [id=655, type=dequantize_per_tensor];
-"656 linear_32_scale_0" [id=656, type=get_attr];
-"657 linear_32_zero_point_0" [id=657, type=get_attr];
-"658 compressed_weight_updated_constant33" [id=658, type=get_attr];
-"659 quantize_per_channel_default_33" [id=659, type=quantize_per_channel];
-"660 dequantize_per_channel_default_33" [id=660, type=dequantize_per_channel];
-"661 _param_constant103_0_0" [id=661, type=get_attr];
-"662 linear_32" [id=662, type=linear];
-"663 unflatten_8" [id=663, type=unflatten];
-"664 unsqueeze_8" [id=664, type=unsqueeze];
-"665 transpose_49" [id=665, type=transpose];
-"666 squeeze_8" [id=666, type=squeeze];
-"667 contiguous_8" [id=667, type=contiguous];
-"668 quantize_per_tensor_default_50" [id=668, type=quantize_per_tensor];
-"669 dequantize_per_tensor_default_50" [id=669, type=dequantize_per_tensor];
-"670 select_24" [id=670, type=select];
-"671 quantize_per_tensor_default_51" [id=671, type=quantize_per_tensor];
-"672 dequantize_per_tensor_default_51" [id=672, type=dequantize_per_tensor];
-"673 select_25" [id=673, type=select];
-"674 select_26" [id=674, type=select];
-"675 view_64" [id=675, type=view];
-"676 transpose_50" [id=676, type=transpose];
-"677 view_65" [id=677, type=view];
-"678 transpose_51" [id=678, type=transpose];
-"679 view_66" [id=679, type=view];
-"680 transpose_52" [id=680, type=transpose];
-"681 view_67" [id=681, type=view];
-"682 view_68" [id=682, type=view];
-"683 view_69" [id=683, type=view];
-"684 scaled_dot_product_attention_8" [id=684, type=scaled_dot_product_attention];
-"685 permute_9" [id=685, type=permute];
-"686 view_70" [id=686, type=view];
-"687 view_70_0_0_nncf_smooth_quant_0" [id=687, type=call_module];
-"688 quantize_per_tensor_default_52" [id=688, type=quantize_per_tensor];
-"689 dequantize_per_tensor_default_52" [id=689, type=dequantize_per_tensor];
-"690 linear_33_scale_0" [id=690, type=get_attr];
-"691 linear_33_zero_point_0" [id=691, type=get_attr];
-"692 compressed_weight_updated_constant34" [id=692, type=get_attr];
-"693 quantize_per_channel_default_34" [id=693, type=quantize_per_channel];
-"694 dequantize_per_channel_default_34" [id=694, type=dequantize_per_channel];
-"695 _param_constant105_0_0" [id=695, type=get_attr];
-"696 linear_33" [id=696, type=linear];
-"697 view_71" [id=697, type=view];
-"698 transpose_53" [id=698, type=transpose];
-"699 dropout_25" [id=699, type=dropout];
-"700 add_17" [id=700, type=add];
-"701 _param_constant106" [id=701, type=get_attr];
-"702 _param_constant107" [id=702, type=get_attr];
-"703 layer_norm_17" [id=703, type=layer_norm];
-"704 layer_norm_17_0_0_nncf_smooth_quant_0" [id=704, type=call_module];
-"705 quantize_per_tensor_default_53" [id=705, type=quantize_per_tensor];
-"706 dequantize_per_tensor_default_53" [id=706, type=dequantize_per_tensor];
-"707 linear_34_scale_0" [id=707, type=get_attr];
-"708 linear_34_zero_point_0" [id=708, type=get_attr];
-"709 compressed_weight_updated_constant35" [id=709, type=get_attr];
-"710 quantize_per_channel_default_35" [id=710, type=quantize_per_channel];
-"711 dequantize_per_channel_default_35" [id=711, type=dequantize_per_channel];
-"712 _param_constant109_0_0" [id=712, type=get_attr];
-"713 linear_34" [id=713, type=linear];
-"714 gelu_8" [id=714, type=gelu];
-"715 dropout_26" [id=715, type=dropout];
-"716 dropout_26_0_0_nncf_smooth_quant_0" [id=716, type=call_module];
-"717 quantize_per_tensor_default_54" [id=717, type=quantize_per_tensor];
-"718 dequantize_per_tensor_default_54" [id=718, type=dequantize_per_tensor];
-"719 linear_35_scale_0" [id=719, type=get_attr];
-"720 linear_35_zero_point_0" [id=720, type=get_attr];
-"721 compressed_weight_updated_constant36" [id=721, type=get_attr];
-"722 quantize_per_channel_default_36" [id=722, type=quantize_per_channel];
-"723 dequantize_per_channel_default_36" [id=723, type=dequantize_per_channel];
-"724 _param_constant111_0_0" [id=724, type=get_attr];
-"725 linear_35" [id=725, type=linear];
-"726 dropout_27" [id=726, type=dropout];
-"727 add_18" [id=727, type=add];
-"728 _param_constant112" [id=728, type=get_attr];
-"729 _param_constant113" [id=729, type=get_attr];
-"730 layer_norm_18" [id=730, type=layer_norm];
-"731 transpose_54" [id=731, type=transpose];
-"732 transpose_54_0_0_nncf_smooth_quant_0" [id=732, type=call_module];
-"733 quantize_per_tensor_default_55" [id=733, type=quantize_per_tensor];
-"734 dequantize_per_tensor_default_55" [id=734, type=dequantize_per_tensor];
-"735 linear_36_scale_0" [id=735, type=get_attr];
-"736 linear_36_zero_point_0" [id=736, type=get_attr];
-"737 compressed_weight_updated_constant37" [id=737, type=get_attr];
-"738 quantize_per_channel_default_37" [id=738, type=quantize_per_channel];
-"739 dequantize_per_channel_default_37" [id=739, type=dequantize_per_channel];
-"740 _param_constant115_0_0" [id=740, type=get_attr];
-"741 linear_36" [id=741, type=linear];
-"742 unflatten_9" [id=742, type=unflatten];
-"743 unsqueeze_9" [id=743, type=unsqueeze];
-"744 transpose_55" [id=744, type=transpose];
-"745 squeeze_9" [id=745, type=squeeze];
-"746 contiguous_9" [id=746, type=contiguous];
-"747 quantize_per_tensor_default_56" [id=747, type=quantize_per_tensor];
-"748 dequantize_per_tensor_default_56" [id=748, type=dequantize_per_tensor];
-"749 select_27" [id=749, type=select];
-"750 quantize_per_tensor_default_57" [id=750, type=quantize_per_tensor];
-"751 dequantize_per_tensor_default_57" [id=751, type=dequantize_per_tensor];
-"752 select_28" [id=752, type=select];
-"753 select_29" [id=753, type=select];
-"754 view_72" [id=754, type=view];
-"755 transpose_56" [id=755, type=transpose];
-"756 view_73" [id=756, type=view];
-"757 transpose_57" [id=757, type=transpose];
-"758 view_74" [id=758, type=view];
-"759 transpose_58" [id=759, type=transpose];
-"760 view_75" [id=760, type=view];
-"761 view_76" [id=761, type=view];
-"762 view_77" [id=762, type=view];
-"763 scaled_dot_product_attention_9" [id=763, type=scaled_dot_product_attention];
-"764 permute_10" [id=764, type=permute];
-"765 view_78" [id=765, type=view];
-"766 view_78_0_0_nncf_smooth_quant_0" [id=766, type=call_module];
-"767 quantize_per_tensor_default_58" [id=767, type=quantize_per_tensor];
-"768 dequantize_per_tensor_default_58" [id=768, type=dequantize_per_tensor];
-"769 linear_37_scale_0" [id=769, type=get_attr];
-"770 linear_37_zero_point_0" [id=770, type=get_attr];
-"771 compressed_weight_updated_constant38" [id=771, type=get_attr];
-"772 quantize_per_channel_default_38" [id=772, type=quantize_per_channel];
-"773 dequantize_per_channel_default_38" [id=773, type=dequantize_per_channel];
-"774 _param_constant117_0_0" [id=774, type=get_attr];
-"775 linear_37" [id=775, type=linear];
-"776 view_79" [id=776, type=view];
-"777 transpose_59" [id=777, type=transpose];
-"778 dropout_28" [id=778, type=dropout];
-"779 add_19" [id=779, type=add];
-"780 _param_constant118" [id=780, type=get_attr];
-"781 _param_constant119" [id=781, type=get_attr];
-"782 layer_norm_19" [id=782, type=layer_norm];
-"783 layer_norm_19_0_0_nncf_smooth_quant_0" [id=783, type=call_module];
-"784 quantize_per_tensor_default_59" [id=784, type=quantize_per_tensor];
-"785 dequantize_per_tensor_default_59" [id=785, type=dequantize_per_tensor];
-"786 linear_38_scale_0" [id=786, type=get_attr];
-"787 linear_38_zero_point_0" [id=787, type=get_attr];
-"788 compressed_weight_updated_constant39" [id=788, type=get_attr];
-"789 quantize_per_channel_default_39" [id=789, type=quantize_per_channel];
-"790 dequantize_per_channel_default_39" [id=790, type=dequantize_per_channel];
-"791 _param_constant121_0_0" [id=791, type=get_attr];
-"792 linear_38" [id=792, type=linear];
-"793 gelu_9" [id=793, type=gelu];
-"794 dropout_29" [id=794, type=dropout];
-"795 dropout_29_0_0_nncf_smooth_quant_0" [id=795, type=call_module];
-"796 quantize_per_tensor_default_60" [id=796, type=quantize_per_tensor];
-"797 dequantize_per_tensor_default_60" [id=797, type=dequantize_per_tensor];
-"798 linear_39_scale_0" [id=798, type=get_attr];
-"799 linear_39_zero_point_0" [id=799, type=get_attr];
-"800 compressed_weight_updated_constant40" [id=800, type=get_attr];
-"801 quantize_per_channel_default_40" [id=801, type=quantize_per_channel];
-"802 dequantize_per_channel_default_40" [id=802, type=dequantize_per_channel];
-"803 _param_constant123_0_0" [id=803, type=get_attr];
-"804 linear_39" [id=804, type=linear];
-"805 dropout_30" [id=805, type=dropout];
-"806 add_20" [id=806, type=add];
-"807 _param_constant124" [id=807, type=get_attr];
-"808 _param_constant125" [id=808, type=get_attr];
-"809 layer_norm_20" [id=809, type=layer_norm];
-"810 transpose_60" [id=810, type=transpose];
-"811 transpose_60_0_0_nncf_smooth_quant_0" [id=811, type=call_module];
-"812 quantize_per_tensor_default_61" [id=812, type=quantize_per_tensor];
-"813 dequantize_per_tensor_default_61" [id=813, type=dequantize_per_tensor];
-"814 linear_40_scale_0" [id=814, type=get_attr];
-"815 linear_40_zero_point_0" [id=815, type=get_attr];
-"816 compressed_weight_updated_constant41" [id=816, type=get_attr];
-"817 quantize_per_channel_default_41" [id=817, type=quantize_per_channel];
-"818 dequantize_per_channel_default_41" [id=818, type=dequantize_per_channel];
-"819 _param_constant127_0_0" [id=819, type=get_attr];
-"820 linear_40" [id=820, type=linear];
-"821 unflatten_10" [id=821, type=unflatten];
-"822 unsqueeze_10" [id=822, type=unsqueeze];
-"823 transpose_61" [id=823, type=transpose];
-"824 squeeze_10" [id=824, type=squeeze];
-"825 contiguous_10" [id=825, type=contiguous];
-"826 quantize_per_tensor_default_62" [id=826, type=quantize_per_tensor];
-"827 dequantize_per_tensor_default_62" [id=827, type=dequantize_per_tensor];
-"828 select_30" [id=828, type=select];
-"829 quantize_per_tensor_default_63" [id=829, type=quantize_per_tensor];
-"830 dequantize_per_tensor_default_63" [id=830, type=dequantize_per_tensor];
-"831 select_31" [id=831, type=select];
-"832 select_32" [id=832, type=select];
-"833 view_80" [id=833, type=view];
-"834 transpose_62" [id=834, type=transpose];
-"835 view_81" [id=835, type=view];
-"836 transpose_63" [id=836, type=transpose];
-"837 view_82" [id=837, type=view];
-"838 transpose_64" [id=838, type=transpose];
-"839 view_83" [id=839, type=view];
-"840 view_84" [id=840, type=view];
-"841 view_85" [id=841, type=view];
-"842 scaled_dot_product_attention_10" [id=842, type=scaled_dot_product_attention];
-"843 permute_11" [id=843, type=permute];
-"844 view_86" [id=844, type=view];
-"845 view_86_0_0_nncf_smooth_quant_0" [id=845, type=call_module];
-"846 quantize_per_tensor_default_64" [id=846, type=quantize_per_tensor];
-"847 dequantize_per_tensor_default_64" [id=847, type=dequantize_per_tensor];
-"848 linear_41_scale_0" [id=848, type=get_attr];
-"849 linear_41_zero_point_0" [id=849, type=get_attr];
-"850 compressed_weight_updated_constant42" [id=850, type=get_attr];
-"851 quantize_per_channel_default_42" [id=851, type=quantize_per_channel];
-"852 dequantize_per_channel_default_42" [id=852, type=dequantize_per_channel];
-"853 _param_constant129_0_0" [id=853, type=get_attr];
-"854 linear_41" [id=854, type=linear];
-"855 view_87" [id=855, type=view];
-"856 transpose_65" [id=856, type=transpose];
-"857 dropout_31" [id=857, type=dropout];
-"858 add_21" [id=858, type=add];
-"859 _param_constant130" [id=859, type=get_attr];
-"860 _param_constant131" [id=860, type=get_attr];
-"861 layer_norm_21" [id=861, type=layer_norm];
-"862 layer_norm_21_0_0_nncf_smooth_quant_0" [id=862, type=call_module];
-"863 quantize_per_tensor_default_65" [id=863, type=quantize_per_tensor];
-"864 dequantize_per_tensor_default_65" [id=864, type=dequantize_per_tensor];
-"865 linear_42_scale_0" [id=865, type=get_attr];
-"866 linear_42_zero_point_0" [id=866, type=get_attr];
-"867 compressed_weight_updated_constant43" [id=867, type=get_attr];
-"868 quantize_per_channel_default_43" [id=868, type=quantize_per_channel];
-"869 dequantize_per_channel_default_43" [id=869, type=dequantize_per_channel];
-"870 _param_constant133_0_0" [id=870, type=get_attr];
-"871 linear_42" [id=871, type=linear];
-"872 gelu_10" [id=872, type=gelu];
-"873 dropout_32" [id=873, type=dropout];
-"874 dropout_32_0_0_nncf_smooth_quant_0" [id=874, type=call_module];
-"875 quantize_per_tensor_default_66" [id=875, type=quantize_per_tensor];
-"876 dequantize_per_tensor_default_66" [id=876, type=dequantize_per_tensor];
-"877 linear_43_scale_0" [id=877, type=get_attr];
-"878 linear_43_zero_point_0" [id=878, type=get_attr];
-"879 compressed_weight_updated_constant44" [id=879, type=get_attr];
-"880 quantize_per_channel_default_44" [id=880, type=quantize_per_channel];
-"881 dequantize_per_channel_default_44" [id=881, type=dequantize_per_channel];
-"882 _param_constant135_0_0" [id=882, type=get_attr];
-"883 linear_43" [id=883, type=linear];
-"884 dropout_33" [id=884, type=dropout];
-"885 add_22" [id=885, type=add];
-"886 _param_constant136" [id=886, type=get_attr];
-"887 _param_constant137" [id=887, type=get_attr];
-"888 layer_norm_22" [id=888, type=layer_norm];
-"889 transpose_66" [id=889, type=transpose];
-"890 transpose_66_0_0_nncf_smooth_quant_0" [id=890, type=call_module];
-"891 quantize_per_tensor_default_67" [id=891, type=quantize_per_tensor];
-"892 dequantize_per_tensor_default_67" [id=892, type=dequantize_per_tensor];
-"893 linear_44_scale_0" [id=893, type=get_attr];
-"894 linear_44_zero_point_0" [id=894, type=get_attr];
-"895 compressed_weight_updated_constant45" [id=895, type=get_attr];
-"896 quantize_per_channel_default_45" [id=896, type=quantize_per_channel];
-"897 dequantize_per_channel_default_45" [id=897, type=dequantize_per_channel];
-"898 _param_constant139_0_0" [id=898, type=get_attr];
-"899 linear_44" [id=899, type=linear];
-"900 unflatten_11" [id=900, type=unflatten];
-"901 unsqueeze_11" [id=901, type=unsqueeze];
-"902 transpose_67" [id=902, type=transpose];
-"903 squeeze_11" [id=903, type=squeeze];
-"904 contiguous_11" [id=904, type=contiguous];
-"905 quantize_per_tensor_default_68" [id=905, type=quantize_per_tensor];
-"906 dequantize_per_tensor_default_68" [id=906, type=dequantize_per_tensor];
-"907 select_33" [id=907, type=select];
-"908 quantize_per_tensor_default_69" [id=908, type=quantize_per_tensor];
-"909 dequantize_per_tensor_default_69" [id=909, type=dequantize_per_tensor];
-"910 select_34" [id=910, type=select];
-"911 select_35" [id=911, type=select];
-"912 view_88" [id=912, type=view];
-"913 transpose_68" [id=913, type=transpose];
-"914 view_89" [id=914, type=view];
-"915 transpose_69" [id=915, type=transpose];
-"916 view_90" [id=916, type=view];
-"917 transpose_70" [id=917, type=transpose];
-"918 view_91" [id=918, type=view];
-"919 view_92" [id=919, type=view];
-"920 view_93" [id=920, type=view];
-"921 scaled_dot_product_attention_11" [id=921, type=scaled_dot_product_attention];
-"922 permute_12" [id=922, type=permute];
-"923 view_94" [id=923, type=view];
-"924 view_94_0_0_nncf_smooth_quant_0" [id=924, type=call_module];
-"925 quantize_per_tensor_default_70" [id=925, type=quantize_per_tensor];
-"926 dequantize_per_tensor_default_70" [id=926, type=dequantize_per_tensor];
-"927 linear_45_scale_0" [id=927, type=get_attr];
-"928 linear_45_zero_point_0" [id=928, type=get_attr];
-"929 compressed_weight_updated_constant46" [id=929, type=get_attr];
-"930 quantize_per_channel_default_46" [id=930, type=quantize_per_channel];
-"931 dequantize_per_channel_default_46" [id=931, type=dequantize_per_channel];
-"932 _param_constant141_0_0" [id=932, type=get_attr];
-"933 linear_45" [id=933, type=linear];
-"934 view_95" [id=934, type=view];
-"935 transpose_71" [id=935, type=transpose];
-"936 dropout_34" [id=936, type=dropout];
-"937 add_23" [id=937, type=add];
-"938 _param_constant142" [id=938, type=get_attr];
-"939 _param_constant143" [id=939, type=get_attr];
-"940 layer_norm_23" [id=940, type=layer_norm];
-"941 layer_norm_23_0_0_nncf_smooth_quant_0" [id=941, type=call_module];
-"942 quantize_per_tensor_default_71" [id=942, type=quantize_per_tensor];
-"943 dequantize_per_tensor_default_71" [id=943, type=dequantize_per_tensor];
-"944 linear_46_scale_0" [id=944, type=get_attr];
-"945 linear_46_zero_point_0" [id=945, type=get_attr];
-"946 compressed_weight_updated_constant47" [id=946, type=get_attr];
-"947 quantize_per_channel_default_47" [id=947, type=quantize_per_channel];
-"948 dequantize_per_channel_default_47" [id=948, type=dequantize_per_channel];
-"949 _param_constant145_0_0" [id=949, type=get_attr];
-"950 linear_46" [id=950, type=linear];
-"951 gelu_11" [id=951, type=gelu];
-"952 dropout_35" [id=952, type=dropout];
-"953 dropout_35_0_0_nncf_smooth_quant_0" [id=953, type=call_module];
-"954 quantize_per_tensor_default_72" [id=954, type=quantize_per_tensor];
-"955 dequantize_per_tensor_default_72" [id=955, type=dequantize_per_tensor];
-"956 linear_47_scale_0" [id=956, type=get_attr];
-"957 linear_47_zero_point_0" [id=957, type=get_attr];
-"958 compressed_weight_updated_constant48" [id=958, type=get_attr];
-"959 quantize_per_channel_default_48" [id=959, type=quantize_per_channel];
-"960 dequantize_per_channel_default_48" [id=960, type=dequantize_per_channel];
-"961 _param_constant147_0_0" [id=961, type=get_attr];
-"962 linear_47" [id=962, type=linear];
-"963 dropout_36" [id=963, type=dropout];
-"964 add_24" [id=964, type=add];
-"965 _param_constant148" [id=965, type=get_attr];
-"966 _param_constant149" [id=966, type=get_attr];
+"0 encoder_pos_embedding" [id=0, type=get_attr];
+"1 encoder_layers_encoder_layer_0_ln_1_weight" [id=1, type=get_attr];
+"2 encoder_layers_encoder_layer_0_ln_1_bias" [id=2, type=get_attr];
+"3 encoder_layers_encoder_layer_0_ln_2_weight" [id=3, type=get_attr];
+"4 encoder_layers_encoder_layer_0_ln_2_bias" [id=4, type=get_attr];
+"5 encoder_layers_encoder_layer_1_ln_1_weight" [id=5, type=get_attr];
+"6 encoder_layers_encoder_layer_1_ln_1_bias" [id=6, type=get_attr];
+"7 encoder_layers_encoder_layer_1_ln_2_weight" [id=7, type=get_attr];
+"8 encoder_layers_encoder_layer_1_ln_2_bias" [id=8, type=get_attr];
+"9 encoder_layers_encoder_layer_2_ln_1_weight" [id=9, type=get_attr];
+"10 encoder_layers_encoder_layer_2_ln_1_bias" [id=10, type=get_attr];
+"11 encoder_layers_encoder_layer_2_ln_2_weight" [id=11, type=get_attr];
+"12 encoder_layers_encoder_layer_2_ln_2_bias" [id=12, type=get_attr];
+"13 encoder_layers_encoder_layer_3_ln_1_weight" [id=13, type=get_attr];
+"14 encoder_layers_encoder_layer_3_ln_1_bias" [id=14, type=get_attr];
+"15 encoder_layers_encoder_layer_3_ln_2_weight" [id=15, type=get_attr];
+"16 encoder_layers_encoder_layer_3_ln_2_bias" [id=16, type=get_attr];
+"17 encoder_layers_encoder_layer_4_ln_1_weight" [id=17, type=get_attr];
+"18 encoder_layers_encoder_layer_4_ln_1_bias" [id=18, type=get_attr];
+"19 encoder_layers_encoder_layer_4_ln_2_weight" [id=19, type=get_attr];
+"20 encoder_layers_encoder_layer_4_ln_2_bias" [id=20, type=get_attr];
+"21 encoder_layers_encoder_layer_5_ln_1_weight" [id=21, type=get_attr];
+"22 encoder_layers_encoder_layer_5_ln_1_bias" [id=22, type=get_attr];
+"23 encoder_layers_encoder_layer_5_ln_2_weight" [id=23, type=get_attr];
+"24 encoder_layers_encoder_layer_5_ln_2_bias" [id=24, type=get_attr];
+"25 encoder_layers_encoder_layer_6_ln_1_weight" [id=25, type=get_attr];
+"26 encoder_layers_encoder_layer_6_ln_1_bias" [id=26, type=get_attr];
+"27 encoder_layers_encoder_layer_6_ln_2_weight" [id=27, type=get_attr];
+"28 encoder_layers_encoder_layer_6_ln_2_bias" [id=28, type=get_attr];
+"29 encoder_layers_encoder_layer_7_ln_1_weight" [id=29, type=get_attr];
+"30 encoder_layers_encoder_layer_7_ln_1_bias" [id=30, type=get_attr];
+"31 encoder_layers_encoder_layer_7_ln_2_weight" [id=31, type=get_attr];
+"32 encoder_layers_encoder_layer_7_ln_2_bias" [id=32, type=get_attr];
+"33 encoder_layers_encoder_layer_8_ln_1_weight" [id=33, type=get_attr];
+"34 encoder_layers_encoder_layer_8_ln_1_bias" [id=34, type=get_attr];
+"35 encoder_layers_encoder_layer_8_ln_2_weight" [id=35, type=get_attr];
+"36 encoder_layers_encoder_layer_8_ln_2_bias" [id=36, type=get_attr];
+"37 encoder_layers_encoder_layer_9_ln_1_weight" [id=37, type=get_attr];
+"38 encoder_layers_encoder_layer_9_ln_1_bias" [id=38, type=get_attr];
+"39 encoder_layers_encoder_layer_9_ln_2_weight" [id=39, type=get_attr];
+"40 encoder_layers_encoder_layer_9_ln_2_bias" [id=40, type=get_attr];
+"41 encoder_layers_encoder_layer_10_ln_1_weight" [id=41, type=get_attr];
+"42 encoder_layers_encoder_layer_10_ln_1_bias" [id=42, type=get_attr];
+"43 encoder_layers_encoder_layer_10_ln_2_weight" [id=43, type=get_attr];
+"44 encoder_layers_encoder_layer_10_ln_2_bias" [id=44, type=get_attr];
+"45 encoder_layers_encoder_layer_11_ln_1_weight" [id=45, type=get_attr];
+"46 encoder_layers_encoder_layer_11_ln_1_bias" [id=46, type=get_attr];
+"47 encoder_layers_encoder_layer_11_ln_2_weight" [id=47, type=get_attr];
+"48 encoder_layers_encoder_layer_11_ln_2_bias" [id=48, type=get_attr];
+"49 encoder_ln_weight" [id=49, type=get_attr];
+"50 encoder_ln_bias" [id=50, type=get_attr];
+"51 x" [id=51, type=input];
+"52 quantize_per_tensor_default" [id=52, type=quantize_per_tensor];
+"53 dequantize_per_tensor_default" [id=53, type=dequantize_per_tensor];
+"54 conv2d_scale_0" [id=54, type=get_attr];
+"55 conv2d_zero_point_0" [id=55, type=get_attr];
+"56 compressed_weight_updated_constant0" [id=56, type=get_attr];
+"57 quantize_per_channel_default" [id=57, type=quantize_per_channel];
+"58 dequantize_per_channel_default" [id=58, type=dequantize_per_channel];
+"59 conv_proj_bias_0_0" [id=59, type=get_attr];
+"60 conv2d" [id=60, type=conv2d];
+"61 reshape" [id=61, type=reshape];
+"62 permute" [id=62, type=permute];
+"63 _frozen_param0" [id=63, type=get_attr];
+"64 cat" [id=64, type=cat];
+"65 add" [id=65, type=add];
+"66 dropout" [id=66, type=dropout];
+"67 layer_norm" [id=67, type=layer_norm];
+"68 transpose" [id=68, type=transpose];
+"69 transpose_0_0_nncf_smooth_quant_0" [id=69, type=call_module];
+"70 quantize_per_tensor_default_1" [id=70, type=quantize_per_tensor];
+"71 dequantize_per_tensor_default_1" [id=71, type=dequantize_per_tensor];
+"72 linear_scale_0" [id=72, type=get_attr];
+"73 linear_zero_point_0" [id=73, type=get_attr];
+"74 compressed_weight_updated_constant1" [id=74, type=get_attr];
+"75 quantize_per_channel_default_1" [id=75, type=quantize_per_channel];
+"76 dequantize_per_channel_default_1" [id=76, type=dequantize_per_channel];
+"77 encoder_layers_encoder_layer_0_self_attention_in_proj_bias_0_0" [id=77, type=get_attr];
+"78 linear" [id=78, type=linear];
+"79 unflatten" [id=79, type=unflatten];
+"80 unsqueeze" [id=80, type=unsqueeze];
+"81 transpose_1" [id=81, type=transpose];
+"82 squeeze" [id=82, type=squeeze];
+"83 contiguous" [id=83, type=contiguous];
+"84 quantize_per_tensor_default_2" [id=84, type=quantize_per_tensor];
+"85 dequantize_per_tensor_default_2" [id=85, type=dequantize_per_tensor];
+"86 select" [id=86, type=select];
+"87 quantize_per_tensor_default_3" [id=87, type=quantize_per_tensor];
+"88 dequantize_per_tensor_default_3" [id=88, type=dequantize_per_tensor];
+"89 select_1" [id=89, type=select];
+"90 select_2" [id=90, type=select];
+"91 view" [id=91, type=view];
+"92 transpose_2" [id=92, type=transpose];
+"93 view_1" [id=93, type=view];
+"94 transpose_3" [id=94, type=transpose];
+"95 view_2" [id=95, type=view];
+"96 transpose_4" [id=96, type=transpose];
+"97 view_3" [id=97, type=view];
+"98 view_4" [id=98, type=view];
+"99 view_5" [id=99, type=view];
+"100 scaled_dot_product_attention" [id=100, type=scaled_dot_product_attention];
+"101 permute_1" [id=101, type=permute];
+"102 view_6" [id=102, type=view];
+"103 view_6_0_0_nncf_smooth_quant_0" [id=103, type=call_module];
+"104 quantize_per_tensor_default_4" [id=104, type=quantize_per_tensor];
+"105 dequantize_per_tensor_default_4" [id=105, type=dequantize_per_tensor];
+"106 linear_1_scale_0" [id=106, type=get_attr];
+"107 linear_1_zero_point_0" [id=107, type=get_attr];
+"108 compressed_weight_updated_constant2" [id=108, type=get_attr];
+"109 quantize_per_channel_default_2" [id=109, type=quantize_per_channel];
+"110 dequantize_per_channel_default_2" [id=110, type=dequantize_per_channel];
+"111 encoder_layers_encoder_layer_0_self_attention_out_proj_bias_0_0" [id=111, type=get_attr];
+"112 linear_1" [id=112, type=linear];
+"113 view_7" [id=113, type=view];
+"114 transpose_5" [id=114, type=transpose];
+"115 dropout_1" [id=115, type=dropout];
+"116 add_1" [id=116, type=add];
+"117 layer_norm_1" [id=117, type=layer_norm];
+"118 layer_norm_1_0_0_nncf_smooth_quant_0" [id=118, type=call_module];
+"119 quantize_per_tensor_default_5" [id=119, type=quantize_per_tensor];
+"120 dequantize_per_tensor_default_5" [id=120, type=dequantize_per_tensor];
+"121 linear_2_scale_0" [id=121, type=get_attr];
+"122 linear_2_zero_point_0" [id=122, type=get_attr];
+"123 compressed_weight_updated_constant3" [id=123, type=get_attr];
+"124 quantize_per_channel_default_3" [id=124, type=quantize_per_channel];
+"125 dequantize_per_channel_default_3" [id=125, type=dequantize_per_channel];
+"126 encoder_layers_encoder_layer_0_mlp_0_bias_0_0" [id=126, type=get_attr];
+"127 linear_2" [id=127, type=linear];
+"128 gelu" [id=128, type=gelu];
+"129 dropout_2" [id=129, type=dropout];
+"130 dropout_2_0_0_nncf_smooth_quant_0" [id=130, type=call_module];
+"131 quantize_per_tensor_default_6" [id=131, type=quantize_per_tensor];
+"132 dequantize_per_tensor_default_6" [id=132, type=dequantize_per_tensor];
+"133 linear_3_scale_0" [id=133, type=get_attr];
+"134 linear_3_zero_point_0" [id=134, type=get_attr];
+"135 compressed_weight_updated_constant4" [id=135, type=get_attr];
+"136 quantize_per_channel_default_4" [id=136, type=quantize_per_channel];
+"137 dequantize_per_channel_default_4" [id=137, type=dequantize_per_channel];
+"138 encoder_layers_encoder_layer_0_mlp_3_bias_0_0" [id=138, type=get_attr];
+"139 linear_3" [id=139, type=linear];
+"140 dropout_3" [id=140, type=dropout];
+"141 add_2" [id=141, type=add];
+"142 layer_norm_2" [id=142, type=layer_norm];
+"143 transpose_6" [id=143, type=transpose];
+"144 transpose_6_0_0_nncf_smooth_quant_0" [id=144, type=call_module];
+"145 quantize_per_tensor_default_7" [id=145, type=quantize_per_tensor];
+"146 dequantize_per_tensor_default_7" [id=146, type=dequantize_per_tensor];
+"147 linear_4_scale_0" [id=147, type=get_attr];
+"148 linear_4_zero_point_0" [id=148, type=get_attr];
+"149 compressed_weight_updated_constant5" [id=149, type=get_attr];
+"150 quantize_per_channel_default_5" [id=150, type=quantize_per_channel];
+"151 dequantize_per_channel_default_5" [id=151, type=dequantize_per_channel];
+"152 encoder_layers_encoder_layer_1_self_attention_in_proj_bias_0_0" [id=152, type=get_attr];
+"153 linear_4" [id=153, type=linear];
+"154 unflatten_1" [id=154, type=unflatten];
+"155 unsqueeze_1" [id=155, type=unsqueeze];
+"156 transpose_7" [id=156, type=transpose];
+"157 squeeze_1" [id=157, type=squeeze];
+"158 contiguous_1" [id=158, type=contiguous];
+"159 quantize_per_tensor_default_8" [id=159, type=quantize_per_tensor];
+"160 dequantize_per_tensor_default_8" [id=160, type=dequantize_per_tensor];
+"161 select_3" [id=161, type=select];
+"162 quantize_per_tensor_default_9" [id=162, type=quantize_per_tensor];
+"163 dequantize_per_tensor_default_9" [id=163, type=dequantize_per_tensor];
+"164 select_4" [id=164, type=select];
+"165 select_5" [id=165, type=select];
+"166 view_8" [id=166, type=view];
+"167 transpose_8" [id=167, type=transpose];
+"168 view_9" [id=168, type=view];
+"169 transpose_9" [id=169, type=transpose];
+"170 view_10" [id=170, type=view];
+"171 transpose_10" [id=171, type=transpose];
+"172 view_11" [id=172, type=view];
+"173 view_12" [id=173, type=view];
+"174 view_13" [id=174, type=view];
+"175 scaled_dot_product_attention_1" [id=175, type=scaled_dot_product_attention];
+"176 permute_2" [id=176, type=permute];
+"177 view_14" [id=177, type=view];
+"178 view_14_0_0_nncf_smooth_quant_0" [id=178, type=call_module];
+"179 quantize_per_tensor_default_10" [id=179, type=quantize_per_tensor];
+"180 dequantize_per_tensor_default_10" [id=180, type=dequantize_per_tensor];
+"181 linear_5_scale_0" [id=181, type=get_attr];
+"182 linear_5_zero_point_0" [id=182, type=get_attr];
+"183 compressed_weight_updated_constant6" [id=183, type=get_attr];
+"184 quantize_per_channel_default_6" [id=184, type=quantize_per_channel];
+"185 dequantize_per_channel_default_6" [id=185, type=dequantize_per_channel];
+"186 encoder_layers_encoder_layer_1_self_attention_out_proj_bias_0_0" [id=186, type=get_attr];
+"187 linear_5" [id=187, type=linear];
+"188 view_15" [id=188, type=view];
+"189 transpose_11" [id=189, type=transpose];
+"190 dropout_4" [id=190, type=dropout];
+"191 add_3" [id=191, type=add];
+"192 layer_norm_3" [id=192, type=layer_norm];
+"193 layer_norm_3_0_0_nncf_smooth_quant_0" [id=193, type=call_module];
+"194 quantize_per_tensor_default_11" [id=194, type=quantize_per_tensor];
+"195 dequantize_per_tensor_default_11" [id=195, type=dequantize_per_tensor];
+"196 linear_6_scale_0" [id=196, type=get_attr];
+"197 linear_6_zero_point_0" [id=197, type=get_attr];
+"198 compressed_weight_updated_constant7" [id=198, type=get_attr];
+"199 quantize_per_channel_default_7" [id=199, type=quantize_per_channel];
+"200 dequantize_per_channel_default_7" [id=200, type=dequantize_per_channel];
+"201 encoder_layers_encoder_layer_1_mlp_0_bias_0_0" [id=201, type=get_attr];
+"202 linear_6" [id=202, type=linear];
+"203 gelu_1" [id=203, type=gelu];
+"204 dropout_5" [id=204, type=dropout];
+"205 dropout_5_0_0_nncf_smooth_quant_0" [id=205, type=call_module];
+"206 quantize_per_tensor_default_12" [id=206, type=quantize_per_tensor];
+"207 dequantize_per_tensor_default_12" [id=207, type=dequantize_per_tensor];
+"208 linear_7_scale_0" [id=208, type=get_attr];
+"209 linear_7_zero_point_0" [id=209, type=get_attr];
+"210 compressed_weight_updated_constant8" [id=210, type=get_attr];
+"211 quantize_per_channel_default_8" [id=211, type=quantize_per_channel];
+"212 dequantize_per_channel_default_8" [id=212, type=dequantize_per_channel];
+"213 encoder_layers_encoder_layer_1_mlp_3_bias_0_0" [id=213, type=get_attr];
+"214 linear_7" [id=214, type=linear];
+"215 dropout_6" [id=215, type=dropout];
+"216 add_4" [id=216, type=add];
+"217 layer_norm_4" [id=217, type=layer_norm];
+"218 transpose_12" [id=218, type=transpose];
+"219 transpose_12_0_0_nncf_smooth_quant_0" [id=219, type=call_module];
+"220 quantize_per_tensor_default_13" [id=220, type=quantize_per_tensor];
+"221 dequantize_per_tensor_default_13" [id=221, type=dequantize_per_tensor];
+"222 linear_8_scale_0" [id=222, type=get_attr];
+"223 linear_8_zero_point_0" [id=223, type=get_attr];
+"224 compressed_weight_updated_constant9" [id=224, type=get_attr];
+"225 quantize_per_channel_default_9" [id=225, type=quantize_per_channel];
+"226 dequantize_per_channel_default_9" [id=226, type=dequantize_per_channel];
+"227 encoder_layers_encoder_layer_2_self_attention_in_proj_bias_0_0" [id=227, type=get_attr];
+"228 linear_8" [id=228, type=linear];
+"229 unflatten_2" [id=229, type=unflatten];
+"230 unsqueeze_2" [id=230, type=unsqueeze];
+"231 transpose_13" [id=231, type=transpose];
+"232 squeeze_2" [id=232, type=squeeze];
+"233 contiguous_2" [id=233, type=contiguous];
+"234 quantize_per_tensor_default_14" [id=234, type=quantize_per_tensor];
+"235 dequantize_per_tensor_default_14" [id=235, type=dequantize_per_tensor];
+"236 select_6" [id=236, type=select];
+"237 quantize_per_tensor_default_15" [id=237, type=quantize_per_tensor];
+"238 dequantize_per_tensor_default_15" [id=238, type=dequantize_per_tensor];
+"239 select_7" [id=239, type=select];
+"240 select_8" [id=240, type=select];
+"241 view_16" [id=241, type=view];
+"242 transpose_14" [id=242, type=transpose];
+"243 view_17" [id=243, type=view];
+"244 transpose_15" [id=244, type=transpose];
+"245 view_18" [id=245, type=view];
+"246 transpose_16" [id=246, type=transpose];
+"247 view_19" [id=247, type=view];
+"248 view_20" [id=248, type=view];
+"249 view_21" [id=249, type=view];
+"250 scaled_dot_product_attention_2" [id=250, type=scaled_dot_product_attention];
+"251 permute_3" [id=251, type=permute];
+"252 view_22" [id=252, type=view];
+"253 view_22_0_0_nncf_smooth_quant_0" [id=253, type=call_module];
+"254 quantize_per_tensor_default_16" [id=254, type=quantize_per_tensor];
+"255 dequantize_per_tensor_default_16" [id=255, type=dequantize_per_tensor];
+"256 linear_9_scale_0" [id=256, type=get_attr];
+"257 linear_9_zero_point_0" [id=257, type=get_attr];
+"258 compressed_weight_updated_constant10" [id=258, type=get_attr];
+"259 quantize_per_channel_default_10" [id=259, type=quantize_per_channel];
+"260 dequantize_per_channel_default_10" [id=260, type=dequantize_per_channel];
+"261 encoder_layers_encoder_layer_2_self_attention_out_proj_bias_0_0" [id=261, type=get_attr];
+"262 linear_9" [id=262, type=linear];
+"263 view_23" [id=263, type=view];
+"264 transpose_17" [id=264, type=transpose];
+"265 dropout_7" [id=265, type=dropout];
+"266 add_5" [id=266, type=add];
+"267 layer_norm_5" [id=267, type=layer_norm];
+"268 layer_norm_5_0_0_nncf_smooth_quant_0" [id=268, type=call_module];
+"269 quantize_per_tensor_default_17" [id=269, type=quantize_per_tensor];
+"270 dequantize_per_tensor_default_17" [id=270, type=dequantize_per_tensor];
+"271 linear_10_scale_0" [id=271, type=get_attr];
+"272 linear_10_zero_point_0" [id=272, type=get_attr];
+"273 compressed_weight_updated_constant11" [id=273, type=get_attr];
+"274 quantize_per_channel_default_11" [id=274, type=quantize_per_channel];
+"275 dequantize_per_channel_default_11" [id=275, type=dequantize_per_channel];
+"276 encoder_layers_encoder_layer_2_mlp_0_bias_0_0" [id=276, type=get_attr];
+"277 linear_10" [id=277, type=linear];
+"278 gelu_2" [id=278, type=gelu];
+"279 dropout_8" [id=279, type=dropout];
+"280 dropout_8_0_0_nncf_smooth_quant_0" [id=280, type=call_module];
+"281 quantize_per_tensor_default_18" [id=281, type=quantize_per_tensor];
+"282 dequantize_per_tensor_default_18" [id=282, type=dequantize_per_tensor];
+"283 linear_11_scale_0" [id=283, type=get_attr];
+"284 linear_11_zero_point_0" [id=284, type=get_attr];
+"285 compressed_weight_updated_constant12" [id=285, type=get_attr];
+"286 quantize_per_channel_default_12" [id=286, type=quantize_per_channel];
+"287 dequantize_per_channel_default_12" [id=287, type=dequantize_per_channel];
+"288 encoder_layers_encoder_layer_2_mlp_3_bias_0_0" [id=288, type=get_attr];
+"289 linear_11" [id=289, type=linear];
+"290 dropout_9" [id=290, type=dropout];
+"291 add_6" [id=291, type=add];
+"292 layer_norm_6" [id=292, type=layer_norm];
+"293 transpose_18" [id=293, type=transpose];
+"294 transpose_18_0_0_nncf_smooth_quant_0" [id=294, type=call_module];
+"295 quantize_per_tensor_default_19" [id=295, type=quantize_per_tensor];
+"296 dequantize_per_tensor_default_19" [id=296, type=dequantize_per_tensor];
+"297 linear_12_scale_0" [id=297, type=get_attr];
+"298 linear_12_zero_point_0" [id=298, type=get_attr];
+"299 compressed_weight_updated_constant13" [id=299, type=get_attr];
+"300 quantize_per_channel_default_13" [id=300, type=quantize_per_channel];
+"301 dequantize_per_channel_default_13" [id=301, type=dequantize_per_channel];
+"302 encoder_layers_encoder_layer_3_self_attention_in_proj_bias_0_0" [id=302, type=get_attr];
+"303 linear_12" [id=303, type=linear];
+"304 unflatten_3" [id=304, type=unflatten];
+"305 unsqueeze_3" [id=305, type=unsqueeze];
+"306 transpose_19" [id=306, type=transpose];
+"307 squeeze_3" [id=307, type=squeeze];
+"308 contiguous_3" [id=308, type=contiguous];
+"309 quantize_per_tensor_default_20" [id=309, type=quantize_per_tensor];
+"310 dequantize_per_tensor_default_20" [id=310, type=dequantize_per_tensor];
+"311 select_9" [id=311, type=select];
+"312 quantize_per_tensor_default_21" [id=312, type=quantize_per_tensor];
+"313 dequantize_per_tensor_default_21" [id=313, type=dequantize_per_tensor];
+"314 select_10" [id=314, type=select];
+"315 select_11" [id=315, type=select];
+"316 view_24" [id=316, type=view];
+"317 transpose_20" [id=317, type=transpose];
+"318 view_25" [id=318, type=view];
+"319 transpose_21" [id=319, type=transpose];
+"320 view_26" [id=320, type=view];
+"321 transpose_22" [id=321, type=transpose];
+"322 view_27" [id=322, type=view];
+"323 view_28" [id=323, type=view];
+"324 view_29" [id=324, type=view];
+"325 scaled_dot_product_attention_3" [id=325, type=scaled_dot_product_attention];
+"326 permute_4" [id=326, type=permute];
+"327 view_30" [id=327, type=view];
+"328 view_30_0_0_nncf_smooth_quant_0" [id=328, type=call_module];
+"329 quantize_per_tensor_default_22" [id=329, type=quantize_per_tensor];
+"330 dequantize_per_tensor_default_22" [id=330, type=dequantize_per_tensor];
+"331 linear_13_scale_0" [id=331, type=get_attr];
+"332 linear_13_zero_point_0" [id=332, type=get_attr];
+"333 compressed_weight_updated_constant14" [id=333, type=get_attr];
+"334 quantize_per_channel_default_14" [id=334, type=quantize_per_channel];
+"335 dequantize_per_channel_default_14" [id=335, type=dequantize_per_channel];
+"336 encoder_layers_encoder_layer_3_self_attention_out_proj_bias_0_0" [id=336, type=get_attr];
+"337 linear_13" [id=337, type=linear];
+"338 view_31" [id=338, type=view];
+"339 transpose_23" [id=339, type=transpose];
+"340 dropout_10" [id=340, type=dropout];
+"341 add_7" [id=341, type=add];
+"342 layer_norm_7" [id=342, type=layer_norm];
+"343 layer_norm_7_0_0_nncf_smooth_quant_0" [id=343, type=call_module];
+"344 quantize_per_tensor_default_23" [id=344, type=quantize_per_tensor];
+"345 dequantize_per_tensor_default_23" [id=345, type=dequantize_per_tensor];
+"346 linear_14_scale_0" [id=346, type=get_attr];
+"347 linear_14_zero_point_0" [id=347, type=get_attr];
+"348 compressed_weight_updated_constant15" [id=348, type=get_attr];
+"349 quantize_per_channel_default_15" [id=349, type=quantize_per_channel];
+"350 dequantize_per_channel_default_15" [id=350, type=dequantize_per_channel];
+"351 encoder_layers_encoder_layer_3_mlp_0_bias_0_0" [id=351, type=get_attr];
+"352 linear_14" [id=352, type=linear];
+"353 gelu_3" [id=353, type=gelu];
+"354 dropout_11" [id=354, type=dropout];
+"355 dropout_11_0_0_nncf_smooth_quant_0" [id=355, type=call_module];
+"356 quantize_per_tensor_default_24" [id=356, type=quantize_per_tensor];
+"357 dequantize_per_tensor_default_24" [id=357, type=dequantize_per_tensor];
+"358 linear_15_scale_0" [id=358, type=get_attr];
+"359 linear_15_zero_point_0" [id=359, type=get_attr];
+"360 compressed_weight_updated_constant16" [id=360, type=get_attr];
+"361 quantize_per_channel_default_16" [id=361, type=quantize_per_channel];
+"362 dequantize_per_channel_default_16" [id=362, type=dequantize_per_channel];
+"363 encoder_layers_encoder_layer_3_mlp_3_bias_0_0" [id=363, type=get_attr];
+"364 linear_15" [id=364, type=linear];
+"365 dropout_12" [id=365, type=dropout];
+"366 add_8" [id=366, type=add];
+"367 layer_norm_8" [id=367, type=layer_norm];
+"368 transpose_24" [id=368, type=transpose];
+"369 transpose_24_0_0_nncf_smooth_quant_0" [id=369, type=call_module];
+"370 quantize_per_tensor_default_25" [id=370, type=quantize_per_tensor];
+"371 dequantize_per_tensor_default_25" [id=371, type=dequantize_per_tensor];
+"372 linear_16_scale_0" [id=372, type=get_attr];
+"373 linear_16_zero_point_0" [id=373, type=get_attr];
+"374 compressed_weight_updated_constant17" [id=374, type=get_attr];
+"375 quantize_per_channel_default_17" [id=375, type=quantize_per_channel];
+"376 dequantize_per_channel_default_17" [id=376, type=dequantize_per_channel];
+"377 encoder_layers_encoder_layer_4_self_attention_in_proj_bias_0_0" [id=377, type=get_attr];
+"378 linear_16" [id=378, type=linear];
+"379 unflatten_4" [id=379, type=unflatten];
+"380 unsqueeze_4" [id=380, type=unsqueeze];
+"381 transpose_25" [id=381, type=transpose];
+"382 squeeze_4" [id=382, type=squeeze];
+"383 contiguous_4" [id=383, type=contiguous];
+"384 quantize_per_tensor_default_26" [id=384, type=quantize_per_tensor];
+"385 dequantize_per_tensor_default_26" [id=385, type=dequantize_per_tensor];
+"386 select_12" [id=386, type=select];
+"387 quantize_per_tensor_default_27" [id=387, type=quantize_per_tensor];
+"388 dequantize_per_tensor_default_27" [id=388, type=dequantize_per_tensor];
+"389 select_13" [id=389, type=select];
+"390 select_14" [id=390, type=select];
+"391 view_32" [id=391, type=view];
+"392 transpose_26" [id=392, type=transpose];
+"393 view_33" [id=393, type=view];
+"394 transpose_27" [id=394, type=transpose];
+"395 view_34" [id=395, type=view];
+"396 transpose_28" [id=396, type=transpose];
+"397 view_35" [id=397, type=view];
+"398 view_36" [id=398, type=view];
+"399 view_37" [id=399, type=view];
+"400 scaled_dot_product_attention_4" [id=400, type=scaled_dot_product_attention];
+"401 permute_5" [id=401, type=permute];
+"402 view_38" [id=402, type=view];
+"403 view_38_0_0_nncf_smooth_quant_0" [id=403, type=call_module];
+"404 quantize_per_tensor_default_28" [id=404, type=quantize_per_tensor];
+"405 dequantize_per_tensor_default_28" [id=405, type=dequantize_per_tensor];
+"406 linear_17_scale_0" [id=406, type=get_attr];
+"407 linear_17_zero_point_0" [id=407, type=get_attr];
+"408 compressed_weight_updated_constant18" [id=408, type=get_attr];
+"409 quantize_per_channel_default_18" [id=409, type=quantize_per_channel];
+"410 dequantize_per_channel_default_18" [id=410, type=dequantize_per_channel];
+"411 encoder_layers_encoder_layer_4_self_attention_out_proj_bias_0_0" [id=411, type=get_attr];
+"412 linear_17" [id=412, type=linear];
+"413 view_39" [id=413, type=view];
+"414 transpose_29" [id=414, type=transpose];
+"415 dropout_13" [id=415, type=dropout];
+"416 add_9" [id=416, type=add];
+"417 layer_norm_9" [id=417, type=layer_norm];
+"418 layer_norm_9_0_0_nncf_smooth_quant_0" [id=418, type=call_module];
+"419 quantize_per_tensor_default_29" [id=419, type=quantize_per_tensor];
+"420 dequantize_per_tensor_default_29" [id=420, type=dequantize_per_tensor];
+"421 linear_18_scale_0" [id=421, type=get_attr];
+"422 linear_18_zero_point_0" [id=422, type=get_attr];
+"423 compressed_weight_updated_constant19" [id=423, type=get_attr];
+"424 quantize_per_channel_default_19" [id=424, type=quantize_per_channel];
+"425 dequantize_per_channel_default_19" [id=425, type=dequantize_per_channel];
+"426 encoder_layers_encoder_layer_4_mlp_0_bias_0_0" [id=426, type=get_attr];
+"427 linear_18" [id=427, type=linear];
+"428 gelu_4" [id=428, type=gelu];
+"429 dropout_14" [id=429, type=dropout];
+"430 dropout_14_0_0_nncf_smooth_quant_0" [id=430, type=call_module];
+"431 quantize_per_tensor_default_30" [id=431, type=quantize_per_tensor];
+"432 dequantize_per_tensor_default_30" [id=432, type=dequantize_per_tensor];
+"433 linear_19_scale_0" [id=433, type=get_attr];
+"434 linear_19_zero_point_0" [id=434, type=get_attr];
+"435 compressed_weight_updated_constant20" [id=435, type=get_attr];
+"436 quantize_per_channel_default_20" [id=436, type=quantize_per_channel];
+"437 dequantize_per_channel_default_20" [id=437, type=dequantize_per_channel];
+"438 encoder_layers_encoder_layer_4_mlp_3_bias_0_0" [id=438, type=get_attr];
+"439 linear_19" [id=439, type=linear];
+"440 dropout_15" [id=440, type=dropout];
+"441 add_10" [id=441, type=add];
+"442 layer_norm_10" [id=442, type=layer_norm];
+"443 transpose_30" [id=443, type=transpose];
+"444 transpose_30_0_0_nncf_smooth_quant_0" [id=444, type=call_module];
+"445 quantize_per_tensor_default_31" [id=445, type=quantize_per_tensor];
+"446 dequantize_per_tensor_default_31" [id=446, type=dequantize_per_tensor];
+"447 linear_20_scale_0" [id=447, type=get_attr];
+"448 linear_20_zero_point_0" [id=448, type=get_attr];
+"449 compressed_weight_updated_constant21" [id=449, type=get_attr];
+"450 quantize_per_channel_default_21" [id=450, type=quantize_per_channel];
+"451 dequantize_per_channel_default_21" [id=451, type=dequantize_per_channel];
+"452 encoder_layers_encoder_layer_5_self_attention_in_proj_bias_0_0" [id=452, type=get_attr];
+"453 linear_20" [id=453, type=linear];
+"454 unflatten_5" [id=454, type=unflatten];
+"455 unsqueeze_5" [id=455, type=unsqueeze];
+"456 transpose_31" [id=456, type=transpose];
+"457 squeeze_5" [id=457, type=squeeze];
+"458 contiguous_5" [id=458, type=contiguous];
+"459 quantize_per_tensor_default_32" [id=459, type=quantize_per_tensor];
+"460 dequantize_per_tensor_default_32" [id=460, type=dequantize_per_tensor];
+"461 select_15" [id=461, type=select];
+"462 quantize_per_tensor_default_33" [id=462, type=quantize_per_tensor];
+"463 dequantize_per_tensor_default_33" [id=463, type=dequantize_per_tensor];
+"464 select_16" [id=464, type=select];
+"465 select_17" [id=465, type=select];
+"466 view_40" [id=466, type=view];
+"467 transpose_32" [id=467, type=transpose];
+"468 view_41" [id=468, type=view];
+"469 transpose_33" [id=469, type=transpose];
+"470 view_42" [id=470, type=view];
+"471 transpose_34" [id=471, type=transpose];
+"472 view_43" [id=472, type=view];
+"473 view_44" [id=473, type=view];
+"474 view_45" [id=474, type=view];
+"475 scaled_dot_product_attention_5" [id=475, type=scaled_dot_product_attention];
+"476 permute_6" [id=476, type=permute];
+"477 view_46" [id=477, type=view];
+"478 view_46_0_0_nncf_smooth_quant_0" [id=478, type=call_module];
+"479 quantize_per_tensor_default_34" [id=479, type=quantize_per_tensor];
+"480 dequantize_per_tensor_default_34" [id=480, type=dequantize_per_tensor];
+"481 linear_21_scale_0" [id=481, type=get_attr];
+"482 linear_21_zero_point_0" [id=482, type=get_attr];
+"483 compressed_weight_updated_constant22" [id=483, type=get_attr];
+"484 quantize_per_channel_default_22" [id=484, type=quantize_per_channel];
+"485 dequantize_per_channel_default_22" [id=485, type=dequantize_per_channel];
+"486 encoder_layers_encoder_layer_5_self_attention_out_proj_bias_0_0" [id=486, type=get_attr];
+"487 linear_21" [id=487, type=linear];
+"488 view_47" [id=488, type=view];
+"489 transpose_35" [id=489, type=transpose];
+"490 dropout_16" [id=490, type=dropout];
+"491 add_11" [id=491, type=add];
+"492 layer_norm_11" [id=492, type=layer_norm];
+"493 layer_norm_11_0_0_nncf_smooth_quant_0" [id=493, type=call_module];
+"494 quantize_per_tensor_default_35" [id=494, type=quantize_per_tensor];
+"495 dequantize_per_tensor_default_35" [id=495, type=dequantize_per_tensor];
+"496 linear_22_scale_0" [id=496, type=get_attr];
+"497 linear_22_zero_point_0" [id=497, type=get_attr];
+"498 compressed_weight_updated_constant23" [id=498, type=get_attr];
+"499 quantize_per_channel_default_23" [id=499, type=quantize_per_channel];
+"500 dequantize_per_channel_default_23" [id=500, type=dequantize_per_channel];
+"501 encoder_layers_encoder_layer_5_mlp_0_bias_0_0" [id=501, type=get_attr];
+"502 linear_22" [id=502, type=linear];
+"503 gelu_5" [id=503, type=gelu];
+"504 dropout_17" [id=504, type=dropout];
+"505 dropout_17_0_0_nncf_smooth_quant_0" [id=505, type=call_module];
+"506 quantize_per_tensor_default_36" [id=506, type=quantize_per_tensor];
+"507 dequantize_per_tensor_default_36" [id=507, type=dequantize_per_tensor];
+"508 linear_23_scale_0" [id=508, type=get_attr];
+"509 linear_23_zero_point_0" [id=509, type=get_attr];
+"510 compressed_weight_updated_constant24" [id=510, type=get_attr];
+"511 quantize_per_channel_default_24" [id=511, type=quantize_per_channel];
+"512 dequantize_per_channel_default_24" [id=512, type=dequantize_per_channel];
+"513 encoder_layers_encoder_layer_5_mlp_3_bias_0_0" [id=513, type=get_attr];
+"514 linear_23" [id=514, type=linear];
+"515 dropout_18" [id=515, type=dropout];
+"516 add_12" [id=516, type=add];
+"517 layer_norm_12" [id=517, type=layer_norm];
+"518 transpose_36" [id=518, type=transpose];
+"519 transpose_36_0_0_nncf_smooth_quant_0" [id=519, type=call_module];
+"520 quantize_per_tensor_default_37" [id=520, type=quantize_per_tensor];
+"521 dequantize_per_tensor_default_37" [id=521, type=dequantize_per_tensor];
+"522 linear_24_scale_0" [id=522, type=get_attr];
+"523 linear_24_zero_point_0" [id=523, type=get_attr];
+"524 compressed_weight_updated_constant25" [id=524, type=get_attr];
+"525 quantize_per_channel_default_25" [id=525, type=quantize_per_channel];
+"526 dequantize_per_channel_default_25" [id=526, type=dequantize_per_channel];
+"527 encoder_layers_encoder_layer_6_self_attention_in_proj_bias_0_0" [id=527, type=get_attr];
+"528 linear_24" [id=528, type=linear];
+"529 unflatten_6" [id=529, type=unflatten];
+"530 unsqueeze_6" [id=530, type=unsqueeze];
+"531 transpose_37" [id=531, type=transpose];
+"532 squeeze_6" [id=532, type=squeeze];
+"533 contiguous_6" [id=533, type=contiguous];
+"534 quantize_per_tensor_default_38" [id=534, type=quantize_per_tensor];
+"535 dequantize_per_tensor_default_38" [id=535, type=dequantize_per_tensor];
+"536 select_18" [id=536, type=select];
+"537 quantize_per_tensor_default_39" [id=537, type=quantize_per_tensor];
+"538 dequantize_per_tensor_default_39" [id=538, type=dequantize_per_tensor];
+"539 select_19" [id=539, type=select];
+"540 select_20" [id=540, type=select];
+"541 view_48" [id=541, type=view];
+"542 transpose_38" [id=542, type=transpose];
+"543 view_49" [id=543, type=view];
+"544 transpose_39" [id=544, type=transpose];
+"545 view_50" [id=545, type=view];
+"546 transpose_40" [id=546, type=transpose];
+"547 view_51" [id=547, type=view];
+"548 view_52" [id=548, type=view];
+"549 view_53" [id=549, type=view];
+"550 scaled_dot_product_attention_6" [id=550, type=scaled_dot_product_attention];
+"551 permute_7" [id=551, type=permute];
+"552 view_54" [id=552, type=view];
+"553 view_54_0_0_nncf_smooth_quant_0" [id=553, type=call_module];
+"554 quantize_per_tensor_default_40" [id=554, type=quantize_per_tensor];
+"555 dequantize_per_tensor_default_40" [id=555, type=dequantize_per_tensor];
+"556 linear_25_scale_0" [id=556, type=get_attr];
+"557 linear_25_zero_point_0" [id=557, type=get_attr];
+"558 compressed_weight_updated_constant26" [id=558, type=get_attr];
+"559 quantize_per_channel_default_26" [id=559, type=quantize_per_channel];
+"560 dequantize_per_channel_default_26" [id=560, type=dequantize_per_channel];
+"561 encoder_layers_encoder_layer_6_self_attention_out_proj_bias_0_0" [id=561, type=get_attr];
+"562 linear_25" [id=562, type=linear];
+"563 view_55" [id=563, type=view];
+"564 transpose_41" [id=564, type=transpose];
+"565 dropout_19" [id=565, type=dropout];
+"566 add_13" [id=566, type=add];
+"567 layer_norm_13" [id=567, type=layer_norm];
+"568 layer_norm_13_0_0_nncf_smooth_quant_0" [id=568, type=call_module];
+"569 quantize_per_tensor_default_41" [id=569, type=quantize_per_tensor];
+"570 dequantize_per_tensor_default_41" [id=570, type=dequantize_per_tensor];
+"571 linear_26_scale_0" [id=571, type=get_attr];
+"572 linear_26_zero_point_0" [id=572, type=get_attr];
+"573 compressed_weight_updated_constant27" [id=573, type=get_attr];
+"574 quantize_per_channel_default_27" [id=574, type=quantize_per_channel];
+"575 dequantize_per_channel_default_27" [id=575, type=dequantize_per_channel];
+"576 encoder_layers_encoder_layer_6_mlp_0_bias_0_0" [id=576, type=get_attr];
+"577 linear_26" [id=577, type=linear];
+"578 gelu_6" [id=578, type=gelu];
+"579 dropout_20" [id=579, type=dropout];
+"580 dropout_20_0_0_nncf_smooth_quant_0" [id=580, type=call_module];
+"581 quantize_per_tensor_default_42" [id=581, type=quantize_per_tensor];
+"582 dequantize_per_tensor_default_42" [id=582, type=dequantize_per_tensor];
+"583 linear_27_scale_0" [id=583, type=get_attr];
+"584 linear_27_zero_point_0" [id=584, type=get_attr];
+"585 compressed_weight_updated_constant28" [id=585, type=get_attr];
+"586 quantize_per_channel_default_28" [id=586, type=quantize_per_channel];
+"587 dequantize_per_channel_default_28" [id=587, type=dequantize_per_channel];
+"588 encoder_layers_encoder_layer_6_mlp_3_bias_0_0" [id=588, type=get_attr];
+"589 linear_27" [id=589, type=linear];
+"590 dropout_21" [id=590, type=dropout];
+"591 add_14" [id=591, type=add];
+"592 layer_norm_14" [id=592, type=layer_norm];
+"593 transpose_42" [id=593, type=transpose];
+"594 transpose_42_0_0_nncf_smooth_quant_0" [id=594, type=call_module];
+"595 quantize_per_tensor_default_43" [id=595, type=quantize_per_tensor];
+"596 dequantize_per_tensor_default_43" [id=596, type=dequantize_per_tensor];
+"597 linear_28_scale_0" [id=597, type=get_attr];
+"598 linear_28_zero_point_0" [id=598, type=get_attr];
+"599 compressed_weight_updated_constant29" [id=599, type=get_attr];
+"600 quantize_per_channel_default_29" [id=600, type=quantize_per_channel];
+"601 dequantize_per_channel_default_29" [id=601, type=dequantize_per_channel];
+"602 encoder_layers_encoder_layer_7_self_attention_in_proj_bias_0_0" [id=602, type=get_attr];
+"603 linear_28" [id=603, type=linear];
+"604 unflatten_7" [id=604, type=unflatten];
+"605 unsqueeze_7" [id=605, type=unsqueeze];
+"606 transpose_43" [id=606, type=transpose];
+"607 squeeze_7" [id=607, type=squeeze];
+"608 contiguous_7" [id=608, type=contiguous];
+"609 quantize_per_tensor_default_44" [id=609, type=quantize_per_tensor];
+"610 dequantize_per_tensor_default_44" [id=610, type=dequantize_per_tensor];
+"611 select_21" [id=611, type=select];
+"612 quantize_per_tensor_default_45" [id=612, type=quantize_per_tensor];
+"613 dequantize_per_tensor_default_45" [id=613, type=dequantize_per_tensor];
+"614 select_22" [id=614, type=select];
+"615 select_23" [id=615, type=select];
+"616 view_56" [id=616, type=view];
+"617 transpose_44" [id=617, type=transpose];
+"618 view_57" [id=618, type=view];
+"619 transpose_45" [id=619, type=transpose];
+"620 view_58" [id=620, type=view];
+"621 transpose_46" [id=621, type=transpose];
+"622 view_59" [id=622, type=view];
+"623 view_60" [id=623, type=view];
+"624 view_61" [id=624, type=view];
+"625 scaled_dot_product_attention_7" [id=625, type=scaled_dot_product_attention];
+"626 permute_8" [id=626, type=permute];
+"627 view_62" [id=627, type=view];
+"628 view_62_0_0_nncf_smooth_quant_0" [id=628, type=call_module];
+"629 quantize_per_tensor_default_46" [id=629, type=quantize_per_tensor];
+"630 dequantize_per_tensor_default_46" [id=630, type=dequantize_per_tensor];
+"631 linear_29_scale_0" [id=631, type=get_attr];
+"632 linear_29_zero_point_0" [id=632, type=get_attr];
+"633 compressed_weight_updated_constant30" [id=633, type=get_attr];
+"634 quantize_per_channel_default_30" [id=634, type=quantize_per_channel];
+"635 dequantize_per_channel_default_30" [id=635, type=dequantize_per_channel];
+"636 encoder_layers_encoder_layer_7_self_attention_out_proj_bias_0_0" [id=636, type=get_attr];
+"637 linear_29" [id=637, type=linear];
+"638 view_63" [id=638, type=view];
+"639 transpose_47" [id=639, type=transpose];
+"640 dropout_22" [id=640, type=dropout];
+"641 add_15" [id=641, type=add];
+"642 layer_norm_15" [id=642, type=layer_norm];
+"643 layer_norm_15_0_0_nncf_smooth_quant_0" [id=643, type=call_module];
+"644 quantize_per_tensor_default_47" [id=644, type=quantize_per_tensor];
+"645 dequantize_per_tensor_default_47" [id=645, type=dequantize_per_tensor];
+"646 linear_30_scale_0" [id=646, type=get_attr];
+"647 linear_30_zero_point_0" [id=647, type=get_attr];
+"648 compressed_weight_updated_constant31" [id=648, type=get_attr];
+"649 quantize_per_channel_default_31" [id=649, type=quantize_per_channel];
+"650 dequantize_per_channel_default_31" [id=650, type=dequantize_per_channel];
+"651 encoder_layers_encoder_layer_7_mlp_0_bias_0_0" [id=651, type=get_attr];
+"652 linear_30" [id=652, type=linear];
+"653 gelu_7" [id=653, type=gelu];
+"654 dropout_23" [id=654, type=dropout];
+"655 dropout_23_0_0_nncf_smooth_quant_0" [id=655, type=call_module];
+"656 quantize_per_tensor_default_48" [id=656, type=quantize_per_tensor];
+"657 dequantize_per_tensor_default_48" [id=657, type=dequantize_per_tensor];
+"658 linear_31_scale_0" [id=658, type=get_attr];
+"659 linear_31_zero_point_0" [id=659, type=get_attr];
+"660 compressed_weight_updated_constant32" [id=660, type=get_attr];
+"661 quantize_per_channel_default_32" [id=661, type=quantize_per_channel];
+"662 dequantize_per_channel_default_32" [id=662, type=dequantize_per_channel];
+"663 encoder_layers_encoder_layer_7_mlp_3_bias_0_0" [id=663, type=get_attr];
+"664 linear_31" [id=664, type=linear];
+"665 dropout_24" [id=665, type=dropout];
+"666 add_16" [id=666, type=add];
+"667 layer_norm_16" [id=667, type=layer_norm];
+"668 transpose_48" [id=668, type=transpose];
+"669 transpose_48_0_0_nncf_smooth_quant_0" [id=669, type=call_module];
+"670 quantize_per_tensor_default_49" [id=670, type=quantize_per_tensor];
+"671 dequantize_per_tensor_default_49" [id=671, type=dequantize_per_tensor];
+"672 linear_32_scale_0" [id=672, type=get_attr];
+"673 linear_32_zero_point_0" [id=673, type=get_attr];
+"674 compressed_weight_updated_constant33" [id=674, type=get_attr];
+"675 quantize_per_channel_default_33" [id=675, type=quantize_per_channel];
+"676 dequantize_per_channel_default_33" [id=676, type=dequantize_per_channel];
+"677 encoder_layers_encoder_layer_8_self_attention_in_proj_bias_0_0" [id=677, type=get_attr];
+"678 linear_32" [id=678, type=linear];
+"679 unflatten_8" [id=679, type=unflatten];
+"680 unsqueeze_8" [id=680, type=unsqueeze];
+"681 transpose_49" [id=681, type=transpose];
+"682 squeeze_8" [id=682, type=squeeze];
+"683 contiguous_8" [id=683, type=contiguous];
+"684 quantize_per_tensor_default_50" [id=684, type=quantize_per_tensor];
+"685 dequantize_per_tensor_default_50" [id=685, type=dequantize_per_tensor];
+"686 select_24" [id=686, type=select];
+"687 quantize_per_tensor_default_51" [id=687, type=quantize_per_tensor];
+"688 dequantize_per_tensor_default_51" [id=688, type=dequantize_per_tensor];
+"689 select_25" [id=689, type=select];
+"690 select_26" [id=690, type=select];
+"691 view_64" [id=691, type=view];
+"692 transpose_50" [id=692, type=transpose];
+"693 view_65" [id=693, type=view];
+"694 transpose_51" [id=694, type=transpose];
+"695 view_66" [id=695, type=view];
+"696 transpose_52" [id=696, type=transpose];
+"697 view_67" [id=697, type=view];
+"698 view_68" [id=698, type=view];
+"699 view_69" [id=699, type=view];
+"700 scaled_dot_product_attention_8" [id=700, type=scaled_dot_product_attention];
+"701 permute_9" [id=701, type=permute];
+"702 view_70" [id=702, type=view];
+"703 view_70_0_0_nncf_smooth_quant_0" [id=703, type=call_module];
+"704 quantize_per_tensor_default_52" [id=704, type=quantize_per_tensor];
+"705 dequantize_per_tensor_default_52" [id=705, type=dequantize_per_tensor];
+"706 linear_33_scale_0" [id=706, type=get_attr];
+"707 linear_33_zero_point_0" [id=707, type=get_attr];
+"708 compressed_weight_updated_constant34" [id=708, type=get_attr];
+"709 quantize_per_channel_default_34" [id=709, type=quantize_per_channel];
+"710 dequantize_per_channel_default_34" [id=710, type=dequantize_per_channel];
+"711 encoder_layers_encoder_layer_8_self_attention_out_proj_bias_0_0" [id=711, type=get_attr];
+"712 linear_33" [id=712, type=linear];
+"713 view_71" [id=713, type=view];
+"714 transpose_53" [id=714, type=transpose];
+"715 dropout_25" [id=715, type=dropout];
+"716 add_17" [id=716, type=add];
+"717 layer_norm_17" [id=717, type=layer_norm];
+"718 layer_norm_17_0_0_nncf_smooth_quant_0" [id=718, type=call_module];
+"719 quantize_per_tensor_default_53" [id=719, type=quantize_per_tensor];
+"720 dequantize_per_tensor_default_53" [id=720, type=dequantize_per_tensor];
+"721 linear_34_scale_0" [id=721, type=get_attr];
+"722 linear_34_zero_point_0" [id=722, type=get_attr];
+"723 compressed_weight_updated_constant35" [id=723, type=get_attr];
+"724 quantize_per_channel_default_35" [id=724, type=quantize_per_channel];
+"725 dequantize_per_channel_default_35" [id=725, type=dequantize_per_channel];
+"726 encoder_layers_encoder_layer_8_mlp_0_bias_0_0" [id=726, type=get_attr];
+"727 linear_34" [id=727, type=linear];
+"728 gelu_8" [id=728, type=gelu];
+"729 dropout_26" [id=729, type=dropout];
+"730 dropout_26_0_0_nncf_smooth_quant_0" [id=730, type=call_module];
+"731 quantize_per_tensor_default_54" [id=731, type=quantize_per_tensor];
+"732 dequantize_per_tensor_default_54" [id=732, type=dequantize_per_tensor];
+"733 linear_35_scale_0" [id=733, type=get_attr];
+"734 linear_35_zero_point_0" [id=734, type=get_attr];
+"735 compressed_weight_updated_constant36" [id=735, type=get_attr];
+"736 quantize_per_channel_default_36" [id=736, type=quantize_per_channel];
+"737 dequantize_per_channel_default_36" [id=737, type=dequantize_per_channel];
+"738 encoder_layers_encoder_layer_8_mlp_3_bias_0_0" [id=738, type=get_attr];
+"739 linear_35" [id=739, type=linear];
+"740 dropout_27" [id=740, type=dropout];
+"741 add_18" [id=741, type=add];
+"742 layer_norm_18" [id=742, type=layer_norm];
+"743 transpose_54" [id=743, type=transpose];
+"744 transpose_54_0_0_nncf_smooth_quant_0" [id=744, type=call_module];
+"745 quantize_per_tensor_default_55" [id=745, type=quantize_per_tensor];
+"746 dequantize_per_tensor_default_55" [id=746, type=dequantize_per_tensor];
+"747 linear_36_scale_0" [id=747, type=get_attr];
+"748 linear_36_zero_point_0" [id=748, type=get_attr];
+"749 compressed_weight_updated_constant37" [id=749, type=get_attr];
+"750 quantize_per_channel_default_37" [id=750, type=quantize_per_channel];
+"751 dequantize_per_channel_default_37" [id=751, type=dequantize_per_channel];
+"752 encoder_layers_encoder_layer_9_self_attention_in_proj_bias_0_0" [id=752, type=get_attr];
+"753 linear_36" [id=753, type=linear];
+"754 unflatten_9" [id=754, type=unflatten];
+"755 unsqueeze_9" [id=755, type=unsqueeze];
+"756 transpose_55" [id=756, type=transpose];
+"757 squeeze_9" [id=757, type=squeeze];
+"758 contiguous_9" [id=758, type=contiguous];
+"759 quantize_per_tensor_default_56" [id=759, type=quantize_per_tensor];
+"760 dequantize_per_tensor_default_56" [id=760, type=dequantize_per_tensor];
+"761 select_27" [id=761, type=select];
+"762 quantize_per_tensor_default_57" [id=762, type=quantize_per_tensor];
+"763 dequantize_per_tensor_default_57" [id=763, type=dequantize_per_tensor];
+"764 select_28" [id=764, type=select];
+"765 select_29" [id=765, type=select];
+"766 view_72" [id=766, type=view];
+"767 transpose_56" [id=767, type=transpose];
+"768 view_73" [id=768, type=view];
+"769 transpose_57" [id=769, type=transpose];
+"770 view_74" [id=770, type=view];
+"771 transpose_58" [id=771, type=transpose];
+"772 view_75" [id=772, type=view];
+"773 view_76" [id=773, type=view];
+"774 view_77" [id=774, type=view];
+"775 scaled_dot_product_attention_9" [id=775, type=scaled_dot_product_attention];
+"776 permute_10" [id=776, type=permute];
+"777 view_78" [id=777, type=view];
+"778 view_78_0_0_nncf_smooth_quant_0" [id=778, type=call_module];
+"779 quantize_per_tensor_default_58" [id=779, type=quantize_per_tensor];
+"780 dequantize_per_tensor_default_58" [id=780, type=dequantize_per_tensor];
+"781 linear_37_scale_0" [id=781, type=get_attr];
+"782 linear_37_zero_point_0" [id=782, type=get_attr];
+"783 compressed_weight_updated_constant38" [id=783, type=get_attr];
+"784 quantize_per_channel_default_38" [id=784, type=quantize_per_channel];
+"785 dequantize_per_channel_default_38" [id=785, type=dequantize_per_channel];
+"786 encoder_layers_encoder_layer_9_self_attention_out_proj_bias_0_0" [id=786, type=get_attr];
+"787 linear_37" [id=787, type=linear];
+"788 view_79" [id=788, type=view];
+"789 transpose_59" [id=789, type=transpose];
+"790 dropout_28" [id=790, type=dropout];
+"791 add_19" [id=791, type=add];
+"792 layer_norm_19" [id=792, type=layer_norm];
+"793 layer_norm_19_0_0_nncf_smooth_quant_0" [id=793, type=call_module];
+"794 quantize_per_tensor_default_59" [id=794, type=quantize_per_tensor];
+"795 dequantize_per_tensor_default_59" [id=795, type=dequantize_per_tensor];
+"796 linear_38_scale_0" [id=796, type=get_attr];
+"797 linear_38_zero_point_0" [id=797, type=get_attr];
+"798 compressed_weight_updated_constant39" [id=798, type=get_attr];
+"799 quantize_per_channel_default_39" [id=799, type=quantize_per_channel];
+"800 dequantize_per_channel_default_39" [id=800, type=dequantize_per_channel];
+"801 encoder_layers_encoder_layer_9_mlp_0_bias_0_0" [id=801, type=get_attr];
+"802 linear_38" [id=802, type=linear];
+"803 gelu_9" [id=803, type=gelu];
+"804 dropout_29" [id=804, type=dropout];
+"805 dropout_29_0_0_nncf_smooth_quant_0" [id=805, type=call_module];
+"806 quantize_per_tensor_default_60" [id=806, type=quantize_per_tensor];
+"807 dequantize_per_tensor_default_60" [id=807, type=dequantize_per_tensor];
+"808 linear_39_scale_0" [id=808, type=get_attr];
+"809 linear_39_zero_point_0" [id=809, type=get_attr];
+"810 compressed_weight_updated_constant40" [id=810, type=get_attr];
+"811 quantize_per_channel_default_40" [id=811, type=quantize_per_channel];
+"812 dequantize_per_channel_default_40" [id=812, type=dequantize_per_channel];
+"813 encoder_layers_encoder_layer_9_mlp_3_bias_0_0" [id=813, type=get_attr];
+"814 linear_39" [id=814, type=linear];
+"815 dropout_30" [id=815, type=dropout];
+"816 add_20" [id=816, type=add];
+"817 layer_norm_20" [id=817, type=layer_norm];
+"818 transpose_60" [id=818, type=transpose];
+"819 transpose_60_0_0_nncf_smooth_quant_0" [id=819, type=call_module];
+"820 quantize_per_tensor_default_61" [id=820, type=quantize_per_tensor];
+"821 dequantize_per_tensor_default_61" [id=821, type=dequantize_per_tensor];
+"822 linear_40_scale_0" [id=822, type=get_attr];
+"823 linear_40_zero_point_0" [id=823, type=get_attr];
+"824 compressed_weight_updated_constant41" [id=824, type=get_attr];
+"825 quantize_per_channel_default_41" [id=825, type=quantize_per_channel];
+"826 dequantize_per_channel_default_41" [id=826, type=dequantize_per_channel];
+"827 encoder_layers_encoder_layer_10_self_attention_in_proj_bias_0_0" [id=827, type=get_attr];
+"828 linear_40" [id=828, type=linear];
+"829 unflatten_10" [id=829, type=unflatten];
+"830 unsqueeze_10" [id=830, type=unsqueeze];
+"831 transpose_61" [id=831, type=transpose];
+"832 squeeze_10" [id=832, type=squeeze];
+"833 contiguous_10" [id=833, type=contiguous];
+"834 quantize_per_tensor_default_62" [id=834, type=quantize_per_tensor];
+"835 dequantize_per_tensor_default_62" [id=835, type=dequantize_per_tensor];
+"836 select_30" [id=836, type=select];
+"837 quantize_per_tensor_default_63" [id=837, type=quantize_per_tensor];
+"838 dequantize_per_tensor_default_63" [id=838, type=dequantize_per_tensor];
+"839 select_31" [id=839, type=select];
+"840 select_32" [id=840, type=select];
+"841 view_80" [id=841, type=view];
+"842 transpose_62" [id=842, type=transpose];
+"843 view_81" [id=843, type=view];
+"844 transpose_63" [id=844, type=transpose];
+"845 view_82" [id=845, type=view];
+"846 transpose_64" [id=846, type=transpose];
+"847 view_83" [id=847, type=view];
+"848 view_84" [id=848, type=view];
+"849 view_85" [id=849, type=view];
+"850 scaled_dot_product_attention_10" [id=850, type=scaled_dot_product_attention];
+"851 permute_11" [id=851, type=permute];
+"852 view_86" [id=852, type=view];
+"853 view_86_0_0_nncf_smooth_quant_0" [id=853, type=call_module];
+"854 quantize_per_tensor_default_64" [id=854, type=quantize_per_tensor];
+"855 dequantize_per_tensor_default_64" [id=855, type=dequantize_per_tensor];
+"856 linear_41_scale_0" [id=856, type=get_attr];
+"857 linear_41_zero_point_0" [id=857, type=get_attr];
+"858 compressed_weight_updated_constant42" [id=858, type=get_attr];
+"859 quantize_per_channel_default_42" [id=859, type=quantize_per_channel];
+"860 dequantize_per_channel_default_42" [id=860, type=dequantize_per_channel];
+"861 encoder_layers_encoder_layer_10_self_attention_out_proj_bias_0_0" [id=861, type=get_attr];
+"862 linear_41" [id=862, type=linear];
+"863 view_87" [id=863, type=view];
+"864 transpose_65" [id=864, type=transpose];
+"865 dropout_31" [id=865, type=dropout];
+"866 add_21" [id=866, type=add];
+"867 layer_norm_21" [id=867, type=layer_norm];
+"868 layer_norm_21_0_0_nncf_smooth_quant_0" [id=868, type=call_module];
+"869 quantize_per_tensor_default_65" [id=869, type=quantize_per_tensor];
+"870 dequantize_per_tensor_default_65" [id=870, type=dequantize_per_tensor];
+"871 linear_42_scale_0" [id=871, type=get_attr];
+"872 linear_42_zero_point_0" [id=872, type=get_attr];
+"873 compressed_weight_updated_constant43" [id=873, type=get_attr];
+"874 quantize_per_channel_default_43" [id=874, type=quantize_per_channel];
+"875 dequantize_per_channel_default_43" [id=875, type=dequantize_per_channel];
+"876 encoder_layers_encoder_layer_10_mlp_0_bias_0_0" [id=876, type=get_attr];
+"877 linear_42" [id=877, type=linear];
+"878 gelu_10" [id=878, type=gelu];
+"879 dropout_32" [id=879, type=dropout];
+"880 dropout_32_0_0_nncf_smooth_quant_0" [id=880, type=call_module];
+"881 quantize_per_tensor_default_66" [id=881, type=quantize_per_tensor];
+"882 dequantize_per_tensor_default_66" [id=882, type=dequantize_per_tensor];
+"883 linear_43_scale_0" [id=883, type=get_attr];
+"884 linear_43_zero_point_0" [id=884, type=get_attr];
+"885 compressed_weight_updated_constant44" [id=885, type=get_attr];
+"886 quantize_per_channel_default_44" [id=886, type=quantize_per_channel];
+"887 dequantize_per_channel_default_44" [id=887, type=dequantize_per_channel];
+"888 encoder_layers_encoder_layer_10_mlp_3_bias_0_0" [id=888, type=get_attr];
+"889 linear_43" [id=889, type=linear];
+"890 dropout_33" [id=890, type=dropout];
+"891 add_22" [id=891, type=add];
+"892 layer_norm_22" [id=892, type=layer_norm];
+"893 transpose_66" [id=893, type=transpose];
+"894 transpose_66_0_0_nncf_smooth_quant_0" [id=894, type=call_module];
+"895 quantize_per_tensor_default_67" [id=895, type=quantize_per_tensor];
+"896 dequantize_per_tensor_default_67" [id=896, type=dequantize_per_tensor];
+"897 linear_44_scale_0" [id=897, type=get_attr];
+"898 linear_44_zero_point_0" [id=898, type=get_attr];
+"899 compressed_weight_updated_constant45" [id=899, type=get_attr];
+"900 quantize_per_channel_default_45" [id=900, type=quantize_per_channel];
+"901 dequantize_per_channel_default_45" [id=901, type=dequantize_per_channel];
+"902 encoder_layers_encoder_layer_11_self_attention_in_proj_bias_0_0" [id=902, type=get_attr];
+"903 linear_44" [id=903, type=linear];
+"904 unflatten_11" [id=904, type=unflatten];
+"905 unsqueeze_11" [id=905, type=unsqueeze];
+"906 transpose_67" [id=906, type=transpose];
+"907 squeeze_11" [id=907, type=squeeze];
+"908 contiguous_11" [id=908, type=contiguous];
+"909 quantize_per_tensor_default_68" [id=909, type=quantize_per_tensor];
+"910 dequantize_per_tensor_default_68" [id=910, type=dequantize_per_tensor];
+"911 select_33" [id=911, type=select];
+"912 quantize_per_tensor_default_69" [id=912, type=quantize_per_tensor];
+"913 dequantize_per_tensor_default_69" [id=913, type=dequantize_per_tensor];
+"914 select_34" [id=914, type=select];
+"915 select_35" [id=915, type=select];
+"916 view_88" [id=916, type=view];
+"917 transpose_68" [id=917, type=transpose];
+"918 view_89" [id=918, type=view];
+"919 transpose_69" [id=919, type=transpose];
+"920 view_90" [id=920, type=view];
+"921 transpose_70" [id=921, type=transpose];
+"922 view_91" [id=922, type=view];
+"923 view_92" [id=923, type=view];
+"924 view_93" [id=924, type=view];
+"925 scaled_dot_product_attention_11" [id=925, type=scaled_dot_product_attention];
+"926 permute_12" [id=926, type=permute];
+"927 view_94" [id=927, type=view];
+"928 view_94_0_0_nncf_smooth_quant_0" [id=928, type=call_module];
+"929 quantize_per_tensor_default_70" [id=929, type=quantize_per_tensor];
+"930 dequantize_per_tensor_default_70" [id=930, type=dequantize_per_tensor];
+"931 linear_45_scale_0" [id=931, type=get_attr];
+"932 linear_45_zero_point_0" [id=932, type=get_attr];
+"933 compressed_weight_updated_constant46" [id=933, type=get_attr];
+"934 quantize_per_channel_default_46" [id=934, type=quantize_per_channel];
+"935 dequantize_per_channel_default_46" [id=935, type=dequantize_per_channel];
+"936 encoder_layers_encoder_layer_11_self_attention_out_proj_bias_0_0" [id=936, type=get_attr];
+"937 linear_45" [id=937, type=linear];
+"938 view_95" [id=938, type=view];
+"939 transpose_71" [id=939, type=transpose];
+"940 dropout_34" [id=940, type=dropout];
+"941 add_23" [id=941, type=add];
+"942 layer_norm_23" [id=942, type=layer_norm];
+"943 layer_norm_23_0_0_nncf_smooth_quant_0" [id=943, type=call_module];
+"944 quantize_per_tensor_default_71" [id=944, type=quantize_per_tensor];
+"945 dequantize_per_tensor_default_71" [id=945, type=dequantize_per_tensor];
+"946 linear_46_scale_0" [id=946, type=get_attr];
+"947 linear_46_zero_point_0" [id=947, type=get_attr];
+"948 compressed_weight_updated_constant47" [id=948, type=get_attr];
+"949 quantize_per_channel_default_47" [id=949, type=quantize_per_channel];
+"950 dequantize_per_channel_default_47" [id=950, type=dequantize_per_channel];
+"951 encoder_layers_encoder_layer_11_mlp_0_bias_0_0" [id=951, type=get_attr];
+"952 linear_46" [id=952, type=linear];
+"953 gelu_11" [id=953, type=gelu];
+"954 dropout_35" [id=954, type=dropout];
+"955 dropout_35_0_0_nncf_smooth_quant_0" [id=955, type=call_module];
+"956 quantize_per_tensor_default_72" [id=956, type=quantize_per_tensor];
+"957 dequantize_per_tensor_default_72" [id=957, type=dequantize_per_tensor];
+"958 linear_47_scale_0" [id=958, type=get_attr];
+"959 linear_47_zero_point_0" [id=959, type=get_attr];
+"960 compressed_weight_updated_constant48" [id=960, type=get_attr];
+"961 quantize_per_channel_default_48" [id=961, type=quantize_per_channel];
+"962 dequantize_per_channel_default_48" [id=962, type=dequantize_per_channel];
+"963 encoder_layers_encoder_layer_11_mlp_3_bias_0_0" [id=963, type=get_attr];
+"964 linear_47" [id=964, type=linear];
+"965 dropout_36" [id=965, type=dropout];
+"966 add_24" [id=966, type=add];
 "967 layer_norm_24" [id=967, type=layer_norm];
 "968 slice_1" [id=968, type=slice];
 "969 select_36" [id=969, type=select];
@@ -977,1122 +977,1122 @@ strict digraph  {
 "975 compressed_weight_updated_constant49" [id=975, type=get_attr];
 "976 quantize_per_channel_default_49" [id=976, type=quantize_per_channel];
 "977 dequantize_per_channel_default_49" [id=977, type=dequantize_per_channel];
-"978 _param_constant151_0_0" [id=978, type=get_attr];
+"978 heads_head_bias_0_0" [id=978, type=get_attr];
 "979 linear_48" [id=979, type=linear];
 "980 output" [id=980, type=output];
-"0 arg0_1" -> "1 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"1 quantize_per_tensor_default" -> "2 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
-"2 dequantize_per_tensor_default" -> "9 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"3 conv2d_scale_0" -> "6 quantize_per_channel_default"  [label="(768,)", style=solid];
-"3 conv2d_scale_0" -> "7 dequantize_per_channel_default"  [label="(768,)", style=solid];
-"4 conv2d_zero_point_0" -> "6 quantize_per_channel_default"  [label="(768,)", style=solid];
-"4 conv2d_zero_point_0" -> "7 dequantize_per_channel_default"  [label="(768,)", style=solid];
-"5 compressed_weight_updated_constant0" -> "6 quantize_per_channel_default"  [label="(768, 3, 16, 16)", style=solid];
-"6 quantize_per_channel_default" -> "7 dequantize_per_channel_default"  [label="(768, 3, 16, 16)", style=solid];
-"7 dequantize_per_channel_default" -> "9 conv2d"  [label="(768, 3, 16, 16)", style=solid];
-"8 _param_constant1_0_0" -> "9 conv2d"  [label="(768,)", style=solid];
-"9 conv2d" -> "10 reshape"  [label="(1, 768, 14, 14)", style=solid];
-"10 reshape" -> "11 permute"  [label="(1, 768, 196)", style=solid];
-"11 permute" -> "13 cat"  [label="(1, 196, 768)", style=solid];
-"12 _frozen_param0" -> "13 cat"  [label="(1, 1, 768)", style=solid];
-"13 cat" -> "15 add"  [label="(1, 197, 768)", style=solid];
-"14 _param_constant3" -> "15 add"  [label="(1, 197, 768)", style=solid];
-"15 add" -> "16 dropout"  [label="(1, 197, 768)", style=solid];
-"16 dropout" -> "19 layer_norm"  [label="(1, 197, 768)", style=solid];
-"16 dropout" -> "68 add_1"  [label="(1, 197, 768)", style=solid];
-"17 _param_constant4" -> "19 layer_norm"  [label="(768,)", style=solid];
-"18 _param_constant5" -> "19 layer_norm"  [label="(768,)", style=solid];
-"19 layer_norm" -> "20 transpose"  [label="(1, 197, 768)", style=solid];
-"20 transpose" -> "21 transpose_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"21 transpose_0_0_nncf_smooth_quant_0" -> "22 quantize_per_tensor_default_1"  [label="(197, 1, 768)", style=solid];
-"22 quantize_per_tensor_default_1" -> "23 dequantize_per_tensor_default_1"  [label="(197, 1, 768)", style=solid];
-"23 dequantize_per_tensor_default_1" -> "30 linear"  [label="(197, 1, 768)", style=solid];
-"24 linear_scale_0" -> "27 quantize_per_channel_default_1"  [label="(2304,)", style=solid];
-"24 linear_scale_0" -> "28 dequantize_per_channel_default_1"  [label="(2304,)", style=solid];
-"25 linear_zero_point_0" -> "27 quantize_per_channel_default_1"  [label="(2304,)", style=solid];
-"25 linear_zero_point_0" -> "28 dequantize_per_channel_default_1"  [label="(2304,)", style=solid];
-"26 compressed_weight_updated_constant1" -> "27 quantize_per_channel_default_1"  [label="(2304, 768)", style=solid];
-"27 quantize_per_channel_default_1" -> "28 dequantize_per_channel_default_1"  [label="(2304, 768)", style=solid];
-"28 dequantize_per_channel_default_1" -> "30 linear"  [label="(2304, 768)", style=solid];
-"29 _param_constant7_0_0" -> "30 linear"  [label="(2304,)", style=solid];
-"30 linear" -> "31 unflatten"  [label="(197, 1, 2304)", style=solid];
-"31 unflatten" -> "32 unsqueeze"  [label="(197, 1, 3, 768)", style=solid];
-"32 unsqueeze" -> "33 transpose_1"  [label="(1, 197, 1, 3, 768)", style=solid];
-"33 transpose_1" -> "34 squeeze"  [label="(3, 197, 1, 1, 768)", style=solid];
-"34 squeeze" -> "35 contiguous"  [label="(3, 197, 1, 768)", style=solid];
-"35 contiguous" -> "36 quantize_per_tensor_default_2"  [label="(3, 197, 1, 768)", style=solid];
-"35 contiguous" -> "39 quantize_per_tensor_default_3"  [label="(3, 197, 1, 768)", style=solid];
-"35 contiguous" -> "42 select_2"  [label="(3, 197, 1, 768)", style=solid];
-"36 quantize_per_tensor_default_2" -> "37 dequantize_per_tensor_default_2"  [label="(3, 197, 1, 768)", style=solid];
-"37 dequantize_per_tensor_default_2" -> "38 select"  [label="(3, 197, 1, 768)", style=solid];
-"38 select" -> "43 view"  [label="(197, 1, 768)", style=solid];
-"39 quantize_per_tensor_default_3" -> "40 dequantize_per_tensor_default_3"  [label="(3, 197, 1, 768)", style=solid];
-"40 dequantize_per_tensor_default_3" -> "41 select_1"  [label="(3, 197, 1, 768)", style=solid];
-"41 select_1" -> "45 view_1"  [label="(197, 1, 768)", style=solid];
-"42 select_2" -> "47 view_2"  [label="(197, 1, 768)", style=solid];
-"43 view" -> "44 transpose_2"  [label="(197, 12, 64)", style=solid];
-"44 transpose_2" -> "49 view_3"  [label="(12, 197, 64)", style=solid];
-"45 view_1" -> "46 transpose_3"  [label="(197, 12, 64)", style=solid];
-"46 transpose_3" -> "50 view_4"  [label="(12, 197, 64)", style=solid];
-"47 view_2" -> "48 transpose_4"  [label="(197, 12, 64)", style=solid];
-"48 transpose_4" -> "51 view_5"  [label="(12, 197, 64)", style=solid];
-"49 view_3" -> "52 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
-"50 view_4" -> "52 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
-"51 view_5" -> "52 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
-"52 scaled_dot_product_attention" -> "53 permute_1"  [label="(1, 12, 197, 64)", style=solid];
-"53 permute_1" -> "54 view_6"  [label="(197, 1, 12, 64)", style=solid];
-"54 view_6" -> "55 view_6_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"55 view_6_0_0_nncf_smooth_quant_0" -> "56 quantize_per_tensor_default_4"  [label="(197, 768)", style=solid];
-"56 quantize_per_tensor_default_4" -> "57 dequantize_per_tensor_default_4"  [label="(197, 768)", style=solid];
-"57 dequantize_per_tensor_default_4" -> "64 linear_1"  [label="(197, 768)", style=solid];
-"58 linear_1_scale_0" -> "61 quantize_per_channel_default_2"  [label="(768,)", style=solid];
-"58 linear_1_scale_0" -> "62 dequantize_per_channel_default_2"  [label="(768,)", style=solid];
-"59 linear_1_zero_point_0" -> "61 quantize_per_channel_default_2"  [label="(768,)", style=solid];
-"59 linear_1_zero_point_0" -> "62 dequantize_per_channel_default_2"  [label="(768,)", style=solid];
-"60 compressed_weight_updated_constant2" -> "61 quantize_per_channel_default_2"  [label="(768, 768)", style=solid];
-"61 quantize_per_channel_default_2" -> "62 dequantize_per_channel_default_2"  [label="(768, 768)", style=solid];
-"62 dequantize_per_channel_default_2" -> "64 linear_1"  [label="(768, 768)", style=solid];
-"63 _param_constant9_0_0" -> "64 linear_1"  [label="(768,)", style=solid];
-"64 linear_1" -> "65 view_7"  [label="(197, 768)", style=solid];
-"65 view_7" -> "66 transpose_5"  [label="(197, 1, 768)", style=solid];
-"66 transpose_5" -> "67 dropout_1"  [label="(1, 197, 768)", style=solid];
-"67 dropout_1" -> "68 add_1"  [label="(1, 197, 768)", style=solid];
-"68 add_1" -> "71 layer_norm_1"  [label="(1, 197, 768)", style=solid];
-"68 add_1" -> "95 add_2"  [label="(1, 197, 768)", style=solid];
-"69 _param_constant10" -> "71 layer_norm_1"  [label="(768,)", style=solid];
-"70 _param_constant11" -> "71 layer_norm_1"  [label="(768,)", style=solid];
-"71 layer_norm_1" -> "72 layer_norm_1_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"72 layer_norm_1_0_0_nncf_smooth_quant_0" -> "73 quantize_per_tensor_default_5"  [label="(1, 197, 768)", style=solid];
-"73 quantize_per_tensor_default_5" -> "74 dequantize_per_tensor_default_5"  [label="(1, 197, 768)", style=solid];
-"74 dequantize_per_tensor_default_5" -> "81 linear_2"  [label="(1, 197, 768)", style=solid];
-"75 linear_2_scale_0" -> "78 quantize_per_channel_default_3"  [label="(3072,)", style=solid];
-"75 linear_2_scale_0" -> "79 dequantize_per_channel_default_3"  [label="(3072,)", style=solid];
-"76 linear_2_zero_point_0" -> "78 quantize_per_channel_default_3"  [label="(3072,)", style=solid];
-"76 linear_2_zero_point_0" -> "79 dequantize_per_channel_default_3"  [label="(3072,)", style=solid];
-"77 compressed_weight_updated_constant3" -> "78 quantize_per_channel_default_3"  [label="(3072, 768)", style=solid];
-"78 quantize_per_channel_default_3" -> "79 dequantize_per_channel_default_3"  [label="(3072, 768)", style=solid];
-"79 dequantize_per_channel_default_3" -> "81 linear_2"  [label="(3072, 768)", style=solid];
-"80 _param_constant13_0_0" -> "81 linear_2"  [label="(3072,)", style=solid];
-"81 linear_2" -> "82 gelu"  [label="(1, 197, 3072)", style=solid];
-"82 gelu" -> "83 dropout_2"  [label="(1, 197, 3072)", style=solid];
-"83 dropout_2" -> "84 dropout_2_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"84 dropout_2_0_0_nncf_smooth_quant_0" -> "85 quantize_per_tensor_default_6"  [label="(1, 197, 3072)", style=solid];
-"85 quantize_per_tensor_default_6" -> "86 dequantize_per_tensor_default_6"  [label="(1, 197, 3072)", style=solid];
-"86 dequantize_per_tensor_default_6" -> "93 linear_3"  [label="(1, 197, 3072)", style=solid];
-"87 linear_3_scale_0" -> "90 quantize_per_channel_default_4"  [label="(768,)", style=solid];
-"87 linear_3_scale_0" -> "91 dequantize_per_channel_default_4"  [label="(768,)", style=solid];
-"88 linear_3_zero_point_0" -> "90 quantize_per_channel_default_4"  [label="(768,)", style=solid];
-"88 linear_3_zero_point_0" -> "91 dequantize_per_channel_default_4"  [label="(768,)", style=solid];
-"89 compressed_weight_updated_constant4" -> "90 quantize_per_channel_default_4"  [label="(768, 3072)", style=solid];
-"90 quantize_per_channel_default_4" -> "91 dequantize_per_channel_default_4"  [label="(768, 3072)", style=solid];
-"91 dequantize_per_channel_default_4" -> "93 linear_3"  [label="(768, 3072)", style=solid];
-"92 _param_constant15_0_0" -> "93 linear_3"  [label="(768,)", style=solid];
-"93 linear_3" -> "94 dropout_3"  [label="(1, 197, 768)", style=solid];
-"94 dropout_3" -> "95 add_2"  [label="(1, 197, 768)", style=solid];
-"95 add_2" -> "98 layer_norm_2"  [label="(1, 197, 768)", style=solid];
-"95 add_2" -> "147 add_3"  [label="(1, 197, 768)", style=solid];
-"96 _param_constant16" -> "98 layer_norm_2"  [label="(768,)", style=solid];
-"97 _param_constant17" -> "98 layer_norm_2"  [label="(768,)", style=solid];
-"98 layer_norm_2" -> "99 transpose_6"  [label="(1, 197, 768)", style=solid];
-"99 transpose_6" -> "100 transpose_6_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"100 transpose_6_0_0_nncf_smooth_quant_0" -> "101 quantize_per_tensor_default_7"  [label="(197, 1, 768)", style=solid];
-"101 quantize_per_tensor_default_7" -> "102 dequantize_per_tensor_default_7"  [label="(197, 1, 768)", style=solid];
-"102 dequantize_per_tensor_default_7" -> "109 linear_4"  [label="(197, 1, 768)", style=solid];
-"103 linear_4_scale_0" -> "106 quantize_per_channel_default_5"  [label="(2304,)", style=solid];
-"103 linear_4_scale_0" -> "107 dequantize_per_channel_default_5"  [label="(2304,)", style=solid];
-"104 linear_4_zero_point_0" -> "106 quantize_per_channel_default_5"  [label="(2304,)", style=solid];
-"104 linear_4_zero_point_0" -> "107 dequantize_per_channel_default_5"  [label="(2304,)", style=solid];
-"105 compressed_weight_updated_constant5" -> "106 quantize_per_channel_default_5"  [label="(2304, 768)", style=solid];
-"106 quantize_per_channel_default_5" -> "107 dequantize_per_channel_default_5"  [label="(2304, 768)", style=solid];
-"107 dequantize_per_channel_default_5" -> "109 linear_4"  [label="(2304, 768)", style=solid];
-"108 _param_constant19_0_0" -> "109 linear_4"  [label="(2304,)", style=solid];
-"109 linear_4" -> "110 unflatten_1"  [label="(197, 1, 2304)", style=solid];
-"110 unflatten_1" -> "111 unsqueeze_1"  [label="(197, 1, 3, 768)", style=solid];
-"111 unsqueeze_1" -> "112 transpose_7"  [label="(1, 197, 1, 3, 768)", style=solid];
-"112 transpose_7" -> "113 squeeze_1"  [label="(3, 197, 1, 1, 768)", style=solid];
-"113 squeeze_1" -> "114 contiguous_1"  [label="(3, 197, 1, 768)", style=solid];
-"114 contiguous_1" -> "115 quantize_per_tensor_default_8"  [label="(3, 197, 1, 768)", style=solid];
-"114 contiguous_1" -> "118 quantize_per_tensor_default_9"  [label="(3, 197, 1, 768)", style=solid];
-"114 contiguous_1" -> "121 select_5"  [label="(3, 197, 1, 768)", style=solid];
-"115 quantize_per_tensor_default_8" -> "116 dequantize_per_tensor_default_8"  [label="(3, 197, 1, 768)", style=solid];
-"116 dequantize_per_tensor_default_8" -> "117 select_3"  [label="(3, 197, 1, 768)", style=solid];
-"117 select_3" -> "122 view_8"  [label="(197, 1, 768)", style=solid];
-"118 quantize_per_tensor_default_9" -> "119 dequantize_per_tensor_default_9"  [label="(3, 197, 1, 768)", style=solid];
-"119 dequantize_per_tensor_default_9" -> "120 select_4"  [label="(3, 197, 1, 768)", style=solid];
-"120 select_4" -> "124 view_9"  [label="(197, 1, 768)", style=solid];
-"121 select_5" -> "126 view_10"  [label="(197, 1, 768)", style=solid];
-"122 view_8" -> "123 transpose_8"  [label="(197, 12, 64)", style=solid];
-"123 transpose_8" -> "128 view_11"  [label="(12, 197, 64)", style=solid];
-"124 view_9" -> "125 transpose_9"  [label="(197, 12, 64)", style=solid];
-"125 transpose_9" -> "129 view_12"  [label="(12, 197, 64)", style=solid];
-"126 view_10" -> "127 transpose_10"  [label="(197, 12, 64)", style=solid];
-"127 transpose_10" -> "130 view_13"  [label="(12, 197, 64)", style=solid];
-"128 view_11" -> "131 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
-"129 view_12" -> "131 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
-"130 view_13" -> "131 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
-"131 scaled_dot_product_attention_1" -> "132 permute_2"  [label="(1, 12, 197, 64)", style=solid];
-"132 permute_2" -> "133 view_14"  [label="(197, 1, 12, 64)", style=solid];
-"133 view_14" -> "134 view_14_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"134 view_14_0_0_nncf_smooth_quant_0" -> "135 quantize_per_tensor_default_10"  [label="(197, 768)", style=solid];
-"135 quantize_per_tensor_default_10" -> "136 dequantize_per_tensor_default_10"  [label="(197, 768)", style=solid];
-"136 dequantize_per_tensor_default_10" -> "143 linear_5"  [label="(197, 768)", style=solid];
-"137 linear_5_scale_0" -> "140 quantize_per_channel_default_6"  [label="(768,)", style=solid];
-"137 linear_5_scale_0" -> "141 dequantize_per_channel_default_6"  [label="(768,)", style=solid];
-"138 linear_5_zero_point_0" -> "140 quantize_per_channel_default_6"  [label="(768,)", style=solid];
-"138 linear_5_zero_point_0" -> "141 dequantize_per_channel_default_6"  [label="(768,)", style=solid];
-"139 compressed_weight_updated_constant6" -> "140 quantize_per_channel_default_6"  [label="(768, 768)", style=solid];
-"140 quantize_per_channel_default_6" -> "141 dequantize_per_channel_default_6"  [label="(768, 768)", style=solid];
-"141 dequantize_per_channel_default_6" -> "143 linear_5"  [label="(768, 768)", style=solid];
-"142 _param_constant21_0_0" -> "143 linear_5"  [label="(768,)", style=solid];
-"143 linear_5" -> "144 view_15"  [label="(197, 768)", style=solid];
-"144 view_15" -> "145 transpose_11"  [label="(197, 1, 768)", style=solid];
-"145 transpose_11" -> "146 dropout_4"  [label="(1, 197, 768)", style=solid];
-"146 dropout_4" -> "147 add_3"  [label="(1, 197, 768)", style=solid];
-"147 add_3" -> "150 layer_norm_3"  [label="(1, 197, 768)", style=solid];
-"147 add_3" -> "174 add_4"  [label="(1, 197, 768)", style=solid];
-"148 _param_constant22" -> "150 layer_norm_3"  [label="(768,)", style=solid];
-"149 _param_constant23" -> "150 layer_norm_3"  [label="(768,)", style=solid];
-"150 layer_norm_3" -> "151 layer_norm_3_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"151 layer_norm_3_0_0_nncf_smooth_quant_0" -> "152 quantize_per_tensor_default_11"  [label="(1, 197, 768)", style=solid];
-"152 quantize_per_tensor_default_11" -> "153 dequantize_per_tensor_default_11"  [label="(1, 197, 768)", style=solid];
-"153 dequantize_per_tensor_default_11" -> "160 linear_6"  [label="(1, 197, 768)", style=solid];
-"154 linear_6_scale_0" -> "157 quantize_per_channel_default_7"  [label="(3072,)", style=solid];
-"154 linear_6_scale_0" -> "158 dequantize_per_channel_default_7"  [label="(3072,)", style=solid];
-"155 linear_6_zero_point_0" -> "157 quantize_per_channel_default_7"  [label="(3072,)", style=solid];
-"155 linear_6_zero_point_0" -> "158 dequantize_per_channel_default_7"  [label="(3072,)", style=solid];
-"156 compressed_weight_updated_constant7" -> "157 quantize_per_channel_default_7"  [label="(3072, 768)", style=solid];
-"157 quantize_per_channel_default_7" -> "158 dequantize_per_channel_default_7"  [label="(3072, 768)", style=solid];
-"158 dequantize_per_channel_default_7" -> "160 linear_6"  [label="(3072, 768)", style=solid];
-"159 _param_constant25_0_0" -> "160 linear_6"  [label="(3072,)", style=solid];
-"160 linear_6" -> "161 gelu_1"  [label="(1, 197, 3072)", style=solid];
-"161 gelu_1" -> "162 dropout_5"  [label="(1, 197, 3072)", style=solid];
-"162 dropout_5" -> "163 dropout_5_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"163 dropout_5_0_0_nncf_smooth_quant_0" -> "164 quantize_per_tensor_default_12"  [label="(1, 197, 3072)", style=solid];
-"164 quantize_per_tensor_default_12" -> "165 dequantize_per_tensor_default_12"  [label="(1, 197, 3072)", style=solid];
-"165 dequantize_per_tensor_default_12" -> "172 linear_7"  [label="(1, 197, 3072)", style=solid];
-"166 linear_7_scale_0" -> "169 quantize_per_channel_default_8"  [label="(768,)", style=solid];
-"166 linear_7_scale_0" -> "170 dequantize_per_channel_default_8"  [label="(768,)", style=solid];
-"167 linear_7_zero_point_0" -> "169 quantize_per_channel_default_8"  [label="(768,)", style=solid];
-"167 linear_7_zero_point_0" -> "170 dequantize_per_channel_default_8"  [label="(768,)", style=solid];
-"168 compressed_weight_updated_constant8" -> "169 quantize_per_channel_default_8"  [label="(768, 3072)", style=solid];
-"169 quantize_per_channel_default_8" -> "170 dequantize_per_channel_default_8"  [label="(768, 3072)", style=solid];
-"170 dequantize_per_channel_default_8" -> "172 linear_7"  [label="(768, 3072)", style=solid];
-"171 _param_constant27_0_0" -> "172 linear_7"  [label="(768,)", style=solid];
-"172 linear_7" -> "173 dropout_6"  [label="(1, 197, 768)", style=solid];
-"173 dropout_6" -> "174 add_4"  [label="(1, 197, 768)", style=solid];
-"174 add_4" -> "177 layer_norm_4"  [label="(1, 197, 768)", style=solid];
-"174 add_4" -> "226 add_5"  [label="(1, 197, 768)", style=solid];
-"175 _param_constant28" -> "177 layer_norm_4"  [label="(768,)", style=solid];
-"176 _param_constant29" -> "177 layer_norm_4"  [label="(768,)", style=solid];
-"177 layer_norm_4" -> "178 transpose_12"  [label="(1, 197, 768)", style=solid];
-"178 transpose_12" -> "179 transpose_12_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"179 transpose_12_0_0_nncf_smooth_quant_0" -> "180 quantize_per_tensor_default_13"  [label="(197, 1, 768)", style=solid];
-"180 quantize_per_tensor_default_13" -> "181 dequantize_per_tensor_default_13"  [label="(197, 1, 768)", style=solid];
-"181 dequantize_per_tensor_default_13" -> "188 linear_8"  [label="(197, 1, 768)", style=solid];
-"182 linear_8_scale_0" -> "185 quantize_per_channel_default_9"  [label="(2304,)", style=solid];
-"182 linear_8_scale_0" -> "186 dequantize_per_channel_default_9"  [label="(2304,)", style=solid];
-"183 linear_8_zero_point_0" -> "185 quantize_per_channel_default_9"  [label="(2304,)", style=solid];
-"183 linear_8_zero_point_0" -> "186 dequantize_per_channel_default_9"  [label="(2304,)", style=solid];
-"184 compressed_weight_updated_constant9" -> "185 quantize_per_channel_default_9"  [label="(2304, 768)", style=solid];
-"185 quantize_per_channel_default_9" -> "186 dequantize_per_channel_default_9"  [label="(2304, 768)", style=solid];
-"186 dequantize_per_channel_default_9" -> "188 linear_8"  [label="(2304, 768)", style=solid];
-"187 _param_constant31_0_0" -> "188 linear_8"  [label="(2304,)", style=solid];
-"188 linear_8" -> "189 unflatten_2"  [label="(197, 1, 2304)", style=solid];
-"189 unflatten_2" -> "190 unsqueeze_2"  [label="(197, 1, 3, 768)", style=solid];
-"190 unsqueeze_2" -> "191 transpose_13"  [label="(1, 197, 1, 3, 768)", style=solid];
-"191 transpose_13" -> "192 squeeze_2"  [label="(3, 197, 1, 1, 768)", style=solid];
-"192 squeeze_2" -> "193 contiguous_2"  [label="(3, 197, 1, 768)", style=solid];
-"193 contiguous_2" -> "194 quantize_per_tensor_default_14"  [label="(3, 197, 1, 768)", style=solid];
-"193 contiguous_2" -> "197 quantize_per_tensor_default_15"  [label="(3, 197, 1, 768)", style=solid];
-"193 contiguous_2" -> "200 select_8"  [label="(3, 197, 1, 768)", style=solid];
-"194 quantize_per_tensor_default_14" -> "195 dequantize_per_tensor_default_14"  [label="(3, 197, 1, 768)", style=solid];
-"195 dequantize_per_tensor_default_14" -> "196 select_6"  [label="(3, 197, 1, 768)", style=solid];
-"196 select_6" -> "201 view_16"  [label="(197, 1, 768)", style=solid];
-"197 quantize_per_tensor_default_15" -> "198 dequantize_per_tensor_default_15"  [label="(3, 197, 1, 768)", style=solid];
-"198 dequantize_per_tensor_default_15" -> "199 select_7"  [label="(3, 197, 1, 768)", style=solid];
-"199 select_7" -> "203 view_17"  [label="(197, 1, 768)", style=solid];
-"200 select_8" -> "205 view_18"  [label="(197, 1, 768)", style=solid];
-"201 view_16" -> "202 transpose_14"  [label="(197, 12, 64)", style=solid];
-"202 transpose_14" -> "207 view_19"  [label="(12, 197, 64)", style=solid];
-"203 view_17" -> "204 transpose_15"  [label="(197, 12, 64)", style=solid];
-"204 transpose_15" -> "208 view_20"  [label="(12, 197, 64)", style=solid];
-"205 view_18" -> "206 transpose_16"  [label="(197, 12, 64)", style=solid];
-"206 transpose_16" -> "209 view_21"  [label="(12, 197, 64)", style=solid];
-"207 view_19" -> "210 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
-"208 view_20" -> "210 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
-"209 view_21" -> "210 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
-"210 scaled_dot_product_attention_2" -> "211 permute_3"  [label="(1, 12, 197, 64)", style=solid];
-"211 permute_3" -> "212 view_22"  [label="(197, 1, 12, 64)", style=solid];
-"212 view_22" -> "213 view_22_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"213 view_22_0_0_nncf_smooth_quant_0" -> "214 quantize_per_tensor_default_16"  [label="(197, 768)", style=solid];
-"214 quantize_per_tensor_default_16" -> "215 dequantize_per_tensor_default_16"  [label="(197, 768)", style=solid];
-"215 dequantize_per_tensor_default_16" -> "222 linear_9"  [label="(197, 768)", style=solid];
-"216 linear_9_scale_0" -> "219 quantize_per_channel_default_10"  [label="(768,)", style=solid];
-"216 linear_9_scale_0" -> "220 dequantize_per_channel_default_10"  [label="(768,)", style=solid];
-"217 linear_9_zero_point_0" -> "219 quantize_per_channel_default_10"  [label="(768,)", style=solid];
-"217 linear_9_zero_point_0" -> "220 dequantize_per_channel_default_10"  [label="(768,)", style=solid];
-"218 compressed_weight_updated_constant10" -> "219 quantize_per_channel_default_10"  [label="(768, 768)", style=solid];
-"219 quantize_per_channel_default_10" -> "220 dequantize_per_channel_default_10"  [label="(768, 768)", style=solid];
-"220 dequantize_per_channel_default_10" -> "222 linear_9"  [label="(768, 768)", style=solid];
-"221 _param_constant33_0_0" -> "222 linear_9"  [label="(768,)", style=solid];
-"222 linear_9" -> "223 view_23"  [label="(197, 768)", style=solid];
-"223 view_23" -> "224 transpose_17"  [label="(197, 1, 768)", style=solid];
-"224 transpose_17" -> "225 dropout_7"  [label="(1, 197, 768)", style=solid];
-"225 dropout_7" -> "226 add_5"  [label="(1, 197, 768)", style=solid];
-"226 add_5" -> "229 layer_norm_5"  [label="(1, 197, 768)", style=solid];
-"226 add_5" -> "253 add_6"  [label="(1, 197, 768)", style=solid];
-"227 _param_constant34" -> "229 layer_norm_5"  [label="(768,)", style=solid];
-"228 _param_constant35" -> "229 layer_norm_5"  [label="(768,)", style=solid];
-"229 layer_norm_5" -> "230 layer_norm_5_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"230 layer_norm_5_0_0_nncf_smooth_quant_0" -> "231 quantize_per_tensor_default_17"  [label="(1, 197, 768)", style=solid];
-"231 quantize_per_tensor_default_17" -> "232 dequantize_per_tensor_default_17"  [label="(1, 197, 768)", style=solid];
-"232 dequantize_per_tensor_default_17" -> "239 linear_10"  [label="(1, 197, 768)", style=solid];
-"233 linear_10_scale_0" -> "236 quantize_per_channel_default_11"  [label="(3072,)", style=solid];
-"233 linear_10_scale_0" -> "237 dequantize_per_channel_default_11"  [label="(3072,)", style=solid];
-"234 linear_10_zero_point_0" -> "236 quantize_per_channel_default_11"  [label="(3072,)", style=solid];
-"234 linear_10_zero_point_0" -> "237 dequantize_per_channel_default_11"  [label="(3072,)", style=solid];
-"235 compressed_weight_updated_constant11" -> "236 quantize_per_channel_default_11"  [label="(3072, 768)", style=solid];
-"236 quantize_per_channel_default_11" -> "237 dequantize_per_channel_default_11"  [label="(3072, 768)", style=solid];
-"237 dequantize_per_channel_default_11" -> "239 linear_10"  [label="(3072, 768)", style=solid];
-"238 _param_constant37_0_0" -> "239 linear_10"  [label="(3072,)", style=solid];
-"239 linear_10" -> "240 gelu_2"  [label="(1, 197, 3072)", style=solid];
-"240 gelu_2" -> "241 dropout_8"  [label="(1, 197, 3072)", style=solid];
-"241 dropout_8" -> "242 dropout_8_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"242 dropout_8_0_0_nncf_smooth_quant_0" -> "243 quantize_per_tensor_default_18"  [label="(1, 197, 3072)", style=solid];
-"243 quantize_per_tensor_default_18" -> "244 dequantize_per_tensor_default_18"  [label="(1, 197, 3072)", style=solid];
-"244 dequantize_per_tensor_default_18" -> "251 linear_11"  [label="(1, 197, 3072)", style=solid];
-"245 linear_11_scale_0" -> "248 quantize_per_channel_default_12"  [label="(768,)", style=solid];
-"245 linear_11_scale_0" -> "249 dequantize_per_channel_default_12"  [label="(768,)", style=solid];
-"246 linear_11_zero_point_0" -> "248 quantize_per_channel_default_12"  [label="(768,)", style=solid];
-"246 linear_11_zero_point_0" -> "249 dequantize_per_channel_default_12"  [label="(768,)", style=solid];
-"247 compressed_weight_updated_constant12" -> "248 quantize_per_channel_default_12"  [label="(768, 3072)", style=solid];
-"248 quantize_per_channel_default_12" -> "249 dequantize_per_channel_default_12"  [label="(768, 3072)", style=solid];
-"249 dequantize_per_channel_default_12" -> "251 linear_11"  [label="(768, 3072)", style=solid];
-"250 _param_constant39_0_0" -> "251 linear_11"  [label="(768,)", style=solid];
-"251 linear_11" -> "252 dropout_9"  [label="(1, 197, 768)", style=solid];
-"252 dropout_9" -> "253 add_6"  [label="(1, 197, 768)", style=solid];
-"253 add_6" -> "256 layer_norm_6"  [label="(1, 197, 768)", style=solid];
-"253 add_6" -> "305 add_7"  [label="(1, 197, 768)", style=solid];
-"254 _param_constant40" -> "256 layer_norm_6"  [label="(768,)", style=solid];
-"255 _param_constant41" -> "256 layer_norm_6"  [label="(768,)", style=solid];
-"256 layer_norm_6" -> "257 transpose_18"  [label="(1, 197, 768)", style=solid];
-"257 transpose_18" -> "258 transpose_18_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"258 transpose_18_0_0_nncf_smooth_quant_0" -> "259 quantize_per_tensor_default_19"  [label="(197, 1, 768)", style=solid];
-"259 quantize_per_tensor_default_19" -> "260 dequantize_per_tensor_default_19"  [label="(197, 1, 768)", style=solid];
-"260 dequantize_per_tensor_default_19" -> "267 linear_12"  [label="(197, 1, 768)", style=solid];
-"261 linear_12_scale_0" -> "264 quantize_per_channel_default_13"  [label="(2304,)", style=solid];
-"261 linear_12_scale_0" -> "265 dequantize_per_channel_default_13"  [label="(2304,)", style=solid];
-"262 linear_12_zero_point_0" -> "264 quantize_per_channel_default_13"  [label="(2304,)", style=solid];
-"262 linear_12_zero_point_0" -> "265 dequantize_per_channel_default_13"  [label="(2304,)", style=solid];
-"263 compressed_weight_updated_constant13" -> "264 quantize_per_channel_default_13"  [label="(2304, 768)", style=solid];
-"264 quantize_per_channel_default_13" -> "265 dequantize_per_channel_default_13"  [label="(2304, 768)", style=solid];
-"265 dequantize_per_channel_default_13" -> "267 linear_12"  [label="(2304, 768)", style=solid];
-"266 _param_constant43_0_0" -> "267 linear_12"  [label="(2304,)", style=solid];
-"267 linear_12" -> "268 unflatten_3"  [label="(197, 1, 2304)", style=solid];
-"268 unflatten_3" -> "269 unsqueeze_3"  [label="(197, 1, 3, 768)", style=solid];
-"269 unsqueeze_3" -> "270 transpose_19"  [label="(1, 197, 1, 3, 768)", style=solid];
-"270 transpose_19" -> "271 squeeze_3"  [label="(3, 197, 1, 1, 768)", style=solid];
-"271 squeeze_3" -> "272 contiguous_3"  [label="(3, 197, 1, 768)", style=solid];
-"272 contiguous_3" -> "273 quantize_per_tensor_default_20"  [label="(3, 197, 1, 768)", style=solid];
-"272 contiguous_3" -> "276 quantize_per_tensor_default_21"  [label="(3, 197, 1, 768)", style=solid];
-"272 contiguous_3" -> "279 select_11"  [label="(3, 197, 1, 768)", style=solid];
-"273 quantize_per_tensor_default_20" -> "274 dequantize_per_tensor_default_20"  [label="(3, 197, 1, 768)", style=solid];
-"274 dequantize_per_tensor_default_20" -> "275 select_9"  [label="(3, 197, 1, 768)", style=solid];
-"275 select_9" -> "280 view_24"  [label="(197, 1, 768)", style=solid];
-"276 quantize_per_tensor_default_21" -> "277 dequantize_per_tensor_default_21"  [label="(3, 197, 1, 768)", style=solid];
-"277 dequantize_per_tensor_default_21" -> "278 select_10"  [label="(3, 197, 1, 768)", style=solid];
-"278 select_10" -> "282 view_25"  [label="(197, 1, 768)", style=solid];
-"279 select_11" -> "284 view_26"  [label="(197, 1, 768)", style=solid];
-"280 view_24" -> "281 transpose_20"  [label="(197, 12, 64)", style=solid];
-"281 transpose_20" -> "286 view_27"  [label="(12, 197, 64)", style=solid];
-"282 view_25" -> "283 transpose_21"  [label="(197, 12, 64)", style=solid];
-"283 transpose_21" -> "287 view_28"  [label="(12, 197, 64)", style=solid];
-"284 view_26" -> "285 transpose_22"  [label="(197, 12, 64)", style=solid];
-"285 transpose_22" -> "288 view_29"  [label="(12, 197, 64)", style=solid];
-"286 view_27" -> "289 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
-"287 view_28" -> "289 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
-"288 view_29" -> "289 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
-"289 scaled_dot_product_attention_3" -> "290 permute_4"  [label="(1, 12, 197, 64)", style=solid];
-"290 permute_4" -> "291 view_30"  [label="(197, 1, 12, 64)", style=solid];
-"291 view_30" -> "292 view_30_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"292 view_30_0_0_nncf_smooth_quant_0" -> "293 quantize_per_tensor_default_22"  [label="(197, 768)", style=solid];
-"293 quantize_per_tensor_default_22" -> "294 dequantize_per_tensor_default_22"  [label="(197, 768)", style=solid];
-"294 dequantize_per_tensor_default_22" -> "301 linear_13"  [label="(197, 768)", style=solid];
-"295 linear_13_scale_0" -> "298 quantize_per_channel_default_14"  [label="(768,)", style=solid];
-"295 linear_13_scale_0" -> "299 dequantize_per_channel_default_14"  [label="(768,)", style=solid];
-"296 linear_13_zero_point_0" -> "298 quantize_per_channel_default_14"  [label="(768,)", style=solid];
-"296 linear_13_zero_point_0" -> "299 dequantize_per_channel_default_14"  [label="(768,)", style=solid];
-"297 compressed_weight_updated_constant14" -> "298 quantize_per_channel_default_14"  [label="(768, 768)", style=solid];
-"298 quantize_per_channel_default_14" -> "299 dequantize_per_channel_default_14"  [label="(768, 768)", style=solid];
-"299 dequantize_per_channel_default_14" -> "301 linear_13"  [label="(768, 768)", style=solid];
-"300 _param_constant45_0_0" -> "301 linear_13"  [label="(768,)", style=solid];
-"301 linear_13" -> "302 view_31"  [label="(197, 768)", style=solid];
-"302 view_31" -> "303 transpose_23"  [label="(197, 1, 768)", style=solid];
-"303 transpose_23" -> "304 dropout_10"  [label="(1, 197, 768)", style=solid];
-"304 dropout_10" -> "305 add_7"  [label="(1, 197, 768)", style=solid];
-"305 add_7" -> "308 layer_norm_7"  [label="(1, 197, 768)", style=solid];
-"305 add_7" -> "332 add_8"  [label="(1, 197, 768)", style=solid];
-"306 _param_constant46" -> "308 layer_norm_7"  [label="(768,)", style=solid];
-"307 _param_constant47" -> "308 layer_norm_7"  [label="(768,)", style=solid];
-"308 layer_norm_7" -> "309 layer_norm_7_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"309 layer_norm_7_0_0_nncf_smooth_quant_0" -> "310 quantize_per_tensor_default_23"  [label="(1, 197, 768)", style=solid];
-"310 quantize_per_tensor_default_23" -> "311 dequantize_per_tensor_default_23"  [label="(1, 197, 768)", style=solid];
-"311 dequantize_per_tensor_default_23" -> "318 linear_14"  [label="(1, 197, 768)", style=solid];
-"312 linear_14_scale_0" -> "315 quantize_per_channel_default_15"  [label="(3072,)", style=solid];
-"312 linear_14_scale_0" -> "316 dequantize_per_channel_default_15"  [label="(3072,)", style=solid];
-"313 linear_14_zero_point_0" -> "315 quantize_per_channel_default_15"  [label="(3072,)", style=solid];
-"313 linear_14_zero_point_0" -> "316 dequantize_per_channel_default_15"  [label="(3072,)", style=solid];
-"314 compressed_weight_updated_constant15" -> "315 quantize_per_channel_default_15"  [label="(3072, 768)", style=solid];
-"315 quantize_per_channel_default_15" -> "316 dequantize_per_channel_default_15"  [label="(3072, 768)", style=solid];
-"316 dequantize_per_channel_default_15" -> "318 linear_14"  [label="(3072, 768)", style=solid];
-"317 _param_constant49_0_0" -> "318 linear_14"  [label="(3072,)", style=solid];
-"318 linear_14" -> "319 gelu_3"  [label="(1, 197, 3072)", style=solid];
-"319 gelu_3" -> "320 dropout_11"  [label="(1, 197, 3072)", style=solid];
-"320 dropout_11" -> "321 dropout_11_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"321 dropout_11_0_0_nncf_smooth_quant_0" -> "322 quantize_per_tensor_default_24"  [label="(1, 197, 3072)", style=solid];
-"322 quantize_per_tensor_default_24" -> "323 dequantize_per_tensor_default_24"  [label="(1, 197, 3072)", style=solid];
-"323 dequantize_per_tensor_default_24" -> "330 linear_15"  [label="(1, 197, 3072)", style=solid];
-"324 linear_15_scale_0" -> "327 quantize_per_channel_default_16"  [label="(768,)", style=solid];
-"324 linear_15_scale_0" -> "328 dequantize_per_channel_default_16"  [label="(768,)", style=solid];
-"325 linear_15_zero_point_0" -> "327 quantize_per_channel_default_16"  [label="(768,)", style=solid];
-"325 linear_15_zero_point_0" -> "328 dequantize_per_channel_default_16"  [label="(768,)", style=solid];
-"326 compressed_weight_updated_constant16" -> "327 quantize_per_channel_default_16"  [label="(768, 3072)", style=solid];
-"327 quantize_per_channel_default_16" -> "328 dequantize_per_channel_default_16"  [label="(768, 3072)", style=solid];
-"328 dequantize_per_channel_default_16" -> "330 linear_15"  [label="(768, 3072)", style=solid];
-"329 _param_constant51_0_0" -> "330 linear_15"  [label="(768,)", style=solid];
-"330 linear_15" -> "331 dropout_12"  [label="(1, 197, 768)", style=solid];
-"331 dropout_12" -> "332 add_8"  [label="(1, 197, 768)", style=solid];
-"332 add_8" -> "335 layer_norm_8"  [label="(1, 197, 768)", style=solid];
-"332 add_8" -> "384 add_9"  [label="(1, 197, 768)", style=solid];
-"333 _param_constant52" -> "335 layer_norm_8"  [label="(768,)", style=solid];
-"334 _param_constant53" -> "335 layer_norm_8"  [label="(768,)", style=solid];
-"335 layer_norm_8" -> "336 transpose_24"  [label="(1, 197, 768)", style=solid];
-"336 transpose_24" -> "337 transpose_24_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"337 transpose_24_0_0_nncf_smooth_quant_0" -> "338 quantize_per_tensor_default_25"  [label="(197, 1, 768)", style=solid];
-"338 quantize_per_tensor_default_25" -> "339 dequantize_per_tensor_default_25"  [label="(197, 1, 768)", style=solid];
-"339 dequantize_per_tensor_default_25" -> "346 linear_16"  [label="(197, 1, 768)", style=solid];
-"340 linear_16_scale_0" -> "343 quantize_per_channel_default_17"  [label="(2304,)", style=solid];
-"340 linear_16_scale_0" -> "344 dequantize_per_channel_default_17"  [label="(2304,)", style=solid];
-"341 linear_16_zero_point_0" -> "343 quantize_per_channel_default_17"  [label="(2304,)", style=solid];
-"341 linear_16_zero_point_0" -> "344 dequantize_per_channel_default_17"  [label="(2304,)", style=solid];
-"342 compressed_weight_updated_constant17" -> "343 quantize_per_channel_default_17"  [label="(2304, 768)", style=solid];
-"343 quantize_per_channel_default_17" -> "344 dequantize_per_channel_default_17"  [label="(2304, 768)", style=solid];
-"344 dequantize_per_channel_default_17" -> "346 linear_16"  [label="(2304, 768)", style=solid];
-"345 _param_constant55_0_0" -> "346 linear_16"  [label="(2304,)", style=solid];
-"346 linear_16" -> "347 unflatten_4"  [label="(197, 1, 2304)", style=solid];
-"347 unflatten_4" -> "348 unsqueeze_4"  [label="(197, 1, 3, 768)", style=solid];
-"348 unsqueeze_4" -> "349 transpose_25"  [label="(1, 197, 1, 3, 768)", style=solid];
-"349 transpose_25" -> "350 squeeze_4"  [label="(3, 197, 1, 1, 768)", style=solid];
-"350 squeeze_4" -> "351 contiguous_4"  [label="(3, 197, 1, 768)", style=solid];
-"351 contiguous_4" -> "352 quantize_per_tensor_default_26"  [label="(3, 197, 1, 768)", style=solid];
-"351 contiguous_4" -> "355 quantize_per_tensor_default_27"  [label="(3, 197, 1, 768)", style=solid];
-"351 contiguous_4" -> "358 select_14"  [label="(3, 197, 1, 768)", style=solid];
-"352 quantize_per_tensor_default_26" -> "353 dequantize_per_tensor_default_26"  [label="(3, 197, 1, 768)", style=solid];
-"353 dequantize_per_tensor_default_26" -> "354 select_12"  [label="(3, 197, 1, 768)", style=solid];
-"354 select_12" -> "359 view_32"  [label="(197, 1, 768)", style=solid];
-"355 quantize_per_tensor_default_27" -> "356 dequantize_per_tensor_default_27"  [label="(3, 197, 1, 768)", style=solid];
-"356 dequantize_per_tensor_default_27" -> "357 select_13"  [label="(3, 197, 1, 768)", style=solid];
-"357 select_13" -> "361 view_33"  [label="(197, 1, 768)", style=solid];
-"358 select_14" -> "363 view_34"  [label="(197, 1, 768)", style=solid];
-"359 view_32" -> "360 transpose_26"  [label="(197, 12, 64)", style=solid];
-"360 transpose_26" -> "365 view_35"  [label="(12, 197, 64)", style=solid];
-"361 view_33" -> "362 transpose_27"  [label="(197, 12, 64)", style=solid];
-"362 transpose_27" -> "366 view_36"  [label="(12, 197, 64)", style=solid];
-"363 view_34" -> "364 transpose_28"  [label="(197, 12, 64)", style=solid];
-"364 transpose_28" -> "367 view_37"  [label="(12, 197, 64)", style=solid];
-"365 view_35" -> "368 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
-"366 view_36" -> "368 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
-"367 view_37" -> "368 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
-"368 scaled_dot_product_attention_4" -> "369 permute_5"  [label="(1, 12, 197, 64)", style=solid];
-"369 permute_5" -> "370 view_38"  [label="(197, 1, 12, 64)", style=solid];
-"370 view_38" -> "371 view_38_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"371 view_38_0_0_nncf_smooth_quant_0" -> "372 quantize_per_tensor_default_28"  [label="(197, 768)", style=solid];
-"372 quantize_per_tensor_default_28" -> "373 dequantize_per_tensor_default_28"  [label="(197, 768)", style=solid];
-"373 dequantize_per_tensor_default_28" -> "380 linear_17"  [label="(197, 768)", style=solid];
-"374 linear_17_scale_0" -> "377 quantize_per_channel_default_18"  [label="(768,)", style=solid];
-"374 linear_17_scale_0" -> "378 dequantize_per_channel_default_18"  [label="(768,)", style=solid];
-"375 linear_17_zero_point_0" -> "377 quantize_per_channel_default_18"  [label="(768,)", style=solid];
-"375 linear_17_zero_point_0" -> "378 dequantize_per_channel_default_18"  [label="(768,)", style=solid];
-"376 compressed_weight_updated_constant18" -> "377 quantize_per_channel_default_18"  [label="(768, 768)", style=solid];
-"377 quantize_per_channel_default_18" -> "378 dequantize_per_channel_default_18"  [label="(768, 768)", style=solid];
-"378 dequantize_per_channel_default_18" -> "380 linear_17"  [label="(768, 768)", style=solid];
-"379 _param_constant57_0_0" -> "380 linear_17"  [label="(768,)", style=solid];
-"380 linear_17" -> "381 view_39"  [label="(197, 768)", style=solid];
-"381 view_39" -> "382 transpose_29"  [label="(197, 1, 768)", style=solid];
-"382 transpose_29" -> "383 dropout_13"  [label="(1, 197, 768)", style=solid];
-"383 dropout_13" -> "384 add_9"  [label="(1, 197, 768)", style=solid];
-"384 add_9" -> "387 layer_norm_9"  [label="(1, 197, 768)", style=solid];
-"384 add_9" -> "411 add_10"  [label="(1, 197, 768)", style=solid];
-"385 _param_constant58" -> "387 layer_norm_9"  [label="(768,)", style=solid];
-"386 _param_constant59" -> "387 layer_norm_9"  [label="(768,)", style=solid];
-"387 layer_norm_9" -> "388 layer_norm_9_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"388 layer_norm_9_0_0_nncf_smooth_quant_0" -> "389 quantize_per_tensor_default_29"  [label="(1, 197, 768)", style=solid];
-"389 quantize_per_tensor_default_29" -> "390 dequantize_per_tensor_default_29"  [label="(1, 197, 768)", style=solid];
-"390 dequantize_per_tensor_default_29" -> "397 linear_18"  [label="(1, 197, 768)", style=solid];
-"391 linear_18_scale_0" -> "394 quantize_per_channel_default_19"  [label="(3072,)", style=solid];
-"391 linear_18_scale_0" -> "395 dequantize_per_channel_default_19"  [label="(3072,)", style=solid];
-"392 linear_18_zero_point_0" -> "394 quantize_per_channel_default_19"  [label="(3072,)", style=solid];
-"392 linear_18_zero_point_0" -> "395 dequantize_per_channel_default_19"  [label="(3072,)", style=solid];
-"393 compressed_weight_updated_constant19" -> "394 quantize_per_channel_default_19"  [label="(3072, 768)", style=solid];
-"394 quantize_per_channel_default_19" -> "395 dequantize_per_channel_default_19"  [label="(3072, 768)", style=solid];
-"395 dequantize_per_channel_default_19" -> "397 linear_18"  [label="(3072, 768)", style=solid];
-"396 _param_constant61_0_0" -> "397 linear_18"  [label="(3072,)", style=solid];
-"397 linear_18" -> "398 gelu_4"  [label="(1, 197, 3072)", style=solid];
-"398 gelu_4" -> "399 dropout_14"  [label="(1, 197, 3072)", style=solid];
-"399 dropout_14" -> "400 dropout_14_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"400 dropout_14_0_0_nncf_smooth_quant_0" -> "401 quantize_per_tensor_default_30"  [label="(1, 197, 3072)", style=solid];
-"401 quantize_per_tensor_default_30" -> "402 dequantize_per_tensor_default_30"  [label="(1, 197, 3072)", style=solid];
-"402 dequantize_per_tensor_default_30" -> "409 linear_19"  [label="(1, 197, 3072)", style=solid];
-"403 linear_19_scale_0" -> "406 quantize_per_channel_default_20"  [label="(768,)", style=solid];
-"403 linear_19_scale_0" -> "407 dequantize_per_channel_default_20"  [label="(768,)", style=solid];
-"404 linear_19_zero_point_0" -> "406 quantize_per_channel_default_20"  [label="(768,)", style=solid];
-"404 linear_19_zero_point_0" -> "407 dequantize_per_channel_default_20"  [label="(768,)", style=solid];
-"405 compressed_weight_updated_constant20" -> "406 quantize_per_channel_default_20"  [label="(768, 3072)", style=solid];
-"406 quantize_per_channel_default_20" -> "407 dequantize_per_channel_default_20"  [label="(768, 3072)", style=solid];
-"407 dequantize_per_channel_default_20" -> "409 linear_19"  [label="(768, 3072)", style=solid];
-"408 _param_constant63_0_0" -> "409 linear_19"  [label="(768,)", style=solid];
-"409 linear_19" -> "410 dropout_15"  [label="(1, 197, 768)", style=solid];
-"410 dropout_15" -> "411 add_10"  [label="(1, 197, 768)", style=solid];
-"411 add_10" -> "414 layer_norm_10"  [label="(1, 197, 768)", style=solid];
-"411 add_10" -> "463 add_11"  [label="(1, 197, 768)", style=solid];
-"412 _param_constant64" -> "414 layer_norm_10"  [label="(768,)", style=solid];
-"413 _param_constant65" -> "414 layer_norm_10"  [label="(768,)", style=solid];
-"414 layer_norm_10" -> "415 transpose_30"  [label="(1, 197, 768)", style=solid];
-"415 transpose_30" -> "416 transpose_30_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"416 transpose_30_0_0_nncf_smooth_quant_0" -> "417 quantize_per_tensor_default_31"  [label="(197, 1, 768)", style=solid];
-"417 quantize_per_tensor_default_31" -> "418 dequantize_per_tensor_default_31"  [label="(197, 1, 768)", style=solid];
-"418 dequantize_per_tensor_default_31" -> "425 linear_20"  [label="(197, 1, 768)", style=solid];
-"419 linear_20_scale_0" -> "422 quantize_per_channel_default_21"  [label="(2304,)", style=solid];
-"419 linear_20_scale_0" -> "423 dequantize_per_channel_default_21"  [label="(2304,)", style=solid];
-"420 linear_20_zero_point_0" -> "422 quantize_per_channel_default_21"  [label="(2304,)", style=solid];
-"420 linear_20_zero_point_0" -> "423 dequantize_per_channel_default_21"  [label="(2304,)", style=solid];
-"421 compressed_weight_updated_constant21" -> "422 quantize_per_channel_default_21"  [label="(2304, 768)", style=solid];
-"422 quantize_per_channel_default_21" -> "423 dequantize_per_channel_default_21"  [label="(2304, 768)", style=solid];
-"423 dequantize_per_channel_default_21" -> "425 linear_20"  [label="(2304, 768)", style=solid];
-"424 _param_constant67_0_0" -> "425 linear_20"  [label="(2304,)", style=solid];
-"425 linear_20" -> "426 unflatten_5"  [label="(197, 1, 2304)", style=solid];
-"426 unflatten_5" -> "427 unsqueeze_5"  [label="(197, 1, 3, 768)", style=solid];
-"427 unsqueeze_5" -> "428 transpose_31"  [label="(1, 197, 1, 3, 768)", style=solid];
-"428 transpose_31" -> "429 squeeze_5"  [label="(3, 197, 1, 1, 768)", style=solid];
-"429 squeeze_5" -> "430 contiguous_5"  [label="(3, 197, 1, 768)", style=solid];
-"430 contiguous_5" -> "431 quantize_per_tensor_default_32"  [label="(3, 197, 1, 768)", style=solid];
-"430 contiguous_5" -> "434 quantize_per_tensor_default_33"  [label="(3, 197, 1, 768)", style=solid];
-"430 contiguous_5" -> "437 select_17"  [label="(3, 197, 1, 768)", style=solid];
-"431 quantize_per_tensor_default_32" -> "432 dequantize_per_tensor_default_32"  [label="(3, 197, 1, 768)", style=solid];
-"432 dequantize_per_tensor_default_32" -> "433 select_15"  [label="(3, 197, 1, 768)", style=solid];
-"433 select_15" -> "438 view_40"  [label="(197, 1, 768)", style=solid];
-"434 quantize_per_tensor_default_33" -> "435 dequantize_per_tensor_default_33"  [label="(3, 197, 1, 768)", style=solid];
-"435 dequantize_per_tensor_default_33" -> "436 select_16"  [label="(3, 197, 1, 768)", style=solid];
-"436 select_16" -> "440 view_41"  [label="(197, 1, 768)", style=solid];
-"437 select_17" -> "442 view_42"  [label="(197, 1, 768)", style=solid];
-"438 view_40" -> "439 transpose_32"  [label="(197, 12, 64)", style=solid];
-"439 transpose_32" -> "444 view_43"  [label="(12, 197, 64)", style=solid];
-"440 view_41" -> "441 transpose_33"  [label="(197, 12, 64)", style=solid];
-"441 transpose_33" -> "445 view_44"  [label="(12, 197, 64)", style=solid];
-"442 view_42" -> "443 transpose_34"  [label="(197, 12, 64)", style=solid];
-"443 transpose_34" -> "446 view_45"  [label="(12, 197, 64)", style=solid];
-"444 view_43" -> "447 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
-"445 view_44" -> "447 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
-"446 view_45" -> "447 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
-"447 scaled_dot_product_attention_5" -> "448 permute_6"  [label="(1, 12, 197, 64)", style=solid];
-"448 permute_6" -> "449 view_46"  [label="(197, 1, 12, 64)", style=solid];
-"449 view_46" -> "450 view_46_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"450 view_46_0_0_nncf_smooth_quant_0" -> "451 quantize_per_tensor_default_34"  [label="(197, 768)", style=solid];
-"451 quantize_per_tensor_default_34" -> "452 dequantize_per_tensor_default_34"  [label="(197, 768)", style=solid];
-"452 dequantize_per_tensor_default_34" -> "459 linear_21"  [label="(197, 768)", style=solid];
-"453 linear_21_scale_0" -> "456 quantize_per_channel_default_22"  [label="(768,)", style=solid];
-"453 linear_21_scale_0" -> "457 dequantize_per_channel_default_22"  [label="(768,)", style=solid];
-"454 linear_21_zero_point_0" -> "456 quantize_per_channel_default_22"  [label="(768,)", style=solid];
-"454 linear_21_zero_point_0" -> "457 dequantize_per_channel_default_22"  [label="(768,)", style=solid];
-"455 compressed_weight_updated_constant22" -> "456 quantize_per_channel_default_22"  [label="(768, 768)", style=solid];
-"456 quantize_per_channel_default_22" -> "457 dequantize_per_channel_default_22"  [label="(768, 768)", style=solid];
-"457 dequantize_per_channel_default_22" -> "459 linear_21"  [label="(768, 768)", style=solid];
-"458 _param_constant69_0_0" -> "459 linear_21"  [label="(768,)", style=solid];
-"459 linear_21" -> "460 view_47"  [label="(197, 768)", style=solid];
-"460 view_47" -> "461 transpose_35"  [label="(197, 1, 768)", style=solid];
-"461 transpose_35" -> "462 dropout_16"  [label="(1, 197, 768)", style=solid];
-"462 dropout_16" -> "463 add_11"  [label="(1, 197, 768)", style=solid];
-"463 add_11" -> "466 layer_norm_11"  [label="(1, 197, 768)", style=solid];
-"463 add_11" -> "490 add_12"  [label="(1, 197, 768)", style=solid];
-"464 _param_constant70" -> "466 layer_norm_11"  [label="(768,)", style=solid];
-"465 _param_constant71" -> "466 layer_norm_11"  [label="(768,)", style=solid];
-"466 layer_norm_11" -> "467 layer_norm_11_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"467 layer_norm_11_0_0_nncf_smooth_quant_0" -> "468 quantize_per_tensor_default_35"  [label="(1, 197, 768)", style=solid];
-"468 quantize_per_tensor_default_35" -> "469 dequantize_per_tensor_default_35"  [label="(1, 197, 768)", style=solid];
-"469 dequantize_per_tensor_default_35" -> "476 linear_22"  [label="(1, 197, 768)", style=solid];
-"470 linear_22_scale_0" -> "473 quantize_per_channel_default_23"  [label="(3072,)", style=solid];
-"470 linear_22_scale_0" -> "474 dequantize_per_channel_default_23"  [label="(3072,)", style=solid];
-"471 linear_22_zero_point_0" -> "473 quantize_per_channel_default_23"  [label="(3072,)", style=solid];
-"471 linear_22_zero_point_0" -> "474 dequantize_per_channel_default_23"  [label="(3072,)", style=solid];
-"472 compressed_weight_updated_constant23" -> "473 quantize_per_channel_default_23"  [label="(3072, 768)", style=solid];
-"473 quantize_per_channel_default_23" -> "474 dequantize_per_channel_default_23"  [label="(3072, 768)", style=solid];
-"474 dequantize_per_channel_default_23" -> "476 linear_22"  [label="(3072, 768)", style=solid];
-"475 _param_constant73_0_0" -> "476 linear_22"  [label="(3072,)", style=solid];
-"476 linear_22" -> "477 gelu_5"  [label="(1, 197, 3072)", style=solid];
-"477 gelu_5" -> "478 dropout_17"  [label="(1, 197, 3072)", style=solid];
-"478 dropout_17" -> "479 dropout_17_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"479 dropout_17_0_0_nncf_smooth_quant_0" -> "480 quantize_per_tensor_default_36"  [label="(1, 197, 3072)", style=solid];
-"480 quantize_per_tensor_default_36" -> "481 dequantize_per_tensor_default_36"  [label="(1, 197, 3072)", style=solid];
-"481 dequantize_per_tensor_default_36" -> "488 linear_23"  [label="(1, 197, 3072)", style=solid];
-"482 linear_23_scale_0" -> "485 quantize_per_channel_default_24"  [label="(768,)", style=solid];
-"482 linear_23_scale_0" -> "486 dequantize_per_channel_default_24"  [label="(768,)", style=solid];
-"483 linear_23_zero_point_0" -> "485 quantize_per_channel_default_24"  [label="(768,)", style=solid];
-"483 linear_23_zero_point_0" -> "486 dequantize_per_channel_default_24"  [label="(768,)", style=solid];
-"484 compressed_weight_updated_constant24" -> "485 quantize_per_channel_default_24"  [label="(768, 3072)", style=solid];
-"485 quantize_per_channel_default_24" -> "486 dequantize_per_channel_default_24"  [label="(768, 3072)", style=solid];
-"486 dequantize_per_channel_default_24" -> "488 linear_23"  [label="(768, 3072)", style=solid];
-"487 _param_constant75_0_0" -> "488 linear_23"  [label="(768,)", style=solid];
-"488 linear_23" -> "489 dropout_18"  [label="(1, 197, 768)", style=solid];
-"489 dropout_18" -> "490 add_12"  [label="(1, 197, 768)", style=solid];
-"490 add_12" -> "493 layer_norm_12"  [label="(1, 197, 768)", style=solid];
-"490 add_12" -> "542 add_13"  [label="(1, 197, 768)", style=solid];
-"491 _param_constant76" -> "493 layer_norm_12"  [label="(768,)", style=solid];
-"492 _param_constant77" -> "493 layer_norm_12"  [label="(768,)", style=solid];
-"493 layer_norm_12" -> "494 transpose_36"  [label="(1, 197, 768)", style=solid];
-"494 transpose_36" -> "495 transpose_36_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"495 transpose_36_0_0_nncf_smooth_quant_0" -> "496 quantize_per_tensor_default_37"  [label="(197, 1, 768)", style=solid];
-"496 quantize_per_tensor_default_37" -> "497 dequantize_per_tensor_default_37"  [label="(197, 1, 768)", style=solid];
-"497 dequantize_per_tensor_default_37" -> "504 linear_24"  [label="(197, 1, 768)", style=solid];
-"498 linear_24_scale_0" -> "501 quantize_per_channel_default_25"  [label="(2304,)", style=solid];
-"498 linear_24_scale_0" -> "502 dequantize_per_channel_default_25"  [label="(2304,)", style=solid];
-"499 linear_24_zero_point_0" -> "501 quantize_per_channel_default_25"  [label="(2304,)", style=solid];
-"499 linear_24_zero_point_0" -> "502 dequantize_per_channel_default_25"  [label="(2304,)", style=solid];
-"500 compressed_weight_updated_constant25" -> "501 quantize_per_channel_default_25"  [label="(2304, 768)", style=solid];
-"501 quantize_per_channel_default_25" -> "502 dequantize_per_channel_default_25"  [label="(2304, 768)", style=solid];
-"502 dequantize_per_channel_default_25" -> "504 linear_24"  [label="(2304, 768)", style=solid];
-"503 _param_constant79_0_0" -> "504 linear_24"  [label="(2304,)", style=solid];
-"504 linear_24" -> "505 unflatten_6"  [label="(197, 1, 2304)", style=solid];
-"505 unflatten_6" -> "506 unsqueeze_6"  [label="(197, 1, 3, 768)", style=solid];
-"506 unsqueeze_6" -> "507 transpose_37"  [label="(1, 197, 1, 3, 768)", style=solid];
-"507 transpose_37" -> "508 squeeze_6"  [label="(3, 197, 1, 1, 768)", style=solid];
-"508 squeeze_6" -> "509 contiguous_6"  [label="(3, 197, 1, 768)", style=solid];
-"509 contiguous_6" -> "510 quantize_per_tensor_default_38"  [label="(3, 197, 1, 768)", style=solid];
-"509 contiguous_6" -> "513 quantize_per_tensor_default_39"  [label="(3, 197, 1, 768)", style=solid];
-"509 contiguous_6" -> "516 select_20"  [label="(3, 197, 1, 768)", style=solid];
-"510 quantize_per_tensor_default_38" -> "511 dequantize_per_tensor_default_38"  [label="(3, 197, 1, 768)", style=solid];
-"511 dequantize_per_tensor_default_38" -> "512 select_18"  [label="(3, 197, 1, 768)", style=solid];
-"512 select_18" -> "517 view_48"  [label="(197, 1, 768)", style=solid];
-"513 quantize_per_tensor_default_39" -> "514 dequantize_per_tensor_default_39"  [label="(3, 197, 1, 768)", style=solid];
-"514 dequantize_per_tensor_default_39" -> "515 select_19"  [label="(3, 197, 1, 768)", style=solid];
-"515 select_19" -> "519 view_49"  [label="(197, 1, 768)", style=solid];
-"516 select_20" -> "521 view_50"  [label="(197, 1, 768)", style=solid];
-"517 view_48" -> "518 transpose_38"  [label="(197, 12, 64)", style=solid];
-"518 transpose_38" -> "523 view_51"  [label="(12, 197, 64)", style=solid];
-"519 view_49" -> "520 transpose_39"  [label="(197, 12, 64)", style=solid];
-"520 transpose_39" -> "524 view_52"  [label="(12, 197, 64)", style=solid];
-"521 view_50" -> "522 transpose_40"  [label="(197, 12, 64)", style=solid];
-"522 transpose_40" -> "525 view_53"  [label="(12, 197, 64)", style=solid];
-"523 view_51" -> "526 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
-"524 view_52" -> "526 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
-"525 view_53" -> "526 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
-"526 scaled_dot_product_attention_6" -> "527 permute_7"  [label="(1, 12, 197, 64)", style=solid];
-"527 permute_7" -> "528 view_54"  [label="(197, 1, 12, 64)", style=solid];
-"528 view_54" -> "529 view_54_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"529 view_54_0_0_nncf_smooth_quant_0" -> "530 quantize_per_tensor_default_40"  [label="(197, 768)", style=solid];
-"530 quantize_per_tensor_default_40" -> "531 dequantize_per_tensor_default_40"  [label="(197, 768)", style=solid];
-"531 dequantize_per_tensor_default_40" -> "538 linear_25"  [label="(197, 768)", style=solid];
-"532 linear_25_scale_0" -> "535 quantize_per_channel_default_26"  [label="(768,)", style=solid];
-"532 linear_25_scale_0" -> "536 dequantize_per_channel_default_26"  [label="(768,)", style=solid];
-"533 linear_25_zero_point_0" -> "535 quantize_per_channel_default_26"  [label="(768,)", style=solid];
-"533 linear_25_zero_point_0" -> "536 dequantize_per_channel_default_26"  [label="(768,)", style=solid];
-"534 compressed_weight_updated_constant26" -> "535 quantize_per_channel_default_26"  [label="(768, 768)", style=solid];
-"535 quantize_per_channel_default_26" -> "536 dequantize_per_channel_default_26"  [label="(768, 768)", style=solid];
-"536 dequantize_per_channel_default_26" -> "538 linear_25"  [label="(768, 768)", style=solid];
-"537 _param_constant81_0_0" -> "538 linear_25"  [label="(768,)", style=solid];
-"538 linear_25" -> "539 view_55"  [label="(197, 768)", style=solid];
-"539 view_55" -> "540 transpose_41"  [label="(197, 1, 768)", style=solid];
-"540 transpose_41" -> "541 dropout_19"  [label="(1, 197, 768)", style=solid];
-"541 dropout_19" -> "542 add_13"  [label="(1, 197, 768)", style=solid];
-"542 add_13" -> "545 layer_norm_13"  [label="(1, 197, 768)", style=solid];
-"542 add_13" -> "569 add_14"  [label="(1, 197, 768)", style=solid];
-"543 _param_constant82" -> "545 layer_norm_13"  [label="(768,)", style=solid];
-"544 _param_constant83" -> "545 layer_norm_13"  [label="(768,)", style=solid];
-"545 layer_norm_13" -> "546 layer_norm_13_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"546 layer_norm_13_0_0_nncf_smooth_quant_0" -> "547 quantize_per_tensor_default_41"  [label="(1, 197, 768)", style=solid];
-"547 quantize_per_tensor_default_41" -> "548 dequantize_per_tensor_default_41"  [label="(1, 197, 768)", style=solid];
-"548 dequantize_per_tensor_default_41" -> "555 linear_26"  [label="(1, 197, 768)", style=solid];
-"549 linear_26_scale_0" -> "552 quantize_per_channel_default_27"  [label="(3072,)", style=solid];
-"549 linear_26_scale_0" -> "553 dequantize_per_channel_default_27"  [label="(3072,)", style=solid];
-"550 linear_26_zero_point_0" -> "552 quantize_per_channel_default_27"  [label="(3072,)", style=solid];
-"550 linear_26_zero_point_0" -> "553 dequantize_per_channel_default_27"  [label="(3072,)", style=solid];
-"551 compressed_weight_updated_constant27" -> "552 quantize_per_channel_default_27"  [label="(3072, 768)", style=solid];
-"552 quantize_per_channel_default_27" -> "553 dequantize_per_channel_default_27"  [label="(3072, 768)", style=solid];
-"553 dequantize_per_channel_default_27" -> "555 linear_26"  [label="(3072, 768)", style=solid];
-"554 _param_constant85_0_0" -> "555 linear_26"  [label="(3072,)", style=solid];
-"555 linear_26" -> "556 gelu_6"  [label="(1, 197, 3072)", style=solid];
-"556 gelu_6" -> "557 dropout_20"  [label="(1, 197, 3072)", style=solid];
-"557 dropout_20" -> "558 dropout_20_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"558 dropout_20_0_0_nncf_smooth_quant_0" -> "559 quantize_per_tensor_default_42"  [label="(1, 197, 3072)", style=solid];
-"559 quantize_per_tensor_default_42" -> "560 dequantize_per_tensor_default_42"  [label="(1, 197, 3072)", style=solid];
-"560 dequantize_per_tensor_default_42" -> "567 linear_27"  [label="(1, 197, 3072)", style=solid];
-"561 linear_27_scale_0" -> "564 quantize_per_channel_default_28"  [label="(768,)", style=solid];
-"561 linear_27_scale_0" -> "565 dequantize_per_channel_default_28"  [label="(768,)", style=solid];
-"562 linear_27_zero_point_0" -> "564 quantize_per_channel_default_28"  [label="(768,)", style=solid];
-"562 linear_27_zero_point_0" -> "565 dequantize_per_channel_default_28"  [label="(768,)", style=solid];
-"563 compressed_weight_updated_constant28" -> "564 quantize_per_channel_default_28"  [label="(768, 3072)", style=solid];
-"564 quantize_per_channel_default_28" -> "565 dequantize_per_channel_default_28"  [label="(768, 3072)", style=solid];
-"565 dequantize_per_channel_default_28" -> "567 linear_27"  [label="(768, 3072)", style=solid];
-"566 _param_constant87_0_0" -> "567 linear_27"  [label="(768,)", style=solid];
-"567 linear_27" -> "568 dropout_21"  [label="(1, 197, 768)", style=solid];
-"568 dropout_21" -> "569 add_14"  [label="(1, 197, 768)", style=solid];
-"569 add_14" -> "572 layer_norm_14"  [label="(1, 197, 768)", style=solid];
-"569 add_14" -> "621 add_15"  [label="(1, 197, 768)", style=solid];
-"570 _param_constant88" -> "572 layer_norm_14"  [label="(768,)", style=solid];
-"571 _param_constant89" -> "572 layer_norm_14"  [label="(768,)", style=solid];
-"572 layer_norm_14" -> "573 transpose_42"  [label="(1, 197, 768)", style=solid];
-"573 transpose_42" -> "574 transpose_42_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"574 transpose_42_0_0_nncf_smooth_quant_0" -> "575 quantize_per_tensor_default_43"  [label="(197, 1, 768)", style=solid];
-"575 quantize_per_tensor_default_43" -> "576 dequantize_per_tensor_default_43"  [label="(197, 1, 768)", style=solid];
-"576 dequantize_per_tensor_default_43" -> "583 linear_28"  [label="(197, 1, 768)", style=solid];
-"577 linear_28_scale_0" -> "580 quantize_per_channel_default_29"  [label="(2304,)", style=solid];
-"577 linear_28_scale_0" -> "581 dequantize_per_channel_default_29"  [label="(2304,)", style=solid];
-"578 linear_28_zero_point_0" -> "580 quantize_per_channel_default_29"  [label="(2304,)", style=solid];
-"578 linear_28_zero_point_0" -> "581 dequantize_per_channel_default_29"  [label="(2304,)", style=solid];
-"579 compressed_weight_updated_constant29" -> "580 quantize_per_channel_default_29"  [label="(2304, 768)", style=solid];
-"580 quantize_per_channel_default_29" -> "581 dequantize_per_channel_default_29"  [label="(2304, 768)", style=solid];
-"581 dequantize_per_channel_default_29" -> "583 linear_28"  [label="(2304, 768)", style=solid];
-"582 _param_constant91_0_0" -> "583 linear_28"  [label="(2304,)", style=solid];
-"583 linear_28" -> "584 unflatten_7"  [label="(197, 1, 2304)", style=solid];
-"584 unflatten_7" -> "585 unsqueeze_7"  [label="(197, 1, 3, 768)", style=solid];
-"585 unsqueeze_7" -> "586 transpose_43"  [label="(1, 197, 1, 3, 768)", style=solid];
-"586 transpose_43" -> "587 squeeze_7"  [label="(3, 197, 1, 1, 768)", style=solid];
-"587 squeeze_7" -> "588 contiguous_7"  [label="(3, 197, 1, 768)", style=solid];
-"588 contiguous_7" -> "589 quantize_per_tensor_default_44"  [label="(3, 197, 1, 768)", style=solid];
-"588 contiguous_7" -> "592 quantize_per_tensor_default_45"  [label="(3, 197, 1, 768)", style=solid];
-"588 contiguous_7" -> "595 select_23"  [label="(3, 197, 1, 768)", style=solid];
-"589 quantize_per_tensor_default_44" -> "590 dequantize_per_tensor_default_44"  [label="(3, 197, 1, 768)", style=solid];
-"590 dequantize_per_tensor_default_44" -> "591 select_21"  [label="(3, 197, 1, 768)", style=solid];
-"591 select_21" -> "596 view_56"  [label="(197, 1, 768)", style=solid];
-"592 quantize_per_tensor_default_45" -> "593 dequantize_per_tensor_default_45"  [label="(3, 197, 1, 768)", style=solid];
-"593 dequantize_per_tensor_default_45" -> "594 select_22"  [label="(3, 197, 1, 768)", style=solid];
-"594 select_22" -> "598 view_57"  [label="(197, 1, 768)", style=solid];
-"595 select_23" -> "600 view_58"  [label="(197, 1, 768)", style=solid];
-"596 view_56" -> "597 transpose_44"  [label="(197, 12, 64)", style=solid];
-"597 transpose_44" -> "602 view_59"  [label="(12, 197, 64)", style=solid];
-"598 view_57" -> "599 transpose_45"  [label="(197, 12, 64)", style=solid];
-"599 transpose_45" -> "603 view_60"  [label="(12, 197, 64)", style=solid];
-"600 view_58" -> "601 transpose_46"  [label="(197, 12, 64)", style=solid];
-"601 transpose_46" -> "604 view_61"  [label="(12, 197, 64)", style=solid];
-"602 view_59" -> "605 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
-"603 view_60" -> "605 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
-"604 view_61" -> "605 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
-"605 scaled_dot_product_attention_7" -> "606 permute_8"  [label="(1, 12, 197, 64)", style=solid];
-"606 permute_8" -> "607 view_62"  [label="(197, 1, 12, 64)", style=solid];
-"607 view_62" -> "608 view_62_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"608 view_62_0_0_nncf_smooth_quant_0" -> "609 quantize_per_tensor_default_46"  [label="(197, 768)", style=solid];
-"609 quantize_per_tensor_default_46" -> "610 dequantize_per_tensor_default_46"  [label="(197, 768)", style=solid];
-"610 dequantize_per_tensor_default_46" -> "617 linear_29"  [label="(197, 768)", style=solid];
-"611 linear_29_scale_0" -> "614 quantize_per_channel_default_30"  [label="(768,)", style=solid];
-"611 linear_29_scale_0" -> "615 dequantize_per_channel_default_30"  [label="(768,)", style=solid];
-"612 linear_29_zero_point_0" -> "614 quantize_per_channel_default_30"  [label="(768,)", style=solid];
-"612 linear_29_zero_point_0" -> "615 dequantize_per_channel_default_30"  [label="(768,)", style=solid];
-"613 compressed_weight_updated_constant30" -> "614 quantize_per_channel_default_30"  [label="(768, 768)", style=solid];
-"614 quantize_per_channel_default_30" -> "615 dequantize_per_channel_default_30"  [label="(768, 768)", style=solid];
-"615 dequantize_per_channel_default_30" -> "617 linear_29"  [label="(768, 768)", style=solid];
-"616 _param_constant93_0_0" -> "617 linear_29"  [label="(768,)", style=solid];
-"617 linear_29" -> "618 view_63"  [label="(197, 768)", style=solid];
-"618 view_63" -> "619 transpose_47"  [label="(197, 1, 768)", style=solid];
-"619 transpose_47" -> "620 dropout_22"  [label="(1, 197, 768)", style=solid];
-"620 dropout_22" -> "621 add_15"  [label="(1, 197, 768)", style=solid];
-"621 add_15" -> "624 layer_norm_15"  [label="(1, 197, 768)", style=solid];
-"621 add_15" -> "648 add_16"  [label="(1, 197, 768)", style=solid];
-"622 _param_constant94" -> "624 layer_norm_15"  [label="(768,)", style=solid];
-"623 _param_constant95" -> "624 layer_norm_15"  [label="(768,)", style=solid];
-"624 layer_norm_15" -> "625 layer_norm_15_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"625 layer_norm_15_0_0_nncf_smooth_quant_0" -> "626 quantize_per_tensor_default_47"  [label="(1, 197, 768)", style=solid];
-"626 quantize_per_tensor_default_47" -> "627 dequantize_per_tensor_default_47"  [label="(1, 197, 768)", style=solid];
-"627 dequantize_per_tensor_default_47" -> "634 linear_30"  [label="(1, 197, 768)", style=solid];
-"628 linear_30_scale_0" -> "631 quantize_per_channel_default_31"  [label="(3072,)", style=solid];
-"628 linear_30_scale_0" -> "632 dequantize_per_channel_default_31"  [label="(3072,)", style=solid];
-"629 linear_30_zero_point_0" -> "631 quantize_per_channel_default_31"  [label="(3072,)", style=solid];
-"629 linear_30_zero_point_0" -> "632 dequantize_per_channel_default_31"  [label="(3072,)", style=solid];
-"630 compressed_weight_updated_constant31" -> "631 quantize_per_channel_default_31"  [label="(3072, 768)", style=solid];
-"631 quantize_per_channel_default_31" -> "632 dequantize_per_channel_default_31"  [label="(3072, 768)", style=solid];
-"632 dequantize_per_channel_default_31" -> "634 linear_30"  [label="(3072, 768)", style=solid];
-"633 _param_constant97_0_0" -> "634 linear_30"  [label="(3072,)", style=solid];
-"634 linear_30" -> "635 gelu_7"  [label="(1, 197, 3072)", style=solid];
-"635 gelu_7" -> "636 dropout_23"  [label="(1, 197, 3072)", style=solid];
-"636 dropout_23" -> "637 dropout_23_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"637 dropout_23_0_0_nncf_smooth_quant_0" -> "638 quantize_per_tensor_default_48"  [label="(1, 197, 3072)", style=solid];
-"638 quantize_per_tensor_default_48" -> "639 dequantize_per_tensor_default_48"  [label="(1, 197, 3072)", style=solid];
-"639 dequantize_per_tensor_default_48" -> "646 linear_31"  [label="(1, 197, 3072)", style=solid];
-"640 linear_31_scale_0" -> "643 quantize_per_channel_default_32"  [label="(768,)", style=solid];
-"640 linear_31_scale_0" -> "644 dequantize_per_channel_default_32"  [label="(768,)", style=solid];
-"641 linear_31_zero_point_0" -> "643 quantize_per_channel_default_32"  [label="(768,)", style=solid];
-"641 linear_31_zero_point_0" -> "644 dequantize_per_channel_default_32"  [label="(768,)", style=solid];
-"642 compressed_weight_updated_constant32" -> "643 quantize_per_channel_default_32"  [label="(768, 3072)", style=solid];
-"643 quantize_per_channel_default_32" -> "644 dequantize_per_channel_default_32"  [label="(768, 3072)", style=solid];
-"644 dequantize_per_channel_default_32" -> "646 linear_31"  [label="(768, 3072)", style=solid];
-"645 _param_constant99_0_0" -> "646 linear_31"  [label="(768,)", style=solid];
-"646 linear_31" -> "647 dropout_24"  [label="(1, 197, 768)", style=solid];
-"647 dropout_24" -> "648 add_16"  [label="(1, 197, 768)", style=solid];
-"648 add_16" -> "651 layer_norm_16"  [label="(1, 197, 768)", style=solid];
-"648 add_16" -> "700 add_17"  [label="(1, 197, 768)", style=solid];
-"649 _param_constant100" -> "651 layer_norm_16"  [label="(768,)", style=solid];
-"650 _param_constant101" -> "651 layer_norm_16"  [label="(768,)", style=solid];
-"651 layer_norm_16" -> "652 transpose_48"  [label="(1, 197, 768)", style=solid];
-"652 transpose_48" -> "653 transpose_48_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"653 transpose_48_0_0_nncf_smooth_quant_0" -> "654 quantize_per_tensor_default_49"  [label="(197, 1, 768)", style=solid];
-"654 quantize_per_tensor_default_49" -> "655 dequantize_per_tensor_default_49"  [label="(197, 1, 768)", style=solid];
-"655 dequantize_per_tensor_default_49" -> "662 linear_32"  [label="(197, 1, 768)", style=solid];
-"656 linear_32_scale_0" -> "659 quantize_per_channel_default_33"  [label="(2304,)", style=solid];
-"656 linear_32_scale_0" -> "660 dequantize_per_channel_default_33"  [label="(2304,)", style=solid];
-"657 linear_32_zero_point_0" -> "659 quantize_per_channel_default_33"  [label="(2304,)", style=solid];
-"657 linear_32_zero_point_0" -> "660 dequantize_per_channel_default_33"  [label="(2304,)", style=solid];
-"658 compressed_weight_updated_constant33" -> "659 quantize_per_channel_default_33"  [label="(2304, 768)", style=solid];
-"659 quantize_per_channel_default_33" -> "660 dequantize_per_channel_default_33"  [label="(2304, 768)", style=solid];
-"660 dequantize_per_channel_default_33" -> "662 linear_32"  [label="(2304, 768)", style=solid];
-"661 _param_constant103_0_0" -> "662 linear_32"  [label="(2304,)", style=solid];
-"662 linear_32" -> "663 unflatten_8"  [label="(197, 1, 2304)", style=solid];
-"663 unflatten_8" -> "664 unsqueeze_8"  [label="(197, 1, 3, 768)", style=solid];
-"664 unsqueeze_8" -> "665 transpose_49"  [label="(1, 197, 1, 3, 768)", style=solid];
-"665 transpose_49" -> "666 squeeze_8"  [label="(3, 197, 1, 1, 768)", style=solid];
-"666 squeeze_8" -> "667 contiguous_8"  [label="(3, 197, 1, 768)", style=solid];
-"667 contiguous_8" -> "668 quantize_per_tensor_default_50"  [label="(3, 197, 1, 768)", style=solid];
-"667 contiguous_8" -> "671 quantize_per_tensor_default_51"  [label="(3, 197, 1, 768)", style=solid];
-"667 contiguous_8" -> "674 select_26"  [label="(3, 197, 1, 768)", style=solid];
-"668 quantize_per_tensor_default_50" -> "669 dequantize_per_tensor_default_50"  [label="(3, 197, 1, 768)", style=solid];
-"669 dequantize_per_tensor_default_50" -> "670 select_24"  [label="(3, 197, 1, 768)", style=solid];
-"670 select_24" -> "675 view_64"  [label="(197, 1, 768)", style=solid];
-"671 quantize_per_tensor_default_51" -> "672 dequantize_per_tensor_default_51"  [label="(3, 197, 1, 768)", style=solid];
-"672 dequantize_per_tensor_default_51" -> "673 select_25"  [label="(3, 197, 1, 768)", style=solid];
-"673 select_25" -> "677 view_65"  [label="(197, 1, 768)", style=solid];
-"674 select_26" -> "679 view_66"  [label="(197, 1, 768)", style=solid];
-"675 view_64" -> "676 transpose_50"  [label="(197, 12, 64)", style=solid];
-"676 transpose_50" -> "681 view_67"  [label="(12, 197, 64)", style=solid];
-"677 view_65" -> "678 transpose_51"  [label="(197, 12, 64)", style=solid];
-"678 transpose_51" -> "682 view_68"  [label="(12, 197, 64)", style=solid];
-"679 view_66" -> "680 transpose_52"  [label="(197, 12, 64)", style=solid];
-"680 transpose_52" -> "683 view_69"  [label="(12, 197, 64)", style=solid];
-"681 view_67" -> "684 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
-"682 view_68" -> "684 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
-"683 view_69" -> "684 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
-"684 scaled_dot_product_attention_8" -> "685 permute_9"  [label="(1, 12, 197, 64)", style=solid];
-"685 permute_9" -> "686 view_70"  [label="(197, 1, 12, 64)", style=solid];
-"686 view_70" -> "687 view_70_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"687 view_70_0_0_nncf_smooth_quant_0" -> "688 quantize_per_tensor_default_52"  [label="(197, 768)", style=solid];
-"688 quantize_per_tensor_default_52" -> "689 dequantize_per_tensor_default_52"  [label="(197, 768)", style=solid];
-"689 dequantize_per_tensor_default_52" -> "696 linear_33"  [label="(197, 768)", style=solid];
-"690 linear_33_scale_0" -> "693 quantize_per_channel_default_34"  [label="(768,)", style=solid];
-"690 linear_33_scale_0" -> "694 dequantize_per_channel_default_34"  [label="(768,)", style=solid];
-"691 linear_33_zero_point_0" -> "693 quantize_per_channel_default_34"  [label="(768,)", style=solid];
-"691 linear_33_zero_point_0" -> "694 dequantize_per_channel_default_34"  [label="(768,)", style=solid];
-"692 compressed_weight_updated_constant34" -> "693 quantize_per_channel_default_34"  [label="(768, 768)", style=solid];
-"693 quantize_per_channel_default_34" -> "694 dequantize_per_channel_default_34"  [label="(768, 768)", style=solid];
-"694 dequantize_per_channel_default_34" -> "696 linear_33"  [label="(768, 768)", style=solid];
-"695 _param_constant105_0_0" -> "696 linear_33"  [label="(768,)", style=solid];
-"696 linear_33" -> "697 view_71"  [label="(197, 768)", style=solid];
-"697 view_71" -> "698 transpose_53"  [label="(197, 1, 768)", style=solid];
-"698 transpose_53" -> "699 dropout_25"  [label="(1, 197, 768)", style=solid];
-"699 dropout_25" -> "700 add_17"  [label="(1, 197, 768)", style=solid];
-"700 add_17" -> "703 layer_norm_17"  [label="(1, 197, 768)", style=solid];
-"700 add_17" -> "727 add_18"  [label="(1, 197, 768)", style=solid];
-"701 _param_constant106" -> "703 layer_norm_17"  [label="(768,)", style=solid];
-"702 _param_constant107" -> "703 layer_norm_17"  [label="(768,)", style=solid];
-"703 layer_norm_17" -> "704 layer_norm_17_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"704 layer_norm_17_0_0_nncf_smooth_quant_0" -> "705 quantize_per_tensor_default_53"  [label="(1, 197, 768)", style=solid];
-"705 quantize_per_tensor_default_53" -> "706 dequantize_per_tensor_default_53"  [label="(1, 197, 768)", style=solid];
-"706 dequantize_per_tensor_default_53" -> "713 linear_34"  [label="(1, 197, 768)", style=solid];
-"707 linear_34_scale_0" -> "710 quantize_per_channel_default_35"  [label="(3072,)", style=solid];
-"707 linear_34_scale_0" -> "711 dequantize_per_channel_default_35"  [label="(3072,)", style=solid];
-"708 linear_34_zero_point_0" -> "710 quantize_per_channel_default_35"  [label="(3072,)", style=solid];
-"708 linear_34_zero_point_0" -> "711 dequantize_per_channel_default_35"  [label="(3072,)", style=solid];
-"709 compressed_weight_updated_constant35" -> "710 quantize_per_channel_default_35"  [label="(3072, 768)", style=solid];
-"710 quantize_per_channel_default_35" -> "711 dequantize_per_channel_default_35"  [label="(3072, 768)", style=solid];
-"711 dequantize_per_channel_default_35" -> "713 linear_34"  [label="(3072, 768)", style=solid];
-"712 _param_constant109_0_0" -> "713 linear_34"  [label="(3072,)", style=solid];
-"713 linear_34" -> "714 gelu_8"  [label="(1, 197, 3072)", style=solid];
-"714 gelu_8" -> "715 dropout_26"  [label="(1, 197, 3072)", style=solid];
-"715 dropout_26" -> "716 dropout_26_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"716 dropout_26_0_0_nncf_smooth_quant_0" -> "717 quantize_per_tensor_default_54"  [label="(1, 197, 3072)", style=solid];
-"717 quantize_per_tensor_default_54" -> "718 dequantize_per_tensor_default_54"  [label="(1, 197, 3072)", style=solid];
-"718 dequantize_per_tensor_default_54" -> "725 linear_35"  [label="(1, 197, 3072)", style=solid];
-"719 linear_35_scale_0" -> "722 quantize_per_channel_default_36"  [label="(768,)", style=solid];
-"719 linear_35_scale_0" -> "723 dequantize_per_channel_default_36"  [label="(768,)", style=solid];
-"720 linear_35_zero_point_0" -> "722 quantize_per_channel_default_36"  [label="(768,)", style=solid];
-"720 linear_35_zero_point_0" -> "723 dequantize_per_channel_default_36"  [label="(768,)", style=solid];
-"721 compressed_weight_updated_constant36" -> "722 quantize_per_channel_default_36"  [label="(768, 3072)", style=solid];
-"722 quantize_per_channel_default_36" -> "723 dequantize_per_channel_default_36"  [label="(768, 3072)", style=solid];
-"723 dequantize_per_channel_default_36" -> "725 linear_35"  [label="(768, 3072)", style=solid];
-"724 _param_constant111_0_0" -> "725 linear_35"  [label="(768,)", style=solid];
-"725 linear_35" -> "726 dropout_27"  [label="(1, 197, 768)", style=solid];
-"726 dropout_27" -> "727 add_18"  [label="(1, 197, 768)", style=solid];
-"727 add_18" -> "730 layer_norm_18"  [label="(1, 197, 768)", style=solid];
-"727 add_18" -> "779 add_19"  [label="(1, 197, 768)", style=solid];
-"728 _param_constant112" -> "730 layer_norm_18"  [label="(768,)", style=solid];
-"729 _param_constant113" -> "730 layer_norm_18"  [label="(768,)", style=solid];
-"730 layer_norm_18" -> "731 transpose_54"  [label="(1, 197, 768)", style=solid];
-"731 transpose_54" -> "732 transpose_54_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"732 transpose_54_0_0_nncf_smooth_quant_0" -> "733 quantize_per_tensor_default_55"  [label="(197, 1, 768)", style=solid];
-"733 quantize_per_tensor_default_55" -> "734 dequantize_per_tensor_default_55"  [label="(197, 1, 768)", style=solid];
-"734 dequantize_per_tensor_default_55" -> "741 linear_36"  [label="(197, 1, 768)", style=solid];
-"735 linear_36_scale_0" -> "738 quantize_per_channel_default_37"  [label="(2304,)", style=solid];
-"735 linear_36_scale_0" -> "739 dequantize_per_channel_default_37"  [label="(2304,)", style=solid];
-"736 linear_36_zero_point_0" -> "738 quantize_per_channel_default_37"  [label="(2304,)", style=solid];
-"736 linear_36_zero_point_0" -> "739 dequantize_per_channel_default_37"  [label="(2304,)", style=solid];
-"737 compressed_weight_updated_constant37" -> "738 quantize_per_channel_default_37"  [label="(2304, 768)", style=solid];
-"738 quantize_per_channel_default_37" -> "739 dequantize_per_channel_default_37"  [label="(2304, 768)", style=solid];
-"739 dequantize_per_channel_default_37" -> "741 linear_36"  [label="(2304, 768)", style=solid];
-"740 _param_constant115_0_0" -> "741 linear_36"  [label="(2304,)", style=solid];
-"741 linear_36" -> "742 unflatten_9"  [label="(197, 1, 2304)", style=solid];
-"742 unflatten_9" -> "743 unsqueeze_9"  [label="(197, 1, 3, 768)", style=solid];
-"743 unsqueeze_9" -> "744 transpose_55"  [label="(1, 197, 1, 3, 768)", style=solid];
-"744 transpose_55" -> "745 squeeze_9"  [label="(3, 197, 1, 1, 768)", style=solid];
-"745 squeeze_9" -> "746 contiguous_9"  [label="(3, 197, 1, 768)", style=solid];
-"746 contiguous_9" -> "747 quantize_per_tensor_default_56"  [label="(3, 197, 1, 768)", style=solid];
-"746 contiguous_9" -> "750 quantize_per_tensor_default_57"  [label="(3, 197, 1, 768)", style=solid];
-"746 contiguous_9" -> "753 select_29"  [label="(3, 197, 1, 768)", style=solid];
-"747 quantize_per_tensor_default_56" -> "748 dequantize_per_tensor_default_56"  [label="(3, 197, 1, 768)", style=solid];
-"748 dequantize_per_tensor_default_56" -> "749 select_27"  [label="(3, 197, 1, 768)", style=solid];
-"749 select_27" -> "754 view_72"  [label="(197, 1, 768)", style=solid];
-"750 quantize_per_tensor_default_57" -> "751 dequantize_per_tensor_default_57"  [label="(3, 197, 1, 768)", style=solid];
-"751 dequantize_per_tensor_default_57" -> "752 select_28"  [label="(3, 197, 1, 768)", style=solid];
-"752 select_28" -> "756 view_73"  [label="(197, 1, 768)", style=solid];
-"753 select_29" -> "758 view_74"  [label="(197, 1, 768)", style=solid];
-"754 view_72" -> "755 transpose_56"  [label="(197, 12, 64)", style=solid];
-"755 transpose_56" -> "760 view_75"  [label="(12, 197, 64)", style=solid];
-"756 view_73" -> "757 transpose_57"  [label="(197, 12, 64)", style=solid];
-"757 transpose_57" -> "761 view_76"  [label="(12, 197, 64)", style=solid];
-"758 view_74" -> "759 transpose_58"  [label="(197, 12, 64)", style=solid];
-"759 transpose_58" -> "762 view_77"  [label="(12, 197, 64)", style=solid];
-"760 view_75" -> "763 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
-"761 view_76" -> "763 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
-"762 view_77" -> "763 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
-"763 scaled_dot_product_attention_9" -> "764 permute_10"  [label="(1, 12, 197, 64)", style=solid];
-"764 permute_10" -> "765 view_78"  [label="(197, 1, 12, 64)", style=solid];
-"765 view_78" -> "766 view_78_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"766 view_78_0_0_nncf_smooth_quant_0" -> "767 quantize_per_tensor_default_58"  [label="(197, 768)", style=solid];
-"767 quantize_per_tensor_default_58" -> "768 dequantize_per_tensor_default_58"  [label="(197, 768)", style=solid];
-"768 dequantize_per_tensor_default_58" -> "775 linear_37"  [label="(197, 768)", style=solid];
-"769 linear_37_scale_0" -> "772 quantize_per_channel_default_38"  [label="(768,)", style=solid];
-"769 linear_37_scale_0" -> "773 dequantize_per_channel_default_38"  [label="(768,)", style=solid];
-"770 linear_37_zero_point_0" -> "772 quantize_per_channel_default_38"  [label="(768,)", style=solid];
-"770 linear_37_zero_point_0" -> "773 dequantize_per_channel_default_38"  [label="(768,)", style=solid];
-"771 compressed_weight_updated_constant38" -> "772 quantize_per_channel_default_38"  [label="(768, 768)", style=solid];
-"772 quantize_per_channel_default_38" -> "773 dequantize_per_channel_default_38"  [label="(768, 768)", style=solid];
-"773 dequantize_per_channel_default_38" -> "775 linear_37"  [label="(768, 768)", style=solid];
-"774 _param_constant117_0_0" -> "775 linear_37"  [label="(768,)", style=solid];
-"775 linear_37" -> "776 view_79"  [label="(197, 768)", style=solid];
-"776 view_79" -> "777 transpose_59"  [label="(197, 1, 768)", style=solid];
-"777 transpose_59" -> "778 dropout_28"  [label="(1, 197, 768)", style=solid];
-"778 dropout_28" -> "779 add_19"  [label="(1, 197, 768)", style=solid];
-"779 add_19" -> "782 layer_norm_19"  [label="(1, 197, 768)", style=solid];
-"779 add_19" -> "806 add_20"  [label="(1, 197, 768)", style=solid];
-"780 _param_constant118" -> "782 layer_norm_19"  [label="(768,)", style=solid];
-"781 _param_constant119" -> "782 layer_norm_19"  [label="(768,)", style=solid];
-"782 layer_norm_19" -> "783 layer_norm_19_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"783 layer_norm_19_0_0_nncf_smooth_quant_0" -> "784 quantize_per_tensor_default_59"  [label="(1, 197, 768)", style=solid];
-"784 quantize_per_tensor_default_59" -> "785 dequantize_per_tensor_default_59"  [label="(1, 197, 768)", style=solid];
-"785 dequantize_per_tensor_default_59" -> "792 linear_38"  [label="(1, 197, 768)", style=solid];
-"786 linear_38_scale_0" -> "789 quantize_per_channel_default_39"  [label="(3072,)", style=solid];
-"786 linear_38_scale_0" -> "790 dequantize_per_channel_default_39"  [label="(3072,)", style=solid];
-"787 linear_38_zero_point_0" -> "789 quantize_per_channel_default_39"  [label="(3072,)", style=solid];
-"787 linear_38_zero_point_0" -> "790 dequantize_per_channel_default_39"  [label="(3072,)", style=solid];
-"788 compressed_weight_updated_constant39" -> "789 quantize_per_channel_default_39"  [label="(3072, 768)", style=solid];
-"789 quantize_per_channel_default_39" -> "790 dequantize_per_channel_default_39"  [label="(3072, 768)", style=solid];
-"790 dequantize_per_channel_default_39" -> "792 linear_38"  [label="(3072, 768)", style=solid];
-"791 _param_constant121_0_0" -> "792 linear_38"  [label="(3072,)", style=solid];
-"792 linear_38" -> "793 gelu_9"  [label="(1, 197, 3072)", style=solid];
-"793 gelu_9" -> "794 dropout_29"  [label="(1, 197, 3072)", style=solid];
-"794 dropout_29" -> "795 dropout_29_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"795 dropout_29_0_0_nncf_smooth_quant_0" -> "796 quantize_per_tensor_default_60"  [label="(1, 197, 3072)", style=solid];
-"796 quantize_per_tensor_default_60" -> "797 dequantize_per_tensor_default_60"  [label="(1, 197, 3072)", style=solid];
-"797 dequantize_per_tensor_default_60" -> "804 linear_39"  [label="(1, 197, 3072)", style=solid];
-"798 linear_39_scale_0" -> "801 quantize_per_channel_default_40"  [label="(768,)", style=solid];
-"798 linear_39_scale_0" -> "802 dequantize_per_channel_default_40"  [label="(768,)", style=solid];
-"799 linear_39_zero_point_0" -> "801 quantize_per_channel_default_40"  [label="(768,)", style=solid];
-"799 linear_39_zero_point_0" -> "802 dequantize_per_channel_default_40"  [label="(768,)", style=solid];
-"800 compressed_weight_updated_constant40" -> "801 quantize_per_channel_default_40"  [label="(768, 3072)", style=solid];
-"801 quantize_per_channel_default_40" -> "802 dequantize_per_channel_default_40"  [label="(768, 3072)", style=solid];
-"802 dequantize_per_channel_default_40" -> "804 linear_39"  [label="(768, 3072)", style=solid];
-"803 _param_constant123_0_0" -> "804 linear_39"  [label="(768,)", style=solid];
-"804 linear_39" -> "805 dropout_30"  [label="(1, 197, 768)", style=solid];
-"805 dropout_30" -> "806 add_20"  [label="(1, 197, 768)", style=solid];
-"806 add_20" -> "809 layer_norm_20"  [label="(1, 197, 768)", style=solid];
-"806 add_20" -> "858 add_21"  [label="(1, 197, 768)", style=solid];
-"807 _param_constant124" -> "809 layer_norm_20"  [label="(768,)", style=solid];
-"808 _param_constant125" -> "809 layer_norm_20"  [label="(768,)", style=solid];
-"809 layer_norm_20" -> "810 transpose_60"  [label="(1, 197, 768)", style=solid];
-"810 transpose_60" -> "811 transpose_60_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"811 transpose_60_0_0_nncf_smooth_quant_0" -> "812 quantize_per_tensor_default_61"  [label="(197, 1, 768)", style=solid];
-"812 quantize_per_tensor_default_61" -> "813 dequantize_per_tensor_default_61"  [label="(197, 1, 768)", style=solid];
-"813 dequantize_per_tensor_default_61" -> "820 linear_40"  [label="(197, 1, 768)", style=solid];
-"814 linear_40_scale_0" -> "817 quantize_per_channel_default_41"  [label="(2304,)", style=solid];
-"814 linear_40_scale_0" -> "818 dequantize_per_channel_default_41"  [label="(2304,)", style=solid];
-"815 linear_40_zero_point_0" -> "817 quantize_per_channel_default_41"  [label="(2304,)", style=solid];
-"815 linear_40_zero_point_0" -> "818 dequantize_per_channel_default_41"  [label="(2304,)", style=solid];
-"816 compressed_weight_updated_constant41" -> "817 quantize_per_channel_default_41"  [label="(2304, 768)", style=solid];
-"817 quantize_per_channel_default_41" -> "818 dequantize_per_channel_default_41"  [label="(2304, 768)", style=solid];
-"818 dequantize_per_channel_default_41" -> "820 linear_40"  [label="(2304, 768)", style=solid];
-"819 _param_constant127_0_0" -> "820 linear_40"  [label="(2304,)", style=solid];
-"820 linear_40" -> "821 unflatten_10"  [label="(197, 1, 2304)", style=solid];
-"821 unflatten_10" -> "822 unsqueeze_10"  [label="(197, 1, 3, 768)", style=solid];
-"822 unsqueeze_10" -> "823 transpose_61"  [label="(1, 197, 1, 3, 768)", style=solid];
-"823 transpose_61" -> "824 squeeze_10"  [label="(3, 197, 1, 1, 768)", style=solid];
-"824 squeeze_10" -> "825 contiguous_10"  [label="(3, 197, 1, 768)", style=solid];
-"825 contiguous_10" -> "826 quantize_per_tensor_default_62"  [label="(3, 197, 1, 768)", style=solid];
-"825 contiguous_10" -> "829 quantize_per_tensor_default_63"  [label="(3, 197, 1, 768)", style=solid];
-"825 contiguous_10" -> "832 select_32"  [label="(3, 197, 1, 768)", style=solid];
-"826 quantize_per_tensor_default_62" -> "827 dequantize_per_tensor_default_62"  [label="(3, 197, 1, 768)", style=solid];
-"827 dequantize_per_tensor_default_62" -> "828 select_30"  [label="(3, 197, 1, 768)", style=solid];
-"828 select_30" -> "833 view_80"  [label="(197, 1, 768)", style=solid];
-"829 quantize_per_tensor_default_63" -> "830 dequantize_per_tensor_default_63"  [label="(3, 197, 1, 768)", style=solid];
-"830 dequantize_per_tensor_default_63" -> "831 select_31"  [label="(3, 197, 1, 768)", style=solid];
-"831 select_31" -> "835 view_81"  [label="(197, 1, 768)", style=solid];
-"832 select_32" -> "837 view_82"  [label="(197, 1, 768)", style=solid];
-"833 view_80" -> "834 transpose_62"  [label="(197, 12, 64)", style=solid];
-"834 transpose_62" -> "839 view_83"  [label="(12, 197, 64)", style=solid];
-"835 view_81" -> "836 transpose_63"  [label="(197, 12, 64)", style=solid];
-"836 transpose_63" -> "840 view_84"  [label="(12, 197, 64)", style=solid];
-"837 view_82" -> "838 transpose_64"  [label="(197, 12, 64)", style=solid];
-"838 transpose_64" -> "841 view_85"  [label="(12, 197, 64)", style=solid];
-"839 view_83" -> "842 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
-"840 view_84" -> "842 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
-"841 view_85" -> "842 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
-"842 scaled_dot_product_attention_10" -> "843 permute_11"  [label="(1, 12, 197, 64)", style=solid];
-"843 permute_11" -> "844 view_86"  [label="(197, 1, 12, 64)", style=solid];
-"844 view_86" -> "845 view_86_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"845 view_86_0_0_nncf_smooth_quant_0" -> "846 quantize_per_tensor_default_64"  [label="(197, 768)", style=solid];
-"846 quantize_per_tensor_default_64" -> "847 dequantize_per_tensor_default_64"  [label="(197, 768)", style=solid];
-"847 dequantize_per_tensor_default_64" -> "854 linear_41"  [label="(197, 768)", style=solid];
-"848 linear_41_scale_0" -> "851 quantize_per_channel_default_42"  [label="(768,)", style=solid];
-"848 linear_41_scale_0" -> "852 dequantize_per_channel_default_42"  [label="(768,)", style=solid];
-"849 linear_41_zero_point_0" -> "851 quantize_per_channel_default_42"  [label="(768,)", style=solid];
-"849 linear_41_zero_point_0" -> "852 dequantize_per_channel_default_42"  [label="(768,)", style=solid];
-"850 compressed_weight_updated_constant42" -> "851 quantize_per_channel_default_42"  [label="(768, 768)", style=solid];
-"851 quantize_per_channel_default_42" -> "852 dequantize_per_channel_default_42"  [label="(768, 768)", style=solid];
-"852 dequantize_per_channel_default_42" -> "854 linear_41"  [label="(768, 768)", style=solid];
-"853 _param_constant129_0_0" -> "854 linear_41"  [label="(768,)", style=solid];
-"854 linear_41" -> "855 view_87"  [label="(197, 768)", style=solid];
-"855 view_87" -> "856 transpose_65"  [label="(197, 1, 768)", style=solid];
-"856 transpose_65" -> "857 dropout_31"  [label="(1, 197, 768)", style=solid];
-"857 dropout_31" -> "858 add_21"  [label="(1, 197, 768)", style=solid];
-"858 add_21" -> "861 layer_norm_21"  [label="(1, 197, 768)", style=solid];
-"858 add_21" -> "885 add_22"  [label="(1, 197, 768)", style=solid];
-"859 _param_constant130" -> "861 layer_norm_21"  [label="(768,)", style=solid];
-"860 _param_constant131" -> "861 layer_norm_21"  [label="(768,)", style=solid];
-"861 layer_norm_21" -> "862 layer_norm_21_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"862 layer_norm_21_0_0_nncf_smooth_quant_0" -> "863 quantize_per_tensor_default_65"  [label="(1, 197, 768)", style=solid];
-"863 quantize_per_tensor_default_65" -> "864 dequantize_per_tensor_default_65"  [label="(1, 197, 768)", style=solid];
-"864 dequantize_per_tensor_default_65" -> "871 linear_42"  [label="(1, 197, 768)", style=solid];
-"865 linear_42_scale_0" -> "868 quantize_per_channel_default_43"  [label="(3072,)", style=solid];
-"865 linear_42_scale_0" -> "869 dequantize_per_channel_default_43"  [label="(3072,)", style=solid];
-"866 linear_42_zero_point_0" -> "868 quantize_per_channel_default_43"  [label="(3072,)", style=solid];
-"866 linear_42_zero_point_0" -> "869 dequantize_per_channel_default_43"  [label="(3072,)", style=solid];
-"867 compressed_weight_updated_constant43" -> "868 quantize_per_channel_default_43"  [label="(3072, 768)", style=solid];
-"868 quantize_per_channel_default_43" -> "869 dequantize_per_channel_default_43"  [label="(3072, 768)", style=solid];
-"869 dequantize_per_channel_default_43" -> "871 linear_42"  [label="(3072, 768)", style=solid];
-"870 _param_constant133_0_0" -> "871 linear_42"  [label="(3072,)", style=solid];
-"871 linear_42" -> "872 gelu_10"  [label="(1, 197, 3072)", style=solid];
-"872 gelu_10" -> "873 dropout_32"  [label="(1, 197, 3072)", style=solid];
-"873 dropout_32" -> "874 dropout_32_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"874 dropout_32_0_0_nncf_smooth_quant_0" -> "875 quantize_per_tensor_default_66"  [label="(1, 197, 3072)", style=solid];
-"875 quantize_per_tensor_default_66" -> "876 dequantize_per_tensor_default_66"  [label="(1, 197, 3072)", style=solid];
-"876 dequantize_per_tensor_default_66" -> "883 linear_43"  [label="(1, 197, 3072)", style=solid];
-"877 linear_43_scale_0" -> "880 quantize_per_channel_default_44"  [label="(768,)", style=solid];
-"877 linear_43_scale_0" -> "881 dequantize_per_channel_default_44"  [label="(768,)", style=solid];
-"878 linear_43_zero_point_0" -> "880 quantize_per_channel_default_44"  [label="(768,)", style=solid];
-"878 linear_43_zero_point_0" -> "881 dequantize_per_channel_default_44"  [label="(768,)", style=solid];
-"879 compressed_weight_updated_constant44" -> "880 quantize_per_channel_default_44"  [label="(768, 3072)", style=solid];
-"880 quantize_per_channel_default_44" -> "881 dequantize_per_channel_default_44"  [label="(768, 3072)", style=solid];
-"881 dequantize_per_channel_default_44" -> "883 linear_43"  [label="(768, 3072)", style=solid];
-"882 _param_constant135_0_0" -> "883 linear_43"  [label="(768,)", style=solid];
-"883 linear_43" -> "884 dropout_33"  [label="(1, 197, 768)", style=solid];
-"884 dropout_33" -> "885 add_22"  [label="(1, 197, 768)", style=solid];
-"885 add_22" -> "888 layer_norm_22"  [label="(1, 197, 768)", style=solid];
-"885 add_22" -> "937 add_23"  [label="(1, 197, 768)", style=solid];
-"886 _param_constant136" -> "888 layer_norm_22"  [label="(768,)", style=solid];
-"887 _param_constant137" -> "888 layer_norm_22"  [label="(768,)", style=solid];
-"888 layer_norm_22" -> "889 transpose_66"  [label="(1, 197, 768)", style=solid];
-"889 transpose_66" -> "890 transpose_66_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
-"890 transpose_66_0_0_nncf_smooth_quant_0" -> "891 quantize_per_tensor_default_67"  [label="(197, 1, 768)", style=solid];
-"891 quantize_per_tensor_default_67" -> "892 dequantize_per_tensor_default_67"  [label="(197, 1, 768)", style=solid];
-"892 dequantize_per_tensor_default_67" -> "899 linear_44"  [label="(197, 1, 768)", style=solid];
-"893 linear_44_scale_0" -> "896 quantize_per_channel_default_45"  [label="(2304,)", style=solid];
-"893 linear_44_scale_0" -> "897 dequantize_per_channel_default_45"  [label="(2304,)", style=solid];
-"894 linear_44_zero_point_0" -> "896 quantize_per_channel_default_45"  [label="(2304,)", style=solid];
-"894 linear_44_zero_point_0" -> "897 dequantize_per_channel_default_45"  [label="(2304,)", style=solid];
-"895 compressed_weight_updated_constant45" -> "896 quantize_per_channel_default_45"  [label="(2304, 768)", style=solid];
-"896 quantize_per_channel_default_45" -> "897 dequantize_per_channel_default_45"  [label="(2304, 768)", style=solid];
-"897 dequantize_per_channel_default_45" -> "899 linear_44"  [label="(2304, 768)", style=solid];
-"898 _param_constant139_0_0" -> "899 linear_44"  [label="(2304,)", style=solid];
-"899 linear_44" -> "900 unflatten_11"  [label="(197, 1, 2304)", style=solid];
-"900 unflatten_11" -> "901 unsqueeze_11"  [label="(197, 1, 3, 768)", style=solid];
-"901 unsqueeze_11" -> "902 transpose_67"  [label="(1, 197, 1, 3, 768)", style=solid];
-"902 transpose_67" -> "903 squeeze_11"  [label="(3, 197, 1, 1, 768)", style=solid];
-"903 squeeze_11" -> "904 contiguous_11"  [label="(3, 197, 1, 768)", style=solid];
-"904 contiguous_11" -> "905 quantize_per_tensor_default_68"  [label="(3, 197, 1, 768)", style=solid];
-"904 contiguous_11" -> "908 quantize_per_tensor_default_69"  [label="(3, 197, 1, 768)", style=solid];
-"904 contiguous_11" -> "911 select_35"  [label="(3, 197, 1, 768)", style=solid];
-"905 quantize_per_tensor_default_68" -> "906 dequantize_per_tensor_default_68"  [label="(3, 197, 1, 768)", style=solid];
-"906 dequantize_per_tensor_default_68" -> "907 select_33"  [label="(3, 197, 1, 768)", style=solid];
-"907 select_33" -> "912 view_88"  [label="(197, 1, 768)", style=solid];
-"908 quantize_per_tensor_default_69" -> "909 dequantize_per_tensor_default_69"  [label="(3, 197, 1, 768)", style=solid];
-"909 dequantize_per_tensor_default_69" -> "910 select_34"  [label="(3, 197, 1, 768)", style=solid];
-"910 select_34" -> "914 view_89"  [label="(197, 1, 768)", style=solid];
-"911 select_35" -> "916 view_90"  [label="(197, 1, 768)", style=solid];
-"912 view_88" -> "913 transpose_68"  [label="(197, 12, 64)", style=solid];
-"913 transpose_68" -> "918 view_91"  [label="(12, 197, 64)", style=solid];
-"914 view_89" -> "915 transpose_69"  [label="(197, 12, 64)", style=solid];
-"915 transpose_69" -> "919 view_92"  [label="(12, 197, 64)", style=solid];
-"916 view_90" -> "917 transpose_70"  [label="(197, 12, 64)", style=solid];
-"917 transpose_70" -> "920 view_93"  [label="(12, 197, 64)", style=solid];
-"918 view_91" -> "921 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
-"919 view_92" -> "921 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
-"920 view_93" -> "921 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
-"921 scaled_dot_product_attention_11" -> "922 permute_12"  [label="(1, 12, 197, 64)", style=solid];
-"922 permute_12" -> "923 view_94"  [label="(197, 1, 12, 64)", style=solid];
-"923 view_94" -> "924 view_94_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
-"924 view_94_0_0_nncf_smooth_quant_0" -> "925 quantize_per_tensor_default_70"  [label="(197, 768)", style=solid];
-"925 quantize_per_tensor_default_70" -> "926 dequantize_per_tensor_default_70"  [label="(197, 768)", style=solid];
-"926 dequantize_per_tensor_default_70" -> "933 linear_45"  [label="(197, 768)", style=solid];
-"927 linear_45_scale_0" -> "930 quantize_per_channel_default_46"  [label="(768,)", style=solid];
-"927 linear_45_scale_0" -> "931 dequantize_per_channel_default_46"  [label="(768,)", style=solid];
-"928 linear_45_zero_point_0" -> "930 quantize_per_channel_default_46"  [label="(768,)", style=solid];
-"928 linear_45_zero_point_0" -> "931 dequantize_per_channel_default_46"  [label="(768,)", style=solid];
-"929 compressed_weight_updated_constant46" -> "930 quantize_per_channel_default_46"  [label="(768, 768)", style=solid];
-"930 quantize_per_channel_default_46" -> "931 dequantize_per_channel_default_46"  [label="(768, 768)", style=solid];
-"931 dequantize_per_channel_default_46" -> "933 linear_45"  [label="(768, 768)", style=solid];
-"932 _param_constant141_0_0" -> "933 linear_45"  [label="(768,)", style=solid];
-"933 linear_45" -> "934 view_95"  [label="(197, 768)", style=solid];
-"934 view_95" -> "935 transpose_71"  [label="(197, 1, 768)", style=solid];
-"935 transpose_71" -> "936 dropout_34"  [label="(1, 197, 768)", style=solid];
-"936 dropout_34" -> "937 add_23"  [label="(1, 197, 768)", style=solid];
-"937 add_23" -> "940 layer_norm_23"  [label="(1, 197, 768)", style=solid];
-"937 add_23" -> "964 add_24"  [label="(1, 197, 768)", style=solid];
-"938 _param_constant142" -> "940 layer_norm_23"  [label="(768,)", style=solid];
-"939 _param_constant143" -> "940 layer_norm_23"  [label="(768,)", style=solid];
-"940 layer_norm_23" -> "941 layer_norm_23_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
-"941 layer_norm_23_0_0_nncf_smooth_quant_0" -> "942 quantize_per_tensor_default_71"  [label="(1, 197, 768)", style=solid];
-"942 quantize_per_tensor_default_71" -> "943 dequantize_per_tensor_default_71"  [label="(1, 197, 768)", style=solid];
-"943 dequantize_per_tensor_default_71" -> "950 linear_46"  [label="(1, 197, 768)", style=solid];
-"944 linear_46_scale_0" -> "947 quantize_per_channel_default_47"  [label="(3072,)", style=solid];
-"944 linear_46_scale_0" -> "948 dequantize_per_channel_default_47"  [label="(3072,)", style=solid];
-"945 linear_46_zero_point_0" -> "947 quantize_per_channel_default_47"  [label="(3072,)", style=solid];
-"945 linear_46_zero_point_0" -> "948 dequantize_per_channel_default_47"  [label="(3072,)", style=solid];
-"946 compressed_weight_updated_constant47" -> "947 quantize_per_channel_default_47"  [label="(3072, 768)", style=solid];
-"947 quantize_per_channel_default_47" -> "948 dequantize_per_channel_default_47"  [label="(3072, 768)", style=solid];
-"948 dequantize_per_channel_default_47" -> "950 linear_46"  [label="(3072, 768)", style=solid];
-"949 _param_constant145_0_0" -> "950 linear_46"  [label="(3072,)", style=solid];
-"950 linear_46" -> "951 gelu_11"  [label="(1, 197, 3072)", style=solid];
-"951 gelu_11" -> "952 dropout_35"  [label="(1, 197, 3072)", style=solid];
-"952 dropout_35" -> "953 dropout_35_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
-"953 dropout_35_0_0_nncf_smooth_quant_0" -> "954 quantize_per_tensor_default_72"  [label="(1, 197, 3072)", style=solid];
-"954 quantize_per_tensor_default_72" -> "955 dequantize_per_tensor_default_72"  [label="(1, 197, 3072)", style=solid];
-"955 dequantize_per_tensor_default_72" -> "962 linear_47"  [label="(1, 197, 3072)", style=solid];
-"956 linear_47_scale_0" -> "959 quantize_per_channel_default_48"  [label="(768,)", style=solid];
-"956 linear_47_scale_0" -> "960 dequantize_per_channel_default_48"  [label="(768,)", style=solid];
-"957 linear_47_zero_point_0" -> "959 quantize_per_channel_default_48"  [label="(768,)", style=solid];
-"957 linear_47_zero_point_0" -> "960 dequantize_per_channel_default_48"  [label="(768,)", style=solid];
-"958 compressed_weight_updated_constant48" -> "959 quantize_per_channel_default_48"  [label="(768, 3072)", style=solid];
-"959 quantize_per_channel_default_48" -> "960 dequantize_per_channel_default_48"  [label="(768, 3072)", style=solid];
-"960 dequantize_per_channel_default_48" -> "962 linear_47"  [label="(768, 3072)", style=solid];
-"961 _param_constant147_0_0" -> "962 linear_47"  [label="(768,)", style=solid];
-"962 linear_47" -> "963 dropout_36"  [label="(1, 197, 768)", style=solid];
-"963 dropout_36" -> "964 add_24"  [label="(1, 197, 768)", style=solid];
-"964 add_24" -> "967 layer_norm_24"  [label="(1, 197, 768)", style=solid];
-"965 _param_constant148" -> "967 layer_norm_24"  [label="(768,)", style=solid];
-"966 _param_constant149" -> "967 layer_norm_24"  [label="(768,)", style=solid];
+"0 encoder_pos_embedding" -> "65 add"  [label="(1, 197, 768)", style=solid];
+"1 encoder_layers_encoder_layer_0_ln_1_weight" -> "67 layer_norm"  [label="(768,)", style=solid];
+"2 encoder_layers_encoder_layer_0_ln_1_bias" -> "67 layer_norm"  [label="(768,)", style=solid];
+"3 encoder_layers_encoder_layer_0_ln_2_weight" -> "117 layer_norm_1"  [label="(768,)", style=solid];
+"4 encoder_layers_encoder_layer_0_ln_2_bias" -> "117 layer_norm_1"  [label="(768,)", style=solid];
+"5 encoder_layers_encoder_layer_1_ln_1_weight" -> "142 layer_norm_2"  [label="(768,)", style=solid];
+"6 encoder_layers_encoder_layer_1_ln_1_bias" -> "142 layer_norm_2"  [label="(768,)", style=solid];
+"7 encoder_layers_encoder_layer_1_ln_2_weight" -> "192 layer_norm_3"  [label="(768,)", style=solid];
+"8 encoder_layers_encoder_layer_1_ln_2_bias" -> "192 layer_norm_3"  [label="(768,)", style=solid];
+"9 encoder_layers_encoder_layer_2_ln_1_weight" -> "217 layer_norm_4"  [label="(768,)", style=solid];
+"10 encoder_layers_encoder_layer_2_ln_1_bias" -> "217 layer_norm_4"  [label="(768,)", style=solid];
+"11 encoder_layers_encoder_layer_2_ln_2_weight" -> "267 layer_norm_5"  [label="(768,)", style=solid];
+"12 encoder_layers_encoder_layer_2_ln_2_bias" -> "267 layer_norm_5"  [label="(768,)", style=solid];
+"13 encoder_layers_encoder_layer_3_ln_1_weight" -> "292 layer_norm_6"  [label="(768,)", style=solid];
+"14 encoder_layers_encoder_layer_3_ln_1_bias" -> "292 layer_norm_6"  [label="(768,)", style=solid];
+"15 encoder_layers_encoder_layer_3_ln_2_weight" -> "342 layer_norm_7"  [label="(768,)", style=solid];
+"16 encoder_layers_encoder_layer_3_ln_2_bias" -> "342 layer_norm_7"  [label="(768,)", style=solid];
+"17 encoder_layers_encoder_layer_4_ln_1_weight" -> "367 layer_norm_8"  [label="(768,)", style=solid];
+"18 encoder_layers_encoder_layer_4_ln_1_bias" -> "367 layer_norm_8"  [label="(768,)", style=solid];
+"19 encoder_layers_encoder_layer_4_ln_2_weight" -> "417 layer_norm_9"  [label="(768,)", style=solid];
+"20 encoder_layers_encoder_layer_4_ln_2_bias" -> "417 layer_norm_9"  [label="(768,)", style=solid];
+"21 encoder_layers_encoder_layer_5_ln_1_weight" -> "442 layer_norm_10"  [label="(768,)", style=solid];
+"22 encoder_layers_encoder_layer_5_ln_1_bias" -> "442 layer_norm_10"  [label="(768,)", style=solid];
+"23 encoder_layers_encoder_layer_5_ln_2_weight" -> "492 layer_norm_11"  [label="(768,)", style=solid];
+"24 encoder_layers_encoder_layer_5_ln_2_bias" -> "492 layer_norm_11"  [label="(768,)", style=solid];
+"25 encoder_layers_encoder_layer_6_ln_1_weight" -> "517 layer_norm_12"  [label="(768,)", style=solid];
+"26 encoder_layers_encoder_layer_6_ln_1_bias" -> "517 layer_norm_12"  [label="(768,)", style=solid];
+"27 encoder_layers_encoder_layer_6_ln_2_weight" -> "567 layer_norm_13"  [label="(768,)", style=solid];
+"28 encoder_layers_encoder_layer_6_ln_2_bias" -> "567 layer_norm_13"  [label="(768,)", style=solid];
+"29 encoder_layers_encoder_layer_7_ln_1_weight" -> "592 layer_norm_14"  [label="(768,)", style=solid];
+"30 encoder_layers_encoder_layer_7_ln_1_bias" -> "592 layer_norm_14"  [label="(768,)", style=solid];
+"31 encoder_layers_encoder_layer_7_ln_2_weight" -> "642 layer_norm_15"  [label="(768,)", style=solid];
+"32 encoder_layers_encoder_layer_7_ln_2_bias" -> "642 layer_norm_15"  [label="(768,)", style=solid];
+"33 encoder_layers_encoder_layer_8_ln_1_weight" -> "667 layer_norm_16"  [label="(768,)", style=solid];
+"34 encoder_layers_encoder_layer_8_ln_1_bias" -> "667 layer_norm_16"  [label="(768,)", style=solid];
+"35 encoder_layers_encoder_layer_8_ln_2_weight" -> "717 layer_norm_17"  [label="(768,)", style=solid];
+"36 encoder_layers_encoder_layer_8_ln_2_bias" -> "717 layer_norm_17"  [label="(768,)", style=solid];
+"37 encoder_layers_encoder_layer_9_ln_1_weight" -> "742 layer_norm_18"  [label="(768,)", style=solid];
+"38 encoder_layers_encoder_layer_9_ln_1_bias" -> "742 layer_norm_18"  [label="(768,)", style=solid];
+"39 encoder_layers_encoder_layer_9_ln_2_weight" -> "792 layer_norm_19"  [label="(768,)", style=solid];
+"40 encoder_layers_encoder_layer_9_ln_2_bias" -> "792 layer_norm_19"  [label="(768,)", style=solid];
+"41 encoder_layers_encoder_layer_10_ln_1_weight" -> "817 layer_norm_20"  [label="(768,)", style=solid];
+"42 encoder_layers_encoder_layer_10_ln_1_bias" -> "817 layer_norm_20"  [label="(768,)", style=solid];
+"43 encoder_layers_encoder_layer_10_ln_2_weight" -> "867 layer_norm_21"  [label="(768,)", style=solid];
+"44 encoder_layers_encoder_layer_10_ln_2_bias" -> "867 layer_norm_21"  [label="(768,)", style=solid];
+"45 encoder_layers_encoder_layer_11_ln_1_weight" -> "892 layer_norm_22"  [label="(768,)", style=solid];
+"46 encoder_layers_encoder_layer_11_ln_1_bias" -> "892 layer_norm_22"  [label="(768,)", style=solid];
+"47 encoder_layers_encoder_layer_11_ln_2_weight" -> "942 layer_norm_23"  [label="(768,)", style=solid];
+"48 encoder_layers_encoder_layer_11_ln_2_bias" -> "942 layer_norm_23"  [label="(768,)", style=solid];
+"49 encoder_ln_weight" -> "967 layer_norm_24"  [label="(768,)", style=solid];
+"50 encoder_ln_bias" -> "967 layer_norm_24"  [label="(768,)", style=solid];
+"51 x" -> "52 quantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"52 quantize_per_tensor_default" -> "53 dequantize_per_tensor_default"  [label="(1, 3, 224, 224)", style=solid];
+"53 dequantize_per_tensor_default" -> "60 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"54 conv2d_scale_0" -> "57 quantize_per_channel_default"  [label="(768,)", style=solid];
+"54 conv2d_scale_0" -> "58 dequantize_per_channel_default"  [label="(768,)", style=solid];
+"55 conv2d_zero_point_0" -> "57 quantize_per_channel_default"  [label="(768,)", style=solid];
+"55 conv2d_zero_point_0" -> "58 dequantize_per_channel_default"  [label="(768,)", style=solid];
+"56 compressed_weight_updated_constant0" -> "57 quantize_per_channel_default"  [label="(768, 3, 16, 16)", style=solid];
+"57 quantize_per_channel_default" -> "58 dequantize_per_channel_default"  [label="(768, 3, 16, 16)", style=solid];
+"58 dequantize_per_channel_default" -> "60 conv2d"  [label="(768, 3, 16, 16)", style=solid];
+"59 conv_proj_bias_0_0" -> "60 conv2d"  [label="(768,)", style=solid];
+"60 conv2d" -> "61 reshape"  [label="(1, 768, 14, 14)", style=solid];
+"61 reshape" -> "62 permute"  [label="(1, 768, 196)", style=solid];
+"62 permute" -> "64 cat"  [label="(1, 196, 768)", style=solid];
+"63 _frozen_param0" -> "64 cat"  [label="(1, 1, 768)", style=solid];
+"64 cat" -> "65 add"  [label="(1, 197, 768)", style=solid];
+"65 add" -> "66 dropout"  [label="(1, 197, 768)", style=solid];
+"66 dropout" -> "67 layer_norm"  [label="(1, 197, 768)", style=solid];
+"66 dropout" -> "116 add_1"  [label="(1, 197, 768)", style=solid];
+"67 layer_norm" -> "68 transpose"  [label="(1, 197, 768)", style=solid];
+"68 transpose" -> "69 transpose_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"69 transpose_0_0_nncf_smooth_quant_0" -> "70 quantize_per_tensor_default_1"  [label="(197, 1, 768)", style=solid];
+"70 quantize_per_tensor_default_1" -> "71 dequantize_per_tensor_default_1"  [label="(197, 1, 768)", style=solid];
+"71 dequantize_per_tensor_default_1" -> "78 linear"  [label="(197, 1, 768)", style=solid];
+"72 linear_scale_0" -> "75 quantize_per_channel_default_1"  [label="(2304,)", style=solid];
+"72 linear_scale_0" -> "76 dequantize_per_channel_default_1"  [label="(2304,)", style=solid];
+"73 linear_zero_point_0" -> "75 quantize_per_channel_default_1"  [label="(2304,)", style=solid];
+"73 linear_zero_point_0" -> "76 dequantize_per_channel_default_1"  [label="(2304,)", style=solid];
+"74 compressed_weight_updated_constant1" -> "75 quantize_per_channel_default_1"  [label="(2304, 768)", style=solid];
+"75 quantize_per_channel_default_1" -> "76 dequantize_per_channel_default_1"  [label="(2304, 768)", style=solid];
+"76 dequantize_per_channel_default_1" -> "78 linear"  [label="(2304, 768)", style=solid];
+"77 encoder_layers_encoder_layer_0_self_attention_in_proj_bias_0_0" -> "78 linear"  [label="(2304,)", style=solid];
+"78 linear" -> "79 unflatten"  [label="(197, 1, 2304)", style=solid];
+"79 unflatten" -> "80 unsqueeze"  [label="(197, 1, 3, 768)", style=solid];
+"80 unsqueeze" -> "81 transpose_1"  [label="(1, 197, 1, 3, 768)", style=solid];
+"81 transpose_1" -> "82 squeeze"  [label="(3, 197, 1, 1, 768)", style=solid];
+"82 squeeze" -> "83 contiguous"  [label="(3, 197, 1, 768)", style=solid];
+"83 contiguous" -> "84 quantize_per_tensor_default_2"  [label="(3, 197, 1, 768)", style=solid];
+"83 contiguous" -> "87 quantize_per_tensor_default_3"  [label="(3, 197, 1, 768)", style=solid];
+"83 contiguous" -> "90 select_2"  [label="(3, 197, 1, 768)", style=solid];
+"84 quantize_per_tensor_default_2" -> "85 dequantize_per_tensor_default_2"  [label="(3, 197, 1, 768)", style=solid];
+"85 dequantize_per_tensor_default_2" -> "86 select"  [label="(3, 197, 1, 768)", style=solid];
+"86 select" -> "91 view"  [label="(197, 1, 768)", style=solid];
+"87 quantize_per_tensor_default_3" -> "88 dequantize_per_tensor_default_3"  [label="(3, 197, 1, 768)", style=solid];
+"88 dequantize_per_tensor_default_3" -> "89 select_1"  [label="(3, 197, 1, 768)", style=solid];
+"89 select_1" -> "93 view_1"  [label="(197, 1, 768)", style=solid];
+"90 select_2" -> "95 view_2"  [label="(197, 1, 768)", style=solid];
+"91 view" -> "92 transpose_2"  [label="(197, 12, 64)", style=solid];
+"92 transpose_2" -> "97 view_3"  [label="(12, 197, 64)", style=solid];
+"93 view_1" -> "94 transpose_3"  [label="(197, 12, 64)", style=solid];
+"94 transpose_3" -> "98 view_4"  [label="(12, 197, 64)", style=solid];
+"95 view_2" -> "96 transpose_4"  [label="(197, 12, 64)", style=solid];
+"96 transpose_4" -> "99 view_5"  [label="(12, 197, 64)", style=solid];
+"97 view_3" -> "100 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
+"98 view_4" -> "100 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
+"99 view_5" -> "100 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
+"100 scaled_dot_product_attention" -> "101 permute_1"  [label="(1, 12, 197, 64)", style=solid];
+"101 permute_1" -> "102 view_6"  [label="(197, 1, 12, 64)", style=solid];
+"102 view_6" -> "103 view_6_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"103 view_6_0_0_nncf_smooth_quant_0" -> "104 quantize_per_tensor_default_4"  [label="(197, 768)", style=solid];
+"104 quantize_per_tensor_default_4" -> "105 dequantize_per_tensor_default_4"  [label="(197, 768)", style=solid];
+"105 dequantize_per_tensor_default_4" -> "112 linear_1"  [label="(197, 768)", style=solid];
+"106 linear_1_scale_0" -> "109 quantize_per_channel_default_2"  [label="(768,)", style=solid];
+"106 linear_1_scale_0" -> "110 dequantize_per_channel_default_2"  [label="(768,)", style=solid];
+"107 linear_1_zero_point_0" -> "109 quantize_per_channel_default_2"  [label="(768,)", style=solid];
+"107 linear_1_zero_point_0" -> "110 dequantize_per_channel_default_2"  [label="(768,)", style=solid];
+"108 compressed_weight_updated_constant2" -> "109 quantize_per_channel_default_2"  [label="(768, 768)", style=solid];
+"109 quantize_per_channel_default_2" -> "110 dequantize_per_channel_default_2"  [label="(768, 768)", style=solid];
+"110 dequantize_per_channel_default_2" -> "112 linear_1"  [label="(768, 768)", style=solid];
+"111 encoder_layers_encoder_layer_0_self_attention_out_proj_bias_0_0" -> "112 linear_1"  [label="(768,)", style=solid];
+"112 linear_1" -> "113 view_7"  [label="(197, 768)", style=solid];
+"113 view_7" -> "114 transpose_5"  [label="(197, 1, 768)", style=solid];
+"114 transpose_5" -> "115 dropout_1"  [label="(1, 197, 768)", style=solid];
+"115 dropout_1" -> "116 add_1"  [label="(1, 197, 768)", style=solid];
+"116 add_1" -> "117 layer_norm_1"  [label="(1, 197, 768)", style=solid];
+"116 add_1" -> "141 add_2"  [label="(1, 197, 768)", style=solid];
+"117 layer_norm_1" -> "118 layer_norm_1_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"118 layer_norm_1_0_0_nncf_smooth_quant_0" -> "119 quantize_per_tensor_default_5"  [label="(1, 197, 768)", style=solid];
+"119 quantize_per_tensor_default_5" -> "120 dequantize_per_tensor_default_5"  [label="(1, 197, 768)", style=solid];
+"120 dequantize_per_tensor_default_5" -> "127 linear_2"  [label="(1, 197, 768)", style=solid];
+"121 linear_2_scale_0" -> "124 quantize_per_channel_default_3"  [label="(3072,)", style=solid];
+"121 linear_2_scale_0" -> "125 dequantize_per_channel_default_3"  [label="(3072,)", style=solid];
+"122 linear_2_zero_point_0" -> "124 quantize_per_channel_default_3"  [label="(3072,)", style=solid];
+"122 linear_2_zero_point_0" -> "125 dequantize_per_channel_default_3"  [label="(3072,)", style=solid];
+"123 compressed_weight_updated_constant3" -> "124 quantize_per_channel_default_3"  [label="(3072, 768)", style=solid];
+"124 quantize_per_channel_default_3" -> "125 dequantize_per_channel_default_3"  [label="(3072, 768)", style=solid];
+"125 dequantize_per_channel_default_3" -> "127 linear_2"  [label="(3072, 768)", style=solid];
+"126 encoder_layers_encoder_layer_0_mlp_0_bias_0_0" -> "127 linear_2"  [label="(3072,)", style=solid];
+"127 linear_2" -> "128 gelu"  [label="(1, 197, 3072)", style=solid];
+"128 gelu" -> "129 dropout_2"  [label="(1, 197, 3072)", style=solid];
+"129 dropout_2" -> "130 dropout_2_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"130 dropout_2_0_0_nncf_smooth_quant_0" -> "131 quantize_per_tensor_default_6"  [label="(1, 197, 3072)", style=solid];
+"131 quantize_per_tensor_default_6" -> "132 dequantize_per_tensor_default_6"  [label="(1, 197, 3072)", style=solid];
+"132 dequantize_per_tensor_default_6" -> "139 linear_3"  [label="(1, 197, 3072)", style=solid];
+"133 linear_3_scale_0" -> "136 quantize_per_channel_default_4"  [label="(768,)", style=solid];
+"133 linear_3_scale_0" -> "137 dequantize_per_channel_default_4"  [label="(768,)", style=solid];
+"134 linear_3_zero_point_0" -> "136 quantize_per_channel_default_4"  [label="(768,)", style=solid];
+"134 linear_3_zero_point_0" -> "137 dequantize_per_channel_default_4"  [label="(768,)", style=solid];
+"135 compressed_weight_updated_constant4" -> "136 quantize_per_channel_default_4"  [label="(768, 3072)", style=solid];
+"136 quantize_per_channel_default_4" -> "137 dequantize_per_channel_default_4"  [label="(768, 3072)", style=solid];
+"137 dequantize_per_channel_default_4" -> "139 linear_3"  [label="(768, 3072)", style=solid];
+"138 encoder_layers_encoder_layer_0_mlp_3_bias_0_0" -> "139 linear_3"  [label="(768,)", style=solid];
+"139 linear_3" -> "140 dropout_3"  [label="(1, 197, 768)", style=solid];
+"140 dropout_3" -> "141 add_2"  [label="(1, 197, 768)", style=solid];
+"141 add_2" -> "142 layer_norm_2"  [label="(1, 197, 768)", style=solid];
+"141 add_2" -> "191 add_3"  [label="(1, 197, 768)", style=solid];
+"142 layer_norm_2" -> "143 transpose_6"  [label="(1, 197, 768)", style=solid];
+"143 transpose_6" -> "144 transpose_6_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"144 transpose_6_0_0_nncf_smooth_quant_0" -> "145 quantize_per_tensor_default_7"  [label="(197, 1, 768)", style=solid];
+"145 quantize_per_tensor_default_7" -> "146 dequantize_per_tensor_default_7"  [label="(197, 1, 768)", style=solid];
+"146 dequantize_per_tensor_default_7" -> "153 linear_4"  [label="(197, 1, 768)", style=solid];
+"147 linear_4_scale_0" -> "150 quantize_per_channel_default_5"  [label="(2304,)", style=solid];
+"147 linear_4_scale_0" -> "151 dequantize_per_channel_default_5"  [label="(2304,)", style=solid];
+"148 linear_4_zero_point_0" -> "150 quantize_per_channel_default_5"  [label="(2304,)", style=solid];
+"148 linear_4_zero_point_0" -> "151 dequantize_per_channel_default_5"  [label="(2304,)", style=solid];
+"149 compressed_weight_updated_constant5" -> "150 quantize_per_channel_default_5"  [label="(2304, 768)", style=solid];
+"150 quantize_per_channel_default_5" -> "151 dequantize_per_channel_default_5"  [label="(2304, 768)", style=solid];
+"151 dequantize_per_channel_default_5" -> "153 linear_4"  [label="(2304, 768)", style=solid];
+"152 encoder_layers_encoder_layer_1_self_attention_in_proj_bias_0_0" -> "153 linear_4"  [label="(2304,)", style=solid];
+"153 linear_4" -> "154 unflatten_1"  [label="(197, 1, 2304)", style=solid];
+"154 unflatten_1" -> "155 unsqueeze_1"  [label="(197, 1, 3, 768)", style=solid];
+"155 unsqueeze_1" -> "156 transpose_7"  [label="(1, 197, 1, 3, 768)", style=solid];
+"156 transpose_7" -> "157 squeeze_1"  [label="(3, 197, 1, 1, 768)", style=solid];
+"157 squeeze_1" -> "158 contiguous_1"  [label="(3, 197, 1, 768)", style=solid];
+"158 contiguous_1" -> "159 quantize_per_tensor_default_8"  [label="(3, 197, 1, 768)", style=solid];
+"158 contiguous_1" -> "162 quantize_per_tensor_default_9"  [label="(3, 197, 1, 768)", style=solid];
+"158 contiguous_1" -> "165 select_5"  [label="(3, 197, 1, 768)", style=solid];
+"159 quantize_per_tensor_default_8" -> "160 dequantize_per_tensor_default_8"  [label="(3, 197, 1, 768)", style=solid];
+"160 dequantize_per_tensor_default_8" -> "161 select_3"  [label="(3, 197, 1, 768)", style=solid];
+"161 select_3" -> "166 view_8"  [label="(197, 1, 768)", style=solid];
+"162 quantize_per_tensor_default_9" -> "163 dequantize_per_tensor_default_9"  [label="(3, 197, 1, 768)", style=solid];
+"163 dequantize_per_tensor_default_9" -> "164 select_4"  [label="(3, 197, 1, 768)", style=solid];
+"164 select_4" -> "168 view_9"  [label="(197, 1, 768)", style=solid];
+"165 select_5" -> "170 view_10"  [label="(197, 1, 768)", style=solid];
+"166 view_8" -> "167 transpose_8"  [label="(197, 12, 64)", style=solid];
+"167 transpose_8" -> "172 view_11"  [label="(12, 197, 64)", style=solid];
+"168 view_9" -> "169 transpose_9"  [label="(197, 12, 64)", style=solid];
+"169 transpose_9" -> "173 view_12"  [label="(12, 197, 64)", style=solid];
+"170 view_10" -> "171 transpose_10"  [label="(197, 12, 64)", style=solid];
+"171 transpose_10" -> "174 view_13"  [label="(12, 197, 64)", style=solid];
+"172 view_11" -> "175 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
+"173 view_12" -> "175 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
+"174 view_13" -> "175 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
+"175 scaled_dot_product_attention_1" -> "176 permute_2"  [label="(1, 12, 197, 64)", style=solid];
+"176 permute_2" -> "177 view_14"  [label="(197, 1, 12, 64)", style=solid];
+"177 view_14" -> "178 view_14_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"178 view_14_0_0_nncf_smooth_quant_0" -> "179 quantize_per_tensor_default_10"  [label="(197, 768)", style=solid];
+"179 quantize_per_tensor_default_10" -> "180 dequantize_per_tensor_default_10"  [label="(197, 768)", style=solid];
+"180 dequantize_per_tensor_default_10" -> "187 linear_5"  [label="(197, 768)", style=solid];
+"181 linear_5_scale_0" -> "184 quantize_per_channel_default_6"  [label="(768,)", style=solid];
+"181 linear_5_scale_0" -> "185 dequantize_per_channel_default_6"  [label="(768,)", style=solid];
+"182 linear_5_zero_point_0" -> "184 quantize_per_channel_default_6"  [label="(768,)", style=solid];
+"182 linear_5_zero_point_0" -> "185 dequantize_per_channel_default_6"  [label="(768,)", style=solid];
+"183 compressed_weight_updated_constant6" -> "184 quantize_per_channel_default_6"  [label="(768, 768)", style=solid];
+"184 quantize_per_channel_default_6" -> "185 dequantize_per_channel_default_6"  [label="(768, 768)", style=solid];
+"185 dequantize_per_channel_default_6" -> "187 linear_5"  [label="(768, 768)", style=solid];
+"186 encoder_layers_encoder_layer_1_self_attention_out_proj_bias_0_0" -> "187 linear_5"  [label="(768,)", style=solid];
+"187 linear_5" -> "188 view_15"  [label="(197, 768)", style=solid];
+"188 view_15" -> "189 transpose_11"  [label="(197, 1, 768)", style=solid];
+"189 transpose_11" -> "190 dropout_4"  [label="(1, 197, 768)", style=solid];
+"190 dropout_4" -> "191 add_3"  [label="(1, 197, 768)", style=solid];
+"191 add_3" -> "192 layer_norm_3"  [label="(1, 197, 768)", style=solid];
+"191 add_3" -> "216 add_4"  [label="(1, 197, 768)", style=solid];
+"192 layer_norm_3" -> "193 layer_norm_3_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"193 layer_norm_3_0_0_nncf_smooth_quant_0" -> "194 quantize_per_tensor_default_11"  [label="(1, 197, 768)", style=solid];
+"194 quantize_per_tensor_default_11" -> "195 dequantize_per_tensor_default_11"  [label="(1, 197, 768)", style=solid];
+"195 dequantize_per_tensor_default_11" -> "202 linear_6"  [label="(1, 197, 768)", style=solid];
+"196 linear_6_scale_0" -> "199 quantize_per_channel_default_7"  [label="(3072,)", style=solid];
+"196 linear_6_scale_0" -> "200 dequantize_per_channel_default_7"  [label="(3072,)", style=solid];
+"197 linear_6_zero_point_0" -> "199 quantize_per_channel_default_7"  [label="(3072,)", style=solid];
+"197 linear_6_zero_point_0" -> "200 dequantize_per_channel_default_7"  [label="(3072,)", style=solid];
+"198 compressed_weight_updated_constant7" -> "199 quantize_per_channel_default_7"  [label="(3072, 768)", style=solid];
+"199 quantize_per_channel_default_7" -> "200 dequantize_per_channel_default_7"  [label="(3072, 768)", style=solid];
+"200 dequantize_per_channel_default_7" -> "202 linear_6"  [label="(3072, 768)", style=solid];
+"201 encoder_layers_encoder_layer_1_mlp_0_bias_0_0" -> "202 linear_6"  [label="(3072,)", style=solid];
+"202 linear_6" -> "203 gelu_1"  [label="(1, 197, 3072)", style=solid];
+"203 gelu_1" -> "204 dropout_5"  [label="(1, 197, 3072)", style=solid];
+"204 dropout_5" -> "205 dropout_5_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"205 dropout_5_0_0_nncf_smooth_quant_0" -> "206 quantize_per_tensor_default_12"  [label="(1, 197, 3072)", style=solid];
+"206 quantize_per_tensor_default_12" -> "207 dequantize_per_tensor_default_12"  [label="(1, 197, 3072)", style=solid];
+"207 dequantize_per_tensor_default_12" -> "214 linear_7"  [label="(1, 197, 3072)", style=solid];
+"208 linear_7_scale_0" -> "211 quantize_per_channel_default_8"  [label="(768,)", style=solid];
+"208 linear_7_scale_0" -> "212 dequantize_per_channel_default_8"  [label="(768,)", style=solid];
+"209 linear_7_zero_point_0" -> "211 quantize_per_channel_default_8"  [label="(768,)", style=solid];
+"209 linear_7_zero_point_0" -> "212 dequantize_per_channel_default_8"  [label="(768,)", style=solid];
+"210 compressed_weight_updated_constant8" -> "211 quantize_per_channel_default_8"  [label="(768, 3072)", style=solid];
+"211 quantize_per_channel_default_8" -> "212 dequantize_per_channel_default_8"  [label="(768, 3072)", style=solid];
+"212 dequantize_per_channel_default_8" -> "214 linear_7"  [label="(768, 3072)", style=solid];
+"213 encoder_layers_encoder_layer_1_mlp_3_bias_0_0" -> "214 linear_7"  [label="(768,)", style=solid];
+"214 linear_7" -> "215 dropout_6"  [label="(1, 197, 768)", style=solid];
+"215 dropout_6" -> "216 add_4"  [label="(1, 197, 768)", style=solid];
+"216 add_4" -> "217 layer_norm_4"  [label="(1, 197, 768)", style=solid];
+"216 add_4" -> "266 add_5"  [label="(1, 197, 768)", style=solid];
+"217 layer_norm_4" -> "218 transpose_12"  [label="(1, 197, 768)", style=solid];
+"218 transpose_12" -> "219 transpose_12_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"219 transpose_12_0_0_nncf_smooth_quant_0" -> "220 quantize_per_tensor_default_13"  [label="(197, 1, 768)", style=solid];
+"220 quantize_per_tensor_default_13" -> "221 dequantize_per_tensor_default_13"  [label="(197, 1, 768)", style=solid];
+"221 dequantize_per_tensor_default_13" -> "228 linear_8"  [label="(197, 1, 768)", style=solid];
+"222 linear_8_scale_0" -> "225 quantize_per_channel_default_9"  [label="(2304,)", style=solid];
+"222 linear_8_scale_0" -> "226 dequantize_per_channel_default_9"  [label="(2304,)", style=solid];
+"223 linear_8_zero_point_0" -> "225 quantize_per_channel_default_9"  [label="(2304,)", style=solid];
+"223 linear_8_zero_point_0" -> "226 dequantize_per_channel_default_9"  [label="(2304,)", style=solid];
+"224 compressed_weight_updated_constant9" -> "225 quantize_per_channel_default_9"  [label="(2304, 768)", style=solid];
+"225 quantize_per_channel_default_9" -> "226 dequantize_per_channel_default_9"  [label="(2304, 768)", style=solid];
+"226 dequantize_per_channel_default_9" -> "228 linear_8"  [label="(2304, 768)", style=solid];
+"227 encoder_layers_encoder_layer_2_self_attention_in_proj_bias_0_0" -> "228 linear_8"  [label="(2304,)", style=solid];
+"228 linear_8" -> "229 unflatten_2"  [label="(197, 1, 2304)", style=solid];
+"229 unflatten_2" -> "230 unsqueeze_2"  [label="(197, 1, 3, 768)", style=solid];
+"230 unsqueeze_2" -> "231 transpose_13"  [label="(1, 197, 1, 3, 768)", style=solid];
+"231 transpose_13" -> "232 squeeze_2"  [label="(3, 197, 1, 1, 768)", style=solid];
+"232 squeeze_2" -> "233 contiguous_2"  [label="(3, 197, 1, 768)", style=solid];
+"233 contiguous_2" -> "234 quantize_per_tensor_default_14"  [label="(3, 197, 1, 768)", style=solid];
+"233 contiguous_2" -> "237 quantize_per_tensor_default_15"  [label="(3, 197, 1, 768)", style=solid];
+"233 contiguous_2" -> "240 select_8"  [label="(3, 197, 1, 768)", style=solid];
+"234 quantize_per_tensor_default_14" -> "235 dequantize_per_tensor_default_14"  [label="(3, 197, 1, 768)", style=solid];
+"235 dequantize_per_tensor_default_14" -> "236 select_6"  [label="(3, 197, 1, 768)", style=solid];
+"236 select_6" -> "241 view_16"  [label="(197, 1, 768)", style=solid];
+"237 quantize_per_tensor_default_15" -> "238 dequantize_per_tensor_default_15"  [label="(3, 197, 1, 768)", style=solid];
+"238 dequantize_per_tensor_default_15" -> "239 select_7"  [label="(3, 197, 1, 768)", style=solid];
+"239 select_7" -> "243 view_17"  [label="(197, 1, 768)", style=solid];
+"240 select_8" -> "245 view_18"  [label="(197, 1, 768)", style=solid];
+"241 view_16" -> "242 transpose_14"  [label="(197, 12, 64)", style=solid];
+"242 transpose_14" -> "247 view_19"  [label="(12, 197, 64)", style=solid];
+"243 view_17" -> "244 transpose_15"  [label="(197, 12, 64)", style=solid];
+"244 transpose_15" -> "248 view_20"  [label="(12, 197, 64)", style=solid];
+"245 view_18" -> "246 transpose_16"  [label="(197, 12, 64)", style=solid];
+"246 transpose_16" -> "249 view_21"  [label="(12, 197, 64)", style=solid];
+"247 view_19" -> "250 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
+"248 view_20" -> "250 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
+"249 view_21" -> "250 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
+"250 scaled_dot_product_attention_2" -> "251 permute_3"  [label="(1, 12, 197, 64)", style=solid];
+"251 permute_3" -> "252 view_22"  [label="(197, 1, 12, 64)", style=solid];
+"252 view_22" -> "253 view_22_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"253 view_22_0_0_nncf_smooth_quant_0" -> "254 quantize_per_tensor_default_16"  [label="(197, 768)", style=solid];
+"254 quantize_per_tensor_default_16" -> "255 dequantize_per_tensor_default_16"  [label="(197, 768)", style=solid];
+"255 dequantize_per_tensor_default_16" -> "262 linear_9"  [label="(197, 768)", style=solid];
+"256 linear_9_scale_0" -> "259 quantize_per_channel_default_10"  [label="(768,)", style=solid];
+"256 linear_9_scale_0" -> "260 dequantize_per_channel_default_10"  [label="(768,)", style=solid];
+"257 linear_9_zero_point_0" -> "259 quantize_per_channel_default_10"  [label="(768,)", style=solid];
+"257 linear_9_zero_point_0" -> "260 dequantize_per_channel_default_10"  [label="(768,)", style=solid];
+"258 compressed_weight_updated_constant10" -> "259 quantize_per_channel_default_10"  [label="(768, 768)", style=solid];
+"259 quantize_per_channel_default_10" -> "260 dequantize_per_channel_default_10"  [label="(768, 768)", style=solid];
+"260 dequantize_per_channel_default_10" -> "262 linear_9"  [label="(768, 768)", style=solid];
+"261 encoder_layers_encoder_layer_2_self_attention_out_proj_bias_0_0" -> "262 linear_9"  [label="(768,)", style=solid];
+"262 linear_9" -> "263 view_23"  [label="(197, 768)", style=solid];
+"263 view_23" -> "264 transpose_17"  [label="(197, 1, 768)", style=solid];
+"264 transpose_17" -> "265 dropout_7"  [label="(1, 197, 768)", style=solid];
+"265 dropout_7" -> "266 add_5"  [label="(1, 197, 768)", style=solid];
+"266 add_5" -> "267 layer_norm_5"  [label="(1, 197, 768)", style=solid];
+"266 add_5" -> "291 add_6"  [label="(1, 197, 768)", style=solid];
+"267 layer_norm_5" -> "268 layer_norm_5_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"268 layer_norm_5_0_0_nncf_smooth_quant_0" -> "269 quantize_per_tensor_default_17"  [label="(1, 197, 768)", style=solid];
+"269 quantize_per_tensor_default_17" -> "270 dequantize_per_tensor_default_17"  [label="(1, 197, 768)", style=solid];
+"270 dequantize_per_tensor_default_17" -> "277 linear_10"  [label="(1, 197, 768)", style=solid];
+"271 linear_10_scale_0" -> "274 quantize_per_channel_default_11"  [label="(3072,)", style=solid];
+"271 linear_10_scale_0" -> "275 dequantize_per_channel_default_11"  [label="(3072,)", style=solid];
+"272 linear_10_zero_point_0" -> "274 quantize_per_channel_default_11"  [label="(3072,)", style=solid];
+"272 linear_10_zero_point_0" -> "275 dequantize_per_channel_default_11"  [label="(3072,)", style=solid];
+"273 compressed_weight_updated_constant11" -> "274 quantize_per_channel_default_11"  [label="(3072, 768)", style=solid];
+"274 quantize_per_channel_default_11" -> "275 dequantize_per_channel_default_11"  [label="(3072, 768)", style=solid];
+"275 dequantize_per_channel_default_11" -> "277 linear_10"  [label="(3072, 768)", style=solid];
+"276 encoder_layers_encoder_layer_2_mlp_0_bias_0_0" -> "277 linear_10"  [label="(3072,)", style=solid];
+"277 linear_10" -> "278 gelu_2"  [label="(1, 197, 3072)", style=solid];
+"278 gelu_2" -> "279 dropout_8"  [label="(1, 197, 3072)", style=solid];
+"279 dropout_8" -> "280 dropout_8_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"280 dropout_8_0_0_nncf_smooth_quant_0" -> "281 quantize_per_tensor_default_18"  [label="(1, 197, 3072)", style=solid];
+"281 quantize_per_tensor_default_18" -> "282 dequantize_per_tensor_default_18"  [label="(1, 197, 3072)", style=solid];
+"282 dequantize_per_tensor_default_18" -> "289 linear_11"  [label="(1, 197, 3072)", style=solid];
+"283 linear_11_scale_0" -> "286 quantize_per_channel_default_12"  [label="(768,)", style=solid];
+"283 linear_11_scale_0" -> "287 dequantize_per_channel_default_12"  [label="(768,)", style=solid];
+"284 linear_11_zero_point_0" -> "286 quantize_per_channel_default_12"  [label="(768,)", style=solid];
+"284 linear_11_zero_point_0" -> "287 dequantize_per_channel_default_12"  [label="(768,)", style=solid];
+"285 compressed_weight_updated_constant12" -> "286 quantize_per_channel_default_12"  [label="(768, 3072)", style=solid];
+"286 quantize_per_channel_default_12" -> "287 dequantize_per_channel_default_12"  [label="(768, 3072)", style=solid];
+"287 dequantize_per_channel_default_12" -> "289 linear_11"  [label="(768, 3072)", style=solid];
+"288 encoder_layers_encoder_layer_2_mlp_3_bias_0_0" -> "289 linear_11"  [label="(768,)", style=solid];
+"289 linear_11" -> "290 dropout_9"  [label="(1, 197, 768)", style=solid];
+"290 dropout_9" -> "291 add_6"  [label="(1, 197, 768)", style=solid];
+"291 add_6" -> "292 layer_norm_6"  [label="(1, 197, 768)", style=solid];
+"291 add_6" -> "341 add_7"  [label="(1, 197, 768)", style=solid];
+"292 layer_norm_6" -> "293 transpose_18"  [label="(1, 197, 768)", style=solid];
+"293 transpose_18" -> "294 transpose_18_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"294 transpose_18_0_0_nncf_smooth_quant_0" -> "295 quantize_per_tensor_default_19"  [label="(197, 1, 768)", style=solid];
+"295 quantize_per_tensor_default_19" -> "296 dequantize_per_tensor_default_19"  [label="(197, 1, 768)", style=solid];
+"296 dequantize_per_tensor_default_19" -> "303 linear_12"  [label="(197, 1, 768)", style=solid];
+"297 linear_12_scale_0" -> "300 quantize_per_channel_default_13"  [label="(2304,)", style=solid];
+"297 linear_12_scale_0" -> "301 dequantize_per_channel_default_13"  [label="(2304,)", style=solid];
+"298 linear_12_zero_point_0" -> "300 quantize_per_channel_default_13"  [label="(2304,)", style=solid];
+"298 linear_12_zero_point_0" -> "301 dequantize_per_channel_default_13"  [label="(2304,)", style=solid];
+"299 compressed_weight_updated_constant13" -> "300 quantize_per_channel_default_13"  [label="(2304, 768)", style=solid];
+"300 quantize_per_channel_default_13" -> "301 dequantize_per_channel_default_13"  [label="(2304, 768)", style=solid];
+"301 dequantize_per_channel_default_13" -> "303 linear_12"  [label="(2304, 768)", style=solid];
+"302 encoder_layers_encoder_layer_3_self_attention_in_proj_bias_0_0" -> "303 linear_12"  [label="(2304,)", style=solid];
+"303 linear_12" -> "304 unflatten_3"  [label="(197, 1, 2304)", style=solid];
+"304 unflatten_3" -> "305 unsqueeze_3"  [label="(197, 1, 3, 768)", style=solid];
+"305 unsqueeze_3" -> "306 transpose_19"  [label="(1, 197, 1, 3, 768)", style=solid];
+"306 transpose_19" -> "307 squeeze_3"  [label="(3, 197, 1, 1, 768)", style=solid];
+"307 squeeze_3" -> "308 contiguous_3"  [label="(3, 197, 1, 768)", style=solid];
+"308 contiguous_3" -> "309 quantize_per_tensor_default_20"  [label="(3, 197, 1, 768)", style=solid];
+"308 contiguous_3" -> "312 quantize_per_tensor_default_21"  [label="(3, 197, 1, 768)", style=solid];
+"308 contiguous_3" -> "315 select_11"  [label="(3, 197, 1, 768)", style=solid];
+"309 quantize_per_tensor_default_20" -> "310 dequantize_per_tensor_default_20"  [label="(3, 197, 1, 768)", style=solid];
+"310 dequantize_per_tensor_default_20" -> "311 select_9"  [label="(3, 197, 1, 768)", style=solid];
+"311 select_9" -> "316 view_24"  [label="(197, 1, 768)", style=solid];
+"312 quantize_per_tensor_default_21" -> "313 dequantize_per_tensor_default_21"  [label="(3, 197, 1, 768)", style=solid];
+"313 dequantize_per_tensor_default_21" -> "314 select_10"  [label="(3, 197, 1, 768)", style=solid];
+"314 select_10" -> "318 view_25"  [label="(197, 1, 768)", style=solid];
+"315 select_11" -> "320 view_26"  [label="(197, 1, 768)", style=solid];
+"316 view_24" -> "317 transpose_20"  [label="(197, 12, 64)", style=solid];
+"317 transpose_20" -> "322 view_27"  [label="(12, 197, 64)", style=solid];
+"318 view_25" -> "319 transpose_21"  [label="(197, 12, 64)", style=solid];
+"319 transpose_21" -> "323 view_28"  [label="(12, 197, 64)", style=solid];
+"320 view_26" -> "321 transpose_22"  [label="(197, 12, 64)", style=solid];
+"321 transpose_22" -> "324 view_29"  [label="(12, 197, 64)", style=solid];
+"322 view_27" -> "325 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
+"323 view_28" -> "325 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
+"324 view_29" -> "325 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
+"325 scaled_dot_product_attention_3" -> "326 permute_4"  [label="(1, 12, 197, 64)", style=solid];
+"326 permute_4" -> "327 view_30"  [label="(197, 1, 12, 64)", style=solid];
+"327 view_30" -> "328 view_30_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"328 view_30_0_0_nncf_smooth_quant_0" -> "329 quantize_per_tensor_default_22"  [label="(197, 768)", style=solid];
+"329 quantize_per_tensor_default_22" -> "330 dequantize_per_tensor_default_22"  [label="(197, 768)", style=solid];
+"330 dequantize_per_tensor_default_22" -> "337 linear_13"  [label="(197, 768)", style=solid];
+"331 linear_13_scale_0" -> "334 quantize_per_channel_default_14"  [label="(768,)", style=solid];
+"331 linear_13_scale_0" -> "335 dequantize_per_channel_default_14"  [label="(768,)", style=solid];
+"332 linear_13_zero_point_0" -> "334 quantize_per_channel_default_14"  [label="(768,)", style=solid];
+"332 linear_13_zero_point_0" -> "335 dequantize_per_channel_default_14"  [label="(768,)", style=solid];
+"333 compressed_weight_updated_constant14" -> "334 quantize_per_channel_default_14"  [label="(768, 768)", style=solid];
+"334 quantize_per_channel_default_14" -> "335 dequantize_per_channel_default_14"  [label="(768, 768)", style=solid];
+"335 dequantize_per_channel_default_14" -> "337 linear_13"  [label="(768, 768)", style=solid];
+"336 encoder_layers_encoder_layer_3_self_attention_out_proj_bias_0_0" -> "337 linear_13"  [label="(768,)", style=solid];
+"337 linear_13" -> "338 view_31"  [label="(197, 768)", style=solid];
+"338 view_31" -> "339 transpose_23"  [label="(197, 1, 768)", style=solid];
+"339 transpose_23" -> "340 dropout_10"  [label="(1, 197, 768)", style=solid];
+"340 dropout_10" -> "341 add_7"  [label="(1, 197, 768)", style=solid];
+"341 add_7" -> "342 layer_norm_7"  [label="(1, 197, 768)", style=solid];
+"341 add_7" -> "366 add_8"  [label="(1, 197, 768)", style=solid];
+"342 layer_norm_7" -> "343 layer_norm_7_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"343 layer_norm_7_0_0_nncf_smooth_quant_0" -> "344 quantize_per_tensor_default_23"  [label="(1, 197, 768)", style=solid];
+"344 quantize_per_tensor_default_23" -> "345 dequantize_per_tensor_default_23"  [label="(1, 197, 768)", style=solid];
+"345 dequantize_per_tensor_default_23" -> "352 linear_14"  [label="(1, 197, 768)", style=solid];
+"346 linear_14_scale_0" -> "349 quantize_per_channel_default_15"  [label="(3072,)", style=solid];
+"346 linear_14_scale_0" -> "350 dequantize_per_channel_default_15"  [label="(3072,)", style=solid];
+"347 linear_14_zero_point_0" -> "349 quantize_per_channel_default_15"  [label="(3072,)", style=solid];
+"347 linear_14_zero_point_0" -> "350 dequantize_per_channel_default_15"  [label="(3072,)", style=solid];
+"348 compressed_weight_updated_constant15" -> "349 quantize_per_channel_default_15"  [label="(3072, 768)", style=solid];
+"349 quantize_per_channel_default_15" -> "350 dequantize_per_channel_default_15"  [label="(3072, 768)", style=solid];
+"350 dequantize_per_channel_default_15" -> "352 linear_14"  [label="(3072, 768)", style=solid];
+"351 encoder_layers_encoder_layer_3_mlp_0_bias_0_0" -> "352 linear_14"  [label="(3072,)", style=solid];
+"352 linear_14" -> "353 gelu_3"  [label="(1, 197, 3072)", style=solid];
+"353 gelu_3" -> "354 dropout_11"  [label="(1, 197, 3072)", style=solid];
+"354 dropout_11" -> "355 dropout_11_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"355 dropout_11_0_0_nncf_smooth_quant_0" -> "356 quantize_per_tensor_default_24"  [label="(1, 197, 3072)", style=solid];
+"356 quantize_per_tensor_default_24" -> "357 dequantize_per_tensor_default_24"  [label="(1, 197, 3072)", style=solid];
+"357 dequantize_per_tensor_default_24" -> "364 linear_15"  [label="(1, 197, 3072)", style=solid];
+"358 linear_15_scale_0" -> "361 quantize_per_channel_default_16"  [label="(768,)", style=solid];
+"358 linear_15_scale_0" -> "362 dequantize_per_channel_default_16"  [label="(768,)", style=solid];
+"359 linear_15_zero_point_0" -> "361 quantize_per_channel_default_16"  [label="(768,)", style=solid];
+"359 linear_15_zero_point_0" -> "362 dequantize_per_channel_default_16"  [label="(768,)", style=solid];
+"360 compressed_weight_updated_constant16" -> "361 quantize_per_channel_default_16"  [label="(768, 3072)", style=solid];
+"361 quantize_per_channel_default_16" -> "362 dequantize_per_channel_default_16"  [label="(768, 3072)", style=solid];
+"362 dequantize_per_channel_default_16" -> "364 linear_15"  [label="(768, 3072)", style=solid];
+"363 encoder_layers_encoder_layer_3_mlp_3_bias_0_0" -> "364 linear_15"  [label="(768,)", style=solid];
+"364 linear_15" -> "365 dropout_12"  [label="(1, 197, 768)", style=solid];
+"365 dropout_12" -> "366 add_8"  [label="(1, 197, 768)", style=solid];
+"366 add_8" -> "367 layer_norm_8"  [label="(1, 197, 768)", style=solid];
+"366 add_8" -> "416 add_9"  [label="(1, 197, 768)", style=solid];
+"367 layer_norm_8" -> "368 transpose_24"  [label="(1, 197, 768)", style=solid];
+"368 transpose_24" -> "369 transpose_24_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"369 transpose_24_0_0_nncf_smooth_quant_0" -> "370 quantize_per_tensor_default_25"  [label="(197, 1, 768)", style=solid];
+"370 quantize_per_tensor_default_25" -> "371 dequantize_per_tensor_default_25"  [label="(197, 1, 768)", style=solid];
+"371 dequantize_per_tensor_default_25" -> "378 linear_16"  [label="(197, 1, 768)", style=solid];
+"372 linear_16_scale_0" -> "375 quantize_per_channel_default_17"  [label="(2304,)", style=solid];
+"372 linear_16_scale_0" -> "376 dequantize_per_channel_default_17"  [label="(2304,)", style=solid];
+"373 linear_16_zero_point_0" -> "375 quantize_per_channel_default_17"  [label="(2304,)", style=solid];
+"373 linear_16_zero_point_0" -> "376 dequantize_per_channel_default_17"  [label="(2304,)", style=solid];
+"374 compressed_weight_updated_constant17" -> "375 quantize_per_channel_default_17"  [label="(2304, 768)", style=solid];
+"375 quantize_per_channel_default_17" -> "376 dequantize_per_channel_default_17"  [label="(2304, 768)", style=solid];
+"376 dequantize_per_channel_default_17" -> "378 linear_16"  [label="(2304, 768)", style=solid];
+"377 encoder_layers_encoder_layer_4_self_attention_in_proj_bias_0_0" -> "378 linear_16"  [label="(2304,)", style=solid];
+"378 linear_16" -> "379 unflatten_4"  [label="(197, 1, 2304)", style=solid];
+"379 unflatten_4" -> "380 unsqueeze_4"  [label="(197, 1, 3, 768)", style=solid];
+"380 unsqueeze_4" -> "381 transpose_25"  [label="(1, 197, 1, 3, 768)", style=solid];
+"381 transpose_25" -> "382 squeeze_4"  [label="(3, 197, 1, 1, 768)", style=solid];
+"382 squeeze_4" -> "383 contiguous_4"  [label="(3, 197, 1, 768)", style=solid];
+"383 contiguous_4" -> "384 quantize_per_tensor_default_26"  [label="(3, 197, 1, 768)", style=solid];
+"383 contiguous_4" -> "387 quantize_per_tensor_default_27"  [label="(3, 197, 1, 768)", style=solid];
+"383 contiguous_4" -> "390 select_14"  [label="(3, 197, 1, 768)", style=solid];
+"384 quantize_per_tensor_default_26" -> "385 dequantize_per_tensor_default_26"  [label="(3, 197, 1, 768)", style=solid];
+"385 dequantize_per_tensor_default_26" -> "386 select_12"  [label="(3, 197, 1, 768)", style=solid];
+"386 select_12" -> "391 view_32"  [label="(197, 1, 768)", style=solid];
+"387 quantize_per_tensor_default_27" -> "388 dequantize_per_tensor_default_27"  [label="(3, 197, 1, 768)", style=solid];
+"388 dequantize_per_tensor_default_27" -> "389 select_13"  [label="(3, 197, 1, 768)", style=solid];
+"389 select_13" -> "393 view_33"  [label="(197, 1, 768)", style=solid];
+"390 select_14" -> "395 view_34"  [label="(197, 1, 768)", style=solid];
+"391 view_32" -> "392 transpose_26"  [label="(197, 12, 64)", style=solid];
+"392 transpose_26" -> "397 view_35"  [label="(12, 197, 64)", style=solid];
+"393 view_33" -> "394 transpose_27"  [label="(197, 12, 64)", style=solid];
+"394 transpose_27" -> "398 view_36"  [label="(12, 197, 64)", style=solid];
+"395 view_34" -> "396 transpose_28"  [label="(197, 12, 64)", style=solid];
+"396 transpose_28" -> "399 view_37"  [label="(12, 197, 64)", style=solid];
+"397 view_35" -> "400 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
+"398 view_36" -> "400 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
+"399 view_37" -> "400 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
+"400 scaled_dot_product_attention_4" -> "401 permute_5"  [label="(1, 12, 197, 64)", style=solid];
+"401 permute_5" -> "402 view_38"  [label="(197, 1, 12, 64)", style=solid];
+"402 view_38" -> "403 view_38_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"403 view_38_0_0_nncf_smooth_quant_0" -> "404 quantize_per_tensor_default_28"  [label="(197, 768)", style=solid];
+"404 quantize_per_tensor_default_28" -> "405 dequantize_per_tensor_default_28"  [label="(197, 768)", style=solid];
+"405 dequantize_per_tensor_default_28" -> "412 linear_17"  [label="(197, 768)", style=solid];
+"406 linear_17_scale_0" -> "409 quantize_per_channel_default_18"  [label="(768,)", style=solid];
+"406 linear_17_scale_0" -> "410 dequantize_per_channel_default_18"  [label="(768,)", style=solid];
+"407 linear_17_zero_point_0" -> "409 quantize_per_channel_default_18"  [label="(768,)", style=solid];
+"407 linear_17_zero_point_0" -> "410 dequantize_per_channel_default_18"  [label="(768,)", style=solid];
+"408 compressed_weight_updated_constant18" -> "409 quantize_per_channel_default_18"  [label="(768, 768)", style=solid];
+"409 quantize_per_channel_default_18" -> "410 dequantize_per_channel_default_18"  [label="(768, 768)", style=solid];
+"410 dequantize_per_channel_default_18" -> "412 linear_17"  [label="(768, 768)", style=solid];
+"411 encoder_layers_encoder_layer_4_self_attention_out_proj_bias_0_0" -> "412 linear_17"  [label="(768,)", style=solid];
+"412 linear_17" -> "413 view_39"  [label="(197, 768)", style=solid];
+"413 view_39" -> "414 transpose_29"  [label="(197, 1, 768)", style=solid];
+"414 transpose_29" -> "415 dropout_13"  [label="(1, 197, 768)", style=solid];
+"415 dropout_13" -> "416 add_9"  [label="(1, 197, 768)", style=solid];
+"416 add_9" -> "417 layer_norm_9"  [label="(1, 197, 768)", style=solid];
+"416 add_9" -> "441 add_10"  [label="(1, 197, 768)", style=solid];
+"417 layer_norm_9" -> "418 layer_norm_9_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"418 layer_norm_9_0_0_nncf_smooth_quant_0" -> "419 quantize_per_tensor_default_29"  [label="(1, 197, 768)", style=solid];
+"419 quantize_per_tensor_default_29" -> "420 dequantize_per_tensor_default_29"  [label="(1, 197, 768)", style=solid];
+"420 dequantize_per_tensor_default_29" -> "427 linear_18"  [label="(1, 197, 768)", style=solid];
+"421 linear_18_scale_0" -> "424 quantize_per_channel_default_19"  [label="(3072,)", style=solid];
+"421 linear_18_scale_0" -> "425 dequantize_per_channel_default_19"  [label="(3072,)", style=solid];
+"422 linear_18_zero_point_0" -> "424 quantize_per_channel_default_19"  [label="(3072,)", style=solid];
+"422 linear_18_zero_point_0" -> "425 dequantize_per_channel_default_19"  [label="(3072,)", style=solid];
+"423 compressed_weight_updated_constant19" -> "424 quantize_per_channel_default_19"  [label="(3072, 768)", style=solid];
+"424 quantize_per_channel_default_19" -> "425 dequantize_per_channel_default_19"  [label="(3072, 768)", style=solid];
+"425 dequantize_per_channel_default_19" -> "427 linear_18"  [label="(3072, 768)", style=solid];
+"426 encoder_layers_encoder_layer_4_mlp_0_bias_0_0" -> "427 linear_18"  [label="(3072,)", style=solid];
+"427 linear_18" -> "428 gelu_4"  [label="(1, 197, 3072)", style=solid];
+"428 gelu_4" -> "429 dropout_14"  [label="(1, 197, 3072)", style=solid];
+"429 dropout_14" -> "430 dropout_14_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"430 dropout_14_0_0_nncf_smooth_quant_0" -> "431 quantize_per_tensor_default_30"  [label="(1, 197, 3072)", style=solid];
+"431 quantize_per_tensor_default_30" -> "432 dequantize_per_tensor_default_30"  [label="(1, 197, 3072)", style=solid];
+"432 dequantize_per_tensor_default_30" -> "439 linear_19"  [label="(1, 197, 3072)", style=solid];
+"433 linear_19_scale_0" -> "436 quantize_per_channel_default_20"  [label="(768,)", style=solid];
+"433 linear_19_scale_0" -> "437 dequantize_per_channel_default_20"  [label="(768,)", style=solid];
+"434 linear_19_zero_point_0" -> "436 quantize_per_channel_default_20"  [label="(768,)", style=solid];
+"434 linear_19_zero_point_0" -> "437 dequantize_per_channel_default_20"  [label="(768,)", style=solid];
+"435 compressed_weight_updated_constant20" -> "436 quantize_per_channel_default_20"  [label="(768, 3072)", style=solid];
+"436 quantize_per_channel_default_20" -> "437 dequantize_per_channel_default_20"  [label="(768, 3072)", style=solid];
+"437 dequantize_per_channel_default_20" -> "439 linear_19"  [label="(768, 3072)", style=solid];
+"438 encoder_layers_encoder_layer_4_mlp_3_bias_0_0" -> "439 linear_19"  [label="(768,)", style=solid];
+"439 linear_19" -> "440 dropout_15"  [label="(1, 197, 768)", style=solid];
+"440 dropout_15" -> "441 add_10"  [label="(1, 197, 768)", style=solid];
+"441 add_10" -> "442 layer_norm_10"  [label="(1, 197, 768)", style=solid];
+"441 add_10" -> "491 add_11"  [label="(1, 197, 768)", style=solid];
+"442 layer_norm_10" -> "443 transpose_30"  [label="(1, 197, 768)", style=solid];
+"443 transpose_30" -> "444 transpose_30_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"444 transpose_30_0_0_nncf_smooth_quant_0" -> "445 quantize_per_tensor_default_31"  [label="(197, 1, 768)", style=solid];
+"445 quantize_per_tensor_default_31" -> "446 dequantize_per_tensor_default_31"  [label="(197, 1, 768)", style=solid];
+"446 dequantize_per_tensor_default_31" -> "453 linear_20"  [label="(197, 1, 768)", style=solid];
+"447 linear_20_scale_0" -> "450 quantize_per_channel_default_21"  [label="(2304,)", style=solid];
+"447 linear_20_scale_0" -> "451 dequantize_per_channel_default_21"  [label="(2304,)", style=solid];
+"448 linear_20_zero_point_0" -> "450 quantize_per_channel_default_21"  [label="(2304,)", style=solid];
+"448 linear_20_zero_point_0" -> "451 dequantize_per_channel_default_21"  [label="(2304,)", style=solid];
+"449 compressed_weight_updated_constant21" -> "450 quantize_per_channel_default_21"  [label="(2304, 768)", style=solid];
+"450 quantize_per_channel_default_21" -> "451 dequantize_per_channel_default_21"  [label="(2304, 768)", style=solid];
+"451 dequantize_per_channel_default_21" -> "453 linear_20"  [label="(2304, 768)", style=solid];
+"452 encoder_layers_encoder_layer_5_self_attention_in_proj_bias_0_0" -> "453 linear_20"  [label="(2304,)", style=solid];
+"453 linear_20" -> "454 unflatten_5"  [label="(197, 1, 2304)", style=solid];
+"454 unflatten_5" -> "455 unsqueeze_5"  [label="(197, 1, 3, 768)", style=solid];
+"455 unsqueeze_5" -> "456 transpose_31"  [label="(1, 197, 1, 3, 768)", style=solid];
+"456 transpose_31" -> "457 squeeze_5"  [label="(3, 197, 1, 1, 768)", style=solid];
+"457 squeeze_5" -> "458 contiguous_5"  [label="(3, 197, 1, 768)", style=solid];
+"458 contiguous_5" -> "459 quantize_per_tensor_default_32"  [label="(3, 197, 1, 768)", style=solid];
+"458 contiguous_5" -> "462 quantize_per_tensor_default_33"  [label="(3, 197, 1, 768)", style=solid];
+"458 contiguous_5" -> "465 select_17"  [label="(3, 197, 1, 768)", style=solid];
+"459 quantize_per_tensor_default_32" -> "460 dequantize_per_tensor_default_32"  [label="(3, 197, 1, 768)", style=solid];
+"460 dequantize_per_tensor_default_32" -> "461 select_15"  [label="(3, 197, 1, 768)", style=solid];
+"461 select_15" -> "466 view_40"  [label="(197, 1, 768)", style=solid];
+"462 quantize_per_tensor_default_33" -> "463 dequantize_per_tensor_default_33"  [label="(3, 197, 1, 768)", style=solid];
+"463 dequantize_per_tensor_default_33" -> "464 select_16"  [label="(3, 197, 1, 768)", style=solid];
+"464 select_16" -> "468 view_41"  [label="(197, 1, 768)", style=solid];
+"465 select_17" -> "470 view_42"  [label="(197, 1, 768)", style=solid];
+"466 view_40" -> "467 transpose_32"  [label="(197, 12, 64)", style=solid];
+"467 transpose_32" -> "472 view_43"  [label="(12, 197, 64)", style=solid];
+"468 view_41" -> "469 transpose_33"  [label="(197, 12, 64)", style=solid];
+"469 transpose_33" -> "473 view_44"  [label="(12, 197, 64)", style=solid];
+"470 view_42" -> "471 transpose_34"  [label="(197, 12, 64)", style=solid];
+"471 transpose_34" -> "474 view_45"  [label="(12, 197, 64)", style=solid];
+"472 view_43" -> "475 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
+"473 view_44" -> "475 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
+"474 view_45" -> "475 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
+"475 scaled_dot_product_attention_5" -> "476 permute_6"  [label="(1, 12, 197, 64)", style=solid];
+"476 permute_6" -> "477 view_46"  [label="(197, 1, 12, 64)", style=solid];
+"477 view_46" -> "478 view_46_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"478 view_46_0_0_nncf_smooth_quant_0" -> "479 quantize_per_tensor_default_34"  [label="(197, 768)", style=solid];
+"479 quantize_per_tensor_default_34" -> "480 dequantize_per_tensor_default_34"  [label="(197, 768)", style=solid];
+"480 dequantize_per_tensor_default_34" -> "487 linear_21"  [label="(197, 768)", style=solid];
+"481 linear_21_scale_0" -> "484 quantize_per_channel_default_22"  [label="(768,)", style=solid];
+"481 linear_21_scale_0" -> "485 dequantize_per_channel_default_22"  [label="(768,)", style=solid];
+"482 linear_21_zero_point_0" -> "484 quantize_per_channel_default_22"  [label="(768,)", style=solid];
+"482 linear_21_zero_point_0" -> "485 dequantize_per_channel_default_22"  [label="(768,)", style=solid];
+"483 compressed_weight_updated_constant22" -> "484 quantize_per_channel_default_22"  [label="(768, 768)", style=solid];
+"484 quantize_per_channel_default_22" -> "485 dequantize_per_channel_default_22"  [label="(768, 768)", style=solid];
+"485 dequantize_per_channel_default_22" -> "487 linear_21"  [label="(768, 768)", style=solid];
+"486 encoder_layers_encoder_layer_5_self_attention_out_proj_bias_0_0" -> "487 linear_21"  [label="(768,)", style=solid];
+"487 linear_21" -> "488 view_47"  [label="(197, 768)", style=solid];
+"488 view_47" -> "489 transpose_35"  [label="(197, 1, 768)", style=solid];
+"489 transpose_35" -> "490 dropout_16"  [label="(1, 197, 768)", style=solid];
+"490 dropout_16" -> "491 add_11"  [label="(1, 197, 768)", style=solid];
+"491 add_11" -> "492 layer_norm_11"  [label="(1, 197, 768)", style=solid];
+"491 add_11" -> "516 add_12"  [label="(1, 197, 768)", style=solid];
+"492 layer_norm_11" -> "493 layer_norm_11_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"493 layer_norm_11_0_0_nncf_smooth_quant_0" -> "494 quantize_per_tensor_default_35"  [label="(1, 197, 768)", style=solid];
+"494 quantize_per_tensor_default_35" -> "495 dequantize_per_tensor_default_35"  [label="(1, 197, 768)", style=solid];
+"495 dequantize_per_tensor_default_35" -> "502 linear_22"  [label="(1, 197, 768)", style=solid];
+"496 linear_22_scale_0" -> "499 quantize_per_channel_default_23"  [label="(3072,)", style=solid];
+"496 linear_22_scale_0" -> "500 dequantize_per_channel_default_23"  [label="(3072,)", style=solid];
+"497 linear_22_zero_point_0" -> "499 quantize_per_channel_default_23"  [label="(3072,)", style=solid];
+"497 linear_22_zero_point_0" -> "500 dequantize_per_channel_default_23"  [label="(3072,)", style=solid];
+"498 compressed_weight_updated_constant23" -> "499 quantize_per_channel_default_23"  [label="(3072, 768)", style=solid];
+"499 quantize_per_channel_default_23" -> "500 dequantize_per_channel_default_23"  [label="(3072, 768)", style=solid];
+"500 dequantize_per_channel_default_23" -> "502 linear_22"  [label="(3072, 768)", style=solid];
+"501 encoder_layers_encoder_layer_5_mlp_0_bias_0_0" -> "502 linear_22"  [label="(3072,)", style=solid];
+"502 linear_22" -> "503 gelu_5"  [label="(1, 197, 3072)", style=solid];
+"503 gelu_5" -> "504 dropout_17"  [label="(1, 197, 3072)", style=solid];
+"504 dropout_17" -> "505 dropout_17_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"505 dropout_17_0_0_nncf_smooth_quant_0" -> "506 quantize_per_tensor_default_36"  [label="(1, 197, 3072)", style=solid];
+"506 quantize_per_tensor_default_36" -> "507 dequantize_per_tensor_default_36"  [label="(1, 197, 3072)", style=solid];
+"507 dequantize_per_tensor_default_36" -> "514 linear_23"  [label="(1, 197, 3072)", style=solid];
+"508 linear_23_scale_0" -> "511 quantize_per_channel_default_24"  [label="(768,)", style=solid];
+"508 linear_23_scale_0" -> "512 dequantize_per_channel_default_24"  [label="(768,)", style=solid];
+"509 linear_23_zero_point_0" -> "511 quantize_per_channel_default_24"  [label="(768,)", style=solid];
+"509 linear_23_zero_point_0" -> "512 dequantize_per_channel_default_24"  [label="(768,)", style=solid];
+"510 compressed_weight_updated_constant24" -> "511 quantize_per_channel_default_24"  [label="(768, 3072)", style=solid];
+"511 quantize_per_channel_default_24" -> "512 dequantize_per_channel_default_24"  [label="(768, 3072)", style=solid];
+"512 dequantize_per_channel_default_24" -> "514 linear_23"  [label="(768, 3072)", style=solid];
+"513 encoder_layers_encoder_layer_5_mlp_3_bias_0_0" -> "514 linear_23"  [label="(768,)", style=solid];
+"514 linear_23" -> "515 dropout_18"  [label="(1, 197, 768)", style=solid];
+"515 dropout_18" -> "516 add_12"  [label="(1, 197, 768)", style=solid];
+"516 add_12" -> "517 layer_norm_12"  [label="(1, 197, 768)", style=solid];
+"516 add_12" -> "566 add_13"  [label="(1, 197, 768)", style=solid];
+"517 layer_norm_12" -> "518 transpose_36"  [label="(1, 197, 768)", style=solid];
+"518 transpose_36" -> "519 transpose_36_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"519 transpose_36_0_0_nncf_smooth_quant_0" -> "520 quantize_per_tensor_default_37"  [label="(197, 1, 768)", style=solid];
+"520 quantize_per_tensor_default_37" -> "521 dequantize_per_tensor_default_37"  [label="(197, 1, 768)", style=solid];
+"521 dequantize_per_tensor_default_37" -> "528 linear_24"  [label="(197, 1, 768)", style=solid];
+"522 linear_24_scale_0" -> "525 quantize_per_channel_default_25"  [label="(2304,)", style=solid];
+"522 linear_24_scale_0" -> "526 dequantize_per_channel_default_25"  [label="(2304,)", style=solid];
+"523 linear_24_zero_point_0" -> "525 quantize_per_channel_default_25"  [label="(2304,)", style=solid];
+"523 linear_24_zero_point_0" -> "526 dequantize_per_channel_default_25"  [label="(2304,)", style=solid];
+"524 compressed_weight_updated_constant25" -> "525 quantize_per_channel_default_25"  [label="(2304, 768)", style=solid];
+"525 quantize_per_channel_default_25" -> "526 dequantize_per_channel_default_25"  [label="(2304, 768)", style=solid];
+"526 dequantize_per_channel_default_25" -> "528 linear_24"  [label="(2304, 768)", style=solid];
+"527 encoder_layers_encoder_layer_6_self_attention_in_proj_bias_0_0" -> "528 linear_24"  [label="(2304,)", style=solid];
+"528 linear_24" -> "529 unflatten_6"  [label="(197, 1, 2304)", style=solid];
+"529 unflatten_6" -> "530 unsqueeze_6"  [label="(197, 1, 3, 768)", style=solid];
+"530 unsqueeze_6" -> "531 transpose_37"  [label="(1, 197, 1, 3, 768)", style=solid];
+"531 transpose_37" -> "532 squeeze_6"  [label="(3, 197, 1, 1, 768)", style=solid];
+"532 squeeze_6" -> "533 contiguous_6"  [label="(3, 197, 1, 768)", style=solid];
+"533 contiguous_6" -> "534 quantize_per_tensor_default_38"  [label="(3, 197, 1, 768)", style=solid];
+"533 contiguous_6" -> "537 quantize_per_tensor_default_39"  [label="(3, 197, 1, 768)", style=solid];
+"533 contiguous_6" -> "540 select_20"  [label="(3, 197, 1, 768)", style=solid];
+"534 quantize_per_tensor_default_38" -> "535 dequantize_per_tensor_default_38"  [label="(3, 197, 1, 768)", style=solid];
+"535 dequantize_per_tensor_default_38" -> "536 select_18"  [label="(3, 197, 1, 768)", style=solid];
+"536 select_18" -> "541 view_48"  [label="(197, 1, 768)", style=solid];
+"537 quantize_per_tensor_default_39" -> "538 dequantize_per_tensor_default_39"  [label="(3, 197, 1, 768)", style=solid];
+"538 dequantize_per_tensor_default_39" -> "539 select_19"  [label="(3, 197, 1, 768)", style=solid];
+"539 select_19" -> "543 view_49"  [label="(197, 1, 768)", style=solid];
+"540 select_20" -> "545 view_50"  [label="(197, 1, 768)", style=solid];
+"541 view_48" -> "542 transpose_38"  [label="(197, 12, 64)", style=solid];
+"542 transpose_38" -> "547 view_51"  [label="(12, 197, 64)", style=solid];
+"543 view_49" -> "544 transpose_39"  [label="(197, 12, 64)", style=solid];
+"544 transpose_39" -> "548 view_52"  [label="(12, 197, 64)", style=solid];
+"545 view_50" -> "546 transpose_40"  [label="(197, 12, 64)", style=solid];
+"546 transpose_40" -> "549 view_53"  [label="(12, 197, 64)", style=solid];
+"547 view_51" -> "550 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
+"548 view_52" -> "550 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
+"549 view_53" -> "550 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
+"550 scaled_dot_product_attention_6" -> "551 permute_7"  [label="(1, 12, 197, 64)", style=solid];
+"551 permute_7" -> "552 view_54"  [label="(197, 1, 12, 64)", style=solid];
+"552 view_54" -> "553 view_54_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"553 view_54_0_0_nncf_smooth_quant_0" -> "554 quantize_per_tensor_default_40"  [label="(197, 768)", style=solid];
+"554 quantize_per_tensor_default_40" -> "555 dequantize_per_tensor_default_40"  [label="(197, 768)", style=solid];
+"555 dequantize_per_tensor_default_40" -> "562 linear_25"  [label="(197, 768)", style=solid];
+"556 linear_25_scale_0" -> "559 quantize_per_channel_default_26"  [label="(768,)", style=solid];
+"556 linear_25_scale_0" -> "560 dequantize_per_channel_default_26"  [label="(768,)", style=solid];
+"557 linear_25_zero_point_0" -> "559 quantize_per_channel_default_26"  [label="(768,)", style=solid];
+"557 linear_25_zero_point_0" -> "560 dequantize_per_channel_default_26"  [label="(768,)", style=solid];
+"558 compressed_weight_updated_constant26" -> "559 quantize_per_channel_default_26"  [label="(768, 768)", style=solid];
+"559 quantize_per_channel_default_26" -> "560 dequantize_per_channel_default_26"  [label="(768, 768)", style=solid];
+"560 dequantize_per_channel_default_26" -> "562 linear_25"  [label="(768, 768)", style=solid];
+"561 encoder_layers_encoder_layer_6_self_attention_out_proj_bias_0_0" -> "562 linear_25"  [label="(768,)", style=solid];
+"562 linear_25" -> "563 view_55"  [label="(197, 768)", style=solid];
+"563 view_55" -> "564 transpose_41"  [label="(197, 1, 768)", style=solid];
+"564 transpose_41" -> "565 dropout_19"  [label="(1, 197, 768)", style=solid];
+"565 dropout_19" -> "566 add_13"  [label="(1, 197, 768)", style=solid];
+"566 add_13" -> "567 layer_norm_13"  [label="(1, 197, 768)", style=solid];
+"566 add_13" -> "591 add_14"  [label="(1, 197, 768)", style=solid];
+"567 layer_norm_13" -> "568 layer_norm_13_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"568 layer_norm_13_0_0_nncf_smooth_quant_0" -> "569 quantize_per_tensor_default_41"  [label="(1, 197, 768)", style=solid];
+"569 quantize_per_tensor_default_41" -> "570 dequantize_per_tensor_default_41"  [label="(1, 197, 768)", style=solid];
+"570 dequantize_per_tensor_default_41" -> "577 linear_26"  [label="(1, 197, 768)", style=solid];
+"571 linear_26_scale_0" -> "574 quantize_per_channel_default_27"  [label="(3072,)", style=solid];
+"571 linear_26_scale_0" -> "575 dequantize_per_channel_default_27"  [label="(3072,)", style=solid];
+"572 linear_26_zero_point_0" -> "574 quantize_per_channel_default_27"  [label="(3072,)", style=solid];
+"572 linear_26_zero_point_0" -> "575 dequantize_per_channel_default_27"  [label="(3072,)", style=solid];
+"573 compressed_weight_updated_constant27" -> "574 quantize_per_channel_default_27"  [label="(3072, 768)", style=solid];
+"574 quantize_per_channel_default_27" -> "575 dequantize_per_channel_default_27"  [label="(3072, 768)", style=solid];
+"575 dequantize_per_channel_default_27" -> "577 linear_26"  [label="(3072, 768)", style=solid];
+"576 encoder_layers_encoder_layer_6_mlp_0_bias_0_0" -> "577 linear_26"  [label="(3072,)", style=solid];
+"577 linear_26" -> "578 gelu_6"  [label="(1, 197, 3072)", style=solid];
+"578 gelu_6" -> "579 dropout_20"  [label="(1, 197, 3072)", style=solid];
+"579 dropout_20" -> "580 dropout_20_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"580 dropout_20_0_0_nncf_smooth_quant_0" -> "581 quantize_per_tensor_default_42"  [label="(1, 197, 3072)", style=solid];
+"581 quantize_per_tensor_default_42" -> "582 dequantize_per_tensor_default_42"  [label="(1, 197, 3072)", style=solid];
+"582 dequantize_per_tensor_default_42" -> "589 linear_27"  [label="(1, 197, 3072)", style=solid];
+"583 linear_27_scale_0" -> "586 quantize_per_channel_default_28"  [label="(768,)", style=solid];
+"583 linear_27_scale_0" -> "587 dequantize_per_channel_default_28"  [label="(768,)", style=solid];
+"584 linear_27_zero_point_0" -> "586 quantize_per_channel_default_28"  [label="(768,)", style=solid];
+"584 linear_27_zero_point_0" -> "587 dequantize_per_channel_default_28"  [label="(768,)", style=solid];
+"585 compressed_weight_updated_constant28" -> "586 quantize_per_channel_default_28"  [label="(768, 3072)", style=solid];
+"586 quantize_per_channel_default_28" -> "587 dequantize_per_channel_default_28"  [label="(768, 3072)", style=solid];
+"587 dequantize_per_channel_default_28" -> "589 linear_27"  [label="(768, 3072)", style=solid];
+"588 encoder_layers_encoder_layer_6_mlp_3_bias_0_0" -> "589 linear_27"  [label="(768,)", style=solid];
+"589 linear_27" -> "590 dropout_21"  [label="(1, 197, 768)", style=solid];
+"590 dropout_21" -> "591 add_14"  [label="(1, 197, 768)", style=solid];
+"591 add_14" -> "592 layer_norm_14"  [label="(1, 197, 768)", style=solid];
+"591 add_14" -> "641 add_15"  [label="(1, 197, 768)", style=solid];
+"592 layer_norm_14" -> "593 transpose_42"  [label="(1, 197, 768)", style=solid];
+"593 transpose_42" -> "594 transpose_42_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"594 transpose_42_0_0_nncf_smooth_quant_0" -> "595 quantize_per_tensor_default_43"  [label="(197, 1, 768)", style=solid];
+"595 quantize_per_tensor_default_43" -> "596 dequantize_per_tensor_default_43"  [label="(197, 1, 768)", style=solid];
+"596 dequantize_per_tensor_default_43" -> "603 linear_28"  [label="(197, 1, 768)", style=solid];
+"597 linear_28_scale_0" -> "600 quantize_per_channel_default_29"  [label="(2304,)", style=solid];
+"597 linear_28_scale_0" -> "601 dequantize_per_channel_default_29"  [label="(2304,)", style=solid];
+"598 linear_28_zero_point_0" -> "600 quantize_per_channel_default_29"  [label="(2304,)", style=solid];
+"598 linear_28_zero_point_0" -> "601 dequantize_per_channel_default_29"  [label="(2304,)", style=solid];
+"599 compressed_weight_updated_constant29" -> "600 quantize_per_channel_default_29"  [label="(2304, 768)", style=solid];
+"600 quantize_per_channel_default_29" -> "601 dequantize_per_channel_default_29"  [label="(2304, 768)", style=solid];
+"601 dequantize_per_channel_default_29" -> "603 linear_28"  [label="(2304, 768)", style=solid];
+"602 encoder_layers_encoder_layer_7_self_attention_in_proj_bias_0_0" -> "603 linear_28"  [label="(2304,)", style=solid];
+"603 linear_28" -> "604 unflatten_7"  [label="(197, 1, 2304)", style=solid];
+"604 unflatten_7" -> "605 unsqueeze_7"  [label="(197, 1, 3, 768)", style=solid];
+"605 unsqueeze_7" -> "606 transpose_43"  [label="(1, 197, 1, 3, 768)", style=solid];
+"606 transpose_43" -> "607 squeeze_7"  [label="(3, 197, 1, 1, 768)", style=solid];
+"607 squeeze_7" -> "608 contiguous_7"  [label="(3, 197, 1, 768)", style=solid];
+"608 contiguous_7" -> "609 quantize_per_tensor_default_44"  [label="(3, 197, 1, 768)", style=solid];
+"608 contiguous_7" -> "612 quantize_per_tensor_default_45"  [label="(3, 197, 1, 768)", style=solid];
+"608 contiguous_7" -> "615 select_23"  [label="(3, 197, 1, 768)", style=solid];
+"609 quantize_per_tensor_default_44" -> "610 dequantize_per_tensor_default_44"  [label="(3, 197, 1, 768)", style=solid];
+"610 dequantize_per_tensor_default_44" -> "611 select_21"  [label="(3, 197, 1, 768)", style=solid];
+"611 select_21" -> "616 view_56"  [label="(197, 1, 768)", style=solid];
+"612 quantize_per_tensor_default_45" -> "613 dequantize_per_tensor_default_45"  [label="(3, 197, 1, 768)", style=solid];
+"613 dequantize_per_tensor_default_45" -> "614 select_22"  [label="(3, 197, 1, 768)", style=solid];
+"614 select_22" -> "618 view_57"  [label="(197, 1, 768)", style=solid];
+"615 select_23" -> "620 view_58"  [label="(197, 1, 768)", style=solid];
+"616 view_56" -> "617 transpose_44"  [label="(197, 12, 64)", style=solid];
+"617 transpose_44" -> "622 view_59"  [label="(12, 197, 64)", style=solid];
+"618 view_57" -> "619 transpose_45"  [label="(197, 12, 64)", style=solid];
+"619 transpose_45" -> "623 view_60"  [label="(12, 197, 64)", style=solid];
+"620 view_58" -> "621 transpose_46"  [label="(197, 12, 64)", style=solid];
+"621 transpose_46" -> "624 view_61"  [label="(12, 197, 64)", style=solid];
+"622 view_59" -> "625 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
+"623 view_60" -> "625 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
+"624 view_61" -> "625 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
+"625 scaled_dot_product_attention_7" -> "626 permute_8"  [label="(1, 12, 197, 64)", style=solid];
+"626 permute_8" -> "627 view_62"  [label="(197, 1, 12, 64)", style=solid];
+"627 view_62" -> "628 view_62_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"628 view_62_0_0_nncf_smooth_quant_0" -> "629 quantize_per_tensor_default_46"  [label="(197, 768)", style=solid];
+"629 quantize_per_tensor_default_46" -> "630 dequantize_per_tensor_default_46"  [label="(197, 768)", style=solid];
+"630 dequantize_per_tensor_default_46" -> "637 linear_29"  [label="(197, 768)", style=solid];
+"631 linear_29_scale_0" -> "634 quantize_per_channel_default_30"  [label="(768,)", style=solid];
+"631 linear_29_scale_0" -> "635 dequantize_per_channel_default_30"  [label="(768,)", style=solid];
+"632 linear_29_zero_point_0" -> "634 quantize_per_channel_default_30"  [label="(768,)", style=solid];
+"632 linear_29_zero_point_0" -> "635 dequantize_per_channel_default_30"  [label="(768,)", style=solid];
+"633 compressed_weight_updated_constant30" -> "634 quantize_per_channel_default_30"  [label="(768, 768)", style=solid];
+"634 quantize_per_channel_default_30" -> "635 dequantize_per_channel_default_30"  [label="(768, 768)", style=solid];
+"635 dequantize_per_channel_default_30" -> "637 linear_29"  [label="(768, 768)", style=solid];
+"636 encoder_layers_encoder_layer_7_self_attention_out_proj_bias_0_0" -> "637 linear_29"  [label="(768,)", style=solid];
+"637 linear_29" -> "638 view_63"  [label="(197, 768)", style=solid];
+"638 view_63" -> "639 transpose_47"  [label="(197, 1, 768)", style=solid];
+"639 transpose_47" -> "640 dropout_22"  [label="(1, 197, 768)", style=solid];
+"640 dropout_22" -> "641 add_15"  [label="(1, 197, 768)", style=solid];
+"641 add_15" -> "642 layer_norm_15"  [label="(1, 197, 768)", style=solid];
+"641 add_15" -> "666 add_16"  [label="(1, 197, 768)", style=solid];
+"642 layer_norm_15" -> "643 layer_norm_15_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"643 layer_norm_15_0_0_nncf_smooth_quant_0" -> "644 quantize_per_tensor_default_47"  [label="(1, 197, 768)", style=solid];
+"644 quantize_per_tensor_default_47" -> "645 dequantize_per_tensor_default_47"  [label="(1, 197, 768)", style=solid];
+"645 dequantize_per_tensor_default_47" -> "652 linear_30"  [label="(1, 197, 768)", style=solid];
+"646 linear_30_scale_0" -> "649 quantize_per_channel_default_31"  [label="(3072,)", style=solid];
+"646 linear_30_scale_0" -> "650 dequantize_per_channel_default_31"  [label="(3072,)", style=solid];
+"647 linear_30_zero_point_0" -> "649 quantize_per_channel_default_31"  [label="(3072,)", style=solid];
+"647 linear_30_zero_point_0" -> "650 dequantize_per_channel_default_31"  [label="(3072,)", style=solid];
+"648 compressed_weight_updated_constant31" -> "649 quantize_per_channel_default_31"  [label="(3072, 768)", style=solid];
+"649 quantize_per_channel_default_31" -> "650 dequantize_per_channel_default_31"  [label="(3072, 768)", style=solid];
+"650 dequantize_per_channel_default_31" -> "652 linear_30"  [label="(3072, 768)", style=solid];
+"651 encoder_layers_encoder_layer_7_mlp_0_bias_0_0" -> "652 linear_30"  [label="(3072,)", style=solid];
+"652 linear_30" -> "653 gelu_7"  [label="(1, 197, 3072)", style=solid];
+"653 gelu_7" -> "654 dropout_23"  [label="(1, 197, 3072)", style=solid];
+"654 dropout_23" -> "655 dropout_23_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"655 dropout_23_0_0_nncf_smooth_quant_0" -> "656 quantize_per_tensor_default_48"  [label="(1, 197, 3072)", style=solid];
+"656 quantize_per_tensor_default_48" -> "657 dequantize_per_tensor_default_48"  [label="(1, 197, 3072)", style=solid];
+"657 dequantize_per_tensor_default_48" -> "664 linear_31"  [label="(1, 197, 3072)", style=solid];
+"658 linear_31_scale_0" -> "661 quantize_per_channel_default_32"  [label="(768,)", style=solid];
+"658 linear_31_scale_0" -> "662 dequantize_per_channel_default_32"  [label="(768,)", style=solid];
+"659 linear_31_zero_point_0" -> "661 quantize_per_channel_default_32"  [label="(768,)", style=solid];
+"659 linear_31_zero_point_0" -> "662 dequantize_per_channel_default_32"  [label="(768,)", style=solid];
+"660 compressed_weight_updated_constant32" -> "661 quantize_per_channel_default_32"  [label="(768, 3072)", style=solid];
+"661 quantize_per_channel_default_32" -> "662 dequantize_per_channel_default_32"  [label="(768, 3072)", style=solid];
+"662 dequantize_per_channel_default_32" -> "664 linear_31"  [label="(768, 3072)", style=solid];
+"663 encoder_layers_encoder_layer_7_mlp_3_bias_0_0" -> "664 linear_31"  [label="(768,)", style=solid];
+"664 linear_31" -> "665 dropout_24"  [label="(1, 197, 768)", style=solid];
+"665 dropout_24" -> "666 add_16"  [label="(1, 197, 768)", style=solid];
+"666 add_16" -> "667 layer_norm_16"  [label="(1, 197, 768)", style=solid];
+"666 add_16" -> "716 add_17"  [label="(1, 197, 768)", style=solid];
+"667 layer_norm_16" -> "668 transpose_48"  [label="(1, 197, 768)", style=solid];
+"668 transpose_48" -> "669 transpose_48_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"669 transpose_48_0_0_nncf_smooth_quant_0" -> "670 quantize_per_tensor_default_49"  [label="(197, 1, 768)", style=solid];
+"670 quantize_per_tensor_default_49" -> "671 dequantize_per_tensor_default_49"  [label="(197, 1, 768)", style=solid];
+"671 dequantize_per_tensor_default_49" -> "678 linear_32"  [label="(197, 1, 768)", style=solid];
+"672 linear_32_scale_0" -> "675 quantize_per_channel_default_33"  [label="(2304,)", style=solid];
+"672 linear_32_scale_0" -> "676 dequantize_per_channel_default_33"  [label="(2304,)", style=solid];
+"673 linear_32_zero_point_0" -> "675 quantize_per_channel_default_33"  [label="(2304,)", style=solid];
+"673 linear_32_zero_point_0" -> "676 dequantize_per_channel_default_33"  [label="(2304,)", style=solid];
+"674 compressed_weight_updated_constant33" -> "675 quantize_per_channel_default_33"  [label="(2304, 768)", style=solid];
+"675 quantize_per_channel_default_33" -> "676 dequantize_per_channel_default_33"  [label="(2304, 768)", style=solid];
+"676 dequantize_per_channel_default_33" -> "678 linear_32"  [label="(2304, 768)", style=solid];
+"677 encoder_layers_encoder_layer_8_self_attention_in_proj_bias_0_0" -> "678 linear_32"  [label="(2304,)", style=solid];
+"678 linear_32" -> "679 unflatten_8"  [label="(197, 1, 2304)", style=solid];
+"679 unflatten_8" -> "680 unsqueeze_8"  [label="(197, 1, 3, 768)", style=solid];
+"680 unsqueeze_8" -> "681 transpose_49"  [label="(1, 197, 1, 3, 768)", style=solid];
+"681 transpose_49" -> "682 squeeze_8"  [label="(3, 197, 1, 1, 768)", style=solid];
+"682 squeeze_8" -> "683 contiguous_8"  [label="(3, 197, 1, 768)", style=solid];
+"683 contiguous_8" -> "684 quantize_per_tensor_default_50"  [label="(3, 197, 1, 768)", style=solid];
+"683 contiguous_8" -> "687 quantize_per_tensor_default_51"  [label="(3, 197, 1, 768)", style=solid];
+"683 contiguous_8" -> "690 select_26"  [label="(3, 197, 1, 768)", style=solid];
+"684 quantize_per_tensor_default_50" -> "685 dequantize_per_tensor_default_50"  [label="(3, 197, 1, 768)", style=solid];
+"685 dequantize_per_tensor_default_50" -> "686 select_24"  [label="(3, 197, 1, 768)", style=solid];
+"686 select_24" -> "691 view_64"  [label="(197, 1, 768)", style=solid];
+"687 quantize_per_tensor_default_51" -> "688 dequantize_per_tensor_default_51"  [label="(3, 197, 1, 768)", style=solid];
+"688 dequantize_per_tensor_default_51" -> "689 select_25"  [label="(3, 197, 1, 768)", style=solid];
+"689 select_25" -> "693 view_65"  [label="(197, 1, 768)", style=solid];
+"690 select_26" -> "695 view_66"  [label="(197, 1, 768)", style=solid];
+"691 view_64" -> "692 transpose_50"  [label="(197, 12, 64)", style=solid];
+"692 transpose_50" -> "697 view_67"  [label="(12, 197, 64)", style=solid];
+"693 view_65" -> "694 transpose_51"  [label="(197, 12, 64)", style=solid];
+"694 transpose_51" -> "698 view_68"  [label="(12, 197, 64)", style=solid];
+"695 view_66" -> "696 transpose_52"  [label="(197, 12, 64)", style=solid];
+"696 transpose_52" -> "699 view_69"  [label="(12, 197, 64)", style=solid];
+"697 view_67" -> "700 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
+"698 view_68" -> "700 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
+"699 view_69" -> "700 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
+"700 scaled_dot_product_attention_8" -> "701 permute_9"  [label="(1, 12, 197, 64)", style=solid];
+"701 permute_9" -> "702 view_70"  [label="(197, 1, 12, 64)", style=solid];
+"702 view_70" -> "703 view_70_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"703 view_70_0_0_nncf_smooth_quant_0" -> "704 quantize_per_tensor_default_52"  [label="(197, 768)", style=solid];
+"704 quantize_per_tensor_default_52" -> "705 dequantize_per_tensor_default_52"  [label="(197, 768)", style=solid];
+"705 dequantize_per_tensor_default_52" -> "712 linear_33"  [label="(197, 768)", style=solid];
+"706 linear_33_scale_0" -> "709 quantize_per_channel_default_34"  [label="(768,)", style=solid];
+"706 linear_33_scale_0" -> "710 dequantize_per_channel_default_34"  [label="(768,)", style=solid];
+"707 linear_33_zero_point_0" -> "709 quantize_per_channel_default_34"  [label="(768,)", style=solid];
+"707 linear_33_zero_point_0" -> "710 dequantize_per_channel_default_34"  [label="(768,)", style=solid];
+"708 compressed_weight_updated_constant34" -> "709 quantize_per_channel_default_34"  [label="(768, 768)", style=solid];
+"709 quantize_per_channel_default_34" -> "710 dequantize_per_channel_default_34"  [label="(768, 768)", style=solid];
+"710 dequantize_per_channel_default_34" -> "712 linear_33"  [label="(768, 768)", style=solid];
+"711 encoder_layers_encoder_layer_8_self_attention_out_proj_bias_0_0" -> "712 linear_33"  [label="(768,)", style=solid];
+"712 linear_33" -> "713 view_71"  [label="(197, 768)", style=solid];
+"713 view_71" -> "714 transpose_53"  [label="(197, 1, 768)", style=solid];
+"714 transpose_53" -> "715 dropout_25"  [label="(1, 197, 768)", style=solid];
+"715 dropout_25" -> "716 add_17"  [label="(1, 197, 768)", style=solid];
+"716 add_17" -> "717 layer_norm_17"  [label="(1, 197, 768)", style=solid];
+"716 add_17" -> "741 add_18"  [label="(1, 197, 768)", style=solid];
+"717 layer_norm_17" -> "718 layer_norm_17_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"718 layer_norm_17_0_0_nncf_smooth_quant_0" -> "719 quantize_per_tensor_default_53"  [label="(1, 197, 768)", style=solid];
+"719 quantize_per_tensor_default_53" -> "720 dequantize_per_tensor_default_53"  [label="(1, 197, 768)", style=solid];
+"720 dequantize_per_tensor_default_53" -> "727 linear_34"  [label="(1, 197, 768)", style=solid];
+"721 linear_34_scale_0" -> "724 quantize_per_channel_default_35"  [label="(3072,)", style=solid];
+"721 linear_34_scale_0" -> "725 dequantize_per_channel_default_35"  [label="(3072,)", style=solid];
+"722 linear_34_zero_point_0" -> "724 quantize_per_channel_default_35"  [label="(3072,)", style=solid];
+"722 linear_34_zero_point_0" -> "725 dequantize_per_channel_default_35"  [label="(3072,)", style=solid];
+"723 compressed_weight_updated_constant35" -> "724 quantize_per_channel_default_35"  [label="(3072, 768)", style=solid];
+"724 quantize_per_channel_default_35" -> "725 dequantize_per_channel_default_35"  [label="(3072, 768)", style=solid];
+"725 dequantize_per_channel_default_35" -> "727 linear_34"  [label="(3072, 768)", style=solid];
+"726 encoder_layers_encoder_layer_8_mlp_0_bias_0_0" -> "727 linear_34"  [label="(3072,)", style=solid];
+"727 linear_34" -> "728 gelu_8"  [label="(1, 197, 3072)", style=solid];
+"728 gelu_8" -> "729 dropout_26"  [label="(1, 197, 3072)", style=solid];
+"729 dropout_26" -> "730 dropout_26_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"730 dropout_26_0_0_nncf_smooth_quant_0" -> "731 quantize_per_tensor_default_54"  [label="(1, 197, 3072)", style=solid];
+"731 quantize_per_tensor_default_54" -> "732 dequantize_per_tensor_default_54"  [label="(1, 197, 3072)", style=solid];
+"732 dequantize_per_tensor_default_54" -> "739 linear_35"  [label="(1, 197, 3072)", style=solid];
+"733 linear_35_scale_0" -> "736 quantize_per_channel_default_36"  [label="(768,)", style=solid];
+"733 linear_35_scale_0" -> "737 dequantize_per_channel_default_36"  [label="(768,)", style=solid];
+"734 linear_35_zero_point_0" -> "736 quantize_per_channel_default_36"  [label="(768,)", style=solid];
+"734 linear_35_zero_point_0" -> "737 dequantize_per_channel_default_36"  [label="(768,)", style=solid];
+"735 compressed_weight_updated_constant36" -> "736 quantize_per_channel_default_36"  [label="(768, 3072)", style=solid];
+"736 quantize_per_channel_default_36" -> "737 dequantize_per_channel_default_36"  [label="(768, 3072)", style=solid];
+"737 dequantize_per_channel_default_36" -> "739 linear_35"  [label="(768, 3072)", style=solid];
+"738 encoder_layers_encoder_layer_8_mlp_3_bias_0_0" -> "739 linear_35"  [label="(768,)", style=solid];
+"739 linear_35" -> "740 dropout_27"  [label="(1, 197, 768)", style=solid];
+"740 dropout_27" -> "741 add_18"  [label="(1, 197, 768)", style=solid];
+"741 add_18" -> "742 layer_norm_18"  [label="(1, 197, 768)", style=solid];
+"741 add_18" -> "791 add_19"  [label="(1, 197, 768)", style=solid];
+"742 layer_norm_18" -> "743 transpose_54"  [label="(1, 197, 768)", style=solid];
+"743 transpose_54" -> "744 transpose_54_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"744 transpose_54_0_0_nncf_smooth_quant_0" -> "745 quantize_per_tensor_default_55"  [label="(197, 1, 768)", style=solid];
+"745 quantize_per_tensor_default_55" -> "746 dequantize_per_tensor_default_55"  [label="(197, 1, 768)", style=solid];
+"746 dequantize_per_tensor_default_55" -> "753 linear_36"  [label="(197, 1, 768)", style=solid];
+"747 linear_36_scale_0" -> "750 quantize_per_channel_default_37"  [label="(2304,)", style=solid];
+"747 linear_36_scale_0" -> "751 dequantize_per_channel_default_37"  [label="(2304,)", style=solid];
+"748 linear_36_zero_point_0" -> "750 quantize_per_channel_default_37"  [label="(2304,)", style=solid];
+"748 linear_36_zero_point_0" -> "751 dequantize_per_channel_default_37"  [label="(2304,)", style=solid];
+"749 compressed_weight_updated_constant37" -> "750 quantize_per_channel_default_37"  [label="(2304, 768)", style=solid];
+"750 quantize_per_channel_default_37" -> "751 dequantize_per_channel_default_37"  [label="(2304, 768)", style=solid];
+"751 dequantize_per_channel_default_37" -> "753 linear_36"  [label="(2304, 768)", style=solid];
+"752 encoder_layers_encoder_layer_9_self_attention_in_proj_bias_0_0" -> "753 linear_36"  [label="(2304,)", style=solid];
+"753 linear_36" -> "754 unflatten_9"  [label="(197, 1, 2304)", style=solid];
+"754 unflatten_9" -> "755 unsqueeze_9"  [label="(197, 1, 3, 768)", style=solid];
+"755 unsqueeze_9" -> "756 transpose_55"  [label="(1, 197, 1, 3, 768)", style=solid];
+"756 transpose_55" -> "757 squeeze_9"  [label="(3, 197, 1, 1, 768)", style=solid];
+"757 squeeze_9" -> "758 contiguous_9"  [label="(3, 197, 1, 768)", style=solid];
+"758 contiguous_9" -> "759 quantize_per_tensor_default_56"  [label="(3, 197, 1, 768)", style=solid];
+"758 contiguous_9" -> "762 quantize_per_tensor_default_57"  [label="(3, 197, 1, 768)", style=solid];
+"758 contiguous_9" -> "765 select_29"  [label="(3, 197, 1, 768)", style=solid];
+"759 quantize_per_tensor_default_56" -> "760 dequantize_per_tensor_default_56"  [label="(3, 197, 1, 768)", style=solid];
+"760 dequantize_per_tensor_default_56" -> "761 select_27"  [label="(3, 197, 1, 768)", style=solid];
+"761 select_27" -> "766 view_72"  [label="(197, 1, 768)", style=solid];
+"762 quantize_per_tensor_default_57" -> "763 dequantize_per_tensor_default_57"  [label="(3, 197, 1, 768)", style=solid];
+"763 dequantize_per_tensor_default_57" -> "764 select_28"  [label="(3, 197, 1, 768)", style=solid];
+"764 select_28" -> "768 view_73"  [label="(197, 1, 768)", style=solid];
+"765 select_29" -> "770 view_74"  [label="(197, 1, 768)", style=solid];
+"766 view_72" -> "767 transpose_56"  [label="(197, 12, 64)", style=solid];
+"767 transpose_56" -> "772 view_75"  [label="(12, 197, 64)", style=solid];
+"768 view_73" -> "769 transpose_57"  [label="(197, 12, 64)", style=solid];
+"769 transpose_57" -> "773 view_76"  [label="(12, 197, 64)", style=solid];
+"770 view_74" -> "771 transpose_58"  [label="(197, 12, 64)", style=solid];
+"771 transpose_58" -> "774 view_77"  [label="(12, 197, 64)", style=solid];
+"772 view_75" -> "775 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
+"773 view_76" -> "775 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
+"774 view_77" -> "775 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
+"775 scaled_dot_product_attention_9" -> "776 permute_10"  [label="(1, 12, 197, 64)", style=solid];
+"776 permute_10" -> "777 view_78"  [label="(197, 1, 12, 64)", style=solid];
+"777 view_78" -> "778 view_78_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"778 view_78_0_0_nncf_smooth_quant_0" -> "779 quantize_per_tensor_default_58"  [label="(197, 768)", style=solid];
+"779 quantize_per_tensor_default_58" -> "780 dequantize_per_tensor_default_58"  [label="(197, 768)", style=solid];
+"780 dequantize_per_tensor_default_58" -> "787 linear_37"  [label="(197, 768)", style=solid];
+"781 linear_37_scale_0" -> "784 quantize_per_channel_default_38"  [label="(768,)", style=solid];
+"781 linear_37_scale_0" -> "785 dequantize_per_channel_default_38"  [label="(768,)", style=solid];
+"782 linear_37_zero_point_0" -> "784 quantize_per_channel_default_38"  [label="(768,)", style=solid];
+"782 linear_37_zero_point_0" -> "785 dequantize_per_channel_default_38"  [label="(768,)", style=solid];
+"783 compressed_weight_updated_constant38" -> "784 quantize_per_channel_default_38"  [label="(768, 768)", style=solid];
+"784 quantize_per_channel_default_38" -> "785 dequantize_per_channel_default_38"  [label="(768, 768)", style=solid];
+"785 dequantize_per_channel_default_38" -> "787 linear_37"  [label="(768, 768)", style=solid];
+"786 encoder_layers_encoder_layer_9_self_attention_out_proj_bias_0_0" -> "787 linear_37"  [label="(768,)", style=solid];
+"787 linear_37" -> "788 view_79"  [label="(197, 768)", style=solid];
+"788 view_79" -> "789 transpose_59"  [label="(197, 1, 768)", style=solid];
+"789 transpose_59" -> "790 dropout_28"  [label="(1, 197, 768)", style=solid];
+"790 dropout_28" -> "791 add_19"  [label="(1, 197, 768)", style=solid];
+"791 add_19" -> "792 layer_norm_19"  [label="(1, 197, 768)", style=solid];
+"791 add_19" -> "816 add_20"  [label="(1, 197, 768)", style=solid];
+"792 layer_norm_19" -> "793 layer_norm_19_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"793 layer_norm_19_0_0_nncf_smooth_quant_0" -> "794 quantize_per_tensor_default_59"  [label="(1, 197, 768)", style=solid];
+"794 quantize_per_tensor_default_59" -> "795 dequantize_per_tensor_default_59"  [label="(1, 197, 768)", style=solid];
+"795 dequantize_per_tensor_default_59" -> "802 linear_38"  [label="(1, 197, 768)", style=solid];
+"796 linear_38_scale_0" -> "799 quantize_per_channel_default_39"  [label="(3072,)", style=solid];
+"796 linear_38_scale_0" -> "800 dequantize_per_channel_default_39"  [label="(3072,)", style=solid];
+"797 linear_38_zero_point_0" -> "799 quantize_per_channel_default_39"  [label="(3072,)", style=solid];
+"797 linear_38_zero_point_0" -> "800 dequantize_per_channel_default_39"  [label="(3072,)", style=solid];
+"798 compressed_weight_updated_constant39" -> "799 quantize_per_channel_default_39"  [label="(3072, 768)", style=solid];
+"799 quantize_per_channel_default_39" -> "800 dequantize_per_channel_default_39"  [label="(3072, 768)", style=solid];
+"800 dequantize_per_channel_default_39" -> "802 linear_38"  [label="(3072, 768)", style=solid];
+"801 encoder_layers_encoder_layer_9_mlp_0_bias_0_0" -> "802 linear_38"  [label="(3072,)", style=solid];
+"802 linear_38" -> "803 gelu_9"  [label="(1, 197, 3072)", style=solid];
+"803 gelu_9" -> "804 dropout_29"  [label="(1, 197, 3072)", style=solid];
+"804 dropout_29" -> "805 dropout_29_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"805 dropout_29_0_0_nncf_smooth_quant_0" -> "806 quantize_per_tensor_default_60"  [label="(1, 197, 3072)", style=solid];
+"806 quantize_per_tensor_default_60" -> "807 dequantize_per_tensor_default_60"  [label="(1, 197, 3072)", style=solid];
+"807 dequantize_per_tensor_default_60" -> "814 linear_39"  [label="(1, 197, 3072)", style=solid];
+"808 linear_39_scale_0" -> "811 quantize_per_channel_default_40"  [label="(768,)", style=solid];
+"808 linear_39_scale_0" -> "812 dequantize_per_channel_default_40"  [label="(768,)", style=solid];
+"809 linear_39_zero_point_0" -> "811 quantize_per_channel_default_40"  [label="(768,)", style=solid];
+"809 linear_39_zero_point_0" -> "812 dequantize_per_channel_default_40"  [label="(768,)", style=solid];
+"810 compressed_weight_updated_constant40" -> "811 quantize_per_channel_default_40"  [label="(768, 3072)", style=solid];
+"811 quantize_per_channel_default_40" -> "812 dequantize_per_channel_default_40"  [label="(768, 3072)", style=solid];
+"812 dequantize_per_channel_default_40" -> "814 linear_39"  [label="(768, 3072)", style=solid];
+"813 encoder_layers_encoder_layer_9_mlp_3_bias_0_0" -> "814 linear_39"  [label="(768,)", style=solid];
+"814 linear_39" -> "815 dropout_30"  [label="(1, 197, 768)", style=solid];
+"815 dropout_30" -> "816 add_20"  [label="(1, 197, 768)", style=solid];
+"816 add_20" -> "817 layer_norm_20"  [label="(1, 197, 768)", style=solid];
+"816 add_20" -> "866 add_21"  [label="(1, 197, 768)", style=solid];
+"817 layer_norm_20" -> "818 transpose_60"  [label="(1, 197, 768)", style=solid];
+"818 transpose_60" -> "819 transpose_60_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"819 transpose_60_0_0_nncf_smooth_quant_0" -> "820 quantize_per_tensor_default_61"  [label="(197, 1, 768)", style=solid];
+"820 quantize_per_tensor_default_61" -> "821 dequantize_per_tensor_default_61"  [label="(197, 1, 768)", style=solid];
+"821 dequantize_per_tensor_default_61" -> "828 linear_40"  [label="(197, 1, 768)", style=solid];
+"822 linear_40_scale_0" -> "825 quantize_per_channel_default_41"  [label="(2304,)", style=solid];
+"822 linear_40_scale_0" -> "826 dequantize_per_channel_default_41"  [label="(2304,)", style=solid];
+"823 linear_40_zero_point_0" -> "825 quantize_per_channel_default_41"  [label="(2304,)", style=solid];
+"823 linear_40_zero_point_0" -> "826 dequantize_per_channel_default_41"  [label="(2304,)", style=solid];
+"824 compressed_weight_updated_constant41" -> "825 quantize_per_channel_default_41"  [label="(2304, 768)", style=solid];
+"825 quantize_per_channel_default_41" -> "826 dequantize_per_channel_default_41"  [label="(2304, 768)", style=solid];
+"826 dequantize_per_channel_default_41" -> "828 linear_40"  [label="(2304, 768)", style=solid];
+"827 encoder_layers_encoder_layer_10_self_attention_in_proj_bias_0_0" -> "828 linear_40"  [label="(2304,)", style=solid];
+"828 linear_40" -> "829 unflatten_10"  [label="(197, 1, 2304)", style=solid];
+"829 unflatten_10" -> "830 unsqueeze_10"  [label="(197, 1, 3, 768)", style=solid];
+"830 unsqueeze_10" -> "831 transpose_61"  [label="(1, 197, 1, 3, 768)", style=solid];
+"831 transpose_61" -> "832 squeeze_10"  [label="(3, 197, 1, 1, 768)", style=solid];
+"832 squeeze_10" -> "833 contiguous_10"  [label="(3, 197, 1, 768)", style=solid];
+"833 contiguous_10" -> "834 quantize_per_tensor_default_62"  [label="(3, 197, 1, 768)", style=solid];
+"833 contiguous_10" -> "837 quantize_per_tensor_default_63"  [label="(3, 197, 1, 768)", style=solid];
+"833 contiguous_10" -> "840 select_32"  [label="(3, 197, 1, 768)", style=solid];
+"834 quantize_per_tensor_default_62" -> "835 dequantize_per_tensor_default_62"  [label="(3, 197, 1, 768)", style=solid];
+"835 dequantize_per_tensor_default_62" -> "836 select_30"  [label="(3, 197, 1, 768)", style=solid];
+"836 select_30" -> "841 view_80"  [label="(197, 1, 768)", style=solid];
+"837 quantize_per_tensor_default_63" -> "838 dequantize_per_tensor_default_63"  [label="(3, 197, 1, 768)", style=solid];
+"838 dequantize_per_tensor_default_63" -> "839 select_31"  [label="(3, 197, 1, 768)", style=solid];
+"839 select_31" -> "843 view_81"  [label="(197, 1, 768)", style=solid];
+"840 select_32" -> "845 view_82"  [label="(197, 1, 768)", style=solid];
+"841 view_80" -> "842 transpose_62"  [label="(197, 12, 64)", style=solid];
+"842 transpose_62" -> "847 view_83"  [label="(12, 197, 64)", style=solid];
+"843 view_81" -> "844 transpose_63"  [label="(197, 12, 64)", style=solid];
+"844 transpose_63" -> "848 view_84"  [label="(12, 197, 64)", style=solid];
+"845 view_82" -> "846 transpose_64"  [label="(197, 12, 64)", style=solid];
+"846 transpose_64" -> "849 view_85"  [label="(12, 197, 64)", style=solid];
+"847 view_83" -> "850 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
+"848 view_84" -> "850 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
+"849 view_85" -> "850 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
+"850 scaled_dot_product_attention_10" -> "851 permute_11"  [label="(1, 12, 197, 64)", style=solid];
+"851 permute_11" -> "852 view_86"  [label="(197, 1, 12, 64)", style=solid];
+"852 view_86" -> "853 view_86_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"853 view_86_0_0_nncf_smooth_quant_0" -> "854 quantize_per_tensor_default_64"  [label="(197, 768)", style=solid];
+"854 quantize_per_tensor_default_64" -> "855 dequantize_per_tensor_default_64"  [label="(197, 768)", style=solid];
+"855 dequantize_per_tensor_default_64" -> "862 linear_41"  [label="(197, 768)", style=solid];
+"856 linear_41_scale_0" -> "859 quantize_per_channel_default_42"  [label="(768,)", style=solid];
+"856 linear_41_scale_0" -> "860 dequantize_per_channel_default_42"  [label="(768,)", style=solid];
+"857 linear_41_zero_point_0" -> "859 quantize_per_channel_default_42"  [label="(768,)", style=solid];
+"857 linear_41_zero_point_0" -> "860 dequantize_per_channel_default_42"  [label="(768,)", style=solid];
+"858 compressed_weight_updated_constant42" -> "859 quantize_per_channel_default_42"  [label="(768, 768)", style=solid];
+"859 quantize_per_channel_default_42" -> "860 dequantize_per_channel_default_42"  [label="(768, 768)", style=solid];
+"860 dequantize_per_channel_default_42" -> "862 linear_41"  [label="(768, 768)", style=solid];
+"861 encoder_layers_encoder_layer_10_self_attention_out_proj_bias_0_0" -> "862 linear_41"  [label="(768,)", style=solid];
+"862 linear_41" -> "863 view_87"  [label="(197, 768)", style=solid];
+"863 view_87" -> "864 transpose_65"  [label="(197, 1, 768)", style=solid];
+"864 transpose_65" -> "865 dropout_31"  [label="(1, 197, 768)", style=solid];
+"865 dropout_31" -> "866 add_21"  [label="(1, 197, 768)", style=solid];
+"866 add_21" -> "867 layer_norm_21"  [label="(1, 197, 768)", style=solid];
+"866 add_21" -> "891 add_22"  [label="(1, 197, 768)", style=solid];
+"867 layer_norm_21" -> "868 layer_norm_21_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"868 layer_norm_21_0_0_nncf_smooth_quant_0" -> "869 quantize_per_tensor_default_65"  [label="(1, 197, 768)", style=solid];
+"869 quantize_per_tensor_default_65" -> "870 dequantize_per_tensor_default_65"  [label="(1, 197, 768)", style=solid];
+"870 dequantize_per_tensor_default_65" -> "877 linear_42"  [label="(1, 197, 768)", style=solid];
+"871 linear_42_scale_0" -> "874 quantize_per_channel_default_43"  [label="(3072,)", style=solid];
+"871 linear_42_scale_0" -> "875 dequantize_per_channel_default_43"  [label="(3072,)", style=solid];
+"872 linear_42_zero_point_0" -> "874 quantize_per_channel_default_43"  [label="(3072,)", style=solid];
+"872 linear_42_zero_point_0" -> "875 dequantize_per_channel_default_43"  [label="(3072,)", style=solid];
+"873 compressed_weight_updated_constant43" -> "874 quantize_per_channel_default_43"  [label="(3072, 768)", style=solid];
+"874 quantize_per_channel_default_43" -> "875 dequantize_per_channel_default_43"  [label="(3072, 768)", style=solid];
+"875 dequantize_per_channel_default_43" -> "877 linear_42"  [label="(3072, 768)", style=solid];
+"876 encoder_layers_encoder_layer_10_mlp_0_bias_0_0" -> "877 linear_42"  [label="(3072,)", style=solid];
+"877 linear_42" -> "878 gelu_10"  [label="(1, 197, 3072)", style=solid];
+"878 gelu_10" -> "879 dropout_32"  [label="(1, 197, 3072)", style=solid];
+"879 dropout_32" -> "880 dropout_32_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"880 dropout_32_0_0_nncf_smooth_quant_0" -> "881 quantize_per_tensor_default_66"  [label="(1, 197, 3072)", style=solid];
+"881 quantize_per_tensor_default_66" -> "882 dequantize_per_tensor_default_66"  [label="(1, 197, 3072)", style=solid];
+"882 dequantize_per_tensor_default_66" -> "889 linear_43"  [label="(1, 197, 3072)", style=solid];
+"883 linear_43_scale_0" -> "886 quantize_per_channel_default_44"  [label="(768,)", style=solid];
+"883 linear_43_scale_0" -> "887 dequantize_per_channel_default_44"  [label="(768,)", style=solid];
+"884 linear_43_zero_point_0" -> "886 quantize_per_channel_default_44"  [label="(768,)", style=solid];
+"884 linear_43_zero_point_0" -> "887 dequantize_per_channel_default_44"  [label="(768,)", style=solid];
+"885 compressed_weight_updated_constant44" -> "886 quantize_per_channel_default_44"  [label="(768, 3072)", style=solid];
+"886 quantize_per_channel_default_44" -> "887 dequantize_per_channel_default_44"  [label="(768, 3072)", style=solid];
+"887 dequantize_per_channel_default_44" -> "889 linear_43"  [label="(768, 3072)", style=solid];
+"888 encoder_layers_encoder_layer_10_mlp_3_bias_0_0" -> "889 linear_43"  [label="(768,)", style=solid];
+"889 linear_43" -> "890 dropout_33"  [label="(1, 197, 768)", style=solid];
+"890 dropout_33" -> "891 add_22"  [label="(1, 197, 768)", style=solid];
+"891 add_22" -> "892 layer_norm_22"  [label="(1, 197, 768)", style=solid];
+"891 add_22" -> "941 add_23"  [label="(1, 197, 768)", style=solid];
+"892 layer_norm_22" -> "893 transpose_66"  [label="(1, 197, 768)", style=solid];
+"893 transpose_66" -> "894 transpose_66_0_0_nncf_smooth_quant_0"  [label="(197, 1, 768)", style=solid];
+"894 transpose_66_0_0_nncf_smooth_quant_0" -> "895 quantize_per_tensor_default_67"  [label="(197, 1, 768)", style=solid];
+"895 quantize_per_tensor_default_67" -> "896 dequantize_per_tensor_default_67"  [label="(197, 1, 768)", style=solid];
+"896 dequantize_per_tensor_default_67" -> "903 linear_44"  [label="(197, 1, 768)", style=solid];
+"897 linear_44_scale_0" -> "900 quantize_per_channel_default_45"  [label="(2304,)", style=solid];
+"897 linear_44_scale_0" -> "901 dequantize_per_channel_default_45"  [label="(2304,)", style=solid];
+"898 linear_44_zero_point_0" -> "900 quantize_per_channel_default_45"  [label="(2304,)", style=solid];
+"898 linear_44_zero_point_0" -> "901 dequantize_per_channel_default_45"  [label="(2304,)", style=solid];
+"899 compressed_weight_updated_constant45" -> "900 quantize_per_channel_default_45"  [label="(2304, 768)", style=solid];
+"900 quantize_per_channel_default_45" -> "901 dequantize_per_channel_default_45"  [label="(2304, 768)", style=solid];
+"901 dequantize_per_channel_default_45" -> "903 linear_44"  [label="(2304, 768)", style=solid];
+"902 encoder_layers_encoder_layer_11_self_attention_in_proj_bias_0_0" -> "903 linear_44"  [label="(2304,)", style=solid];
+"903 linear_44" -> "904 unflatten_11"  [label="(197, 1, 2304)", style=solid];
+"904 unflatten_11" -> "905 unsqueeze_11"  [label="(197, 1, 3, 768)", style=solid];
+"905 unsqueeze_11" -> "906 transpose_67"  [label="(1, 197, 1, 3, 768)", style=solid];
+"906 transpose_67" -> "907 squeeze_11"  [label="(3, 197, 1, 1, 768)", style=solid];
+"907 squeeze_11" -> "908 contiguous_11"  [label="(3, 197, 1, 768)", style=solid];
+"908 contiguous_11" -> "909 quantize_per_tensor_default_68"  [label="(3, 197, 1, 768)", style=solid];
+"908 contiguous_11" -> "912 quantize_per_tensor_default_69"  [label="(3, 197, 1, 768)", style=solid];
+"908 contiguous_11" -> "915 select_35"  [label="(3, 197, 1, 768)", style=solid];
+"909 quantize_per_tensor_default_68" -> "910 dequantize_per_tensor_default_68"  [label="(3, 197, 1, 768)", style=solid];
+"910 dequantize_per_tensor_default_68" -> "911 select_33"  [label="(3, 197, 1, 768)", style=solid];
+"911 select_33" -> "916 view_88"  [label="(197, 1, 768)", style=solid];
+"912 quantize_per_tensor_default_69" -> "913 dequantize_per_tensor_default_69"  [label="(3, 197, 1, 768)", style=solid];
+"913 dequantize_per_tensor_default_69" -> "914 select_34"  [label="(3, 197, 1, 768)", style=solid];
+"914 select_34" -> "918 view_89"  [label="(197, 1, 768)", style=solid];
+"915 select_35" -> "920 view_90"  [label="(197, 1, 768)", style=solid];
+"916 view_88" -> "917 transpose_68"  [label="(197, 12, 64)", style=solid];
+"917 transpose_68" -> "922 view_91"  [label="(12, 197, 64)", style=solid];
+"918 view_89" -> "919 transpose_69"  [label="(197, 12, 64)", style=solid];
+"919 transpose_69" -> "923 view_92"  [label="(12, 197, 64)", style=solid];
+"920 view_90" -> "921 transpose_70"  [label="(197, 12, 64)", style=solid];
+"921 transpose_70" -> "924 view_93"  [label="(12, 197, 64)", style=solid];
+"922 view_91" -> "925 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
+"923 view_92" -> "925 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
+"924 view_93" -> "925 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
+"925 scaled_dot_product_attention_11" -> "926 permute_12"  [label="(1, 12, 197, 64)", style=solid];
+"926 permute_12" -> "927 view_94"  [label="(197, 1, 12, 64)", style=solid];
+"927 view_94" -> "928 view_94_0_0_nncf_smooth_quant_0"  [label="(197, 768)", style=solid];
+"928 view_94_0_0_nncf_smooth_quant_0" -> "929 quantize_per_tensor_default_70"  [label="(197, 768)", style=solid];
+"929 quantize_per_tensor_default_70" -> "930 dequantize_per_tensor_default_70"  [label="(197, 768)", style=solid];
+"930 dequantize_per_tensor_default_70" -> "937 linear_45"  [label="(197, 768)", style=solid];
+"931 linear_45_scale_0" -> "934 quantize_per_channel_default_46"  [label="(768,)", style=solid];
+"931 linear_45_scale_0" -> "935 dequantize_per_channel_default_46"  [label="(768,)", style=solid];
+"932 linear_45_zero_point_0" -> "934 quantize_per_channel_default_46"  [label="(768,)", style=solid];
+"932 linear_45_zero_point_0" -> "935 dequantize_per_channel_default_46"  [label="(768,)", style=solid];
+"933 compressed_weight_updated_constant46" -> "934 quantize_per_channel_default_46"  [label="(768, 768)", style=solid];
+"934 quantize_per_channel_default_46" -> "935 dequantize_per_channel_default_46"  [label="(768, 768)", style=solid];
+"935 dequantize_per_channel_default_46" -> "937 linear_45"  [label="(768, 768)", style=solid];
+"936 encoder_layers_encoder_layer_11_self_attention_out_proj_bias_0_0" -> "937 linear_45"  [label="(768,)", style=solid];
+"937 linear_45" -> "938 view_95"  [label="(197, 768)", style=solid];
+"938 view_95" -> "939 transpose_71"  [label="(197, 1, 768)", style=solid];
+"939 transpose_71" -> "940 dropout_34"  [label="(1, 197, 768)", style=solid];
+"940 dropout_34" -> "941 add_23"  [label="(1, 197, 768)", style=solid];
+"941 add_23" -> "942 layer_norm_23"  [label="(1, 197, 768)", style=solid];
+"941 add_23" -> "966 add_24"  [label="(1, 197, 768)", style=solid];
+"942 layer_norm_23" -> "943 layer_norm_23_0_0_nncf_smooth_quant_0"  [label="(1, 197, 768)", style=solid];
+"943 layer_norm_23_0_0_nncf_smooth_quant_0" -> "944 quantize_per_tensor_default_71"  [label="(1, 197, 768)", style=solid];
+"944 quantize_per_tensor_default_71" -> "945 dequantize_per_tensor_default_71"  [label="(1, 197, 768)", style=solid];
+"945 dequantize_per_tensor_default_71" -> "952 linear_46"  [label="(1, 197, 768)", style=solid];
+"946 linear_46_scale_0" -> "949 quantize_per_channel_default_47"  [label="(3072,)", style=solid];
+"946 linear_46_scale_0" -> "950 dequantize_per_channel_default_47"  [label="(3072,)", style=solid];
+"947 linear_46_zero_point_0" -> "949 quantize_per_channel_default_47"  [label="(3072,)", style=solid];
+"947 linear_46_zero_point_0" -> "950 dequantize_per_channel_default_47"  [label="(3072,)", style=solid];
+"948 compressed_weight_updated_constant47" -> "949 quantize_per_channel_default_47"  [label="(3072, 768)", style=solid];
+"949 quantize_per_channel_default_47" -> "950 dequantize_per_channel_default_47"  [label="(3072, 768)", style=solid];
+"950 dequantize_per_channel_default_47" -> "952 linear_46"  [label="(3072, 768)", style=solid];
+"951 encoder_layers_encoder_layer_11_mlp_0_bias_0_0" -> "952 linear_46"  [label="(3072,)", style=solid];
+"952 linear_46" -> "953 gelu_11"  [label="(1, 197, 3072)", style=solid];
+"953 gelu_11" -> "954 dropout_35"  [label="(1, 197, 3072)", style=solid];
+"954 dropout_35" -> "955 dropout_35_0_0_nncf_smooth_quant_0"  [label="(1, 197, 3072)", style=solid];
+"955 dropout_35_0_0_nncf_smooth_quant_0" -> "956 quantize_per_tensor_default_72"  [label="(1, 197, 3072)", style=solid];
+"956 quantize_per_tensor_default_72" -> "957 dequantize_per_tensor_default_72"  [label="(1, 197, 3072)", style=solid];
+"957 dequantize_per_tensor_default_72" -> "964 linear_47"  [label="(1, 197, 3072)", style=solid];
+"958 linear_47_scale_0" -> "961 quantize_per_channel_default_48"  [label="(768,)", style=solid];
+"958 linear_47_scale_0" -> "962 dequantize_per_channel_default_48"  [label="(768,)", style=solid];
+"959 linear_47_zero_point_0" -> "961 quantize_per_channel_default_48"  [label="(768,)", style=solid];
+"959 linear_47_zero_point_0" -> "962 dequantize_per_channel_default_48"  [label="(768,)", style=solid];
+"960 compressed_weight_updated_constant48" -> "961 quantize_per_channel_default_48"  [label="(768, 3072)", style=solid];
+"961 quantize_per_channel_default_48" -> "962 dequantize_per_channel_default_48"  [label="(768, 3072)", style=solid];
+"962 dequantize_per_channel_default_48" -> "964 linear_47"  [label="(768, 3072)", style=solid];
+"963 encoder_layers_encoder_layer_11_mlp_3_bias_0_0" -> "964 linear_47"  [label="(768,)", style=solid];
+"964 linear_47" -> "965 dropout_36"  [label="(1, 197, 768)", style=solid];
+"965 dropout_36" -> "966 add_24"  [label="(1, 197, 768)", style=solid];
+"966 add_24" -> "967 layer_norm_24"  [label="(1, 197, 768)", style=solid];
 "967 layer_norm_24" -> "968 slice_1"  [label="(1, 197, 768)", style=solid];
 "968 slice_1" -> "969 select_36"  [label="(1, 197, 768)", style=solid];
 "969 select_36" -> "970 select_36_0_0_nncf_smooth_quant_0"  [label="(1, 768)", style=solid];
@@ -2106,6 +2106,6 @@ strict digraph  {
 "975 compressed_weight_updated_constant49" -> "976 quantize_per_channel_default_49"  [label="(1000, 768)", style=solid];
 "976 quantize_per_channel_default_49" -> "977 dequantize_per_channel_default_49"  [label="(1000, 768)", style=solid];
 "977 dequantize_per_channel_default_49" -> "979 linear_48"  [label="(1000, 768)", style=solid];
-"978 _param_constant151_0_0" -> "979 linear_48"  [label="(1000,)", style=solid];
+"978 heads_head_bias_0_0" -> "979 linear_48"  [label="(1000,)", style=solid];
 "979 linear_48" -> "980 output"  [label="(1, 1000)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/reference_attributes/default_shared_attribute_test_model.json b/tests/torch/data/reference_graphs/fx/reference_attributes/default_shared_attribute_test_model.json
new file mode 100644
index 00000000000..0ea3c861ab7
--- /dev/null
+++ b/tests/torch/data/reference_graphs/fx/reference_attributes/default_shared_attribute_test_model.json
@@ -0,0 +1,18 @@
+{
+    "conv_a_weight": false,
+    "conv_a_bias": false,
+    "conv_b_weight": false,
+    "conv_b_bias": false,
+    "conv_c_weight": false,
+    "conv_c_bias": false,
+    "bias": true,
+    "x": false,
+    "conv2d": false,
+    "conv2d_1": false,
+    "add_": false,
+    "add__1": false,
+    "add": false,
+    "conv2d_2": false,
+    "add_1": false,
+    "output_1": false
+}
\ No newline at end of file
diff --git a/tests/torch/data/reference_graphs/fx/reference_attributes/not_unified_shared_attribute_test_model.json b/tests/torch/data/reference_graphs/fx/reference_attributes/not_unified_shared_attribute_test_model.json
deleted file mode 100644
index fd4489ff8b6..00000000000
--- a/tests/torch/data/reference_graphs/fx/reference_attributes/not_unified_shared_attribute_test_model.json
+++ /dev/null
@@ -1,20 +0,0 @@
-{
-    "arg0_1": false,
-    "_param_constant0": false,
-    "_param_constant1": false,
-    "conv2d": false,
-    "_param_constant2": false,
-    "_param_constant3": false,
-    "conv2d_1": false,
-    "_tensor_constant0": true,
-    "add_": false,
-    "_tensor_constant0_1": true,
-    "add__1": false,
-    "add": false,
-    "_param_constant4": false,
-    "_param_constant5": false,
-    "conv2d_2": false,
-    "_tensor_constant0_2": true,
-    "add_1": false,
-    "output": false
-}
\ No newline at end of file
diff --git a/tests/torch/data/reference_graphs/fx/reference_attributes/unified_shared_attribute_test_model.json b/tests/torch/data/reference_graphs/fx/reference_attributes/unified_shared_attribute_test_model.json
deleted file mode 100644
index 4c57c9317d2..00000000000
--- a/tests/torch/data/reference_graphs/fx/reference_attributes/unified_shared_attribute_test_model.json
+++ /dev/null
@@ -1,18 +0,0 @@
-{
-    "arg0_1": false,
-    "_param_constant0": false,
-    "_param_constant1": false,
-    "conv2d": false,
-    "_param_constant2": false,
-    "_param_constant3": false,
-    "conv2d_1": false,
-    "_tensor_constant0": true,
-    "add_": false,
-    "add__1": false,
-    "add": false,
-    "_param_constant4": false,
-    "_param_constant5": false,
-    "conv2d_2": false,
-    "add_1": false,
-    "output": false
-}
\ No newline at end of file
diff --git a/tests/torch/data/reference_graphs/fx/reference_metatypes/mobilenet_v3_small.json b/tests/torch/data/reference_graphs/fx/reference_metatypes/mobilenet_v3_small.json
index 64c7d3b0e66..f5112693c40 100644
--- a/tests/torch/data/reference_graphs/fx/reference_metatypes/mobilenet_v3_small.json
+++ b/tests/torch/data/reference_graphs/fx/reference_metatypes/mobilenet_v3_small.json
@@ -1,507 +1,405 @@
 {
-    "arg0_1": "PTInputNoopMetatype",
-    "_param_constant0": "PTConstNoopMetatype",
+    "features_0_0_weight": "PTConstNoopMetatype",
+    "features_0_1_weight": "PTConstNoopMetatype",
+    "features_0_1_bias": "PTConstNoopMetatype",
+    "features_1_block_0_0_weight": "PTConstNoopMetatype",
+    "features_1_block_0_1_weight": "PTConstNoopMetatype",
+    "features_1_block_0_1_bias": "PTConstNoopMetatype",
+    "features_1_block_1_fc1_weight": "PTConstNoopMetatype",
+    "features_1_block_1_fc1_bias": "PTConstNoopMetatype",
+    "features_1_block_1_fc2_weight": "PTConstNoopMetatype",
+    "features_1_block_1_fc2_bias": "PTConstNoopMetatype",
+    "features_1_block_2_0_weight": "PTConstNoopMetatype",
+    "features_1_block_2_1_weight": "PTConstNoopMetatype",
+    "features_1_block_2_1_bias": "PTConstNoopMetatype",
+    "features_2_block_0_0_weight": "PTConstNoopMetatype",
+    "features_2_block_0_1_weight": "PTConstNoopMetatype",
+    "features_2_block_0_1_bias": "PTConstNoopMetatype",
+    "features_2_block_1_0_weight": "PTConstNoopMetatype",
+    "features_2_block_1_1_weight": "PTConstNoopMetatype",
+    "features_2_block_1_1_bias": "PTConstNoopMetatype",
+    "features_2_block_2_0_weight": "PTConstNoopMetatype",
+    "features_2_block_2_1_weight": "PTConstNoopMetatype",
+    "features_2_block_2_1_bias": "PTConstNoopMetatype",
+    "features_3_block_0_0_weight": "PTConstNoopMetatype",
+    "features_3_block_0_1_weight": "PTConstNoopMetatype",
+    "features_3_block_0_1_bias": "PTConstNoopMetatype",
+    "features_3_block_1_0_weight": "PTConstNoopMetatype",
+    "features_3_block_1_1_weight": "PTConstNoopMetatype",
+    "features_3_block_1_1_bias": "PTConstNoopMetatype",
+    "features_3_block_2_0_weight": "PTConstNoopMetatype",
+    "features_3_block_2_1_weight": "PTConstNoopMetatype",
+    "features_3_block_2_1_bias": "PTConstNoopMetatype",
+    "features_4_block_0_0_weight": "PTConstNoopMetatype",
+    "features_4_block_0_1_weight": "PTConstNoopMetatype",
+    "features_4_block_0_1_bias": "PTConstNoopMetatype",
+    "features_4_block_1_0_weight": "PTConstNoopMetatype",
+    "features_4_block_1_1_weight": "PTConstNoopMetatype",
+    "features_4_block_1_1_bias": "PTConstNoopMetatype",
+    "features_4_block_2_fc1_weight": "PTConstNoopMetatype",
+    "features_4_block_2_fc1_bias": "PTConstNoopMetatype",
+    "features_4_block_2_fc2_weight": "PTConstNoopMetatype",
+    "features_4_block_2_fc2_bias": "PTConstNoopMetatype",
+    "features_4_block_3_0_weight": "PTConstNoopMetatype",
+    "features_4_block_3_1_weight": "PTConstNoopMetatype",
+    "features_4_block_3_1_bias": "PTConstNoopMetatype",
+    "features_5_block_0_0_weight": "PTConstNoopMetatype",
+    "features_5_block_0_1_weight": "PTConstNoopMetatype",
+    "features_5_block_0_1_bias": "PTConstNoopMetatype",
+    "features_5_block_1_0_weight": "PTConstNoopMetatype",
+    "features_5_block_1_1_weight": "PTConstNoopMetatype",
+    "features_5_block_1_1_bias": "PTConstNoopMetatype",
+    "features_5_block_2_fc1_weight": "PTConstNoopMetatype",
+    "features_5_block_2_fc1_bias": "PTConstNoopMetatype",
+    "features_5_block_2_fc2_weight": "PTConstNoopMetatype",
+    "features_5_block_2_fc2_bias": "PTConstNoopMetatype",
+    "features_5_block_3_0_weight": "PTConstNoopMetatype",
+    "features_5_block_3_1_weight": "PTConstNoopMetatype",
+    "features_5_block_3_1_bias": "PTConstNoopMetatype",
+    "features_6_block_0_0_weight": "PTConstNoopMetatype",
+    "features_6_block_0_1_weight": "PTConstNoopMetatype",
+    "features_6_block_0_1_bias": "PTConstNoopMetatype",
+    "features_6_block_1_0_weight": "PTConstNoopMetatype",
+    "features_6_block_1_1_weight": "PTConstNoopMetatype",
+    "features_6_block_1_1_bias": "PTConstNoopMetatype",
+    "features_6_block_2_fc1_weight": "PTConstNoopMetatype",
+    "features_6_block_2_fc1_bias": "PTConstNoopMetatype",
+    "features_6_block_2_fc2_weight": "PTConstNoopMetatype",
+    "features_6_block_2_fc2_bias": "PTConstNoopMetatype",
+    "features_6_block_3_0_weight": "PTConstNoopMetatype",
+    "features_6_block_3_1_weight": "PTConstNoopMetatype",
+    "features_6_block_3_1_bias": "PTConstNoopMetatype",
+    "features_7_block_0_0_weight": "PTConstNoopMetatype",
+    "features_7_block_0_1_weight": "PTConstNoopMetatype",
+    "features_7_block_0_1_bias": "PTConstNoopMetatype",
+    "features_7_block_1_0_weight": "PTConstNoopMetatype",
+    "features_7_block_1_1_weight": "PTConstNoopMetatype",
+    "features_7_block_1_1_bias": "PTConstNoopMetatype",
+    "features_7_block_2_fc1_weight": "PTConstNoopMetatype",
+    "features_7_block_2_fc1_bias": "PTConstNoopMetatype",
+    "features_7_block_2_fc2_weight": "PTConstNoopMetatype",
+    "features_7_block_2_fc2_bias": "PTConstNoopMetatype",
+    "features_7_block_3_0_weight": "PTConstNoopMetatype",
+    "features_7_block_3_1_weight": "PTConstNoopMetatype",
+    "features_7_block_3_1_bias": "PTConstNoopMetatype",
+    "features_8_block_0_0_weight": "PTConstNoopMetatype",
+    "features_8_block_0_1_weight": "PTConstNoopMetatype",
+    "features_8_block_0_1_bias": "PTConstNoopMetatype",
+    "features_8_block_1_0_weight": "PTConstNoopMetatype",
+    "features_8_block_1_1_weight": "PTConstNoopMetatype",
+    "features_8_block_1_1_bias": "PTConstNoopMetatype",
+    "features_8_block_2_fc1_weight": "PTConstNoopMetatype",
+    "features_8_block_2_fc1_bias": "PTConstNoopMetatype",
+    "features_8_block_2_fc2_weight": "PTConstNoopMetatype",
+    "features_8_block_2_fc2_bias": "PTConstNoopMetatype",
+    "features_8_block_3_0_weight": "PTConstNoopMetatype",
+    "features_8_block_3_1_weight": "PTConstNoopMetatype",
+    "features_8_block_3_1_bias": "PTConstNoopMetatype",
+    "features_9_block_0_0_weight": "PTConstNoopMetatype",
+    "features_9_block_0_1_weight": "PTConstNoopMetatype",
+    "features_9_block_0_1_bias": "PTConstNoopMetatype",
+    "features_9_block_1_0_weight": "PTConstNoopMetatype",
+    "features_9_block_1_1_weight": "PTConstNoopMetatype",
+    "features_9_block_1_1_bias": "PTConstNoopMetatype",
+    "features_9_block_2_fc1_weight": "PTConstNoopMetatype",
+    "features_9_block_2_fc1_bias": "PTConstNoopMetatype",
+    "features_9_block_2_fc2_weight": "PTConstNoopMetatype",
+    "features_9_block_2_fc2_bias": "PTConstNoopMetatype",
+    "features_9_block_3_0_weight": "PTConstNoopMetatype",
+    "features_9_block_3_1_weight": "PTConstNoopMetatype",
+    "features_9_block_3_1_bias": "PTConstNoopMetatype",
+    "features_10_block_0_0_weight": "PTConstNoopMetatype",
+    "features_10_block_0_1_weight": "PTConstNoopMetatype",
+    "features_10_block_0_1_bias": "PTConstNoopMetatype",
+    "features_10_block_1_0_weight": "PTConstNoopMetatype",
+    "features_10_block_1_1_weight": "PTConstNoopMetatype",
+    "features_10_block_1_1_bias": "PTConstNoopMetatype",
+    "features_10_block_2_fc1_weight": "PTConstNoopMetatype",
+    "features_10_block_2_fc1_bias": "PTConstNoopMetatype",
+    "features_10_block_2_fc2_weight": "PTConstNoopMetatype",
+    "features_10_block_2_fc2_bias": "PTConstNoopMetatype",
+    "features_10_block_3_0_weight": "PTConstNoopMetatype",
+    "features_10_block_3_1_weight": "PTConstNoopMetatype",
+    "features_10_block_3_1_bias": "PTConstNoopMetatype",
+    "features_11_block_0_0_weight": "PTConstNoopMetatype",
+    "features_11_block_0_1_weight": "PTConstNoopMetatype",
+    "features_11_block_0_1_bias": "PTConstNoopMetatype",
+    "features_11_block_1_0_weight": "PTConstNoopMetatype",
+    "features_11_block_1_1_weight": "PTConstNoopMetatype",
+    "features_11_block_1_1_bias": "PTConstNoopMetatype",
+    "features_11_block_2_fc1_weight": "PTConstNoopMetatype",
+    "features_11_block_2_fc1_bias": "PTConstNoopMetatype",
+    "features_11_block_2_fc2_weight": "PTConstNoopMetatype",
+    "features_11_block_2_fc2_bias": "PTConstNoopMetatype",
+    "features_11_block_3_0_weight": "PTConstNoopMetatype",
+    "features_11_block_3_1_weight": "PTConstNoopMetatype",
+    "features_11_block_3_1_bias": "PTConstNoopMetatype",
+    "features_12_0_weight": "PTConstNoopMetatype",
+    "features_12_1_weight": "PTConstNoopMetatype",
+    "features_12_1_bias": "PTConstNoopMetatype",
+    "classifier_0_weight": "PTConstNoopMetatype",
+    "classifier_0_bias": "PTConstNoopMetatype",
+    "classifier_3_weight": "PTConstNoopMetatype",
+    "classifier_3_bias": "PTConstNoopMetatype",
+    "features_0_1_running_mean": "PTConstNoopMetatype",
+    "features_0_1_running_var": "PTConstNoopMetatype",
+    "features_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_1_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_1_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_1_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_1_block_2_1_running_mean": "PTConstNoopMetatype",
+    "features_1_block_2_1_running_var": "PTConstNoopMetatype",
+    "features_1_block_2_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_2_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_2_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_2_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_2_block_1_1_running_mean": "PTConstNoopMetatype",
+    "features_2_block_1_1_running_var": "PTConstNoopMetatype",
+    "features_2_block_1_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_2_block_2_1_running_mean": "PTConstNoopMetatype",
+    "features_2_block_2_1_running_var": "PTConstNoopMetatype",
+    "features_2_block_2_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_3_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_3_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_3_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_3_block_1_1_running_mean": "PTConstNoopMetatype",
+    "features_3_block_1_1_running_var": "PTConstNoopMetatype",
+    "features_3_block_1_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_3_block_2_1_running_mean": "PTConstNoopMetatype",
+    "features_3_block_2_1_running_var": "PTConstNoopMetatype",
+    "features_3_block_2_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_4_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_4_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_4_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_4_block_1_1_running_mean": "PTConstNoopMetatype",
+    "features_4_block_1_1_running_var": "PTConstNoopMetatype",
+    "features_4_block_1_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_4_block_3_1_running_mean": "PTConstNoopMetatype",
+    "features_4_block_3_1_running_var": "PTConstNoopMetatype",
+    "features_4_block_3_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_5_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_5_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_5_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_5_block_1_1_running_mean": "PTConstNoopMetatype",
+    "features_5_block_1_1_running_var": "PTConstNoopMetatype",
+    "features_5_block_1_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_5_block_3_1_running_mean": "PTConstNoopMetatype",
+    "features_5_block_3_1_running_var": "PTConstNoopMetatype",
+    "features_5_block_3_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_6_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_6_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_6_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_6_block_1_1_running_mean": "PTConstNoopMetatype",
+    "features_6_block_1_1_running_var": "PTConstNoopMetatype",
+    "features_6_block_1_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_6_block_3_1_running_mean": "PTConstNoopMetatype",
+    "features_6_block_3_1_running_var": "PTConstNoopMetatype",
+    "features_6_block_3_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_7_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_7_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_7_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_7_block_1_1_running_mean": "PTConstNoopMetatype",
+    "features_7_block_1_1_running_var": "PTConstNoopMetatype",
+    "features_7_block_1_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_7_block_3_1_running_mean": "PTConstNoopMetatype",
+    "features_7_block_3_1_running_var": "PTConstNoopMetatype",
+    "features_7_block_3_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_8_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_8_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_8_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_8_block_1_1_running_mean": "PTConstNoopMetatype",
+    "features_8_block_1_1_running_var": "PTConstNoopMetatype",
+    "features_8_block_1_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_8_block_3_1_running_mean": "PTConstNoopMetatype",
+    "features_8_block_3_1_running_var": "PTConstNoopMetatype",
+    "features_8_block_3_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_9_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_9_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_9_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_9_block_1_1_running_mean": "PTConstNoopMetatype",
+    "features_9_block_1_1_running_var": "PTConstNoopMetatype",
+    "features_9_block_1_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_9_block_3_1_running_mean": "PTConstNoopMetatype",
+    "features_9_block_3_1_running_var": "PTConstNoopMetatype",
+    "features_9_block_3_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_10_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_10_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_10_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_10_block_1_1_running_mean": "PTConstNoopMetatype",
+    "features_10_block_1_1_running_var": "PTConstNoopMetatype",
+    "features_10_block_1_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_10_block_3_1_running_mean": "PTConstNoopMetatype",
+    "features_10_block_3_1_running_var": "PTConstNoopMetatype",
+    "features_10_block_3_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_11_block_0_1_running_mean": "PTConstNoopMetatype",
+    "features_11_block_0_1_running_var": "PTConstNoopMetatype",
+    "features_11_block_0_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_11_block_1_1_running_mean": "PTConstNoopMetatype",
+    "features_11_block_1_1_running_var": "PTConstNoopMetatype",
+    "features_11_block_1_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_11_block_3_1_running_mean": "PTConstNoopMetatype",
+    "features_11_block_3_1_running_var": "PTConstNoopMetatype",
+    "features_11_block_3_1_num_batches_tracked": "PTConstNoopMetatype",
+    "features_12_1_running_mean": "PTConstNoopMetatype",
+    "features_12_1_running_var": "PTConstNoopMetatype",
+    "features_12_1_num_batches_tracked": "PTConstNoopMetatype",
+    "x": "PTInputNoopMetatype",
     "conv2d": "PTConv2dMetatype",
-    "empty": "UnknownMetatype",
-    "_param_constant1": "PTConstNoopMetatype",
-    "_param_constant2": "PTConstNoopMetatype",
-    "_tensor_constant0": "PTConstNoopMetatype",
-    "_tensor_constant1": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training": "PTBatchNormMetatype",
-    "getitem": "PTGatherMetatype",
-    "getitem_1": "PTGatherMetatype",
-    "getitem_2": "PTGatherMetatype",
+    "batch_norm": "PTBatchNormMetatype",
     "hardswish_": "PTHardSwishMetatype",
-    "_param_constant3": "PTConstNoopMetatype",
     "conv2d_1": "PTDepthwiseConv2dSubtype",
-    "empty_1": "UnknownMetatype",
-    "_param_constant4": "PTConstNoopMetatype",
-    "_param_constant5": "PTConstNoopMetatype",
-    "_tensor_constant2": "PTConstNoopMetatype",
-    "_tensor_constant3": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_1": "PTBatchNormMetatype",
-    "getitem_3": "PTGatherMetatype",
-    "getitem_4": "PTGatherMetatype",
-    "getitem_5": "PTGatherMetatype",
+    "batch_norm_1": "PTBatchNormMetatype",
     "relu_": "PTRELUMetatype",
     "adaptive_avg_pool2d": "PTAvgPool2dMetatype",
-    "_param_constant6": "PTConstNoopMetatype",
-    "_param_constant7": "PTConstNoopMetatype",
     "conv2d_2": "PTConv2dMetatype",
     "relu": "PTRELUMetatype",
-    "_param_constant8": "PTConstNoopMetatype",
-    "_param_constant9": "PTConstNoopMetatype",
     "conv2d_3": "PTConv2dMetatype",
     "hardsigmoid": "PTHardSigmoidMetatype",
     "mul": "PTMulMetatype",
-    "_param_constant10": "PTConstNoopMetatype",
     "conv2d_4": "PTConv2dMetatype",
-    "empty_2": "UnknownMetatype",
-    "_param_constant11": "PTConstNoopMetatype",
-    "_param_constant12": "PTConstNoopMetatype",
-    "_tensor_constant4": "PTConstNoopMetatype",
-    "_tensor_constant5": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_2": "PTBatchNormMetatype",
-    "getitem_6": "PTGatherMetatype",
-    "getitem_7": "PTGatherMetatype",
-    "getitem_8": "PTGatherMetatype",
-    "_param_constant13": "PTConstNoopMetatype",
+    "batch_norm_2": "PTBatchNormMetatype",
     "conv2d_5": "PTConv2dMetatype",
-    "empty_3": "UnknownMetatype",
-    "_param_constant14": "PTConstNoopMetatype",
-    "_param_constant15": "PTConstNoopMetatype",
-    "_tensor_constant6": "PTConstNoopMetatype",
-    "_tensor_constant7": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_3": "PTBatchNormMetatype",
-    "getitem_9": "PTGatherMetatype",
-    "getitem_10": "PTGatherMetatype",
-    "getitem_11": "PTGatherMetatype",
+    "batch_norm_3": "PTBatchNormMetatype",
     "relu__1": "PTRELUMetatype",
-    "_param_constant16": "PTConstNoopMetatype",
     "conv2d_6": "PTDepthwiseConv2dSubtype",
-    "empty_4": "UnknownMetatype",
-    "_param_constant17": "PTConstNoopMetatype",
-    "_param_constant18": "PTConstNoopMetatype",
-    "_tensor_constant8": "PTConstNoopMetatype",
-    "_tensor_constant9": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_4": "PTBatchNormMetatype",
-    "getitem_12": "PTGatherMetatype",
-    "getitem_13": "PTGatherMetatype",
-    "getitem_14": "PTGatherMetatype",
+    "batch_norm_4": "PTBatchNormMetatype",
     "relu__2": "PTRELUMetatype",
-    "_param_constant19": "PTConstNoopMetatype",
     "conv2d_7": "PTConv2dMetatype",
-    "empty_5": "UnknownMetatype",
-    "_param_constant20": "PTConstNoopMetatype",
-    "_param_constant21": "PTConstNoopMetatype",
-    "_tensor_constant10": "PTConstNoopMetatype",
-    "_tensor_constant11": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_5": "PTBatchNormMetatype",
-    "getitem_15": "PTGatherMetatype",
-    "getitem_16": "PTGatherMetatype",
-    "getitem_17": "PTGatherMetatype",
-    "_param_constant22": "PTConstNoopMetatype",
+    "batch_norm_5": "PTBatchNormMetatype",
     "conv2d_8": "PTConv2dMetatype",
-    "empty_6": "UnknownMetatype",
-    "_param_constant23": "PTConstNoopMetatype",
-    "_param_constant24": "PTConstNoopMetatype",
-    "_tensor_constant12": "PTConstNoopMetatype",
-    "_tensor_constant13": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_6": "PTBatchNormMetatype",
-    "getitem_18": "PTGatherMetatype",
-    "getitem_19": "PTGatherMetatype",
-    "getitem_20": "PTGatherMetatype",
+    "batch_norm_6": "PTBatchNormMetatype",
     "relu__3": "PTRELUMetatype",
-    "_param_constant25": "PTConstNoopMetatype",
     "conv2d_9": "PTDepthwiseConv2dSubtype",
-    "empty_7": "UnknownMetatype",
-    "_param_constant26": "PTConstNoopMetatype",
-    "_param_constant27": "PTConstNoopMetatype",
-    "_tensor_constant14": "PTConstNoopMetatype",
-    "_tensor_constant15": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_7": "PTBatchNormMetatype",
-    "getitem_21": "PTGatherMetatype",
-    "getitem_22": "PTGatherMetatype",
-    "getitem_23": "PTGatherMetatype",
+    "batch_norm_7": "PTBatchNormMetatype",
     "relu__4": "PTRELUMetatype",
-    "_param_constant28": "PTConstNoopMetatype",
     "conv2d_10": "PTConv2dMetatype",
-    "empty_8": "UnknownMetatype",
-    "_param_constant29": "PTConstNoopMetatype",
-    "_param_constant30": "PTConstNoopMetatype",
-    "_tensor_constant16": "PTConstNoopMetatype",
-    "_tensor_constant17": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_8": "PTBatchNormMetatype",
-    "getitem_24": "PTGatherMetatype",
-    "getitem_25": "PTGatherMetatype",
-    "getitem_26": "PTGatherMetatype",
+    "batch_norm_8": "PTBatchNormMetatype",
     "add_": "PTAddMetatype",
-    "_param_constant31": "PTConstNoopMetatype",
     "conv2d_11": "PTConv2dMetatype",
-    "empty_9": "UnknownMetatype",
-    "_param_constant32": "PTConstNoopMetatype",
-    "_param_constant33": "PTConstNoopMetatype",
-    "_tensor_constant18": "PTConstNoopMetatype",
-    "_tensor_constant19": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_9": "PTBatchNormMetatype",
-    "getitem_27": "PTGatherMetatype",
-    "getitem_28": "PTGatherMetatype",
-    "getitem_29": "PTGatherMetatype",
+    "batch_norm_9": "PTBatchNormMetatype",
     "hardswish__1": "PTHardSwishMetatype",
-    "_param_constant34": "PTConstNoopMetatype",
     "conv2d_12": "PTDepthwiseConv2dSubtype",
-    "empty_10": "UnknownMetatype",
-    "_param_constant35": "PTConstNoopMetatype",
-    "_param_constant36": "PTConstNoopMetatype",
-    "_tensor_constant20": "PTConstNoopMetatype",
-    "_tensor_constant21": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_10": "PTBatchNormMetatype",
-    "getitem_30": "PTGatherMetatype",
-    "getitem_31": "PTGatherMetatype",
-    "getitem_32": "PTGatherMetatype",
+    "batch_norm_10": "PTBatchNormMetatype",
     "hardswish__2": "PTHardSwishMetatype",
     "adaptive_avg_pool2d_1": "PTAvgPool2dMetatype",
-    "_param_constant37": "PTConstNoopMetatype",
-    "_param_constant38": "PTConstNoopMetatype",
     "conv2d_13": "PTConv2dMetatype",
     "relu_1": "PTRELUMetatype",
-    "_param_constant39": "PTConstNoopMetatype",
-    "_param_constant40": "PTConstNoopMetatype",
     "conv2d_14": "PTConv2dMetatype",
     "hardsigmoid_1": "PTHardSigmoidMetatype",
     "mul_1": "PTMulMetatype",
-    "_param_constant41": "PTConstNoopMetatype",
     "conv2d_15": "PTConv2dMetatype",
-    "empty_11": "UnknownMetatype",
-    "_param_constant42": "PTConstNoopMetatype",
-    "_param_constant43": "PTConstNoopMetatype",
-    "_tensor_constant22": "PTConstNoopMetatype",
-    "_tensor_constant23": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_11": "PTBatchNormMetatype",
-    "getitem_33": "PTGatherMetatype",
-    "getitem_34": "PTGatherMetatype",
-    "getitem_35": "PTGatherMetatype",
-    "_param_constant44": "PTConstNoopMetatype",
+    "batch_norm_11": "PTBatchNormMetatype",
     "conv2d_16": "PTConv2dMetatype",
-    "empty_12": "UnknownMetatype",
-    "_param_constant45": "PTConstNoopMetatype",
-    "_param_constant46": "PTConstNoopMetatype",
-    "_tensor_constant24": "PTConstNoopMetatype",
-    "_tensor_constant25": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_12": "PTBatchNormMetatype",
-    "getitem_36": "PTGatherMetatype",
-    "getitem_37": "PTGatherMetatype",
-    "getitem_38": "PTGatherMetatype",
+    "batch_norm_12": "PTBatchNormMetatype",
     "hardswish__3": "PTHardSwishMetatype",
-    "_param_constant47": "PTConstNoopMetatype",
     "conv2d_17": "PTDepthwiseConv2dSubtype",
-    "empty_13": "UnknownMetatype",
-    "_param_constant48": "PTConstNoopMetatype",
-    "_param_constant49": "PTConstNoopMetatype",
-    "_tensor_constant26": "PTConstNoopMetatype",
-    "_tensor_constant27": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_13": "PTBatchNormMetatype",
-    "getitem_39": "PTGatherMetatype",
-    "getitem_40": "PTGatherMetatype",
-    "getitem_41": "PTGatherMetatype",
+    "batch_norm_13": "PTBatchNormMetatype",
     "hardswish__4": "PTHardSwishMetatype",
     "adaptive_avg_pool2d_2": "PTAvgPool2dMetatype",
-    "_param_constant50": "PTConstNoopMetatype",
-    "_param_constant51": "PTConstNoopMetatype",
     "conv2d_18": "PTConv2dMetatype",
     "relu_2": "PTRELUMetatype",
-    "_param_constant52": "PTConstNoopMetatype",
-    "_param_constant53": "PTConstNoopMetatype",
     "conv2d_19": "PTConv2dMetatype",
     "hardsigmoid_2": "PTHardSigmoidMetatype",
     "mul_2": "PTMulMetatype",
-    "_param_constant54": "PTConstNoopMetatype",
     "conv2d_20": "PTConv2dMetatype",
-    "empty_14": "UnknownMetatype",
-    "_param_constant55": "PTConstNoopMetatype",
-    "_param_constant56": "PTConstNoopMetatype",
-    "_tensor_constant28": "PTConstNoopMetatype",
-    "_tensor_constant29": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_14": "PTBatchNormMetatype",
-    "getitem_42": "PTGatherMetatype",
-    "getitem_43": "PTGatherMetatype",
-    "getitem_44": "PTGatherMetatype",
+    "batch_norm_14": "PTBatchNormMetatype",
     "add__1": "PTAddMetatype",
-    "_param_constant57": "PTConstNoopMetatype",
     "conv2d_21": "PTConv2dMetatype",
-    "empty_15": "UnknownMetatype",
-    "_param_constant58": "PTConstNoopMetatype",
-    "_param_constant59": "PTConstNoopMetatype",
-    "_tensor_constant30": "PTConstNoopMetatype",
-    "_tensor_constant31": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_15": "PTBatchNormMetatype",
-    "getitem_45": "PTGatherMetatype",
-    "getitem_46": "PTGatherMetatype",
-    "getitem_47": "PTGatherMetatype",
+    "batch_norm_15": "PTBatchNormMetatype",
     "hardswish__5": "PTHardSwishMetatype",
-    "_param_constant60": "PTConstNoopMetatype",
     "conv2d_22": "PTDepthwiseConv2dSubtype",
-    "empty_16": "UnknownMetatype",
-    "_param_constant61": "PTConstNoopMetatype",
-    "_param_constant62": "PTConstNoopMetatype",
-    "_tensor_constant32": "PTConstNoopMetatype",
-    "_tensor_constant33": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_16": "PTBatchNormMetatype",
-    "getitem_48": "PTGatherMetatype",
-    "getitem_49": "PTGatherMetatype",
-    "getitem_50": "PTGatherMetatype",
+    "batch_norm_16": "PTBatchNormMetatype",
     "hardswish__6": "PTHardSwishMetatype",
     "adaptive_avg_pool2d_3": "PTAvgPool2dMetatype",
-    "_param_constant63": "PTConstNoopMetatype",
-    "_param_constant64": "PTConstNoopMetatype",
     "conv2d_23": "PTConv2dMetatype",
     "relu_3": "PTRELUMetatype",
-    "_param_constant65": "PTConstNoopMetatype",
-    "_param_constant66": "PTConstNoopMetatype",
     "conv2d_24": "PTConv2dMetatype",
     "hardsigmoid_3": "PTHardSigmoidMetatype",
     "mul_3": "PTMulMetatype",
-    "_param_constant67": "PTConstNoopMetatype",
     "conv2d_25": "PTConv2dMetatype",
-    "empty_17": "UnknownMetatype",
-    "_param_constant68": "PTConstNoopMetatype",
-    "_param_constant69": "PTConstNoopMetatype",
-    "_tensor_constant34": "PTConstNoopMetatype",
-    "_tensor_constant35": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_17": "PTBatchNormMetatype",
-    "getitem_51": "PTGatherMetatype",
-    "getitem_52": "PTGatherMetatype",
-    "getitem_53": "PTGatherMetatype",
+    "batch_norm_17": "PTBatchNormMetatype",
     "add__2": "PTAddMetatype",
-    "_param_constant70": "PTConstNoopMetatype",
     "conv2d_26": "PTConv2dMetatype",
-    "empty_18": "UnknownMetatype",
-    "_param_constant71": "PTConstNoopMetatype",
-    "_param_constant72": "PTConstNoopMetatype",
-    "_tensor_constant36": "PTConstNoopMetatype",
-    "_tensor_constant37": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_18": "PTBatchNormMetatype",
-    "getitem_54": "PTGatherMetatype",
-    "getitem_55": "PTGatherMetatype",
-    "getitem_56": "PTGatherMetatype",
+    "batch_norm_18": "PTBatchNormMetatype",
     "hardswish__7": "PTHardSwishMetatype",
-    "_param_constant73": "PTConstNoopMetatype",
     "conv2d_27": "PTDepthwiseConv2dSubtype",
-    "empty_19": "UnknownMetatype",
-    "_param_constant74": "PTConstNoopMetatype",
-    "_param_constant75": "PTConstNoopMetatype",
-    "_tensor_constant38": "PTConstNoopMetatype",
-    "_tensor_constant39": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_19": "PTBatchNormMetatype",
-    "getitem_57": "PTGatherMetatype",
-    "getitem_58": "PTGatherMetatype",
-    "getitem_59": "PTGatherMetatype",
+    "batch_norm_19": "PTBatchNormMetatype",
     "hardswish__8": "PTHardSwishMetatype",
     "adaptive_avg_pool2d_4": "PTAvgPool2dMetatype",
-    "_param_constant76": "PTConstNoopMetatype",
-    "_param_constant77": "PTConstNoopMetatype",
     "conv2d_28": "PTConv2dMetatype",
     "relu_4": "PTRELUMetatype",
-    "_param_constant78": "PTConstNoopMetatype",
-    "_param_constant79": "PTConstNoopMetatype",
     "conv2d_29": "PTConv2dMetatype",
     "hardsigmoid_4": "PTHardSigmoidMetatype",
     "mul_4": "PTMulMetatype",
-    "_param_constant80": "PTConstNoopMetatype",
     "conv2d_30": "PTConv2dMetatype",
-    "empty_20": "UnknownMetatype",
-    "_param_constant81": "PTConstNoopMetatype",
-    "_param_constant82": "PTConstNoopMetatype",
-    "_tensor_constant40": "PTConstNoopMetatype",
-    "_tensor_constant41": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_20": "PTBatchNormMetatype",
-    "getitem_60": "PTGatherMetatype",
-    "getitem_61": "PTGatherMetatype",
-    "getitem_62": "PTGatherMetatype",
-    "_param_constant83": "PTConstNoopMetatype",
+    "batch_norm_20": "PTBatchNormMetatype",
     "conv2d_31": "PTConv2dMetatype",
-    "empty_21": "UnknownMetatype",
-    "_param_constant84": "PTConstNoopMetatype",
-    "_param_constant85": "PTConstNoopMetatype",
-    "_tensor_constant42": "PTConstNoopMetatype",
-    "_tensor_constant43": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_21": "PTBatchNormMetatype",
-    "getitem_63": "PTGatherMetatype",
-    "getitem_64": "PTGatherMetatype",
-    "getitem_65": "PTGatherMetatype",
+    "batch_norm_21": "PTBatchNormMetatype",
     "hardswish__9": "PTHardSwishMetatype",
-    "_param_constant86": "PTConstNoopMetatype",
     "conv2d_32": "PTDepthwiseConv2dSubtype",
-    "empty_22": "UnknownMetatype",
-    "_param_constant87": "PTConstNoopMetatype",
-    "_param_constant88": "PTConstNoopMetatype",
-    "_tensor_constant44": "PTConstNoopMetatype",
-    "_tensor_constant45": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_22": "PTBatchNormMetatype",
-    "getitem_66": "PTGatherMetatype",
-    "getitem_67": "PTGatherMetatype",
-    "getitem_68": "PTGatherMetatype",
+    "batch_norm_22": "PTBatchNormMetatype",
     "hardswish__10": "PTHardSwishMetatype",
     "adaptive_avg_pool2d_5": "PTAvgPool2dMetatype",
-    "_param_constant89": "PTConstNoopMetatype",
-    "_param_constant90": "PTConstNoopMetatype",
     "conv2d_33": "PTConv2dMetatype",
     "relu_5": "PTRELUMetatype",
-    "_param_constant91": "PTConstNoopMetatype",
-    "_param_constant92": "PTConstNoopMetatype",
     "conv2d_34": "PTConv2dMetatype",
     "hardsigmoid_5": "PTHardSigmoidMetatype",
     "mul_5": "PTMulMetatype",
-    "_param_constant93": "PTConstNoopMetatype",
     "conv2d_35": "PTConv2dMetatype",
-    "empty_23": "UnknownMetatype",
-    "_param_constant94": "PTConstNoopMetatype",
-    "_param_constant95": "PTConstNoopMetatype",
-    "_tensor_constant46": "PTConstNoopMetatype",
-    "_tensor_constant47": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_23": "PTBatchNormMetatype",
-    "getitem_69": "PTGatherMetatype",
-    "getitem_70": "PTGatherMetatype",
-    "getitem_71": "PTGatherMetatype",
+    "batch_norm_23": "PTBatchNormMetatype",
     "add__3": "PTAddMetatype",
-    "_param_constant96": "PTConstNoopMetatype",
     "conv2d_36": "PTConv2dMetatype",
-    "empty_24": "UnknownMetatype",
-    "_param_constant97": "PTConstNoopMetatype",
-    "_param_constant98": "PTConstNoopMetatype",
-    "_tensor_constant48": "PTConstNoopMetatype",
-    "_tensor_constant49": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_24": "PTBatchNormMetatype",
-    "getitem_72": "PTGatherMetatype",
-    "getitem_73": "PTGatherMetatype",
-    "getitem_74": "PTGatherMetatype",
+    "batch_norm_24": "PTBatchNormMetatype",
     "hardswish__11": "PTHardSwishMetatype",
-    "_param_constant99": "PTConstNoopMetatype",
     "conv2d_37": "PTDepthwiseConv2dSubtype",
-    "empty_25": "UnknownMetatype",
-    "_param_constant100": "PTConstNoopMetatype",
-    "_param_constant101": "PTConstNoopMetatype",
-    "_tensor_constant50": "PTConstNoopMetatype",
-    "_tensor_constant51": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_25": "PTBatchNormMetatype",
-    "getitem_75": "PTGatherMetatype",
-    "getitem_76": "PTGatherMetatype",
-    "getitem_77": "PTGatherMetatype",
+    "batch_norm_25": "PTBatchNormMetatype",
     "hardswish__12": "PTHardSwishMetatype",
     "adaptive_avg_pool2d_6": "PTAvgPool2dMetatype",
-    "_param_constant102": "PTConstNoopMetatype",
-    "_param_constant103": "PTConstNoopMetatype",
     "conv2d_38": "PTConv2dMetatype",
     "relu_6": "PTRELUMetatype",
-    "_param_constant104": "PTConstNoopMetatype",
-    "_param_constant105": "PTConstNoopMetatype",
     "conv2d_39": "PTConv2dMetatype",
     "hardsigmoid_6": "PTHardSigmoidMetatype",
     "mul_6": "PTMulMetatype",
-    "_param_constant106": "PTConstNoopMetatype",
     "conv2d_40": "PTConv2dMetatype",
-    "empty_26": "UnknownMetatype",
-    "_param_constant107": "PTConstNoopMetatype",
-    "_param_constant108": "PTConstNoopMetatype",
-    "_tensor_constant52": "PTConstNoopMetatype",
-    "_tensor_constant53": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_26": "PTBatchNormMetatype",
-    "getitem_78": "PTGatherMetatype",
-    "getitem_79": "PTGatherMetatype",
-    "getitem_80": "PTGatherMetatype",
-    "_param_constant109": "PTConstNoopMetatype",
+    "batch_norm_26": "PTBatchNormMetatype",
     "conv2d_41": "PTConv2dMetatype",
-    "empty_27": "UnknownMetatype",
-    "_param_constant110": "PTConstNoopMetatype",
-    "_param_constant111": "PTConstNoopMetatype",
-    "_tensor_constant54": "PTConstNoopMetatype",
-    "_tensor_constant55": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_27": "PTBatchNormMetatype",
-    "getitem_81": "PTGatherMetatype",
-    "getitem_82": "PTGatherMetatype",
-    "getitem_83": "PTGatherMetatype",
+    "batch_norm_27": "PTBatchNormMetatype",
     "hardswish__13": "PTHardSwishMetatype",
-    "_param_constant112": "PTConstNoopMetatype",
     "conv2d_42": "PTDepthwiseConv2dSubtype",
-    "empty_28": "UnknownMetatype",
-    "_param_constant113": "PTConstNoopMetatype",
-    "_param_constant114": "PTConstNoopMetatype",
-    "_tensor_constant56": "PTConstNoopMetatype",
-    "_tensor_constant57": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_28": "PTBatchNormMetatype",
-    "getitem_84": "PTGatherMetatype",
-    "getitem_85": "PTGatherMetatype",
-    "getitem_86": "PTGatherMetatype",
+    "batch_norm_28": "PTBatchNormMetatype",
     "hardswish__14": "PTHardSwishMetatype",
     "adaptive_avg_pool2d_7": "PTAvgPool2dMetatype",
-    "_param_constant115": "PTConstNoopMetatype",
-    "_param_constant116": "PTConstNoopMetatype",
     "conv2d_43": "PTConv2dMetatype",
     "relu_7": "PTRELUMetatype",
-    "_param_constant117": "PTConstNoopMetatype",
-    "_param_constant118": "PTConstNoopMetatype",
     "conv2d_44": "PTConv2dMetatype",
     "hardsigmoid_7": "PTHardSigmoidMetatype",
     "mul_7": "PTMulMetatype",
-    "_param_constant119": "PTConstNoopMetatype",
     "conv2d_45": "PTConv2dMetatype",
-    "empty_29": "UnknownMetatype",
-    "_param_constant120": "PTConstNoopMetatype",
-    "_param_constant121": "PTConstNoopMetatype",
-    "_tensor_constant58": "PTConstNoopMetatype",
-    "_tensor_constant59": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_29": "PTBatchNormMetatype",
-    "getitem_87": "PTGatherMetatype",
-    "getitem_88": "PTGatherMetatype",
-    "getitem_89": "PTGatherMetatype",
+    "batch_norm_29": "PTBatchNormMetatype",
     "add__4": "PTAddMetatype",
-    "_param_constant122": "PTConstNoopMetatype",
     "conv2d_46": "PTConv2dMetatype",
-    "empty_30": "UnknownMetatype",
-    "_param_constant123": "PTConstNoopMetatype",
-    "_param_constant124": "PTConstNoopMetatype",
-    "_tensor_constant60": "PTConstNoopMetatype",
-    "_tensor_constant61": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_30": "PTBatchNormMetatype",
-    "getitem_90": "PTGatherMetatype",
-    "getitem_91": "PTGatherMetatype",
-    "getitem_92": "PTGatherMetatype",
+    "batch_norm_30": "PTBatchNormMetatype",
     "hardswish__15": "PTHardSwishMetatype",
-    "_param_constant125": "PTConstNoopMetatype",
     "conv2d_47": "PTDepthwiseConv2dSubtype",
-    "empty_31": "UnknownMetatype",
-    "_param_constant126": "PTConstNoopMetatype",
-    "_param_constant127": "PTConstNoopMetatype",
-    "_tensor_constant62": "PTConstNoopMetatype",
-    "_tensor_constant63": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_31": "PTBatchNormMetatype",
-    "getitem_93": "PTGatherMetatype",
-    "getitem_94": "PTGatherMetatype",
-    "getitem_95": "PTGatherMetatype",
+    "batch_norm_31": "PTBatchNormMetatype",
     "hardswish__16": "PTHardSwishMetatype",
     "adaptive_avg_pool2d_8": "PTAvgPool2dMetatype",
-    "_param_constant128": "PTConstNoopMetatype",
-    "_param_constant129": "PTConstNoopMetatype",
     "conv2d_48": "PTConv2dMetatype",
     "relu_8": "PTRELUMetatype",
-    "_param_constant130": "PTConstNoopMetatype",
-    "_param_constant131": "PTConstNoopMetatype",
     "conv2d_49": "PTConv2dMetatype",
     "hardsigmoid_8": "PTHardSigmoidMetatype",
     "mul_8": "PTMulMetatype",
-    "_param_constant132": "PTConstNoopMetatype",
     "conv2d_50": "PTConv2dMetatype",
-    "empty_32": "UnknownMetatype",
-    "_param_constant133": "PTConstNoopMetatype",
-    "_param_constant134": "PTConstNoopMetatype",
-    "_tensor_constant64": "PTConstNoopMetatype",
-    "_tensor_constant65": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_32": "PTBatchNormMetatype",
-    "getitem_96": "PTGatherMetatype",
-    "getitem_97": "PTGatherMetatype",
-    "getitem_98": "PTGatherMetatype",
+    "batch_norm_32": "PTBatchNormMetatype",
     "add__5": "PTAddMetatype",
-    "_param_constant135": "PTConstNoopMetatype",
     "conv2d_51": "PTConv2dMetatype",
-    "empty_33": "UnknownMetatype",
-    "_param_constant136": "PTConstNoopMetatype",
-    "_param_constant137": "PTConstNoopMetatype",
-    "_tensor_constant66": "PTConstNoopMetatype",
-    "_tensor_constant67": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_33": "PTBatchNormMetatype",
-    "getitem_99": "PTGatherMetatype",
-    "getitem_100": "PTGatherMetatype",
-    "getitem_101": "PTGatherMetatype",
+    "batch_norm_33": "PTBatchNormMetatype",
     "hardswish__17": "PTHardSwishMetatype",
     "adaptive_avg_pool2d_9": "PTAvgPool2dMetatype",
     "flatten": "PTReshapeMetatype",
-    "_param_constant138": "PTConstNoopMetatype",
-    "_param_constant139": "PTConstNoopMetatype",
     "linear": "PTLinearMetatype",
     "hardswish__18": "PTHardSwishMetatype",
     "dropout_": "PTDropoutMetatype",
-    "_param_constant140": "PTConstNoopMetatype",
-    "_param_constant141": "PTConstNoopMetatype",
     "linear_1": "PTLinearMetatype",
-    "output": "PTOutputNoopMetatype"
+    "output_1": "PTOutputNoopMetatype"
 }
\ No newline at end of file
diff --git a/tests/torch/data/reference_graphs/fx/reference_metatypes/resnet18.json b/tests/torch/data/reference_graphs/fx/reference_metatypes/resnet18.json
index 1ae38c3acae..9544050ce71 100644
--- a/tests/torch/data/reference_graphs/fx/reference_metatypes/resnet18.json
+++ b/tests/torch/data/reference_graphs/fx/reference_metatypes/resnet18.json
@@ -1,255 +1,195 @@
 {
-    "arg0_1": "PTInputNoopMetatype",
-    "_param_constant0": "PTConstNoopMetatype",
+    "conv1_weight": "PTConstNoopMetatype",
+    "bn1_weight": "PTConstNoopMetatype",
+    "bn1_bias": "PTConstNoopMetatype",
+    "layer1_0_conv1_weight": "PTConstNoopMetatype",
+    "layer1_0_bn1_weight": "PTConstNoopMetatype",
+    "layer1_0_bn1_bias": "PTConstNoopMetatype",
+    "layer1_0_conv2_weight": "PTConstNoopMetatype",
+    "layer1_0_bn2_weight": "PTConstNoopMetatype",
+    "layer1_0_bn2_bias": "PTConstNoopMetatype",
+    "layer1_1_conv1_weight": "PTConstNoopMetatype",
+    "layer1_1_bn1_weight": "PTConstNoopMetatype",
+    "layer1_1_bn1_bias": "PTConstNoopMetatype",
+    "layer1_1_conv2_weight": "PTConstNoopMetatype",
+    "layer1_1_bn2_weight": "PTConstNoopMetatype",
+    "layer1_1_bn2_bias": "PTConstNoopMetatype",
+    "layer2_0_conv1_weight": "PTConstNoopMetatype",
+    "layer2_0_bn1_weight": "PTConstNoopMetatype",
+    "layer2_0_bn1_bias": "PTConstNoopMetatype",
+    "layer2_0_conv2_weight": "PTConstNoopMetatype",
+    "layer2_0_bn2_weight": "PTConstNoopMetatype",
+    "layer2_0_bn2_bias": "PTConstNoopMetatype",
+    "layer2_0_downsample_0_weight": "PTConstNoopMetatype",
+    "layer2_0_downsample_1_weight": "PTConstNoopMetatype",
+    "layer2_0_downsample_1_bias": "PTConstNoopMetatype",
+    "layer2_1_conv1_weight": "PTConstNoopMetatype",
+    "layer2_1_bn1_weight": "PTConstNoopMetatype",
+    "layer2_1_bn1_bias": "PTConstNoopMetatype",
+    "layer2_1_conv2_weight": "PTConstNoopMetatype",
+    "layer2_1_bn2_weight": "PTConstNoopMetatype",
+    "layer2_1_bn2_bias": "PTConstNoopMetatype",
+    "layer3_0_conv1_weight": "PTConstNoopMetatype",
+    "layer3_0_bn1_weight": "PTConstNoopMetatype",
+    "layer3_0_bn1_bias": "PTConstNoopMetatype",
+    "layer3_0_conv2_weight": "PTConstNoopMetatype",
+    "layer3_0_bn2_weight": "PTConstNoopMetatype",
+    "layer3_0_bn2_bias": "PTConstNoopMetatype",
+    "layer3_0_downsample_0_weight": "PTConstNoopMetatype",
+    "layer3_0_downsample_1_weight": "PTConstNoopMetatype",
+    "layer3_0_downsample_1_bias": "PTConstNoopMetatype",
+    "layer3_1_conv1_weight": "PTConstNoopMetatype",
+    "layer3_1_bn1_weight": "PTConstNoopMetatype",
+    "layer3_1_bn1_bias": "PTConstNoopMetatype",
+    "layer3_1_conv2_weight": "PTConstNoopMetatype",
+    "layer3_1_bn2_weight": "PTConstNoopMetatype",
+    "layer3_1_bn2_bias": "PTConstNoopMetatype",
+    "layer4_0_conv1_weight": "PTConstNoopMetatype",
+    "layer4_0_bn1_weight": "PTConstNoopMetatype",
+    "layer4_0_bn1_bias": "PTConstNoopMetatype",
+    "layer4_0_conv2_weight": "PTConstNoopMetatype",
+    "layer4_0_bn2_weight": "PTConstNoopMetatype",
+    "layer4_0_bn2_bias": "PTConstNoopMetatype",
+    "layer4_0_downsample_0_weight": "PTConstNoopMetatype",
+    "layer4_0_downsample_1_weight": "PTConstNoopMetatype",
+    "layer4_0_downsample_1_bias": "PTConstNoopMetatype",
+    "layer4_1_conv1_weight": "PTConstNoopMetatype",
+    "layer4_1_bn1_weight": "PTConstNoopMetatype",
+    "layer4_1_bn1_bias": "PTConstNoopMetatype",
+    "layer4_1_conv2_weight": "PTConstNoopMetatype",
+    "layer4_1_bn2_weight": "PTConstNoopMetatype",
+    "layer4_1_bn2_bias": "PTConstNoopMetatype",
+    "fc_weight": "PTConstNoopMetatype",
+    "fc_bias": "PTConstNoopMetatype",
+    "bn1_running_mean": "PTConstNoopMetatype",
+    "bn1_running_var": "PTConstNoopMetatype",
+    "bn1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer1_0_bn1_running_mean": "PTConstNoopMetatype",
+    "layer1_0_bn1_running_var": "PTConstNoopMetatype",
+    "layer1_0_bn1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer1_0_bn2_running_mean": "PTConstNoopMetatype",
+    "layer1_0_bn2_running_var": "PTConstNoopMetatype",
+    "layer1_0_bn2_num_batches_tracked": "PTConstNoopMetatype",
+    "layer1_1_bn1_running_mean": "PTConstNoopMetatype",
+    "layer1_1_bn1_running_var": "PTConstNoopMetatype",
+    "layer1_1_bn1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer1_1_bn2_running_mean": "PTConstNoopMetatype",
+    "layer1_1_bn2_running_var": "PTConstNoopMetatype",
+    "layer1_1_bn2_num_batches_tracked": "PTConstNoopMetatype",
+    "layer2_0_bn1_running_mean": "PTConstNoopMetatype",
+    "layer2_0_bn1_running_var": "PTConstNoopMetatype",
+    "layer2_0_bn1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer2_0_bn2_running_mean": "PTConstNoopMetatype",
+    "layer2_0_bn2_running_var": "PTConstNoopMetatype",
+    "layer2_0_bn2_num_batches_tracked": "PTConstNoopMetatype",
+    "layer2_0_downsample_1_running_mean": "PTConstNoopMetatype",
+    "layer2_0_downsample_1_running_var": "PTConstNoopMetatype",
+    "layer2_0_downsample_1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer2_1_bn1_running_mean": "PTConstNoopMetatype",
+    "layer2_1_bn1_running_var": "PTConstNoopMetatype",
+    "layer2_1_bn1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer2_1_bn2_running_mean": "PTConstNoopMetatype",
+    "layer2_1_bn2_running_var": "PTConstNoopMetatype",
+    "layer2_1_bn2_num_batches_tracked": "PTConstNoopMetatype",
+    "layer3_0_bn1_running_mean": "PTConstNoopMetatype",
+    "layer3_0_bn1_running_var": "PTConstNoopMetatype",
+    "layer3_0_bn1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer3_0_bn2_running_mean": "PTConstNoopMetatype",
+    "layer3_0_bn2_running_var": "PTConstNoopMetatype",
+    "layer3_0_bn2_num_batches_tracked": "PTConstNoopMetatype",
+    "layer3_0_downsample_1_running_mean": "PTConstNoopMetatype",
+    "layer3_0_downsample_1_running_var": "PTConstNoopMetatype",
+    "layer3_0_downsample_1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer3_1_bn1_running_mean": "PTConstNoopMetatype",
+    "layer3_1_bn1_running_var": "PTConstNoopMetatype",
+    "layer3_1_bn1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer3_1_bn2_running_mean": "PTConstNoopMetatype",
+    "layer3_1_bn2_running_var": "PTConstNoopMetatype",
+    "layer3_1_bn2_num_batches_tracked": "PTConstNoopMetatype",
+    "layer4_0_bn1_running_mean": "PTConstNoopMetatype",
+    "layer4_0_bn1_running_var": "PTConstNoopMetatype",
+    "layer4_0_bn1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer4_0_bn2_running_mean": "PTConstNoopMetatype",
+    "layer4_0_bn2_running_var": "PTConstNoopMetatype",
+    "layer4_0_bn2_num_batches_tracked": "PTConstNoopMetatype",
+    "layer4_0_downsample_1_running_mean": "PTConstNoopMetatype",
+    "layer4_0_downsample_1_running_var": "PTConstNoopMetatype",
+    "layer4_0_downsample_1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer4_1_bn1_running_mean": "PTConstNoopMetatype",
+    "layer4_1_bn1_running_var": "PTConstNoopMetatype",
+    "layer4_1_bn1_num_batches_tracked": "PTConstNoopMetatype",
+    "layer4_1_bn2_running_mean": "PTConstNoopMetatype",
+    "layer4_1_bn2_running_var": "PTConstNoopMetatype",
+    "layer4_1_bn2_num_batches_tracked": "PTConstNoopMetatype",
+    "x": "PTInputNoopMetatype",
     "conv2d": "PTConv2dMetatype",
-    "empty": "UnknownMetatype",
-    "_param_constant1": "PTConstNoopMetatype",
-    "_param_constant2": "PTConstNoopMetatype",
-    "_tensor_constant0": "PTConstNoopMetatype",
-    "_tensor_constant1": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training": "PTBatchNormMetatype",
-    "getitem": "PTGatherMetatype",
-    "getitem_1": "PTGatherMetatype",
-    "getitem_2": "PTGatherMetatype",
+    "batch_norm": "PTBatchNormMetatype",
     "relu_": "PTRELUMetatype",
     "max_pool2d": "PTMaxPool2dMetatype",
-    "_param_constant3": "PTConstNoopMetatype",
     "conv2d_1": "PTConv2dMetatype",
-    "empty_1": "UnknownMetatype",
-    "_param_constant4": "PTConstNoopMetatype",
-    "_param_constant5": "PTConstNoopMetatype",
-    "_tensor_constant2": "PTConstNoopMetatype",
-    "_tensor_constant3": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_1": "PTBatchNormMetatype",
-    "getitem_3": "PTGatherMetatype",
-    "getitem_4": "PTGatherMetatype",
-    "getitem_5": "PTGatherMetatype",
+    "batch_norm_1": "PTBatchNormMetatype",
     "relu__1": "PTRELUMetatype",
-    "_param_constant6": "PTConstNoopMetatype",
     "conv2d_2": "PTConv2dMetatype",
-    "empty_2": "UnknownMetatype",
-    "_param_constant7": "PTConstNoopMetatype",
-    "_param_constant8": "PTConstNoopMetatype",
-    "_tensor_constant4": "PTConstNoopMetatype",
-    "_tensor_constant5": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_2": "PTBatchNormMetatype",
-    "getitem_6": "PTGatherMetatype",
-    "getitem_7": "PTGatherMetatype",
-    "getitem_8": "PTGatherMetatype",
+    "batch_norm_2": "PTBatchNormMetatype",
     "add_": "PTAddMetatype",
     "relu__2": "PTRELUMetatype",
-    "_param_constant9": "PTConstNoopMetatype",
     "conv2d_3": "PTConv2dMetatype",
-    "empty_3": "UnknownMetatype",
-    "_param_constant10": "PTConstNoopMetatype",
-    "_param_constant11": "PTConstNoopMetatype",
-    "_tensor_constant6": "PTConstNoopMetatype",
-    "_tensor_constant7": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_3": "PTBatchNormMetatype",
-    "getitem_9": "PTGatherMetatype",
-    "getitem_10": "PTGatherMetatype",
-    "getitem_11": "PTGatherMetatype",
+    "batch_norm_3": "PTBatchNormMetatype",
     "relu__3": "PTRELUMetatype",
-    "_param_constant12": "PTConstNoopMetatype",
     "conv2d_4": "PTConv2dMetatype",
-    "empty_4": "UnknownMetatype",
-    "_param_constant13": "PTConstNoopMetatype",
-    "_param_constant14": "PTConstNoopMetatype",
-    "_tensor_constant8": "PTConstNoopMetatype",
-    "_tensor_constant9": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_4": "PTBatchNormMetatype",
-    "getitem_12": "PTGatherMetatype",
-    "getitem_13": "PTGatherMetatype",
-    "getitem_14": "PTGatherMetatype",
+    "batch_norm_4": "PTBatchNormMetatype",
     "add__1": "PTAddMetatype",
     "relu__4": "PTRELUMetatype",
-    "_param_constant15": "PTConstNoopMetatype",
     "conv2d_5": "PTConv2dMetatype",
-    "empty_5": "UnknownMetatype",
-    "_param_constant16": "PTConstNoopMetatype",
-    "_param_constant17": "PTConstNoopMetatype",
-    "_tensor_constant10": "PTConstNoopMetatype",
-    "_tensor_constant11": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_5": "PTBatchNormMetatype",
-    "getitem_15": "PTGatherMetatype",
-    "getitem_16": "PTGatherMetatype",
-    "getitem_17": "PTGatherMetatype",
+    "batch_norm_5": "PTBatchNormMetatype",
     "relu__5": "PTRELUMetatype",
-    "_param_constant18": "PTConstNoopMetatype",
     "conv2d_6": "PTConv2dMetatype",
-    "empty_6": "UnknownMetatype",
-    "_param_constant19": "PTConstNoopMetatype",
-    "_param_constant20": "PTConstNoopMetatype",
-    "_tensor_constant12": "PTConstNoopMetatype",
-    "_tensor_constant13": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_6": "PTBatchNormMetatype",
-    "getitem_18": "PTGatherMetatype",
-    "getitem_19": "PTGatherMetatype",
-    "getitem_20": "PTGatherMetatype",
-    "_param_constant21": "PTConstNoopMetatype",
+    "batch_norm_6": "PTBatchNormMetatype",
     "conv2d_7": "PTConv2dMetatype",
-    "empty_7": "UnknownMetatype",
-    "_param_constant22": "PTConstNoopMetatype",
-    "_param_constant23": "PTConstNoopMetatype",
-    "_tensor_constant14": "PTConstNoopMetatype",
-    "_tensor_constant15": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_7": "PTBatchNormMetatype",
-    "getitem_21": "PTGatherMetatype",
-    "getitem_22": "PTGatherMetatype",
-    "getitem_23": "PTGatherMetatype",
+    "batch_norm_7": "PTBatchNormMetatype",
     "add__2": "PTAddMetatype",
     "relu__6": "PTRELUMetatype",
-    "_param_constant24": "PTConstNoopMetatype",
     "conv2d_8": "PTConv2dMetatype",
-    "empty_8": "UnknownMetatype",
-    "_param_constant25": "PTConstNoopMetatype",
-    "_param_constant26": "PTConstNoopMetatype",
-    "_tensor_constant16": "PTConstNoopMetatype",
-    "_tensor_constant17": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_8": "PTBatchNormMetatype",
-    "getitem_24": "PTGatherMetatype",
-    "getitem_25": "PTGatherMetatype",
-    "getitem_26": "PTGatherMetatype",
+    "batch_norm_8": "PTBatchNormMetatype",
     "relu__7": "PTRELUMetatype",
-    "_param_constant27": "PTConstNoopMetatype",
     "conv2d_9": "PTConv2dMetatype",
-    "empty_9": "UnknownMetatype",
-    "_param_constant28": "PTConstNoopMetatype",
-    "_param_constant29": "PTConstNoopMetatype",
-    "_tensor_constant18": "PTConstNoopMetatype",
-    "_tensor_constant19": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_9": "PTBatchNormMetatype",
-    "getitem_27": "PTGatherMetatype",
-    "getitem_28": "PTGatherMetatype",
-    "getitem_29": "PTGatherMetatype",
+    "batch_norm_9": "PTBatchNormMetatype",
     "add__3": "PTAddMetatype",
     "relu__8": "PTRELUMetatype",
-    "_param_constant30": "PTConstNoopMetatype",
     "conv2d_10": "PTConv2dMetatype",
-    "empty_10": "UnknownMetatype",
-    "_param_constant31": "PTConstNoopMetatype",
-    "_param_constant32": "PTConstNoopMetatype",
-    "_tensor_constant20": "PTConstNoopMetatype",
-    "_tensor_constant21": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_10": "PTBatchNormMetatype",
-    "getitem_30": "PTGatherMetatype",
-    "getitem_31": "PTGatherMetatype",
-    "getitem_32": "PTGatherMetatype",
+    "batch_norm_10": "PTBatchNormMetatype",
     "relu__9": "PTRELUMetatype",
-    "_param_constant33": "PTConstNoopMetatype",
     "conv2d_11": "PTConv2dMetatype",
-    "empty_11": "UnknownMetatype",
-    "_param_constant34": "PTConstNoopMetatype",
-    "_param_constant35": "PTConstNoopMetatype",
-    "_tensor_constant22": "PTConstNoopMetatype",
-    "_tensor_constant23": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_11": "PTBatchNormMetatype",
-    "getitem_33": "PTGatherMetatype",
-    "getitem_34": "PTGatherMetatype",
-    "getitem_35": "PTGatherMetatype",
-    "_param_constant36": "PTConstNoopMetatype",
+    "batch_norm_11": "PTBatchNormMetatype",
     "conv2d_12": "PTConv2dMetatype",
-    "empty_12": "UnknownMetatype",
-    "_param_constant37": "PTConstNoopMetatype",
-    "_param_constant38": "PTConstNoopMetatype",
-    "_tensor_constant24": "PTConstNoopMetatype",
-    "_tensor_constant25": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_12": "PTBatchNormMetatype",
-    "getitem_36": "PTGatherMetatype",
-    "getitem_37": "PTGatherMetatype",
-    "getitem_38": "PTGatherMetatype",
+    "batch_norm_12": "PTBatchNormMetatype",
     "add__4": "PTAddMetatype",
     "relu__10": "PTRELUMetatype",
-    "_param_constant39": "PTConstNoopMetatype",
     "conv2d_13": "PTConv2dMetatype",
-    "empty_13": "UnknownMetatype",
-    "_param_constant40": "PTConstNoopMetatype",
-    "_param_constant41": "PTConstNoopMetatype",
-    "_tensor_constant26": "PTConstNoopMetatype",
-    "_tensor_constant27": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_13": "PTBatchNormMetatype",
-    "getitem_39": "PTGatherMetatype",
-    "getitem_40": "PTGatherMetatype",
-    "getitem_41": "PTGatherMetatype",
+    "batch_norm_13": "PTBatchNormMetatype",
     "relu__11": "PTRELUMetatype",
-    "_param_constant42": "PTConstNoopMetatype",
     "conv2d_14": "PTConv2dMetatype",
-    "empty_14": "UnknownMetatype",
-    "_param_constant43": "PTConstNoopMetatype",
-    "_param_constant44": "PTConstNoopMetatype",
-    "_tensor_constant28": "PTConstNoopMetatype",
-    "_tensor_constant29": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_14": "PTBatchNormMetatype",
-    "getitem_42": "PTGatherMetatype",
-    "getitem_43": "PTGatherMetatype",
-    "getitem_44": "PTGatherMetatype",
+    "batch_norm_14": "PTBatchNormMetatype",
     "add__5": "PTAddMetatype",
     "relu__12": "PTRELUMetatype",
-    "_param_constant45": "PTConstNoopMetatype",
     "conv2d_15": "PTConv2dMetatype",
-    "empty_15": "UnknownMetatype",
-    "_param_constant46": "PTConstNoopMetatype",
-    "_param_constant47": "PTConstNoopMetatype",
-    "_tensor_constant30": "PTConstNoopMetatype",
-    "_tensor_constant31": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_15": "PTBatchNormMetatype",
-    "getitem_45": "PTGatherMetatype",
-    "getitem_46": "PTGatherMetatype",
-    "getitem_47": "PTGatherMetatype",
+    "batch_norm_15": "PTBatchNormMetatype",
     "relu__13": "PTRELUMetatype",
-    "_param_constant48": "PTConstNoopMetatype",
     "conv2d_16": "PTConv2dMetatype",
-    "empty_16": "UnknownMetatype",
-    "_param_constant49": "PTConstNoopMetatype",
-    "_param_constant50": "PTConstNoopMetatype",
-    "_tensor_constant32": "PTConstNoopMetatype",
-    "_tensor_constant33": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_16": "PTBatchNormMetatype",
-    "getitem_48": "PTGatherMetatype",
-    "getitem_49": "PTGatherMetatype",
-    "getitem_50": "PTGatherMetatype",
-    "_param_constant51": "PTConstNoopMetatype",
+    "batch_norm_16": "PTBatchNormMetatype",
     "conv2d_17": "PTConv2dMetatype",
-    "empty_17": "UnknownMetatype",
-    "_param_constant52": "PTConstNoopMetatype",
-    "_param_constant53": "PTConstNoopMetatype",
-    "_tensor_constant34": "PTConstNoopMetatype",
-    "_tensor_constant35": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_17": "PTBatchNormMetatype",
-    "getitem_51": "PTGatherMetatype",
-    "getitem_52": "PTGatherMetatype",
-    "getitem_53": "PTGatherMetatype",
+    "batch_norm_17": "PTBatchNormMetatype",
     "add__6": "PTAddMetatype",
     "relu__14": "PTRELUMetatype",
-    "_param_constant54": "PTConstNoopMetatype",
     "conv2d_18": "PTConv2dMetatype",
-    "empty_18": "UnknownMetatype",
-    "_param_constant55": "PTConstNoopMetatype",
-    "_param_constant56": "PTConstNoopMetatype",
-    "_tensor_constant36": "PTConstNoopMetatype",
-    "_tensor_constant37": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_18": "PTBatchNormMetatype",
-    "getitem_54": "PTGatherMetatype",
-    "getitem_55": "PTGatherMetatype",
-    "getitem_56": "PTGatherMetatype",
+    "batch_norm_18": "PTBatchNormMetatype",
     "relu__15": "PTRELUMetatype",
-    "_param_constant57": "PTConstNoopMetatype",
     "conv2d_19": "PTConv2dMetatype",
-    "empty_19": "UnknownMetatype",
-    "_param_constant58": "PTConstNoopMetatype",
-    "_param_constant59": "PTConstNoopMetatype",
-    "_tensor_constant38": "PTConstNoopMetatype",
-    "_tensor_constant39": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_19": "PTBatchNormMetatype",
-    "getitem_57": "PTGatherMetatype",
-    "getitem_58": "PTGatherMetatype",
-    "getitem_59": "PTGatherMetatype",
+    "batch_norm_19": "PTBatchNormMetatype",
     "add__7": "PTAddMetatype",
     "relu__16": "PTRELUMetatype",
     "adaptive_avg_pool2d": "PTAvgPool2dMetatype",
     "flatten": "PTReshapeMetatype",
-    "_param_constant60": "PTConstNoopMetatype",
-    "_param_constant61": "PTConstNoopMetatype",
     "linear": "PTLinearMetatype",
-    "output": "PTOutputNoopMetatype"
+    "output_1": "PTOutputNoopMetatype"
 }
\ No newline at end of file
diff --git a/tests/torch/data/reference_graphs/fx/reference_metatypes/swin_v2_s.json b/tests/torch/data/reference_graphs/fx/reference_metatypes/swin_v2_s.json
index 6f038197c8e..b4788d60229 100644
--- a/tests/torch/data/reference_graphs/fx/reference_metatypes/swin_v2_s.json
+++ b/tests/torch/data/reference_graphs/fx/reference_metatypes/swin_v2_s.json
@@ -1,21 +1,560 @@
 {
-    "arg0_1": "PTInputNoopMetatype",
-    "_param_constant0": "PTConstNoopMetatype",
-    "_param_constant1": "PTConstNoopMetatype",
+    "features_1_0_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_1_0_attn_proj_weight": "PTConstNoopMetatype",
+    "features_1_0_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_1_0_attn_proj_bias": "PTConstNoopMetatype",
+    "features_1_0_attn_logit_scale": "PTConstNoopMetatype",
+    "features_1_1_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_1_1_attn_proj_weight": "PTConstNoopMetatype",
+    "features_1_1_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_1_1_attn_proj_bias": "PTConstNoopMetatype",
+    "features_1_1_attn_logit_scale": "PTConstNoopMetatype",
+    "features_3_0_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_3_0_attn_proj_weight": "PTConstNoopMetatype",
+    "features_3_0_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_3_0_attn_proj_bias": "PTConstNoopMetatype",
+    "features_3_0_attn_logit_scale": "PTConstNoopMetatype",
+    "features_3_1_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_3_1_attn_proj_weight": "PTConstNoopMetatype",
+    "features_3_1_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_3_1_attn_proj_bias": "PTConstNoopMetatype",
+    "features_3_1_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_0_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_0_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_0_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_0_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_0_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_1_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_1_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_1_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_1_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_1_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_2_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_2_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_2_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_2_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_2_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_3_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_3_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_3_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_3_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_3_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_4_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_4_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_4_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_4_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_4_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_5_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_5_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_5_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_5_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_5_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_6_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_6_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_6_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_6_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_6_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_7_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_7_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_7_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_7_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_7_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_8_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_8_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_8_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_8_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_8_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_9_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_9_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_9_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_9_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_9_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_10_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_10_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_10_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_10_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_10_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_11_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_11_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_11_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_11_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_11_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_12_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_12_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_12_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_12_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_12_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_13_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_13_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_13_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_13_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_13_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_14_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_14_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_14_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_14_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_14_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_15_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_15_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_15_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_15_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_15_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_16_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_16_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_16_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_16_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_16_attn_logit_scale": "PTConstNoopMetatype",
+    "features_5_17_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_5_17_attn_proj_weight": "PTConstNoopMetatype",
+    "features_5_17_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_5_17_attn_proj_bias": "PTConstNoopMetatype",
+    "features_5_17_attn_logit_scale": "PTConstNoopMetatype",
+    "features_7_0_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_7_0_attn_proj_weight": "PTConstNoopMetatype",
+    "features_7_0_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_7_0_attn_proj_bias": "PTConstNoopMetatype",
+    "features_7_0_attn_logit_scale": "PTConstNoopMetatype",
+    "features_7_1_attn_qkv_weight": "PTConstNoopMetatype",
+    "features_7_1_attn_proj_weight": "PTConstNoopMetatype",
+    "features_7_1_attn_qkv_bias": "PTConstNoopMetatype",
+    "features_7_1_attn_proj_bias": "PTConstNoopMetatype",
+    "features_7_1_attn_logit_scale": "PTConstNoopMetatype",
+    "features_0_0_weight": "PTConstNoopMetatype",
+    "features_0_0_bias": "PTConstNoopMetatype",
+    "features_0_2_weight": "PTConstNoopMetatype",
+    "features_0_2_bias": "PTConstNoopMetatype",
+    "features_1_0_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_1_0_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_1_0_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_1_0_norm1_weight": "PTConstNoopMetatype",
+    "features_1_0_norm1_bias": "PTConstNoopMetatype",
+    "features_1_0_mlp_0_weight": "PTConstNoopMetatype",
+    "features_1_0_mlp_0_bias": "PTConstNoopMetatype",
+    "features_1_0_mlp_3_weight": "PTConstNoopMetatype",
+    "features_1_0_mlp_3_bias": "PTConstNoopMetatype",
+    "features_1_0_norm2_weight": "PTConstNoopMetatype",
+    "features_1_0_norm2_bias": "PTConstNoopMetatype",
+    "features_1_1_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_1_1_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_1_1_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_1_1_norm1_weight": "PTConstNoopMetatype",
+    "features_1_1_norm1_bias": "PTConstNoopMetatype",
+    "features_1_1_mlp_0_weight": "PTConstNoopMetatype",
+    "features_1_1_mlp_0_bias": "PTConstNoopMetatype",
+    "features_1_1_mlp_3_weight": "PTConstNoopMetatype",
+    "features_1_1_mlp_3_bias": "PTConstNoopMetatype",
+    "features_1_1_norm2_weight": "PTConstNoopMetatype",
+    "features_1_1_norm2_bias": "PTConstNoopMetatype",
+    "features_2_reduction_weight": "PTConstNoopMetatype",
+    "features_2_norm_weight": "PTConstNoopMetatype",
+    "features_2_norm_bias": "PTConstNoopMetatype",
+    "features_3_0_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_3_0_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_3_0_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_3_0_norm1_weight": "PTConstNoopMetatype",
+    "features_3_0_norm1_bias": "PTConstNoopMetatype",
+    "features_3_0_mlp_0_weight": "PTConstNoopMetatype",
+    "features_3_0_mlp_0_bias": "PTConstNoopMetatype",
+    "features_3_0_mlp_3_weight": "PTConstNoopMetatype",
+    "features_3_0_mlp_3_bias": "PTConstNoopMetatype",
+    "features_3_0_norm2_weight": "PTConstNoopMetatype",
+    "features_3_0_norm2_bias": "PTConstNoopMetatype",
+    "features_3_1_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_3_1_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_3_1_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_3_1_norm1_weight": "PTConstNoopMetatype",
+    "features_3_1_norm1_bias": "PTConstNoopMetatype",
+    "features_3_1_mlp_0_weight": "PTConstNoopMetatype",
+    "features_3_1_mlp_0_bias": "PTConstNoopMetatype",
+    "features_3_1_mlp_3_weight": "PTConstNoopMetatype",
+    "features_3_1_mlp_3_bias": "PTConstNoopMetatype",
+    "features_3_1_norm2_weight": "PTConstNoopMetatype",
+    "features_3_1_norm2_bias": "PTConstNoopMetatype",
+    "features_4_reduction_weight": "PTConstNoopMetatype",
+    "features_4_norm_weight": "PTConstNoopMetatype",
+    "features_4_norm_bias": "PTConstNoopMetatype",
+    "features_5_0_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_0_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_0_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_0_norm1_weight": "PTConstNoopMetatype",
+    "features_5_0_norm1_bias": "PTConstNoopMetatype",
+    "features_5_0_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_0_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_0_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_0_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_0_norm2_weight": "PTConstNoopMetatype",
+    "features_5_0_norm2_bias": "PTConstNoopMetatype",
+    "features_5_1_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_1_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_1_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_1_norm1_weight": "PTConstNoopMetatype",
+    "features_5_1_norm1_bias": "PTConstNoopMetatype",
+    "features_5_1_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_1_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_1_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_1_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_1_norm2_weight": "PTConstNoopMetatype",
+    "features_5_1_norm2_bias": "PTConstNoopMetatype",
+    "features_5_2_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_2_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_2_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_2_norm1_weight": "PTConstNoopMetatype",
+    "features_5_2_norm1_bias": "PTConstNoopMetatype",
+    "features_5_2_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_2_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_2_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_2_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_2_norm2_weight": "PTConstNoopMetatype",
+    "features_5_2_norm2_bias": "PTConstNoopMetatype",
+    "features_5_3_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_3_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_3_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_3_norm1_weight": "PTConstNoopMetatype",
+    "features_5_3_norm1_bias": "PTConstNoopMetatype",
+    "features_5_3_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_3_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_3_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_3_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_3_norm2_weight": "PTConstNoopMetatype",
+    "features_5_3_norm2_bias": "PTConstNoopMetatype",
+    "features_5_4_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_4_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_4_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_4_norm1_weight": "PTConstNoopMetatype",
+    "features_5_4_norm1_bias": "PTConstNoopMetatype",
+    "features_5_4_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_4_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_4_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_4_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_4_norm2_weight": "PTConstNoopMetatype",
+    "features_5_4_norm2_bias": "PTConstNoopMetatype",
+    "features_5_5_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_5_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_5_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_5_norm1_weight": "PTConstNoopMetatype",
+    "features_5_5_norm1_bias": "PTConstNoopMetatype",
+    "features_5_5_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_5_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_5_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_5_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_5_norm2_weight": "PTConstNoopMetatype",
+    "features_5_5_norm2_bias": "PTConstNoopMetatype",
+    "features_5_6_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_6_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_6_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_6_norm1_weight": "PTConstNoopMetatype",
+    "features_5_6_norm1_bias": "PTConstNoopMetatype",
+    "features_5_6_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_6_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_6_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_6_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_6_norm2_weight": "PTConstNoopMetatype",
+    "features_5_6_norm2_bias": "PTConstNoopMetatype",
+    "features_5_7_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_7_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_7_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_7_norm1_weight": "PTConstNoopMetatype",
+    "features_5_7_norm1_bias": "PTConstNoopMetatype",
+    "features_5_7_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_7_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_7_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_7_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_7_norm2_weight": "PTConstNoopMetatype",
+    "features_5_7_norm2_bias": "PTConstNoopMetatype",
+    "features_5_8_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_8_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_8_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_8_norm1_weight": "PTConstNoopMetatype",
+    "features_5_8_norm1_bias": "PTConstNoopMetatype",
+    "features_5_8_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_8_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_8_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_8_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_8_norm2_weight": "PTConstNoopMetatype",
+    "features_5_8_norm2_bias": "PTConstNoopMetatype",
+    "features_5_9_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_9_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_9_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_9_norm1_weight": "PTConstNoopMetatype",
+    "features_5_9_norm1_bias": "PTConstNoopMetatype",
+    "features_5_9_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_9_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_9_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_9_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_9_norm2_weight": "PTConstNoopMetatype",
+    "features_5_9_norm2_bias": "PTConstNoopMetatype",
+    "features_5_10_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_10_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_10_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_10_norm1_weight": "PTConstNoopMetatype",
+    "features_5_10_norm1_bias": "PTConstNoopMetatype",
+    "features_5_10_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_10_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_10_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_10_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_10_norm2_weight": "PTConstNoopMetatype",
+    "features_5_10_norm2_bias": "PTConstNoopMetatype",
+    "features_5_11_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_11_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_11_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_11_norm1_weight": "PTConstNoopMetatype",
+    "features_5_11_norm1_bias": "PTConstNoopMetatype",
+    "features_5_11_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_11_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_11_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_11_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_11_norm2_weight": "PTConstNoopMetatype",
+    "features_5_11_norm2_bias": "PTConstNoopMetatype",
+    "features_5_12_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_12_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_12_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_12_norm1_weight": "PTConstNoopMetatype",
+    "features_5_12_norm1_bias": "PTConstNoopMetatype",
+    "features_5_12_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_12_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_12_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_12_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_12_norm2_weight": "PTConstNoopMetatype",
+    "features_5_12_norm2_bias": "PTConstNoopMetatype",
+    "features_5_13_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_13_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_13_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_13_norm1_weight": "PTConstNoopMetatype",
+    "features_5_13_norm1_bias": "PTConstNoopMetatype",
+    "features_5_13_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_13_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_13_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_13_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_13_norm2_weight": "PTConstNoopMetatype",
+    "features_5_13_norm2_bias": "PTConstNoopMetatype",
+    "features_5_14_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_14_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_14_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_14_norm1_weight": "PTConstNoopMetatype",
+    "features_5_14_norm1_bias": "PTConstNoopMetatype",
+    "features_5_14_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_14_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_14_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_14_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_14_norm2_weight": "PTConstNoopMetatype",
+    "features_5_14_norm2_bias": "PTConstNoopMetatype",
+    "features_5_15_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_15_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_15_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_15_norm1_weight": "PTConstNoopMetatype",
+    "features_5_15_norm1_bias": "PTConstNoopMetatype",
+    "features_5_15_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_15_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_15_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_15_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_15_norm2_weight": "PTConstNoopMetatype",
+    "features_5_15_norm2_bias": "PTConstNoopMetatype",
+    "features_5_16_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_16_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_16_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_16_norm1_weight": "PTConstNoopMetatype",
+    "features_5_16_norm1_bias": "PTConstNoopMetatype",
+    "features_5_16_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_16_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_16_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_16_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_16_norm2_weight": "PTConstNoopMetatype",
+    "features_5_16_norm2_bias": "PTConstNoopMetatype",
+    "features_5_17_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_17_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_17_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_5_17_norm1_weight": "PTConstNoopMetatype",
+    "features_5_17_norm1_bias": "PTConstNoopMetatype",
+    "features_5_17_mlp_0_weight": "PTConstNoopMetatype",
+    "features_5_17_mlp_0_bias": "PTConstNoopMetatype",
+    "features_5_17_mlp_3_weight": "PTConstNoopMetatype",
+    "features_5_17_mlp_3_bias": "PTConstNoopMetatype",
+    "features_5_17_norm2_weight": "PTConstNoopMetatype",
+    "features_5_17_norm2_bias": "PTConstNoopMetatype",
+    "features_6_reduction_weight": "PTConstNoopMetatype",
+    "features_6_norm_weight": "PTConstNoopMetatype",
+    "features_6_norm_bias": "PTConstNoopMetatype",
+    "features_7_0_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_7_0_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_7_0_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_7_0_norm1_weight": "PTConstNoopMetatype",
+    "features_7_0_norm1_bias": "PTConstNoopMetatype",
+    "features_7_0_mlp_0_weight": "PTConstNoopMetatype",
+    "features_7_0_mlp_0_bias": "PTConstNoopMetatype",
+    "features_7_0_mlp_3_weight": "PTConstNoopMetatype",
+    "features_7_0_mlp_3_bias": "PTConstNoopMetatype",
+    "features_7_0_norm2_weight": "PTConstNoopMetatype",
+    "features_7_0_norm2_bias": "PTConstNoopMetatype",
+    "features_7_1_attn_cpb_mlp_0_weight": "PTConstNoopMetatype",
+    "features_7_1_attn_cpb_mlp_0_bias": "PTConstNoopMetatype",
+    "features_7_1_attn_cpb_mlp_2_weight": "PTConstNoopMetatype",
+    "features_7_1_norm1_weight": "PTConstNoopMetatype",
+    "features_7_1_norm1_bias": "PTConstNoopMetatype",
+    "features_7_1_mlp_0_weight": "PTConstNoopMetatype",
+    "features_7_1_mlp_0_bias": "PTConstNoopMetatype",
+    "features_7_1_mlp_3_weight": "PTConstNoopMetatype",
+    "features_7_1_mlp_3_bias": "PTConstNoopMetatype",
+    "features_7_1_norm2_weight": "PTConstNoopMetatype",
+    "features_7_1_norm2_bias": "PTConstNoopMetatype",
+    "norm_weight": "PTConstNoopMetatype",
+    "norm_bias": "PTConstNoopMetatype",
+    "head_weight": "PTConstNoopMetatype",
+    "head_bias": "PTConstNoopMetatype",
+    "features_1_0_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_1_0_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_1_1_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_1_1_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_3_0_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_3_0_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_3_1_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_3_1_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_0_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_0_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_1_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_1_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_2_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_2_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_3_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_3_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_4_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_4_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_5_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_5_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_6_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_6_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_7_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_7_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_8_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_8_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_9_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_9_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_10_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_10_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_11_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_11_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_12_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_12_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_13_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_13_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_14_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_14_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_15_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_15_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_16_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_16_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_5_17_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_5_17_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_7_0_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_7_0_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_7_1_attn_relative_coords_table": "PTConstNoopMetatype",
+    "features_7_1_attn_relative_position_index": "PTConstNoopMetatype",
+    "features_1_1_attn_lifted_tensor_0": "PTConstNoopMetatype",
+    "features_1_1_attn_lifted_tensor_1": "PTConstNoopMetatype",
+    "features_1_1_attn_lifted_tensor_2": "PTConstNoopMetatype",
+    "features_1_1_attn_lifted_tensor_3": "PTConstNoopMetatype",
+    "features_1_1_attn_lifted_tensor_4": "PTConstNoopMetatype",
+    "features_1_1_attn_lifted_tensor_5": "PTConstNoopMetatype",
+    "features_1_1_attn_lifted_tensor_6": "PTConstNoopMetatype",
+    "features_1_1_attn_lifted_tensor_7": "PTConstNoopMetatype",
+    "features_1_1_attn_lifted_tensor_8": "PTConstNoopMetatype",
+    "features_3_1_attn_lifted_tensor_9": "PTConstNoopMetatype",
+    "features_3_1_attn_lifted_tensor_10": "PTConstNoopMetatype",
+    "features_3_1_attn_lifted_tensor_11": "PTConstNoopMetatype",
+    "features_3_1_attn_lifted_tensor_12": "PTConstNoopMetatype",
+    "features_3_1_attn_lifted_tensor_13": "PTConstNoopMetatype",
+    "features_3_1_attn_lifted_tensor_14": "PTConstNoopMetatype",
+    "features_3_1_attn_lifted_tensor_15": "PTConstNoopMetatype",
+    "features_3_1_attn_lifted_tensor_16": "PTConstNoopMetatype",
+    "features_3_1_attn_lifted_tensor_17": "PTConstNoopMetatype",
+    "features_5_1_attn_lifted_tensor_18": "PTConstNoopMetatype",
+    "features_5_1_attn_lifted_tensor_19": "PTConstNoopMetatype",
+    "features_5_1_attn_lifted_tensor_20": "PTConstNoopMetatype",
+    "features_5_1_attn_lifted_tensor_21": "PTConstNoopMetatype",
+    "features_5_1_attn_lifted_tensor_22": "PTConstNoopMetatype",
+    "features_5_1_attn_lifted_tensor_23": "PTConstNoopMetatype",
+    "features_5_1_attn_lifted_tensor_24": "PTConstNoopMetatype",
+    "features_5_1_attn_lifted_tensor_25": "PTConstNoopMetatype",
+    "features_5_1_attn_lifted_tensor_26": "PTConstNoopMetatype",
+    "features_5_3_attn_lifted_tensor_27": "PTConstNoopMetatype",
+    "features_5_3_attn_lifted_tensor_28": "PTConstNoopMetatype",
+    "features_5_3_attn_lifted_tensor_29": "PTConstNoopMetatype",
+    "features_5_3_attn_lifted_tensor_30": "PTConstNoopMetatype",
+    "features_5_3_attn_lifted_tensor_31": "PTConstNoopMetatype",
+    "features_5_3_attn_lifted_tensor_32": "PTConstNoopMetatype",
+    "features_5_3_attn_lifted_tensor_33": "PTConstNoopMetatype",
+    "features_5_3_attn_lifted_tensor_34": "PTConstNoopMetatype",
+    "features_5_3_attn_lifted_tensor_35": "PTConstNoopMetatype",
+    "features_5_5_attn_lifted_tensor_36": "PTConstNoopMetatype",
+    "features_5_5_attn_lifted_tensor_37": "PTConstNoopMetatype",
+    "features_5_5_attn_lifted_tensor_38": "PTConstNoopMetatype",
+    "features_5_5_attn_lifted_tensor_39": "PTConstNoopMetatype",
+    "features_5_5_attn_lifted_tensor_40": "PTConstNoopMetatype",
+    "features_5_5_attn_lifted_tensor_41": "PTConstNoopMetatype",
+    "features_5_5_attn_lifted_tensor_42": "PTConstNoopMetatype",
+    "features_5_5_attn_lifted_tensor_43": "PTConstNoopMetatype",
+    "features_5_5_attn_lifted_tensor_44": "PTConstNoopMetatype",
+    "features_5_7_attn_lifted_tensor_45": "PTConstNoopMetatype",
+    "features_5_7_attn_lifted_tensor_46": "PTConstNoopMetatype",
+    "features_5_7_attn_lifted_tensor_47": "PTConstNoopMetatype",
+    "features_5_7_attn_lifted_tensor_48": "PTConstNoopMetatype",
+    "features_5_7_attn_lifted_tensor_49": "PTConstNoopMetatype",
+    "features_5_7_attn_lifted_tensor_50": "PTConstNoopMetatype",
+    "features_5_7_attn_lifted_tensor_51": "PTConstNoopMetatype",
+    "features_5_7_attn_lifted_tensor_52": "PTConstNoopMetatype",
+    "features_5_7_attn_lifted_tensor_53": "PTConstNoopMetatype",
+    "features_5_9_attn_lifted_tensor_54": "PTConstNoopMetatype",
+    "features_5_9_attn_lifted_tensor_55": "PTConstNoopMetatype",
+    "features_5_9_attn_lifted_tensor_56": "PTConstNoopMetatype",
+    "features_5_9_attn_lifted_tensor_57": "PTConstNoopMetatype",
+    "features_5_9_attn_lifted_tensor_58": "PTConstNoopMetatype",
+    "features_5_9_attn_lifted_tensor_59": "PTConstNoopMetatype",
+    "features_5_9_attn_lifted_tensor_60": "PTConstNoopMetatype",
+    "features_5_9_attn_lifted_tensor_61": "PTConstNoopMetatype",
+    "features_5_9_attn_lifted_tensor_62": "PTConstNoopMetatype",
+    "features_5_11_attn_lifted_tensor_63": "PTConstNoopMetatype",
+    "features_5_11_attn_lifted_tensor_64": "PTConstNoopMetatype",
+    "features_5_11_attn_lifted_tensor_65": "PTConstNoopMetatype",
+    "features_5_11_attn_lifted_tensor_66": "PTConstNoopMetatype",
+    "features_5_11_attn_lifted_tensor_67": "PTConstNoopMetatype",
+    "features_5_11_attn_lifted_tensor_68": "PTConstNoopMetatype",
+    "features_5_11_attn_lifted_tensor_69": "PTConstNoopMetatype",
+    "features_5_11_attn_lifted_tensor_70": "PTConstNoopMetatype",
+    "features_5_11_attn_lifted_tensor_71": "PTConstNoopMetatype",
+    "features_5_13_attn_lifted_tensor_72": "PTConstNoopMetatype",
+    "features_5_13_attn_lifted_tensor_73": "PTConstNoopMetatype",
+    "features_5_13_attn_lifted_tensor_74": "PTConstNoopMetatype",
+    "features_5_13_attn_lifted_tensor_75": "PTConstNoopMetatype",
+    "features_5_13_attn_lifted_tensor_76": "PTConstNoopMetatype",
+    "features_5_13_attn_lifted_tensor_77": "PTConstNoopMetatype",
+    "features_5_13_attn_lifted_tensor_78": "PTConstNoopMetatype",
+    "features_5_13_attn_lifted_tensor_79": "PTConstNoopMetatype",
+    "features_5_13_attn_lifted_tensor_80": "PTConstNoopMetatype",
+    "features_5_15_attn_lifted_tensor_81": "PTConstNoopMetatype",
+    "features_5_15_attn_lifted_tensor_82": "PTConstNoopMetatype",
+    "features_5_15_attn_lifted_tensor_83": "PTConstNoopMetatype",
+    "features_5_15_attn_lifted_tensor_84": "PTConstNoopMetatype",
+    "features_5_15_attn_lifted_tensor_85": "PTConstNoopMetatype",
+    "features_5_15_attn_lifted_tensor_86": "PTConstNoopMetatype",
+    "features_5_15_attn_lifted_tensor_87": "PTConstNoopMetatype",
+    "features_5_15_attn_lifted_tensor_88": "PTConstNoopMetatype",
+    "features_5_15_attn_lifted_tensor_89": "PTConstNoopMetatype",
+    "features_5_17_attn_lifted_tensor_90": "PTConstNoopMetatype",
+    "features_5_17_attn_lifted_tensor_91": "PTConstNoopMetatype",
+    "features_5_17_attn_lifted_tensor_92": "PTConstNoopMetatype",
+    "features_5_17_attn_lifted_tensor_93": "PTConstNoopMetatype",
+    "features_5_17_attn_lifted_tensor_94": "PTConstNoopMetatype",
+    "features_5_17_attn_lifted_tensor_95": "PTConstNoopMetatype",
+    "features_5_17_attn_lifted_tensor_96": "PTConstNoopMetatype",
+    "features_5_17_attn_lifted_tensor_97": "PTConstNoopMetatype",
+    "features_5_17_attn_lifted_tensor_98": "PTConstNoopMetatype",
+    "x": "PTInputNoopMetatype",
     "conv2d": "PTConv2dMetatype",
     "permute": "PTTransposeMetatype",
-    "_param_constant2": "PTConstNoopMetatype",
-    "_param_constant3": "PTConstNoopMetatype",
     "layer_norm": "PTLayerNormMetatype",
-    "_tensor_constant0": "PTConstNoopMetatype",
-    "_param_constant4": "PTConstNoopMetatype",
-    "_param_constant5": "PTConstNoopMetatype",
     "linear": "PTLinearMetatype",
     "relu_": "PTRELUMetatype",
-    "_param_constant6": "PTConstNoopMetatype",
     "linear_1": "PTLinearMetatype",
     "view": "PTReshapeMetatype",
-    "_tensor_constant1": "PTConstNoopMetatype",
     "index": "UnknownMetatype",
     "view_1": "PTReshapeMetatype",
     "permute_1": "PTTransposeMetatype",
@@ -27,11 +566,9 @@
     "view_2": "PTReshapeMetatype",
     "permute_2": "PTTransposeMetatype",
     "reshape": "PTReshapeMetatype",
-    "_param_constant7": "PTConstNoopMetatype",
     "clone": "PTNoopMetatype",
     "slice_1": "PTGatherMetatype",
     "zero_": "UnknownMetatype",
-    "_param_constant8": "PTConstNoopMetatype",
     "linear_2": "PTLinearMetatype",
     "reshape_1": "PTReshapeMetatype",
     "permute_3": "PTTransposeMetatype",
@@ -48,7 +585,6 @@
     "div_1": "PTDivMetatype",
     "transpose": "PTTransposeMetatype",
     "matmul": "PTMatMulMetatype",
-    "_param_constant9": "PTConstNoopMetatype",
     "clamp": "UnknownMetatype",
     "exp": "PTExpMetatype",
     "mul_1": "PTMulMetatype",
@@ -58,8 +594,6 @@
     "matmul_1": "PTMatMulMetatype",
     "transpose_1": "PTTransposeMetatype",
     "reshape_2": "PTReshapeMetatype",
-    "_param_constant10": "PTConstNoopMetatype",
-    "_param_constant11": "PTConstNoopMetatype",
     "linear_3": "PTLinearMetatype",
     "dropout_1": "PTDropoutMetatype",
     "view_3": "PTReshapeMetatype",
@@ -67,32 +601,19 @@
     "reshape_3": "PTReshapeMetatype",
     "slice_2": "PTGatherMetatype",
     "slice_3": "PTGatherMetatype",
-    "_param_constant12": "PTConstNoopMetatype",
-    "_param_constant13": "PTConstNoopMetatype",
     "layer_norm_1": "PTLayerNormMetatype",
     "add_1": "PTAddMetatype",
-    "_param_constant14": "PTConstNoopMetatype",
-    "_param_constant15": "PTConstNoopMetatype",
     "linear_4": "PTLinearMetatype",
     "gelu": "PTGELUMetatype",
     "dropout_2": "PTDropoutMetatype",
-    "_param_constant16": "PTConstNoopMetatype",
-    "_param_constant17": "PTConstNoopMetatype",
     "linear_5": "PTLinearMetatype",
     "dropout_3": "PTDropoutMetatype",
-    "_param_constant18": "PTConstNoopMetatype",
-    "_param_constant19": "PTConstNoopMetatype",
     "layer_norm_2": "PTLayerNormMetatype",
     "add_2": "PTAddMetatype",
-    "_tensor_constant2": "PTConstNoopMetatype",
-    "_param_constant20": "PTConstNoopMetatype",
-    "_param_constant21": "PTConstNoopMetatype",
     "linear_6": "PTLinearMetatype",
     "relu__1": "PTRELUMetatype",
-    "_param_constant22": "PTConstNoopMetatype",
     "linear_7": "PTLinearMetatype",
     "view_4": "PTReshapeMetatype",
-    "_tensor_constant3": "PTConstNoopMetatype",
     "index_1": "UnknownMetatype",
     "view_5": "PTReshapeMetatype",
     "permute_5": "PTTransposeMetatype",
@@ -105,11 +626,9 @@
     "view_6": "PTReshapeMetatype",
     "permute_6": "PTTransposeMetatype",
     "reshape_4": "PTReshapeMetatype",
-    "_param_constant23": "PTConstNoopMetatype",
     "clone_1": "PTNoopMetatype",
     "slice_4": "PTGatherMetatype",
     "zero__1": "UnknownMetatype",
-    "_param_constant24": "PTConstNoopMetatype",
     "linear_8": "PTLinearMetatype",
     "reshape_5": "PTReshapeMetatype",
     "permute_7": "PTTransposeMetatype",
@@ -126,53 +645,43 @@
     "div_3": "PTDivMetatype",
     "transpose_2": "PTTransposeMetatype",
     "matmul_2": "PTMatMulMetatype",
-    "_param_constant25": "PTConstNoopMetatype",
     "clamp_1": "UnknownMetatype",
     "exp_1": "PTExpMetatype",
     "mul_3": "PTMulMetatype",
     "add_3": "PTAddMetatype",
     "new_zeros": "UnknownMetatype",
-    "_tensor_constant4": "PTConstNoopMetatype",
     "lift_fresh_copy": "UnknownMetatype",
     "slice_5": "PTGatherMetatype",
     "slice_6": "PTGatherMetatype",
     "fill_": "UnknownMetatype",
-    "_tensor_constant5": "PTConstNoopMetatype",
     "lift_fresh_copy_1": "UnknownMetatype",
     "slice_7": "PTGatherMetatype",
     "slice_8": "PTGatherMetatype",
     "fill__1": "UnknownMetatype",
-    "_tensor_constant6": "PTConstNoopMetatype",
     "lift_fresh_copy_2": "UnknownMetatype",
     "slice_9": "PTGatherMetatype",
     "slice_10": "PTGatherMetatype",
     "fill__2": "UnknownMetatype",
-    "_tensor_constant7": "PTConstNoopMetatype",
     "lift_fresh_copy_3": "UnknownMetatype",
     "slice_11": "PTGatherMetatype",
     "slice_12": "PTGatherMetatype",
     "fill__3": "UnknownMetatype",
-    "_tensor_constant8": "PTConstNoopMetatype",
     "lift_fresh_copy_4": "UnknownMetatype",
     "slice_13": "PTGatherMetatype",
     "slice_14": "PTGatherMetatype",
     "fill__4": "UnknownMetatype",
-    "_tensor_constant9": "PTConstNoopMetatype",
     "lift_fresh_copy_5": "UnknownMetatype",
     "slice_15": "PTGatherMetatype",
     "slice_16": "PTGatherMetatype",
     "fill__5": "UnknownMetatype",
-    "_tensor_constant10": "PTConstNoopMetatype",
     "lift_fresh_copy_6": "UnknownMetatype",
     "slice_17": "PTGatherMetatype",
     "slice_18": "PTGatherMetatype",
     "fill__6": "UnknownMetatype",
-    "_tensor_constant11": "PTConstNoopMetatype",
     "lift_fresh_copy_7": "UnknownMetatype",
     "slice_19": "PTGatherMetatype",
     "slice_20": "PTGatherMetatype",
     "fill__7": "UnknownMetatype",
-    "_tensor_constant12": "PTConstNoopMetatype",
     "lift_fresh_copy_8": "UnknownMetatype",
     "slice_21": "PTGatherMetatype",
     "slice_22": "PTGatherMetatype",
@@ -197,8 +706,6 @@
     "matmul_3": "PTMatMulMetatype",
     "transpose_3": "PTTransposeMetatype",
     "reshape_7": "PTReshapeMetatype",
-    "_param_constant26": "PTConstNoopMetatype",
-    "_param_constant27": "PTConstNoopMetatype",
     "linear_9": "PTLinearMetatype",
     "dropout_5": "PTDropoutMetatype",
     "view_10": "PTReshapeMetatype",
@@ -207,21 +714,13 @@
     "roll_1": "UnknownMetatype",
     "slice_23": "PTGatherMetatype",
     "slice_24": "PTGatherMetatype",
-    "_param_constant28": "PTConstNoopMetatype",
-    "_param_constant29": "PTConstNoopMetatype",
     "layer_norm_3": "PTLayerNormMetatype",
     "add_5": "PTAddMetatype",
-    "_param_constant30": "PTConstNoopMetatype",
-    "_param_constant31": "PTConstNoopMetatype",
     "linear_10": "PTLinearMetatype",
     "gelu_1": "PTGELUMetatype",
     "dropout_6": "PTDropoutMetatype",
-    "_param_constant32": "PTConstNoopMetatype",
-    "_param_constant33": "PTConstNoopMetatype",
     "linear_11": "PTLinearMetatype",
     "dropout_7": "PTDropoutMetatype",
-    "_param_constant34": "PTConstNoopMetatype",
-    "_param_constant35": "PTConstNoopMetatype",
     "layer_norm_4": "PTLayerNormMetatype",
     "add_6": "PTAddMetatype",
     "pad_2": "PTPadMetatype",
@@ -238,20 +737,12 @@
     "slice_35": "PTGatherMetatype",
     "slice_36": "PTGatherMetatype",
     "cat": "PTCatMetatype",
-    "_param_constant36": "PTConstNoopMetatype",
     "linear_12": "PTLinearMetatype",
-    "_param_constant37": "PTConstNoopMetatype",
-    "_param_constant38": "PTConstNoopMetatype",
     "layer_norm_5": "PTLayerNormMetatype",
-    "_tensor_constant13": "PTConstNoopMetatype",
-    "_param_constant39": "PTConstNoopMetatype",
-    "_param_constant40": "PTConstNoopMetatype",
     "linear_13": "PTLinearMetatype",
     "relu__2": "PTRELUMetatype",
-    "_param_constant41": "PTConstNoopMetatype",
     "linear_14": "PTLinearMetatype",
     "view_11": "PTReshapeMetatype",
-    "_tensor_constant14": "PTConstNoopMetatype",
     "index_2": "UnknownMetatype",
     "view_12": "PTReshapeMetatype",
     "permute_10": "PTTransposeMetatype",
@@ -263,11 +754,9 @@
     "view_13": "PTReshapeMetatype",
     "permute_11": "PTTransposeMetatype",
     "reshape_9": "PTReshapeMetatype",
-    "_param_constant42": "PTConstNoopMetatype",
     "clone_2": "PTNoopMetatype",
     "slice_37": "PTGatherMetatype",
     "zero__2": "UnknownMetatype",
-    "_param_constant43": "PTConstNoopMetatype",
     "linear_15": "PTLinearMetatype",
     "reshape_10": "PTReshapeMetatype",
     "permute_12": "PTTransposeMetatype",
@@ -284,7 +773,6 @@
     "div_5": "PTDivMetatype",
     "transpose_4": "PTTransposeMetatype",
     "matmul_4": "PTMatMulMetatype",
-    "_param_constant44": "PTConstNoopMetatype",
     "clamp_2": "UnknownMetatype",
     "exp_2": "PTExpMetatype",
     "mul_5": "PTMulMetatype",
@@ -294,8 +782,6 @@
     "matmul_5": "PTMatMulMetatype",
     "transpose_5": "PTTransposeMetatype",
     "reshape_11": "PTReshapeMetatype",
-    "_param_constant45": "PTConstNoopMetatype",
-    "_param_constant46": "PTConstNoopMetatype",
     "linear_16": "PTLinearMetatype",
     "dropout_9": "PTDropoutMetatype",
     "view_14": "PTReshapeMetatype",
@@ -306,32 +792,19 @@
     "slice_40": "PTGatherMetatype",
     "slice_41": "PTGatherMetatype",
     "contiguous_3": "PTNoopMetatype",
-    "_param_constant47": "PTConstNoopMetatype",
-    "_param_constant48": "PTConstNoopMetatype",
     "layer_norm_6": "PTLayerNormMetatype",
     "add_8": "PTAddMetatype",
-    "_param_constant49": "PTConstNoopMetatype",
-    "_param_constant50": "PTConstNoopMetatype",
     "linear_17": "PTLinearMetatype",
     "gelu_2": "PTGELUMetatype",
     "dropout_10": "PTDropoutMetatype",
-    "_param_constant51": "PTConstNoopMetatype",
-    "_param_constant52": "PTConstNoopMetatype",
     "linear_18": "PTLinearMetatype",
     "dropout_11": "PTDropoutMetatype",
-    "_param_constant53": "PTConstNoopMetatype",
-    "_param_constant54": "PTConstNoopMetatype",
     "layer_norm_7": "PTLayerNormMetatype",
     "add_9": "PTAddMetatype",
-    "_tensor_constant15": "PTConstNoopMetatype",
-    "_param_constant55": "PTConstNoopMetatype",
-    "_param_constant56": "PTConstNoopMetatype",
     "linear_19": "PTLinearMetatype",
     "relu__3": "PTRELUMetatype",
-    "_param_constant57": "PTConstNoopMetatype",
     "linear_20": "PTLinearMetatype",
     "view_15": "PTReshapeMetatype",
-    "_tensor_constant16": "PTConstNoopMetatype",
     "index_3": "UnknownMetatype",
     "view_16": "PTReshapeMetatype",
     "permute_14": "PTTransposeMetatype",
@@ -344,11 +817,9 @@
     "view_17": "PTReshapeMetatype",
     "permute_15": "PTTransposeMetatype",
     "reshape_13": "PTReshapeMetatype",
-    "_param_constant58": "PTConstNoopMetatype",
     "clone_3": "PTNoopMetatype",
     "slice_42": "PTGatherMetatype",
     "zero__3": "UnknownMetatype",
-    "_param_constant59": "PTConstNoopMetatype",
     "linear_21": "PTLinearMetatype",
     "reshape_14": "PTReshapeMetatype",
     "permute_16": "PTTransposeMetatype",
@@ -365,53 +836,43 @@
     "div_7": "PTDivMetatype",
     "transpose_6": "PTTransposeMetatype",
     "matmul_6": "PTMatMulMetatype",
-    "_param_constant60": "PTConstNoopMetatype",
     "clamp_3": "UnknownMetatype",
     "exp_3": "PTExpMetatype",
     "mul_7": "PTMulMetatype",
     "add_10": "PTAddMetatype",
     "new_zeros_1": "UnknownMetatype",
-    "_tensor_constant17": "PTConstNoopMetatype",
     "lift_fresh_copy_9": "UnknownMetatype",
     "slice_43": "PTGatherMetatype",
     "slice_44": "PTGatherMetatype",
     "fill__9": "UnknownMetatype",
-    "_tensor_constant18": "PTConstNoopMetatype",
     "lift_fresh_copy_10": "UnknownMetatype",
     "slice_45": "PTGatherMetatype",
     "slice_46": "PTGatherMetatype",
     "fill__10": "UnknownMetatype",
-    "_tensor_constant19": "PTConstNoopMetatype",
     "lift_fresh_copy_11": "UnknownMetatype",
     "slice_47": "PTGatherMetatype",
     "slice_48": "PTGatherMetatype",
     "fill__11": "UnknownMetatype",
-    "_tensor_constant20": "PTConstNoopMetatype",
     "lift_fresh_copy_12": "UnknownMetatype",
     "slice_49": "PTGatherMetatype",
     "slice_50": "PTGatherMetatype",
     "fill__12": "UnknownMetatype",
-    "_tensor_constant21": "PTConstNoopMetatype",
     "lift_fresh_copy_13": "UnknownMetatype",
     "slice_51": "PTGatherMetatype",
     "slice_52": "PTGatherMetatype",
     "fill__13": "UnknownMetatype",
-    "_tensor_constant22": "PTConstNoopMetatype",
     "lift_fresh_copy_14": "UnknownMetatype",
     "slice_53": "PTGatherMetatype",
     "slice_54": "PTGatherMetatype",
     "fill__14": "UnknownMetatype",
-    "_tensor_constant23": "PTConstNoopMetatype",
     "lift_fresh_copy_15": "UnknownMetatype",
     "slice_55": "PTGatherMetatype",
     "slice_56": "PTGatherMetatype",
     "fill__15": "UnknownMetatype",
-    "_tensor_constant24": "PTConstNoopMetatype",
     "lift_fresh_copy_16": "UnknownMetatype",
     "slice_57": "PTGatherMetatype",
     "slice_58": "PTGatherMetatype",
     "fill__16": "UnknownMetatype",
-    "_tensor_constant25": "PTConstNoopMetatype",
     "lift_fresh_copy_17": "UnknownMetatype",
     "slice_59": "PTGatherMetatype",
     "slice_60": "PTGatherMetatype",
@@ -436,8 +897,6 @@
     "matmul_7": "PTMatMulMetatype",
     "transpose_7": "PTTransposeMetatype",
     "reshape_16": "PTReshapeMetatype",
-    "_param_constant61": "PTConstNoopMetatype",
-    "_param_constant62": "PTConstNoopMetatype",
     "linear_22": "PTLinearMetatype",
     "dropout_13": "PTDropoutMetatype",
     "view_21": "PTReshapeMetatype",
@@ -449,21 +908,13 @@
     "slice_63": "PTGatherMetatype",
     "slice_64": "PTGatherMetatype",
     "contiguous_5": "PTNoopMetatype",
-    "_param_constant63": "PTConstNoopMetatype",
-    "_param_constant64": "PTConstNoopMetatype",
     "layer_norm_8": "PTLayerNormMetatype",
     "add_12": "PTAddMetatype",
-    "_param_constant65": "PTConstNoopMetatype",
-    "_param_constant66": "PTConstNoopMetatype",
     "linear_23": "PTLinearMetatype",
     "gelu_3": "PTGELUMetatype",
     "dropout_14": "PTDropoutMetatype",
-    "_param_constant67": "PTConstNoopMetatype",
-    "_param_constant68": "PTConstNoopMetatype",
     "linear_24": "PTLinearMetatype",
     "dropout_15": "PTDropoutMetatype",
-    "_param_constant69": "PTConstNoopMetatype",
-    "_param_constant70": "PTConstNoopMetatype",
     "layer_norm_9": "PTLayerNormMetatype",
     "add_13": "PTAddMetatype",
     "pad_5": "PTPadMetatype",
@@ -480,20 +931,12 @@
     "slice_75": "PTGatherMetatype",
     "slice_76": "PTGatherMetatype",
     "cat_1": "PTCatMetatype",
-    "_param_constant71": "PTConstNoopMetatype",
     "linear_25": "PTLinearMetatype",
-    "_param_constant72": "PTConstNoopMetatype",
-    "_param_constant73": "PTConstNoopMetatype",
     "layer_norm_10": "PTLayerNormMetatype",
-    "_tensor_constant26": "PTConstNoopMetatype",
-    "_param_constant74": "PTConstNoopMetatype",
-    "_param_constant75": "PTConstNoopMetatype",
     "linear_26": "PTLinearMetatype",
     "relu__4": "PTRELUMetatype",
-    "_param_constant76": "PTConstNoopMetatype",
     "linear_27": "PTLinearMetatype",
     "view_22": "PTReshapeMetatype",
-    "_tensor_constant27": "PTConstNoopMetatype",
     "index_4": "UnknownMetatype",
     "view_23": "PTReshapeMetatype",
     "permute_19": "PTTransposeMetatype",
@@ -505,11 +948,9 @@
     "view_24": "PTReshapeMetatype",
     "permute_20": "PTTransposeMetatype",
     "reshape_18": "PTReshapeMetatype",
-    "_param_constant77": "PTConstNoopMetatype",
     "clone_4": "PTNoopMetatype",
     "slice_77": "PTGatherMetatype",
     "zero__4": "UnknownMetatype",
-    "_param_constant78": "PTConstNoopMetatype",
     "linear_28": "PTLinearMetatype",
     "reshape_19": "PTReshapeMetatype",
     "permute_21": "PTTransposeMetatype",
@@ -526,7 +967,6 @@
     "div_9": "PTDivMetatype",
     "transpose_8": "PTTransposeMetatype",
     "matmul_8": "PTMatMulMetatype",
-    "_param_constant79": "PTConstNoopMetatype",
     "clamp_4": "UnknownMetatype",
     "exp_4": "PTExpMetatype",
     "mul_9": "PTMulMetatype",
@@ -536,8 +976,6 @@
     "matmul_9": "PTMatMulMetatype",
     "transpose_9": "PTTransposeMetatype",
     "reshape_20": "PTReshapeMetatype",
-    "_param_constant80": "PTConstNoopMetatype",
-    "_param_constant81": "PTConstNoopMetatype",
     "linear_29": "PTLinearMetatype",
     "dropout_17": "PTDropoutMetatype",
     "view_25": "PTReshapeMetatype",
@@ -548,32 +986,19 @@
     "slice_80": "PTGatherMetatype",
     "slice_81": "PTGatherMetatype",
     "contiguous_7": "PTNoopMetatype",
-    "_param_constant82": "PTConstNoopMetatype",
-    "_param_constant83": "PTConstNoopMetatype",
     "layer_norm_11": "PTLayerNormMetatype",
     "add_15": "PTAddMetatype",
-    "_param_constant84": "PTConstNoopMetatype",
-    "_param_constant85": "PTConstNoopMetatype",
     "linear_30": "PTLinearMetatype",
     "gelu_4": "PTGELUMetatype",
     "dropout_18": "PTDropoutMetatype",
-    "_param_constant86": "PTConstNoopMetatype",
-    "_param_constant87": "PTConstNoopMetatype",
     "linear_31": "PTLinearMetatype",
     "dropout_19": "PTDropoutMetatype",
-    "_param_constant88": "PTConstNoopMetatype",
-    "_param_constant89": "PTConstNoopMetatype",
     "layer_norm_12": "PTLayerNormMetatype",
     "add_16": "PTAddMetatype",
-    "_tensor_constant28": "PTConstNoopMetatype",
-    "_param_constant90": "PTConstNoopMetatype",
-    "_param_constant91": "PTConstNoopMetatype",
     "linear_32": "PTLinearMetatype",
     "relu__5": "PTRELUMetatype",
-    "_param_constant92": "PTConstNoopMetatype",
     "linear_33": "PTLinearMetatype",
     "view_26": "PTReshapeMetatype",
-    "_tensor_constant29": "PTConstNoopMetatype",
     "index_5": "UnknownMetatype",
     "view_27": "PTReshapeMetatype",
     "permute_23": "PTTransposeMetatype",
@@ -586,11 +1011,9 @@
     "view_28": "PTReshapeMetatype",
     "permute_24": "PTTransposeMetatype",
     "reshape_22": "PTReshapeMetatype",
-    "_param_constant93": "PTConstNoopMetatype",
     "clone_5": "PTNoopMetatype",
     "slice_82": "PTGatherMetatype",
     "zero__5": "UnknownMetatype",
-    "_param_constant94": "PTConstNoopMetatype",
     "linear_34": "PTLinearMetatype",
     "reshape_23": "PTReshapeMetatype",
     "permute_25": "PTTransposeMetatype",
@@ -607,53 +1030,43 @@
     "div_11": "PTDivMetatype",
     "transpose_10": "PTTransposeMetatype",
     "matmul_10": "PTMatMulMetatype",
-    "_param_constant95": "PTConstNoopMetatype",
     "clamp_5": "UnknownMetatype",
     "exp_5": "PTExpMetatype",
     "mul_11": "PTMulMetatype",
     "add_17": "PTAddMetatype",
     "new_zeros_2": "UnknownMetatype",
-    "_tensor_constant30": "PTConstNoopMetatype",
     "lift_fresh_copy_18": "UnknownMetatype",
     "slice_83": "PTGatherMetatype",
     "slice_84": "PTGatherMetatype",
     "fill__18": "UnknownMetatype",
-    "_tensor_constant31": "PTConstNoopMetatype",
     "lift_fresh_copy_19": "UnknownMetatype",
     "slice_85": "PTGatherMetatype",
     "slice_86": "PTGatherMetatype",
     "fill__19": "UnknownMetatype",
-    "_tensor_constant32": "PTConstNoopMetatype",
     "lift_fresh_copy_20": "UnknownMetatype",
     "slice_87": "PTGatherMetatype",
     "slice_88": "PTGatherMetatype",
     "fill__20": "UnknownMetatype",
-    "_tensor_constant33": "PTConstNoopMetatype",
     "lift_fresh_copy_21": "UnknownMetatype",
     "slice_89": "PTGatherMetatype",
     "slice_90": "PTGatherMetatype",
     "fill__21": "UnknownMetatype",
-    "_tensor_constant34": "PTConstNoopMetatype",
     "lift_fresh_copy_22": "UnknownMetatype",
     "slice_91": "PTGatherMetatype",
     "slice_92": "PTGatherMetatype",
     "fill__22": "UnknownMetatype",
-    "_tensor_constant35": "PTConstNoopMetatype",
     "lift_fresh_copy_23": "UnknownMetatype",
     "slice_93": "PTGatherMetatype",
     "slice_94": "PTGatherMetatype",
     "fill__23": "UnknownMetatype",
-    "_tensor_constant36": "PTConstNoopMetatype",
     "lift_fresh_copy_24": "UnknownMetatype",
     "slice_95": "PTGatherMetatype",
     "slice_96": "PTGatherMetatype",
     "fill__24": "UnknownMetatype",
-    "_tensor_constant37": "PTConstNoopMetatype",
     "lift_fresh_copy_25": "UnknownMetatype",
     "slice_97": "PTGatherMetatype",
     "slice_98": "PTGatherMetatype",
     "fill__25": "UnknownMetatype",
-    "_tensor_constant38": "PTConstNoopMetatype",
     "lift_fresh_copy_26": "UnknownMetatype",
     "slice_99": "PTGatherMetatype",
     "slice_100": "PTGatherMetatype",
@@ -678,8 +1091,6 @@
     "matmul_11": "PTMatMulMetatype",
     "transpose_11": "PTTransposeMetatype",
     "reshape_25": "PTReshapeMetatype",
-    "_param_constant96": "PTConstNoopMetatype",
-    "_param_constant97": "PTConstNoopMetatype",
     "linear_35": "PTLinearMetatype",
     "dropout_21": "PTDropoutMetatype",
     "view_32": "PTReshapeMetatype",
@@ -691,32 +1102,19 @@
     "slice_103": "PTGatherMetatype",
     "slice_104": "PTGatherMetatype",
     "contiguous_9": "PTNoopMetatype",
-    "_param_constant98": "PTConstNoopMetatype",
-    "_param_constant99": "PTConstNoopMetatype",
     "layer_norm_13": "PTLayerNormMetatype",
     "add_19": "PTAddMetatype",
-    "_param_constant100": "PTConstNoopMetatype",
-    "_param_constant101": "PTConstNoopMetatype",
     "linear_36": "PTLinearMetatype",
     "gelu_5": "PTGELUMetatype",
     "dropout_22": "PTDropoutMetatype",
-    "_param_constant102": "PTConstNoopMetatype",
-    "_param_constant103": "PTConstNoopMetatype",
     "linear_37": "PTLinearMetatype",
     "dropout_23": "PTDropoutMetatype",
-    "_param_constant104": "PTConstNoopMetatype",
-    "_param_constant105": "PTConstNoopMetatype",
     "layer_norm_14": "PTLayerNormMetatype",
     "add_20": "PTAddMetatype",
-    "_tensor_constant39": "PTConstNoopMetatype",
-    "_param_constant106": "PTConstNoopMetatype",
-    "_param_constant107": "PTConstNoopMetatype",
     "linear_38": "PTLinearMetatype",
     "relu__6": "PTRELUMetatype",
-    "_param_constant108": "PTConstNoopMetatype",
     "linear_39": "PTLinearMetatype",
     "view_33": "PTReshapeMetatype",
-    "_tensor_constant40": "PTConstNoopMetatype",
     "index_6": "UnknownMetatype",
     "view_34": "PTReshapeMetatype",
     "permute_28": "PTTransposeMetatype",
@@ -728,11 +1126,9 @@
     "view_35": "PTReshapeMetatype",
     "permute_29": "PTTransposeMetatype",
     "reshape_27": "PTReshapeMetatype",
-    "_param_constant109": "PTConstNoopMetatype",
     "clone_6": "PTNoopMetatype",
     "slice_105": "PTGatherMetatype",
     "zero__6": "UnknownMetatype",
-    "_param_constant110": "PTConstNoopMetatype",
     "linear_40": "PTLinearMetatype",
     "reshape_28": "PTReshapeMetatype",
     "permute_30": "PTTransposeMetatype",
@@ -749,7 +1145,6 @@
     "div_13": "PTDivMetatype",
     "transpose_12": "PTTransposeMetatype",
     "matmul_12": "PTMatMulMetatype",
-    "_param_constant111": "PTConstNoopMetatype",
     "clamp_6": "UnknownMetatype",
     "exp_6": "PTExpMetatype",
     "mul_13": "PTMulMetatype",
@@ -759,8 +1154,6 @@
     "matmul_13": "PTMatMulMetatype",
     "transpose_13": "PTTransposeMetatype",
     "reshape_29": "PTReshapeMetatype",
-    "_param_constant112": "PTConstNoopMetatype",
-    "_param_constant113": "PTConstNoopMetatype",
     "linear_41": "PTLinearMetatype",
     "dropout_25": "PTDropoutMetatype",
     "view_36": "PTReshapeMetatype",
@@ -771,32 +1164,19 @@
     "slice_108": "PTGatherMetatype",
     "slice_109": "PTGatherMetatype",
     "contiguous_11": "PTNoopMetatype",
-    "_param_constant114": "PTConstNoopMetatype",
-    "_param_constant115": "PTConstNoopMetatype",
     "layer_norm_15": "PTLayerNormMetatype",
     "add_22": "PTAddMetatype",
-    "_param_constant116": "PTConstNoopMetatype",
-    "_param_constant117": "PTConstNoopMetatype",
     "linear_42": "PTLinearMetatype",
     "gelu_6": "PTGELUMetatype",
     "dropout_26": "PTDropoutMetatype",
-    "_param_constant118": "PTConstNoopMetatype",
-    "_param_constant119": "PTConstNoopMetatype",
     "linear_43": "PTLinearMetatype",
     "dropout_27": "PTDropoutMetatype",
-    "_param_constant120": "PTConstNoopMetatype",
-    "_param_constant121": "PTConstNoopMetatype",
     "layer_norm_16": "PTLayerNormMetatype",
     "add_23": "PTAddMetatype",
-    "_tensor_constant41": "PTConstNoopMetatype",
-    "_param_constant122": "PTConstNoopMetatype",
-    "_param_constant123": "PTConstNoopMetatype",
     "linear_44": "PTLinearMetatype",
     "relu__7": "PTRELUMetatype",
-    "_param_constant124": "PTConstNoopMetatype",
     "linear_45": "PTLinearMetatype",
     "view_37": "PTReshapeMetatype",
-    "_tensor_constant42": "PTConstNoopMetatype",
     "index_7": "UnknownMetatype",
     "view_38": "PTReshapeMetatype",
     "permute_32": "PTTransposeMetatype",
@@ -809,11 +1189,9 @@
     "view_39": "PTReshapeMetatype",
     "permute_33": "PTTransposeMetatype",
     "reshape_31": "PTReshapeMetatype",
-    "_param_constant125": "PTConstNoopMetatype",
     "clone_7": "PTNoopMetatype",
     "slice_110": "PTGatherMetatype",
     "zero__7": "UnknownMetatype",
-    "_param_constant126": "PTConstNoopMetatype",
     "linear_46": "PTLinearMetatype",
     "reshape_32": "PTReshapeMetatype",
     "permute_34": "PTTransposeMetatype",
@@ -830,53 +1208,43 @@
     "div_15": "PTDivMetatype",
     "transpose_14": "PTTransposeMetatype",
     "matmul_14": "PTMatMulMetatype",
-    "_param_constant127": "PTConstNoopMetatype",
     "clamp_7": "UnknownMetatype",
     "exp_7": "PTExpMetatype",
     "mul_15": "PTMulMetatype",
     "add_24": "PTAddMetatype",
     "new_zeros_3": "UnknownMetatype",
-    "_tensor_constant43": "PTConstNoopMetatype",
     "lift_fresh_copy_27": "UnknownMetatype",
     "slice_111": "PTGatherMetatype",
     "slice_112": "PTGatherMetatype",
     "fill__27": "UnknownMetatype",
-    "_tensor_constant44": "PTConstNoopMetatype",
     "lift_fresh_copy_28": "UnknownMetatype",
     "slice_113": "PTGatherMetatype",
     "slice_114": "PTGatherMetatype",
     "fill__28": "UnknownMetatype",
-    "_tensor_constant45": "PTConstNoopMetatype",
     "lift_fresh_copy_29": "UnknownMetatype",
     "slice_115": "PTGatherMetatype",
     "slice_116": "PTGatherMetatype",
     "fill__29": "UnknownMetatype",
-    "_tensor_constant46": "PTConstNoopMetatype",
     "lift_fresh_copy_30": "UnknownMetatype",
     "slice_117": "PTGatherMetatype",
     "slice_118": "PTGatherMetatype",
     "fill__30": "UnknownMetatype",
-    "_tensor_constant47": "PTConstNoopMetatype",
     "lift_fresh_copy_31": "UnknownMetatype",
     "slice_119": "PTGatherMetatype",
     "slice_120": "PTGatherMetatype",
     "fill__31": "UnknownMetatype",
-    "_tensor_constant48": "PTConstNoopMetatype",
     "lift_fresh_copy_32": "UnknownMetatype",
     "slice_121": "PTGatherMetatype",
     "slice_122": "PTGatherMetatype",
     "fill__32": "UnknownMetatype",
-    "_tensor_constant49": "PTConstNoopMetatype",
     "lift_fresh_copy_33": "UnknownMetatype",
     "slice_123": "PTGatherMetatype",
     "slice_124": "PTGatherMetatype",
     "fill__33": "UnknownMetatype",
-    "_tensor_constant50": "PTConstNoopMetatype",
     "lift_fresh_copy_34": "UnknownMetatype",
     "slice_125": "PTGatherMetatype",
     "slice_126": "PTGatherMetatype",
     "fill__34": "UnknownMetatype",
-    "_tensor_constant51": "PTConstNoopMetatype",
     "lift_fresh_copy_35": "UnknownMetatype",
     "slice_127": "PTGatherMetatype",
     "slice_128": "PTGatherMetatype",
@@ -901,8 +1269,6 @@
     "matmul_15": "PTMatMulMetatype",
     "transpose_15": "PTTransposeMetatype",
     "reshape_34": "PTReshapeMetatype",
-    "_param_constant128": "PTConstNoopMetatype",
-    "_param_constant129": "PTConstNoopMetatype",
     "linear_47": "PTLinearMetatype",
     "dropout_29": "PTDropoutMetatype",
     "view_43": "PTReshapeMetatype",
@@ -914,32 +1280,19 @@
     "slice_131": "PTGatherMetatype",
     "slice_132": "PTGatherMetatype",
     "contiguous_13": "PTNoopMetatype",
-    "_param_constant130": "PTConstNoopMetatype",
-    "_param_constant131": "PTConstNoopMetatype",
     "layer_norm_17": "PTLayerNormMetatype",
     "add_26": "PTAddMetatype",
-    "_param_constant132": "PTConstNoopMetatype",
-    "_param_constant133": "PTConstNoopMetatype",
     "linear_48": "PTLinearMetatype",
     "gelu_7": "PTGELUMetatype",
     "dropout_30": "PTDropoutMetatype",
-    "_param_constant134": "PTConstNoopMetatype",
-    "_param_constant135": "PTConstNoopMetatype",
     "linear_49": "PTLinearMetatype",
     "dropout_31": "PTDropoutMetatype",
-    "_param_constant136": "PTConstNoopMetatype",
-    "_param_constant137": "PTConstNoopMetatype",
     "layer_norm_18": "PTLayerNormMetatype",
     "add_27": "PTAddMetatype",
-    "_tensor_constant52": "PTConstNoopMetatype",
-    "_param_constant138": "PTConstNoopMetatype",
-    "_param_constant139": "PTConstNoopMetatype",
     "linear_50": "PTLinearMetatype",
     "relu__8": "PTRELUMetatype",
-    "_param_constant140": "PTConstNoopMetatype",
     "linear_51": "PTLinearMetatype",
     "view_44": "PTReshapeMetatype",
-    "_tensor_constant53": "PTConstNoopMetatype",
     "index_8": "UnknownMetatype",
     "view_45": "PTReshapeMetatype",
     "permute_37": "PTTransposeMetatype",
@@ -951,11 +1304,9 @@
     "view_46": "PTReshapeMetatype",
     "permute_38": "PTTransposeMetatype",
     "reshape_36": "PTReshapeMetatype",
-    "_param_constant141": "PTConstNoopMetatype",
     "clone_8": "PTNoopMetatype",
     "slice_133": "PTGatherMetatype",
     "zero__8": "UnknownMetatype",
-    "_param_constant142": "PTConstNoopMetatype",
     "linear_52": "PTLinearMetatype",
     "reshape_37": "PTReshapeMetatype",
     "permute_39": "PTTransposeMetatype",
@@ -972,7 +1323,6 @@
     "div_17": "PTDivMetatype",
     "transpose_16": "PTTransposeMetatype",
     "matmul_16": "PTMatMulMetatype",
-    "_param_constant143": "PTConstNoopMetatype",
     "clamp_8": "UnknownMetatype",
     "exp_8": "PTExpMetatype",
     "mul_17": "PTMulMetatype",
@@ -982,8 +1332,6 @@
     "matmul_17": "PTMatMulMetatype",
     "transpose_17": "PTTransposeMetatype",
     "reshape_38": "PTReshapeMetatype",
-    "_param_constant144": "PTConstNoopMetatype",
-    "_param_constant145": "PTConstNoopMetatype",
     "linear_53": "PTLinearMetatype",
     "dropout_33": "PTDropoutMetatype",
     "view_47": "PTReshapeMetatype",
@@ -994,32 +1342,19 @@
     "slice_136": "PTGatherMetatype",
     "slice_137": "PTGatherMetatype",
     "contiguous_15": "PTNoopMetatype",
-    "_param_constant146": "PTConstNoopMetatype",
-    "_param_constant147": "PTConstNoopMetatype",
     "layer_norm_19": "PTLayerNormMetatype",
     "add_29": "PTAddMetatype",
-    "_param_constant148": "PTConstNoopMetatype",
-    "_param_constant149": "PTConstNoopMetatype",
     "linear_54": "PTLinearMetatype",
     "gelu_8": "PTGELUMetatype",
     "dropout_34": "PTDropoutMetatype",
-    "_param_constant150": "PTConstNoopMetatype",
-    "_param_constant151": "PTConstNoopMetatype",
     "linear_55": "PTLinearMetatype",
     "dropout_35": "PTDropoutMetatype",
-    "_param_constant152": "PTConstNoopMetatype",
-    "_param_constant153": "PTConstNoopMetatype",
     "layer_norm_20": "PTLayerNormMetatype",
     "add_30": "PTAddMetatype",
-    "_tensor_constant54": "PTConstNoopMetatype",
-    "_param_constant154": "PTConstNoopMetatype",
-    "_param_constant155": "PTConstNoopMetatype",
     "linear_56": "PTLinearMetatype",
     "relu__9": "PTRELUMetatype",
-    "_param_constant156": "PTConstNoopMetatype",
     "linear_57": "PTLinearMetatype",
     "view_48": "PTReshapeMetatype",
-    "_tensor_constant55": "PTConstNoopMetatype",
     "index_9": "UnknownMetatype",
     "view_49": "PTReshapeMetatype",
     "permute_41": "PTTransposeMetatype",
@@ -1032,11 +1367,9 @@
     "view_50": "PTReshapeMetatype",
     "permute_42": "PTTransposeMetatype",
     "reshape_40": "PTReshapeMetatype",
-    "_param_constant157": "PTConstNoopMetatype",
     "clone_9": "PTNoopMetatype",
     "slice_138": "PTGatherMetatype",
     "zero__9": "UnknownMetatype",
-    "_param_constant158": "PTConstNoopMetatype",
     "linear_58": "PTLinearMetatype",
     "reshape_41": "PTReshapeMetatype",
     "permute_43": "PTTransposeMetatype",
@@ -1053,53 +1386,43 @@
     "div_19": "PTDivMetatype",
     "transpose_18": "PTTransposeMetatype",
     "matmul_18": "PTMatMulMetatype",
-    "_param_constant159": "PTConstNoopMetatype",
     "clamp_9": "UnknownMetatype",
     "exp_9": "PTExpMetatype",
     "mul_19": "PTMulMetatype",
     "add_31": "PTAddMetatype",
     "new_zeros_4": "UnknownMetatype",
-    "_tensor_constant56": "PTConstNoopMetatype",
     "lift_fresh_copy_36": "UnknownMetatype",
     "slice_139": "PTGatherMetatype",
     "slice_140": "PTGatherMetatype",
     "fill__36": "UnknownMetatype",
-    "_tensor_constant57": "PTConstNoopMetatype",
     "lift_fresh_copy_37": "UnknownMetatype",
     "slice_141": "PTGatherMetatype",
     "slice_142": "PTGatherMetatype",
     "fill__37": "UnknownMetatype",
-    "_tensor_constant58": "PTConstNoopMetatype",
     "lift_fresh_copy_38": "UnknownMetatype",
     "slice_143": "PTGatherMetatype",
     "slice_144": "PTGatherMetatype",
     "fill__38": "UnknownMetatype",
-    "_tensor_constant59": "PTConstNoopMetatype",
     "lift_fresh_copy_39": "UnknownMetatype",
     "slice_145": "PTGatherMetatype",
     "slice_146": "PTGatherMetatype",
     "fill__39": "UnknownMetatype",
-    "_tensor_constant60": "PTConstNoopMetatype",
     "lift_fresh_copy_40": "UnknownMetatype",
     "slice_147": "PTGatherMetatype",
     "slice_148": "PTGatherMetatype",
     "fill__40": "UnknownMetatype",
-    "_tensor_constant61": "PTConstNoopMetatype",
     "lift_fresh_copy_41": "UnknownMetatype",
     "slice_149": "PTGatherMetatype",
     "slice_150": "PTGatherMetatype",
     "fill__41": "UnknownMetatype",
-    "_tensor_constant62": "PTConstNoopMetatype",
     "lift_fresh_copy_42": "UnknownMetatype",
     "slice_151": "PTGatherMetatype",
     "slice_152": "PTGatherMetatype",
     "fill__42": "UnknownMetatype",
-    "_tensor_constant63": "PTConstNoopMetatype",
     "lift_fresh_copy_43": "UnknownMetatype",
     "slice_153": "PTGatherMetatype",
     "slice_154": "PTGatherMetatype",
     "fill__43": "UnknownMetatype",
-    "_tensor_constant64": "PTConstNoopMetatype",
     "lift_fresh_copy_44": "UnknownMetatype",
     "slice_155": "PTGatherMetatype",
     "slice_156": "PTGatherMetatype",
@@ -1124,8 +1447,6 @@
     "matmul_19": "PTMatMulMetatype",
     "transpose_19": "PTTransposeMetatype",
     "reshape_43": "PTReshapeMetatype",
-    "_param_constant160": "PTConstNoopMetatype",
-    "_param_constant161": "PTConstNoopMetatype",
     "linear_59": "PTLinearMetatype",
     "dropout_37": "PTDropoutMetatype",
     "view_54": "PTReshapeMetatype",
@@ -1137,32 +1458,19 @@
     "slice_159": "PTGatherMetatype",
     "slice_160": "PTGatherMetatype",
     "contiguous_17": "PTNoopMetatype",
-    "_param_constant162": "PTConstNoopMetatype",
-    "_param_constant163": "PTConstNoopMetatype",
     "layer_norm_21": "PTLayerNormMetatype",
     "add_33": "PTAddMetatype",
-    "_param_constant164": "PTConstNoopMetatype",
-    "_param_constant165": "PTConstNoopMetatype",
     "linear_60": "PTLinearMetatype",
     "gelu_9": "PTGELUMetatype",
     "dropout_38": "PTDropoutMetatype",
-    "_param_constant166": "PTConstNoopMetatype",
-    "_param_constant167": "PTConstNoopMetatype",
     "linear_61": "PTLinearMetatype",
     "dropout_39": "PTDropoutMetatype",
-    "_param_constant168": "PTConstNoopMetatype",
-    "_param_constant169": "PTConstNoopMetatype",
     "layer_norm_22": "PTLayerNormMetatype",
     "add_34": "PTAddMetatype",
-    "_tensor_constant65": "PTConstNoopMetatype",
-    "_param_constant170": "PTConstNoopMetatype",
-    "_param_constant171": "PTConstNoopMetatype",
     "linear_62": "PTLinearMetatype",
     "relu__10": "PTRELUMetatype",
-    "_param_constant172": "PTConstNoopMetatype",
     "linear_63": "PTLinearMetatype",
     "view_55": "PTReshapeMetatype",
-    "_tensor_constant66": "PTConstNoopMetatype",
     "index_10": "UnknownMetatype",
     "view_56": "PTReshapeMetatype",
     "permute_46": "PTTransposeMetatype",
@@ -1174,11 +1482,9 @@
     "view_57": "PTReshapeMetatype",
     "permute_47": "PTTransposeMetatype",
     "reshape_45": "PTReshapeMetatype",
-    "_param_constant173": "PTConstNoopMetatype",
     "clone_10": "PTNoopMetatype",
     "slice_161": "PTGatherMetatype",
     "zero__10": "UnknownMetatype",
-    "_param_constant174": "PTConstNoopMetatype",
     "linear_64": "PTLinearMetatype",
     "reshape_46": "PTReshapeMetatype",
     "permute_48": "PTTransposeMetatype",
@@ -1195,7 +1501,6 @@
     "div_21": "PTDivMetatype",
     "transpose_20": "PTTransposeMetatype",
     "matmul_20": "PTMatMulMetatype",
-    "_param_constant175": "PTConstNoopMetatype",
     "clamp_10": "UnknownMetatype",
     "exp_10": "PTExpMetatype",
     "mul_21": "PTMulMetatype",
@@ -1205,8 +1510,6 @@
     "matmul_21": "PTMatMulMetatype",
     "transpose_21": "PTTransposeMetatype",
     "reshape_47": "PTReshapeMetatype",
-    "_param_constant176": "PTConstNoopMetatype",
-    "_param_constant177": "PTConstNoopMetatype",
     "linear_65": "PTLinearMetatype",
     "dropout_41": "PTDropoutMetatype",
     "view_58": "PTReshapeMetatype",
@@ -1217,32 +1520,19 @@
     "slice_164": "PTGatherMetatype",
     "slice_165": "PTGatherMetatype",
     "contiguous_19": "PTNoopMetatype",
-    "_param_constant178": "PTConstNoopMetatype",
-    "_param_constant179": "PTConstNoopMetatype",
     "layer_norm_23": "PTLayerNormMetatype",
     "add_36": "PTAddMetatype",
-    "_param_constant180": "PTConstNoopMetatype",
-    "_param_constant181": "PTConstNoopMetatype",
     "linear_66": "PTLinearMetatype",
     "gelu_10": "PTGELUMetatype",
     "dropout_42": "PTDropoutMetatype",
-    "_param_constant182": "PTConstNoopMetatype",
-    "_param_constant183": "PTConstNoopMetatype",
     "linear_67": "PTLinearMetatype",
     "dropout_43": "PTDropoutMetatype",
-    "_param_constant184": "PTConstNoopMetatype",
-    "_param_constant185": "PTConstNoopMetatype",
     "layer_norm_24": "PTLayerNormMetatype",
     "add_37": "PTAddMetatype",
-    "_tensor_constant67": "PTConstNoopMetatype",
-    "_param_constant186": "PTConstNoopMetatype",
-    "_param_constant187": "PTConstNoopMetatype",
     "linear_68": "PTLinearMetatype",
     "relu__11": "PTRELUMetatype",
-    "_param_constant188": "PTConstNoopMetatype",
     "linear_69": "PTLinearMetatype",
     "view_59": "PTReshapeMetatype",
-    "_tensor_constant68": "PTConstNoopMetatype",
     "index_11": "UnknownMetatype",
     "view_60": "PTReshapeMetatype",
     "permute_50": "PTTransposeMetatype",
@@ -1255,11 +1545,9 @@
     "view_61": "PTReshapeMetatype",
     "permute_51": "PTTransposeMetatype",
     "reshape_49": "PTReshapeMetatype",
-    "_param_constant189": "PTConstNoopMetatype",
     "clone_11": "PTNoopMetatype",
     "slice_166": "PTGatherMetatype",
     "zero__11": "UnknownMetatype",
-    "_param_constant190": "PTConstNoopMetatype",
     "linear_70": "PTLinearMetatype",
     "reshape_50": "PTReshapeMetatype",
     "permute_52": "PTTransposeMetatype",
@@ -1276,53 +1564,43 @@
     "div_23": "PTDivMetatype",
     "transpose_22": "PTTransposeMetatype",
     "matmul_22": "PTMatMulMetatype",
-    "_param_constant191": "PTConstNoopMetatype",
     "clamp_11": "UnknownMetatype",
     "exp_11": "PTExpMetatype",
     "mul_23": "PTMulMetatype",
     "add_38": "PTAddMetatype",
     "new_zeros_5": "UnknownMetatype",
-    "_tensor_constant69": "PTConstNoopMetatype",
     "lift_fresh_copy_45": "UnknownMetatype",
     "slice_167": "PTGatherMetatype",
     "slice_168": "PTGatherMetatype",
     "fill__45": "UnknownMetatype",
-    "_tensor_constant70": "PTConstNoopMetatype",
     "lift_fresh_copy_46": "UnknownMetatype",
     "slice_169": "PTGatherMetatype",
     "slice_170": "PTGatherMetatype",
     "fill__46": "UnknownMetatype",
-    "_tensor_constant71": "PTConstNoopMetatype",
     "lift_fresh_copy_47": "UnknownMetatype",
     "slice_171": "PTGatherMetatype",
     "slice_172": "PTGatherMetatype",
     "fill__47": "UnknownMetatype",
-    "_tensor_constant72": "PTConstNoopMetatype",
     "lift_fresh_copy_48": "UnknownMetatype",
     "slice_173": "PTGatherMetatype",
     "slice_174": "PTGatherMetatype",
     "fill__48": "UnknownMetatype",
-    "_tensor_constant73": "PTConstNoopMetatype",
     "lift_fresh_copy_49": "UnknownMetatype",
     "slice_175": "PTGatherMetatype",
     "slice_176": "PTGatherMetatype",
     "fill__49": "UnknownMetatype",
-    "_tensor_constant74": "PTConstNoopMetatype",
     "lift_fresh_copy_50": "UnknownMetatype",
     "slice_177": "PTGatherMetatype",
     "slice_178": "PTGatherMetatype",
     "fill__50": "UnknownMetatype",
-    "_tensor_constant75": "PTConstNoopMetatype",
     "lift_fresh_copy_51": "UnknownMetatype",
     "slice_179": "PTGatherMetatype",
     "slice_180": "PTGatherMetatype",
     "fill__51": "UnknownMetatype",
-    "_tensor_constant76": "PTConstNoopMetatype",
     "lift_fresh_copy_52": "UnknownMetatype",
     "slice_181": "PTGatherMetatype",
     "slice_182": "PTGatherMetatype",
     "fill__52": "UnknownMetatype",
-    "_tensor_constant77": "PTConstNoopMetatype",
     "lift_fresh_copy_53": "UnknownMetatype",
     "slice_183": "PTGatherMetatype",
     "slice_184": "PTGatherMetatype",
@@ -1347,8 +1625,6 @@
     "matmul_23": "PTMatMulMetatype",
     "transpose_23": "PTTransposeMetatype",
     "reshape_52": "PTReshapeMetatype",
-    "_param_constant192": "PTConstNoopMetatype",
-    "_param_constant193": "PTConstNoopMetatype",
     "linear_71": "PTLinearMetatype",
     "dropout_45": "PTDropoutMetatype",
     "view_65": "PTReshapeMetatype",
@@ -1360,32 +1636,19 @@
     "slice_187": "PTGatherMetatype",
     "slice_188": "PTGatherMetatype",
     "contiguous_21": "PTNoopMetatype",
-    "_param_constant194": "PTConstNoopMetatype",
-    "_param_constant195": "PTConstNoopMetatype",
     "layer_norm_25": "PTLayerNormMetatype",
     "add_40": "PTAddMetatype",
-    "_param_constant196": "PTConstNoopMetatype",
-    "_param_constant197": "PTConstNoopMetatype",
     "linear_72": "PTLinearMetatype",
     "gelu_11": "PTGELUMetatype",
     "dropout_46": "PTDropoutMetatype",
-    "_param_constant198": "PTConstNoopMetatype",
-    "_param_constant199": "PTConstNoopMetatype",
     "linear_73": "PTLinearMetatype",
     "dropout_47": "PTDropoutMetatype",
-    "_param_constant200": "PTConstNoopMetatype",
-    "_param_constant201": "PTConstNoopMetatype",
     "layer_norm_26": "PTLayerNormMetatype",
     "add_41": "PTAddMetatype",
-    "_tensor_constant78": "PTConstNoopMetatype",
-    "_param_constant202": "PTConstNoopMetatype",
-    "_param_constant203": "PTConstNoopMetatype",
     "linear_74": "PTLinearMetatype",
     "relu__12": "PTRELUMetatype",
-    "_param_constant204": "PTConstNoopMetatype",
     "linear_75": "PTLinearMetatype",
     "view_66": "PTReshapeMetatype",
-    "_tensor_constant79": "PTConstNoopMetatype",
     "index_12": "UnknownMetatype",
     "view_67": "PTReshapeMetatype",
     "permute_55": "PTTransposeMetatype",
@@ -1397,11 +1660,9 @@
     "view_68": "PTReshapeMetatype",
     "permute_56": "PTTransposeMetatype",
     "reshape_54": "PTReshapeMetatype",
-    "_param_constant205": "PTConstNoopMetatype",
     "clone_12": "PTNoopMetatype",
     "slice_189": "PTGatherMetatype",
     "zero__12": "UnknownMetatype",
-    "_param_constant206": "PTConstNoopMetatype",
     "linear_76": "PTLinearMetatype",
     "reshape_55": "PTReshapeMetatype",
     "permute_57": "PTTransposeMetatype",
@@ -1418,7 +1679,6 @@
     "div_25": "PTDivMetatype",
     "transpose_24": "PTTransposeMetatype",
     "matmul_24": "PTMatMulMetatype",
-    "_param_constant207": "PTConstNoopMetatype",
     "clamp_12": "UnknownMetatype",
     "exp_12": "PTExpMetatype",
     "mul_25": "PTMulMetatype",
@@ -1428,8 +1688,6 @@
     "matmul_25": "PTMatMulMetatype",
     "transpose_25": "PTTransposeMetatype",
     "reshape_56": "PTReshapeMetatype",
-    "_param_constant208": "PTConstNoopMetatype",
-    "_param_constant209": "PTConstNoopMetatype",
     "linear_77": "PTLinearMetatype",
     "dropout_49": "PTDropoutMetatype",
     "view_69": "PTReshapeMetatype",
@@ -1440,32 +1698,19 @@
     "slice_192": "PTGatherMetatype",
     "slice_193": "PTGatherMetatype",
     "contiguous_23": "PTNoopMetatype",
-    "_param_constant210": "PTConstNoopMetatype",
-    "_param_constant211": "PTConstNoopMetatype",
     "layer_norm_27": "PTLayerNormMetatype",
     "add_43": "PTAddMetatype",
-    "_param_constant212": "PTConstNoopMetatype",
-    "_param_constant213": "PTConstNoopMetatype",
     "linear_78": "PTLinearMetatype",
     "gelu_12": "PTGELUMetatype",
     "dropout_50": "PTDropoutMetatype",
-    "_param_constant214": "PTConstNoopMetatype",
-    "_param_constant215": "PTConstNoopMetatype",
     "linear_79": "PTLinearMetatype",
     "dropout_51": "PTDropoutMetatype",
-    "_param_constant216": "PTConstNoopMetatype",
-    "_param_constant217": "PTConstNoopMetatype",
     "layer_norm_28": "PTLayerNormMetatype",
     "add_44": "PTAddMetatype",
-    "_tensor_constant80": "PTConstNoopMetatype",
-    "_param_constant218": "PTConstNoopMetatype",
-    "_param_constant219": "PTConstNoopMetatype",
     "linear_80": "PTLinearMetatype",
     "relu__13": "PTRELUMetatype",
-    "_param_constant220": "PTConstNoopMetatype",
     "linear_81": "PTLinearMetatype",
     "view_70": "PTReshapeMetatype",
-    "_tensor_constant81": "PTConstNoopMetatype",
     "index_13": "UnknownMetatype",
     "view_71": "PTReshapeMetatype",
     "permute_59": "PTTransposeMetatype",
@@ -1478,11 +1723,9 @@
     "view_72": "PTReshapeMetatype",
     "permute_60": "PTTransposeMetatype",
     "reshape_58": "PTReshapeMetatype",
-    "_param_constant221": "PTConstNoopMetatype",
     "clone_13": "PTNoopMetatype",
     "slice_194": "PTGatherMetatype",
     "zero__13": "UnknownMetatype",
-    "_param_constant222": "PTConstNoopMetatype",
     "linear_82": "PTLinearMetatype",
     "reshape_59": "PTReshapeMetatype",
     "permute_61": "PTTransposeMetatype",
@@ -1499,53 +1742,43 @@
     "div_27": "PTDivMetatype",
     "transpose_26": "PTTransposeMetatype",
     "matmul_26": "PTMatMulMetatype",
-    "_param_constant223": "PTConstNoopMetatype",
     "clamp_13": "UnknownMetatype",
     "exp_13": "PTExpMetatype",
     "mul_27": "PTMulMetatype",
     "add_45": "PTAddMetatype",
     "new_zeros_6": "UnknownMetatype",
-    "_tensor_constant82": "PTConstNoopMetatype",
     "lift_fresh_copy_54": "UnknownMetatype",
     "slice_195": "PTGatherMetatype",
     "slice_196": "PTGatherMetatype",
     "fill__54": "UnknownMetatype",
-    "_tensor_constant83": "PTConstNoopMetatype",
     "lift_fresh_copy_55": "UnknownMetatype",
     "slice_197": "PTGatherMetatype",
     "slice_198": "PTGatherMetatype",
     "fill__55": "UnknownMetatype",
-    "_tensor_constant84": "PTConstNoopMetatype",
     "lift_fresh_copy_56": "UnknownMetatype",
     "slice_199": "PTGatherMetatype",
     "slice_200": "PTGatherMetatype",
     "fill__56": "UnknownMetatype",
-    "_tensor_constant85": "PTConstNoopMetatype",
     "lift_fresh_copy_57": "UnknownMetatype",
     "slice_201": "PTGatherMetatype",
     "slice_202": "PTGatherMetatype",
     "fill__57": "UnknownMetatype",
-    "_tensor_constant86": "PTConstNoopMetatype",
     "lift_fresh_copy_58": "UnknownMetatype",
     "slice_203": "PTGatherMetatype",
     "slice_204": "PTGatherMetatype",
     "fill__58": "UnknownMetatype",
-    "_tensor_constant87": "PTConstNoopMetatype",
     "lift_fresh_copy_59": "UnknownMetatype",
     "slice_205": "PTGatherMetatype",
     "slice_206": "PTGatherMetatype",
     "fill__59": "UnknownMetatype",
-    "_tensor_constant88": "PTConstNoopMetatype",
     "lift_fresh_copy_60": "UnknownMetatype",
     "slice_207": "PTGatherMetatype",
     "slice_208": "PTGatherMetatype",
     "fill__60": "UnknownMetatype",
-    "_tensor_constant89": "PTConstNoopMetatype",
     "lift_fresh_copy_61": "UnknownMetatype",
     "slice_209": "PTGatherMetatype",
     "slice_210": "PTGatherMetatype",
     "fill__61": "UnknownMetatype",
-    "_tensor_constant90": "PTConstNoopMetatype",
     "lift_fresh_copy_62": "UnknownMetatype",
     "slice_211": "PTGatherMetatype",
     "slice_212": "PTGatherMetatype",
@@ -1570,8 +1803,6 @@
     "matmul_27": "PTMatMulMetatype",
     "transpose_27": "PTTransposeMetatype",
     "reshape_61": "PTReshapeMetatype",
-    "_param_constant224": "PTConstNoopMetatype",
-    "_param_constant225": "PTConstNoopMetatype",
     "linear_83": "PTLinearMetatype",
     "dropout_53": "PTDropoutMetatype",
     "view_76": "PTReshapeMetatype",
@@ -1583,32 +1814,19 @@
     "slice_215": "PTGatherMetatype",
     "slice_216": "PTGatherMetatype",
     "contiguous_25": "PTNoopMetatype",
-    "_param_constant226": "PTConstNoopMetatype",
-    "_param_constant227": "PTConstNoopMetatype",
     "layer_norm_29": "PTLayerNormMetatype",
     "add_47": "PTAddMetatype",
-    "_param_constant228": "PTConstNoopMetatype",
-    "_param_constant229": "PTConstNoopMetatype",
     "linear_84": "PTLinearMetatype",
     "gelu_13": "PTGELUMetatype",
     "dropout_54": "PTDropoutMetatype",
-    "_param_constant230": "PTConstNoopMetatype",
-    "_param_constant231": "PTConstNoopMetatype",
     "linear_85": "PTLinearMetatype",
     "dropout_55": "PTDropoutMetatype",
-    "_param_constant232": "PTConstNoopMetatype",
-    "_param_constant233": "PTConstNoopMetatype",
     "layer_norm_30": "PTLayerNormMetatype",
     "add_48": "PTAddMetatype",
-    "_tensor_constant91": "PTConstNoopMetatype",
-    "_param_constant234": "PTConstNoopMetatype",
-    "_param_constant235": "PTConstNoopMetatype",
     "linear_86": "PTLinearMetatype",
     "relu__14": "PTRELUMetatype",
-    "_param_constant236": "PTConstNoopMetatype",
     "linear_87": "PTLinearMetatype",
     "view_77": "PTReshapeMetatype",
-    "_tensor_constant92": "PTConstNoopMetatype",
     "index_14": "UnknownMetatype",
     "view_78": "PTReshapeMetatype",
     "permute_64": "PTTransposeMetatype",
@@ -1620,11 +1838,9 @@
     "view_79": "PTReshapeMetatype",
     "permute_65": "PTTransposeMetatype",
     "reshape_63": "PTReshapeMetatype",
-    "_param_constant237": "PTConstNoopMetatype",
     "clone_14": "PTNoopMetatype",
     "slice_217": "PTGatherMetatype",
     "zero__14": "UnknownMetatype",
-    "_param_constant238": "PTConstNoopMetatype",
     "linear_88": "PTLinearMetatype",
     "reshape_64": "PTReshapeMetatype",
     "permute_66": "PTTransposeMetatype",
@@ -1641,7 +1857,6 @@
     "div_29": "PTDivMetatype",
     "transpose_28": "PTTransposeMetatype",
     "matmul_28": "PTMatMulMetatype",
-    "_param_constant239": "PTConstNoopMetatype",
     "clamp_14": "UnknownMetatype",
     "exp_14": "PTExpMetatype",
     "mul_29": "PTMulMetatype",
@@ -1651,8 +1866,6 @@
     "matmul_29": "PTMatMulMetatype",
     "transpose_29": "PTTransposeMetatype",
     "reshape_65": "PTReshapeMetatype",
-    "_param_constant240": "PTConstNoopMetatype",
-    "_param_constant241": "PTConstNoopMetatype",
     "linear_89": "PTLinearMetatype",
     "dropout_57": "PTDropoutMetatype",
     "view_80": "PTReshapeMetatype",
@@ -1663,32 +1876,19 @@
     "slice_220": "PTGatherMetatype",
     "slice_221": "PTGatherMetatype",
     "contiguous_27": "PTNoopMetatype",
-    "_param_constant242": "PTConstNoopMetatype",
-    "_param_constant243": "PTConstNoopMetatype",
     "layer_norm_31": "PTLayerNormMetatype",
     "add_50": "PTAddMetatype",
-    "_param_constant244": "PTConstNoopMetatype",
-    "_param_constant245": "PTConstNoopMetatype",
     "linear_90": "PTLinearMetatype",
     "gelu_14": "PTGELUMetatype",
     "dropout_58": "PTDropoutMetatype",
-    "_param_constant246": "PTConstNoopMetatype",
-    "_param_constant247": "PTConstNoopMetatype",
     "linear_91": "PTLinearMetatype",
     "dropout_59": "PTDropoutMetatype",
-    "_param_constant248": "PTConstNoopMetatype",
-    "_param_constant249": "PTConstNoopMetatype",
     "layer_norm_32": "PTLayerNormMetatype",
     "add_51": "PTAddMetatype",
-    "_tensor_constant93": "PTConstNoopMetatype",
-    "_param_constant250": "PTConstNoopMetatype",
-    "_param_constant251": "PTConstNoopMetatype",
     "linear_92": "PTLinearMetatype",
     "relu__15": "PTRELUMetatype",
-    "_param_constant252": "PTConstNoopMetatype",
     "linear_93": "PTLinearMetatype",
     "view_81": "PTReshapeMetatype",
-    "_tensor_constant94": "PTConstNoopMetatype",
     "index_15": "UnknownMetatype",
     "view_82": "PTReshapeMetatype",
     "permute_68": "PTTransposeMetatype",
@@ -1701,11 +1901,9 @@
     "view_83": "PTReshapeMetatype",
     "permute_69": "PTTransposeMetatype",
     "reshape_67": "PTReshapeMetatype",
-    "_param_constant253": "PTConstNoopMetatype",
     "clone_15": "PTNoopMetatype",
     "slice_222": "PTGatherMetatype",
     "zero__15": "UnknownMetatype",
-    "_param_constant254": "PTConstNoopMetatype",
     "linear_94": "PTLinearMetatype",
     "reshape_68": "PTReshapeMetatype",
     "permute_70": "PTTransposeMetatype",
@@ -1722,53 +1920,43 @@
     "div_31": "PTDivMetatype",
     "transpose_30": "PTTransposeMetatype",
     "matmul_30": "PTMatMulMetatype",
-    "_param_constant255": "PTConstNoopMetatype",
     "clamp_15": "UnknownMetatype",
     "exp_15": "PTExpMetatype",
     "mul_31": "PTMulMetatype",
     "add_52": "PTAddMetatype",
     "new_zeros_7": "UnknownMetatype",
-    "_tensor_constant95": "PTConstNoopMetatype",
     "lift_fresh_copy_63": "UnknownMetatype",
     "slice_223": "PTGatherMetatype",
     "slice_224": "PTGatherMetatype",
     "fill__63": "UnknownMetatype",
-    "_tensor_constant96": "PTConstNoopMetatype",
     "lift_fresh_copy_64": "UnknownMetatype",
     "slice_225": "PTGatherMetatype",
     "slice_226": "PTGatherMetatype",
     "fill__64": "UnknownMetatype",
-    "_tensor_constant97": "PTConstNoopMetatype",
     "lift_fresh_copy_65": "UnknownMetatype",
     "slice_227": "PTGatherMetatype",
     "slice_228": "PTGatherMetatype",
     "fill__65": "UnknownMetatype",
-    "_tensor_constant98": "PTConstNoopMetatype",
     "lift_fresh_copy_66": "UnknownMetatype",
     "slice_229": "PTGatherMetatype",
     "slice_230": "PTGatherMetatype",
     "fill__66": "UnknownMetatype",
-    "_tensor_constant99": "PTConstNoopMetatype",
     "lift_fresh_copy_67": "UnknownMetatype",
     "slice_231": "PTGatherMetatype",
     "slice_232": "PTGatherMetatype",
     "fill__67": "UnknownMetatype",
-    "_tensor_constant100": "PTConstNoopMetatype",
     "lift_fresh_copy_68": "UnknownMetatype",
     "slice_233": "PTGatherMetatype",
     "slice_234": "PTGatherMetatype",
     "fill__68": "UnknownMetatype",
-    "_tensor_constant101": "PTConstNoopMetatype",
     "lift_fresh_copy_69": "UnknownMetatype",
     "slice_235": "PTGatherMetatype",
     "slice_236": "PTGatherMetatype",
     "fill__69": "UnknownMetatype",
-    "_tensor_constant102": "PTConstNoopMetatype",
     "lift_fresh_copy_70": "UnknownMetatype",
     "slice_237": "PTGatherMetatype",
     "slice_238": "PTGatherMetatype",
     "fill__70": "UnknownMetatype",
-    "_tensor_constant103": "PTConstNoopMetatype",
     "lift_fresh_copy_71": "UnknownMetatype",
     "slice_239": "PTGatherMetatype",
     "slice_240": "PTGatherMetatype",
@@ -1793,8 +1981,6 @@
     "matmul_31": "PTMatMulMetatype",
     "transpose_31": "PTTransposeMetatype",
     "reshape_70": "PTReshapeMetatype",
-    "_param_constant256": "PTConstNoopMetatype",
-    "_param_constant257": "PTConstNoopMetatype",
     "linear_95": "PTLinearMetatype",
     "dropout_61": "PTDropoutMetatype",
     "view_87": "PTReshapeMetatype",
@@ -1806,32 +1992,19 @@
     "slice_243": "PTGatherMetatype",
     "slice_244": "PTGatherMetatype",
     "contiguous_29": "PTNoopMetatype",
-    "_param_constant258": "PTConstNoopMetatype",
-    "_param_constant259": "PTConstNoopMetatype",
     "layer_norm_33": "PTLayerNormMetatype",
     "add_54": "PTAddMetatype",
-    "_param_constant260": "PTConstNoopMetatype",
-    "_param_constant261": "PTConstNoopMetatype",
     "linear_96": "PTLinearMetatype",
     "gelu_15": "PTGELUMetatype",
     "dropout_62": "PTDropoutMetatype",
-    "_param_constant262": "PTConstNoopMetatype",
-    "_param_constant263": "PTConstNoopMetatype",
     "linear_97": "PTLinearMetatype",
     "dropout_63": "PTDropoutMetatype",
-    "_param_constant264": "PTConstNoopMetatype",
-    "_param_constant265": "PTConstNoopMetatype",
     "layer_norm_34": "PTLayerNormMetatype",
     "add_55": "PTAddMetatype",
-    "_tensor_constant104": "PTConstNoopMetatype",
-    "_param_constant266": "PTConstNoopMetatype",
-    "_param_constant267": "PTConstNoopMetatype",
     "linear_98": "PTLinearMetatype",
     "relu__16": "PTRELUMetatype",
-    "_param_constant268": "PTConstNoopMetatype",
     "linear_99": "PTLinearMetatype",
     "view_88": "PTReshapeMetatype",
-    "_tensor_constant105": "PTConstNoopMetatype",
     "index_16": "UnknownMetatype",
     "view_89": "PTReshapeMetatype",
     "permute_73": "PTTransposeMetatype",
@@ -1843,11 +2016,9 @@
     "view_90": "PTReshapeMetatype",
     "permute_74": "PTTransposeMetatype",
     "reshape_72": "PTReshapeMetatype",
-    "_param_constant269": "PTConstNoopMetatype",
     "clone_16": "PTNoopMetatype",
     "slice_245": "PTGatherMetatype",
     "zero__16": "UnknownMetatype",
-    "_param_constant270": "PTConstNoopMetatype",
     "linear_100": "PTLinearMetatype",
     "reshape_73": "PTReshapeMetatype",
     "permute_75": "PTTransposeMetatype",
@@ -1864,7 +2035,6 @@
     "div_33": "PTDivMetatype",
     "transpose_32": "PTTransposeMetatype",
     "matmul_32": "PTMatMulMetatype",
-    "_param_constant271": "PTConstNoopMetatype",
     "clamp_16": "UnknownMetatype",
     "exp_16": "PTExpMetatype",
     "mul_33": "PTMulMetatype",
@@ -1874,8 +2044,6 @@
     "matmul_33": "PTMatMulMetatype",
     "transpose_33": "PTTransposeMetatype",
     "reshape_74": "PTReshapeMetatype",
-    "_param_constant272": "PTConstNoopMetatype",
-    "_param_constant273": "PTConstNoopMetatype",
     "linear_101": "PTLinearMetatype",
     "dropout_65": "PTDropoutMetatype",
     "view_91": "PTReshapeMetatype",
@@ -1886,32 +2054,19 @@
     "slice_248": "PTGatherMetatype",
     "slice_249": "PTGatherMetatype",
     "contiguous_31": "PTNoopMetatype",
-    "_param_constant274": "PTConstNoopMetatype",
-    "_param_constant275": "PTConstNoopMetatype",
     "layer_norm_35": "PTLayerNormMetatype",
     "add_57": "PTAddMetatype",
-    "_param_constant276": "PTConstNoopMetatype",
-    "_param_constant277": "PTConstNoopMetatype",
     "linear_102": "PTLinearMetatype",
     "gelu_16": "PTGELUMetatype",
     "dropout_66": "PTDropoutMetatype",
-    "_param_constant278": "PTConstNoopMetatype",
-    "_param_constant279": "PTConstNoopMetatype",
     "linear_103": "PTLinearMetatype",
     "dropout_67": "PTDropoutMetatype",
-    "_param_constant280": "PTConstNoopMetatype",
-    "_param_constant281": "PTConstNoopMetatype",
     "layer_norm_36": "PTLayerNormMetatype",
     "add_58": "PTAddMetatype",
-    "_tensor_constant106": "PTConstNoopMetatype",
-    "_param_constant282": "PTConstNoopMetatype",
-    "_param_constant283": "PTConstNoopMetatype",
     "linear_104": "PTLinearMetatype",
     "relu__17": "PTRELUMetatype",
-    "_param_constant284": "PTConstNoopMetatype",
     "linear_105": "PTLinearMetatype",
     "view_92": "PTReshapeMetatype",
-    "_tensor_constant107": "PTConstNoopMetatype",
     "index_17": "UnknownMetatype",
     "view_93": "PTReshapeMetatype",
     "permute_77": "PTTransposeMetatype",
@@ -1924,11 +2079,9 @@
     "view_94": "PTReshapeMetatype",
     "permute_78": "PTTransposeMetatype",
     "reshape_76": "PTReshapeMetatype",
-    "_param_constant285": "PTConstNoopMetatype",
     "clone_17": "PTNoopMetatype",
     "slice_250": "PTGatherMetatype",
     "zero__17": "UnknownMetatype",
-    "_param_constant286": "PTConstNoopMetatype",
     "linear_106": "PTLinearMetatype",
     "reshape_77": "PTReshapeMetatype",
     "permute_79": "PTTransposeMetatype",
@@ -1945,53 +2098,43 @@
     "div_35": "PTDivMetatype",
     "transpose_34": "PTTransposeMetatype",
     "matmul_34": "PTMatMulMetatype",
-    "_param_constant287": "PTConstNoopMetatype",
     "clamp_17": "UnknownMetatype",
     "exp_17": "PTExpMetatype",
     "mul_35": "PTMulMetatype",
     "add_59": "PTAddMetatype",
     "new_zeros_8": "UnknownMetatype",
-    "_tensor_constant108": "PTConstNoopMetatype",
     "lift_fresh_copy_72": "UnknownMetatype",
     "slice_251": "PTGatherMetatype",
     "slice_252": "PTGatherMetatype",
     "fill__72": "UnknownMetatype",
-    "_tensor_constant109": "PTConstNoopMetatype",
     "lift_fresh_copy_73": "UnknownMetatype",
     "slice_253": "PTGatherMetatype",
     "slice_254": "PTGatherMetatype",
     "fill__73": "UnknownMetatype",
-    "_tensor_constant110": "PTConstNoopMetatype",
     "lift_fresh_copy_74": "UnknownMetatype",
     "slice_255": "PTGatherMetatype",
     "slice_256": "PTGatherMetatype",
     "fill__74": "UnknownMetatype",
-    "_tensor_constant111": "PTConstNoopMetatype",
     "lift_fresh_copy_75": "UnknownMetatype",
     "slice_257": "PTGatherMetatype",
     "slice_258": "PTGatherMetatype",
     "fill__75": "UnknownMetatype",
-    "_tensor_constant112": "PTConstNoopMetatype",
     "lift_fresh_copy_76": "UnknownMetatype",
     "slice_259": "PTGatherMetatype",
     "slice_260": "PTGatherMetatype",
     "fill__76": "UnknownMetatype",
-    "_tensor_constant113": "PTConstNoopMetatype",
     "lift_fresh_copy_77": "UnknownMetatype",
     "slice_261": "PTGatherMetatype",
     "slice_262": "PTGatherMetatype",
     "fill__77": "UnknownMetatype",
-    "_tensor_constant114": "PTConstNoopMetatype",
     "lift_fresh_copy_78": "UnknownMetatype",
     "slice_263": "PTGatherMetatype",
     "slice_264": "PTGatherMetatype",
     "fill__78": "UnknownMetatype",
-    "_tensor_constant115": "PTConstNoopMetatype",
     "lift_fresh_copy_79": "UnknownMetatype",
     "slice_265": "PTGatherMetatype",
     "slice_266": "PTGatherMetatype",
     "fill__79": "UnknownMetatype",
-    "_tensor_constant116": "PTConstNoopMetatype",
     "lift_fresh_copy_80": "UnknownMetatype",
     "slice_267": "PTGatherMetatype",
     "slice_268": "PTGatherMetatype",
@@ -2016,8 +2159,6 @@
     "matmul_35": "PTMatMulMetatype",
     "transpose_35": "PTTransposeMetatype",
     "reshape_79": "PTReshapeMetatype",
-    "_param_constant288": "PTConstNoopMetatype",
-    "_param_constant289": "PTConstNoopMetatype",
     "linear_107": "PTLinearMetatype",
     "dropout_69": "PTDropoutMetatype",
     "view_98": "PTReshapeMetatype",
@@ -2029,32 +2170,19 @@
     "slice_271": "PTGatherMetatype",
     "slice_272": "PTGatherMetatype",
     "contiguous_33": "PTNoopMetatype",
-    "_param_constant290": "PTConstNoopMetatype",
-    "_param_constant291": "PTConstNoopMetatype",
     "layer_norm_37": "PTLayerNormMetatype",
     "add_61": "PTAddMetatype",
-    "_param_constant292": "PTConstNoopMetatype",
-    "_param_constant293": "PTConstNoopMetatype",
     "linear_108": "PTLinearMetatype",
     "gelu_17": "PTGELUMetatype",
     "dropout_70": "PTDropoutMetatype",
-    "_param_constant294": "PTConstNoopMetatype",
-    "_param_constant295": "PTConstNoopMetatype",
     "linear_109": "PTLinearMetatype",
     "dropout_71": "PTDropoutMetatype",
-    "_param_constant296": "PTConstNoopMetatype",
-    "_param_constant297": "PTConstNoopMetatype",
     "layer_norm_38": "PTLayerNormMetatype",
     "add_62": "PTAddMetatype",
-    "_tensor_constant117": "PTConstNoopMetatype",
-    "_param_constant298": "PTConstNoopMetatype",
-    "_param_constant299": "PTConstNoopMetatype",
     "linear_110": "PTLinearMetatype",
     "relu__18": "PTRELUMetatype",
-    "_param_constant300": "PTConstNoopMetatype",
     "linear_111": "PTLinearMetatype",
     "view_99": "PTReshapeMetatype",
-    "_tensor_constant118": "PTConstNoopMetatype",
     "index_18": "UnknownMetatype",
     "view_100": "PTReshapeMetatype",
     "permute_82": "PTTransposeMetatype",
@@ -2066,11 +2194,9 @@
     "view_101": "PTReshapeMetatype",
     "permute_83": "PTTransposeMetatype",
     "reshape_81": "PTReshapeMetatype",
-    "_param_constant301": "PTConstNoopMetatype",
     "clone_18": "PTNoopMetatype",
     "slice_273": "PTGatherMetatype",
     "zero__18": "UnknownMetatype",
-    "_param_constant302": "PTConstNoopMetatype",
     "linear_112": "PTLinearMetatype",
     "reshape_82": "PTReshapeMetatype",
     "permute_84": "PTTransposeMetatype",
@@ -2087,7 +2213,6 @@
     "div_37": "PTDivMetatype",
     "transpose_36": "PTTransposeMetatype",
     "matmul_36": "PTMatMulMetatype",
-    "_param_constant303": "PTConstNoopMetatype",
     "clamp_18": "UnknownMetatype",
     "exp_18": "PTExpMetatype",
     "mul_37": "PTMulMetatype",
@@ -2097,8 +2222,6 @@
     "matmul_37": "PTMatMulMetatype",
     "transpose_37": "PTTransposeMetatype",
     "reshape_83": "PTReshapeMetatype",
-    "_param_constant304": "PTConstNoopMetatype",
-    "_param_constant305": "PTConstNoopMetatype",
     "linear_113": "PTLinearMetatype",
     "dropout_73": "PTDropoutMetatype",
     "view_102": "PTReshapeMetatype",
@@ -2109,32 +2232,19 @@
     "slice_276": "PTGatherMetatype",
     "slice_277": "PTGatherMetatype",
     "contiguous_35": "PTNoopMetatype",
-    "_param_constant306": "PTConstNoopMetatype",
-    "_param_constant307": "PTConstNoopMetatype",
     "layer_norm_39": "PTLayerNormMetatype",
     "add_64": "PTAddMetatype",
-    "_param_constant308": "PTConstNoopMetatype",
-    "_param_constant309": "PTConstNoopMetatype",
     "linear_114": "PTLinearMetatype",
     "gelu_18": "PTGELUMetatype",
     "dropout_74": "PTDropoutMetatype",
-    "_param_constant310": "PTConstNoopMetatype",
-    "_param_constant311": "PTConstNoopMetatype",
     "linear_115": "PTLinearMetatype",
     "dropout_75": "PTDropoutMetatype",
-    "_param_constant312": "PTConstNoopMetatype",
-    "_param_constant313": "PTConstNoopMetatype",
     "layer_norm_40": "PTLayerNormMetatype",
     "add_65": "PTAddMetatype",
-    "_tensor_constant119": "PTConstNoopMetatype",
-    "_param_constant314": "PTConstNoopMetatype",
-    "_param_constant315": "PTConstNoopMetatype",
     "linear_116": "PTLinearMetatype",
     "relu__19": "PTRELUMetatype",
-    "_param_constant316": "PTConstNoopMetatype",
     "linear_117": "PTLinearMetatype",
     "view_103": "PTReshapeMetatype",
-    "_tensor_constant120": "PTConstNoopMetatype",
     "index_19": "UnknownMetatype",
     "view_104": "PTReshapeMetatype",
     "permute_86": "PTTransposeMetatype",
@@ -2147,11 +2257,9 @@
     "view_105": "PTReshapeMetatype",
     "permute_87": "PTTransposeMetatype",
     "reshape_85": "PTReshapeMetatype",
-    "_param_constant317": "PTConstNoopMetatype",
     "clone_19": "PTNoopMetatype",
     "slice_278": "PTGatherMetatype",
     "zero__19": "UnknownMetatype",
-    "_param_constant318": "PTConstNoopMetatype",
     "linear_118": "PTLinearMetatype",
     "reshape_86": "PTReshapeMetatype",
     "permute_88": "PTTransposeMetatype",
@@ -2168,53 +2276,43 @@
     "div_39": "PTDivMetatype",
     "transpose_38": "PTTransposeMetatype",
     "matmul_38": "PTMatMulMetatype",
-    "_param_constant319": "PTConstNoopMetatype",
     "clamp_19": "UnknownMetatype",
     "exp_19": "PTExpMetatype",
     "mul_39": "PTMulMetatype",
     "add_66": "PTAddMetatype",
     "new_zeros_9": "UnknownMetatype",
-    "_tensor_constant121": "PTConstNoopMetatype",
     "lift_fresh_copy_81": "UnknownMetatype",
     "slice_279": "PTGatherMetatype",
     "slice_280": "PTGatherMetatype",
     "fill__81": "UnknownMetatype",
-    "_tensor_constant122": "PTConstNoopMetatype",
     "lift_fresh_copy_82": "UnknownMetatype",
     "slice_281": "PTGatherMetatype",
     "slice_282": "PTGatherMetatype",
     "fill__82": "UnknownMetatype",
-    "_tensor_constant123": "PTConstNoopMetatype",
     "lift_fresh_copy_83": "UnknownMetatype",
     "slice_283": "PTGatherMetatype",
     "slice_284": "PTGatherMetatype",
     "fill__83": "UnknownMetatype",
-    "_tensor_constant124": "PTConstNoopMetatype",
     "lift_fresh_copy_84": "UnknownMetatype",
     "slice_285": "PTGatherMetatype",
     "slice_286": "PTGatherMetatype",
     "fill__84": "UnknownMetatype",
-    "_tensor_constant125": "PTConstNoopMetatype",
     "lift_fresh_copy_85": "UnknownMetatype",
     "slice_287": "PTGatherMetatype",
     "slice_288": "PTGatherMetatype",
     "fill__85": "UnknownMetatype",
-    "_tensor_constant126": "PTConstNoopMetatype",
     "lift_fresh_copy_86": "UnknownMetatype",
     "slice_289": "PTGatherMetatype",
     "slice_290": "PTGatherMetatype",
     "fill__86": "UnknownMetatype",
-    "_tensor_constant127": "PTConstNoopMetatype",
     "lift_fresh_copy_87": "UnknownMetatype",
     "slice_291": "PTGatherMetatype",
     "slice_292": "PTGatherMetatype",
     "fill__87": "UnknownMetatype",
-    "_tensor_constant128": "PTConstNoopMetatype",
     "lift_fresh_copy_88": "UnknownMetatype",
     "slice_293": "PTGatherMetatype",
     "slice_294": "PTGatherMetatype",
     "fill__88": "UnknownMetatype",
-    "_tensor_constant129": "PTConstNoopMetatype",
     "lift_fresh_copy_89": "UnknownMetatype",
     "slice_295": "PTGatherMetatype",
     "slice_296": "PTGatherMetatype",
@@ -2239,8 +2337,6 @@
     "matmul_39": "PTMatMulMetatype",
     "transpose_39": "PTTransposeMetatype",
     "reshape_88": "PTReshapeMetatype",
-    "_param_constant320": "PTConstNoopMetatype",
-    "_param_constant321": "PTConstNoopMetatype",
     "linear_119": "PTLinearMetatype",
     "dropout_77": "PTDropoutMetatype",
     "view_109": "PTReshapeMetatype",
@@ -2252,32 +2348,19 @@
     "slice_299": "PTGatherMetatype",
     "slice_300": "PTGatherMetatype",
     "contiguous_37": "PTNoopMetatype",
-    "_param_constant322": "PTConstNoopMetatype",
-    "_param_constant323": "PTConstNoopMetatype",
     "layer_norm_41": "PTLayerNormMetatype",
     "add_68": "PTAddMetatype",
-    "_param_constant324": "PTConstNoopMetatype",
-    "_param_constant325": "PTConstNoopMetatype",
     "linear_120": "PTLinearMetatype",
     "gelu_19": "PTGELUMetatype",
     "dropout_78": "PTDropoutMetatype",
-    "_param_constant326": "PTConstNoopMetatype",
-    "_param_constant327": "PTConstNoopMetatype",
     "linear_121": "PTLinearMetatype",
     "dropout_79": "PTDropoutMetatype",
-    "_param_constant328": "PTConstNoopMetatype",
-    "_param_constant329": "PTConstNoopMetatype",
     "layer_norm_42": "PTLayerNormMetatype",
     "add_69": "PTAddMetatype",
-    "_tensor_constant130": "PTConstNoopMetatype",
-    "_param_constant330": "PTConstNoopMetatype",
-    "_param_constant331": "PTConstNoopMetatype",
     "linear_122": "PTLinearMetatype",
     "relu__20": "PTRELUMetatype",
-    "_param_constant332": "PTConstNoopMetatype",
     "linear_123": "PTLinearMetatype",
     "view_110": "PTReshapeMetatype",
-    "_tensor_constant131": "PTConstNoopMetatype",
     "index_20": "UnknownMetatype",
     "view_111": "PTReshapeMetatype",
     "permute_91": "PTTransposeMetatype",
@@ -2289,11 +2372,9 @@
     "view_112": "PTReshapeMetatype",
     "permute_92": "PTTransposeMetatype",
     "reshape_90": "PTReshapeMetatype",
-    "_param_constant333": "PTConstNoopMetatype",
     "clone_20": "PTNoopMetatype",
     "slice_301": "PTGatherMetatype",
     "zero__20": "UnknownMetatype",
-    "_param_constant334": "PTConstNoopMetatype",
     "linear_124": "PTLinearMetatype",
     "reshape_91": "PTReshapeMetatype",
     "permute_93": "PTTransposeMetatype",
@@ -2310,7 +2391,6 @@
     "div_41": "PTDivMetatype",
     "transpose_40": "PTTransposeMetatype",
     "matmul_40": "PTMatMulMetatype",
-    "_param_constant335": "PTConstNoopMetatype",
     "clamp_20": "UnknownMetatype",
     "exp_20": "PTExpMetatype",
     "mul_41": "PTMulMetatype",
@@ -2320,8 +2400,6 @@
     "matmul_41": "PTMatMulMetatype",
     "transpose_41": "PTTransposeMetatype",
     "reshape_92": "PTReshapeMetatype",
-    "_param_constant336": "PTConstNoopMetatype",
-    "_param_constant337": "PTConstNoopMetatype",
     "linear_125": "PTLinearMetatype",
     "dropout_81": "PTDropoutMetatype",
     "view_113": "PTReshapeMetatype",
@@ -2332,32 +2410,19 @@
     "slice_304": "PTGatherMetatype",
     "slice_305": "PTGatherMetatype",
     "contiguous_39": "PTNoopMetatype",
-    "_param_constant338": "PTConstNoopMetatype",
-    "_param_constant339": "PTConstNoopMetatype",
     "layer_norm_43": "PTLayerNormMetatype",
     "add_71": "PTAddMetatype",
-    "_param_constant340": "PTConstNoopMetatype",
-    "_param_constant341": "PTConstNoopMetatype",
     "linear_126": "PTLinearMetatype",
     "gelu_20": "PTGELUMetatype",
     "dropout_82": "PTDropoutMetatype",
-    "_param_constant342": "PTConstNoopMetatype",
-    "_param_constant343": "PTConstNoopMetatype",
     "linear_127": "PTLinearMetatype",
     "dropout_83": "PTDropoutMetatype",
-    "_param_constant344": "PTConstNoopMetatype",
-    "_param_constant345": "PTConstNoopMetatype",
     "layer_norm_44": "PTLayerNormMetatype",
     "add_72": "PTAddMetatype",
-    "_tensor_constant132": "PTConstNoopMetatype",
-    "_param_constant346": "PTConstNoopMetatype",
-    "_param_constant347": "PTConstNoopMetatype",
     "linear_128": "PTLinearMetatype",
     "relu__21": "PTRELUMetatype",
-    "_param_constant348": "PTConstNoopMetatype",
     "linear_129": "PTLinearMetatype",
     "view_114": "PTReshapeMetatype",
-    "_tensor_constant133": "PTConstNoopMetatype",
     "index_21": "UnknownMetatype",
     "view_115": "PTReshapeMetatype",
     "permute_95": "PTTransposeMetatype",
@@ -2370,11 +2435,9 @@
     "view_116": "PTReshapeMetatype",
     "permute_96": "PTTransposeMetatype",
     "reshape_94": "PTReshapeMetatype",
-    "_param_constant349": "PTConstNoopMetatype",
     "clone_21": "PTNoopMetatype",
     "slice_306": "PTGatherMetatype",
     "zero__21": "UnknownMetatype",
-    "_param_constant350": "PTConstNoopMetatype",
     "linear_130": "PTLinearMetatype",
     "reshape_95": "PTReshapeMetatype",
     "permute_97": "PTTransposeMetatype",
@@ -2391,53 +2454,43 @@
     "div_43": "PTDivMetatype",
     "transpose_42": "PTTransposeMetatype",
     "matmul_42": "PTMatMulMetatype",
-    "_param_constant351": "PTConstNoopMetatype",
     "clamp_21": "UnknownMetatype",
     "exp_21": "PTExpMetatype",
     "mul_43": "PTMulMetatype",
     "add_73": "PTAddMetatype",
     "new_zeros_10": "UnknownMetatype",
-    "_tensor_constant134": "PTConstNoopMetatype",
     "lift_fresh_copy_90": "UnknownMetatype",
     "slice_307": "PTGatherMetatype",
     "slice_308": "PTGatherMetatype",
     "fill__90": "UnknownMetatype",
-    "_tensor_constant135": "PTConstNoopMetatype",
     "lift_fresh_copy_91": "UnknownMetatype",
     "slice_309": "PTGatherMetatype",
     "slice_310": "PTGatherMetatype",
     "fill__91": "UnknownMetatype",
-    "_tensor_constant136": "PTConstNoopMetatype",
     "lift_fresh_copy_92": "UnknownMetatype",
     "slice_311": "PTGatherMetatype",
     "slice_312": "PTGatherMetatype",
     "fill__92": "UnknownMetatype",
-    "_tensor_constant137": "PTConstNoopMetatype",
     "lift_fresh_copy_93": "UnknownMetatype",
     "slice_313": "PTGatherMetatype",
     "slice_314": "PTGatherMetatype",
     "fill__93": "UnknownMetatype",
-    "_tensor_constant138": "PTConstNoopMetatype",
     "lift_fresh_copy_94": "UnknownMetatype",
     "slice_315": "PTGatherMetatype",
     "slice_316": "PTGatherMetatype",
     "fill__94": "UnknownMetatype",
-    "_tensor_constant139": "PTConstNoopMetatype",
     "lift_fresh_copy_95": "UnknownMetatype",
     "slice_317": "PTGatherMetatype",
     "slice_318": "PTGatherMetatype",
     "fill__95": "UnknownMetatype",
-    "_tensor_constant140": "PTConstNoopMetatype",
     "lift_fresh_copy_96": "UnknownMetatype",
     "slice_319": "PTGatherMetatype",
     "slice_320": "PTGatherMetatype",
     "fill__96": "UnknownMetatype",
-    "_tensor_constant141": "PTConstNoopMetatype",
     "lift_fresh_copy_97": "UnknownMetatype",
     "slice_321": "PTGatherMetatype",
     "slice_322": "PTGatherMetatype",
     "fill__97": "UnknownMetatype",
-    "_tensor_constant142": "PTConstNoopMetatype",
     "lift_fresh_copy_98": "UnknownMetatype",
     "slice_323": "PTGatherMetatype",
     "slice_324": "PTGatherMetatype",
@@ -2462,8 +2515,6 @@
     "matmul_43": "PTMatMulMetatype",
     "transpose_43": "PTTransposeMetatype",
     "reshape_97": "PTReshapeMetatype",
-    "_param_constant352": "PTConstNoopMetatype",
-    "_param_constant353": "PTConstNoopMetatype",
     "linear_131": "PTLinearMetatype",
     "dropout_85": "PTDropoutMetatype",
     "view_120": "PTReshapeMetatype",
@@ -2475,21 +2526,13 @@
     "slice_327": "PTGatherMetatype",
     "slice_328": "PTGatherMetatype",
     "contiguous_41": "PTNoopMetatype",
-    "_param_constant354": "PTConstNoopMetatype",
-    "_param_constant355": "PTConstNoopMetatype",
     "layer_norm_45": "PTLayerNormMetatype",
     "add_75": "PTAddMetatype",
-    "_param_constant356": "PTConstNoopMetatype",
-    "_param_constant357": "PTConstNoopMetatype",
     "linear_132": "PTLinearMetatype",
     "gelu_21": "PTGELUMetatype",
     "dropout_86": "PTDropoutMetatype",
-    "_param_constant358": "PTConstNoopMetatype",
-    "_param_constant359": "PTConstNoopMetatype",
     "linear_133": "PTLinearMetatype",
     "dropout_87": "PTDropoutMetatype",
-    "_param_constant360": "PTConstNoopMetatype",
-    "_param_constant361": "PTConstNoopMetatype",
     "layer_norm_46": "PTLayerNormMetatype",
     "add_76": "PTAddMetatype",
     "pad_24": "PTPadMetatype",
@@ -2506,20 +2549,12 @@
     "slice_339": "PTGatherMetatype",
     "slice_340": "PTGatherMetatype",
     "cat_2": "PTCatMetatype",
-    "_param_constant362": "PTConstNoopMetatype",
     "linear_134": "PTLinearMetatype",
-    "_param_constant363": "PTConstNoopMetatype",
-    "_param_constant364": "PTConstNoopMetatype",
     "layer_norm_47": "PTLayerNormMetatype",
-    "_tensor_constant143": "PTConstNoopMetatype",
-    "_param_constant365": "PTConstNoopMetatype",
-    "_param_constant366": "PTConstNoopMetatype",
     "linear_135": "PTLinearMetatype",
     "relu__22": "PTRELUMetatype",
-    "_param_constant367": "PTConstNoopMetatype",
     "linear_136": "PTLinearMetatype",
     "view_121": "PTReshapeMetatype",
-    "_tensor_constant144": "PTConstNoopMetatype",
     "index_22": "UnknownMetatype",
     "view_122": "PTReshapeMetatype",
     "permute_100": "PTTransposeMetatype",
@@ -2531,11 +2566,9 @@
     "view_123": "PTReshapeMetatype",
     "permute_101": "PTTransposeMetatype",
     "reshape_99": "PTReshapeMetatype",
-    "_param_constant368": "PTConstNoopMetatype",
     "clone_22": "PTNoopMetatype",
     "slice_341": "PTGatherMetatype",
     "zero__22": "UnknownMetatype",
-    "_param_constant369": "PTConstNoopMetatype",
     "linear_137": "PTLinearMetatype",
     "reshape_100": "PTReshapeMetatype",
     "permute_102": "PTTransposeMetatype",
@@ -2552,7 +2585,6 @@
     "div_45": "PTDivMetatype",
     "transpose_44": "PTTransposeMetatype",
     "matmul_44": "PTMatMulMetatype",
-    "_param_constant370": "PTConstNoopMetatype",
     "clamp_22": "UnknownMetatype",
     "exp_22": "PTExpMetatype",
     "mul_45": "PTMulMetatype",
@@ -2562,8 +2594,6 @@
     "matmul_45": "PTMatMulMetatype",
     "transpose_45": "PTTransposeMetatype",
     "reshape_101": "PTReshapeMetatype",
-    "_param_constant371": "PTConstNoopMetatype",
-    "_param_constant372": "PTConstNoopMetatype",
     "linear_138": "PTLinearMetatype",
     "dropout_89": "PTDropoutMetatype",
     "view_124": "PTReshapeMetatype",
@@ -2574,32 +2604,19 @@
     "slice_344": "PTGatherMetatype",
     "slice_345": "PTGatherMetatype",
     "contiguous_43": "PTNoopMetatype",
-    "_param_constant373": "PTConstNoopMetatype",
-    "_param_constant374": "PTConstNoopMetatype",
     "layer_norm_48": "PTLayerNormMetatype",
     "add_78": "PTAddMetatype",
-    "_param_constant375": "PTConstNoopMetatype",
-    "_param_constant376": "PTConstNoopMetatype",
     "linear_139": "PTLinearMetatype",
     "gelu_22": "PTGELUMetatype",
     "dropout_90": "PTDropoutMetatype",
-    "_param_constant377": "PTConstNoopMetatype",
-    "_param_constant378": "PTConstNoopMetatype",
     "linear_140": "PTLinearMetatype",
     "dropout_91": "PTDropoutMetatype",
-    "_param_constant379": "PTConstNoopMetatype",
-    "_param_constant380": "PTConstNoopMetatype",
     "layer_norm_49": "PTLayerNormMetatype",
     "add_79": "PTAddMetatype",
-    "_tensor_constant145": "PTConstNoopMetatype",
-    "_param_constant381": "PTConstNoopMetatype",
-    "_param_constant382": "PTConstNoopMetatype",
     "linear_141": "PTLinearMetatype",
     "relu__23": "PTRELUMetatype",
-    "_param_constant383": "PTConstNoopMetatype",
     "linear_142": "PTLinearMetatype",
     "view_125": "PTReshapeMetatype",
-    "_tensor_constant146": "PTConstNoopMetatype",
     "index_23": "UnknownMetatype",
     "view_126": "PTReshapeMetatype",
     "permute_104": "PTTransposeMetatype",
@@ -2611,11 +2628,9 @@
     "view_127": "PTReshapeMetatype",
     "permute_105": "PTTransposeMetatype",
     "reshape_103": "PTReshapeMetatype",
-    "_param_constant384": "PTConstNoopMetatype",
     "clone_23": "PTNoopMetatype",
     "slice_346": "PTGatherMetatype",
     "zero__23": "UnknownMetatype",
-    "_param_constant385": "PTConstNoopMetatype",
     "linear_143": "PTLinearMetatype",
     "reshape_104": "PTReshapeMetatype",
     "permute_106": "PTTransposeMetatype",
@@ -2632,7 +2647,6 @@
     "div_47": "PTDivMetatype",
     "transpose_46": "PTTransposeMetatype",
     "matmul_46": "PTMatMulMetatype",
-    "_param_constant386": "PTConstNoopMetatype",
     "clamp_23": "UnknownMetatype",
     "exp_23": "PTExpMetatype",
     "mul_47": "PTMulMetatype",
@@ -2642,8 +2656,6 @@
     "matmul_47": "PTMatMulMetatype",
     "transpose_47": "PTTransposeMetatype",
     "reshape_105": "PTReshapeMetatype",
-    "_param_constant387": "PTConstNoopMetatype",
-    "_param_constant388": "PTConstNoopMetatype",
     "linear_144": "PTLinearMetatype",
     "dropout_93": "PTDropoutMetatype",
     "view_128": "PTReshapeMetatype",
@@ -2654,31 +2666,19 @@
     "slice_349": "PTGatherMetatype",
     "slice_350": "PTGatherMetatype",
     "contiguous_45": "PTNoopMetatype",
-    "_param_constant389": "PTConstNoopMetatype",
-    "_param_constant390": "PTConstNoopMetatype",
     "layer_norm_50": "PTLayerNormMetatype",
     "add_81": "PTAddMetatype",
-    "_param_constant391": "PTConstNoopMetatype",
-    "_param_constant392": "PTConstNoopMetatype",
     "linear_145": "PTLinearMetatype",
     "gelu_23": "PTGELUMetatype",
     "dropout_94": "PTDropoutMetatype",
-    "_param_constant393": "PTConstNoopMetatype",
-    "_param_constant394": "PTConstNoopMetatype",
     "linear_146": "PTLinearMetatype",
     "dropout_95": "PTDropoutMetatype",
-    "_param_constant395": "PTConstNoopMetatype",
-    "_param_constant396": "PTConstNoopMetatype",
     "layer_norm_51": "PTLayerNormMetatype",
     "add_82": "PTAddMetatype",
-    "_param_constant397": "PTConstNoopMetatype",
-    "_param_constant398": "PTConstNoopMetatype",
     "layer_norm_52": "PTLayerNormMetatype",
     "permute_108": "PTTransposeMetatype",
     "adaptive_avg_pool2d": "PTAvgPool2dMetatype",
     "flatten": "PTReshapeMetatype",
-    "_param_constant399": "PTConstNoopMetatype",
-    "_param_constant400": "PTConstNoopMetatype",
     "linear_147": "PTLinearMetatype",
-    "output": "PTOutputNoopMetatype"
+    "output_1": "PTOutputNoopMetatype"
 }
\ No newline at end of file
diff --git a/tests/torch/data/reference_graphs/fx/reference_metatypes/synthetic_transformer.json b/tests/torch/data/reference_graphs/fx/reference_metatypes/synthetic_transformer.json
index c8375399b97..ba475c16e61 100644
--- a/tests/torch/data/reference_graphs/fx/reference_metatypes/synthetic_transformer.json
+++ b/tests/torch/data/reference_graphs/fx/reference_metatypes/synthetic_transformer.json
@@ -1,12 +1,12 @@
 {
-    "arg0_1": "PTInputNoopMetatype",
-    "_param_constant0": "PTConstNoopMetatype",
+    "wte_weight": "PTConstNoopMetatype",
+    "linear_weight": "PTConstNoopMetatype",
+    "linear_bias": "PTConstNoopMetatype",
+    "lm_head_weight": "PTConstNoopMetatype",
+    "lm_head_bias": "PTConstNoopMetatype",
+    "input_ids": "PTInputNoopMetatype",
     "embedding": "PTAtenEmbeddingMetatype",
-    "_param_constant1": "PTConstNoopMetatype",
-    "_param_constant2": "PTConstNoopMetatype",
     "linear": "PTLinearMetatype",
-    "_param_constant3": "PTConstNoopMetatype",
-    "_param_constant4": "PTConstNoopMetatype",
     "linear_1": "PTLinearMetatype",
-    "output": "PTOutputNoopMetatype"
+    "output_1": "PTOutputNoopMetatype"
 }
\ No newline at end of file
diff --git a/tests/torch/data/reference_graphs/fx/reference_metatypes/unet.json b/tests/torch/data/reference_graphs/fx/reference_metatypes/unet.json
index fb26ec71994..cf5ba08fc69 100644
--- a/tests/torch/data/reference_graphs/fx/reference_metatypes/unet.json
+++ b/tests/torch/data/reference_graphs/fx/reference_metatypes/unet.json
@@ -1,277 +1,223 @@
 {
-    "arg0_1": "PTInputNoopMetatype",
-    "_param_constant0": "PTConstNoopMetatype",
-    "_param_constant1": "PTConstNoopMetatype",
+    "down_path_0_block_0_weight": "PTConstNoopMetatype",
+    "down_path_0_block_0_bias": "PTConstNoopMetatype",
+    "down_path_0_block_1_weight": "PTConstNoopMetatype",
+    "down_path_0_block_1_bias": "PTConstNoopMetatype",
+    "down_path_0_block_3_weight": "PTConstNoopMetatype",
+    "down_path_0_block_3_bias": "PTConstNoopMetatype",
+    "down_path_0_block_4_weight": "PTConstNoopMetatype",
+    "down_path_0_block_4_bias": "PTConstNoopMetatype",
+    "down_path_1_block_0_weight": "PTConstNoopMetatype",
+    "down_path_1_block_0_bias": "PTConstNoopMetatype",
+    "down_path_1_block_1_weight": "PTConstNoopMetatype",
+    "down_path_1_block_1_bias": "PTConstNoopMetatype",
+    "down_path_1_block_3_weight": "PTConstNoopMetatype",
+    "down_path_1_block_3_bias": "PTConstNoopMetatype",
+    "down_path_1_block_4_weight": "PTConstNoopMetatype",
+    "down_path_1_block_4_bias": "PTConstNoopMetatype",
+    "down_path_2_block_0_weight": "PTConstNoopMetatype",
+    "down_path_2_block_0_bias": "PTConstNoopMetatype",
+    "down_path_2_block_1_weight": "PTConstNoopMetatype",
+    "down_path_2_block_1_bias": "PTConstNoopMetatype",
+    "down_path_2_block_3_weight": "PTConstNoopMetatype",
+    "down_path_2_block_3_bias": "PTConstNoopMetatype",
+    "down_path_2_block_4_weight": "PTConstNoopMetatype",
+    "down_path_2_block_4_bias": "PTConstNoopMetatype",
+    "down_path_3_block_0_weight": "PTConstNoopMetatype",
+    "down_path_3_block_0_bias": "PTConstNoopMetatype",
+    "down_path_3_block_1_weight": "PTConstNoopMetatype",
+    "down_path_3_block_1_bias": "PTConstNoopMetatype",
+    "down_path_3_block_3_weight": "PTConstNoopMetatype",
+    "down_path_3_block_3_bias": "PTConstNoopMetatype",
+    "down_path_3_block_4_weight": "PTConstNoopMetatype",
+    "down_path_3_block_4_bias": "PTConstNoopMetatype",
+    "down_path_4_block_0_weight": "PTConstNoopMetatype",
+    "down_path_4_block_0_bias": "PTConstNoopMetatype",
+    "down_path_4_block_1_weight": "PTConstNoopMetatype",
+    "down_path_4_block_1_bias": "PTConstNoopMetatype",
+    "down_path_4_block_3_weight": "PTConstNoopMetatype",
+    "down_path_4_block_3_bias": "PTConstNoopMetatype",
+    "down_path_4_block_4_weight": "PTConstNoopMetatype",
+    "down_path_4_block_4_bias": "PTConstNoopMetatype",
+    "up_path_0_up_weight": "PTConstNoopMetatype",
+    "up_path_0_up_bias": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_0_weight": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_0_bias": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_1_weight": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_1_bias": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_3_weight": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_3_bias": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_4_weight": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_4_bias": "PTConstNoopMetatype",
+    "up_path_1_up_weight": "PTConstNoopMetatype",
+    "up_path_1_up_bias": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_0_weight": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_0_bias": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_1_weight": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_1_bias": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_3_weight": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_3_bias": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_4_weight": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_4_bias": "PTConstNoopMetatype",
+    "up_path_2_up_weight": "PTConstNoopMetatype",
+    "up_path_2_up_bias": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_0_weight": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_0_bias": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_1_weight": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_1_bias": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_3_weight": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_3_bias": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_4_weight": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_4_bias": "PTConstNoopMetatype",
+    "up_path_3_up_weight": "PTConstNoopMetatype",
+    "up_path_3_up_bias": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_0_weight": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_0_bias": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_1_weight": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_1_bias": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_3_weight": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_3_bias": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_4_weight": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_4_bias": "PTConstNoopMetatype",
+    "last_weight": "PTConstNoopMetatype",
+    "last_bias": "PTConstNoopMetatype",
+    "down_path_0_block_1_running_mean": "PTConstNoopMetatype",
+    "down_path_0_block_1_running_var": "PTConstNoopMetatype",
+    "down_path_0_block_1_num_batches_tracked": "PTConstNoopMetatype",
+    "down_path_0_block_4_running_mean": "PTConstNoopMetatype",
+    "down_path_0_block_4_running_var": "PTConstNoopMetatype",
+    "down_path_0_block_4_num_batches_tracked": "PTConstNoopMetatype",
+    "down_path_1_block_1_running_mean": "PTConstNoopMetatype",
+    "down_path_1_block_1_running_var": "PTConstNoopMetatype",
+    "down_path_1_block_1_num_batches_tracked": "PTConstNoopMetatype",
+    "down_path_1_block_4_running_mean": "PTConstNoopMetatype",
+    "down_path_1_block_4_running_var": "PTConstNoopMetatype",
+    "down_path_1_block_4_num_batches_tracked": "PTConstNoopMetatype",
+    "down_path_2_block_1_running_mean": "PTConstNoopMetatype",
+    "down_path_2_block_1_running_var": "PTConstNoopMetatype",
+    "down_path_2_block_1_num_batches_tracked": "PTConstNoopMetatype",
+    "down_path_2_block_4_running_mean": "PTConstNoopMetatype",
+    "down_path_2_block_4_running_var": "PTConstNoopMetatype",
+    "down_path_2_block_4_num_batches_tracked": "PTConstNoopMetatype",
+    "down_path_3_block_1_running_mean": "PTConstNoopMetatype",
+    "down_path_3_block_1_running_var": "PTConstNoopMetatype",
+    "down_path_3_block_1_num_batches_tracked": "PTConstNoopMetatype",
+    "down_path_3_block_4_running_mean": "PTConstNoopMetatype",
+    "down_path_3_block_4_running_var": "PTConstNoopMetatype",
+    "down_path_3_block_4_num_batches_tracked": "PTConstNoopMetatype",
+    "down_path_4_block_1_running_mean": "PTConstNoopMetatype",
+    "down_path_4_block_1_running_var": "PTConstNoopMetatype",
+    "down_path_4_block_1_num_batches_tracked": "PTConstNoopMetatype",
+    "down_path_4_block_4_running_mean": "PTConstNoopMetatype",
+    "down_path_4_block_4_running_var": "PTConstNoopMetatype",
+    "down_path_4_block_4_num_batches_tracked": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_1_running_mean": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_1_running_var": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_1_num_batches_tracked": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_4_running_mean": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_4_running_var": "PTConstNoopMetatype",
+    "up_path_0_conv_block_block_4_num_batches_tracked": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_1_running_mean": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_1_running_var": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_1_num_batches_tracked": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_4_running_mean": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_4_running_var": "PTConstNoopMetatype",
+    "up_path_1_conv_block_block_4_num_batches_tracked": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_1_running_mean": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_1_running_var": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_1_num_batches_tracked": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_4_running_mean": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_4_running_var": "PTConstNoopMetatype",
+    "up_path_2_conv_block_block_4_num_batches_tracked": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_1_running_mean": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_1_running_var": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_1_num_batches_tracked": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_4_running_mean": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_4_running_var": "PTConstNoopMetatype",
+    "up_path_3_conv_block_block_4_num_batches_tracked": "PTConstNoopMetatype",
+    "x": "PTInputNoopMetatype",
     "conv2d": "PTConv2dMetatype",
-    "empty": "UnknownMetatype",
-    "_param_constant2": "PTConstNoopMetatype",
-    "_param_constant3": "PTConstNoopMetatype",
-    "_tensor_constant0": "PTConstNoopMetatype",
-    "_tensor_constant1": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training": "PTBatchNormMetatype",
-    "getitem": "PTGatherMetatype",
-    "getitem_1": "PTGatherMetatype",
-    "getitem_2": "PTGatherMetatype",
+    "batch_norm": "PTBatchNormMetatype",
     "relu": "PTRELUMetatype",
-    "_param_constant4": "PTConstNoopMetatype",
-    "_param_constant5": "PTConstNoopMetatype",
     "conv2d_1": "PTConv2dMetatype",
-    "empty_1": "UnknownMetatype",
-    "_param_constant6": "PTConstNoopMetatype",
-    "_param_constant7": "PTConstNoopMetatype",
-    "_tensor_constant2": "PTConstNoopMetatype",
-    "_tensor_constant3": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_1": "PTBatchNormMetatype",
-    "getitem_3": "PTGatherMetatype",
-    "getitem_4": "PTGatherMetatype",
-    "getitem_5": "PTGatherMetatype",
+    "batch_norm_1": "PTBatchNormMetatype",
     "relu_1": "PTRELUMetatype",
     "max_pool2d": "PTMaxPool2dMetatype",
-    "_param_constant8": "PTConstNoopMetatype",
-    "_param_constant9": "PTConstNoopMetatype",
     "conv2d_2": "PTConv2dMetatype",
-    "empty_2": "UnknownMetatype",
-    "_param_constant10": "PTConstNoopMetatype",
-    "_param_constant11": "PTConstNoopMetatype",
-    "_tensor_constant4": "PTConstNoopMetatype",
-    "_tensor_constant5": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_2": "PTBatchNormMetatype",
-    "getitem_6": "PTGatherMetatype",
-    "getitem_7": "PTGatherMetatype",
-    "getitem_8": "PTGatherMetatype",
+    "batch_norm_2": "PTBatchNormMetatype",
     "relu_2": "PTRELUMetatype",
-    "_param_constant12": "PTConstNoopMetatype",
-    "_param_constant13": "PTConstNoopMetatype",
     "conv2d_3": "PTConv2dMetatype",
-    "empty_3": "UnknownMetatype",
-    "_param_constant14": "PTConstNoopMetatype",
-    "_param_constant15": "PTConstNoopMetatype",
-    "_tensor_constant6": "PTConstNoopMetatype",
-    "_tensor_constant7": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_3": "PTBatchNormMetatype",
-    "getitem_9": "PTGatherMetatype",
-    "getitem_10": "PTGatherMetatype",
-    "getitem_11": "PTGatherMetatype",
+    "batch_norm_3": "PTBatchNormMetatype",
     "relu_3": "PTRELUMetatype",
     "max_pool2d_1": "PTMaxPool2dMetatype",
-    "_param_constant16": "PTConstNoopMetatype",
-    "_param_constant17": "PTConstNoopMetatype",
     "conv2d_4": "PTConv2dMetatype",
-    "empty_4": "UnknownMetatype",
-    "_param_constant18": "PTConstNoopMetatype",
-    "_param_constant19": "PTConstNoopMetatype",
-    "_tensor_constant8": "PTConstNoopMetatype",
-    "_tensor_constant9": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_4": "PTBatchNormMetatype",
-    "getitem_12": "PTGatherMetatype",
-    "getitem_13": "PTGatherMetatype",
-    "getitem_14": "PTGatherMetatype",
+    "batch_norm_4": "PTBatchNormMetatype",
     "relu_4": "PTRELUMetatype",
-    "_param_constant20": "PTConstNoopMetatype",
-    "_param_constant21": "PTConstNoopMetatype",
     "conv2d_5": "PTConv2dMetatype",
-    "empty_5": "UnknownMetatype",
-    "_param_constant22": "PTConstNoopMetatype",
-    "_param_constant23": "PTConstNoopMetatype",
-    "_tensor_constant10": "PTConstNoopMetatype",
-    "_tensor_constant11": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_5": "PTBatchNormMetatype",
-    "getitem_15": "PTGatherMetatype",
-    "getitem_16": "PTGatherMetatype",
-    "getitem_17": "PTGatherMetatype",
+    "batch_norm_5": "PTBatchNormMetatype",
     "relu_5": "PTRELUMetatype",
     "max_pool2d_2": "PTMaxPool2dMetatype",
-    "_param_constant24": "PTConstNoopMetatype",
-    "_param_constant25": "PTConstNoopMetatype",
     "conv2d_6": "PTConv2dMetatype",
-    "empty_6": "UnknownMetatype",
-    "_param_constant26": "PTConstNoopMetatype",
-    "_param_constant27": "PTConstNoopMetatype",
-    "_tensor_constant12": "PTConstNoopMetatype",
-    "_tensor_constant13": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_6": "PTBatchNormMetatype",
-    "getitem_18": "PTGatherMetatype",
-    "getitem_19": "PTGatherMetatype",
-    "getitem_20": "PTGatherMetatype",
+    "batch_norm_6": "PTBatchNormMetatype",
     "relu_6": "PTRELUMetatype",
-    "_param_constant28": "PTConstNoopMetatype",
-    "_param_constant29": "PTConstNoopMetatype",
     "conv2d_7": "PTConv2dMetatype",
-    "empty_7": "UnknownMetatype",
-    "_param_constant30": "PTConstNoopMetatype",
-    "_param_constant31": "PTConstNoopMetatype",
-    "_tensor_constant14": "PTConstNoopMetatype",
-    "_tensor_constant15": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_7": "PTBatchNormMetatype",
-    "getitem_21": "PTGatherMetatype",
-    "getitem_22": "PTGatherMetatype",
-    "getitem_23": "PTGatherMetatype",
+    "batch_norm_7": "PTBatchNormMetatype",
     "relu_7": "PTRELUMetatype",
     "max_pool2d_3": "PTMaxPool2dMetatype",
-    "_param_constant32": "PTConstNoopMetatype",
-    "_param_constant33": "PTConstNoopMetatype",
     "conv2d_8": "PTConv2dMetatype",
-    "empty_8": "UnknownMetatype",
-    "_param_constant34": "PTConstNoopMetatype",
-    "_param_constant35": "PTConstNoopMetatype",
-    "_tensor_constant16": "PTConstNoopMetatype",
-    "_tensor_constant17": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_8": "PTBatchNormMetatype",
-    "getitem_24": "PTGatherMetatype",
-    "getitem_25": "PTGatherMetatype",
-    "getitem_26": "PTGatherMetatype",
+    "batch_norm_8": "PTBatchNormMetatype",
     "relu_8": "PTRELUMetatype",
-    "_param_constant36": "PTConstNoopMetatype",
-    "_param_constant37": "PTConstNoopMetatype",
     "conv2d_9": "PTConv2dMetatype",
-    "empty_9": "UnknownMetatype",
-    "_param_constant38": "PTConstNoopMetatype",
-    "_param_constant39": "PTConstNoopMetatype",
-    "_tensor_constant18": "PTConstNoopMetatype",
-    "_tensor_constant19": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_9": "PTBatchNormMetatype",
-    "getitem_27": "PTGatherMetatype",
-    "getitem_28": "PTGatherMetatype",
-    "getitem_29": "PTGatherMetatype",
+    "batch_norm_9": "PTBatchNormMetatype",
     "relu_9": "PTRELUMetatype",
-    "_param_constant40": "PTConstNoopMetatype",
-    "_param_constant41": "PTConstNoopMetatype",
     "conv_transpose2d": "PTConvTranspose2dMetatype",
     "slice_1": "PTGatherMetatype",
     "slice_2": "PTGatherMetatype",
     "slice_3": "PTGatherMetatype",
     "slice_4": "PTGatherMetatype",
     "cat": "PTCatMetatype",
-    "_param_constant42": "PTConstNoopMetatype",
-    "_param_constant43": "PTConstNoopMetatype",
     "conv2d_10": "PTConv2dMetatype",
-    "empty_10": "UnknownMetatype",
-    "_param_constant44": "PTConstNoopMetatype",
-    "_param_constant45": "PTConstNoopMetatype",
-    "_tensor_constant20": "PTConstNoopMetatype",
-    "_tensor_constant21": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_10": "PTBatchNormMetatype",
-    "getitem_30": "PTGatherMetatype",
-    "getitem_31": "PTGatherMetatype",
-    "getitem_32": "PTGatherMetatype",
+    "batch_norm_10": "PTBatchNormMetatype",
     "relu_10": "PTRELUMetatype",
-    "_param_constant46": "PTConstNoopMetatype",
-    "_param_constant47": "PTConstNoopMetatype",
     "conv2d_11": "PTConv2dMetatype",
-    "empty_11": "UnknownMetatype",
-    "_param_constant48": "PTConstNoopMetatype",
-    "_param_constant49": "PTConstNoopMetatype",
-    "_tensor_constant22": "PTConstNoopMetatype",
-    "_tensor_constant23": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_11": "PTBatchNormMetatype",
-    "getitem_33": "PTGatherMetatype",
-    "getitem_34": "PTGatherMetatype",
-    "getitem_35": "PTGatherMetatype",
+    "batch_norm_11": "PTBatchNormMetatype",
     "relu_11": "PTRELUMetatype",
-    "_param_constant50": "PTConstNoopMetatype",
-    "_param_constant51": "PTConstNoopMetatype",
     "conv_transpose2d_1": "PTConvTranspose2dMetatype",
     "slice_5": "PTGatherMetatype",
     "slice_6": "PTGatherMetatype",
     "slice_7": "PTGatherMetatype",
     "slice_8": "PTGatherMetatype",
     "cat_1": "PTCatMetatype",
-    "_param_constant52": "PTConstNoopMetatype",
-    "_param_constant53": "PTConstNoopMetatype",
     "conv2d_12": "PTConv2dMetatype",
-    "empty_12": "UnknownMetatype",
-    "_param_constant54": "PTConstNoopMetatype",
-    "_param_constant55": "PTConstNoopMetatype",
-    "_tensor_constant24": "PTConstNoopMetatype",
-    "_tensor_constant25": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_12": "PTBatchNormMetatype",
-    "getitem_36": "PTGatherMetatype",
-    "getitem_37": "PTGatherMetatype",
-    "getitem_38": "PTGatherMetatype",
+    "batch_norm_12": "PTBatchNormMetatype",
     "relu_12": "PTRELUMetatype",
-    "_param_constant56": "PTConstNoopMetatype",
-    "_param_constant57": "PTConstNoopMetatype",
     "conv2d_13": "PTConv2dMetatype",
-    "empty_13": "UnknownMetatype",
-    "_param_constant58": "PTConstNoopMetatype",
-    "_param_constant59": "PTConstNoopMetatype",
-    "_tensor_constant26": "PTConstNoopMetatype",
-    "_tensor_constant27": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_13": "PTBatchNormMetatype",
-    "getitem_39": "PTGatherMetatype",
-    "getitem_40": "PTGatherMetatype",
-    "getitem_41": "PTGatherMetatype",
+    "batch_norm_13": "PTBatchNormMetatype",
     "relu_13": "PTRELUMetatype",
-    "_param_constant60": "PTConstNoopMetatype",
-    "_param_constant61": "PTConstNoopMetatype",
     "conv_transpose2d_2": "PTConvTranspose2dMetatype",
     "slice_9": "PTGatherMetatype",
     "slice_10": "PTGatherMetatype",
     "slice_11": "PTGatherMetatype",
     "slice_12": "PTGatherMetatype",
     "cat_2": "PTCatMetatype",
-    "_param_constant62": "PTConstNoopMetatype",
-    "_param_constant63": "PTConstNoopMetatype",
     "conv2d_14": "PTConv2dMetatype",
-    "empty_14": "UnknownMetatype",
-    "_param_constant64": "PTConstNoopMetatype",
-    "_param_constant65": "PTConstNoopMetatype",
-    "_tensor_constant28": "PTConstNoopMetatype",
-    "_tensor_constant29": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_14": "PTBatchNormMetatype",
-    "getitem_42": "PTGatherMetatype",
-    "getitem_43": "PTGatherMetatype",
-    "getitem_44": "PTGatherMetatype",
+    "batch_norm_14": "PTBatchNormMetatype",
     "relu_14": "PTRELUMetatype",
-    "_param_constant66": "PTConstNoopMetatype",
-    "_param_constant67": "PTConstNoopMetatype",
     "conv2d_15": "PTConv2dMetatype",
-    "empty_15": "UnknownMetatype",
-    "_param_constant68": "PTConstNoopMetatype",
-    "_param_constant69": "PTConstNoopMetatype",
-    "_tensor_constant30": "PTConstNoopMetatype",
-    "_tensor_constant31": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_15": "PTBatchNormMetatype",
-    "getitem_45": "PTGatherMetatype",
-    "getitem_46": "PTGatherMetatype",
-    "getitem_47": "PTGatherMetatype",
+    "batch_norm_15": "PTBatchNormMetatype",
     "relu_15": "PTRELUMetatype",
-    "_param_constant70": "PTConstNoopMetatype",
-    "_param_constant71": "PTConstNoopMetatype",
     "conv_transpose2d_3": "PTConvTranspose2dMetatype",
     "slice_13": "PTGatherMetatype",
     "slice_14": "PTGatherMetatype",
     "slice_15": "PTGatherMetatype",
     "slice_16": "PTGatherMetatype",
     "cat_3": "PTCatMetatype",
-    "_param_constant72": "PTConstNoopMetatype",
-    "_param_constant73": "PTConstNoopMetatype",
     "conv2d_16": "PTConv2dMetatype",
-    "empty_16": "UnknownMetatype",
-    "_param_constant74": "PTConstNoopMetatype",
-    "_param_constant75": "PTConstNoopMetatype",
-    "_tensor_constant32": "PTConstNoopMetatype",
-    "_tensor_constant33": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_16": "PTBatchNormMetatype",
-    "getitem_48": "PTGatherMetatype",
-    "getitem_49": "PTGatherMetatype",
-    "getitem_50": "PTGatherMetatype",
+    "batch_norm_16": "PTBatchNormMetatype",
     "relu_16": "PTRELUMetatype",
-    "_param_constant76": "PTConstNoopMetatype",
-    "_param_constant77": "PTConstNoopMetatype",
     "conv2d_17": "PTConv2dMetatype",
-    "empty_17": "UnknownMetatype",
-    "_param_constant78": "PTConstNoopMetatype",
-    "_param_constant79": "PTConstNoopMetatype",
-    "_tensor_constant34": "PTConstNoopMetatype",
-    "_tensor_constant35": "PTConstNoopMetatype",
-    "_native_batch_norm_legit_no_training_17": "PTBatchNormMetatype",
-    "getitem_51": "PTGatherMetatype",
-    "getitem_52": "PTGatherMetatype",
-    "getitem_53": "PTGatherMetatype",
+    "batch_norm_17": "PTBatchNormMetatype",
     "relu_17": "PTRELUMetatype",
-    "_param_constant80": "PTConstNoopMetatype",
-    "_param_constant81": "PTConstNoopMetatype",
     "conv2d_18": "PTConv2dMetatype",
-    "output": "PTOutputNoopMetatype"
+    "output_1": "PTOutputNoopMetatype"
 }
\ No newline at end of file
diff --git a/tests/torch/data/reference_graphs/fx/reference_metatypes/vit_b_16.json b/tests/torch/data/reference_graphs/fx/reference_metatypes/vit_b_16.json
index 2b9eb7fa783..42484583d7a 100644
--- a/tests/torch/data/reference_graphs/fx/reference_metatypes/vit_b_16.json
+++ b/tests/torch/data/reference_graphs/fx/reference_metatypes/vit_b_16.json
@@ -1,22 +1,166 @@
 {
-    "arg0_1": "PTInputNoopMetatype",
-    "_param_constant0": "PTConstNoopMetatype",
-    "_param_constant1": "PTConstNoopMetatype",
+    "class_token": "PTConstNoopMetatype",
+    "encoder_pos_embedding": "PTConstNoopMetatype",
+    "conv_proj_weight": "PTConstNoopMetatype",
+    "conv_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_0_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_1_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_2_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_3_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_4_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_5_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_6_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_7_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_8_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_9_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_10_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_ln_1_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_ln_1_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_self_attention_in_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_self_attention_in_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_self_attention_out_proj_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_self_attention_out_proj_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_ln_2_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_ln_2_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_mlp_0_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_mlp_0_bias": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_mlp_3_weight": "PTConstNoopMetatype",
+    "encoder_layers_encoder_layer_11_mlp_3_bias": "PTConstNoopMetatype",
+    "encoder_ln_weight": "PTConstNoopMetatype",
+    "encoder_ln_bias": "PTConstNoopMetatype",
+    "heads_head_weight": "PTConstNoopMetatype",
+    "heads_head_bias": "PTConstNoopMetatype",
+    "x": "PTInputNoopMetatype",
     "conv2d": "PTConv2dMetatype",
     "reshape": "PTReshapeMetatype",
     "permute": "PTTransposeMetatype",
-    "_param_constant2": "PTConstNoopMetatype",
     "expand": "PTExpandMetatype",
     "cat": "PTCatMetatype",
-    "_param_constant3": "PTConstNoopMetatype",
     "add": "PTAddMetatype",
     "dropout": "PTDropoutMetatype",
-    "_param_constant4": "PTConstNoopMetatype",
-    "_param_constant5": "PTConstNoopMetatype",
     "layer_norm": "PTLayerNormMetatype",
     "transpose": "PTTransposeMetatype",
-    "_param_constant6": "PTConstNoopMetatype",
-    "_param_constant7": "PTConstNoopMetatype",
     "linear": "PTLinearMetatype",
     "unflatten": "PTReshapeMetatype",
     "unsqueeze": "PTReshapeMetatype",
@@ -38,32 +182,20 @@
     "scaled_dot_product_attention": "PTScaledDotProductAttentionMetatype",
     "permute_1": "PTTransposeMetatype",
     "view_6": "PTReshapeMetatype",
-    "_param_constant8": "PTConstNoopMetatype",
-    "_param_constant9": "PTConstNoopMetatype",
     "linear_1": "PTLinearMetatype",
     "view_7": "PTReshapeMetatype",
     "transpose_5": "PTTransposeMetatype",
     "dropout_1": "PTDropoutMetatype",
     "add_1": "PTAddMetatype",
-    "_param_constant10": "PTConstNoopMetatype",
-    "_param_constant11": "PTConstNoopMetatype",
     "layer_norm_1": "PTLayerNormMetatype",
-    "_param_constant12": "PTConstNoopMetatype",
-    "_param_constant13": "PTConstNoopMetatype",
     "linear_2": "PTLinearMetatype",
     "gelu": "PTGELUMetatype",
     "dropout_2": "PTDropoutMetatype",
-    "_param_constant14": "PTConstNoopMetatype",
-    "_param_constant15": "PTConstNoopMetatype",
     "linear_3": "PTLinearMetatype",
     "dropout_3": "PTDropoutMetatype",
     "add_2": "PTAddMetatype",
-    "_param_constant16": "PTConstNoopMetatype",
-    "_param_constant17": "PTConstNoopMetatype",
     "layer_norm_2": "PTLayerNormMetatype",
     "transpose_6": "PTTransposeMetatype",
-    "_param_constant18": "PTConstNoopMetatype",
-    "_param_constant19": "PTConstNoopMetatype",
     "linear_4": "PTLinearMetatype",
     "unflatten_1": "PTReshapeMetatype",
     "unsqueeze_1": "PTReshapeMetatype",
@@ -85,32 +217,20 @@
     "scaled_dot_product_attention_1": "PTScaledDotProductAttentionMetatype",
     "permute_2": "PTTransposeMetatype",
     "view_14": "PTReshapeMetatype",
-    "_param_constant20": "PTConstNoopMetatype",
-    "_param_constant21": "PTConstNoopMetatype",
     "linear_5": "PTLinearMetatype",
     "view_15": "PTReshapeMetatype",
     "transpose_11": "PTTransposeMetatype",
     "dropout_4": "PTDropoutMetatype",
     "add_3": "PTAddMetatype",
-    "_param_constant22": "PTConstNoopMetatype",
-    "_param_constant23": "PTConstNoopMetatype",
     "layer_norm_3": "PTLayerNormMetatype",
-    "_param_constant24": "PTConstNoopMetatype",
-    "_param_constant25": "PTConstNoopMetatype",
     "linear_6": "PTLinearMetatype",
     "gelu_1": "PTGELUMetatype",
     "dropout_5": "PTDropoutMetatype",
-    "_param_constant26": "PTConstNoopMetatype",
-    "_param_constant27": "PTConstNoopMetatype",
     "linear_7": "PTLinearMetatype",
     "dropout_6": "PTDropoutMetatype",
     "add_4": "PTAddMetatype",
-    "_param_constant28": "PTConstNoopMetatype",
-    "_param_constant29": "PTConstNoopMetatype",
     "layer_norm_4": "PTLayerNormMetatype",
     "transpose_12": "PTTransposeMetatype",
-    "_param_constant30": "PTConstNoopMetatype",
-    "_param_constant31": "PTConstNoopMetatype",
     "linear_8": "PTLinearMetatype",
     "unflatten_2": "PTReshapeMetatype",
     "unsqueeze_2": "PTReshapeMetatype",
@@ -132,32 +252,20 @@
     "scaled_dot_product_attention_2": "PTScaledDotProductAttentionMetatype",
     "permute_3": "PTTransposeMetatype",
     "view_22": "PTReshapeMetatype",
-    "_param_constant32": "PTConstNoopMetatype",
-    "_param_constant33": "PTConstNoopMetatype",
     "linear_9": "PTLinearMetatype",
     "view_23": "PTReshapeMetatype",
     "transpose_17": "PTTransposeMetatype",
     "dropout_7": "PTDropoutMetatype",
     "add_5": "PTAddMetatype",
-    "_param_constant34": "PTConstNoopMetatype",
-    "_param_constant35": "PTConstNoopMetatype",
     "layer_norm_5": "PTLayerNormMetatype",
-    "_param_constant36": "PTConstNoopMetatype",
-    "_param_constant37": "PTConstNoopMetatype",
     "linear_10": "PTLinearMetatype",
     "gelu_2": "PTGELUMetatype",
     "dropout_8": "PTDropoutMetatype",
-    "_param_constant38": "PTConstNoopMetatype",
-    "_param_constant39": "PTConstNoopMetatype",
     "linear_11": "PTLinearMetatype",
     "dropout_9": "PTDropoutMetatype",
     "add_6": "PTAddMetatype",
-    "_param_constant40": "PTConstNoopMetatype",
-    "_param_constant41": "PTConstNoopMetatype",
     "layer_norm_6": "PTLayerNormMetatype",
     "transpose_18": "PTTransposeMetatype",
-    "_param_constant42": "PTConstNoopMetatype",
-    "_param_constant43": "PTConstNoopMetatype",
     "linear_12": "PTLinearMetatype",
     "unflatten_3": "PTReshapeMetatype",
     "unsqueeze_3": "PTReshapeMetatype",
@@ -179,32 +287,20 @@
     "scaled_dot_product_attention_3": "PTScaledDotProductAttentionMetatype",
     "permute_4": "PTTransposeMetatype",
     "view_30": "PTReshapeMetatype",
-    "_param_constant44": "PTConstNoopMetatype",
-    "_param_constant45": "PTConstNoopMetatype",
     "linear_13": "PTLinearMetatype",
     "view_31": "PTReshapeMetatype",
     "transpose_23": "PTTransposeMetatype",
     "dropout_10": "PTDropoutMetatype",
     "add_7": "PTAddMetatype",
-    "_param_constant46": "PTConstNoopMetatype",
-    "_param_constant47": "PTConstNoopMetatype",
     "layer_norm_7": "PTLayerNormMetatype",
-    "_param_constant48": "PTConstNoopMetatype",
-    "_param_constant49": "PTConstNoopMetatype",
     "linear_14": "PTLinearMetatype",
     "gelu_3": "PTGELUMetatype",
     "dropout_11": "PTDropoutMetatype",
-    "_param_constant50": "PTConstNoopMetatype",
-    "_param_constant51": "PTConstNoopMetatype",
     "linear_15": "PTLinearMetatype",
     "dropout_12": "PTDropoutMetatype",
     "add_8": "PTAddMetatype",
-    "_param_constant52": "PTConstNoopMetatype",
-    "_param_constant53": "PTConstNoopMetatype",
     "layer_norm_8": "PTLayerNormMetatype",
     "transpose_24": "PTTransposeMetatype",
-    "_param_constant54": "PTConstNoopMetatype",
-    "_param_constant55": "PTConstNoopMetatype",
     "linear_16": "PTLinearMetatype",
     "unflatten_4": "PTReshapeMetatype",
     "unsqueeze_4": "PTReshapeMetatype",
@@ -226,32 +322,20 @@
     "scaled_dot_product_attention_4": "PTScaledDotProductAttentionMetatype",
     "permute_5": "PTTransposeMetatype",
     "view_38": "PTReshapeMetatype",
-    "_param_constant56": "PTConstNoopMetatype",
-    "_param_constant57": "PTConstNoopMetatype",
     "linear_17": "PTLinearMetatype",
     "view_39": "PTReshapeMetatype",
     "transpose_29": "PTTransposeMetatype",
     "dropout_13": "PTDropoutMetatype",
     "add_9": "PTAddMetatype",
-    "_param_constant58": "PTConstNoopMetatype",
-    "_param_constant59": "PTConstNoopMetatype",
     "layer_norm_9": "PTLayerNormMetatype",
-    "_param_constant60": "PTConstNoopMetatype",
-    "_param_constant61": "PTConstNoopMetatype",
     "linear_18": "PTLinearMetatype",
     "gelu_4": "PTGELUMetatype",
     "dropout_14": "PTDropoutMetatype",
-    "_param_constant62": "PTConstNoopMetatype",
-    "_param_constant63": "PTConstNoopMetatype",
     "linear_19": "PTLinearMetatype",
     "dropout_15": "PTDropoutMetatype",
     "add_10": "PTAddMetatype",
-    "_param_constant64": "PTConstNoopMetatype",
-    "_param_constant65": "PTConstNoopMetatype",
     "layer_norm_10": "PTLayerNormMetatype",
     "transpose_30": "PTTransposeMetatype",
-    "_param_constant66": "PTConstNoopMetatype",
-    "_param_constant67": "PTConstNoopMetatype",
     "linear_20": "PTLinearMetatype",
     "unflatten_5": "PTReshapeMetatype",
     "unsqueeze_5": "PTReshapeMetatype",
@@ -273,32 +357,20 @@
     "scaled_dot_product_attention_5": "PTScaledDotProductAttentionMetatype",
     "permute_6": "PTTransposeMetatype",
     "view_46": "PTReshapeMetatype",
-    "_param_constant68": "PTConstNoopMetatype",
-    "_param_constant69": "PTConstNoopMetatype",
     "linear_21": "PTLinearMetatype",
     "view_47": "PTReshapeMetatype",
     "transpose_35": "PTTransposeMetatype",
     "dropout_16": "PTDropoutMetatype",
     "add_11": "PTAddMetatype",
-    "_param_constant70": "PTConstNoopMetatype",
-    "_param_constant71": "PTConstNoopMetatype",
     "layer_norm_11": "PTLayerNormMetatype",
-    "_param_constant72": "PTConstNoopMetatype",
-    "_param_constant73": "PTConstNoopMetatype",
     "linear_22": "PTLinearMetatype",
     "gelu_5": "PTGELUMetatype",
     "dropout_17": "PTDropoutMetatype",
-    "_param_constant74": "PTConstNoopMetatype",
-    "_param_constant75": "PTConstNoopMetatype",
     "linear_23": "PTLinearMetatype",
     "dropout_18": "PTDropoutMetatype",
     "add_12": "PTAddMetatype",
-    "_param_constant76": "PTConstNoopMetatype",
-    "_param_constant77": "PTConstNoopMetatype",
     "layer_norm_12": "PTLayerNormMetatype",
     "transpose_36": "PTTransposeMetatype",
-    "_param_constant78": "PTConstNoopMetatype",
-    "_param_constant79": "PTConstNoopMetatype",
     "linear_24": "PTLinearMetatype",
     "unflatten_6": "PTReshapeMetatype",
     "unsqueeze_6": "PTReshapeMetatype",
@@ -320,32 +392,20 @@
     "scaled_dot_product_attention_6": "PTScaledDotProductAttentionMetatype",
     "permute_7": "PTTransposeMetatype",
     "view_54": "PTReshapeMetatype",
-    "_param_constant80": "PTConstNoopMetatype",
-    "_param_constant81": "PTConstNoopMetatype",
     "linear_25": "PTLinearMetatype",
     "view_55": "PTReshapeMetatype",
     "transpose_41": "PTTransposeMetatype",
     "dropout_19": "PTDropoutMetatype",
     "add_13": "PTAddMetatype",
-    "_param_constant82": "PTConstNoopMetatype",
-    "_param_constant83": "PTConstNoopMetatype",
     "layer_norm_13": "PTLayerNormMetatype",
-    "_param_constant84": "PTConstNoopMetatype",
-    "_param_constant85": "PTConstNoopMetatype",
     "linear_26": "PTLinearMetatype",
     "gelu_6": "PTGELUMetatype",
     "dropout_20": "PTDropoutMetatype",
-    "_param_constant86": "PTConstNoopMetatype",
-    "_param_constant87": "PTConstNoopMetatype",
     "linear_27": "PTLinearMetatype",
     "dropout_21": "PTDropoutMetatype",
     "add_14": "PTAddMetatype",
-    "_param_constant88": "PTConstNoopMetatype",
-    "_param_constant89": "PTConstNoopMetatype",
     "layer_norm_14": "PTLayerNormMetatype",
     "transpose_42": "PTTransposeMetatype",
-    "_param_constant90": "PTConstNoopMetatype",
-    "_param_constant91": "PTConstNoopMetatype",
     "linear_28": "PTLinearMetatype",
     "unflatten_7": "PTReshapeMetatype",
     "unsqueeze_7": "PTReshapeMetatype",
@@ -367,32 +427,20 @@
     "scaled_dot_product_attention_7": "PTScaledDotProductAttentionMetatype",
     "permute_8": "PTTransposeMetatype",
     "view_62": "PTReshapeMetatype",
-    "_param_constant92": "PTConstNoopMetatype",
-    "_param_constant93": "PTConstNoopMetatype",
     "linear_29": "PTLinearMetatype",
     "view_63": "PTReshapeMetatype",
     "transpose_47": "PTTransposeMetatype",
     "dropout_22": "PTDropoutMetatype",
     "add_15": "PTAddMetatype",
-    "_param_constant94": "PTConstNoopMetatype",
-    "_param_constant95": "PTConstNoopMetatype",
     "layer_norm_15": "PTLayerNormMetatype",
-    "_param_constant96": "PTConstNoopMetatype",
-    "_param_constant97": "PTConstNoopMetatype",
     "linear_30": "PTLinearMetatype",
     "gelu_7": "PTGELUMetatype",
     "dropout_23": "PTDropoutMetatype",
-    "_param_constant98": "PTConstNoopMetatype",
-    "_param_constant99": "PTConstNoopMetatype",
     "linear_31": "PTLinearMetatype",
     "dropout_24": "PTDropoutMetatype",
     "add_16": "PTAddMetatype",
-    "_param_constant100": "PTConstNoopMetatype",
-    "_param_constant101": "PTConstNoopMetatype",
     "layer_norm_16": "PTLayerNormMetatype",
     "transpose_48": "PTTransposeMetatype",
-    "_param_constant102": "PTConstNoopMetatype",
-    "_param_constant103": "PTConstNoopMetatype",
     "linear_32": "PTLinearMetatype",
     "unflatten_8": "PTReshapeMetatype",
     "unsqueeze_8": "PTReshapeMetatype",
@@ -414,32 +462,20 @@
     "scaled_dot_product_attention_8": "PTScaledDotProductAttentionMetatype",
     "permute_9": "PTTransposeMetatype",
     "view_70": "PTReshapeMetatype",
-    "_param_constant104": "PTConstNoopMetatype",
-    "_param_constant105": "PTConstNoopMetatype",
     "linear_33": "PTLinearMetatype",
     "view_71": "PTReshapeMetatype",
     "transpose_53": "PTTransposeMetatype",
     "dropout_25": "PTDropoutMetatype",
     "add_17": "PTAddMetatype",
-    "_param_constant106": "PTConstNoopMetatype",
-    "_param_constant107": "PTConstNoopMetatype",
     "layer_norm_17": "PTLayerNormMetatype",
-    "_param_constant108": "PTConstNoopMetatype",
-    "_param_constant109": "PTConstNoopMetatype",
     "linear_34": "PTLinearMetatype",
     "gelu_8": "PTGELUMetatype",
     "dropout_26": "PTDropoutMetatype",
-    "_param_constant110": "PTConstNoopMetatype",
-    "_param_constant111": "PTConstNoopMetatype",
     "linear_35": "PTLinearMetatype",
     "dropout_27": "PTDropoutMetatype",
     "add_18": "PTAddMetatype",
-    "_param_constant112": "PTConstNoopMetatype",
-    "_param_constant113": "PTConstNoopMetatype",
     "layer_norm_18": "PTLayerNormMetatype",
     "transpose_54": "PTTransposeMetatype",
-    "_param_constant114": "PTConstNoopMetatype",
-    "_param_constant115": "PTConstNoopMetatype",
     "linear_36": "PTLinearMetatype",
     "unflatten_9": "PTReshapeMetatype",
     "unsqueeze_9": "PTReshapeMetatype",
@@ -461,32 +497,20 @@
     "scaled_dot_product_attention_9": "PTScaledDotProductAttentionMetatype",
     "permute_10": "PTTransposeMetatype",
     "view_78": "PTReshapeMetatype",
-    "_param_constant116": "PTConstNoopMetatype",
-    "_param_constant117": "PTConstNoopMetatype",
     "linear_37": "PTLinearMetatype",
     "view_79": "PTReshapeMetatype",
     "transpose_59": "PTTransposeMetatype",
     "dropout_28": "PTDropoutMetatype",
     "add_19": "PTAddMetatype",
-    "_param_constant118": "PTConstNoopMetatype",
-    "_param_constant119": "PTConstNoopMetatype",
     "layer_norm_19": "PTLayerNormMetatype",
-    "_param_constant120": "PTConstNoopMetatype",
-    "_param_constant121": "PTConstNoopMetatype",
     "linear_38": "PTLinearMetatype",
     "gelu_9": "PTGELUMetatype",
     "dropout_29": "PTDropoutMetatype",
-    "_param_constant122": "PTConstNoopMetatype",
-    "_param_constant123": "PTConstNoopMetatype",
     "linear_39": "PTLinearMetatype",
     "dropout_30": "PTDropoutMetatype",
     "add_20": "PTAddMetatype",
-    "_param_constant124": "PTConstNoopMetatype",
-    "_param_constant125": "PTConstNoopMetatype",
     "layer_norm_20": "PTLayerNormMetatype",
     "transpose_60": "PTTransposeMetatype",
-    "_param_constant126": "PTConstNoopMetatype",
-    "_param_constant127": "PTConstNoopMetatype",
     "linear_40": "PTLinearMetatype",
     "unflatten_10": "PTReshapeMetatype",
     "unsqueeze_10": "PTReshapeMetatype",
@@ -508,32 +532,20 @@
     "scaled_dot_product_attention_10": "PTScaledDotProductAttentionMetatype",
     "permute_11": "PTTransposeMetatype",
     "view_86": "PTReshapeMetatype",
-    "_param_constant128": "PTConstNoopMetatype",
-    "_param_constant129": "PTConstNoopMetatype",
     "linear_41": "PTLinearMetatype",
     "view_87": "PTReshapeMetatype",
     "transpose_65": "PTTransposeMetatype",
     "dropout_31": "PTDropoutMetatype",
     "add_21": "PTAddMetatype",
-    "_param_constant130": "PTConstNoopMetatype",
-    "_param_constant131": "PTConstNoopMetatype",
     "layer_norm_21": "PTLayerNormMetatype",
-    "_param_constant132": "PTConstNoopMetatype",
-    "_param_constant133": "PTConstNoopMetatype",
     "linear_42": "PTLinearMetatype",
     "gelu_10": "PTGELUMetatype",
     "dropout_32": "PTDropoutMetatype",
-    "_param_constant134": "PTConstNoopMetatype",
-    "_param_constant135": "PTConstNoopMetatype",
     "linear_43": "PTLinearMetatype",
     "dropout_33": "PTDropoutMetatype",
     "add_22": "PTAddMetatype",
-    "_param_constant136": "PTConstNoopMetatype",
-    "_param_constant137": "PTConstNoopMetatype",
     "layer_norm_22": "PTLayerNormMetatype",
     "transpose_66": "PTTransposeMetatype",
-    "_param_constant138": "PTConstNoopMetatype",
-    "_param_constant139": "PTConstNoopMetatype",
     "linear_44": "PTLinearMetatype",
     "unflatten_11": "PTReshapeMetatype",
     "unsqueeze_11": "PTReshapeMetatype",
@@ -555,33 +567,21 @@
     "scaled_dot_product_attention_11": "PTScaledDotProductAttentionMetatype",
     "permute_12": "PTTransposeMetatype",
     "view_94": "PTReshapeMetatype",
-    "_param_constant140": "PTConstNoopMetatype",
-    "_param_constant141": "PTConstNoopMetatype",
     "linear_45": "PTLinearMetatype",
     "view_95": "PTReshapeMetatype",
     "transpose_71": "PTTransposeMetatype",
     "dropout_34": "PTDropoutMetatype",
     "add_23": "PTAddMetatype",
-    "_param_constant142": "PTConstNoopMetatype",
-    "_param_constant143": "PTConstNoopMetatype",
     "layer_norm_23": "PTLayerNormMetatype",
-    "_param_constant144": "PTConstNoopMetatype",
-    "_param_constant145": "PTConstNoopMetatype",
     "linear_46": "PTLinearMetatype",
     "gelu_11": "PTGELUMetatype",
     "dropout_35": "PTDropoutMetatype",
-    "_param_constant146": "PTConstNoopMetatype",
-    "_param_constant147": "PTConstNoopMetatype",
     "linear_47": "PTLinearMetatype",
     "dropout_36": "PTDropoutMetatype",
     "add_24": "PTAddMetatype",
-    "_param_constant148": "PTConstNoopMetatype",
-    "_param_constant149": "PTConstNoopMetatype",
     "layer_norm_24": "PTLayerNormMetatype",
     "slice_1": "PTGatherMetatype",
     "select_36": "PTGatherMetatype",
-    "_param_constant150": "PTConstNoopMetatype",
-    "_param_constant151": "PTConstNoopMetatype",
     "linear_48": "PTLinearMetatype",
-    "output": "PTOutputNoopMetatype"
+    "output_1": "PTOutputNoopMetatype"
 }
\ No newline at end of file
diff --git a/tests/torch/data/reference_graphs/fx/resnet18.dot b/tests/torch/data/reference_graphs/fx/resnet18.dot
index 53a4ea32a8e..a566ccc6d7a 100644
--- a/tests/torch/data/reference_graphs/fx/resnet18.dot
+++ b/tests/torch/data/reference_graphs/fx/resnet18.dot
@@ -1,495 +1,375 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 conv2d" [id=2, type=conv2d];
-"3 empty" [id=3, type=empty];
-"4 _param_constant1" [id=4, type=get_attr];
-"5 _param_constant2" [id=5, type=get_attr];
-"6 _tensor_constant0" [id=6, type=get_attr];
-"7 _tensor_constant1" [id=7, type=get_attr];
-"8 _native_batch_norm_legit_no_training" [id=8, type=_native_batch_norm_legit_no_training];
-"9 getitem" [id=9, type=__getitem__];
-"10 getitem_1" [id=10, type=__getitem__];
-"11 getitem_2" [id=11, type=__getitem__];
-"12 relu_" [id=12, type=relu_];
-"13 max_pool2d" [id=13, type=max_pool2d];
-"14 _param_constant3" [id=14, type=get_attr];
-"15 conv2d_1" [id=15, type=conv2d];
-"16 empty_1" [id=16, type=empty];
-"17 _param_constant4" [id=17, type=get_attr];
-"18 _param_constant5" [id=18, type=get_attr];
-"19 _tensor_constant2" [id=19, type=get_attr];
-"20 _tensor_constant3" [id=20, type=get_attr];
-"21 _native_batch_norm_legit_no_training_1" [id=21, type=_native_batch_norm_legit_no_training];
-"22 getitem_3" [id=22, type=__getitem__];
-"23 getitem_4" [id=23, type=__getitem__];
-"24 getitem_5" [id=24, type=__getitem__];
-"25 relu__1" [id=25, type=relu_];
-"26 _param_constant6" [id=26, type=get_attr];
-"27 conv2d_2" [id=27, type=conv2d];
-"28 empty_2" [id=28, type=empty];
-"29 _param_constant7" [id=29, type=get_attr];
-"30 _param_constant8" [id=30, type=get_attr];
-"31 _tensor_constant4" [id=31, type=get_attr];
-"32 _tensor_constant5" [id=32, type=get_attr];
-"33 _native_batch_norm_legit_no_training_2" [id=33, type=_native_batch_norm_legit_no_training];
-"34 getitem_6" [id=34, type=__getitem__];
-"35 getitem_7" [id=35, type=__getitem__];
-"36 getitem_8" [id=36, type=__getitem__];
-"37 add_" [id=37, type=add_];
-"38 relu__2" [id=38, type=relu_];
-"39 _param_constant9" [id=39, type=get_attr];
-"40 conv2d_3" [id=40, type=conv2d];
-"41 empty_3" [id=41, type=empty];
-"42 _param_constant10" [id=42, type=get_attr];
-"43 _param_constant11" [id=43, type=get_attr];
-"44 _tensor_constant6" [id=44, type=get_attr];
-"45 _tensor_constant7" [id=45, type=get_attr];
-"46 _native_batch_norm_legit_no_training_3" [id=46, type=_native_batch_norm_legit_no_training];
-"47 getitem_9" [id=47, type=__getitem__];
-"48 getitem_10" [id=48, type=__getitem__];
-"49 getitem_11" [id=49, type=__getitem__];
-"50 relu__3" [id=50, type=relu_];
-"51 _param_constant12" [id=51, type=get_attr];
-"52 conv2d_4" [id=52, type=conv2d];
-"53 empty_4" [id=53, type=empty];
-"54 _param_constant13" [id=54, type=get_attr];
-"55 _param_constant14" [id=55, type=get_attr];
-"56 _tensor_constant8" [id=56, type=get_attr];
-"57 _tensor_constant9" [id=57, type=get_attr];
-"58 _native_batch_norm_legit_no_training_4" [id=58, type=_native_batch_norm_legit_no_training];
-"59 getitem_12" [id=59, type=__getitem__];
-"60 getitem_13" [id=60, type=__getitem__];
-"61 getitem_14" [id=61, type=__getitem__];
-"62 add__1" [id=62, type=add_];
-"63 relu__4" [id=63, type=relu_];
-"64 _param_constant15" [id=64, type=get_attr];
-"65 conv2d_5" [id=65, type=conv2d];
-"66 empty_5" [id=66, type=empty];
-"67 _param_constant16" [id=67, type=get_attr];
-"68 _param_constant17" [id=68, type=get_attr];
-"69 _tensor_constant10" [id=69, type=get_attr];
-"70 _tensor_constant11" [id=70, type=get_attr];
-"71 _native_batch_norm_legit_no_training_5" [id=71, type=_native_batch_norm_legit_no_training];
-"72 getitem_15" [id=72, type=__getitem__];
-"73 getitem_16" [id=73, type=__getitem__];
-"74 getitem_17" [id=74, type=__getitem__];
-"75 relu__5" [id=75, type=relu_];
-"76 _param_constant18" [id=76, type=get_attr];
-"77 conv2d_6" [id=77, type=conv2d];
-"78 empty_6" [id=78, type=empty];
-"79 _param_constant19" [id=79, type=get_attr];
-"80 _param_constant20" [id=80, type=get_attr];
-"81 _tensor_constant12" [id=81, type=get_attr];
-"82 _tensor_constant13" [id=82, type=get_attr];
-"83 _native_batch_norm_legit_no_training_6" [id=83, type=_native_batch_norm_legit_no_training];
-"84 getitem_18" [id=84, type=__getitem__];
-"85 getitem_19" [id=85, type=__getitem__];
-"86 getitem_20" [id=86, type=__getitem__];
-"87 _param_constant21" [id=87, type=get_attr];
-"88 conv2d_7" [id=88, type=conv2d];
-"89 empty_7" [id=89, type=empty];
-"90 _param_constant22" [id=90, type=get_attr];
-"91 _param_constant23" [id=91, type=get_attr];
-"92 _tensor_constant14" [id=92, type=get_attr];
-"93 _tensor_constant15" [id=93, type=get_attr];
-"94 _native_batch_norm_legit_no_training_7" [id=94, type=_native_batch_norm_legit_no_training];
-"95 getitem_21" [id=95, type=__getitem__];
-"96 getitem_22" [id=96, type=__getitem__];
-"97 getitem_23" [id=97, type=__getitem__];
-"98 add__2" [id=98, type=add_];
-"99 relu__6" [id=99, type=relu_];
-"100 _param_constant24" [id=100, type=get_attr];
-"101 conv2d_8" [id=101, type=conv2d];
-"102 empty_8" [id=102, type=empty];
-"103 _param_constant25" [id=103, type=get_attr];
-"104 _param_constant26" [id=104, type=get_attr];
-"105 _tensor_constant16" [id=105, type=get_attr];
-"106 _tensor_constant17" [id=106, type=get_attr];
-"107 _native_batch_norm_legit_no_training_8" [id=107, type=_native_batch_norm_legit_no_training];
-"108 getitem_24" [id=108, type=__getitem__];
-"109 getitem_25" [id=109, type=__getitem__];
-"110 getitem_26" [id=110, type=__getitem__];
-"111 relu__7" [id=111, type=relu_];
-"112 _param_constant27" [id=112, type=get_attr];
-"113 conv2d_9" [id=113, type=conv2d];
-"114 empty_9" [id=114, type=empty];
-"115 _param_constant28" [id=115, type=get_attr];
-"116 _param_constant29" [id=116, type=get_attr];
-"117 _tensor_constant18" [id=117, type=get_attr];
-"118 _tensor_constant19" [id=118, type=get_attr];
-"119 _native_batch_norm_legit_no_training_9" [id=119, type=_native_batch_norm_legit_no_training];
-"120 getitem_27" [id=120, type=__getitem__];
-"121 getitem_28" [id=121, type=__getitem__];
-"122 getitem_29" [id=122, type=__getitem__];
-"123 add__3" [id=123, type=add_];
-"124 relu__8" [id=124, type=relu_];
-"125 _param_constant30" [id=125, type=get_attr];
-"126 conv2d_10" [id=126, type=conv2d];
-"127 empty_10" [id=127, type=empty];
-"128 _param_constant31" [id=128, type=get_attr];
-"129 _param_constant32" [id=129, type=get_attr];
-"130 _tensor_constant20" [id=130, type=get_attr];
-"131 _tensor_constant21" [id=131, type=get_attr];
-"132 _native_batch_norm_legit_no_training_10" [id=132, type=_native_batch_norm_legit_no_training];
-"133 getitem_30" [id=133, type=__getitem__];
-"134 getitem_31" [id=134, type=__getitem__];
-"135 getitem_32" [id=135, type=__getitem__];
-"136 relu__9" [id=136, type=relu_];
-"137 _param_constant33" [id=137, type=get_attr];
-"138 conv2d_11" [id=138, type=conv2d];
-"139 empty_11" [id=139, type=empty];
-"140 _param_constant34" [id=140, type=get_attr];
-"141 _param_constant35" [id=141, type=get_attr];
-"142 _tensor_constant22" [id=142, type=get_attr];
-"143 _tensor_constant23" [id=143, type=get_attr];
-"144 _native_batch_norm_legit_no_training_11" [id=144, type=_native_batch_norm_legit_no_training];
-"145 getitem_33" [id=145, type=__getitem__];
-"146 getitem_34" [id=146, type=__getitem__];
-"147 getitem_35" [id=147, type=__getitem__];
-"148 _param_constant36" [id=148, type=get_attr];
-"149 conv2d_12" [id=149, type=conv2d];
-"150 empty_12" [id=150, type=empty];
-"151 _param_constant37" [id=151, type=get_attr];
-"152 _param_constant38" [id=152, type=get_attr];
-"153 _tensor_constant24" [id=153, type=get_attr];
-"154 _tensor_constant25" [id=154, type=get_attr];
-"155 _native_batch_norm_legit_no_training_12" [id=155, type=_native_batch_norm_legit_no_training];
-"156 getitem_36" [id=156, type=__getitem__];
-"157 getitem_37" [id=157, type=__getitem__];
-"158 getitem_38" [id=158, type=__getitem__];
-"159 add__4" [id=159, type=add_];
-"160 relu__10" [id=160, type=relu_];
-"161 _param_constant39" [id=161, type=get_attr];
-"162 conv2d_13" [id=162, type=conv2d];
-"163 empty_13" [id=163, type=empty];
-"164 _param_constant40" [id=164, type=get_attr];
-"165 _param_constant41" [id=165, type=get_attr];
-"166 _tensor_constant26" [id=166, type=get_attr];
-"167 _tensor_constant27" [id=167, type=get_attr];
-"168 _native_batch_norm_legit_no_training_13" [id=168, type=_native_batch_norm_legit_no_training];
-"169 getitem_39" [id=169, type=__getitem__];
-"170 getitem_40" [id=170, type=__getitem__];
-"171 getitem_41" [id=171, type=__getitem__];
-"172 relu__11" [id=172, type=relu_];
-"173 _param_constant42" [id=173, type=get_attr];
-"174 conv2d_14" [id=174, type=conv2d];
-"175 empty_14" [id=175, type=empty];
-"176 _param_constant43" [id=176, type=get_attr];
-"177 _param_constant44" [id=177, type=get_attr];
-"178 _tensor_constant28" [id=178, type=get_attr];
-"179 _tensor_constant29" [id=179, type=get_attr];
-"180 _native_batch_norm_legit_no_training_14" [id=180, type=_native_batch_norm_legit_no_training];
-"181 getitem_42" [id=181, type=__getitem__];
-"182 getitem_43" [id=182, type=__getitem__];
-"183 getitem_44" [id=183, type=__getitem__];
-"184 add__5" [id=184, type=add_];
-"185 relu__12" [id=185, type=relu_];
-"186 _param_constant45" [id=186, type=get_attr];
-"187 conv2d_15" [id=187, type=conv2d];
-"188 empty_15" [id=188, type=empty];
-"189 _param_constant46" [id=189, type=get_attr];
-"190 _param_constant47" [id=190, type=get_attr];
-"191 _tensor_constant30" [id=191, type=get_attr];
-"192 _tensor_constant31" [id=192, type=get_attr];
-"193 _native_batch_norm_legit_no_training_15" [id=193, type=_native_batch_norm_legit_no_training];
-"194 getitem_45" [id=194, type=__getitem__];
-"195 getitem_46" [id=195, type=__getitem__];
-"196 getitem_47" [id=196, type=__getitem__];
-"197 relu__13" [id=197, type=relu_];
-"198 _param_constant48" [id=198, type=get_attr];
-"199 conv2d_16" [id=199, type=conv2d];
-"200 empty_16" [id=200, type=empty];
-"201 _param_constant49" [id=201, type=get_attr];
-"202 _param_constant50" [id=202, type=get_attr];
-"203 _tensor_constant32" [id=203, type=get_attr];
-"204 _tensor_constant33" [id=204, type=get_attr];
-"205 _native_batch_norm_legit_no_training_16" [id=205, type=_native_batch_norm_legit_no_training];
-"206 getitem_48" [id=206, type=__getitem__];
-"207 getitem_49" [id=207, type=__getitem__];
-"208 getitem_50" [id=208, type=__getitem__];
-"209 _param_constant51" [id=209, type=get_attr];
-"210 conv2d_17" [id=210, type=conv2d];
-"211 empty_17" [id=211, type=empty];
-"212 _param_constant52" [id=212, type=get_attr];
-"213 _param_constant53" [id=213, type=get_attr];
-"214 _tensor_constant34" [id=214, type=get_attr];
-"215 _tensor_constant35" [id=215, type=get_attr];
-"216 _native_batch_norm_legit_no_training_17" [id=216, type=_native_batch_norm_legit_no_training];
-"217 getitem_51" [id=217, type=__getitem__];
-"218 getitem_52" [id=218, type=__getitem__];
-"219 getitem_53" [id=219, type=__getitem__];
-"220 add__6" [id=220, type=add_];
-"221 relu__14" [id=221, type=relu_];
-"222 _param_constant54" [id=222, type=get_attr];
-"223 conv2d_18" [id=223, type=conv2d];
-"224 empty_18" [id=224, type=empty];
-"225 _param_constant55" [id=225, type=get_attr];
-"226 _param_constant56" [id=226, type=get_attr];
-"227 _tensor_constant36" [id=227, type=get_attr];
-"228 _tensor_constant37" [id=228, type=get_attr];
-"229 _native_batch_norm_legit_no_training_18" [id=229, type=_native_batch_norm_legit_no_training];
-"230 getitem_54" [id=230, type=__getitem__];
-"231 getitem_55" [id=231, type=__getitem__];
-"232 getitem_56" [id=232, type=__getitem__];
-"233 relu__15" [id=233, type=relu_];
-"234 _param_constant57" [id=234, type=get_attr];
-"235 conv2d_19" [id=235, type=conv2d];
-"236 empty_19" [id=236, type=empty];
-"237 _param_constant58" [id=237, type=get_attr];
-"238 _param_constant59" [id=238, type=get_attr];
-"239 _tensor_constant38" [id=239, type=get_attr];
-"240 _tensor_constant39" [id=240, type=get_attr];
-"241 _native_batch_norm_legit_no_training_19" [id=241, type=_native_batch_norm_legit_no_training];
-"242 getitem_57" [id=242, type=__getitem__];
-"243 getitem_58" [id=243, type=__getitem__];
-"244 getitem_59" [id=244, type=__getitem__];
-"245 add__7" [id=245, type=add_];
-"246 relu__16" [id=246, type=relu_];
-"247 adaptive_avg_pool2d" [id=247, type=adaptive_avg_pool2d];
-"248 flatten" [id=248, type=flatten];
-"249 _param_constant60" [id=249, type=get_attr];
-"250 _param_constant61" [id=250, type=get_attr];
-"251 linear" [id=251, type=linear];
-"252 output" [id=252, type=output];
-"0 arg0_1" -> "2 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"1 _param_constant0" -> "2 conv2d"  [label="(64, 3, 7, 7)", style=solid];
-"2 conv2d" -> "8 _native_batch_norm_legit_no_training"  [label="(1, 64, 112, 112)", style=solid];
-"4 _param_constant1" -> "8 _native_batch_norm_legit_no_training"  [label="(64,)", style=solid];
-"5 _param_constant2" -> "8 _native_batch_norm_legit_no_training"  [label="(64,)", style=solid];
-"6 _tensor_constant0" -> "8 _native_batch_norm_legit_no_training"  [label="(64,)", style=solid];
-"7 _tensor_constant1" -> "8 _native_batch_norm_legit_no_training"  [label="(64,)", style=solid];
-"8 _native_batch_norm_legit_no_training" -> "9 getitem"  [label="(1, 64, 112, 112)", style=solid];
-"8 _native_batch_norm_legit_no_training" -> "10 getitem_1"  [label="(1, 64, 112, 112)", style=solid];
-"8 _native_batch_norm_legit_no_training" -> "11 getitem_2"  [label="(1, 64, 112, 112)", style=solid];
-"9 getitem" -> "12 relu_"  [label="(1, 64, 112, 112)", style=solid];
-"12 relu_" -> "13 max_pool2d"  [label="(1, 64, 112, 112)", style=solid];
-"13 max_pool2d" -> "15 conv2d_1"  [label="(1, 64, 56, 56)", style=solid];
-"13 max_pool2d" -> "37 add_"  [label="(1, 64, 56, 56)", style=solid];
-"14 _param_constant3" -> "15 conv2d_1"  [label="(64, 64, 3, 3)", style=solid];
-"15 conv2d_1" -> "21 _native_batch_norm_legit_no_training_1"  [label="(1, 64, 56, 56)", style=solid];
-"17 _param_constant4" -> "21 _native_batch_norm_legit_no_training_1"  [label="(64,)", style=solid];
-"18 _param_constant5" -> "21 _native_batch_norm_legit_no_training_1"  [label="(64,)", style=solid];
-"19 _tensor_constant2" -> "21 _native_batch_norm_legit_no_training_1"  [label="(64,)", style=solid];
-"20 _tensor_constant3" -> "21 _native_batch_norm_legit_no_training_1"  [label="(64,)", style=solid];
-"21 _native_batch_norm_legit_no_training_1" -> "22 getitem_3"  [label="(1, 64, 56, 56)", style=solid];
-"21 _native_batch_norm_legit_no_training_1" -> "23 getitem_4"  [label="(1, 64, 56, 56)", style=solid];
-"21 _native_batch_norm_legit_no_training_1" -> "24 getitem_5"  [label="(1, 64, 56, 56)", style=solid];
-"22 getitem_3" -> "25 relu__1"  [label="(1, 64, 56, 56)", style=solid];
-"25 relu__1" -> "27 conv2d_2"  [label="(1, 64, 56, 56)", style=solid];
-"26 _param_constant6" -> "27 conv2d_2"  [label="(64, 64, 3, 3)", style=solid];
-"27 conv2d_2" -> "33 _native_batch_norm_legit_no_training_2"  [label="(1, 64, 56, 56)", style=solid];
-"29 _param_constant7" -> "33 _native_batch_norm_legit_no_training_2"  [label="(64,)", style=solid];
-"30 _param_constant8" -> "33 _native_batch_norm_legit_no_training_2"  [label="(64,)", style=solid];
-"31 _tensor_constant4" -> "33 _native_batch_norm_legit_no_training_2"  [label="(64,)", style=solid];
-"32 _tensor_constant5" -> "33 _native_batch_norm_legit_no_training_2"  [label="(64,)", style=solid];
-"33 _native_batch_norm_legit_no_training_2" -> "34 getitem_6"  [label="(1, 64, 56, 56)", style=solid];
-"33 _native_batch_norm_legit_no_training_2" -> "35 getitem_7"  [label="(1, 64, 56, 56)", style=solid];
-"33 _native_batch_norm_legit_no_training_2" -> "36 getitem_8"  [label="(1, 64, 56, 56)", style=solid];
-"34 getitem_6" -> "37 add_"  [label="(1, 64, 56, 56)", style=solid];
-"37 add_" -> "38 relu__2"  [label="(1, 64, 56, 56)", style=solid];
-"38 relu__2" -> "40 conv2d_3"  [label="(1, 64, 56, 56)", style=solid];
-"38 relu__2" -> "62 add__1"  [label="(1, 64, 56, 56)", style=solid];
-"39 _param_constant9" -> "40 conv2d_3"  [label="(64, 64, 3, 3)", style=solid];
-"40 conv2d_3" -> "46 _native_batch_norm_legit_no_training_3"  [label="(1, 64, 56, 56)", style=solid];
-"42 _param_constant10" -> "46 _native_batch_norm_legit_no_training_3"  [label="(64,)", style=solid];
-"43 _param_constant11" -> "46 _native_batch_norm_legit_no_training_3"  [label="(64,)", style=solid];
-"44 _tensor_constant6" -> "46 _native_batch_norm_legit_no_training_3"  [label="(64,)", style=solid];
-"45 _tensor_constant7" -> "46 _native_batch_norm_legit_no_training_3"  [label="(64,)", style=solid];
-"46 _native_batch_norm_legit_no_training_3" -> "47 getitem_9"  [label="(1, 64, 56, 56)", style=solid];
-"46 _native_batch_norm_legit_no_training_3" -> "48 getitem_10"  [label="(1, 64, 56, 56)", style=solid];
-"46 _native_batch_norm_legit_no_training_3" -> "49 getitem_11"  [label="(1, 64, 56, 56)", style=solid];
-"47 getitem_9" -> "50 relu__3"  [label="(1, 64, 56, 56)", style=solid];
-"50 relu__3" -> "52 conv2d_4"  [label="(1, 64, 56, 56)", style=solid];
-"51 _param_constant12" -> "52 conv2d_4"  [label="(64, 64, 3, 3)", style=solid];
-"52 conv2d_4" -> "58 _native_batch_norm_legit_no_training_4"  [label="(1, 64, 56, 56)", style=solid];
-"54 _param_constant13" -> "58 _native_batch_norm_legit_no_training_4"  [label="(64,)", style=solid];
-"55 _param_constant14" -> "58 _native_batch_norm_legit_no_training_4"  [label="(64,)", style=solid];
-"56 _tensor_constant8" -> "58 _native_batch_norm_legit_no_training_4"  [label="(64,)", style=solid];
-"57 _tensor_constant9" -> "58 _native_batch_norm_legit_no_training_4"  [label="(64,)", style=solid];
-"58 _native_batch_norm_legit_no_training_4" -> "59 getitem_12"  [label="(1, 64, 56, 56)", style=solid];
-"58 _native_batch_norm_legit_no_training_4" -> "60 getitem_13"  [label="(1, 64, 56, 56)", style=solid];
-"58 _native_batch_norm_legit_no_training_4" -> "61 getitem_14"  [label="(1, 64, 56, 56)", style=solid];
-"59 getitem_12" -> "62 add__1"  [label="(1, 64, 56, 56)", style=solid];
-"62 add__1" -> "63 relu__4"  [label="(1, 64, 56, 56)", style=solid];
-"63 relu__4" -> "65 conv2d_5"  [label="(1, 64, 56, 56)", style=solid];
-"63 relu__4" -> "88 conv2d_7"  [label="(1, 64, 56, 56)", style=solid];
-"64 _param_constant15" -> "65 conv2d_5"  [label="(128, 64, 3, 3)", style=solid];
-"65 conv2d_5" -> "71 _native_batch_norm_legit_no_training_5"  [label="(1, 128, 28, 28)", style=solid];
-"67 _param_constant16" -> "71 _native_batch_norm_legit_no_training_5"  [label="(128,)", style=solid];
-"68 _param_constant17" -> "71 _native_batch_norm_legit_no_training_5"  [label="(128,)", style=solid];
-"69 _tensor_constant10" -> "71 _native_batch_norm_legit_no_training_5"  [label="(128,)", style=solid];
-"70 _tensor_constant11" -> "71 _native_batch_norm_legit_no_training_5"  [label="(128,)", style=solid];
-"71 _native_batch_norm_legit_no_training_5" -> "72 getitem_15"  [label="(1, 128, 28, 28)", style=solid];
-"71 _native_batch_norm_legit_no_training_5" -> "73 getitem_16"  [label="(1, 128, 28, 28)", style=solid];
-"71 _native_batch_norm_legit_no_training_5" -> "74 getitem_17"  [label="(1, 128, 28, 28)", style=solid];
-"72 getitem_15" -> "75 relu__5"  [label="(1, 128, 28, 28)", style=solid];
-"75 relu__5" -> "77 conv2d_6"  [label="(1, 128, 28, 28)", style=solid];
-"76 _param_constant18" -> "77 conv2d_6"  [label="(128, 128, 3, 3)", style=solid];
-"77 conv2d_6" -> "83 _native_batch_norm_legit_no_training_6"  [label="(1, 128, 28, 28)", style=solid];
-"79 _param_constant19" -> "83 _native_batch_norm_legit_no_training_6"  [label="(128,)", style=solid];
-"80 _param_constant20" -> "83 _native_batch_norm_legit_no_training_6"  [label="(128,)", style=solid];
-"81 _tensor_constant12" -> "83 _native_batch_norm_legit_no_training_6"  [label="(128,)", style=solid];
-"82 _tensor_constant13" -> "83 _native_batch_norm_legit_no_training_6"  [label="(128,)", style=solid];
-"83 _native_batch_norm_legit_no_training_6" -> "84 getitem_18"  [label="(1, 128, 28, 28)", style=solid];
-"83 _native_batch_norm_legit_no_training_6" -> "85 getitem_19"  [label="(1, 128, 28, 28)", style=solid];
-"83 _native_batch_norm_legit_no_training_6" -> "86 getitem_20"  [label="(1, 128, 28, 28)", style=solid];
-"84 getitem_18" -> "98 add__2"  [label="(1, 128, 28, 28)", style=solid];
-"87 _param_constant21" -> "88 conv2d_7"  [label="(128, 64, 1, 1)", style=solid];
-"88 conv2d_7" -> "94 _native_batch_norm_legit_no_training_7"  [label="(1, 128, 28, 28)", style=solid];
-"90 _param_constant22" -> "94 _native_batch_norm_legit_no_training_7"  [label="(128,)", style=solid];
-"91 _param_constant23" -> "94 _native_batch_norm_legit_no_training_7"  [label="(128,)", style=solid];
-"92 _tensor_constant14" -> "94 _native_batch_norm_legit_no_training_7"  [label="(128,)", style=solid];
-"93 _tensor_constant15" -> "94 _native_batch_norm_legit_no_training_7"  [label="(128,)", style=solid];
-"94 _native_batch_norm_legit_no_training_7" -> "95 getitem_21"  [label="(1, 128, 28, 28)", style=solid];
-"94 _native_batch_norm_legit_no_training_7" -> "96 getitem_22"  [label="(1, 128, 28, 28)", style=solid];
-"94 _native_batch_norm_legit_no_training_7" -> "97 getitem_23"  [label="(1, 128, 28, 28)", style=solid];
-"95 getitem_21" -> "98 add__2"  [label="(1, 128, 28, 28)", style=solid];
-"98 add__2" -> "99 relu__6"  [label="(1, 128, 28, 28)", style=solid];
-"99 relu__6" -> "101 conv2d_8"  [label="(1, 128, 28, 28)", style=solid];
-"99 relu__6" -> "123 add__3"  [label="(1, 128, 28, 28)", style=solid];
-"100 _param_constant24" -> "101 conv2d_8"  [label="(128, 128, 3, 3)", style=solid];
-"101 conv2d_8" -> "107 _native_batch_norm_legit_no_training_8"  [label="(1, 128, 28, 28)", style=solid];
-"103 _param_constant25" -> "107 _native_batch_norm_legit_no_training_8"  [label="(128,)", style=solid];
-"104 _param_constant26" -> "107 _native_batch_norm_legit_no_training_8"  [label="(128,)", style=solid];
-"105 _tensor_constant16" -> "107 _native_batch_norm_legit_no_training_8"  [label="(128,)", style=solid];
-"106 _tensor_constant17" -> "107 _native_batch_norm_legit_no_training_8"  [label="(128,)", style=solid];
-"107 _native_batch_norm_legit_no_training_8" -> "108 getitem_24"  [label="(1, 128, 28, 28)", style=solid];
-"107 _native_batch_norm_legit_no_training_8" -> "109 getitem_25"  [label="(1, 128, 28, 28)", style=solid];
-"107 _native_batch_norm_legit_no_training_8" -> "110 getitem_26"  [label="(1, 128, 28, 28)", style=solid];
-"108 getitem_24" -> "111 relu__7"  [label="(1, 128, 28, 28)", style=solid];
-"111 relu__7" -> "113 conv2d_9"  [label="(1, 128, 28, 28)", style=solid];
-"112 _param_constant27" -> "113 conv2d_9"  [label="(128, 128, 3, 3)", style=solid];
-"113 conv2d_9" -> "119 _native_batch_norm_legit_no_training_9"  [label="(1, 128, 28, 28)", style=solid];
-"115 _param_constant28" -> "119 _native_batch_norm_legit_no_training_9"  [label="(128,)", style=solid];
-"116 _param_constant29" -> "119 _native_batch_norm_legit_no_training_9"  [label="(128,)", style=solid];
-"117 _tensor_constant18" -> "119 _native_batch_norm_legit_no_training_9"  [label="(128,)", style=solid];
-"118 _tensor_constant19" -> "119 _native_batch_norm_legit_no_training_9"  [label="(128,)", style=solid];
-"119 _native_batch_norm_legit_no_training_9" -> "120 getitem_27"  [label="(1, 128, 28, 28)", style=solid];
-"119 _native_batch_norm_legit_no_training_9" -> "121 getitem_28"  [label="(1, 128, 28, 28)", style=solid];
-"119 _native_batch_norm_legit_no_training_9" -> "122 getitem_29"  [label="(1, 128, 28, 28)", style=solid];
-"120 getitem_27" -> "123 add__3"  [label="(1, 128, 28, 28)", style=solid];
-"123 add__3" -> "124 relu__8"  [label="(1, 128, 28, 28)", style=solid];
-"124 relu__8" -> "126 conv2d_10"  [label="(1, 128, 28, 28)", style=solid];
-"124 relu__8" -> "149 conv2d_12"  [label="(1, 128, 28, 28)", style=solid];
-"125 _param_constant30" -> "126 conv2d_10"  [label="(256, 128, 3, 3)", style=solid];
-"126 conv2d_10" -> "132 _native_batch_norm_legit_no_training_10"  [label="(1, 256, 14, 14)", style=solid];
-"128 _param_constant31" -> "132 _native_batch_norm_legit_no_training_10"  [label="(256,)", style=solid];
-"129 _param_constant32" -> "132 _native_batch_norm_legit_no_training_10"  [label="(256,)", style=solid];
-"130 _tensor_constant20" -> "132 _native_batch_norm_legit_no_training_10"  [label="(256,)", style=solid];
-"131 _tensor_constant21" -> "132 _native_batch_norm_legit_no_training_10"  [label="(256,)", style=solid];
-"132 _native_batch_norm_legit_no_training_10" -> "133 getitem_30"  [label="(1, 256, 14, 14)", style=solid];
-"132 _native_batch_norm_legit_no_training_10" -> "134 getitem_31"  [label="(1, 256, 14, 14)", style=solid];
-"132 _native_batch_norm_legit_no_training_10" -> "135 getitem_32"  [label="(1, 256, 14, 14)", style=solid];
-"133 getitem_30" -> "136 relu__9"  [label="(1, 256, 14, 14)", style=solid];
-"136 relu__9" -> "138 conv2d_11"  [label="(1, 256, 14, 14)", style=solid];
-"137 _param_constant33" -> "138 conv2d_11"  [label="(256, 256, 3, 3)", style=solid];
-"138 conv2d_11" -> "144 _native_batch_norm_legit_no_training_11"  [label="(1, 256, 14, 14)", style=solid];
-"140 _param_constant34" -> "144 _native_batch_norm_legit_no_training_11"  [label="(256,)", style=solid];
-"141 _param_constant35" -> "144 _native_batch_norm_legit_no_training_11"  [label="(256,)", style=solid];
-"142 _tensor_constant22" -> "144 _native_batch_norm_legit_no_training_11"  [label="(256,)", style=solid];
-"143 _tensor_constant23" -> "144 _native_batch_norm_legit_no_training_11"  [label="(256,)", style=solid];
-"144 _native_batch_norm_legit_no_training_11" -> "145 getitem_33"  [label="(1, 256, 14, 14)", style=solid];
-"144 _native_batch_norm_legit_no_training_11" -> "146 getitem_34"  [label="(1, 256, 14, 14)", style=solid];
-"144 _native_batch_norm_legit_no_training_11" -> "147 getitem_35"  [label="(1, 256, 14, 14)", style=solid];
-"145 getitem_33" -> "159 add__4"  [label="(1, 256, 14, 14)", style=solid];
-"148 _param_constant36" -> "149 conv2d_12"  [label="(256, 128, 1, 1)", style=solid];
-"149 conv2d_12" -> "155 _native_batch_norm_legit_no_training_12"  [label="(1, 256, 14, 14)", style=solid];
-"151 _param_constant37" -> "155 _native_batch_norm_legit_no_training_12"  [label="(256,)", style=solid];
-"152 _param_constant38" -> "155 _native_batch_norm_legit_no_training_12"  [label="(256,)", style=solid];
-"153 _tensor_constant24" -> "155 _native_batch_norm_legit_no_training_12"  [label="(256,)", style=solid];
-"154 _tensor_constant25" -> "155 _native_batch_norm_legit_no_training_12"  [label="(256,)", style=solid];
-"155 _native_batch_norm_legit_no_training_12" -> "156 getitem_36"  [label="(1, 256, 14, 14)", style=solid];
-"155 _native_batch_norm_legit_no_training_12" -> "157 getitem_37"  [label="(1, 256, 14, 14)", style=solid];
-"155 _native_batch_norm_legit_no_training_12" -> "158 getitem_38"  [label="(1, 256, 14, 14)", style=solid];
-"156 getitem_36" -> "159 add__4"  [label="(1, 256, 14, 14)", style=solid];
-"159 add__4" -> "160 relu__10"  [label="(1, 256, 14, 14)", style=solid];
-"160 relu__10" -> "162 conv2d_13"  [label="(1, 256, 14, 14)", style=solid];
-"160 relu__10" -> "184 add__5"  [label="(1, 256, 14, 14)", style=solid];
-"161 _param_constant39" -> "162 conv2d_13"  [label="(256, 256, 3, 3)", style=solid];
-"162 conv2d_13" -> "168 _native_batch_norm_legit_no_training_13"  [label="(1, 256, 14, 14)", style=solid];
-"164 _param_constant40" -> "168 _native_batch_norm_legit_no_training_13"  [label="(256,)", style=solid];
-"165 _param_constant41" -> "168 _native_batch_norm_legit_no_training_13"  [label="(256,)", style=solid];
-"166 _tensor_constant26" -> "168 _native_batch_norm_legit_no_training_13"  [label="(256,)", style=solid];
-"167 _tensor_constant27" -> "168 _native_batch_norm_legit_no_training_13"  [label="(256,)", style=solid];
-"168 _native_batch_norm_legit_no_training_13" -> "169 getitem_39"  [label="(1, 256, 14, 14)", style=solid];
-"168 _native_batch_norm_legit_no_training_13" -> "170 getitem_40"  [label="(1, 256, 14, 14)", style=solid];
-"168 _native_batch_norm_legit_no_training_13" -> "171 getitem_41"  [label="(1, 256, 14, 14)", style=solid];
-"169 getitem_39" -> "172 relu__11"  [label="(1, 256, 14, 14)", style=solid];
-"172 relu__11" -> "174 conv2d_14"  [label="(1, 256, 14, 14)", style=solid];
-"173 _param_constant42" -> "174 conv2d_14"  [label="(256, 256, 3, 3)", style=solid];
-"174 conv2d_14" -> "180 _native_batch_norm_legit_no_training_14"  [label="(1, 256, 14, 14)", style=solid];
-"176 _param_constant43" -> "180 _native_batch_norm_legit_no_training_14"  [label="(256,)", style=solid];
-"177 _param_constant44" -> "180 _native_batch_norm_legit_no_training_14"  [label="(256,)", style=solid];
-"178 _tensor_constant28" -> "180 _native_batch_norm_legit_no_training_14"  [label="(256,)", style=solid];
-"179 _tensor_constant29" -> "180 _native_batch_norm_legit_no_training_14"  [label="(256,)", style=solid];
-"180 _native_batch_norm_legit_no_training_14" -> "181 getitem_42"  [label="(1, 256, 14, 14)", style=solid];
-"180 _native_batch_norm_legit_no_training_14" -> "182 getitem_43"  [label="(1, 256, 14, 14)", style=solid];
-"180 _native_batch_norm_legit_no_training_14" -> "183 getitem_44"  [label="(1, 256, 14, 14)", style=solid];
-"181 getitem_42" -> "184 add__5"  [label="(1, 256, 14, 14)", style=solid];
-"184 add__5" -> "185 relu__12"  [label="(1, 256, 14, 14)", style=solid];
-"185 relu__12" -> "187 conv2d_15"  [label="(1, 256, 14, 14)", style=solid];
-"185 relu__12" -> "210 conv2d_17"  [label="(1, 256, 14, 14)", style=solid];
-"186 _param_constant45" -> "187 conv2d_15"  [label="(512, 256, 3, 3)", style=solid];
-"187 conv2d_15" -> "193 _native_batch_norm_legit_no_training_15"  [label="(1, 512, 7, 7)", style=solid];
-"189 _param_constant46" -> "193 _native_batch_norm_legit_no_training_15"  [label="(512,)", style=solid];
-"190 _param_constant47" -> "193 _native_batch_norm_legit_no_training_15"  [label="(512,)", style=solid];
-"191 _tensor_constant30" -> "193 _native_batch_norm_legit_no_training_15"  [label="(512,)", style=solid];
-"192 _tensor_constant31" -> "193 _native_batch_norm_legit_no_training_15"  [label="(512,)", style=solid];
-"193 _native_batch_norm_legit_no_training_15" -> "194 getitem_45"  [label="(1, 512, 7, 7)", style=solid];
-"193 _native_batch_norm_legit_no_training_15" -> "195 getitem_46"  [label="(1, 512, 7, 7)", style=solid];
-"193 _native_batch_norm_legit_no_training_15" -> "196 getitem_47"  [label="(1, 512, 7, 7)", style=solid];
-"194 getitem_45" -> "197 relu__13"  [label="(1, 512, 7, 7)", style=solid];
-"197 relu__13" -> "199 conv2d_16"  [label="(1, 512, 7, 7)", style=solid];
-"198 _param_constant48" -> "199 conv2d_16"  [label="(512, 512, 3, 3)", style=solid];
-"199 conv2d_16" -> "205 _native_batch_norm_legit_no_training_16"  [label="(1, 512, 7, 7)", style=solid];
-"201 _param_constant49" -> "205 _native_batch_norm_legit_no_training_16"  [label="(512,)", style=solid];
-"202 _param_constant50" -> "205 _native_batch_norm_legit_no_training_16"  [label="(512,)", style=solid];
-"203 _tensor_constant32" -> "205 _native_batch_norm_legit_no_training_16"  [label="(512,)", style=solid];
-"204 _tensor_constant33" -> "205 _native_batch_norm_legit_no_training_16"  [label="(512,)", style=solid];
-"205 _native_batch_norm_legit_no_training_16" -> "206 getitem_48"  [label="(1, 512, 7, 7)", style=solid];
-"205 _native_batch_norm_legit_no_training_16" -> "207 getitem_49"  [label="(1, 512, 7, 7)", style=solid];
-"205 _native_batch_norm_legit_no_training_16" -> "208 getitem_50"  [label="(1, 512, 7, 7)", style=solid];
-"206 getitem_48" -> "220 add__6"  [label="(1, 512, 7, 7)", style=solid];
-"209 _param_constant51" -> "210 conv2d_17"  [label="(512, 256, 1, 1)", style=solid];
-"210 conv2d_17" -> "216 _native_batch_norm_legit_no_training_17"  [label="(1, 512, 7, 7)", style=solid];
-"212 _param_constant52" -> "216 _native_batch_norm_legit_no_training_17"  [label="(512,)", style=solid];
-"213 _param_constant53" -> "216 _native_batch_norm_legit_no_training_17"  [label="(512,)", style=solid];
-"214 _tensor_constant34" -> "216 _native_batch_norm_legit_no_training_17"  [label="(512,)", style=solid];
-"215 _tensor_constant35" -> "216 _native_batch_norm_legit_no_training_17"  [label="(512,)", style=solid];
-"216 _native_batch_norm_legit_no_training_17" -> "217 getitem_51"  [label="(1, 512, 7, 7)", style=solid];
-"216 _native_batch_norm_legit_no_training_17" -> "218 getitem_52"  [label="(1, 512, 7, 7)", style=solid];
-"216 _native_batch_norm_legit_no_training_17" -> "219 getitem_53"  [label="(1, 512, 7, 7)", style=solid];
-"217 getitem_51" -> "220 add__6"  [label="(1, 512, 7, 7)", style=solid];
-"220 add__6" -> "221 relu__14"  [label="(1, 512, 7, 7)", style=solid];
-"221 relu__14" -> "223 conv2d_18"  [label="(1, 512, 7, 7)", style=solid];
-"221 relu__14" -> "245 add__7"  [label="(1, 512, 7, 7)", style=solid];
-"222 _param_constant54" -> "223 conv2d_18"  [label="(512, 512, 3, 3)", style=solid];
-"223 conv2d_18" -> "229 _native_batch_norm_legit_no_training_18"  [label="(1, 512, 7, 7)", style=solid];
-"225 _param_constant55" -> "229 _native_batch_norm_legit_no_training_18"  [label="(512,)", style=solid];
-"226 _param_constant56" -> "229 _native_batch_norm_legit_no_training_18"  [label="(512,)", style=solid];
-"227 _tensor_constant36" -> "229 _native_batch_norm_legit_no_training_18"  [label="(512,)", style=solid];
-"228 _tensor_constant37" -> "229 _native_batch_norm_legit_no_training_18"  [label="(512,)", style=solid];
-"229 _native_batch_norm_legit_no_training_18" -> "230 getitem_54"  [label="(1, 512, 7, 7)", style=solid];
-"229 _native_batch_norm_legit_no_training_18" -> "231 getitem_55"  [label="(1, 512, 7, 7)", style=solid];
-"229 _native_batch_norm_legit_no_training_18" -> "232 getitem_56"  [label="(1, 512, 7, 7)", style=solid];
-"230 getitem_54" -> "233 relu__15"  [label="(1, 512, 7, 7)", style=solid];
-"233 relu__15" -> "235 conv2d_19"  [label="(1, 512, 7, 7)", style=solid];
-"234 _param_constant57" -> "235 conv2d_19"  [label="(512, 512, 3, 3)", style=solid];
-"235 conv2d_19" -> "241 _native_batch_norm_legit_no_training_19"  [label="(1, 512, 7, 7)", style=solid];
-"237 _param_constant58" -> "241 _native_batch_norm_legit_no_training_19"  [label="(512,)", style=solid];
-"238 _param_constant59" -> "241 _native_batch_norm_legit_no_training_19"  [label="(512,)", style=solid];
-"239 _tensor_constant38" -> "241 _native_batch_norm_legit_no_training_19"  [label="(512,)", style=solid];
-"240 _tensor_constant39" -> "241 _native_batch_norm_legit_no_training_19"  [label="(512,)", style=solid];
-"241 _native_batch_norm_legit_no_training_19" -> "242 getitem_57"  [label="(1, 512, 7, 7)", style=solid];
-"241 _native_batch_norm_legit_no_training_19" -> "243 getitem_58"  [label="(1, 512, 7, 7)", style=solid];
-"241 _native_batch_norm_legit_no_training_19" -> "244 getitem_59"  [label="(1, 512, 7, 7)", style=solid];
-"242 getitem_57" -> "245 add__7"  [label="(1, 512, 7, 7)", style=solid];
-"245 add__7" -> "246 relu__16"  [label="(1, 512, 7, 7)", style=solid];
-"246 relu__16" -> "247 adaptive_avg_pool2d"  [label="(1, 512, 7, 7)", style=solid];
-"247 adaptive_avg_pool2d" -> "248 flatten"  [label="(1, 512, 1, 1)", style=solid];
-"248 flatten" -> "251 linear"  [label="(1, 512)", style=solid];
-"249 _param_constant60" -> "251 linear"  [label="(1000, 512)", style=solid];
-"250 _param_constant61" -> "251 linear"  [label="(1000,)", style=solid];
-"251 linear" -> "252 output"  [label="(1, 1000)", style=solid];
+"0 conv1_weight" [id=0, type=get_attr];
+"1 bn1_weight" [id=1, type=get_attr];
+"2 bn1_bias" [id=2, type=get_attr];
+"3 layer1_0_conv1_weight" [id=3, type=get_attr];
+"4 layer1_0_bn1_weight" [id=4, type=get_attr];
+"5 layer1_0_bn1_bias" [id=5, type=get_attr];
+"6 layer1_0_conv2_weight" [id=6, type=get_attr];
+"7 layer1_0_bn2_weight" [id=7, type=get_attr];
+"8 layer1_0_bn2_bias" [id=8, type=get_attr];
+"9 layer1_1_conv1_weight" [id=9, type=get_attr];
+"10 layer1_1_bn1_weight" [id=10, type=get_attr];
+"11 layer1_1_bn1_bias" [id=11, type=get_attr];
+"12 layer1_1_conv2_weight" [id=12, type=get_attr];
+"13 layer1_1_bn2_weight" [id=13, type=get_attr];
+"14 layer1_1_bn2_bias" [id=14, type=get_attr];
+"15 layer2_0_conv1_weight" [id=15, type=get_attr];
+"16 layer2_0_bn1_weight" [id=16, type=get_attr];
+"17 layer2_0_bn1_bias" [id=17, type=get_attr];
+"18 layer2_0_conv2_weight" [id=18, type=get_attr];
+"19 layer2_0_bn2_weight" [id=19, type=get_attr];
+"20 layer2_0_bn2_bias" [id=20, type=get_attr];
+"21 layer2_0_downsample_0_weight" [id=21, type=get_attr];
+"22 layer2_0_downsample_1_weight" [id=22, type=get_attr];
+"23 layer2_0_downsample_1_bias" [id=23, type=get_attr];
+"24 layer2_1_conv1_weight" [id=24, type=get_attr];
+"25 layer2_1_bn1_weight" [id=25, type=get_attr];
+"26 layer2_1_bn1_bias" [id=26, type=get_attr];
+"27 layer2_1_conv2_weight" [id=27, type=get_attr];
+"28 layer2_1_bn2_weight" [id=28, type=get_attr];
+"29 layer2_1_bn2_bias" [id=29, type=get_attr];
+"30 layer3_0_conv1_weight" [id=30, type=get_attr];
+"31 layer3_0_bn1_weight" [id=31, type=get_attr];
+"32 layer3_0_bn1_bias" [id=32, type=get_attr];
+"33 layer3_0_conv2_weight" [id=33, type=get_attr];
+"34 layer3_0_bn2_weight" [id=34, type=get_attr];
+"35 layer3_0_bn2_bias" [id=35, type=get_attr];
+"36 layer3_0_downsample_0_weight" [id=36, type=get_attr];
+"37 layer3_0_downsample_1_weight" [id=37, type=get_attr];
+"38 layer3_0_downsample_1_bias" [id=38, type=get_attr];
+"39 layer3_1_conv1_weight" [id=39, type=get_attr];
+"40 layer3_1_bn1_weight" [id=40, type=get_attr];
+"41 layer3_1_bn1_bias" [id=41, type=get_attr];
+"42 layer3_1_conv2_weight" [id=42, type=get_attr];
+"43 layer3_1_bn2_weight" [id=43, type=get_attr];
+"44 layer3_1_bn2_bias" [id=44, type=get_attr];
+"45 layer4_0_conv1_weight" [id=45, type=get_attr];
+"46 layer4_0_bn1_weight" [id=46, type=get_attr];
+"47 layer4_0_bn1_bias" [id=47, type=get_attr];
+"48 layer4_0_conv2_weight" [id=48, type=get_attr];
+"49 layer4_0_bn2_weight" [id=49, type=get_attr];
+"50 layer4_0_bn2_bias" [id=50, type=get_attr];
+"51 layer4_0_downsample_0_weight" [id=51, type=get_attr];
+"52 layer4_0_downsample_1_weight" [id=52, type=get_attr];
+"53 layer4_0_downsample_1_bias" [id=53, type=get_attr];
+"54 layer4_1_conv1_weight" [id=54, type=get_attr];
+"55 layer4_1_bn1_weight" [id=55, type=get_attr];
+"56 layer4_1_bn1_bias" [id=56, type=get_attr];
+"57 layer4_1_conv2_weight" [id=57, type=get_attr];
+"58 layer4_1_bn2_weight" [id=58, type=get_attr];
+"59 layer4_1_bn2_bias" [id=59, type=get_attr];
+"60 fc_weight" [id=60, type=get_attr];
+"61 fc_bias" [id=61, type=get_attr];
+"62 bn1_running_mean" [id=62, type=get_attr];
+"63 bn1_running_var" [id=63, type=get_attr];
+"64 bn1_num_batches_tracked" [id=64, type=get_attr];
+"65 layer1_0_bn1_running_mean" [id=65, type=get_attr];
+"66 layer1_0_bn1_running_var" [id=66, type=get_attr];
+"67 layer1_0_bn1_num_batches_tracked" [id=67, type=get_attr];
+"68 layer1_0_bn2_running_mean" [id=68, type=get_attr];
+"69 layer1_0_bn2_running_var" [id=69, type=get_attr];
+"70 layer1_0_bn2_num_batches_tracked" [id=70, type=get_attr];
+"71 layer1_1_bn1_running_mean" [id=71, type=get_attr];
+"72 layer1_1_bn1_running_var" [id=72, type=get_attr];
+"73 layer1_1_bn1_num_batches_tracked" [id=73, type=get_attr];
+"74 layer1_1_bn2_running_mean" [id=74, type=get_attr];
+"75 layer1_1_bn2_running_var" [id=75, type=get_attr];
+"76 layer1_1_bn2_num_batches_tracked" [id=76, type=get_attr];
+"77 layer2_0_bn1_running_mean" [id=77, type=get_attr];
+"78 layer2_0_bn1_running_var" [id=78, type=get_attr];
+"79 layer2_0_bn1_num_batches_tracked" [id=79, type=get_attr];
+"80 layer2_0_bn2_running_mean" [id=80, type=get_attr];
+"81 layer2_0_bn2_running_var" [id=81, type=get_attr];
+"82 layer2_0_bn2_num_batches_tracked" [id=82, type=get_attr];
+"83 layer2_0_downsample_1_running_mean" [id=83, type=get_attr];
+"84 layer2_0_downsample_1_running_var" [id=84, type=get_attr];
+"85 layer2_0_downsample_1_num_batches_tracked" [id=85, type=get_attr];
+"86 layer2_1_bn1_running_mean" [id=86, type=get_attr];
+"87 layer2_1_bn1_running_var" [id=87, type=get_attr];
+"88 layer2_1_bn1_num_batches_tracked" [id=88, type=get_attr];
+"89 layer2_1_bn2_running_mean" [id=89, type=get_attr];
+"90 layer2_1_bn2_running_var" [id=90, type=get_attr];
+"91 layer2_1_bn2_num_batches_tracked" [id=91, type=get_attr];
+"92 layer3_0_bn1_running_mean" [id=92, type=get_attr];
+"93 layer3_0_bn1_running_var" [id=93, type=get_attr];
+"94 layer3_0_bn1_num_batches_tracked" [id=94, type=get_attr];
+"95 layer3_0_bn2_running_mean" [id=95, type=get_attr];
+"96 layer3_0_bn2_running_var" [id=96, type=get_attr];
+"97 layer3_0_bn2_num_batches_tracked" [id=97, type=get_attr];
+"98 layer3_0_downsample_1_running_mean" [id=98, type=get_attr];
+"99 layer3_0_downsample_1_running_var" [id=99, type=get_attr];
+"100 layer3_0_downsample_1_num_batches_tracked" [id=100, type=get_attr];
+"101 layer3_1_bn1_running_mean" [id=101, type=get_attr];
+"102 layer3_1_bn1_running_var" [id=102, type=get_attr];
+"103 layer3_1_bn1_num_batches_tracked" [id=103, type=get_attr];
+"104 layer3_1_bn2_running_mean" [id=104, type=get_attr];
+"105 layer3_1_bn2_running_var" [id=105, type=get_attr];
+"106 layer3_1_bn2_num_batches_tracked" [id=106, type=get_attr];
+"107 layer4_0_bn1_running_mean" [id=107, type=get_attr];
+"108 layer4_0_bn1_running_var" [id=108, type=get_attr];
+"109 layer4_0_bn1_num_batches_tracked" [id=109, type=get_attr];
+"110 layer4_0_bn2_running_mean" [id=110, type=get_attr];
+"111 layer4_0_bn2_running_var" [id=111, type=get_attr];
+"112 layer4_0_bn2_num_batches_tracked" [id=112, type=get_attr];
+"113 layer4_0_downsample_1_running_mean" [id=113, type=get_attr];
+"114 layer4_0_downsample_1_running_var" [id=114, type=get_attr];
+"115 layer4_0_downsample_1_num_batches_tracked" [id=115, type=get_attr];
+"116 layer4_1_bn1_running_mean" [id=116, type=get_attr];
+"117 layer4_1_bn1_running_var" [id=117, type=get_attr];
+"118 layer4_1_bn1_num_batches_tracked" [id=118, type=get_attr];
+"119 layer4_1_bn2_running_mean" [id=119, type=get_attr];
+"120 layer4_1_bn2_running_var" [id=120, type=get_attr];
+"121 layer4_1_bn2_num_batches_tracked" [id=121, type=get_attr];
+"122 x" [id=122, type=input];
+"123 conv2d" [id=123, type=conv2d];
+"124 batch_norm" [id=124, type=batch_norm];
+"125 relu_" [id=125, type=relu_];
+"126 max_pool2d" [id=126, type=max_pool2d];
+"127 conv2d_1" [id=127, type=conv2d];
+"128 batch_norm_1" [id=128, type=batch_norm];
+"129 relu__1" [id=129, type=relu_];
+"130 conv2d_2" [id=130, type=conv2d];
+"131 batch_norm_2" [id=131, type=batch_norm];
+"132 add_" [id=132, type=add_];
+"133 relu__2" [id=133, type=relu_];
+"134 conv2d_3" [id=134, type=conv2d];
+"135 batch_norm_3" [id=135, type=batch_norm];
+"136 relu__3" [id=136, type=relu_];
+"137 conv2d_4" [id=137, type=conv2d];
+"138 batch_norm_4" [id=138, type=batch_norm];
+"139 add__1" [id=139, type=add_];
+"140 relu__4" [id=140, type=relu_];
+"141 conv2d_5" [id=141, type=conv2d];
+"142 batch_norm_5" [id=142, type=batch_norm];
+"143 relu__5" [id=143, type=relu_];
+"144 conv2d_6" [id=144, type=conv2d];
+"145 batch_norm_6" [id=145, type=batch_norm];
+"146 conv2d_7" [id=146, type=conv2d];
+"147 batch_norm_7" [id=147, type=batch_norm];
+"148 add__2" [id=148, type=add_];
+"149 relu__6" [id=149, type=relu_];
+"150 conv2d_8" [id=150, type=conv2d];
+"151 batch_norm_8" [id=151, type=batch_norm];
+"152 relu__7" [id=152, type=relu_];
+"153 conv2d_9" [id=153, type=conv2d];
+"154 batch_norm_9" [id=154, type=batch_norm];
+"155 add__3" [id=155, type=add_];
+"156 relu__8" [id=156, type=relu_];
+"157 conv2d_10" [id=157, type=conv2d];
+"158 batch_norm_10" [id=158, type=batch_norm];
+"159 relu__9" [id=159, type=relu_];
+"160 conv2d_11" [id=160, type=conv2d];
+"161 batch_norm_11" [id=161, type=batch_norm];
+"162 conv2d_12" [id=162, type=conv2d];
+"163 batch_norm_12" [id=163, type=batch_norm];
+"164 add__4" [id=164, type=add_];
+"165 relu__10" [id=165, type=relu_];
+"166 conv2d_13" [id=166, type=conv2d];
+"167 batch_norm_13" [id=167, type=batch_norm];
+"168 relu__11" [id=168, type=relu_];
+"169 conv2d_14" [id=169, type=conv2d];
+"170 batch_norm_14" [id=170, type=batch_norm];
+"171 add__5" [id=171, type=add_];
+"172 relu__12" [id=172, type=relu_];
+"173 conv2d_15" [id=173, type=conv2d];
+"174 batch_norm_15" [id=174, type=batch_norm];
+"175 relu__13" [id=175, type=relu_];
+"176 conv2d_16" [id=176, type=conv2d];
+"177 batch_norm_16" [id=177, type=batch_norm];
+"178 conv2d_17" [id=178, type=conv2d];
+"179 batch_norm_17" [id=179, type=batch_norm];
+"180 add__6" [id=180, type=add_];
+"181 relu__14" [id=181, type=relu_];
+"182 conv2d_18" [id=182, type=conv2d];
+"183 batch_norm_18" [id=183, type=batch_norm];
+"184 relu__15" [id=184, type=relu_];
+"185 conv2d_19" [id=185, type=conv2d];
+"186 batch_norm_19" [id=186, type=batch_norm];
+"187 add__7" [id=187, type=add_];
+"188 relu__16" [id=188, type=relu_];
+"189 adaptive_avg_pool2d" [id=189, type=adaptive_avg_pool2d];
+"190 flatten" [id=190, type=flatten];
+"191 linear" [id=191, type=linear];
+"192 output_1" [id=192, type=output];
+"0 conv1_weight" -> "123 conv2d"  [label="(64, 3, 7, 7)", style=solid];
+"1 bn1_weight" -> "124 batch_norm"  [label="(64,)", style=solid];
+"2 bn1_bias" -> "124 batch_norm"  [label="(64,)", style=solid];
+"3 layer1_0_conv1_weight" -> "127 conv2d_1"  [label="(64, 64, 3, 3)", style=solid];
+"4 layer1_0_bn1_weight" -> "128 batch_norm_1"  [label="(64,)", style=solid];
+"5 layer1_0_bn1_bias" -> "128 batch_norm_1"  [label="(64,)", style=solid];
+"6 layer1_0_conv2_weight" -> "130 conv2d_2"  [label="(64, 64, 3, 3)", style=solid];
+"7 layer1_0_bn2_weight" -> "131 batch_norm_2"  [label="(64,)", style=solid];
+"8 layer1_0_bn2_bias" -> "131 batch_norm_2"  [label="(64,)", style=solid];
+"9 layer1_1_conv1_weight" -> "134 conv2d_3"  [label="(64, 64, 3, 3)", style=solid];
+"10 layer1_1_bn1_weight" -> "135 batch_norm_3"  [label="(64,)", style=solid];
+"11 layer1_1_bn1_bias" -> "135 batch_norm_3"  [label="(64,)", style=solid];
+"12 layer1_1_conv2_weight" -> "137 conv2d_4"  [label="(64, 64, 3, 3)", style=solid];
+"13 layer1_1_bn2_weight" -> "138 batch_norm_4"  [label="(64,)", style=solid];
+"14 layer1_1_bn2_bias" -> "138 batch_norm_4"  [label="(64,)", style=solid];
+"15 layer2_0_conv1_weight" -> "141 conv2d_5"  [label="(128, 64, 3, 3)", style=solid];
+"16 layer2_0_bn1_weight" -> "142 batch_norm_5"  [label="(128,)", style=solid];
+"17 layer2_0_bn1_bias" -> "142 batch_norm_5"  [label="(128,)", style=solid];
+"18 layer2_0_conv2_weight" -> "144 conv2d_6"  [label="(128, 128, 3, 3)", style=solid];
+"19 layer2_0_bn2_weight" -> "145 batch_norm_6"  [label="(128,)", style=solid];
+"20 layer2_0_bn2_bias" -> "145 batch_norm_6"  [label="(128,)", style=solid];
+"21 layer2_0_downsample_0_weight" -> "146 conv2d_7"  [label="(128, 64, 1, 1)", style=solid];
+"22 layer2_0_downsample_1_weight" -> "147 batch_norm_7"  [label="(128,)", style=solid];
+"23 layer2_0_downsample_1_bias" -> "147 batch_norm_7"  [label="(128,)", style=solid];
+"24 layer2_1_conv1_weight" -> "150 conv2d_8"  [label="(128, 128, 3, 3)", style=solid];
+"25 layer2_1_bn1_weight" -> "151 batch_norm_8"  [label="(128,)", style=solid];
+"26 layer2_1_bn1_bias" -> "151 batch_norm_8"  [label="(128,)", style=solid];
+"27 layer2_1_conv2_weight" -> "153 conv2d_9"  [label="(128, 128, 3, 3)", style=solid];
+"28 layer2_1_bn2_weight" -> "154 batch_norm_9"  [label="(128,)", style=solid];
+"29 layer2_1_bn2_bias" -> "154 batch_norm_9"  [label="(128,)", style=solid];
+"30 layer3_0_conv1_weight" -> "157 conv2d_10"  [label="(256, 128, 3, 3)", style=solid];
+"31 layer3_0_bn1_weight" -> "158 batch_norm_10"  [label="(256,)", style=solid];
+"32 layer3_0_bn1_bias" -> "158 batch_norm_10"  [label="(256,)", style=solid];
+"33 layer3_0_conv2_weight" -> "160 conv2d_11"  [label="(256, 256, 3, 3)", style=solid];
+"34 layer3_0_bn2_weight" -> "161 batch_norm_11"  [label="(256,)", style=solid];
+"35 layer3_0_bn2_bias" -> "161 batch_norm_11"  [label="(256,)", style=solid];
+"36 layer3_0_downsample_0_weight" -> "162 conv2d_12"  [label="(256, 128, 1, 1)", style=solid];
+"37 layer3_0_downsample_1_weight" -> "163 batch_norm_12"  [label="(256,)", style=solid];
+"38 layer3_0_downsample_1_bias" -> "163 batch_norm_12"  [label="(256,)", style=solid];
+"39 layer3_1_conv1_weight" -> "166 conv2d_13"  [label="(256, 256, 3, 3)", style=solid];
+"40 layer3_1_bn1_weight" -> "167 batch_norm_13"  [label="(256,)", style=solid];
+"41 layer3_1_bn1_bias" -> "167 batch_norm_13"  [label="(256,)", style=solid];
+"42 layer3_1_conv2_weight" -> "169 conv2d_14"  [label="(256, 256, 3, 3)", style=solid];
+"43 layer3_1_bn2_weight" -> "170 batch_norm_14"  [label="(256,)", style=solid];
+"44 layer3_1_bn2_bias" -> "170 batch_norm_14"  [label="(256,)", style=solid];
+"45 layer4_0_conv1_weight" -> "173 conv2d_15"  [label="(512, 256, 3, 3)", style=solid];
+"46 layer4_0_bn1_weight" -> "174 batch_norm_15"  [label="(512,)", style=solid];
+"47 layer4_0_bn1_bias" -> "174 batch_norm_15"  [label="(512,)", style=solid];
+"48 layer4_0_conv2_weight" -> "176 conv2d_16"  [label="(512, 512, 3, 3)", style=solid];
+"49 layer4_0_bn2_weight" -> "177 batch_norm_16"  [label="(512,)", style=solid];
+"50 layer4_0_bn2_bias" -> "177 batch_norm_16"  [label="(512,)", style=solid];
+"51 layer4_0_downsample_0_weight" -> "178 conv2d_17"  [label="(512, 256, 1, 1)", style=solid];
+"52 layer4_0_downsample_1_weight" -> "179 batch_norm_17"  [label="(512,)", style=solid];
+"53 layer4_0_downsample_1_bias" -> "179 batch_norm_17"  [label="(512,)", style=solid];
+"54 layer4_1_conv1_weight" -> "182 conv2d_18"  [label="(512, 512, 3, 3)", style=solid];
+"55 layer4_1_bn1_weight" -> "183 batch_norm_18"  [label="(512,)", style=solid];
+"56 layer4_1_bn1_bias" -> "183 batch_norm_18"  [label="(512,)", style=solid];
+"57 layer4_1_conv2_weight" -> "185 conv2d_19"  [label="(512, 512, 3, 3)", style=solid];
+"58 layer4_1_bn2_weight" -> "186 batch_norm_19"  [label="(512,)", style=solid];
+"59 layer4_1_bn2_bias" -> "186 batch_norm_19"  [label="(512,)", style=solid];
+"60 fc_weight" -> "191 linear"  [label="(1000, 512)", style=solid];
+"61 fc_bias" -> "191 linear"  [label="(1000,)", style=solid];
+"62 bn1_running_mean" -> "124 batch_norm"  [label="(64,)", style=solid];
+"63 bn1_running_var" -> "124 batch_norm"  [label="(64,)", style=solid];
+"65 layer1_0_bn1_running_mean" -> "128 batch_norm_1"  [label="(64,)", style=solid];
+"66 layer1_0_bn1_running_var" -> "128 batch_norm_1"  [label="(64,)", style=solid];
+"68 layer1_0_bn2_running_mean" -> "131 batch_norm_2"  [label="(64,)", style=solid];
+"69 layer1_0_bn2_running_var" -> "131 batch_norm_2"  [label="(64,)", style=solid];
+"71 layer1_1_bn1_running_mean" -> "135 batch_norm_3"  [label="(64,)", style=solid];
+"72 layer1_1_bn1_running_var" -> "135 batch_norm_3"  [label="(64,)", style=solid];
+"74 layer1_1_bn2_running_mean" -> "138 batch_norm_4"  [label="(64,)", style=solid];
+"75 layer1_1_bn2_running_var" -> "138 batch_norm_4"  [label="(64,)", style=solid];
+"77 layer2_0_bn1_running_mean" -> "142 batch_norm_5"  [label="(128,)", style=solid];
+"78 layer2_0_bn1_running_var" -> "142 batch_norm_5"  [label="(128,)", style=solid];
+"80 layer2_0_bn2_running_mean" -> "145 batch_norm_6"  [label="(128,)", style=solid];
+"81 layer2_0_bn2_running_var" -> "145 batch_norm_6"  [label="(128,)", style=solid];
+"83 layer2_0_downsample_1_running_mean" -> "147 batch_norm_7"  [label="(128,)", style=solid];
+"84 layer2_0_downsample_1_running_var" -> "147 batch_norm_7"  [label="(128,)", style=solid];
+"86 layer2_1_bn1_running_mean" -> "151 batch_norm_8"  [label="(128,)", style=solid];
+"87 layer2_1_bn1_running_var" -> "151 batch_norm_8"  [label="(128,)", style=solid];
+"89 layer2_1_bn2_running_mean" -> "154 batch_norm_9"  [label="(128,)", style=solid];
+"90 layer2_1_bn2_running_var" -> "154 batch_norm_9"  [label="(128,)", style=solid];
+"92 layer3_0_bn1_running_mean" -> "158 batch_norm_10"  [label="(256,)", style=solid];
+"93 layer3_0_bn1_running_var" -> "158 batch_norm_10"  [label="(256,)", style=solid];
+"95 layer3_0_bn2_running_mean" -> "161 batch_norm_11"  [label="(256,)", style=solid];
+"96 layer3_0_bn2_running_var" -> "161 batch_norm_11"  [label="(256,)", style=solid];
+"98 layer3_0_downsample_1_running_mean" -> "163 batch_norm_12"  [label="(256,)", style=solid];
+"99 layer3_0_downsample_1_running_var" -> "163 batch_norm_12"  [label="(256,)", style=solid];
+"101 layer3_1_bn1_running_mean" -> "167 batch_norm_13"  [label="(256,)", style=solid];
+"102 layer3_1_bn1_running_var" -> "167 batch_norm_13"  [label="(256,)", style=solid];
+"104 layer3_1_bn2_running_mean" -> "170 batch_norm_14"  [label="(256,)", style=solid];
+"105 layer3_1_bn2_running_var" -> "170 batch_norm_14"  [label="(256,)", style=solid];
+"107 layer4_0_bn1_running_mean" -> "174 batch_norm_15"  [label="(512,)", style=solid];
+"108 layer4_0_bn1_running_var" -> "174 batch_norm_15"  [label="(512,)", style=solid];
+"110 layer4_0_bn2_running_mean" -> "177 batch_norm_16"  [label="(512,)", style=solid];
+"111 layer4_0_bn2_running_var" -> "177 batch_norm_16"  [label="(512,)", style=solid];
+"113 layer4_0_downsample_1_running_mean" -> "179 batch_norm_17"  [label="(512,)", style=solid];
+"114 layer4_0_downsample_1_running_var" -> "179 batch_norm_17"  [label="(512,)", style=solid];
+"116 layer4_1_bn1_running_mean" -> "183 batch_norm_18"  [label="(512,)", style=solid];
+"117 layer4_1_bn1_running_var" -> "183 batch_norm_18"  [label="(512,)", style=solid];
+"119 layer4_1_bn2_running_mean" -> "186 batch_norm_19"  [label="(512,)", style=solid];
+"120 layer4_1_bn2_running_var" -> "186 batch_norm_19"  [label="(512,)", style=solid];
+"122 x" -> "123 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"123 conv2d" -> "124 batch_norm"  [label="(1, 64, 112, 112)", style=solid];
+"124 batch_norm" -> "125 relu_"  [label="(1, 64, 112, 112)", style=solid];
+"125 relu_" -> "126 max_pool2d"  [label="(1, 64, 112, 112)", style=solid];
+"126 max_pool2d" -> "127 conv2d_1"  [label="(1, 64, 56, 56)", style=solid];
+"126 max_pool2d" -> "132 add_"  [label="(1, 64, 56, 56)", style=solid];
+"127 conv2d_1" -> "128 batch_norm_1"  [label="(1, 64, 56, 56)", style=solid];
+"128 batch_norm_1" -> "129 relu__1"  [label="(1, 64, 56, 56)", style=solid];
+"129 relu__1" -> "130 conv2d_2"  [label="(1, 64, 56, 56)", style=solid];
+"130 conv2d_2" -> "131 batch_norm_2"  [label="(1, 64, 56, 56)", style=solid];
+"131 batch_norm_2" -> "132 add_"  [label="(1, 64, 56, 56)", style=solid];
+"132 add_" -> "133 relu__2"  [label="(1, 64, 56, 56)", style=solid];
+"133 relu__2" -> "134 conv2d_3"  [label="(1, 64, 56, 56)", style=solid];
+"133 relu__2" -> "139 add__1"  [label="(1, 64, 56, 56)", style=solid];
+"134 conv2d_3" -> "135 batch_norm_3"  [label="(1, 64, 56, 56)", style=solid];
+"135 batch_norm_3" -> "136 relu__3"  [label="(1, 64, 56, 56)", style=solid];
+"136 relu__3" -> "137 conv2d_4"  [label="(1, 64, 56, 56)", style=solid];
+"137 conv2d_4" -> "138 batch_norm_4"  [label="(1, 64, 56, 56)", style=solid];
+"138 batch_norm_4" -> "139 add__1"  [label="(1, 64, 56, 56)", style=solid];
+"139 add__1" -> "140 relu__4"  [label="(1, 64, 56, 56)", style=solid];
+"140 relu__4" -> "141 conv2d_5"  [label="(1, 64, 56, 56)", style=solid];
+"140 relu__4" -> "146 conv2d_7"  [label="(1, 64, 56, 56)", style=solid];
+"141 conv2d_5" -> "142 batch_norm_5"  [label="(1, 128, 28, 28)", style=solid];
+"142 batch_norm_5" -> "143 relu__5"  [label="(1, 128, 28, 28)", style=solid];
+"143 relu__5" -> "144 conv2d_6"  [label="(1, 128, 28, 28)", style=solid];
+"144 conv2d_6" -> "145 batch_norm_6"  [label="(1, 128, 28, 28)", style=solid];
+"145 batch_norm_6" -> "148 add__2"  [label="(1, 128, 28, 28)", style=solid];
+"146 conv2d_7" -> "147 batch_norm_7"  [label="(1, 128, 28, 28)", style=solid];
+"147 batch_norm_7" -> "148 add__2"  [label="(1, 128, 28, 28)", style=solid];
+"148 add__2" -> "149 relu__6"  [label="(1, 128, 28, 28)", style=solid];
+"149 relu__6" -> "150 conv2d_8"  [label="(1, 128, 28, 28)", style=solid];
+"149 relu__6" -> "155 add__3"  [label="(1, 128, 28, 28)", style=solid];
+"150 conv2d_8" -> "151 batch_norm_8"  [label="(1, 128, 28, 28)", style=solid];
+"151 batch_norm_8" -> "152 relu__7"  [label="(1, 128, 28, 28)", style=solid];
+"152 relu__7" -> "153 conv2d_9"  [label="(1, 128, 28, 28)", style=solid];
+"153 conv2d_9" -> "154 batch_norm_9"  [label="(1, 128, 28, 28)", style=solid];
+"154 batch_norm_9" -> "155 add__3"  [label="(1, 128, 28, 28)", style=solid];
+"155 add__3" -> "156 relu__8"  [label="(1, 128, 28, 28)", style=solid];
+"156 relu__8" -> "157 conv2d_10"  [label="(1, 128, 28, 28)", style=solid];
+"156 relu__8" -> "162 conv2d_12"  [label="(1, 128, 28, 28)", style=solid];
+"157 conv2d_10" -> "158 batch_norm_10"  [label="(1, 256, 14, 14)", style=solid];
+"158 batch_norm_10" -> "159 relu__9"  [label="(1, 256, 14, 14)", style=solid];
+"159 relu__9" -> "160 conv2d_11"  [label="(1, 256, 14, 14)", style=solid];
+"160 conv2d_11" -> "161 batch_norm_11"  [label="(1, 256, 14, 14)", style=solid];
+"161 batch_norm_11" -> "164 add__4"  [label="(1, 256, 14, 14)", style=solid];
+"162 conv2d_12" -> "163 batch_norm_12"  [label="(1, 256, 14, 14)", style=solid];
+"163 batch_norm_12" -> "164 add__4"  [label="(1, 256, 14, 14)", style=solid];
+"164 add__4" -> "165 relu__10"  [label="(1, 256, 14, 14)", style=solid];
+"165 relu__10" -> "166 conv2d_13"  [label="(1, 256, 14, 14)", style=solid];
+"165 relu__10" -> "171 add__5"  [label="(1, 256, 14, 14)", style=solid];
+"166 conv2d_13" -> "167 batch_norm_13"  [label="(1, 256, 14, 14)", style=solid];
+"167 batch_norm_13" -> "168 relu__11"  [label="(1, 256, 14, 14)", style=solid];
+"168 relu__11" -> "169 conv2d_14"  [label="(1, 256, 14, 14)", style=solid];
+"169 conv2d_14" -> "170 batch_norm_14"  [label="(1, 256, 14, 14)", style=solid];
+"170 batch_norm_14" -> "171 add__5"  [label="(1, 256, 14, 14)", style=solid];
+"171 add__5" -> "172 relu__12"  [label="(1, 256, 14, 14)", style=solid];
+"172 relu__12" -> "173 conv2d_15"  [label="(1, 256, 14, 14)", style=solid];
+"172 relu__12" -> "178 conv2d_17"  [label="(1, 256, 14, 14)", style=solid];
+"173 conv2d_15" -> "174 batch_norm_15"  [label="(1, 512, 7, 7)", style=solid];
+"174 batch_norm_15" -> "175 relu__13"  [label="(1, 512, 7, 7)", style=solid];
+"175 relu__13" -> "176 conv2d_16"  [label="(1, 512, 7, 7)", style=solid];
+"176 conv2d_16" -> "177 batch_norm_16"  [label="(1, 512, 7, 7)", style=solid];
+"177 batch_norm_16" -> "180 add__6"  [label="(1, 512, 7, 7)", style=solid];
+"178 conv2d_17" -> "179 batch_norm_17"  [label="(1, 512, 7, 7)", style=solid];
+"179 batch_norm_17" -> "180 add__6"  [label="(1, 512, 7, 7)", style=solid];
+"180 add__6" -> "181 relu__14"  [label="(1, 512, 7, 7)", style=solid];
+"181 relu__14" -> "182 conv2d_18"  [label="(1, 512, 7, 7)", style=solid];
+"181 relu__14" -> "187 add__7"  [label="(1, 512, 7, 7)", style=solid];
+"182 conv2d_18" -> "183 batch_norm_18"  [label="(1, 512, 7, 7)", style=solid];
+"183 batch_norm_18" -> "184 relu__15"  [label="(1, 512, 7, 7)", style=solid];
+"184 relu__15" -> "185 conv2d_19"  [label="(1, 512, 7, 7)", style=solid];
+"185 conv2d_19" -> "186 batch_norm_19"  [label="(1, 512, 7, 7)", style=solid];
+"186 batch_norm_19" -> "187 add__7"  [label="(1, 512, 7, 7)", style=solid];
+"187 add__7" -> "188 relu__16"  [label="(1, 512, 7, 7)", style=solid];
+"188 relu__16" -> "189 adaptive_avg_pool2d"  [label="(1, 512, 7, 7)", style=solid];
+"189 adaptive_avg_pool2d" -> "190 flatten"  [label="(1, 512, 1, 1)", style=solid];
+"190 flatten" -> "191 linear"  [label="(1, 512)", style=solid];
+"191 linear" -> "192 output_1"  [label="(1, 1000)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/swin_v2_s.dot b/tests/torch/data/reference_graphs/fx/swin_v2_s.dot
index 52e2bfdc398..1371ee266f8 100644
--- a/tests/torch/data/reference_graphs/fx/swin_v2_s.dot
+++ b/tests/torch/data/reference_graphs/fx/swin_v2_s.dot
@@ -1,5610 +1,5610 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d" [id=3, type=conv2d];
-"4 permute" [id=4, type=permute];
-"5 _param_constant2" [id=5, type=get_attr];
-"6 _param_constant3" [id=6, type=get_attr];
-"7 layer_norm" [id=7, type=layer_norm];
-"8 _tensor_constant0" [id=8, type=get_attr];
-"9 _param_constant4" [id=9, type=get_attr];
-"10 _param_constant5" [id=10, type=get_attr];
-"11 linear" [id=11, type=linear];
-"12 relu_" [id=12, type=relu_];
-"13 _param_constant6" [id=13, type=get_attr];
-"14 linear_1" [id=14, type=linear];
-"15 view" [id=15, type=view];
-"16 _tensor_constant1" [id=16, type=get_attr];
-"17 index" [id=17, type=index];
-"18 view_1" [id=18, type=view];
-"19 permute_1" [id=19, type=permute];
-"20 contiguous" [id=20, type=contiguous];
-"21 unsqueeze" [id=21, type=unsqueeze];
-"22 sigmoid" [id=22, type=sigmoid];
-"23 mul" [id=23, type=mul];
-"24 pad" [id=24, type=pad];
-"25 view_2" [id=25, type=view];
-"26 permute_2" [id=26, type=permute];
-"27 reshape" [id=27, type=reshape];
-"28 _param_constant7" [id=28, type=get_attr];
-"29 clone" [id=29, type=clone];
-"30 slice_1" [id=30, type=slice];
-"31 zero_" [id=31, type=zero_];
-"32 _param_constant8" [id=32, type=get_attr];
-"33 linear_2" [id=33, type=linear];
-"34 reshape_1" [id=34, type=reshape];
-"35 permute_3" [id=35, type=permute];
-"36 select" [id=36, type=select];
-"37 select_1" [id=37, type=select];
-"38 select_2" [id=38, type=select];
-"39 linalg_vector_norm" [id=39, type=linalg_vector_norm];
-"40 clamp_min" [id=40, type=clamp_min];
-"41 expand_as" [id=41, type=expand_as];
-"42 div" [id=42, type=div];
-"43 linalg_vector_norm_1" [id=43, type=linalg_vector_norm];
-"44 clamp_min_1" [id=44, type=clamp_min];
-"45 expand_as_1" [id=45, type=expand_as];
-"46 div_1" [id=46, type=div];
-"47 transpose" [id=47, type=transpose];
-"48 matmul" [id=48, type=matmul];
-"49 _param_constant9" [id=49, type=get_attr];
-"50 clamp" [id=50, type=clamp];
-"51 exp" [id=51, type=exp];
-"52 mul_1" [id=52, type=mul];
-"53 add" [id=53, type=add];
-"54 softmax" [id=54, type=softmax];
-"55 dropout" [id=55, type=dropout];
-"56 matmul_1" [id=56, type=matmul];
-"57 transpose_1" [id=57, type=transpose];
-"58 reshape_2" [id=58, type=reshape];
-"59 _param_constant10" [id=59, type=get_attr];
-"60 _param_constant11" [id=60, type=get_attr];
-"61 linear_3" [id=61, type=linear];
-"62 dropout_1" [id=62, type=dropout];
-"63 view_3" [id=63, type=view];
-"64 permute_4" [id=64, type=permute];
-"65 reshape_3" [id=65, type=reshape];
-"66 slice_2" [id=66, type=slice];
-"67 slice_3" [id=67, type=slice];
-"68 _param_constant12" [id=68, type=get_attr];
-"69 _param_constant13" [id=69, type=get_attr];
-"70 layer_norm_1" [id=70, type=layer_norm];
-"71 add_1" [id=71, type=add];
-"72 _param_constant14" [id=72, type=get_attr];
-"73 _param_constant15" [id=73, type=get_attr];
-"74 linear_4" [id=74, type=linear];
-"75 gelu" [id=75, type=gelu];
-"76 dropout_2" [id=76, type=dropout];
-"77 _param_constant16" [id=77, type=get_attr];
-"78 _param_constant17" [id=78, type=get_attr];
-"79 linear_5" [id=79, type=linear];
-"80 dropout_3" [id=80, type=dropout];
-"81 _param_constant18" [id=81, type=get_attr];
-"82 _param_constant19" [id=82, type=get_attr];
-"83 layer_norm_2" [id=83, type=layer_norm];
-"84 add_2" [id=84, type=add];
-"85 _tensor_constant2" [id=85, type=get_attr];
-"86 _param_constant20" [id=86, type=get_attr];
-"87 _param_constant21" [id=87, type=get_attr];
-"88 linear_6" [id=88, type=linear];
-"89 relu__1" [id=89, type=relu_];
-"90 _param_constant22" [id=90, type=get_attr];
-"91 linear_7" [id=91, type=linear];
-"92 view_4" [id=92, type=view];
-"93 _tensor_constant3" [id=93, type=get_attr];
-"94 index_1" [id=94, type=index];
-"95 view_5" [id=95, type=view];
-"96 permute_5" [id=96, type=permute];
-"97 contiguous_1" [id=97, type=contiguous];
-"98 unsqueeze_1" [id=98, type=unsqueeze];
-"99 sigmoid_1" [id=99, type=sigmoid];
-"100 mul_2" [id=100, type=mul];
-"101 pad_1" [id=101, type=pad];
-"102 roll" [id=102, type=roll];
-"103 view_6" [id=103, type=view];
-"104 permute_6" [id=104, type=permute];
-"105 reshape_4" [id=105, type=reshape];
-"106 _param_constant23" [id=106, type=get_attr];
-"107 clone_1" [id=107, type=clone];
-"108 slice_4" [id=108, type=slice];
-"109 zero__1" [id=109, type=zero_];
-"110 _param_constant24" [id=110, type=get_attr];
-"111 linear_8" [id=111, type=linear];
-"112 reshape_5" [id=112, type=reshape];
-"113 permute_7" [id=113, type=permute];
-"114 select_3" [id=114, type=select];
-"115 select_4" [id=115, type=select];
-"116 select_5" [id=116, type=select];
-"117 linalg_vector_norm_2" [id=117, type=linalg_vector_norm];
-"118 clamp_min_2" [id=118, type=clamp_min];
-"119 expand_as_2" [id=119, type=expand_as];
-"120 div_2" [id=120, type=div];
-"121 linalg_vector_norm_3" [id=121, type=linalg_vector_norm];
-"122 clamp_min_3" [id=122, type=clamp_min];
-"123 expand_as_3" [id=123, type=expand_as];
-"124 div_3" [id=124, type=div];
-"125 transpose_2" [id=125, type=transpose];
-"126 matmul_2" [id=126, type=matmul];
-"127 _param_constant25" [id=127, type=get_attr];
-"128 clamp_1" [id=128, type=clamp];
-"129 exp_1" [id=129, type=exp];
-"130 mul_3" [id=130, type=mul];
-"131 add_3" [id=131, type=add];
-"132 new_zeros" [id=132, type=new_zeros];
-"133 _tensor_constant4" [id=133, type=get_attr];
-"134 lift_fresh_copy" [id=134, type=lift_fresh_copy];
-"135 slice_5" [id=135, type=slice];
-"136 slice_6" [id=136, type=slice];
-"137 fill_" [id=137, type=fill_];
-"138 _tensor_constant5" [id=138, type=get_attr];
-"139 lift_fresh_copy_1" [id=139, type=lift_fresh_copy];
-"140 slice_7" [id=140, type=slice];
-"141 slice_8" [id=141, type=slice];
-"142 fill__1" [id=142, type=fill_];
-"143 _tensor_constant6" [id=143, type=get_attr];
-"144 lift_fresh_copy_2" [id=144, type=lift_fresh_copy];
-"145 slice_9" [id=145, type=slice];
-"146 slice_10" [id=146, type=slice];
-"147 fill__2" [id=147, type=fill_];
-"148 _tensor_constant7" [id=148, type=get_attr];
-"149 lift_fresh_copy_3" [id=149, type=lift_fresh_copy];
-"150 slice_11" [id=150, type=slice];
-"151 slice_12" [id=151, type=slice];
-"152 fill__3" [id=152, type=fill_];
-"153 _tensor_constant8" [id=153, type=get_attr];
-"154 lift_fresh_copy_4" [id=154, type=lift_fresh_copy];
-"155 slice_13" [id=155, type=slice];
-"156 slice_14" [id=156, type=slice];
-"157 fill__4" [id=157, type=fill_];
-"158 _tensor_constant9" [id=158, type=get_attr];
-"159 lift_fresh_copy_5" [id=159, type=lift_fresh_copy];
-"160 slice_15" [id=160, type=slice];
-"161 slice_16" [id=161, type=slice];
-"162 fill__5" [id=162, type=fill_];
-"163 _tensor_constant10" [id=163, type=get_attr];
-"164 lift_fresh_copy_6" [id=164, type=lift_fresh_copy];
-"165 slice_17" [id=165, type=slice];
-"166 slice_18" [id=166, type=slice];
-"167 fill__6" [id=167, type=fill_];
-"168 _tensor_constant11" [id=168, type=get_attr];
-"169 lift_fresh_copy_7" [id=169, type=lift_fresh_copy];
-"170 slice_19" [id=170, type=slice];
-"171 slice_20" [id=171, type=slice];
-"172 fill__7" [id=172, type=fill_];
-"173 _tensor_constant12" [id=173, type=get_attr];
-"174 lift_fresh_copy_8" [id=174, type=lift_fresh_copy];
-"175 slice_21" [id=175, type=slice];
-"176 slice_22" [id=176, type=slice];
-"177 fill__8" [id=177, type=fill_];
-"178 view_7" [id=178, type=view];
-"179 permute_8" [id=179, type=permute];
-"180 reshape_6" [id=180, type=reshape];
-"181 unsqueeze_2" [id=181, type=unsqueeze];
-"182 unsqueeze_3" [id=182, type=unsqueeze];
-"183 sub" [id=183, type=sub];
-"184 ne" [id=184, type=ne];
-"185 masked_fill" [id=185, type=masked_fill];
-"186 eq" [id=186, type=eq];
-"187 masked_fill_1" [id=187, type=masked_fill];
-"188 view_8" [id=188, type=view];
-"189 unsqueeze_4" [id=189, type=unsqueeze];
-"190 unsqueeze_5" [id=190, type=unsqueeze];
-"191 add_4" [id=191, type=add];
-"192 view_9" [id=192, type=view];
-"193 softmax_1" [id=193, type=softmax];
-"194 dropout_4" [id=194, type=dropout];
-"195 matmul_3" [id=195, type=matmul];
-"196 transpose_3" [id=196, type=transpose];
-"197 reshape_7" [id=197, type=reshape];
-"198 _param_constant26" [id=198, type=get_attr];
-"199 _param_constant27" [id=199, type=get_attr];
-"200 linear_9" [id=200, type=linear];
-"201 dropout_5" [id=201, type=dropout];
-"202 view_10" [id=202, type=view];
-"203 permute_9" [id=203, type=permute];
-"204 reshape_8" [id=204, type=reshape];
-"205 roll_1" [id=205, type=roll];
-"206 slice_23" [id=206, type=slice];
-"207 slice_24" [id=207, type=slice];
-"208 _param_constant28" [id=208, type=get_attr];
-"209 _param_constant29" [id=209, type=get_attr];
-"210 layer_norm_3" [id=210, type=layer_norm];
-"211 add_5" [id=211, type=add];
-"212 _param_constant30" [id=212, type=get_attr];
-"213 _param_constant31" [id=213, type=get_attr];
-"214 linear_10" [id=214, type=linear];
-"215 gelu_1" [id=215, type=gelu];
-"216 dropout_6" [id=216, type=dropout];
-"217 _param_constant32" [id=217, type=get_attr];
-"218 _param_constant33" [id=218, type=get_attr];
-"219 linear_11" [id=219, type=linear];
-"220 dropout_7" [id=220, type=dropout];
-"221 _param_constant34" [id=221, type=get_attr];
-"222 _param_constant35" [id=222, type=get_attr];
-"223 layer_norm_4" [id=223, type=layer_norm];
-"224 add_6" [id=224, type=add];
-"225 pad_2" [id=225, type=pad];
-"226 slice_25" [id=226, type=slice];
-"227 slice_26" [id=227, type=slice];
-"228 slice_27" [id=228, type=slice];
-"229 slice_28" [id=229, type=slice];
-"230 slice_29" [id=230, type=slice];
-"231 slice_30" [id=231, type=slice];
-"232 slice_31" [id=232, type=slice];
-"233 slice_32" [id=233, type=slice];
-"234 slice_33" [id=234, type=slice];
-"235 slice_34" [id=235, type=slice];
-"236 slice_35" [id=236, type=slice];
-"237 slice_36" [id=237, type=slice];
-"238 cat" [id=238, type=cat];
-"239 _param_constant36" [id=239, type=get_attr];
-"240 linear_12" [id=240, type=linear];
-"241 _param_constant37" [id=241, type=get_attr];
-"242 _param_constant38" [id=242, type=get_attr];
-"243 layer_norm_5" [id=243, type=layer_norm];
-"244 _tensor_constant13" [id=244, type=get_attr];
-"245 _param_constant39" [id=245, type=get_attr];
-"246 _param_constant40" [id=246, type=get_attr];
-"247 linear_13" [id=247, type=linear];
-"248 relu__2" [id=248, type=relu_];
-"249 _param_constant41" [id=249, type=get_attr];
-"250 linear_14" [id=250, type=linear];
-"251 view_11" [id=251, type=view];
-"252 _tensor_constant14" [id=252, type=get_attr];
-"253 index_2" [id=253, type=index];
-"254 view_12" [id=254, type=view];
-"255 permute_10" [id=255, type=permute];
-"256 contiguous_2" [id=256, type=contiguous];
-"257 unsqueeze_6" [id=257, type=unsqueeze];
-"258 sigmoid_2" [id=258, type=sigmoid];
-"259 mul_4" [id=259, type=mul];
-"260 pad_3" [id=260, type=pad];
-"261 view_13" [id=261, type=view];
-"262 permute_11" [id=262, type=permute];
-"263 reshape_9" [id=263, type=reshape];
-"264 _param_constant42" [id=264, type=get_attr];
-"265 clone_2" [id=265, type=clone];
-"266 slice_37" [id=266, type=slice];
-"267 zero__2" [id=267, type=zero_];
-"268 _param_constant43" [id=268, type=get_attr];
-"269 linear_15" [id=269, type=linear];
-"270 reshape_10" [id=270, type=reshape];
-"271 permute_12" [id=271, type=permute];
-"272 select_6" [id=272, type=select];
-"273 select_7" [id=273, type=select];
-"274 select_8" [id=274, type=select];
-"275 linalg_vector_norm_4" [id=275, type=linalg_vector_norm];
-"276 clamp_min_4" [id=276, type=clamp_min];
-"277 expand_as_4" [id=277, type=expand_as];
-"278 div_4" [id=278, type=div];
-"279 linalg_vector_norm_5" [id=279, type=linalg_vector_norm];
-"280 clamp_min_5" [id=280, type=clamp_min];
-"281 expand_as_5" [id=281, type=expand_as];
-"282 div_5" [id=282, type=div];
-"283 transpose_4" [id=283, type=transpose];
-"284 matmul_4" [id=284, type=matmul];
-"285 _param_constant44" [id=285, type=get_attr];
-"286 clamp_2" [id=286, type=clamp];
-"287 exp_2" [id=287, type=exp];
-"288 mul_5" [id=288, type=mul];
-"289 add_7" [id=289, type=add];
-"290 softmax_2" [id=290, type=softmax];
-"291 dropout_8" [id=291, type=dropout];
-"292 matmul_5" [id=292, type=matmul];
-"293 transpose_5" [id=293, type=transpose];
-"294 reshape_11" [id=294, type=reshape];
-"295 _param_constant45" [id=295, type=get_attr];
-"296 _param_constant46" [id=296, type=get_attr];
-"297 linear_16" [id=297, type=linear];
-"298 dropout_9" [id=298, type=dropout];
-"299 view_14" [id=299, type=view];
-"300 permute_13" [id=300, type=permute];
-"301 reshape_12" [id=301, type=reshape];
-"302 slice_38" [id=302, type=slice];
-"303 slice_39" [id=303, type=slice];
-"304 slice_40" [id=304, type=slice];
-"305 slice_41" [id=305, type=slice];
-"306 contiguous_3" [id=306, type=contiguous];
-"307 _param_constant47" [id=307, type=get_attr];
-"308 _param_constant48" [id=308, type=get_attr];
-"309 layer_norm_6" [id=309, type=layer_norm];
-"310 add_8" [id=310, type=add];
-"311 _param_constant49" [id=311, type=get_attr];
-"312 _param_constant50" [id=312, type=get_attr];
-"313 linear_17" [id=313, type=linear];
-"314 gelu_2" [id=314, type=gelu];
-"315 dropout_10" [id=315, type=dropout];
-"316 _param_constant51" [id=316, type=get_attr];
-"317 _param_constant52" [id=317, type=get_attr];
-"318 linear_18" [id=318, type=linear];
-"319 dropout_11" [id=319, type=dropout];
-"320 _param_constant53" [id=320, type=get_attr];
-"321 _param_constant54" [id=321, type=get_attr];
-"322 layer_norm_7" [id=322, type=layer_norm];
-"323 add_9" [id=323, type=add];
-"324 _tensor_constant15" [id=324, type=get_attr];
-"325 _param_constant55" [id=325, type=get_attr];
-"326 _param_constant56" [id=326, type=get_attr];
-"327 linear_19" [id=327, type=linear];
-"328 relu__3" [id=328, type=relu_];
-"329 _param_constant57" [id=329, type=get_attr];
-"330 linear_20" [id=330, type=linear];
-"331 view_15" [id=331, type=view];
-"332 _tensor_constant16" [id=332, type=get_attr];
-"333 index_3" [id=333, type=index];
-"334 view_16" [id=334, type=view];
-"335 permute_14" [id=335, type=permute];
-"336 contiguous_4" [id=336, type=contiguous];
-"337 unsqueeze_7" [id=337, type=unsqueeze];
-"338 sigmoid_3" [id=338, type=sigmoid];
-"339 mul_6" [id=339, type=mul];
-"340 pad_4" [id=340, type=pad];
-"341 roll_2" [id=341, type=roll];
-"342 view_17" [id=342, type=view];
-"343 permute_15" [id=343, type=permute];
-"344 reshape_13" [id=344, type=reshape];
-"345 _param_constant58" [id=345, type=get_attr];
-"346 clone_3" [id=346, type=clone];
-"347 slice_42" [id=347, type=slice];
-"348 zero__3" [id=348, type=zero_];
-"349 _param_constant59" [id=349, type=get_attr];
-"350 linear_21" [id=350, type=linear];
-"351 reshape_14" [id=351, type=reshape];
-"352 permute_16" [id=352, type=permute];
-"353 select_9" [id=353, type=select];
-"354 select_10" [id=354, type=select];
-"355 select_11" [id=355, type=select];
-"356 linalg_vector_norm_6" [id=356, type=linalg_vector_norm];
-"357 clamp_min_6" [id=357, type=clamp_min];
-"358 expand_as_6" [id=358, type=expand_as];
-"359 div_6" [id=359, type=div];
-"360 linalg_vector_norm_7" [id=360, type=linalg_vector_norm];
-"361 clamp_min_7" [id=361, type=clamp_min];
-"362 expand_as_7" [id=362, type=expand_as];
-"363 div_7" [id=363, type=div];
-"364 transpose_6" [id=364, type=transpose];
-"365 matmul_6" [id=365, type=matmul];
-"366 _param_constant60" [id=366, type=get_attr];
-"367 clamp_3" [id=367, type=clamp];
-"368 exp_3" [id=368, type=exp];
-"369 mul_7" [id=369, type=mul];
-"370 add_10" [id=370, type=add];
-"371 new_zeros_1" [id=371, type=new_zeros];
-"372 _tensor_constant17" [id=372, type=get_attr];
-"373 lift_fresh_copy_9" [id=373, type=lift_fresh_copy];
-"374 slice_43" [id=374, type=slice];
-"375 slice_44" [id=375, type=slice];
-"376 fill__9" [id=376, type=fill_];
-"377 _tensor_constant18" [id=377, type=get_attr];
-"378 lift_fresh_copy_10" [id=378, type=lift_fresh_copy];
-"379 slice_45" [id=379, type=slice];
-"380 slice_46" [id=380, type=slice];
-"381 fill__10" [id=381, type=fill_];
-"382 _tensor_constant19" [id=382, type=get_attr];
-"383 lift_fresh_copy_11" [id=383, type=lift_fresh_copy];
-"384 slice_47" [id=384, type=slice];
-"385 slice_48" [id=385, type=slice];
-"386 fill__11" [id=386, type=fill_];
-"387 _tensor_constant20" [id=387, type=get_attr];
-"388 lift_fresh_copy_12" [id=388, type=lift_fresh_copy];
-"389 slice_49" [id=389, type=slice];
-"390 slice_50" [id=390, type=slice];
-"391 fill__12" [id=391, type=fill_];
-"392 _tensor_constant21" [id=392, type=get_attr];
-"393 lift_fresh_copy_13" [id=393, type=lift_fresh_copy];
-"394 slice_51" [id=394, type=slice];
-"395 slice_52" [id=395, type=slice];
-"396 fill__13" [id=396, type=fill_];
-"397 _tensor_constant22" [id=397, type=get_attr];
-"398 lift_fresh_copy_14" [id=398, type=lift_fresh_copy];
-"399 slice_53" [id=399, type=slice];
-"400 slice_54" [id=400, type=slice];
-"401 fill__14" [id=401, type=fill_];
-"402 _tensor_constant23" [id=402, type=get_attr];
-"403 lift_fresh_copy_15" [id=403, type=lift_fresh_copy];
-"404 slice_55" [id=404, type=slice];
-"405 slice_56" [id=405, type=slice];
-"406 fill__15" [id=406, type=fill_];
-"407 _tensor_constant24" [id=407, type=get_attr];
-"408 lift_fresh_copy_16" [id=408, type=lift_fresh_copy];
-"409 slice_57" [id=409, type=slice];
-"410 slice_58" [id=410, type=slice];
-"411 fill__16" [id=411, type=fill_];
-"412 _tensor_constant25" [id=412, type=get_attr];
-"413 lift_fresh_copy_17" [id=413, type=lift_fresh_copy];
-"414 slice_59" [id=414, type=slice];
-"415 slice_60" [id=415, type=slice];
-"416 fill__17" [id=416, type=fill_];
-"417 view_18" [id=417, type=view];
-"418 permute_17" [id=418, type=permute];
-"419 reshape_15" [id=419, type=reshape];
-"420 unsqueeze_8" [id=420, type=unsqueeze];
-"421 unsqueeze_9" [id=421, type=unsqueeze];
-"422 sub_1" [id=422, type=sub];
-"423 ne_1" [id=423, type=ne];
-"424 masked_fill_2" [id=424, type=masked_fill];
-"425 eq_1" [id=425, type=eq];
-"426 masked_fill_3" [id=426, type=masked_fill];
-"427 view_19" [id=427, type=view];
-"428 unsqueeze_10" [id=428, type=unsqueeze];
-"429 unsqueeze_11" [id=429, type=unsqueeze];
-"430 add_11" [id=430, type=add];
-"431 view_20" [id=431, type=view];
-"432 softmax_3" [id=432, type=softmax];
-"433 dropout_12" [id=433, type=dropout];
-"434 matmul_7" [id=434, type=matmul];
-"435 transpose_7" [id=435, type=transpose];
-"436 reshape_16" [id=436, type=reshape];
-"437 _param_constant61" [id=437, type=get_attr];
-"438 _param_constant62" [id=438, type=get_attr];
-"439 linear_22" [id=439, type=linear];
-"440 dropout_13" [id=440, type=dropout];
-"441 view_21" [id=441, type=view];
-"442 permute_18" [id=442, type=permute];
-"443 reshape_17" [id=443, type=reshape];
-"444 roll_3" [id=444, type=roll];
-"445 slice_61" [id=445, type=slice];
-"446 slice_62" [id=446, type=slice];
-"447 slice_63" [id=447, type=slice];
-"448 slice_64" [id=448, type=slice];
-"449 contiguous_5" [id=449, type=contiguous];
-"450 _param_constant63" [id=450, type=get_attr];
-"451 _param_constant64" [id=451, type=get_attr];
-"452 layer_norm_8" [id=452, type=layer_norm];
-"453 add_12" [id=453, type=add];
-"454 _param_constant65" [id=454, type=get_attr];
-"455 _param_constant66" [id=455, type=get_attr];
-"456 linear_23" [id=456, type=linear];
-"457 gelu_3" [id=457, type=gelu];
-"458 dropout_14" [id=458, type=dropout];
-"459 _param_constant67" [id=459, type=get_attr];
-"460 _param_constant68" [id=460, type=get_attr];
-"461 linear_24" [id=461, type=linear];
-"462 dropout_15" [id=462, type=dropout];
-"463 _param_constant69" [id=463, type=get_attr];
-"464 _param_constant70" [id=464, type=get_attr];
-"465 layer_norm_9" [id=465, type=layer_norm];
-"466 add_13" [id=466, type=add];
-"467 pad_5" [id=467, type=pad];
-"468 slice_65" [id=468, type=slice];
-"469 slice_66" [id=469, type=slice];
-"470 slice_67" [id=470, type=slice];
-"471 slice_68" [id=471, type=slice];
-"472 slice_69" [id=472, type=slice];
-"473 slice_70" [id=473, type=slice];
-"474 slice_71" [id=474, type=slice];
-"475 slice_72" [id=475, type=slice];
-"476 slice_73" [id=476, type=slice];
-"477 slice_74" [id=477, type=slice];
-"478 slice_75" [id=478, type=slice];
-"479 slice_76" [id=479, type=slice];
-"480 cat_1" [id=480, type=cat];
-"481 _param_constant71" [id=481, type=get_attr];
-"482 linear_25" [id=482, type=linear];
-"483 _param_constant72" [id=483, type=get_attr];
-"484 _param_constant73" [id=484, type=get_attr];
-"485 layer_norm_10" [id=485, type=layer_norm];
-"486 _tensor_constant26" [id=486, type=get_attr];
-"487 _param_constant74" [id=487, type=get_attr];
-"488 _param_constant75" [id=488, type=get_attr];
-"489 linear_26" [id=489, type=linear];
-"490 relu__4" [id=490, type=relu_];
-"491 _param_constant76" [id=491, type=get_attr];
-"492 linear_27" [id=492, type=linear];
-"493 view_22" [id=493, type=view];
-"494 _tensor_constant27" [id=494, type=get_attr];
-"495 index_4" [id=495, type=index];
-"496 view_23" [id=496, type=view];
-"497 permute_19" [id=497, type=permute];
-"498 contiguous_6" [id=498, type=contiguous];
-"499 unsqueeze_12" [id=499, type=unsqueeze];
-"500 sigmoid_4" [id=500, type=sigmoid];
-"501 mul_8" [id=501, type=mul];
-"502 pad_6" [id=502, type=pad];
-"503 view_24" [id=503, type=view];
-"504 permute_20" [id=504, type=permute];
-"505 reshape_18" [id=505, type=reshape];
-"506 _param_constant77" [id=506, type=get_attr];
-"507 clone_4" [id=507, type=clone];
-"508 slice_77" [id=508, type=slice];
-"509 zero__4" [id=509, type=zero_];
-"510 _param_constant78" [id=510, type=get_attr];
-"511 linear_28" [id=511, type=linear];
-"512 reshape_19" [id=512, type=reshape];
-"513 permute_21" [id=513, type=permute];
-"514 select_12" [id=514, type=select];
-"515 select_13" [id=515, type=select];
-"516 select_14" [id=516, type=select];
-"517 linalg_vector_norm_8" [id=517, type=linalg_vector_norm];
-"518 clamp_min_8" [id=518, type=clamp_min];
-"519 expand_as_8" [id=519, type=expand_as];
-"520 div_8" [id=520, type=div];
-"521 linalg_vector_norm_9" [id=521, type=linalg_vector_norm];
-"522 clamp_min_9" [id=522, type=clamp_min];
-"523 expand_as_9" [id=523, type=expand_as];
-"524 div_9" [id=524, type=div];
-"525 transpose_8" [id=525, type=transpose];
-"526 matmul_8" [id=526, type=matmul];
-"527 _param_constant79" [id=527, type=get_attr];
-"528 clamp_4" [id=528, type=clamp];
-"529 exp_4" [id=529, type=exp];
-"530 mul_9" [id=530, type=mul];
-"531 add_14" [id=531, type=add];
-"532 softmax_4" [id=532, type=softmax];
-"533 dropout_16" [id=533, type=dropout];
-"534 matmul_9" [id=534, type=matmul];
-"535 transpose_9" [id=535, type=transpose];
-"536 reshape_20" [id=536, type=reshape];
-"537 _param_constant80" [id=537, type=get_attr];
-"538 _param_constant81" [id=538, type=get_attr];
-"539 linear_29" [id=539, type=linear];
-"540 dropout_17" [id=540, type=dropout];
-"541 view_25" [id=541, type=view];
-"542 permute_22" [id=542, type=permute];
-"543 reshape_21" [id=543, type=reshape];
-"544 slice_78" [id=544, type=slice];
-"545 slice_79" [id=545, type=slice];
-"546 slice_80" [id=546, type=slice];
-"547 slice_81" [id=547, type=slice];
-"548 contiguous_7" [id=548, type=contiguous];
-"549 _param_constant82" [id=549, type=get_attr];
-"550 _param_constant83" [id=550, type=get_attr];
-"551 layer_norm_11" [id=551, type=layer_norm];
-"552 add_15" [id=552, type=add];
-"553 _param_constant84" [id=553, type=get_attr];
-"554 _param_constant85" [id=554, type=get_attr];
-"555 linear_30" [id=555, type=linear];
-"556 gelu_4" [id=556, type=gelu];
-"557 dropout_18" [id=557, type=dropout];
-"558 _param_constant86" [id=558, type=get_attr];
-"559 _param_constant87" [id=559, type=get_attr];
-"560 linear_31" [id=560, type=linear];
-"561 dropout_19" [id=561, type=dropout];
-"562 _param_constant88" [id=562, type=get_attr];
-"563 _param_constant89" [id=563, type=get_attr];
-"564 layer_norm_12" [id=564, type=layer_norm];
-"565 add_16" [id=565, type=add];
-"566 _tensor_constant28" [id=566, type=get_attr];
-"567 _param_constant90" [id=567, type=get_attr];
-"568 _param_constant91" [id=568, type=get_attr];
-"569 linear_32" [id=569, type=linear];
-"570 relu__5" [id=570, type=relu_];
-"571 _param_constant92" [id=571, type=get_attr];
-"572 linear_33" [id=572, type=linear];
-"573 view_26" [id=573, type=view];
-"574 _tensor_constant29" [id=574, type=get_attr];
-"575 index_5" [id=575, type=index];
-"576 view_27" [id=576, type=view];
-"577 permute_23" [id=577, type=permute];
-"578 contiguous_8" [id=578, type=contiguous];
-"579 unsqueeze_13" [id=579, type=unsqueeze];
-"580 sigmoid_5" [id=580, type=sigmoid];
-"581 mul_10" [id=581, type=mul];
-"582 pad_7" [id=582, type=pad];
-"583 roll_4" [id=583, type=roll];
-"584 view_28" [id=584, type=view];
-"585 permute_24" [id=585, type=permute];
-"586 reshape_22" [id=586, type=reshape];
-"587 _param_constant93" [id=587, type=get_attr];
-"588 clone_5" [id=588, type=clone];
-"589 slice_82" [id=589, type=slice];
-"590 zero__5" [id=590, type=zero_];
-"591 _param_constant94" [id=591, type=get_attr];
-"592 linear_34" [id=592, type=linear];
-"593 reshape_23" [id=593, type=reshape];
-"594 permute_25" [id=594, type=permute];
-"595 select_15" [id=595, type=select];
-"596 select_16" [id=596, type=select];
-"597 select_17" [id=597, type=select];
-"598 linalg_vector_norm_10" [id=598, type=linalg_vector_norm];
-"599 clamp_min_10" [id=599, type=clamp_min];
-"600 expand_as_10" [id=600, type=expand_as];
-"601 div_10" [id=601, type=div];
-"602 linalg_vector_norm_11" [id=602, type=linalg_vector_norm];
-"603 clamp_min_11" [id=603, type=clamp_min];
-"604 expand_as_11" [id=604, type=expand_as];
-"605 div_11" [id=605, type=div];
-"606 transpose_10" [id=606, type=transpose];
-"607 matmul_10" [id=607, type=matmul];
-"608 _param_constant95" [id=608, type=get_attr];
-"609 clamp_5" [id=609, type=clamp];
-"610 exp_5" [id=610, type=exp];
-"611 mul_11" [id=611, type=mul];
-"612 add_17" [id=612, type=add];
-"613 new_zeros_2" [id=613, type=new_zeros];
-"614 _tensor_constant30" [id=614, type=get_attr];
-"615 lift_fresh_copy_18" [id=615, type=lift_fresh_copy];
-"616 slice_83" [id=616, type=slice];
-"617 slice_84" [id=617, type=slice];
-"618 fill__18" [id=618, type=fill_];
-"619 _tensor_constant31" [id=619, type=get_attr];
-"620 lift_fresh_copy_19" [id=620, type=lift_fresh_copy];
-"621 slice_85" [id=621, type=slice];
-"622 slice_86" [id=622, type=slice];
-"623 fill__19" [id=623, type=fill_];
-"624 _tensor_constant32" [id=624, type=get_attr];
-"625 lift_fresh_copy_20" [id=625, type=lift_fresh_copy];
-"626 slice_87" [id=626, type=slice];
-"627 slice_88" [id=627, type=slice];
-"628 fill__20" [id=628, type=fill_];
-"629 _tensor_constant33" [id=629, type=get_attr];
-"630 lift_fresh_copy_21" [id=630, type=lift_fresh_copy];
-"631 slice_89" [id=631, type=slice];
-"632 slice_90" [id=632, type=slice];
-"633 fill__21" [id=633, type=fill_];
-"634 _tensor_constant34" [id=634, type=get_attr];
-"635 lift_fresh_copy_22" [id=635, type=lift_fresh_copy];
-"636 slice_91" [id=636, type=slice];
-"637 slice_92" [id=637, type=slice];
-"638 fill__22" [id=638, type=fill_];
-"639 _tensor_constant35" [id=639, type=get_attr];
-"640 lift_fresh_copy_23" [id=640, type=lift_fresh_copy];
-"641 slice_93" [id=641, type=slice];
-"642 slice_94" [id=642, type=slice];
-"643 fill__23" [id=643, type=fill_];
-"644 _tensor_constant36" [id=644, type=get_attr];
-"645 lift_fresh_copy_24" [id=645, type=lift_fresh_copy];
-"646 slice_95" [id=646, type=slice];
-"647 slice_96" [id=647, type=slice];
-"648 fill__24" [id=648, type=fill_];
-"649 _tensor_constant37" [id=649, type=get_attr];
-"650 lift_fresh_copy_25" [id=650, type=lift_fresh_copy];
-"651 slice_97" [id=651, type=slice];
-"652 slice_98" [id=652, type=slice];
-"653 fill__25" [id=653, type=fill_];
-"654 _tensor_constant38" [id=654, type=get_attr];
-"655 lift_fresh_copy_26" [id=655, type=lift_fresh_copy];
-"656 slice_99" [id=656, type=slice];
-"657 slice_100" [id=657, type=slice];
-"658 fill__26" [id=658, type=fill_];
-"659 view_29" [id=659, type=view];
-"660 permute_26" [id=660, type=permute];
-"661 reshape_24" [id=661, type=reshape];
-"662 unsqueeze_14" [id=662, type=unsqueeze];
-"663 unsqueeze_15" [id=663, type=unsqueeze];
-"664 sub_2" [id=664, type=sub];
-"665 ne_2" [id=665, type=ne];
-"666 masked_fill_4" [id=666, type=masked_fill];
-"667 eq_2" [id=667, type=eq];
-"668 masked_fill_5" [id=668, type=masked_fill];
-"669 view_30" [id=669, type=view];
-"670 unsqueeze_16" [id=670, type=unsqueeze];
-"671 unsqueeze_17" [id=671, type=unsqueeze];
-"672 add_18" [id=672, type=add];
-"673 view_31" [id=673, type=view];
-"674 softmax_5" [id=674, type=softmax];
-"675 dropout_20" [id=675, type=dropout];
-"676 matmul_11" [id=676, type=matmul];
-"677 transpose_11" [id=677, type=transpose];
-"678 reshape_25" [id=678, type=reshape];
-"679 _param_constant96" [id=679, type=get_attr];
-"680 _param_constant97" [id=680, type=get_attr];
-"681 linear_35" [id=681, type=linear];
-"682 dropout_21" [id=682, type=dropout];
-"683 view_32" [id=683, type=view];
-"684 permute_27" [id=684, type=permute];
-"685 reshape_26" [id=685, type=reshape];
-"686 roll_5" [id=686, type=roll];
-"687 slice_101" [id=687, type=slice];
-"688 slice_102" [id=688, type=slice];
-"689 slice_103" [id=689, type=slice];
-"690 slice_104" [id=690, type=slice];
-"691 contiguous_9" [id=691, type=contiguous];
-"692 _param_constant98" [id=692, type=get_attr];
-"693 _param_constant99" [id=693, type=get_attr];
-"694 layer_norm_13" [id=694, type=layer_norm];
-"695 add_19" [id=695, type=add];
-"696 _param_constant100" [id=696, type=get_attr];
-"697 _param_constant101" [id=697, type=get_attr];
-"698 linear_36" [id=698, type=linear];
-"699 gelu_5" [id=699, type=gelu];
-"700 dropout_22" [id=700, type=dropout];
-"701 _param_constant102" [id=701, type=get_attr];
-"702 _param_constant103" [id=702, type=get_attr];
-"703 linear_37" [id=703, type=linear];
-"704 dropout_23" [id=704, type=dropout];
-"705 _param_constant104" [id=705, type=get_attr];
-"706 _param_constant105" [id=706, type=get_attr];
-"707 layer_norm_14" [id=707, type=layer_norm];
-"708 add_20" [id=708, type=add];
-"709 _tensor_constant39" [id=709, type=get_attr];
-"710 _param_constant106" [id=710, type=get_attr];
-"711 _param_constant107" [id=711, type=get_attr];
-"712 linear_38" [id=712, type=linear];
-"713 relu__6" [id=713, type=relu_];
-"714 _param_constant108" [id=714, type=get_attr];
-"715 linear_39" [id=715, type=linear];
-"716 view_33" [id=716, type=view];
-"717 _tensor_constant40" [id=717, type=get_attr];
-"718 index_6" [id=718, type=index];
-"719 view_34" [id=719, type=view];
-"720 permute_28" [id=720, type=permute];
-"721 contiguous_10" [id=721, type=contiguous];
-"722 unsqueeze_18" [id=722, type=unsqueeze];
-"723 sigmoid_6" [id=723, type=sigmoid];
-"724 mul_12" [id=724, type=mul];
-"725 pad_8" [id=725, type=pad];
-"726 view_35" [id=726, type=view];
-"727 permute_29" [id=727, type=permute];
-"728 reshape_27" [id=728, type=reshape];
-"729 _param_constant109" [id=729, type=get_attr];
-"730 clone_6" [id=730, type=clone];
-"731 slice_105" [id=731, type=slice];
-"732 zero__6" [id=732, type=zero_];
-"733 _param_constant110" [id=733, type=get_attr];
-"734 linear_40" [id=734, type=linear];
-"735 reshape_28" [id=735, type=reshape];
-"736 permute_30" [id=736, type=permute];
-"737 select_18" [id=737, type=select];
-"738 select_19" [id=738, type=select];
-"739 select_20" [id=739, type=select];
-"740 linalg_vector_norm_12" [id=740, type=linalg_vector_norm];
-"741 clamp_min_12" [id=741, type=clamp_min];
-"742 expand_as_12" [id=742, type=expand_as];
-"743 div_12" [id=743, type=div];
-"744 linalg_vector_norm_13" [id=744, type=linalg_vector_norm];
-"745 clamp_min_13" [id=745, type=clamp_min];
-"746 expand_as_13" [id=746, type=expand_as];
-"747 div_13" [id=747, type=div];
-"748 transpose_12" [id=748, type=transpose];
-"749 matmul_12" [id=749, type=matmul];
-"750 _param_constant111" [id=750, type=get_attr];
-"751 clamp_6" [id=751, type=clamp];
-"752 exp_6" [id=752, type=exp];
-"753 mul_13" [id=753, type=mul];
-"754 add_21" [id=754, type=add];
-"755 softmax_6" [id=755, type=softmax];
-"756 dropout_24" [id=756, type=dropout];
-"757 matmul_13" [id=757, type=matmul];
-"758 transpose_13" [id=758, type=transpose];
-"759 reshape_29" [id=759, type=reshape];
-"760 _param_constant112" [id=760, type=get_attr];
-"761 _param_constant113" [id=761, type=get_attr];
-"762 linear_41" [id=762, type=linear];
-"763 dropout_25" [id=763, type=dropout];
-"764 view_36" [id=764, type=view];
-"765 permute_31" [id=765, type=permute];
-"766 reshape_30" [id=766, type=reshape];
-"767 slice_106" [id=767, type=slice];
-"768 slice_107" [id=768, type=slice];
-"769 slice_108" [id=769, type=slice];
-"770 slice_109" [id=770, type=slice];
-"771 contiguous_11" [id=771, type=contiguous];
-"772 _param_constant114" [id=772, type=get_attr];
-"773 _param_constant115" [id=773, type=get_attr];
-"774 layer_norm_15" [id=774, type=layer_norm];
-"775 add_22" [id=775, type=add];
-"776 _param_constant116" [id=776, type=get_attr];
-"777 _param_constant117" [id=777, type=get_attr];
-"778 linear_42" [id=778, type=linear];
-"779 gelu_6" [id=779, type=gelu];
-"780 dropout_26" [id=780, type=dropout];
-"781 _param_constant118" [id=781, type=get_attr];
-"782 _param_constant119" [id=782, type=get_attr];
-"783 linear_43" [id=783, type=linear];
-"784 dropout_27" [id=784, type=dropout];
-"785 _param_constant120" [id=785, type=get_attr];
-"786 _param_constant121" [id=786, type=get_attr];
-"787 layer_norm_16" [id=787, type=layer_norm];
-"788 add_23" [id=788, type=add];
-"789 _tensor_constant41" [id=789, type=get_attr];
-"790 _param_constant122" [id=790, type=get_attr];
-"791 _param_constant123" [id=791, type=get_attr];
-"792 linear_44" [id=792, type=linear];
-"793 relu__7" [id=793, type=relu_];
-"794 _param_constant124" [id=794, type=get_attr];
-"795 linear_45" [id=795, type=linear];
-"796 view_37" [id=796, type=view];
-"797 _tensor_constant42" [id=797, type=get_attr];
-"798 index_7" [id=798, type=index];
-"799 view_38" [id=799, type=view];
-"800 permute_32" [id=800, type=permute];
-"801 contiguous_12" [id=801, type=contiguous];
-"802 unsqueeze_19" [id=802, type=unsqueeze];
-"803 sigmoid_7" [id=803, type=sigmoid];
-"804 mul_14" [id=804, type=mul];
-"805 pad_9" [id=805, type=pad];
-"806 roll_6" [id=806, type=roll];
-"807 view_39" [id=807, type=view];
-"808 permute_33" [id=808, type=permute];
-"809 reshape_31" [id=809, type=reshape];
-"810 _param_constant125" [id=810, type=get_attr];
-"811 clone_7" [id=811, type=clone];
-"812 slice_110" [id=812, type=slice];
-"813 zero__7" [id=813, type=zero_];
-"814 _param_constant126" [id=814, type=get_attr];
-"815 linear_46" [id=815, type=linear];
-"816 reshape_32" [id=816, type=reshape];
-"817 permute_34" [id=817, type=permute];
-"818 select_21" [id=818, type=select];
-"819 select_22" [id=819, type=select];
-"820 select_23" [id=820, type=select];
-"821 linalg_vector_norm_14" [id=821, type=linalg_vector_norm];
-"822 clamp_min_14" [id=822, type=clamp_min];
-"823 expand_as_14" [id=823, type=expand_as];
-"824 div_14" [id=824, type=div];
-"825 linalg_vector_norm_15" [id=825, type=linalg_vector_norm];
-"826 clamp_min_15" [id=826, type=clamp_min];
-"827 expand_as_15" [id=827, type=expand_as];
-"828 div_15" [id=828, type=div];
-"829 transpose_14" [id=829, type=transpose];
-"830 matmul_14" [id=830, type=matmul];
-"831 _param_constant127" [id=831, type=get_attr];
-"832 clamp_7" [id=832, type=clamp];
-"833 exp_7" [id=833, type=exp];
-"834 mul_15" [id=834, type=mul];
-"835 add_24" [id=835, type=add];
-"836 new_zeros_3" [id=836, type=new_zeros];
-"837 _tensor_constant43" [id=837, type=get_attr];
-"838 lift_fresh_copy_27" [id=838, type=lift_fresh_copy];
-"839 slice_111" [id=839, type=slice];
-"840 slice_112" [id=840, type=slice];
-"841 fill__27" [id=841, type=fill_];
-"842 _tensor_constant44" [id=842, type=get_attr];
-"843 lift_fresh_copy_28" [id=843, type=lift_fresh_copy];
-"844 slice_113" [id=844, type=slice];
-"845 slice_114" [id=845, type=slice];
-"846 fill__28" [id=846, type=fill_];
-"847 _tensor_constant45" [id=847, type=get_attr];
-"848 lift_fresh_copy_29" [id=848, type=lift_fresh_copy];
-"849 slice_115" [id=849, type=slice];
-"850 slice_116" [id=850, type=slice];
-"851 fill__29" [id=851, type=fill_];
-"852 _tensor_constant46" [id=852, type=get_attr];
-"853 lift_fresh_copy_30" [id=853, type=lift_fresh_copy];
-"854 slice_117" [id=854, type=slice];
-"855 slice_118" [id=855, type=slice];
-"856 fill__30" [id=856, type=fill_];
-"857 _tensor_constant47" [id=857, type=get_attr];
-"858 lift_fresh_copy_31" [id=858, type=lift_fresh_copy];
-"859 slice_119" [id=859, type=slice];
-"860 slice_120" [id=860, type=slice];
-"861 fill__31" [id=861, type=fill_];
-"862 _tensor_constant48" [id=862, type=get_attr];
-"863 lift_fresh_copy_32" [id=863, type=lift_fresh_copy];
-"864 slice_121" [id=864, type=slice];
-"865 slice_122" [id=865, type=slice];
-"866 fill__32" [id=866, type=fill_];
-"867 _tensor_constant49" [id=867, type=get_attr];
-"868 lift_fresh_copy_33" [id=868, type=lift_fresh_copy];
-"869 slice_123" [id=869, type=slice];
-"870 slice_124" [id=870, type=slice];
-"871 fill__33" [id=871, type=fill_];
-"872 _tensor_constant50" [id=872, type=get_attr];
-"873 lift_fresh_copy_34" [id=873, type=lift_fresh_copy];
-"874 slice_125" [id=874, type=slice];
-"875 slice_126" [id=875, type=slice];
-"876 fill__34" [id=876, type=fill_];
-"877 _tensor_constant51" [id=877, type=get_attr];
-"878 lift_fresh_copy_35" [id=878, type=lift_fresh_copy];
-"879 slice_127" [id=879, type=slice];
-"880 slice_128" [id=880, type=slice];
-"881 fill__35" [id=881, type=fill_];
-"882 view_40" [id=882, type=view];
-"883 permute_35" [id=883, type=permute];
-"884 reshape_33" [id=884, type=reshape];
-"885 unsqueeze_20" [id=885, type=unsqueeze];
-"886 unsqueeze_21" [id=886, type=unsqueeze];
-"887 sub_3" [id=887, type=sub];
-"888 ne_3" [id=888, type=ne];
-"889 masked_fill_6" [id=889, type=masked_fill];
-"890 eq_3" [id=890, type=eq];
-"891 masked_fill_7" [id=891, type=masked_fill];
-"892 view_41" [id=892, type=view];
-"893 unsqueeze_22" [id=893, type=unsqueeze];
-"894 unsqueeze_23" [id=894, type=unsqueeze];
-"895 add_25" [id=895, type=add];
-"896 view_42" [id=896, type=view];
-"897 softmax_7" [id=897, type=softmax];
-"898 dropout_28" [id=898, type=dropout];
-"899 matmul_15" [id=899, type=matmul];
-"900 transpose_15" [id=900, type=transpose];
-"901 reshape_34" [id=901, type=reshape];
-"902 _param_constant128" [id=902, type=get_attr];
-"903 _param_constant129" [id=903, type=get_attr];
-"904 linear_47" [id=904, type=linear];
-"905 dropout_29" [id=905, type=dropout];
-"906 view_43" [id=906, type=view];
-"907 permute_36" [id=907, type=permute];
-"908 reshape_35" [id=908, type=reshape];
-"909 roll_7" [id=909, type=roll];
-"910 slice_129" [id=910, type=slice];
-"911 slice_130" [id=911, type=slice];
-"912 slice_131" [id=912, type=slice];
-"913 slice_132" [id=913, type=slice];
-"914 contiguous_13" [id=914, type=contiguous];
-"915 _param_constant130" [id=915, type=get_attr];
-"916 _param_constant131" [id=916, type=get_attr];
-"917 layer_norm_17" [id=917, type=layer_norm];
-"918 add_26" [id=918, type=add];
-"919 _param_constant132" [id=919, type=get_attr];
-"920 _param_constant133" [id=920, type=get_attr];
-"921 linear_48" [id=921, type=linear];
-"922 gelu_7" [id=922, type=gelu];
-"923 dropout_30" [id=923, type=dropout];
-"924 _param_constant134" [id=924, type=get_attr];
-"925 _param_constant135" [id=925, type=get_attr];
-"926 linear_49" [id=926, type=linear];
-"927 dropout_31" [id=927, type=dropout];
-"928 _param_constant136" [id=928, type=get_attr];
-"929 _param_constant137" [id=929, type=get_attr];
-"930 layer_norm_18" [id=930, type=layer_norm];
-"931 add_27" [id=931, type=add];
-"932 _tensor_constant52" [id=932, type=get_attr];
-"933 _param_constant138" [id=933, type=get_attr];
-"934 _param_constant139" [id=934, type=get_attr];
-"935 linear_50" [id=935, type=linear];
-"936 relu__8" [id=936, type=relu_];
-"937 _param_constant140" [id=937, type=get_attr];
-"938 linear_51" [id=938, type=linear];
-"939 view_44" [id=939, type=view];
-"940 _tensor_constant53" [id=940, type=get_attr];
-"941 index_8" [id=941, type=index];
-"942 view_45" [id=942, type=view];
-"943 permute_37" [id=943, type=permute];
-"944 contiguous_14" [id=944, type=contiguous];
-"945 unsqueeze_24" [id=945, type=unsqueeze];
-"946 sigmoid_8" [id=946, type=sigmoid];
-"947 mul_16" [id=947, type=mul];
-"948 pad_10" [id=948, type=pad];
-"949 view_46" [id=949, type=view];
-"950 permute_38" [id=950, type=permute];
-"951 reshape_36" [id=951, type=reshape];
-"952 _param_constant141" [id=952, type=get_attr];
-"953 clone_8" [id=953, type=clone];
-"954 slice_133" [id=954, type=slice];
-"955 zero__8" [id=955, type=zero_];
-"956 _param_constant142" [id=956, type=get_attr];
-"957 linear_52" [id=957, type=linear];
-"958 reshape_37" [id=958, type=reshape];
-"959 permute_39" [id=959, type=permute];
-"960 select_24" [id=960, type=select];
-"961 select_25" [id=961, type=select];
-"962 select_26" [id=962, type=select];
-"963 linalg_vector_norm_16" [id=963, type=linalg_vector_norm];
-"964 clamp_min_16" [id=964, type=clamp_min];
-"965 expand_as_16" [id=965, type=expand_as];
-"966 div_16" [id=966, type=div];
-"967 linalg_vector_norm_17" [id=967, type=linalg_vector_norm];
-"968 clamp_min_17" [id=968, type=clamp_min];
-"969 expand_as_17" [id=969, type=expand_as];
-"970 div_17" [id=970, type=div];
-"971 transpose_16" [id=971, type=transpose];
-"972 matmul_16" [id=972, type=matmul];
-"973 _param_constant143" [id=973, type=get_attr];
-"974 clamp_8" [id=974, type=clamp];
-"975 exp_8" [id=975, type=exp];
-"976 mul_17" [id=976, type=mul];
-"977 add_28" [id=977, type=add];
-"978 softmax_8" [id=978, type=softmax];
-"979 dropout_32" [id=979, type=dropout];
-"980 matmul_17" [id=980, type=matmul];
-"981 transpose_17" [id=981, type=transpose];
-"982 reshape_38" [id=982, type=reshape];
-"983 _param_constant144" [id=983, type=get_attr];
-"984 _param_constant145" [id=984, type=get_attr];
-"985 linear_53" [id=985, type=linear];
-"986 dropout_33" [id=986, type=dropout];
-"987 view_47" [id=987, type=view];
-"988 permute_40" [id=988, type=permute];
-"989 reshape_39" [id=989, type=reshape];
-"990 slice_134" [id=990, type=slice];
-"991 slice_135" [id=991, type=slice];
-"992 slice_136" [id=992, type=slice];
-"993 slice_137" [id=993, type=slice];
-"994 contiguous_15" [id=994, type=contiguous];
-"995 _param_constant146" [id=995, type=get_attr];
-"996 _param_constant147" [id=996, type=get_attr];
-"997 layer_norm_19" [id=997, type=layer_norm];
-"998 add_29" [id=998, type=add];
-"999 _param_constant148" [id=999, type=get_attr];
-"1000 _param_constant149" [id=1000, type=get_attr];
-"1001 linear_54" [id=1001, type=linear];
-"1002 gelu_8" [id=1002, type=gelu];
-"1003 dropout_34" [id=1003, type=dropout];
-"1004 _param_constant150" [id=1004, type=get_attr];
-"1005 _param_constant151" [id=1005, type=get_attr];
-"1006 linear_55" [id=1006, type=linear];
-"1007 dropout_35" [id=1007, type=dropout];
-"1008 _param_constant152" [id=1008, type=get_attr];
-"1009 _param_constant153" [id=1009, type=get_attr];
-"1010 layer_norm_20" [id=1010, type=layer_norm];
-"1011 add_30" [id=1011, type=add];
-"1012 _tensor_constant54" [id=1012, type=get_attr];
-"1013 _param_constant154" [id=1013, type=get_attr];
-"1014 _param_constant155" [id=1014, type=get_attr];
-"1015 linear_56" [id=1015, type=linear];
-"1016 relu__9" [id=1016, type=relu_];
-"1017 _param_constant156" [id=1017, type=get_attr];
-"1018 linear_57" [id=1018, type=linear];
-"1019 view_48" [id=1019, type=view];
-"1020 _tensor_constant55" [id=1020, type=get_attr];
-"1021 index_9" [id=1021, type=index];
-"1022 view_49" [id=1022, type=view];
-"1023 permute_41" [id=1023, type=permute];
-"1024 contiguous_16" [id=1024, type=contiguous];
-"1025 unsqueeze_25" [id=1025, type=unsqueeze];
-"1026 sigmoid_9" [id=1026, type=sigmoid];
-"1027 mul_18" [id=1027, type=mul];
-"1028 pad_11" [id=1028, type=pad];
-"1029 roll_8" [id=1029, type=roll];
-"1030 view_50" [id=1030, type=view];
-"1031 permute_42" [id=1031, type=permute];
-"1032 reshape_40" [id=1032, type=reshape];
-"1033 _param_constant157" [id=1033, type=get_attr];
-"1034 clone_9" [id=1034, type=clone];
-"1035 slice_138" [id=1035, type=slice];
-"1036 zero__9" [id=1036, type=zero_];
-"1037 _param_constant158" [id=1037, type=get_attr];
-"1038 linear_58" [id=1038, type=linear];
-"1039 reshape_41" [id=1039, type=reshape];
-"1040 permute_43" [id=1040, type=permute];
-"1041 select_27" [id=1041, type=select];
-"1042 select_28" [id=1042, type=select];
-"1043 select_29" [id=1043, type=select];
-"1044 linalg_vector_norm_18" [id=1044, type=linalg_vector_norm];
-"1045 clamp_min_18" [id=1045, type=clamp_min];
-"1046 expand_as_18" [id=1046, type=expand_as];
-"1047 div_18" [id=1047, type=div];
-"1048 linalg_vector_norm_19" [id=1048, type=linalg_vector_norm];
-"1049 clamp_min_19" [id=1049, type=clamp_min];
-"1050 expand_as_19" [id=1050, type=expand_as];
-"1051 div_19" [id=1051, type=div];
-"1052 transpose_18" [id=1052, type=transpose];
-"1053 matmul_18" [id=1053, type=matmul];
-"1054 _param_constant159" [id=1054, type=get_attr];
-"1055 clamp_9" [id=1055, type=clamp];
-"1056 exp_9" [id=1056, type=exp];
-"1057 mul_19" [id=1057, type=mul];
-"1058 add_31" [id=1058, type=add];
-"1059 new_zeros_4" [id=1059, type=new_zeros];
-"1060 _tensor_constant56" [id=1060, type=get_attr];
-"1061 lift_fresh_copy_36" [id=1061, type=lift_fresh_copy];
-"1062 slice_139" [id=1062, type=slice];
-"1063 slice_140" [id=1063, type=slice];
-"1064 fill__36" [id=1064, type=fill_];
-"1065 _tensor_constant57" [id=1065, type=get_attr];
-"1066 lift_fresh_copy_37" [id=1066, type=lift_fresh_copy];
-"1067 slice_141" [id=1067, type=slice];
-"1068 slice_142" [id=1068, type=slice];
-"1069 fill__37" [id=1069, type=fill_];
-"1070 _tensor_constant58" [id=1070, type=get_attr];
-"1071 lift_fresh_copy_38" [id=1071, type=lift_fresh_copy];
-"1072 slice_143" [id=1072, type=slice];
-"1073 slice_144" [id=1073, type=slice];
-"1074 fill__38" [id=1074, type=fill_];
-"1075 _tensor_constant59" [id=1075, type=get_attr];
-"1076 lift_fresh_copy_39" [id=1076, type=lift_fresh_copy];
-"1077 slice_145" [id=1077, type=slice];
-"1078 slice_146" [id=1078, type=slice];
-"1079 fill__39" [id=1079, type=fill_];
-"1080 _tensor_constant60" [id=1080, type=get_attr];
-"1081 lift_fresh_copy_40" [id=1081, type=lift_fresh_copy];
-"1082 slice_147" [id=1082, type=slice];
-"1083 slice_148" [id=1083, type=slice];
-"1084 fill__40" [id=1084, type=fill_];
-"1085 _tensor_constant61" [id=1085, type=get_attr];
-"1086 lift_fresh_copy_41" [id=1086, type=lift_fresh_copy];
-"1087 slice_149" [id=1087, type=slice];
-"1088 slice_150" [id=1088, type=slice];
-"1089 fill__41" [id=1089, type=fill_];
-"1090 _tensor_constant62" [id=1090, type=get_attr];
-"1091 lift_fresh_copy_42" [id=1091, type=lift_fresh_copy];
-"1092 slice_151" [id=1092, type=slice];
-"1093 slice_152" [id=1093, type=slice];
-"1094 fill__42" [id=1094, type=fill_];
-"1095 _tensor_constant63" [id=1095, type=get_attr];
-"1096 lift_fresh_copy_43" [id=1096, type=lift_fresh_copy];
-"1097 slice_153" [id=1097, type=slice];
-"1098 slice_154" [id=1098, type=slice];
-"1099 fill__43" [id=1099, type=fill_];
-"1100 _tensor_constant64" [id=1100, type=get_attr];
-"1101 lift_fresh_copy_44" [id=1101, type=lift_fresh_copy];
-"1102 slice_155" [id=1102, type=slice];
-"1103 slice_156" [id=1103, type=slice];
-"1104 fill__44" [id=1104, type=fill_];
-"1105 view_51" [id=1105, type=view];
-"1106 permute_44" [id=1106, type=permute];
-"1107 reshape_42" [id=1107, type=reshape];
-"1108 unsqueeze_26" [id=1108, type=unsqueeze];
-"1109 unsqueeze_27" [id=1109, type=unsqueeze];
-"1110 sub_4" [id=1110, type=sub];
-"1111 ne_4" [id=1111, type=ne];
-"1112 masked_fill_8" [id=1112, type=masked_fill];
-"1113 eq_4" [id=1113, type=eq];
-"1114 masked_fill_9" [id=1114, type=masked_fill];
-"1115 view_52" [id=1115, type=view];
-"1116 unsqueeze_28" [id=1116, type=unsqueeze];
-"1117 unsqueeze_29" [id=1117, type=unsqueeze];
-"1118 add_32" [id=1118, type=add];
-"1119 view_53" [id=1119, type=view];
-"1120 softmax_9" [id=1120, type=softmax];
-"1121 dropout_36" [id=1121, type=dropout];
-"1122 matmul_19" [id=1122, type=matmul];
-"1123 transpose_19" [id=1123, type=transpose];
-"1124 reshape_43" [id=1124, type=reshape];
-"1125 _param_constant160" [id=1125, type=get_attr];
-"1126 _param_constant161" [id=1126, type=get_attr];
-"1127 linear_59" [id=1127, type=linear];
-"1128 dropout_37" [id=1128, type=dropout];
-"1129 view_54" [id=1129, type=view];
-"1130 permute_45" [id=1130, type=permute];
-"1131 reshape_44" [id=1131, type=reshape];
-"1132 roll_9" [id=1132, type=roll];
-"1133 slice_157" [id=1133, type=slice];
-"1134 slice_158" [id=1134, type=slice];
-"1135 slice_159" [id=1135, type=slice];
-"1136 slice_160" [id=1136, type=slice];
-"1137 contiguous_17" [id=1137, type=contiguous];
-"1138 _param_constant162" [id=1138, type=get_attr];
-"1139 _param_constant163" [id=1139, type=get_attr];
-"1140 layer_norm_21" [id=1140, type=layer_norm];
-"1141 add_33" [id=1141, type=add];
-"1142 _param_constant164" [id=1142, type=get_attr];
-"1143 _param_constant165" [id=1143, type=get_attr];
-"1144 linear_60" [id=1144, type=linear];
-"1145 gelu_9" [id=1145, type=gelu];
-"1146 dropout_38" [id=1146, type=dropout];
-"1147 _param_constant166" [id=1147, type=get_attr];
-"1148 _param_constant167" [id=1148, type=get_attr];
-"1149 linear_61" [id=1149, type=linear];
-"1150 dropout_39" [id=1150, type=dropout];
-"1151 _param_constant168" [id=1151, type=get_attr];
-"1152 _param_constant169" [id=1152, type=get_attr];
-"1153 layer_norm_22" [id=1153, type=layer_norm];
-"1154 add_34" [id=1154, type=add];
-"1155 _tensor_constant65" [id=1155, type=get_attr];
-"1156 _param_constant170" [id=1156, type=get_attr];
-"1157 _param_constant171" [id=1157, type=get_attr];
-"1158 linear_62" [id=1158, type=linear];
-"1159 relu__10" [id=1159, type=relu_];
-"1160 _param_constant172" [id=1160, type=get_attr];
-"1161 linear_63" [id=1161, type=linear];
-"1162 view_55" [id=1162, type=view];
-"1163 _tensor_constant66" [id=1163, type=get_attr];
-"1164 index_10" [id=1164, type=index];
-"1165 view_56" [id=1165, type=view];
-"1166 permute_46" [id=1166, type=permute];
-"1167 contiguous_18" [id=1167, type=contiguous];
-"1168 unsqueeze_30" [id=1168, type=unsqueeze];
-"1169 sigmoid_10" [id=1169, type=sigmoid];
-"1170 mul_20" [id=1170, type=mul];
-"1171 pad_12" [id=1171, type=pad];
-"1172 view_57" [id=1172, type=view];
-"1173 permute_47" [id=1173, type=permute];
-"1174 reshape_45" [id=1174, type=reshape];
-"1175 _param_constant173" [id=1175, type=get_attr];
-"1176 clone_10" [id=1176, type=clone];
-"1177 slice_161" [id=1177, type=slice];
-"1178 zero__10" [id=1178, type=zero_];
-"1179 _param_constant174" [id=1179, type=get_attr];
-"1180 linear_64" [id=1180, type=linear];
-"1181 reshape_46" [id=1181, type=reshape];
-"1182 permute_48" [id=1182, type=permute];
-"1183 select_30" [id=1183, type=select];
-"1184 select_31" [id=1184, type=select];
-"1185 select_32" [id=1185, type=select];
-"1186 linalg_vector_norm_20" [id=1186, type=linalg_vector_norm];
-"1187 clamp_min_20" [id=1187, type=clamp_min];
-"1188 expand_as_20" [id=1188, type=expand_as];
-"1189 div_20" [id=1189, type=div];
-"1190 linalg_vector_norm_21" [id=1190, type=linalg_vector_norm];
-"1191 clamp_min_21" [id=1191, type=clamp_min];
-"1192 expand_as_21" [id=1192, type=expand_as];
-"1193 div_21" [id=1193, type=div];
-"1194 transpose_20" [id=1194, type=transpose];
-"1195 matmul_20" [id=1195, type=matmul];
-"1196 _param_constant175" [id=1196, type=get_attr];
-"1197 clamp_10" [id=1197, type=clamp];
-"1198 exp_10" [id=1198, type=exp];
-"1199 mul_21" [id=1199, type=mul];
-"1200 add_35" [id=1200, type=add];
-"1201 softmax_10" [id=1201, type=softmax];
-"1202 dropout_40" [id=1202, type=dropout];
-"1203 matmul_21" [id=1203, type=matmul];
-"1204 transpose_21" [id=1204, type=transpose];
-"1205 reshape_47" [id=1205, type=reshape];
-"1206 _param_constant176" [id=1206, type=get_attr];
-"1207 _param_constant177" [id=1207, type=get_attr];
-"1208 linear_65" [id=1208, type=linear];
-"1209 dropout_41" [id=1209, type=dropout];
-"1210 view_58" [id=1210, type=view];
-"1211 permute_49" [id=1211, type=permute];
-"1212 reshape_48" [id=1212, type=reshape];
-"1213 slice_162" [id=1213, type=slice];
-"1214 slice_163" [id=1214, type=slice];
-"1215 slice_164" [id=1215, type=slice];
-"1216 slice_165" [id=1216, type=slice];
-"1217 contiguous_19" [id=1217, type=contiguous];
-"1218 _param_constant178" [id=1218, type=get_attr];
-"1219 _param_constant179" [id=1219, type=get_attr];
-"1220 layer_norm_23" [id=1220, type=layer_norm];
-"1221 add_36" [id=1221, type=add];
-"1222 _param_constant180" [id=1222, type=get_attr];
-"1223 _param_constant181" [id=1223, type=get_attr];
-"1224 linear_66" [id=1224, type=linear];
-"1225 gelu_10" [id=1225, type=gelu];
-"1226 dropout_42" [id=1226, type=dropout];
-"1227 _param_constant182" [id=1227, type=get_attr];
-"1228 _param_constant183" [id=1228, type=get_attr];
-"1229 linear_67" [id=1229, type=linear];
-"1230 dropout_43" [id=1230, type=dropout];
-"1231 _param_constant184" [id=1231, type=get_attr];
-"1232 _param_constant185" [id=1232, type=get_attr];
-"1233 layer_norm_24" [id=1233, type=layer_norm];
-"1234 add_37" [id=1234, type=add];
-"1235 _tensor_constant67" [id=1235, type=get_attr];
-"1236 _param_constant186" [id=1236, type=get_attr];
-"1237 _param_constant187" [id=1237, type=get_attr];
-"1238 linear_68" [id=1238, type=linear];
-"1239 relu__11" [id=1239, type=relu_];
-"1240 _param_constant188" [id=1240, type=get_attr];
-"1241 linear_69" [id=1241, type=linear];
-"1242 view_59" [id=1242, type=view];
-"1243 _tensor_constant68" [id=1243, type=get_attr];
-"1244 index_11" [id=1244, type=index];
-"1245 view_60" [id=1245, type=view];
-"1246 permute_50" [id=1246, type=permute];
-"1247 contiguous_20" [id=1247, type=contiguous];
-"1248 unsqueeze_31" [id=1248, type=unsqueeze];
-"1249 sigmoid_11" [id=1249, type=sigmoid];
-"1250 mul_22" [id=1250, type=mul];
-"1251 pad_13" [id=1251, type=pad];
-"1252 roll_10" [id=1252, type=roll];
-"1253 view_61" [id=1253, type=view];
-"1254 permute_51" [id=1254, type=permute];
-"1255 reshape_49" [id=1255, type=reshape];
-"1256 _param_constant189" [id=1256, type=get_attr];
-"1257 clone_11" [id=1257, type=clone];
-"1258 slice_166" [id=1258, type=slice];
-"1259 zero__11" [id=1259, type=zero_];
-"1260 _param_constant190" [id=1260, type=get_attr];
-"1261 linear_70" [id=1261, type=linear];
-"1262 reshape_50" [id=1262, type=reshape];
-"1263 permute_52" [id=1263, type=permute];
-"1264 select_33" [id=1264, type=select];
-"1265 select_34" [id=1265, type=select];
-"1266 select_35" [id=1266, type=select];
-"1267 linalg_vector_norm_22" [id=1267, type=linalg_vector_norm];
-"1268 clamp_min_22" [id=1268, type=clamp_min];
-"1269 expand_as_22" [id=1269, type=expand_as];
-"1270 div_22" [id=1270, type=div];
-"1271 linalg_vector_norm_23" [id=1271, type=linalg_vector_norm];
-"1272 clamp_min_23" [id=1272, type=clamp_min];
-"1273 expand_as_23" [id=1273, type=expand_as];
-"1274 div_23" [id=1274, type=div];
-"1275 transpose_22" [id=1275, type=transpose];
-"1276 matmul_22" [id=1276, type=matmul];
-"1277 _param_constant191" [id=1277, type=get_attr];
-"1278 clamp_11" [id=1278, type=clamp];
-"1279 exp_11" [id=1279, type=exp];
-"1280 mul_23" [id=1280, type=mul];
-"1281 add_38" [id=1281, type=add];
-"1282 new_zeros_5" [id=1282, type=new_zeros];
-"1283 _tensor_constant69" [id=1283, type=get_attr];
-"1284 lift_fresh_copy_45" [id=1284, type=lift_fresh_copy];
-"1285 slice_167" [id=1285, type=slice];
-"1286 slice_168" [id=1286, type=slice];
-"1287 fill__45" [id=1287, type=fill_];
-"1288 _tensor_constant70" [id=1288, type=get_attr];
-"1289 lift_fresh_copy_46" [id=1289, type=lift_fresh_copy];
-"1290 slice_169" [id=1290, type=slice];
-"1291 slice_170" [id=1291, type=slice];
-"1292 fill__46" [id=1292, type=fill_];
-"1293 _tensor_constant71" [id=1293, type=get_attr];
-"1294 lift_fresh_copy_47" [id=1294, type=lift_fresh_copy];
-"1295 slice_171" [id=1295, type=slice];
-"1296 slice_172" [id=1296, type=slice];
-"1297 fill__47" [id=1297, type=fill_];
-"1298 _tensor_constant72" [id=1298, type=get_attr];
-"1299 lift_fresh_copy_48" [id=1299, type=lift_fresh_copy];
-"1300 slice_173" [id=1300, type=slice];
-"1301 slice_174" [id=1301, type=slice];
-"1302 fill__48" [id=1302, type=fill_];
-"1303 _tensor_constant73" [id=1303, type=get_attr];
-"1304 lift_fresh_copy_49" [id=1304, type=lift_fresh_copy];
-"1305 slice_175" [id=1305, type=slice];
-"1306 slice_176" [id=1306, type=slice];
-"1307 fill__49" [id=1307, type=fill_];
-"1308 _tensor_constant74" [id=1308, type=get_attr];
-"1309 lift_fresh_copy_50" [id=1309, type=lift_fresh_copy];
-"1310 slice_177" [id=1310, type=slice];
-"1311 slice_178" [id=1311, type=slice];
-"1312 fill__50" [id=1312, type=fill_];
-"1313 _tensor_constant75" [id=1313, type=get_attr];
-"1314 lift_fresh_copy_51" [id=1314, type=lift_fresh_copy];
-"1315 slice_179" [id=1315, type=slice];
-"1316 slice_180" [id=1316, type=slice];
-"1317 fill__51" [id=1317, type=fill_];
-"1318 _tensor_constant76" [id=1318, type=get_attr];
-"1319 lift_fresh_copy_52" [id=1319, type=lift_fresh_copy];
-"1320 slice_181" [id=1320, type=slice];
-"1321 slice_182" [id=1321, type=slice];
-"1322 fill__52" [id=1322, type=fill_];
-"1323 _tensor_constant77" [id=1323, type=get_attr];
-"1324 lift_fresh_copy_53" [id=1324, type=lift_fresh_copy];
-"1325 slice_183" [id=1325, type=slice];
-"1326 slice_184" [id=1326, type=slice];
-"1327 fill__53" [id=1327, type=fill_];
-"1328 view_62" [id=1328, type=view];
-"1329 permute_53" [id=1329, type=permute];
-"1330 reshape_51" [id=1330, type=reshape];
-"1331 unsqueeze_32" [id=1331, type=unsqueeze];
-"1332 unsqueeze_33" [id=1332, type=unsqueeze];
-"1333 sub_5" [id=1333, type=sub];
-"1334 ne_5" [id=1334, type=ne];
-"1335 masked_fill_10" [id=1335, type=masked_fill];
-"1336 eq_5" [id=1336, type=eq];
-"1337 masked_fill_11" [id=1337, type=masked_fill];
-"1338 view_63" [id=1338, type=view];
-"1339 unsqueeze_34" [id=1339, type=unsqueeze];
-"1340 unsqueeze_35" [id=1340, type=unsqueeze];
-"1341 add_39" [id=1341, type=add];
-"1342 view_64" [id=1342, type=view];
-"1343 softmax_11" [id=1343, type=softmax];
-"1344 dropout_44" [id=1344, type=dropout];
-"1345 matmul_23" [id=1345, type=matmul];
-"1346 transpose_23" [id=1346, type=transpose];
-"1347 reshape_52" [id=1347, type=reshape];
-"1348 _param_constant192" [id=1348, type=get_attr];
-"1349 _param_constant193" [id=1349, type=get_attr];
-"1350 linear_71" [id=1350, type=linear];
-"1351 dropout_45" [id=1351, type=dropout];
-"1352 view_65" [id=1352, type=view];
-"1353 permute_54" [id=1353, type=permute];
-"1354 reshape_53" [id=1354, type=reshape];
-"1355 roll_11" [id=1355, type=roll];
-"1356 slice_185" [id=1356, type=slice];
-"1357 slice_186" [id=1357, type=slice];
-"1358 slice_187" [id=1358, type=slice];
-"1359 slice_188" [id=1359, type=slice];
-"1360 contiguous_21" [id=1360, type=contiguous];
-"1361 _param_constant194" [id=1361, type=get_attr];
-"1362 _param_constant195" [id=1362, type=get_attr];
-"1363 layer_norm_25" [id=1363, type=layer_norm];
-"1364 add_40" [id=1364, type=add];
-"1365 _param_constant196" [id=1365, type=get_attr];
-"1366 _param_constant197" [id=1366, type=get_attr];
-"1367 linear_72" [id=1367, type=linear];
-"1368 gelu_11" [id=1368, type=gelu];
-"1369 dropout_46" [id=1369, type=dropout];
-"1370 _param_constant198" [id=1370, type=get_attr];
-"1371 _param_constant199" [id=1371, type=get_attr];
-"1372 linear_73" [id=1372, type=linear];
-"1373 dropout_47" [id=1373, type=dropout];
-"1374 _param_constant200" [id=1374, type=get_attr];
-"1375 _param_constant201" [id=1375, type=get_attr];
-"1376 layer_norm_26" [id=1376, type=layer_norm];
-"1377 add_41" [id=1377, type=add];
-"1378 _tensor_constant78" [id=1378, type=get_attr];
-"1379 _param_constant202" [id=1379, type=get_attr];
-"1380 _param_constant203" [id=1380, type=get_attr];
-"1381 linear_74" [id=1381, type=linear];
-"1382 relu__12" [id=1382, type=relu_];
-"1383 _param_constant204" [id=1383, type=get_attr];
-"1384 linear_75" [id=1384, type=linear];
-"1385 view_66" [id=1385, type=view];
-"1386 _tensor_constant79" [id=1386, type=get_attr];
-"1387 index_12" [id=1387, type=index];
-"1388 view_67" [id=1388, type=view];
-"1389 permute_55" [id=1389, type=permute];
-"1390 contiguous_22" [id=1390, type=contiguous];
-"1391 unsqueeze_36" [id=1391, type=unsqueeze];
-"1392 sigmoid_12" [id=1392, type=sigmoid];
-"1393 mul_24" [id=1393, type=mul];
-"1394 pad_14" [id=1394, type=pad];
-"1395 view_68" [id=1395, type=view];
-"1396 permute_56" [id=1396, type=permute];
-"1397 reshape_54" [id=1397, type=reshape];
-"1398 _param_constant205" [id=1398, type=get_attr];
-"1399 clone_12" [id=1399, type=clone];
-"1400 slice_189" [id=1400, type=slice];
-"1401 zero__12" [id=1401, type=zero_];
-"1402 _param_constant206" [id=1402, type=get_attr];
-"1403 linear_76" [id=1403, type=linear];
-"1404 reshape_55" [id=1404, type=reshape];
-"1405 permute_57" [id=1405, type=permute];
-"1406 select_36" [id=1406, type=select];
-"1407 select_37" [id=1407, type=select];
-"1408 select_38" [id=1408, type=select];
-"1409 linalg_vector_norm_24" [id=1409, type=linalg_vector_norm];
-"1410 clamp_min_24" [id=1410, type=clamp_min];
-"1411 expand_as_24" [id=1411, type=expand_as];
-"1412 div_24" [id=1412, type=div];
-"1413 linalg_vector_norm_25" [id=1413, type=linalg_vector_norm];
-"1414 clamp_min_25" [id=1414, type=clamp_min];
-"1415 expand_as_25" [id=1415, type=expand_as];
-"1416 div_25" [id=1416, type=div];
-"1417 transpose_24" [id=1417, type=transpose];
-"1418 matmul_24" [id=1418, type=matmul];
-"1419 _param_constant207" [id=1419, type=get_attr];
-"1420 clamp_12" [id=1420, type=clamp];
-"1421 exp_12" [id=1421, type=exp];
-"1422 mul_25" [id=1422, type=mul];
-"1423 add_42" [id=1423, type=add];
-"1424 softmax_12" [id=1424, type=softmax];
-"1425 dropout_48" [id=1425, type=dropout];
-"1426 matmul_25" [id=1426, type=matmul];
-"1427 transpose_25" [id=1427, type=transpose];
-"1428 reshape_56" [id=1428, type=reshape];
-"1429 _param_constant208" [id=1429, type=get_attr];
-"1430 _param_constant209" [id=1430, type=get_attr];
-"1431 linear_77" [id=1431, type=linear];
-"1432 dropout_49" [id=1432, type=dropout];
-"1433 view_69" [id=1433, type=view];
-"1434 permute_58" [id=1434, type=permute];
-"1435 reshape_57" [id=1435, type=reshape];
-"1436 slice_190" [id=1436, type=slice];
-"1437 slice_191" [id=1437, type=slice];
-"1438 slice_192" [id=1438, type=slice];
-"1439 slice_193" [id=1439, type=slice];
-"1440 contiguous_23" [id=1440, type=contiguous];
-"1441 _param_constant210" [id=1441, type=get_attr];
-"1442 _param_constant211" [id=1442, type=get_attr];
-"1443 layer_norm_27" [id=1443, type=layer_norm];
-"1444 add_43" [id=1444, type=add];
-"1445 _param_constant212" [id=1445, type=get_attr];
-"1446 _param_constant213" [id=1446, type=get_attr];
-"1447 linear_78" [id=1447, type=linear];
-"1448 gelu_12" [id=1448, type=gelu];
-"1449 dropout_50" [id=1449, type=dropout];
-"1450 _param_constant214" [id=1450, type=get_attr];
-"1451 _param_constant215" [id=1451, type=get_attr];
-"1452 linear_79" [id=1452, type=linear];
-"1453 dropout_51" [id=1453, type=dropout];
-"1454 _param_constant216" [id=1454, type=get_attr];
-"1455 _param_constant217" [id=1455, type=get_attr];
-"1456 layer_norm_28" [id=1456, type=layer_norm];
-"1457 add_44" [id=1457, type=add];
-"1458 _tensor_constant80" [id=1458, type=get_attr];
-"1459 _param_constant218" [id=1459, type=get_attr];
-"1460 _param_constant219" [id=1460, type=get_attr];
-"1461 linear_80" [id=1461, type=linear];
-"1462 relu__13" [id=1462, type=relu_];
-"1463 _param_constant220" [id=1463, type=get_attr];
-"1464 linear_81" [id=1464, type=linear];
-"1465 view_70" [id=1465, type=view];
-"1466 _tensor_constant81" [id=1466, type=get_attr];
-"1467 index_13" [id=1467, type=index];
-"1468 view_71" [id=1468, type=view];
-"1469 permute_59" [id=1469, type=permute];
-"1470 contiguous_24" [id=1470, type=contiguous];
-"1471 unsqueeze_37" [id=1471, type=unsqueeze];
-"1472 sigmoid_13" [id=1472, type=sigmoid];
-"1473 mul_26" [id=1473, type=mul];
-"1474 pad_15" [id=1474, type=pad];
-"1475 roll_12" [id=1475, type=roll];
-"1476 view_72" [id=1476, type=view];
-"1477 permute_60" [id=1477, type=permute];
-"1478 reshape_58" [id=1478, type=reshape];
-"1479 _param_constant221" [id=1479, type=get_attr];
-"1480 clone_13" [id=1480, type=clone];
-"1481 slice_194" [id=1481, type=slice];
-"1482 zero__13" [id=1482, type=zero_];
-"1483 _param_constant222" [id=1483, type=get_attr];
-"1484 linear_82" [id=1484, type=linear];
-"1485 reshape_59" [id=1485, type=reshape];
-"1486 permute_61" [id=1486, type=permute];
-"1487 select_39" [id=1487, type=select];
-"1488 select_40" [id=1488, type=select];
-"1489 select_41" [id=1489, type=select];
-"1490 linalg_vector_norm_26" [id=1490, type=linalg_vector_norm];
-"1491 clamp_min_26" [id=1491, type=clamp_min];
-"1492 expand_as_26" [id=1492, type=expand_as];
-"1493 div_26" [id=1493, type=div];
-"1494 linalg_vector_norm_27" [id=1494, type=linalg_vector_norm];
-"1495 clamp_min_27" [id=1495, type=clamp_min];
-"1496 expand_as_27" [id=1496, type=expand_as];
-"1497 div_27" [id=1497, type=div];
-"1498 transpose_26" [id=1498, type=transpose];
-"1499 matmul_26" [id=1499, type=matmul];
-"1500 _param_constant223" [id=1500, type=get_attr];
-"1501 clamp_13" [id=1501, type=clamp];
-"1502 exp_13" [id=1502, type=exp];
-"1503 mul_27" [id=1503, type=mul];
-"1504 add_45" [id=1504, type=add];
-"1505 new_zeros_6" [id=1505, type=new_zeros];
-"1506 _tensor_constant82" [id=1506, type=get_attr];
-"1507 lift_fresh_copy_54" [id=1507, type=lift_fresh_copy];
-"1508 slice_195" [id=1508, type=slice];
-"1509 slice_196" [id=1509, type=slice];
-"1510 fill__54" [id=1510, type=fill_];
-"1511 _tensor_constant83" [id=1511, type=get_attr];
-"1512 lift_fresh_copy_55" [id=1512, type=lift_fresh_copy];
-"1513 slice_197" [id=1513, type=slice];
-"1514 slice_198" [id=1514, type=slice];
-"1515 fill__55" [id=1515, type=fill_];
-"1516 _tensor_constant84" [id=1516, type=get_attr];
-"1517 lift_fresh_copy_56" [id=1517, type=lift_fresh_copy];
-"1518 slice_199" [id=1518, type=slice];
-"1519 slice_200" [id=1519, type=slice];
-"1520 fill__56" [id=1520, type=fill_];
-"1521 _tensor_constant85" [id=1521, type=get_attr];
-"1522 lift_fresh_copy_57" [id=1522, type=lift_fresh_copy];
-"1523 slice_201" [id=1523, type=slice];
-"1524 slice_202" [id=1524, type=slice];
-"1525 fill__57" [id=1525, type=fill_];
-"1526 _tensor_constant86" [id=1526, type=get_attr];
-"1527 lift_fresh_copy_58" [id=1527, type=lift_fresh_copy];
-"1528 slice_203" [id=1528, type=slice];
-"1529 slice_204" [id=1529, type=slice];
-"1530 fill__58" [id=1530, type=fill_];
-"1531 _tensor_constant87" [id=1531, type=get_attr];
-"1532 lift_fresh_copy_59" [id=1532, type=lift_fresh_copy];
-"1533 slice_205" [id=1533, type=slice];
-"1534 slice_206" [id=1534, type=slice];
-"1535 fill__59" [id=1535, type=fill_];
-"1536 _tensor_constant88" [id=1536, type=get_attr];
-"1537 lift_fresh_copy_60" [id=1537, type=lift_fresh_copy];
-"1538 slice_207" [id=1538, type=slice];
-"1539 slice_208" [id=1539, type=slice];
-"1540 fill__60" [id=1540, type=fill_];
-"1541 _tensor_constant89" [id=1541, type=get_attr];
-"1542 lift_fresh_copy_61" [id=1542, type=lift_fresh_copy];
-"1543 slice_209" [id=1543, type=slice];
-"1544 slice_210" [id=1544, type=slice];
-"1545 fill__61" [id=1545, type=fill_];
-"1546 _tensor_constant90" [id=1546, type=get_attr];
-"1547 lift_fresh_copy_62" [id=1547, type=lift_fresh_copy];
-"1548 slice_211" [id=1548, type=slice];
-"1549 slice_212" [id=1549, type=slice];
-"1550 fill__62" [id=1550, type=fill_];
-"1551 view_73" [id=1551, type=view];
-"1552 permute_62" [id=1552, type=permute];
-"1553 reshape_60" [id=1553, type=reshape];
-"1554 unsqueeze_38" [id=1554, type=unsqueeze];
-"1555 unsqueeze_39" [id=1555, type=unsqueeze];
-"1556 sub_6" [id=1556, type=sub];
-"1557 ne_6" [id=1557, type=ne];
-"1558 masked_fill_12" [id=1558, type=masked_fill];
-"1559 eq_6" [id=1559, type=eq];
-"1560 masked_fill_13" [id=1560, type=masked_fill];
-"1561 view_74" [id=1561, type=view];
-"1562 unsqueeze_40" [id=1562, type=unsqueeze];
-"1563 unsqueeze_41" [id=1563, type=unsqueeze];
-"1564 add_46" [id=1564, type=add];
-"1565 view_75" [id=1565, type=view];
-"1566 softmax_13" [id=1566, type=softmax];
-"1567 dropout_52" [id=1567, type=dropout];
-"1568 matmul_27" [id=1568, type=matmul];
-"1569 transpose_27" [id=1569, type=transpose];
-"1570 reshape_61" [id=1570, type=reshape];
-"1571 _param_constant224" [id=1571, type=get_attr];
-"1572 _param_constant225" [id=1572, type=get_attr];
-"1573 linear_83" [id=1573, type=linear];
-"1574 dropout_53" [id=1574, type=dropout];
-"1575 view_76" [id=1575, type=view];
-"1576 permute_63" [id=1576, type=permute];
-"1577 reshape_62" [id=1577, type=reshape];
-"1578 roll_13" [id=1578, type=roll];
-"1579 slice_213" [id=1579, type=slice];
-"1580 slice_214" [id=1580, type=slice];
-"1581 slice_215" [id=1581, type=slice];
-"1582 slice_216" [id=1582, type=slice];
-"1583 contiguous_25" [id=1583, type=contiguous];
-"1584 _param_constant226" [id=1584, type=get_attr];
-"1585 _param_constant227" [id=1585, type=get_attr];
-"1586 layer_norm_29" [id=1586, type=layer_norm];
-"1587 add_47" [id=1587, type=add];
-"1588 _param_constant228" [id=1588, type=get_attr];
-"1589 _param_constant229" [id=1589, type=get_attr];
-"1590 linear_84" [id=1590, type=linear];
-"1591 gelu_13" [id=1591, type=gelu];
-"1592 dropout_54" [id=1592, type=dropout];
-"1593 _param_constant230" [id=1593, type=get_attr];
-"1594 _param_constant231" [id=1594, type=get_attr];
-"1595 linear_85" [id=1595, type=linear];
-"1596 dropout_55" [id=1596, type=dropout];
-"1597 _param_constant232" [id=1597, type=get_attr];
-"1598 _param_constant233" [id=1598, type=get_attr];
-"1599 layer_norm_30" [id=1599, type=layer_norm];
-"1600 add_48" [id=1600, type=add];
-"1601 _tensor_constant91" [id=1601, type=get_attr];
-"1602 _param_constant234" [id=1602, type=get_attr];
-"1603 _param_constant235" [id=1603, type=get_attr];
-"1604 linear_86" [id=1604, type=linear];
-"1605 relu__14" [id=1605, type=relu_];
-"1606 _param_constant236" [id=1606, type=get_attr];
-"1607 linear_87" [id=1607, type=linear];
-"1608 view_77" [id=1608, type=view];
-"1609 _tensor_constant92" [id=1609, type=get_attr];
-"1610 index_14" [id=1610, type=index];
-"1611 view_78" [id=1611, type=view];
-"1612 permute_64" [id=1612, type=permute];
-"1613 contiguous_26" [id=1613, type=contiguous];
-"1614 unsqueeze_42" [id=1614, type=unsqueeze];
-"1615 sigmoid_14" [id=1615, type=sigmoid];
-"1616 mul_28" [id=1616, type=mul];
-"1617 pad_16" [id=1617, type=pad];
-"1618 view_79" [id=1618, type=view];
-"1619 permute_65" [id=1619, type=permute];
-"1620 reshape_63" [id=1620, type=reshape];
-"1621 _param_constant237" [id=1621, type=get_attr];
-"1622 clone_14" [id=1622, type=clone];
-"1623 slice_217" [id=1623, type=slice];
-"1624 zero__14" [id=1624, type=zero_];
-"1625 _param_constant238" [id=1625, type=get_attr];
-"1626 linear_88" [id=1626, type=linear];
-"1627 reshape_64" [id=1627, type=reshape];
-"1628 permute_66" [id=1628, type=permute];
-"1629 select_42" [id=1629, type=select];
-"1630 select_43" [id=1630, type=select];
-"1631 select_44" [id=1631, type=select];
-"1632 linalg_vector_norm_28" [id=1632, type=linalg_vector_norm];
-"1633 clamp_min_28" [id=1633, type=clamp_min];
-"1634 expand_as_28" [id=1634, type=expand_as];
-"1635 div_28" [id=1635, type=div];
-"1636 linalg_vector_norm_29" [id=1636, type=linalg_vector_norm];
-"1637 clamp_min_29" [id=1637, type=clamp_min];
-"1638 expand_as_29" [id=1638, type=expand_as];
-"1639 div_29" [id=1639, type=div];
-"1640 transpose_28" [id=1640, type=transpose];
-"1641 matmul_28" [id=1641, type=matmul];
-"1642 _param_constant239" [id=1642, type=get_attr];
-"1643 clamp_14" [id=1643, type=clamp];
-"1644 exp_14" [id=1644, type=exp];
-"1645 mul_29" [id=1645, type=mul];
-"1646 add_49" [id=1646, type=add];
-"1647 softmax_14" [id=1647, type=softmax];
-"1648 dropout_56" [id=1648, type=dropout];
-"1649 matmul_29" [id=1649, type=matmul];
-"1650 transpose_29" [id=1650, type=transpose];
-"1651 reshape_65" [id=1651, type=reshape];
-"1652 _param_constant240" [id=1652, type=get_attr];
-"1653 _param_constant241" [id=1653, type=get_attr];
-"1654 linear_89" [id=1654, type=linear];
-"1655 dropout_57" [id=1655, type=dropout];
-"1656 view_80" [id=1656, type=view];
-"1657 permute_67" [id=1657, type=permute];
-"1658 reshape_66" [id=1658, type=reshape];
-"1659 slice_218" [id=1659, type=slice];
-"1660 slice_219" [id=1660, type=slice];
-"1661 slice_220" [id=1661, type=slice];
-"1662 slice_221" [id=1662, type=slice];
-"1663 contiguous_27" [id=1663, type=contiguous];
-"1664 _param_constant242" [id=1664, type=get_attr];
-"1665 _param_constant243" [id=1665, type=get_attr];
-"1666 layer_norm_31" [id=1666, type=layer_norm];
-"1667 add_50" [id=1667, type=add];
-"1668 _param_constant244" [id=1668, type=get_attr];
-"1669 _param_constant245" [id=1669, type=get_attr];
-"1670 linear_90" [id=1670, type=linear];
-"1671 gelu_14" [id=1671, type=gelu];
-"1672 dropout_58" [id=1672, type=dropout];
-"1673 _param_constant246" [id=1673, type=get_attr];
-"1674 _param_constant247" [id=1674, type=get_attr];
-"1675 linear_91" [id=1675, type=linear];
-"1676 dropout_59" [id=1676, type=dropout];
-"1677 _param_constant248" [id=1677, type=get_attr];
-"1678 _param_constant249" [id=1678, type=get_attr];
-"1679 layer_norm_32" [id=1679, type=layer_norm];
-"1680 add_51" [id=1680, type=add];
-"1681 _tensor_constant93" [id=1681, type=get_attr];
-"1682 _param_constant250" [id=1682, type=get_attr];
-"1683 _param_constant251" [id=1683, type=get_attr];
-"1684 linear_92" [id=1684, type=linear];
-"1685 relu__15" [id=1685, type=relu_];
-"1686 _param_constant252" [id=1686, type=get_attr];
-"1687 linear_93" [id=1687, type=linear];
-"1688 view_81" [id=1688, type=view];
-"1689 _tensor_constant94" [id=1689, type=get_attr];
-"1690 index_15" [id=1690, type=index];
-"1691 view_82" [id=1691, type=view];
-"1692 permute_68" [id=1692, type=permute];
-"1693 contiguous_28" [id=1693, type=contiguous];
-"1694 unsqueeze_43" [id=1694, type=unsqueeze];
-"1695 sigmoid_15" [id=1695, type=sigmoid];
-"1696 mul_30" [id=1696, type=mul];
-"1697 pad_17" [id=1697, type=pad];
-"1698 roll_14" [id=1698, type=roll];
-"1699 view_83" [id=1699, type=view];
-"1700 permute_69" [id=1700, type=permute];
-"1701 reshape_67" [id=1701, type=reshape];
-"1702 _param_constant253" [id=1702, type=get_attr];
-"1703 clone_15" [id=1703, type=clone];
-"1704 slice_222" [id=1704, type=slice];
-"1705 zero__15" [id=1705, type=zero_];
-"1706 _param_constant254" [id=1706, type=get_attr];
-"1707 linear_94" [id=1707, type=linear];
-"1708 reshape_68" [id=1708, type=reshape];
-"1709 permute_70" [id=1709, type=permute];
-"1710 select_45" [id=1710, type=select];
-"1711 select_46" [id=1711, type=select];
-"1712 select_47" [id=1712, type=select];
-"1713 linalg_vector_norm_30" [id=1713, type=linalg_vector_norm];
-"1714 clamp_min_30" [id=1714, type=clamp_min];
-"1715 expand_as_30" [id=1715, type=expand_as];
-"1716 div_30" [id=1716, type=div];
-"1717 linalg_vector_norm_31" [id=1717, type=linalg_vector_norm];
-"1718 clamp_min_31" [id=1718, type=clamp_min];
-"1719 expand_as_31" [id=1719, type=expand_as];
-"1720 div_31" [id=1720, type=div];
-"1721 transpose_30" [id=1721, type=transpose];
-"1722 matmul_30" [id=1722, type=matmul];
-"1723 _param_constant255" [id=1723, type=get_attr];
-"1724 clamp_15" [id=1724, type=clamp];
-"1725 exp_15" [id=1725, type=exp];
-"1726 mul_31" [id=1726, type=mul];
-"1727 add_52" [id=1727, type=add];
-"1728 new_zeros_7" [id=1728, type=new_zeros];
-"1729 _tensor_constant95" [id=1729, type=get_attr];
-"1730 lift_fresh_copy_63" [id=1730, type=lift_fresh_copy];
-"1731 slice_223" [id=1731, type=slice];
-"1732 slice_224" [id=1732, type=slice];
-"1733 fill__63" [id=1733, type=fill_];
-"1734 _tensor_constant96" [id=1734, type=get_attr];
-"1735 lift_fresh_copy_64" [id=1735, type=lift_fresh_copy];
-"1736 slice_225" [id=1736, type=slice];
-"1737 slice_226" [id=1737, type=slice];
-"1738 fill__64" [id=1738, type=fill_];
-"1739 _tensor_constant97" [id=1739, type=get_attr];
-"1740 lift_fresh_copy_65" [id=1740, type=lift_fresh_copy];
-"1741 slice_227" [id=1741, type=slice];
-"1742 slice_228" [id=1742, type=slice];
-"1743 fill__65" [id=1743, type=fill_];
-"1744 _tensor_constant98" [id=1744, type=get_attr];
-"1745 lift_fresh_copy_66" [id=1745, type=lift_fresh_copy];
-"1746 slice_229" [id=1746, type=slice];
-"1747 slice_230" [id=1747, type=slice];
-"1748 fill__66" [id=1748, type=fill_];
-"1749 _tensor_constant99" [id=1749, type=get_attr];
-"1750 lift_fresh_copy_67" [id=1750, type=lift_fresh_copy];
-"1751 slice_231" [id=1751, type=slice];
-"1752 slice_232" [id=1752, type=slice];
-"1753 fill__67" [id=1753, type=fill_];
-"1754 _tensor_constant100" [id=1754, type=get_attr];
-"1755 lift_fresh_copy_68" [id=1755, type=lift_fresh_copy];
-"1756 slice_233" [id=1756, type=slice];
-"1757 slice_234" [id=1757, type=slice];
-"1758 fill__68" [id=1758, type=fill_];
-"1759 _tensor_constant101" [id=1759, type=get_attr];
-"1760 lift_fresh_copy_69" [id=1760, type=lift_fresh_copy];
-"1761 slice_235" [id=1761, type=slice];
-"1762 slice_236" [id=1762, type=slice];
-"1763 fill__69" [id=1763, type=fill_];
-"1764 _tensor_constant102" [id=1764, type=get_attr];
-"1765 lift_fresh_copy_70" [id=1765, type=lift_fresh_copy];
-"1766 slice_237" [id=1766, type=slice];
-"1767 slice_238" [id=1767, type=slice];
-"1768 fill__70" [id=1768, type=fill_];
-"1769 _tensor_constant103" [id=1769, type=get_attr];
-"1770 lift_fresh_copy_71" [id=1770, type=lift_fresh_copy];
-"1771 slice_239" [id=1771, type=slice];
-"1772 slice_240" [id=1772, type=slice];
-"1773 fill__71" [id=1773, type=fill_];
-"1774 view_84" [id=1774, type=view];
-"1775 permute_71" [id=1775, type=permute];
-"1776 reshape_69" [id=1776, type=reshape];
-"1777 unsqueeze_44" [id=1777, type=unsqueeze];
-"1778 unsqueeze_45" [id=1778, type=unsqueeze];
-"1779 sub_7" [id=1779, type=sub];
-"1780 ne_7" [id=1780, type=ne];
-"1781 masked_fill_14" [id=1781, type=masked_fill];
-"1782 eq_7" [id=1782, type=eq];
-"1783 masked_fill_15" [id=1783, type=masked_fill];
-"1784 view_85" [id=1784, type=view];
-"1785 unsqueeze_46" [id=1785, type=unsqueeze];
-"1786 unsqueeze_47" [id=1786, type=unsqueeze];
-"1787 add_53" [id=1787, type=add];
-"1788 view_86" [id=1788, type=view];
-"1789 softmax_15" [id=1789, type=softmax];
-"1790 dropout_60" [id=1790, type=dropout];
-"1791 matmul_31" [id=1791, type=matmul];
-"1792 transpose_31" [id=1792, type=transpose];
-"1793 reshape_70" [id=1793, type=reshape];
-"1794 _param_constant256" [id=1794, type=get_attr];
-"1795 _param_constant257" [id=1795, type=get_attr];
-"1796 linear_95" [id=1796, type=linear];
-"1797 dropout_61" [id=1797, type=dropout];
-"1798 view_87" [id=1798, type=view];
-"1799 permute_72" [id=1799, type=permute];
-"1800 reshape_71" [id=1800, type=reshape];
-"1801 roll_15" [id=1801, type=roll];
-"1802 slice_241" [id=1802, type=slice];
-"1803 slice_242" [id=1803, type=slice];
-"1804 slice_243" [id=1804, type=slice];
-"1805 slice_244" [id=1805, type=slice];
-"1806 contiguous_29" [id=1806, type=contiguous];
-"1807 _param_constant258" [id=1807, type=get_attr];
-"1808 _param_constant259" [id=1808, type=get_attr];
-"1809 layer_norm_33" [id=1809, type=layer_norm];
-"1810 add_54" [id=1810, type=add];
-"1811 _param_constant260" [id=1811, type=get_attr];
-"1812 _param_constant261" [id=1812, type=get_attr];
-"1813 linear_96" [id=1813, type=linear];
-"1814 gelu_15" [id=1814, type=gelu];
-"1815 dropout_62" [id=1815, type=dropout];
-"1816 _param_constant262" [id=1816, type=get_attr];
-"1817 _param_constant263" [id=1817, type=get_attr];
-"1818 linear_97" [id=1818, type=linear];
-"1819 dropout_63" [id=1819, type=dropout];
-"1820 _param_constant264" [id=1820, type=get_attr];
-"1821 _param_constant265" [id=1821, type=get_attr];
-"1822 layer_norm_34" [id=1822, type=layer_norm];
-"1823 add_55" [id=1823, type=add];
-"1824 _tensor_constant104" [id=1824, type=get_attr];
-"1825 _param_constant266" [id=1825, type=get_attr];
-"1826 _param_constant267" [id=1826, type=get_attr];
-"1827 linear_98" [id=1827, type=linear];
-"1828 relu__16" [id=1828, type=relu_];
-"1829 _param_constant268" [id=1829, type=get_attr];
-"1830 linear_99" [id=1830, type=linear];
-"1831 view_88" [id=1831, type=view];
-"1832 _tensor_constant105" [id=1832, type=get_attr];
-"1833 index_16" [id=1833, type=index];
-"1834 view_89" [id=1834, type=view];
-"1835 permute_73" [id=1835, type=permute];
-"1836 contiguous_30" [id=1836, type=contiguous];
-"1837 unsqueeze_48" [id=1837, type=unsqueeze];
-"1838 sigmoid_16" [id=1838, type=sigmoid];
-"1839 mul_32" [id=1839, type=mul];
-"1840 pad_18" [id=1840, type=pad];
-"1841 view_90" [id=1841, type=view];
-"1842 permute_74" [id=1842, type=permute];
-"1843 reshape_72" [id=1843, type=reshape];
-"1844 _param_constant269" [id=1844, type=get_attr];
-"1845 clone_16" [id=1845, type=clone];
-"1846 slice_245" [id=1846, type=slice];
-"1847 zero__16" [id=1847, type=zero_];
-"1848 _param_constant270" [id=1848, type=get_attr];
-"1849 linear_100" [id=1849, type=linear];
-"1850 reshape_73" [id=1850, type=reshape];
-"1851 permute_75" [id=1851, type=permute];
-"1852 select_48" [id=1852, type=select];
-"1853 select_49" [id=1853, type=select];
-"1854 select_50" [id=1854, type=select];
-"1855 linalg_vector_norm_32" [id=1855, type=linalg_vector_norm];
-"1856 clamp_min_32" [id=1856, type=clamp_min];
-"1857 expand_as_32" [id=1857, type=expand_as];
-"1858 div_32" [id=1858, type=div];
-"1859 linalg_vector_norm_33" [id=1859, type=linalg_vector_norm];
-"1860 clamp_min_33" [id=1860, type=clamp_min];
-"1861 expand_as_33" [id=1861, type=expand_as];
-"1862 div_33" [id=1862, type=div];
-"1863 transpose_32" [id=1863, type=transpose];
-"1864 matmul_32" [id=1864, type=matmul];
-"1865 _param_constant271" [id=1865, type=get_attr];
-"1866 clamp_16" [id=1866, type=clamp];
-"1867 exp_16" [id=1867, type=exp];
-"1868 mul_33" [id=1868, type=mul];
-"1869 add_56" [id=1869, type=add];
-"1870 softmax_16" [id=1870, type=softmax];
-"1871 dropout_64" [id=1871, type=dropout];
-"1872 matmul_33" [id=1872, type=matmul];
-"1873 transpose_33" [id=1873, type=transpose];
-"1874 reshape_74" [id=1874, type=reshape];
-"1875 _param_constant272" [id=1875, type=get_attr];
-"1876 _param_constant273" [id=1876, type=get_attr];
-"1877 linear_101" [id=1877, type=linear];
-"1878 dropout_65" [id=1878, type=dropout];
-"1879 view_91" [id=1879, type=view];
-"1880 permute_76" [id=1880, type=permute];
-"1881 reshape_75" [id=1881, type=reshape];
-"1882 slice_246" [id=1882, type=slice];
-"1883 slice_247" [id=1883, type=slice];
-"1884 slice_248" [id=1884, type=slice];
-"1885 slice_249" [id=1885, type=slice];
-"1886 contiguous_31" [id=1886, type=contiguous];
-"1887 _param_constant274" [id=1887, type=get_attr];
-"1888 _param_constant275" [id=1888, type=get_attr];
-"1889 layer_norm_35" [id=1889, type=layer_norm];
-"1890 add_57" [id=1890, type=add];
-"1891 _param_constant276" [id=1891, type=get_attr];
-"1892 _param_constant277" [id=1892, type=get_attr];
-"1893 linear_102" [id=1893, type=linear];
-"1894 gelu_16" [id=1894, type=gelu];
-"1895 dropout_66" [id=1895, type=dropout];
-"1896 _param_constant278" [id=1896, type=get_attr];
-"1897 _param_constant279" [id=1897, type=get_attr];
-"1898 linear_103" [id=1898, type=linear];
-"1899 dropout_67" [id=1899, type=dropout];
-"1900 _param_constant280" [id=1900, type=get_attr];
-"1901 _param_constant281" [id=1901, type=get_attr];
-"1902 layer_norm_36" [id=1902, type=layer_norm];
-"1903 add_58" [id=1903, type=add];
-"1904 _tensor_constant106" [id=1904, type=get_attr];
-"1905 _param_constant282" [id=1905, type=get_attr];
-"1906 _param_constant283" [id=1906, type=get_attr];
-"1907 linear_104" [id=1907, type=linear];
-"1908 relu__17" [id=1908, type=relu_];
-"1909 _param_constant284" [id=1909, type=get_attr];
-"1910 linear_105" [id=1910, type=linear];
-"1911 view_92" [id=1911, type=view];
-"1912 _tensor_constant107" [id=1912, type=get_attr];
-"1913 index_17" [id=1913, type=index];
-"1914 view_93" [id=1914, type=view];
-"1915 permute_77" [id=1915, type=permute];
-"1916 contiguous_32" [id=1916, type=contiguous];
-"1917 unsqueeze_49" [id=1917, type=unsqueeze];
-"1918 sigmoid_17" [id=1918, type=sigmoid];
-"1919 mul_34" [id=1919, type=mul];
-"1920 pad_19" [id=1920, type=pad];
-"1921 roll_16" [id=1921, type=roll];
-"1922 view_94" [id=1922, type=view];
-"1923 permute_78" [id=1923, type=permute];
-"1924 reshape_76" [id=1924, type=reshape];
-"1925 _param_constant285" [id=1925, type=get_attr];
-"1926 clone_17" [id=1926, type=clone];
-"1927 slice_250" [id=1927, type=slice];
-"1928 zero__17" [id=1928, type=zero_];
-"1929 _param_constant286" [id=1929, type=get_attr];
-"1930 linear_106" [id=1930, type=linear];
-"1931 reshape_77" [id=1931, type=reshape];
-"1932 permute_79" [id=1932, type=permute];
-"1933 select_51" [id=1933, type=select];
-"1934 select_52" [id=1934, type=select];
-"1935 select_53" [id=1935, type=select];
-"1936 linalg_vector_norm_34" [id=1936, type=linalg_vector_norm];
-"1937 clamp_min_34" [id=1937, type=clamp_min];
-"1938 expand_as_34" [id=1938, type=expand_as];
-"1939 div_34" [id=1939, type=div];
-"1940 linalg_vector_norm_35" [id=1940, type=linalg_vector_norm];
-"1941 clamp_min_35" [id=1941, type=clamp_min];
-"1942 expand_as_35" [id=1942, type=expand_as];
-"1943 div_35" [id=1943, type=div];
-"1944 transpose_34" [id=1944, type=transpose];
-"1945 matmul_34" [id=1945, type=matmul];
-"1946 _param_constant287" [id=1946, type=get_attr];
-"1947 clamp_17" [id=1947, type=clamp];
-"1948 exp_17" [id=1948, type=exp];
-"1949 mul_35" [id=1949, type=mul];
-"1950 add_59" [id=1950, type=add];
-"1951 new_zeros_8" [id=1951, type=new_zeros];
-"1952 _tensor_constant108" [id=1952, type=get_attr];
-"1953 lift_fresh_copy_72" [id=1953, type=lift_fresh_copy];
-"1954 slice_251" [id=1954, type=slice];
-"1955 slice_252" [id=1955, type=slice];
-"1956 fill__72" [id=1956, type=fill_];
-"1957 _tensor_constant109" [id=1957, type=get_attr];
-"1958 lift_fresh_copy_73" [id=1958, type=lift_fresh_copy];
-"1959 slice_253" [id=1959, type=slice];
-"1960 slice_254" [id=1960, type=slice];
-"1961 fill__73" [id=1961, type=fill_];
-"1962 _tensor_constant110" [id=1962, type=get_attr];
-"1963 lift_fresh_copy_74" [id=1963, type=lift_fresh_copy];
-"1964 slice_255" [id=1964, type=slice];
-"1965 slice_256" [id=1965, type=slice];
-"1966 fill__74" [id=1966, type=fill_];
-"1967 _tensor_constant111" [id=1967, type=get_attr];
-"1968 lift_fresh_copy_75" [id=1968, type=lift_fresh_copy];
-"1969 slice_257" [id=1969, type=slice];
-"1970 slice_258" [id=1970, type=slice];
-"1971 fill__75" [id=1971, type=fill_];
-"1972 _tensor_constant112" [id=1972, type=get_attr];
-"1973 lift_fresh_copy_76" [id=1973, type=lift_fresh_copy];
-"1974 slice_259" [id=1974, type=slice];
-"1975 slice_260" [id=1975, type=slice];
-"1976 fill__76" [id=1976, type=fill_];
-"1977 _tensor_constant113" [id=1977, type=get_attr];
-"1978 lift_fresh_copy_77" [id=1978, type=lift_fresh_copy];
-"1979 slice_261" [id=1979, type=slice];
-"1980 slice_262" [id=1980, type=slice];
-"1981 fill__77" [id=1981, type=fill_];
-"1982 _tensor_constant114" [id=1982, type=get_attr];
-"1983 lift_fresh_copy_78" [id=1983, type=lift_fresh_copy];
-"1984 slice_263" [id=1984, type=slice];
-"1985 slice_264" [id=1985, type=slice];
-"1986 fill__78" [id=1986, type=fill_];
-"1987 _tensor_constant115" [id=1987, type=get_attr];
-"1988 lift_fresh_copy_79" [id=1988, type=lift_fresh_copy];
-"1989 slice_265" [id=1989, type=slice];
-"1990 slice_266" [id=1990, type=slice];
-"1991 fill__79" [id=1991, type=fill_];
-"1992 _tensor_constant116" [id=1992, type=get_attr];
-"1993 lift_fresh_copy_80" [id=1993, type=lift_fresh_copy];
-"1994 slice_267" [id=1994, type=slice];
-"1995 slice_268" [id=1995, type=slice];
-"1996 fill__80" [id=1996, type=fill_];
-"1997 view_95" [id=1997, type=view];
-"1998 permute_80" [id=1998, type=permute];
-"1999 reshape_78" [id=1999, type=reshape];
-"2000 unsqueeze_50" [id=2000, type=unsqueeze];
-"2001 unsqueeze_51" [id=2001, type=unsqueeze];
-"2002 sub_8" [id=2002, type=sub];
-"2003 ne_8" [id=2003, type=ne];
-"2004 masked_fill_16" [id=2004, type=masked_fill];
-"2005 eq_8" [id=2005, type=eq];
-"2006 masked_fill_17" [id=2006, type=masked_fill];
-"2007 view_96" [id=2007, type=view];
-"2008 unsqueeze_52" [id=2008, type=unsqueeze];
-"2009 unsqueeze_53" [id=2009, type=unsqueeze];
-"2010 add_60" [id=2010, type=add];
-"2011 view_97" [id=2011, type=view];
-"2012 softmax_17" [id=2012, type=softmax];
-"2013 dropout_68" [id=2013, type=dropout];
-"2014 matmul_35" [id=2014, type=matmul];
-"2015 transpose_35" [id=2015, type=transpose];
-"2016 reshape_79" [id=2016, type=reshape];
-"2017 _param_constant288" [id=2017, type=get_attr];
-"2018 _param_constant289" [id=2018, type=get_attr];
-"2019 linear_107" [id=2019, type=linear];
-"2020 dropout_69" [id=2020, type=dropout];
-"2021 view_98" [id=2021, type=view];
-"2022 permute_81" [id=2022, type=permute];
-"2023 reshape_80" [id=2023, type=reshape];
-"2024 roll_17" [id=2024, type=roll];
-"2025 slice_269" [id=2025, type=slice];
-"2026 slice_270" [id=2026, type=slice];
-"2027 slice_271" [id=2027, type=slice];
-"2028 slice_272" [id=2028, type=slice];
-"2029 contiguous_33" [id=2029, type=contiguous];
-"2030 _param_constant290" [id=2030, type=get_attr];
-"2031 _param_constant291" [id=2031, type=get_attr];
-"2032 layer_norm_37" [id=2032, type=layer_norm];
-"2033 add_61" [id=2033, type=add];
-"2034 _param_constant292" [id=2034, type=get_attr];
-"2035 _param_constant293" [id=2035, type=get_attr];
-"2036 linear_108" [id=2036, type=linear];
-"2037 gelu_17" [id=2037, type=gelu];
-"2038 dropout_70" [id=2038, type=dropout];
-"2039 _param_constant294" [id=2039, type=get_attr];
-"2040 _param_constant295" [id=2040, type=get_attr];
-"2041 linear_109" [id=2041, type=linear];
-"2042 dropout_71" [id=2042, type=dropout];
-"2043 _param_constant296" [id=2043, type=get_attr];
-"2044 _param_constant297" [id=2044, type=get_attr];
-"2045 layer_norm_38" [id=2045, type=layer_norm];
-"2046 add_62" [id=2046, type=add];
-"2047 _tensor_constant117" [id=2047, type=get_attr];
-"2048 _param_constant298" [id=2048, type=get_attr];
-"2049 _param_constant299" [id=2049, type=get_attr];
-"2050 linear_110" [id=2050, type=linear];
-"2051 relu__18" [id=2051, type=relu_];
-"2052 _param_constant300" [id=2052, type=get_attr];
-"2053 linear_111" [id=2053, type=linear];
-"2054 view_99" [id=2054, type=view];
-"2055 _tensor_constant118" [id=2055, type=get_attr];
-"2056 index_18" [id=2056, type=index];
-"2057 view_100" [id=2057, type=view];
-"2058 permute_82" [id=2058, type=permute];
-"2059 contiguous_34" [id=2059, type=contiguous];
-"2060 unsqueeze_54" [id=2060, type=unsqueeze];
-"2061 sigmoid_18" [id=2061, type=sigmoid];
-"2062 mul_36" [id=2062, type=mul];
-"2063 pad_20" [id=2063, type=pad];
-"2064 view_101" [id=2064, type=view];
-"2065 permute_83" [id=2065, type=permute];
-"2066 reshape_81" [id=2066, type=reshape];
-"2067 _param_constant301" [id=2067, type=get_attr];
-"2068 clone_18" [id=2068, type=clone];
-"2069 slice_273" [id=2069, type=slice];
-"2070 zero__18" [id=2070, type=zero_];
-"2071 _param_constant302" [id=2071, type=get_attr];
-"2072 linear_112" [id=2072, type=linear];
-"2073 reshape_82" [id=2073, type=reshape];
-"2074 permute_84" [id=2074, type=permute];
-"2075 select_54" [id=2075, type=select];
-"2076 select_55" [id=2076, type=select];
-"2077 select_56" [id=2077, type=select];
-"2078 linalg_vector_norm_36" [id=2078, type=linalg_vector_norm];
-"2079 clamp_min_36" [id=2079, type=clamp_min];
-"2080 expand_as_36" [id=2080, type=expand_as];
-"2081 div_36" [id=2081, type=div];
-"2082 linalg_vector_norm_37" [id=2082, type=linalg_vector_norm];
-"2083 clamp_min_37" [id=2083, type=clamp_min];
-"2084 expand_as_37" [id=2084, type=expand_as];
-"2085 div_37" [id=2085, type=div];
-"2086 transpose_36" [id=2086, type=transpose];
-"2087 matmul_36" [id=2087, type=matmul];
-"2088 _param_constant303" [id=2088, type=get_attr];
-"2089 clamp_18" [id=2089, type=clamp];
-"2090 exp_18" [id=2090, type=exp];
-"2091 mul_37" [id=2091, type=mul];
-"2092 add_63" [id=2092, type=add];
-"2093 softmax_18" [id=2093, type=softmax];
-"2094 dropout_72" [id=2094, type=dropout];
-"2095 matmul_37" [id=2095, type=matmul];
-"2096 transpose_37" [id=2096, type=transpose];
-"2097 reshape_83" [id=2097, type=reshape];
-"2098 _param_constant304" [id=2098, type=get_attr];
-"2099 _param_constant305" [id=2099, type=get_attr];
-"2100 linear_113" [id=2100, type=linear];
-"2101 dropout_73" [id=2101, type=dropout];
-"2102 view_102" [id=2102, type=view];
-"2103 permute_85" [id=2103, type=permute];
-"2104 reshape_84" [id=2104, type=reshape];
-"2105 slice_274" [id=2105, type=slice];
-"2106 slice_275" [id=2106, type=slice];
-"2107 slice_276" [id=2107, type=slice];
-"2108 slice_277" [id=2108, type=slice];
-"2109 contiguous_35" [id=2109, type=contiguous];
-"2110 _param_constant306" [id=2110, type=get_attr];
-"2111 _param_constant307" [id=2111, type=get_attr];
-"2112 layer_norm_39" [id=2112, type=layer_norm];
-"2113 add_64" [id=2113, type=add];
-"2114 _param_constant308" [id=2114, type=get_attr];
-"2115 _param_constant309" [id=2115, type=get_attr];
-"2116 linear_114" [id=2116, type=linear];
-"2117 gelu_18" [id=2117, type=gelu];
-"2118 dropout_74" [id=2118, type=dropout];
-"2119 _param_constant310" [id=2119, type=get_attr];
-"2120 _param_constant311" [id=2120, type=get_attr];
-"2121 linear_115" [id=2121, type=linear];
-"2122 dropout_75" [id=2122, type=dropout];
-"2123 _param_constant312" [id=2123, type=get_attr];
-"2124 _param_constant313" [id=2124, type=get_attr];
-"2125 layer_norm_40" [id=2125, type=layer_norm];
-"2126 add_65" [id=2126, type=add];
-"2127 _tensor_constant119" [id=2127, type=get_attr];
-"2128 _param_constant314" [id=2128, type=get_attr];
-"2129 _param_constant315" [id=2129, type=get_attr];
-"2130 linear_116" [id=2130, type=linear];
-"2131 relu__19" [id=2131, type=relu_];
-"2132 _param_constant316" [id=2132, type=get_attr];
-"2133 linear_117" [id=2133, type=linear];
-"2134 view_103" [id=2134, type=view];
-"2135 _tensor_constant120" [id=2135, type=get_attr];
-"2136 index_19" [id=2136, type=index];
-"2137 view_104" [id=2137, type=view];
-"2138 permute_86" [id=2138, type=permute];
-"2139 contiguous_36" [id=2139, type=contiguous];
-"2140 unsqueeze_55" [id=2140, type=unsqueeze];
-"2141 sigmoid_19" [id=2141, type=sigmoid];
-"2142 mul_38" [id=2142, type=mul];
-"2143 pad_21" [id=2143, type=pad];
-"2144 roll_18" [id=2144, type=roll];
-"2145 view_105" [id=2145, type=view];
-"2146 permute_87" [id=2146, type=permute];
-"2147 reshape_85" [id=2147, type=reshape];
-"2148 _param_constant317" [id=2148, type=get_attr];
-"2149 clone_19" [id=2149, type=clone];
-"2150 slice_278" [id=2150, type=slice];
-"2151 zero__19" [id=2151, type=zero_];
-"2152 _param_constant318" [id=2152, type=get_attr];
-"2153 linear_118" [id=2153, type=linear];
-"2154 reshape_86" [id=2154, type=reshape];
-"2155 permute_88" [id=2155, type=permute];
-"2156 select_57" [id=2156, type=select];
-"2157 select_58" [id=2157, type=select];
-"2158 select_59" [id=2158, type=select];
-"2159 linalg_vector_norm_38" [id=2159, type=linalg_vector_norm];
-"2160 clamp_min_38" [id=2160, type=clamp_min];
-"2161 expand_as_38" [id=2161, type=expand_as];
-"2162 div_38" [id=2162, type=div];
-"2163 linalg_vector_norm_39" [id=2163, type=linalg_vector_norm];
-"2164 clamp_min_39" [id=2164, type=clamp_min];
-"2165 expand_as_39" [id=2165, type=expand_as];
-"2166 div_39" [id=2166, type=div];
-"2167 transpose_38" [id=2167, type=transpose];
-"2168 matmul_38" [id=2168, type=matmul];
-"2169 _param_constant319" [id=2169, type=get_attr];
-"2170 clamp_19" [id=2170, type=clamp];
-"2171 exp_19" [id=2171, type=exp];
-"2172 mul_39" [id=2172, type=mul];
-"2173 add_66" [id=2173, type=add];
-"2174 new_zeros_9" [id=2174, type=new_zeros];
-"2175 _tensor_constant121" [id=2175, type=get_attr];
-"2176 lift_fresh_copy_81" [id=2176, type=lift_fresh_copy];
-"2177 slice_279" [id=2177, type=slice];
-"2178 slice_280" [id=2178, type=slice];
-"2179 fill__81" [id=2179, type=fill_];
-"2180 _tensor_constant122" [id=2180, type=get_attr];
-"2181 lift_fresh_copy_82" [id=2181, type=lift_fresh_copy];
-"2182 slice_281" [id=2182, type=slice];
-"2183 slice_282" [id=2183, type=slice];
-"2184 fill__82" [id=2184, type=fill_];
-"2185 _tensor_constant123" [id=2185, type=get_attr];
-"2186 lift_fresh_copy_83" [id=2186, type=lift_fresh_copy];
-"2187 slice_283" [id=2187, type=slice];
-"2188 slice_284" [id=2188, type=slice];
-"2189 fill__83" [id=2189, type=fill_];
-"2190 _tensor_constant124" [id=2190, type=get_attr];
-"2191 lift_fresh_copy_84" [id=2191, type=lift_fresh_copy];
-"2192 slice_285" [id=2192, type=slice];
-"2193 slice_286" [id=2193, type=slice];
-"2194 fill__84" [id=2194, type=fill_];
-"2195 _tensor_constant125" [id=2195, type=get_attr];
-"2196 lift_fresh_copy_85" [id=2196, type=lift_fresh_copy];
-"2197 slice_287" [id=2197, type=slice];
-"2198 slice_288" [id=2198, type=slice];
-"2199 fill__85" [id=2199, type=fill_];
-"2200 _tensor_constant126" [id=2200, type=get_attr];
-"2201 lift_fresh_copy_86" [id=2201, type=lift_fresh_copy];
-"2202 slice_289" [id=2202, type=slice];
-"2203 slice_290" [id=2203, type=slice];
-"2204 fill__86" [id=2204, type=fill_];
-"2205 _tensor_constant127" [id=2205, type=get_attr];
-"2206 lift_fresh_copy_87" [id=2206, type=lift_fresh_copy];
-"2207 slice_291" [id=2207, type=slice];
-"2208 slice_292" [id=2208, type=slice];
-"2209 fill__87" [id=2209, type=fill_];
-"2210 _tensor_constant128" [id=2210, type=get_attr];
-"2211 lift_fresh_copy_88" [id=2211, type=lift_fresh_copy];
-"2212 slice_293" [id=2212, type=slice];
-"2213 slice_294" [id=2213, type=slice];
-"2214 fill__88" [id=2214, type=fill_];
-"2215 _tensor_constant129" [id=2215, type=get_attr];
-"2216 lift_fresh_copy_89" [id=2216, type=lift_fresh_copy];
-"2217 slice_295" [id=2217, type=slice];
-"2218 slice_296" [id=2218, type=slice];
-"2219 fill__89" [id=2219, type=fill_];
-"2220 view_106" [id=2220, type=view];
-"2221 permute_89" [id=2221, type=permute];
-"2222 reshape_87" [id=2222, type=reshape];
-"2223 unsqueeze_56" [id=2223, type=unsqueeze];
-"2224 unsqueeze_57" [id=2224, type=unsqueeze];
-"2225 sub_9" [id=2225, type=sub];
-"2226 ne_9" [id=2226, type=ne];
-"2227 masked_fill_18" [id=2227, type=masked_fill];
-"2228 eq_9" [id=2228, type=eq];
-"2229 masked_fill_19" [id=2229, type=masked_fill];
-"2230 view_107" [id=2230, type=view];
-"2231 unsqueeze_58" [id=2231, type=unsqueeze];
-"2232 unsqueeze_59" [id=2232, type=unsqueeze];
-"2233 add_67" [id=2233, type=add];
-"2234 view_108" [id=2234, type=view];
-"2235 softmax_19" [id=2235, type=softmax];
-"2236 dropout_76" [id=2236, type=dropout];
-"2237 matmul_39" [id=2237, type=matmul];
-"2238 transpose_39" [id=2238, type=transpose];
-"2239 reshape_88" [id=2239, type=reshape];
-"2240 _param_constant320" [id=2240, type=get_attr];
-"2241 _param_constant321" [id=2241, type=get_attr];
-"2242 linear_119" [id=2242, type=linear];
-"2243 dropout_77" [id=2243, type=dropout];
-"2244 view_109" [id=2244, type=view];
-"2245 permute_90" [id=2245, type=permute];
-"2246 reshape_89" [id=2246, type=reshape];
-"2247 roll_19" [id=2247, type=roll];
-"2248 slice_297" [id=2248, type=slice];
-"2249 slice_298" [id=2249, type=slice];
-"2250 slice_299" [id=2250, type=slice];
-"2251 slice_300" [id=2251, type=slice];
-"2252 contiguous_37" [id=2252, type=contiguous];
-"2253 _param_constant322" [id=2253, type=get_attr];
-"2254 _param_constant323" [id=2254, type=get_attr];
-"2255 layer_norm_41" [id=2255, type=layer_norm];
-"2256 add_68" [id=2256, type=add];
-"2257 _param_constant324" [id=2257, type=get_attr];
-"2258 _param_constant325" [id=2258, type=get_attr];
-"2259 linear_120" [id=2259, type=linear];
-"2260 gelu_19" [id=2260, type=gelu];
-"2261 dropout_78" [id=2261, type=dropout];
-"2262 _param_constant326" [id=2262, type=get_attr];
-"2263 _param_constant327" [id=2263, type=get_attr];
-"2264 linear_121" [id=2264, type=linear];
-"2265 dropout_79" [id=2265, type=dropout];
-"2266 _param_constant328" [id=2266, type=get_attr];
-"2267 _param_constant329" [id=2267, type=get_attr];
-"2268 layer_norm_42" [id=2268, type=layer_norm];
-"2269 add_69" [id=2269, type=add];
-"2270 _tensor_constant130" [id=2270, type=get_attr];
-"2271 _param_constant330" [id=2271, type=get_attr];
-"2272 _param_constant331" [id=2272, type=get_attr];
-"2273 linear_122" [id=2273, type=linear];
-"2274 relu__20" [id=2274, type=relu_];
-"2275 _param_constant332" [id=2275, type=get_attr];
-"2276 linear_123" [id=2276, type=linear];
-"2277 view_110" [id=2277, type=view];
-"2278 _tensor_constant131" [id=2278, type=get_attr];
-"2279 index_20" [id=2279, type=index];
-"2280 view_111" [id=2280, type=view];
-"2281 permute_91" [id=2281, type=permute];
-"2282 contiguous_38" [id=2282, type=contiguous];
-"2283 unsqueeze_60" [id=2283, type=unsqueeze];
-"2284 sigmoid_20" [id=2284, type=sigmoid];
-"2285 mul_40" [id=2285, type=mul];
-"2286 pad_22" [id=2286, type=pad];
-"2287 view_112" [id=2287, type=view];
-"2288 permute_92" [id=2288, type=permute];
-"2289 reshape_90" [id=2289, type=reshape];
-"2290 _param_constant333" [id=2290, type=get_attr];
-"2291 clone_20" [id=2291, type=clone];
-"2292 slice_301" [id=2292, type=slice];
-"2293 zero__20" [id=2293, type=zero_];
-"2294 _param_constant334" [id=2294, type=get_attr];
-"2295 linear_124" [id=2295, type=linear];
-"2296 reshape_91" [id=2296, type=reshape];
-"2297 permute_93" [id=2297, type=permute];
-"2298 select_60" [id=2298, type=select];
-"2299 select_61" [id=2299, type=select];
-"2300 select_62" [id=2300, type=select];
-"2301 linalg_vector_norm_40" [id=2301, type=linalg_vector_norm];
-"2302 clamp_min_40" [id=2302, type=clamp_min];
-"2303 expand_as_40" [id=2303, type=expand_as];
-"2304 div_40" [id=2304, type=div];
-"2305 linalg_vector_norm_41" [id=2305, type=linalg_vector_norm];
-"2306 clamp_min_41" [id=2306, type=clamp_min];
-"2307 expand_as_41" [id=2307, type=expand_as];
-"2308 div_41" [id=2308, type=div];
-"2309 transpose_40" [id=2309, type=transpose];
-"2310 matmul_40" [id=2310, type=matmul];
-"2311 _param_constant335" [id=2311, type=get_attr];
-"2312 clamp_20" [id=2312, type=clamp];
-"2313 exp_20" [id=2313, type=exp];
-"2314 mul_41" [id=2314, type=mul];
-"2315 add_70" [id=2315, type=add];
-"2316 softmax_20" [id=2316, type=softmax];
-"2317 dropout_80" [id=2317, type=dropout];
-"2318 matmul_41" [id=2318, type=matmul];
-"2319 transpose_41" [id=2319, type=transpose];
-"2320 reshape_92" [id=2320, type=reshape];
-"2321 _param_constant336" [id=2321, type=get_attr];
-"2322 _param_constant337" [id=2322, type=get_attr];
-"2323 linear_125" [id=2323, type=linear];
-"2324 dropout_81" [id=2324, type=dropout];
-"2325 view_113" [id=2325, type=view];
-"2326 permute_94" [id=2326, type=permute];
-"2327 reshape_93" [id=2327, type=reshape];
-"2328 slice_302" [id=2328, type=slice];
-"2329 slice_303" [id=2329, type=slice];
-"2330 slice_304" [id=2330, type=slice];
-"2331 slice_305" [id=2331, type=slice];
-"2332 contiguous_39" [id=2332, type=contiguous];
-"2333 _param_constant338" [id=2333, type=get_attr];
-"2334 _param_constant339" [id=2334, type=get_attr];
-"2335 layer_norm_43" [id=2335, type=layer_norm];
-"2336 add_71" [id=2336, type=add];
-"2337 _param_constant340" [id=2337, type=get_attr];
-"2338 _param_constant341" [id=2338, type=get_attr];
-"2339 linear_126" [id=2339, type=linear];
-"2340 gelu_20" [id=2340, type=gelu];
-"2341 dropout_82" [id=2341, type=dropout];
-"2342 _param_constant342" [id=2342, type=get_attr];
-"2343 _param_constant343" [id=2343, type=get_attr];
-"2344 linear_127" [id=2344, type=linear];
-"2345 dropout_83" [id=2345, type=dropout];
-"2346 _param_constant344" [id=2346, type=get_attr];
-"2347 _param_constant345" [id=2347, type=get_attr];
-"2348 layer_norm_44" [id=2348, type=layer_norm];
-"2349 add_72" [id=2349, type=add];
-"2350 _tensor_constant132" [id=2350, type=get_attr];
-"2351 _param_constant346" [id=2351, type=get_attr];
-"2352 _param_constant347" [id=2352, type=get_attr];
-"2353 linear_128" [id=2353, type=linear];
-"2354 relu__21" [id=2354, type=relu_];
-"2355 _param_constant348" [id=2355, type=get_attr];
-"2356 linear_129" [id=2356, type=linear];
-"2357 view_114" [id=2357, type=view];
-"2358 _tensor_constant133" [id=2358, type=get_attr];
-"2359 index_21" [id=2359, type=index];
-"2360 view_115" [id=2360, type=view];
-"2361 permute_95" [id=2361, type=permute];
-"2362 contiguous_40" [id=2362, type=contiguous];
-"2363 unsqueeze_61" [id=2363, type=unsqueeze];
-"2364 sigmoid_21" [id=2364, type=sigmoid];
-"2365 mul_42" [id=2365, type=mul];
-"2366 pad_23" [id=2366, type=pad];
-"2367 roll_20" [id=2367, type=roll];
-"2368 view_116" [id=2368, type=view];
-"2369 permute_96" [id=2369, type=permute];
-"2370 reshape_94" [id=2370, type=reshape];
-"2371 _param_constant349" [id=2371, type=get_attr];
-"2372 clone_21" [id=2372, type=clone];
-"2373 slice_306" [id=2373, type=slice];
-"2374 zero__21" [id=2374, type=zero_];
-"2375 _param_constant350" [id=2375, type=get_attr];
-"2376 linear_130" [id=2376, type=linear];
-"2377 reshape_95" [id=2377, type=reshape];
-"2378 permute_97" [id=2378, type=permute];
-"2379 select_63" [id=2379, type=select];
-"2380 select_64" [id=2380, type=select];
-"2381 select_65" [id=2381, type=select];
-"2382 linalg_vector_norm_42" [id=2382, type=linalg_vector_norm];
-"2383 clamp_min_42" [id=2383, type=clamp_min];
-"2384 expand_as_42" [id=2384, type=expand_as];
-"2385 div_42" [id=2385, type=div];
-"2386 linalg_vector_norm_43" [id=2386, type=linalg_vector_norm];
-"2387 clamp_min_43" [id=2387, type=clamp_min];
-"2388 expand_as_43" [id=2388, type=expand_as];
-"2389 div_43" [id=2389, type=div];
-"2390 transpose_42" [id=2390, type=transpose];
-"2391 matmul_42" [id=2391, type=matmul];
-"2392 _param_constant351" [id=2392, type=get_attr];
-"2393 clamp_21" [id=2393, type=clamp];
-"2394 exp_21" [id=2394, type=exp];
-"2395 mul_43" [id=2395, type=mul];
-"2396 add_73" [id=2396, type=add];
-"2397 new_zeros_10" [id=2397, type=new_zeros];
-"2398 _tensor_constant134" [id=2398, type=get_attr];
-"2399 lift_fresh_copy_90" [id=2399, type=lift_fresh_copy];
-"2400 slice_307" [id=2400, type=slice];
-"2401 slice_308" [id=2401, type=slice];
-"2402 fill__90" [id=2402, type=fill_];
-"2403 _tensor_constant135" [id=2403, type=get_attr];
-"2404 lift_fresh_copy_91" [id=2404, type=lift_fresh_copy];
-"2405 slice_309" [id=2405, type=slice];
-"2406 slice_310" [id=2406, type=slice];
-"2407 fill__91" [id=2407, type=fill_];
-"2408 _tensor_constant136" [id=2408, type=get_attr];
-"2409 lift_fresh_copy_92" [id=2409, type=lift_fresh_copy];
-"2410 slice_311" [id=2410, type=slice];
-"2411 slice_312" [id=2411, type=slice];
-"2412 fill__92" [id=2412, type=fill_];
-"2413 _tensor_constant137" [id=2413, type=get_attr];
-"2414 lift_fresh_copy_93" [id=2414, type=lift_fresh_copy];
-"2415 slice_313" [id=2415, type=slice];
-"2416 slice_314" [id=2416, type=slice];
-"2417 fill__93" [id=2417, type=fill_];
-"2418 _tensor_constant138" [id=2418, type=get_attr];
-"2419 lift_fresh_copy_94" [id=2419, type=lift_fresh_copy];
-"2420 slice_315" [id=2420, type=slice];
-"2421 slice_316" [id=2421, type=slice];
-"2422 fill__94" [id=2422, type=fill_];
-"2423 _tensor_constant139" [id=2423, type=get_attr];
-"2424 lift_fresh_copy_95" [id=2424, type=lift_fresh_copy];
-"2425 slice_317" [id=2425, type=slice];
-"2426 slice_318" [id=2426, type=slice];
-"2427 fill__95" [id=2427, type=fill_];
-"2428 _tensor_constant140" [id=2428, type=get_attr];
-"2429 lift_fresh_copy_96" [id=2429, type=lift_fresh_copy];
-"2430 slice_319" [id=2430, type=slice];
-"2431 slice_320" [id=2431, type=slice];
-"2432 fill__96" [id=2432, type=fill_];
-"2433 _tensor_constant141" [id=2433, type=get_attr];
-"2434 lift_fresh_copy_97" [id=2434, type=lift_fresh_copy];
-"2435 slice_321" [id=2435, type=slice];
-"2436 slice_322" [id=2436, type=slice];
-"2437 fill__97" [id=2437, type=fill_];
-"2438 _tensor_constant142" [id=2438, type=get_attr];
-"2439 lift_fresh_copy_98" [id=2439, type=lift_fresh_copy];
-"2440 slice_323" [id=2440, type=slice];
-"2441 slice_324" [id=2441, type=slice];
-"2442 fill__98" [id=2442, type=fill_];
-"2443 view_117" [id=2443, type=view];
-"2444 permute_98" [id=2444, type=permute];
-"2445 reshape_96" [id=2445, type=reshape];
-"2446 unsqueeze_62" [id=2446, type=unsqueeze];
-"2447 unsqueeze_63" [id=2447, type=unsqueeze];
-"2448 sub_10" [id=2448, type=sub];
-"2449 ne_10" [id=2449, type=ne];
-"2450 masked_fill_20" [id=2450, type=masked_fill];
-"2451 eq_10" [id=2451, type=eq];
-"2452 masked_fill_21" [id=2452, type=masked_fill];
-"2453 view_118" [id=2453, type=view];
-"2454 unsqueeze_64" [id=2454, type=unsqueeze];
-"2455 unsqueeze_65" [id=2455, type=unsqueeze];
-"2456 add_74" [id=2456, type=add];
-"2457 view_119" [id=2457, type=view];
-"2458 softmax_21" [id=2458, type=softmax];
-"2459 dropout_84" [id=2459, type=dropout];
-"2460 matmul_43" [id=2460, type=matmul];
-"2461 transpose_43" [id=2461, type=transpose];
-"2462 reshape_97" [id=2462, type=reshape];
-"2463 _param_constant352" [id=2463, type=get_attr];
-"2464 _param_constant353" [id=2464, type=get_attr];
-"2465 linear_131" [id=2465, type=linear];
-"2466 dropout_85" [id=2466, type=dropout];
-"2467 view_120" [id=2467, type=view];
-"2468 permute_99" [id=2468, type=permute];
-"2469 reshape_98" [id=2469, type=reshape];
-"2470 roll_21" [id=2470, type=roll];
-"2471 slice_325" [id=2471, type=slice];
-"2472 slice_326" [id=2472, type=slice];
-"2473 slice_327" [id=2473, type=slice];
-"2474 slice_328" [id=2474, type=slice];
-"2475 contiguous_41" [id=2475, type=contiguous];
-"2476 _param_constant354" [id=2476, type=get_attr];
-"2477 _param_constant355" [id=2477, type=get_attr];
-"2478 layer_norm_45" [id=2478, type=layer_norm];
-"2479 add_75" [id=2479, type=add];
-"2480 _param_constant356" [id=2480, type=get_attr];
-"2481 _param_constant357" [id=2481, type=get_attr];
-"2482 linear_132" [id=2482, type=linear];
-"2483 gelu_21" [id=2483, type=gelu];
-"2484 dropout_86" [id=2484, type=dropout];
-"2485 _param_constant358" [id=2485, type=get_attr];
-"2486 _param_constant359" [id=2486, type=get_attr];
-"2487 linear_133" [id=2487, type=linear];
-"2488 dropout_87" [id=2488, type=dropout];
-"2489 _param_constant360" [id=2489, type=get_attr];
-"2490 _param_constant361" [id=2490, type=get_attr];
-"2491 layer_norm_46" [id=2491, type=layer_norm];
-"2492 add_76" [id=2492, type=add];
-"2493 pad_24" [id=2493, type=pad];
-"2494 slice_329" [id=2494, type=slice];
-"2495 slice_330" [id=2495, type=slice];
-"2496 slice_331" [id=2496, type=slice];
-"2497 slice_332" [id=2497, type=slice];
-"2498 slice_333" [id=2498, type=slice];
-"2499 slice_334" [id=2499, type=slice];
-"2500 slice_335" [id=2500, type=slice];
-"2501 slice_336" [id=2501, type=slice];
-"2502 slice_337" [id=2502, type=slice];
-"2503 slice_338" [id=2503, type=slice];
-"2504 slice_339" [id=2504, type=slice];
-"2505 slice_340" [id=2505, type=slice];
-"2506 cat_2" [id=2506, type=cat];
-"2507 _param_constant362" [id=2507, type=get_attr];
-"2508 linear_134" [id=2508, type=linear];
-"2509 _param_constant363" [id=2509, type=get_attr];
-"2510 _param_constant364" [id=2510, type=get_attr];
-"2511 layer_norm_47" [id=2511, type=layer_norm];
-"2512 _tensor_constant143" [id=2512, type=get_attr];
-"2513 _param_constant365" [id=2513, type=get_attr];
-"2514 _param_constant366" [id=2514, type=get_attr];
-"2515 linear_135" [id=2515, type=linear];
-"2516 relu__22" [id=2516, type=relu_];
-"2517 _param_constant367" [id=2517, type=get_attr];
-"2518 linear_136" [id=2518, type=linear];
-"2519 view_121" [id=2519, type=view];
-"2520 _tensor_constant144" [id=2520, type=get_attr];
-"2521 index_22" [id=2521, type=index];
-"2522 view_122" [id=2522, type=view];
-"2523 permute_100" [id=2523, type=permute];
-"2524 contiguous_42" [id=2524, type=contiguous];
-"2525 unsqueeze_66" [id=2525, type=unsqueeze];
-"2526 sigmoid_22" [id=2526, type=sigmoid];
-"2527 mul_44" [id=2527, type=mul];
-"2528 pad_25" [id=2528, type=pad];
-"2529 view_123" [id=2529, type=view];
-"2530 permute_101" [id=2530, type=permute];
-"2531 reshape_99" [id=2531, type=reshape];
-"2532 _param_constant368" [id=2532, type=get_attr];
-"2533 clone_22" [id=2533, type=clone];
-"2534 slice_341" [id=2534, type=slice];
-"2535 zero__22" [id=2535, type=zero_];
-"2536 _param_constant369" [id=2536, type=get_attr];
-"2537 linear_137" [id=2537, type=linear];
-"2538 reshape_100" [id=2538, type=reshape];
-"2539 permute_102" [id=2539, type=permute];
-"2540 select_66" [id=2540, type=select];
-"2541 select_67" [id=2541, type=select];
-"2542 select_68" [id=2542, type=select];
-"2543 linalg_vector_norm_44" [id=2543, type=linalg_vector_norm];
-"2544 clamp_min_44" [id=2544, type=clamp_min];
-"2545 expand_as_44" [id=2545, type=expand_as];
-"2546 div_44" [id=2546, type=div];
-"2547 linalg_vector_norm_45" [id=2547, type=linalg_vector_norm];
-"2548 clamp_min_45" [id=2548, type=clamp_min];
-"2549 expand_as_45" [id=2549, type=expand_as];
-"2550 div_45" [id=2550, type=div];
-"2551 transpose_44" [id=2551, type=transpose];
-"2552 matmul_44" [id=2552, type=matmul];
-"2553 _param_constant370" [id=2553, type=get_attr];
-"2554 clamp_22" [id=2554, type=clamp];
-"2555 exp_22" [id=2555, type=exp];
-"2556 mul_45" [id=2556, type=mul];
-"2557 add_77" [id=2557, type=add];
-"2558 softmax_22" [id=2558, type=softmax];
-"2559 dropout_88" [id=2559, type=dropout];
-"2560 matmul_45" [id=2560, type=matmul];
-"2561 transpose_45" [id=2561, type=transpose];
-"2562 reshape_101" [id=2562, type=reshape];
-"2563 _param_constant371" [id=2563, type=get_attr];
-"2564 _param_constant372" [id=2564, type=get_attr];
-"2565 linear_138" [id=2565, type=linear];
-"2566 dropout_89" [id=2566, type=dropout];
-"2567 view_124" [id=2567, type=view];
-"2568 permute_103" [id=2568, type=permute];
-"2569 reshape_102" [id=2569, type=reshape];
-"2570 slice_342" [id=2570, type=slice];
-"2571 slice_343" [id=2571, type=slice];
-"2572 slice_344" [id=2572, type=slice];
-"2573 slice_345" [id=2573, type=slice];
-"2574 contiguous_43" [id=2574, type=contiguous];
-"2575 _param_constant373" [id=2575, type=get_attr];
-"2576 _param_constant374" [id=2576, type=get_attr];
-"2577 layer_norm_48" [id=2577, type=layer_norm];
-"2578 add_78" [id=2578, type=add];
-"2579 _param_constant375" [id=2579, type=get_attr];
-"2580 _param_constant376" [id=2580, type=get_attr];
-"2581 linear_139" [id=2581, type=linear];
-"2582 gelu_22" [id=2582, type=gelu];
-"2583 dropout_90" [id=2583, type=dropout];
-"2584 _param_constant377" [id=2584, type=get_attr];
-"2585 _param_constant378" [id=2585, type=get_attr];
-"2586 linear_140" [id=2586, type=linear];
-"2587 dropout_91" [id=2587, type=dropout];
-"2588 _param_constant379" [id=2588, type=get_attr];
-"2589 _param_constant380" [id=2589, type=get_attr];
-"2590 layer_norm_49" [id=2590, type=layer_norm];
-"2591 add_79" [id=2591, type=add];
-"2592 _tensor_constant145" [id=2592, type=get_attr];
-"2593 _param_constant381" [id=2593, type=get_attr];
-"2594 _param_constant382" [id=2594, type=get_attr];
-"2595 linear_141" [id=2595, type=linear];
-"2596 relu__23" [id=2596, type=relu_];
-"2597 _param_constant383" [id=2597, type=get_attr];
-"2598 linear_142" [id=2598, type=linear];
-"2599 view_125" [id=2599, type=view];
-"2600 _tensor_constant146" [id=2600, type=get_attr];
-"2601 index_23" [id=2601, type=index];
-"2602 view_126" [id=2602, type=view];
-"2603 permute_104" [id=2603, type=permute];
-"2604 contiguous_44" [id=2604, type=contiguous];
-"2605 unsqueeze_67" [id=2605, type=unsqueeze];
-"2606 sigmoid_23" [id=2606, type=sigmoid];
-"2607 mul_46" [id=2607, type=mul];
-"2608 pad_26" [id=2608, type=pad];
-"2609 view_127" [id=2609, type=view];
-"2610 permute_105" [id=2610, type=permute];
-"2611 reshape_103" [id=2611, type=reshape];
-"2612 _param_constant384" [id=2612, type=get_attr];
-"2613 clone_23" [id=2613, type=clone];
-"2614 slice_346" [id=2614, type=slice];
-"2615 zero__23" [id=2615, type=zero_];
-"2616 _param_constant385" [id=2616, type=get_attr];
-"2617 linear_143" [id=2617, type=linear];
-"2618 reshape_104" [id=2618, type=reshape];
-"2619 permute_106" [id=2619, type=permute];
-"2620 select_69" [id=2620, type=select];
-"2621 select_70" [id=2621, type=select];
-"2622 select_71" [id=2622, type=select];
-"2623 linalg_vector_norm_46" [id=2623, type=linalg_vector_norm];
-"2624 clamp_min_46" [id=2624, type=clamp_min];
-"2625 expand_as_46" [id=2625, type=expand_as];
-"2626 div_46" [id=2626, type=div];
-"2627 linalg_vector_norm_47" [id=2627, type=linalg_vector_norm];
-"2628 clamp_min_47" [id=2628, type=clamp_min];
-"2629 expand_as_47" [id=2629, type=expand_as];
-"2630 div_47" [id=2630, type=div];
-"2631 transpose_46" [id=2631, type=transpose];
-"2632 matmul_46" [id=2632, type=matmul];
-"2633 _param_constant386" [id=2633, type=get_attr];
-"2634 clamp_23" [id=2634, type=clamp];
-"2635 exp_23" [id=2635, type=exp];
-"2636 mul_47" [id=2636, type=mul];
-"2637 add_80" [id=2637, type=add];
-"2638 softmax_23" [id=2638, type=softmax];
-"2639 dropout_92" [id=2639, type=dropout];
-"2640 matmul_47" [id=2640, type=matmul];
-"2641 transpose_47" [id=2641, type=transpose];
-"2642 reshape_105" [id=2642, type=reshape];
-"2643 _param_constant387" [id=2643, type=get_attr];
-"2644 _param_constant388" [id=2644, type=get_attr];
-"2645 linear_144" [id=2645, type=linear];
-"2646 dropout_93" [id=2646, type=dropout];
-"2647 view_128" [id=2647, type=view];
-"2648 permute_107" [id=2648, type=permute];
-"2649 reshape_106" [id=2649, type=reshape];
-"2650 slice_347" [id=2650, type=slice];
-"2651 slice_348" [id=2651, type=slice];
-"2652 slice_349" [id=2652, type=slice];
-"2653 slice_350" [id=2653, type=slice];
-"2654 contiguous_45" [id=2654, type=contiguous];
-"2655 _param_constant389" [id=2655, type=get_attr];
-"2656 _param_constant390" [id=2656, type=get_attr];
-"2657 layer_norm_50" [id=2657, type=layer_norm];
-"2658 add_81" [id=2658, type=add];
-"2659 _param_constant391" [id=2659, type=get_attr];
-"2660 _param_constant392" [id=2660, type=get_attr];
-"2661 linear_145" [id=2661, type=linear];
-"2662 gelu_23" [id=2662, type=gelu];
-"2663 dropout_94" [id=2663, type=dropout];
-"2664 _param_constant393" [id=2664, type=get_attr];
-"2665 _param_constant394" [id=2665, type=get_attr];
-"2666 linear_146" [id=2666, type=linear];
-"2667 dropout_95" [id=2667, type=dropout];
-"2668 _param_constant395" [id=2668, type=get_attr];
-"2669 _param_constant396" [id=2669, type=get_attr];
-"2670 layer_norm_51" [id=2670, type=layer_norm];
-"2671 add_82" [id=2671, type=add];
-"2672 _param_constant397" [id=2672, type=get_attr];
-"2673 _param_constant398" [id=2673, type=get_attr];
-"2674 layer_norm_52" [id=2674, type=layer_norm];
-"2675 permute_108" [id=2675, type=permute];
-"2676 adaptive_avg_pool2d" [id=2676, type=adaptive_avg_pool2d];
-"2677 flatten" [id=2677, type=flatten];
-"2678 _param_constant399" [id=2678, type=get_attr];
-"2679 _param_constant400" [id=2679, type=get_attr];
+"0 features_1_0_attn_qkv_weight" [id=0, type=get_attr];
+"1 features_1_0_attn_proj_weight" [id=1, type=get_attr];
+"2 features_1_0_attn_qkv_bias" [id=2, type=get_attr];
+"3 features_1_0_attn_proj_bias" [id=3, type=get_attr];
+"4 features_1_0_attn_logit_scale" [id=4, type=get_attr];
+"5 features_1_1_attn_qkv_weight" [id=5, type=get_attr];
+"6 features_1_1_attn_proj_weight" [id=6, type=get_attr];
+"7 features_1_1_attn_qkv_bias" [id=7, type=get_attr];
+"8 features_1_1_attn_proj_bias" [id=8, type=get_attr];
+"9 features_1_1_attn_logit_scale" [id=9, type=get_attr];
+"10 features_3_0_attn_qkv_weight" [id=10, type=get_attr];
+"11 features_3_0_attn_proj_weight" [id=11, type=get_attr];
+"12 features_3_0_attn_qkv_bias" [id=12, type=get_attr];
+"13 features_3_0_attn_proj_bias" [id=13, type=get_attr];
+"14 features_3_0_attn_logit_scale" [id=14, type=get_attr];
+"15 features_3_1_attn_qkv_weight" [id=15, type=get_attr];
+"16 features_3_1_attn_proj_weight" [id=16, type=get_attr];
+"17 features_3_1_attn_qkv_bias" [id=17, type=get_attr];
+"18 features_3_1_attn_proj_bias" [id=18, type=get_attr];
+"19 features_3_1_attn_logit_scale" [id=19, type=get_attr];
+"20 features_5_0_attn_qkv_weight" [id=20, type=get_attr];
+"21 features_5_0_attn_proj_weight" [id=21, type=get_attr];
+"22 features_5_0_attn_qkv_bias" [id=22, type=get_attr];
+"23 features_5_0_attn_proj_bias" [id=23, type=get_attr];
+"24 features_5_0_attn_logit_scale" [id=24, type=get_attr];
+"25 features_5_1_attn_qkv_weight" [id=25, type=get_attr];
+"26 features_5_1_attn_proj_weight" [id=26, type=get_attr];
+"27 features_5_1_attn_qkv_bias" [id=27, type=get_attr];
+"28 features_5_1_attn_proj_bias" [id=28, type=get_attr];
+"29 features_5_1_attn_logit_scale" [id=29, type=get_attr];
+"30 features_5_2_attn_qkv_weight" [id=30, type=get_attr];
+"31 features_5_2_attn_proj_weight" [id=31, type=get_attr];
+"32 features_5_2_attn_qkv_bias" [id=32, type=get_attr];
+"33 features_5_2_attn_proj_bias" [id=33, type=get_attr];
+"34 features_5_2_attn_logit_scale" [id=34, type=get_attr];
+"35 features_5_3_attn_qkv_weight" [id=35, type=get_attr];
+"36 features_5_3_attn_proj_weight" [id=36, type=get_attr];
+"37 features_5_3_attn_qkv_bias" [id=37, type=get_attr];
+"38 features_5_3_attn_proj_bias" [id=38, type=get_attr];
+"39 features_5_3_attn_logit_scale" [id=39, type=get_attr];
+"40 features_5_4_attn_qkv_weight" [id=40, type=get_attr];
+"41 features_5_4_attn_proj_weight" [id=41, type=get_attr];
+"42 features_5_4_attn_qkv_bias" [id=42, type=get_attr];
+"43 features_5_4_attn_proj_bias" [id=43, type=get_attr];
+"44 features_5_4_attn_logit_scale" [id=44, type=get_attr];
+"45 features_5_5_attn_qkv_weight" [id=45, type=get_attr];
+"46 features_5_5_attn_proj_weight" [id=46, type=get_attr];
+"47 features_5_5_attn_qkv_bias" [id=47, type=get_attr];
+"48 features_5_5_attn_proj_bias" [id=48, type=get_attr];
+"49 features_5_5_attn_logit_scale" [id=49, type=get_attr];
+"50 features_5_6_attn_qkv_weight" [id=50, type=get_attr];
+"51 features_5_6_attn_proj_weight" [id=51, type=get_attr];
+"52 features_5_6_attn_qkv_bias" [id=52, type=get_attr];
+"53 features_5_6_attn_proj_bias" [id=53, type=get_attr];
+"54 features_5_6_attn_logit_scale" [id=54, type=get_attr];
+"55 features_5_7_attn_qkv_weight" [id=55, type=get_attr];
+"56 features_5_7_attn_proj_weight" [id=56, type=get_attr];
+"57 features_5_7_attn_qkv_bias" [id=57, type=get_attr];
+"58 features_5_7_attn_proj_bias" [id=58, type=get_attr];
+"59 features_5_7_attn_logit_scale" [id=59, type=get_attr];
+"60 features_5_8_attn_qkv_weight" [id=60, type=get_attr];
+"61 features_5_8_attn_proj_weight" [id=61, type=get_attr];
+"62 features_5_8_attn_qkv_bias" [id=62, type=get_attr];
+"63 features_5_8_attn_proj_bias" [id=63, type=get_attr];
+"64 features_5_8_attn_logit_scale" [id=64, type=get_attr];
+"65 features_5_9_attn_qkv_weight" [id=65, type=get_attr];
+"66 features_5_9_attn_proj_weight" [id=66, type=get_attr];
+"67 features_5_9_attn_qkv_bias" [id=67, type=get_attr];
+"68 features_5_9_attn_proj_bias" [id=68, type=get_attr];
+"69 features_5_9_attn_logit_scale" [id=69, type=get_attr];
+"70 features_5_10_attn_qkv_weight" [id=70, type=get_attr];
+"71 features_5_10_attn_proj_weight" [id=71, type=get_attr];
+"72 features_5_10_attn_qkv_bias" [id=72, type=get_attr];
+"73 features_5_10_attn_proj_bias" [id=73, type=get_attr];
+"74 features_5_10_attn_logit_scale" [id=74, type=get_attr];
+"75 features_5_11_attn_qkv_weight" [id=75, type=get_attr];
+"76 features_5_11_attn_proj_weight" [id=76, type=get_attr];
+"77 features_5_11_attn_qkv_bias" [id=77, type=get_attr];
+"78 features_5_11_attn_proj_bias" [id=78, type=get_attr];
+"79 features_5_11_attn_logit_scale" [id=79, type=get_attr];
+"80 features_5_12_attn_qkv_weight" [id=80, type=get_attr];
+"81 features_5_12_attn_proj_weight" [id=81, type=get_attr];
+"82 features_5_12_attn_qkv_bias" [id=82, type=get_attr];
+"83 features_5_12_attn_proj_bias" [id=83, type=get_attr];
+"84 features_5_12_attn_logit_scale" [id=84, type=get_attr];
+"85 features_5_13_attn_qkv_weight" [id=85, type=get_attr];
+"86 features_5_13_attn_proj_weight" [id=86, type=get_attr];
+"87 features_5_13_attn_qkv_bias" [id=87, type=get_attr];
+"88 features_5_13_attn_proj_bias" [id=88, type=get_attr];
+"89 features_5_13_attn_logit_scale" [id=89, type=get_attr];
+"90 features_5_14_attn_qkv_weight" [id=90, type=get_attr];
+"91 features_5_14_attn_proj_weight" [id=91, type=get_attr];
+"92 features_5_14_attn_qkv_bias" [id=92, type=get_attr];
+"93 features_5_14_attn_proj_bias" [id=93, type=get_attr];
+"94 features_5_14_attn_logit_scale" [id=94, type=get_attr];
+"95 features_5_15_attn_qkv_weight" [id=95, type=get_attr];
+"96 features_5_15_attn_proj_weight" [id=96, type=get_attr];
+"97 features_5_15_attn_qkv_bias" [id=97, type=get_attr];
+"98 features_5_15_attn_proj_bias" [id=98, type=get_attr];
+"99 features_5_15_attn_logit_scale" [id=99, type=get_attr];
+"100 features_5_16_attn_qkv_weight" [id=100, type=get_attr];
+"101 features_5_16_attn_proj_weight" [id=101, type=get_attr];
+"102 features_5_16_attn_qkv_bias" [id=102, type=get_attr];
+"103 features_5_16_attn_proj_bias" [id=103, type=get_attr];
+"104 features_5_16_attn_logit_scale" [id=104, type=get_attr];
+"105 features_5_17_attn_qkv_weight" [id=105, type=get_attr];
+"106 features_5_17_attn_proj_weight" [id=106, type=get_attr];
+"107 features_5_17_attn_qkv_bias" [id=107, type=get_attr];
+"108 features_5_17_attn_proj_bias" [id=108, type=get_attr];
+"109 features_5_17_attn_logit_scale" [id=109, type=get_attr];
+"110 features_7_0_attn_qkv_weight" [id=110, type=get_attr];
+"111 features_7_0_attn_proj_weight" [id=111, type=get_attr];
+"112 features_7_0_attn_qkv_bias" [id=112, type=get_attr];
+"113 features_7_0_attn_proj_bias" [id=113, type=get_attr];
+"114 features_7_0_attn_logit_scale" [id=114, type=get_attr];
+"115 features_7_1_attn_qkv_weight" [id=115, type=get_attr];
+"116 features_7_1_attn_proj_weight" [id=116, type=get_attr];
+"117 features_7_1_attn_qkv_bias" [id=117, type=get_attr];
+"118 features_7_1_attn_proj_bias" [id=118, type=get_attr];
+"119 features_7_1_attn_logit_scale" [id=119, type=get_attr];
+"120 features_0_0_weight" [id=120, type=get_attr];
+"121 features_0_0_bias" [id=121, type=get_attr];
+"122 features_0_2_weight" [id=122, type=get_attr];
+"123 features_0_2_bias" [id=123, type=get_attr];
+"124 features_1_0_attn_cpb_mlp_0_weight" [id=124, type=get_attr];
+"125 features_1_0_attn_cpb_mlp_0_bias" [id=125, type=get_attr];
+"126 features_1_0_attn_cpb_mlp_2_weight" [id=126, type=get_attr];
+"127 features_1_0_norm1_weight" [id=127, type=get_attr];
+"128 features_1_0_norm1_bias" [id=128, type=get_attr];
+"129 features_1_0_mlp_0_weight" [id=129, type=get_attr];
+"130 features_1_0_mlp_0_bias" [id=130, type=get_attr];
+"131 features_1_0_mlp_3_weight" [id=131, type=get_attr];
+"132 features_1_0_mlp_3_bias" [id=132, type=get_attr];
+"133 features_1_0_norm2_weight" [id=133, type=get_attr];
+"134 features_1_0_norm2_bias" [id=134, type=get_attr];
+"135 features_1_1_attn_cpb_mlp_0_weight" [id=135, type=get_attr];
+"136 features_1_1_attn_cpb_mlp_0_bias" [id=136, type=get_attr];
+"137 features_1_1_attn_cpb_mlp_2_weight" [id=137, type=get_attr];
+"138 features_1_1_norm1_weight" [id=138, type=get_attr];
+"139 features_1_1_norm1_bias" [id=139, type=get_attr];
+"140 features_1_1_mlp_0_weight" [id=140, type=get_attr];
+"141 features_1_1_mlp_0_bias" [id=141, type=get_attr];
+"142 features_1_1_mlp_3_weight" [id=142, type=get_attr];
+"143 features_1_1_mlp_3_bias" [id=143, type=get_attr];
+"144 features_1_1_norm2_weight" [id=144, type=get_attr];
+"145 features_1_1_norm2_bias" [id=145, type=get_attr];
+"146 features_2_reduction_weight" [id=146, type=get_attr];
+"147 features_2_norm_weight" [id=147, type=get_attr];
+"148 features_2_norm_bias" [id=148, type=get_attr];
+"149 features_3_0_attn_cpb_mlp_0_weight" [id=149, type=get_attr];
+"150 features_3_0_attn_cpb_mlp_0_bias" [id=150, type=get_attr];
+"151 features_3_0_attn_cpb_mlp_2_weight" [id=151, type=get_attr];
+"152 features_3_0_norm1_weight" [id=152, type=get_attr];
+"153 features_3_0_norm1_bias" [id=153, type=get_attr];
+"154 features_3_0_mlp_0_weight" [id=154, type=get_attr];
+"155 features_3_0_mlp_0_bias" [id=155, type=get_attr];
+"156 features_3_0_mlp_3_weight" [id=156, type=get_attr];
+"157 features_3_0_mlp_3_bias" [id=157, type=get_attr];
+"158 features_3_0_norm2_weight" [id=158, type=get_attr];
+"159 features_3_0_norm2_bias" [id=159, type=get_attr];
+"160 features_3_1_attn_cpb_mlp_0_weight" [id=160, type=get_attr];
+"161 features_3_1_attn_cpb_mlp_0_bias" [id=161, type=get_attr];
+"162 features_3_1_attn_cpb_mlp_2_weight" [id=162, type=get_attr];
+"163 features_3_1_norm1_weight" [id=163, type=get_attr];
+"164 features_3_1_norm1_bias" [id=164, type=get_attr];
+"165 features_3_1_mlp_0_weight" [id=165, type=get_attr];
+"166 features_3_1_mlp_0_bias" [id=166, type=get_attr];
+"167 features_3_1_mlp_3_weight" [id=167, type=get_attr];
+"168 features_3_1_mlp_3_bias" [id=168, type=get_attr];
+"169 features_3_1_norm2_weight" [id=169, type=get_attr];
+"170 features_3_1_norm2_bias" [id=170, type=get_attr];
+"171 features_4_reduction_weight" [id=171, type=get_attr];
+"172 features_4_norm_weight" [id=172, type=get_attr];
+"173 features_4_norm_bias" [id=173, type=get_attr];
+"174 features_5_0_attn_cpb_mlp_0_weight" [id=174, type=get_attr];
+"175 features_5_0_attn_cpb_mlp_0_bias" [id=175, type=get_attr];
+"176 features_5_0_attn_cpb_mlp_2_weight" [id=176, type=get_attr];
+"177 features_5_0_norm1_weight" [id=177, type=get_attr];
+"178 features_5_0_norm1_bias" [id=178, type=get_attr];
+"179 features_5_0_mlp_0_weight" [id=179, type=get_attr];
+"180 features_5_0_mlp_0_bias" [id=180, type=get_attr];
+"181 features_5_0_mlp_3_weight" [id=181, type=get_attr];
+"182 features_5_0_mlp_3_bias" [id=182, type=get_attr];
+"183 features_5_0_norm2_weight" [id=183, type=get_attr];
+"184 features_5_0_norm2_bias" [id=184, type=get_attr];
+"185 features_5_1_attn_cpb_mlp_0_weight" [id=185, type=get_attr];
+"186 features_5_1_attn_cpb_mlp_0_bias" [id=186, type=get_attr];
+"187 features_5_1_attn_cpb_mlp_2_weight" [id=187, type=get_attr];
+"188 features_5_1_norm1_weight" [id=188, type=get_attr];
+"189 features_5_1_norm1_bias" [id=189, type=get_attr];
+"190 features_5_1_mlp_0_weight" [id=190, type=get_attr];
+"191 features_5_1_mlp_0_bias" [id=191, type=get_attr];
+"192 features_5_1_mlp_3_weight" [id=192, type=get_attr];
+"193 features_5_1_mlp_3_bias" [id=193, type=get_attr];
+"194 features_5_1_norm2_weight" [id=194, type=get_attr];
+"195 features_5_1_norm2_bias" [id=195, type=get_attr];
+"196 features_5_2_attn_cpb_mlp_0_weight" [id=196, type=get_attr];
+"197 features_5_2_attn_cpb_mlp_0_bias" [id=197, type=get_attr];
+"198 features_5_2_attn_cpb_mlp_2_weight" [id=198, type=get_attr];
+"199 features_5_2_norm1_weight" [id=199, type=get_attr];
+"200 features_5_2_norm1_bias" [id=200, type=get_attr];
+"201 features_5_2_mlp_0_weight" [id=201, type=get_attr];
+"202 features_5_2_mlp_0_bias" [id=202, type=get_attr];
+"203 features_5_2_mlp_3_weight" [id=203, type=get_attr];
+"204 features_5_2_mlp_3_bias" [id=204, type=get_attr];
+"205 features_5_2_norm2_weight" [id=205, type=get_attr];
+"206 features_5_2_norm2_bias" [id=206, type=get_attr];
+"207 features_5_3_attn_cpb_mlp_0_weight" [id=207, type=get_attr];
+"208 features_5_3_attn_cpb_mlp_0_bias" [id=208, type=get_attr];
+"209 features_5_3_attn_cpb_mlp_2_weight" [id=209, type=get_attr];
+"210 features_5_3_norm1_weight" [id=210, type=get_attr];
+"211 features_5_3_norm1_bias" [id=211, type=get_attr];
+"212 features_5_3_mlp_0_weight" [id=212, type=get_attr];
+"213 features_5_3_mlp_0_bias" [id=213, type=get_attr];
+"214 features_5_3_mlp_3_weight" [id=214, type=get_attr];
+"215 features_5_3_mlp_3_bias" [id=215, type=get_attr];
+"216 features_5_3_norm2_weight" [id=216, type=get_attr];
+"217 features_5_3_norm2_bias" [id=217, type=get_attr];
+"218 features_5_4_attn_cpb_mlp_0_weight" [id=218, type=get_attr];
+"219 features_5_4_attn_cpb_mlp_0_bias" [id=219, type=get_attr];
+"220 features_5_4_attn_cpb_mlp_2_weight" [id=220, type=get_attr];
+"221 features_5_4_norm1_weight" [id=221, type=get_attr];
+"222 features_5_4_norm1_bias" [id=222, type=get_attr];
+"223 features_5_4_mlp_0_weight" [id=223, type=get_attr];
+"224 features_5_4_mlp_0_bias" [id=224, type=get_attr];
+"225 features_5_4_mlp_3_weight" [id=225, type=get_attr];
+"226 features_5_4_mlp_3_bias" [id=226, type=get_attr];
+"227 features_5_4_norm2_weight" [id=227, type=get_attr];
+"228 features_5_4_norm2_bias" [id=228, type=get_attr];
+"229 features_5_5_attn_cpb_mlp_0_weight" [id=229, type=get_attr];
+"230 features_5_5_attn_cpb_mlp_0_bias" [id=230, type=get_attr];
+"231 features_5_5_attn_cpb_mlp_2_weight" [id=231, type=get_attr];
+"232 features_5_5_norm1_weight" [id=232, type=get_attr];
+"233 features_5_5_norm1_bias" [id=233, type=get_attr];
+"234 features_5_5_mlp_0_weight" [id=234, type=get_attr];
+"235 features_5_5_mlp_0_bias" [id=235, type=get_attr];
+"236 features_5_5_mlp_3_weight" [id=236, type=get_attr];
+"237 features_5_5_mlp_3_bias" [id=237, type=get_attr];
+"238 features_5_5_norm2_weight" [id=238, type=get_attr];
+"239 features_5_5_norm2_bias" [id=239, type=get_attr];
+"240 features_5_6_attn_cpb_mlp_0_weight" [id=240, type=get_attr];
+"241 features_5_6_attn_cpb_mlp_0_bias" [id=241, type=get_attr];
+"242 features_5_6_attn_cpb_mlp_2_weight" [id=242, type=get_attr];
+"243 features_5_6_norm1_weight" [id=243, type=get_attr];
+"244 features_5_6_norm1_bias" [id=244, type=get_attr];
+"245 features_5_6_mlp_0_weight" [id=245, type=get_attr];
+"246 features_5_6_mlp_0_bias" [id=246, type=get_attr];
+"247 features_5_6_mlp_3_weight" [id=247, type=get_attr];
+"248 features_5_6_mlp_3_bias" [id=248, type=get_attr];
+"249 features_5_6_norm2_weight" [id=249, type=get_attr];
+"250 features_5_6_norm2_bias" [id=250, type=get_attr];
+"251 features_5_7_attn_cpb_mlp_0_weight" [id=251, type=get_attr];
+"252 features_5_7_attn_cpb_mlp_0_bias" [id=252, type=get_attr];
+"253 features_5_7_attn_cpb_mlp_2_weight" [id=253, type=get_attr];
+"254 features_5_7_norm1_weight" [id=254, type=get_attr];
+"255 features_5_7_norm1_bias" [id=255, type=get_attr];
+"256 features_5_7_mlp_0_weight" [id=256, type=get_attr];
+"257 features_5_7_mlp_0_bias" [id=257, type=get_attr];
+"258 features_5_7_mlp_3_weight" [id=258, type=get_attr];
+"259 features_5_7_mlp_3_bias" [id=259, type=get_attr];
+"260 features_5_7_norm2_weight" [id=260, type=get_attr];
+"261 features_5_7_norm2_bias" [id=261, type=get_attr];
+"262 features_5_8_attn_cpb_mlp_0_weight" [id=262, type=get_attr];
+"263 features_5_8_attn_cpb_mlp_0_bias" [id=263, type=get_attr];
+"264 features_5_8_attn_cpb_mlp_2_weight" [id=264, type=get_attr];
+"265 features_5_8_norm1_weight" [id=265, type=get_attr];
+"266 features_5_8_norm1_bias" [id=266, type=get_attr];
+"267 features_5_8_mlp_0_weight" [id=267, type=get_attr];
+"268 features_5_8_mlp_0_bias" [id=268, type=get_attr];
+"269 features_5_8_mlp_3_weight" [id=269, type=get_attr];
+"270 features_5_8_mlp_3_bias" [id=270, type=get_attr];
+"271 features_5_8_norm2_weight" [id=271, type=get_attr];
+"272 features_5_8_norm2_bias" [id=272, type=get_attr];
+"273 features_5_9_attn_cpb_mlp_0_weight" [id=273, type=get_attr];
+"274 features_5_9_attn_cpb_mlp_0_bias" [id=274, type=get_attr];
+"275 features_5_9_attn_cpb_mlp_2_weight" [id=275, type=get_attr];
+"276 features_5_9_norm1_weight" [id=276, type=get_attr];
+"277 features_5_9_norm1_bias" [id=277, type=get_attr];
+"278 features_5_9_mlp_0_weight" [id=278, type=get_attr];
+"279 features_5_9_mlp_0_bias" [id=279, type=get_attr];
+"280 features_5_9_mlp_3_weight" [id=280, type=get_attr];
+"281 features_5_9_mlp_3_bias" [id=281, type=get_attr];
+"282 features_5_9_norm2_weight" [id=282, type=get_attr];
+"283 features_5_9_norm2_bias" [id=283, type=get_attr];
+"284 features_5_10_attn_cpb_mlp_0_weight" [id=284, type=get_attr];
+"285 features_5_10_attn_cpb_mlp_0_bias" [id=285, type=get_attr];
+"286 features_5_10_attn_cpb_mlp_2_weight" [id=286, type=get_attr];
+"287 features_5_10_norm1_weight" [id=287, type=get_attr];
+"288 features_5_10_norm1_bias" [id=288, type=get_attr];
+"289 features_5_10_mlp_0_weight" [id=289, type=get_attr];
+"290 features_5_10_mlp_0_bias" [id=290, type=get_attr];
+"291 features_5_10_mlp_3_weight" [id=291, type=get_attr];
+"292 features_5_10_mlp_3_bias" [id=292, type=get_attr];
+"293 features_5_10_norm2_weight" [id=293, type=get_attr];
+"294 features_5_10_norm2_bias" [id=294, type=get_attr];
+"295 features_5_11_attn_cpb_mlp_0_weight" [id=295, type=get_attr];
+"296 features_5_11_attn_cpb_mlp_0_bias" [id=296, type=get_attr];
+"297 features_5_11_attn_cpb_mlp_2_weight" [id=297, type=get_attr];
+"298 features_5_11_norm1_weight" [id=298, type=get_attr];
+"299 features_5_11_norm1_bias" [id=299, type=get_attr];
+"300 features_5_11_mlp_0_weight" [id=300, type=get_attr];
+"301 features_5_11_mlp_0_bias" [id=301, type=get_attr];
+"302 features_5_11_mlp_3_weight" [id=302, type=get_attr];
+"303 features_5_11_mlp_3_bias" [id=303, type=get_attr];
+"304 features_5_11_norm2_weight" [id=304, type=get_attr];
+"305 features_5_11_norm2_bias" [id=305, type=get_attr];
+"306 features_5_12_attn_cpb_mlp_0_weight" [id=306, type=get_attr];
+"307 features_5_12_attn_cpb_mlp_0_bias" [id=307, type=get_attr];
+"308 features_5_12_attn_cpb_mlp_2_weight" [id=308, type=get_attr];
+"309 features_5_12_norm1_weight" [id=309, type=get_attr];
+"310 features_5_12_norm1_bias" [id=310, type=get_attr];
+"311 features_5_12_mlp_0_weight" [id=311, type=get_attr];
+"312 features_5_12_mlp_0_bias" [id=312, type=get_attr];
+"313 features_5_12_mlp_3_weight" [id=313, type=get_attr];
+"314 features_5_12_mlp_3_bias" [id=314, type=get_attr];
+"315 features_5_12_norm2_weight" [id=315, type=get_attr];
+"316 features_5_12_norm2_bias" [id=316, type=get_attr];
+"317 features_5_13_attn_cpb_mlp_0_weight" [id=317, type=get_attr];
+"318 features_5_13_attn_cpb_mlp_0_bias" [id=318, type=get_attr];
+"319 features_5_13_attn_cpb_mlp_2_weight" [id=319, type=get_attr];
+"320 features_5_13_norm1_weight" [id=320, type=get_attr];
+"321 features_5_13_norm1_bias" [id=321, type=get_attr];
+"322 features_5_13_mlp_0_weight" [id=322, type=get_attr];
+"323 features_5_13_mlp_0_bias" [id=323, type=get_attr];
+"324 features_5_13_mlp_3_weight" [id=324, type=get_attr];
+"325 features_5_13_mlp_3_bias" [id=325, type=get_attr];
+"326 features_5_13_norm2_weight" [id=326, type=get_attr];
+"327 features_5_13_norm2_bias" [id=327, type=get_attr];
+"328 features_5_14_attn_cpb_mlp_0_weight" [id=328, type=get_attr];
+"329 features_5_14_attn_cpb_mlp_0_bias" [id=329, type=get_attr];
+"330 features_5_14_attn_cpb_mlp_2_weight" [id=330, type=get_attr];
+"331 features_5_14_norm1_weight" [id=331, type=get_attr];
+"332 features_5_14_norm1_bias" [id=332, type=get_attr];
+"333 features_5_14_mlp_0_weight" [id=333, type=get_attr];
+"334 features_5_14_mlp_0_bias" [id=334, type=get_attr];
+"335 features_5_14_mlp_3_weight" [id=335, type=get_attr];
+"336 features_5_14_mlp_3_bias" [id=336, type=get_attr];
+"337 features_5_14_norm2_weight" [id=337, type=get_attr];
+"338 features_5_14_norm2_bias" [id=338, type=get_attr];
+"339 features_5_15_attn_cpb_mlp_0_weight" [id=339, type=get_attr];
+"340 features_5_15_attn_cpb_mlp_0_bias" [id=340, type=get_attr];
+"341 features_5_15_attn_cpb_mlp_2_weight" [id=341, type=get_attr];
+"342 features_5_15_norm1_weight" [id=342, type=get_attr];
+"343 features_5_15_norm1_bias" [id=343, type=get_attr];
+"344 features_5_15_mlp_0_weight" [id=344, type=get_attr];
+"345 features_5_15_mlp_0_bias" [id=345, type=get_attr];
+"346 features_5_15_mlp_3_weight" [id=346, type=get_attr];
+"347 features_5_15_mlp_3_bias" [id=347, type=get_attr];
+"348 features_5_15_norm2_weight" [id=348, type=get_attr];
+"349 features_5_15_norm2_bias" [id=349, type=get_attr];
+"350 features_5_16_attn_cpb_mlp_0_weight" [id=350, type=get_attr];
+"351 features_5_16_attn_cpb_mlp_0_bias" [id=351, type=get_attr];
+"352 features_5_16_attn_cpb_mlp_2_weight" [id=352, type=get_attr];
+"353 features_5_16_norm1_weight" [id=353, type=get_attr];
+"354 features_5_16_norm1_bias" [id=354, type=get_attr];
+"355 features_5_16_mlp_0_weight" [id=355, type=get_attr];
+"356 features_5_16_mlp_0_bias" [id=356, type=get_attr];
+"357 features_5_16_mlp_3_weight" [id=357, type=get_attr];
+"358 features_5_16_mlp_3_bias" [id=358, type=get_attr];
+"359 features_5_16_norm2_weight" [id=359, type=get_attr];
+"360 features_5_16_norm2_bias" [id=360, type=get_attr];
+"361 features_5_17_attn_cpb_mlp_0_weight" [id=361, type=get_attr];
+"362 features_5_17_attn_cpb_mlp_0_bias" [id=362, type=get_attr];
+"363 features_5_17_attn_cpb_mlp_2_weight" [id=363, type=get_attr];
+"364 features_5_17_norm1_weight" [id=364, type=get_attr];
+"365 features_5_17_norm1_bias" [id=365, type=get_attr];
+"366 features_5_17_mlp_0_weight" [id=366, type=get_attr];
+"367 features_5_17_mlp_0_bias" [id=367, type=get_attr];
+"368 features_5_17_mlp_3_weight" [id=368, type=get_attr];
+"369 features_5_17_mlp_3_bias" [id=369, type=get_attr];
+"370 features_5_17_norm2_weight" [id=370, type=get_attr];
+"371 features_5_17_norm2_bias" [id=371, type=get_attr];
+"372 features_6_reduction_weight" [id=372, type=get_attr];
+"373 features_6_norm_weight" [id=373, type=get_attr];
+"374 features_6_norm_bias" [id=374, type=get_attr];
+"375 features_7_0_attn_cpb_mlp_0_weight" [id=375, type=get_attr];
+"376 features_7_0_attn_cpb_mlp_0_bias" [id=376, type=get_attr];
+"377 features_7_0_attn_cpb_mlp_2_weight" [id=377, type=get_attr];
+"378 features_7_0_norm1_weight" [id=378, type=get_attr];
+"379 features_7_0_norm1_bias" [id=379, type=get_attr];
+"380 features_7_0_mlp_0_weight" [id=380, type=get_attr];
+"381 features_7_0_mlp_0_bias" [id=381, type=get_attr];
+"382 features_7_0_mlp_3_weight" [id=382, type=get_attr];
+"383 features_7_0_mlp_3_bias" [id=383, type=get_attr];
+"384 features_7_0_norm2_weight" [id=384, type=get_attr];
+"385 features_7_0_norm2_bias" [id=385, type=get_attr];
+"386 features_7_1_attn_cpb_mlp_0_weight" [id=386, type=get_attr];
+"387 features_7_1_attn_cpb_mlp_0_bias" [id=387, type=get_attr];
+"388 features_7_1_attn_cpb_mlp_2_weight" [id=388, type=get_attr];
+"389 features_7_1_norm1_weight" [id=389, type=get_attr];
+"390 features_7_1_norm1_bias" [id=390, type=get_attr];
+"391 features_7_1_mlp_0_weight" [id=391, type=get_attr];
+"392 features_7_1_mlp_0_bias" [id=392, type=get_attr];
+"393 features_7_1_mlp_3_weight" [id=393, type=get_attr];
+"394 features_7_1_mlp_3_bias" [id=394, type=get_attr];
+"395 features_7_1_norm2_weight" [id=395, type=get_attr];
+"396 features_7_1_norm2_bias" [id=396, type=get_attr];
+"397 norm_weight" [id=397, type=get_attr];
+"398 norm_bias" [id=398, type=get_attr];
+"399 head_weight" [id=399, type=get_attr];
+"400 head_bias" [id=400, type=get_attr];
+"401 features_1_0_attn_relative_coords_table" [id=401, type=get_attr];
+"402 features_1_0_attn_relative_position_index" [id=402, type=get_attr];
+"403 features_1_1_attn_relative_coords_table" [id=403, type=get_attr];
+"404 features_1_1_attn_relative_position_index" [id=404, type=get_attr];
+"405 features_3_0_attn_relative_coords_table" [id=405, type=get_attr];
+"406 features_3_0_attn_relative_position_index" [id=406, type=get_attr];
+"407 features_3_1_attn_relative_coords_table" [id=407, type=get_attr];
+"408 features_3_1_attn_relative_position_index" [id=408, type=get_attr];
+"409 features_5_0_attn_relative_coords_table" [id=409, type=get_attr];
+"410 features_5_0_attn_relative_position_index" [id=410, type=get_attr];
+"411 features_5_1_attn_relative_coords_table" [id=411, type=get_attr];
+"412 features_5_1_attn_relative_position_index" [id=412, type=get_attr];
+"413 features_5_2_attn_relative_coords_table" [id=413, type=get_attr];
+"414 features_5_2_attn_relative_position_index" [id=414, type=get_attr];
+"415 features_5_3_attn_relative_coords_table" [id=415, type=get_attr];
+"416 features_5_3_attn_relative_position_index" [id=416, type=get_attr];
+"417 features_5_4_attn_relative_coords_table" [id=417, type=get_attr];
+"418 features_5_4_attn_relative_position_index" [id=418, type=get_attr];
+"419 features_5_5_attn_relative_coords_table" [id=419, type=get_attr];
+"420 features_5_5_attn_relative_position_index" [id=420, type=get_attr];
+"421 features_5_6_attn_relative_coords_table" [id=421, type=get_attr];
+"422 features_5_6_attn_relative_position_index" [id=422, type=get_attr];
+"423 features_5_7_attn_relative_coords_table" [id=423, type=get_attr];
+"424 features_5_7_attn_relative_position_index" [id=424, type=get_attr];
+"425 features_5_8_attn_relative_coords_table" [id=425, type=get_attr];
+"426 features_5_8_attn_relative_position_index" [id=426, type=get_attr];
+"427 features_5_9_attn_relative_coords_table" [id=427, type=get_attr];
+"428 features_5_9_attn_relative_position_index" [id=428, type=get_attr];
+"429 features_5_10_attn_relative_coords_table" [id=429, type=get_attr];
+"430 features_5_10_attn_relative_position_index" [id=430, type=get_attr];
+"431 features_5_11_attn_relative_coords_table" [id=431, type=get_attr];
+"432 features_5_11_attn_relative_position_index" [id=432, type=get_attr];
+"433 features_5_12_attn_relative_coords_table" [id=433, type=get_attr];
+"434 features_5_12_attn_relative_position_index" [id=434, type=get_attr];
+"435 features_5_13_attn_relative_coords_table" [id=435, type=get_attr];
+"436 features_5_13_attn_relative_position_index" [id=436, type=get_attr];
+"437 features_5_14_attn_relative_coords_table" [id=437, type=get_attr];
+"438 features_5_14_attn_relative_position_index" [id=438, type=get_attr];
+"439 features_5_15_attn_relative_coords_table" [id=439, type=get_attr];
+"440 features_5_15_attn_relative_position_index" [id=440, type=get_attr];
+"441 features_5_16_attn_relative_coords_table" [id=441, type=get_attr];
+"442 features_5_16_attn_relative_position_index" [id=442, type=get_attr];
+"443 features_5_17_attn_relative_coords_table" [id=443, type=get_attr];
+"444 features_5_17_attn_relative_position_index" [id=444, type=get_attr];
+"445 features_7_0_attn_relative_coords_table" [id=445, type=get_attr];
+"446 features_7_0_attn_relative_position_index" [id=446, type=get_attr];
+"447 features_7_1_attn_relative_coords_table" [id=447, type=get_attr];
+"448 features_7_1_attn_relative_position_index" [id=448, type=get_attr];
+"449 features_1_1_attn_lifted_tensor_0" [id=449, type=get_attr];
+"450 features_1_1_attn_lifted_tensor_1" [id=450, type=get_attr];
+"451 features_1_1_attn_lifted_tensor_2" [id=451, type=get_attr];
+"452 features_1_1_attn_lifted_tensor_3" [id=452, type=get_attr];
+"453 features_1_1_attn_lifted_tensor_4" [id=453, type=get_attr];
+"454 features_1_1_attn_lifted_tensor_5" [id=454, type=get_attr];
+"455 features_1_1_attn_lifted_tensor_6" [id=455, type=get_attr];
+"456 features_1_1_attn_lifted_tensor_7" [id=456, type=get_attr];
+"457 features_1_1_attn_lifted_tensor_8" [id=457, type=get_attr];
+"458 features_3_1_attn_lifted_tensor_9" [id=458, type=get_attr];
+"459 features_3_1_attn_lifted_tensor_10" [id=459, type=get_attr];
+"460 features_3_1_attn_lifted_tensor_11" [id=460, type=get_attr];
+"461 features_3_1_attn_lifted_tensor_12" [id=461, type=get_attr];
+"462 features_3_1_attn_lifted_tensor_13" [id=462, type=get_attr];
+"463 features_3_1_attn_lifted_tensor_14" [id=463, type=get_attr];
+"464 features_3_1_attn_lifted_tensor_15" [id=464, type=get_attr];
+"465 features_3_1_attn_lifted_tensor_16" [id=465, type=get_attr];
+"466 features_3_1_attn_lifted_tensor_17" [id=466, type=get_attr];
+"467 features_5_1_attn_lifted_tensor_18" [id=467, type=get_attr];
+"468 features_5_1_attn_lifted_tensor_19" [id=468, type=get_attr];
+"469 features_5_1_attn_lifted_tensor_20" [id=469, type=get_attr];
+"470 features_5_1_attn_lifted_tensor_21" [id=470, type=get_attr];
+"471 features_5_1_attn_lifted_tensor_22" [id=471, type=get_attr];
+"472 features_5_1_attn_lifted_tensor_23" [id=472, type=get_attr];
+"473 features_5_1_attn_lifted_tensor_24" [id=473, type=get_attr];
+"474 features_5_1_attn_lifted_tensor_25" [id=474, type=get_attr];
+"475 features_5_1_attn_lifted_tensor_26" [id=475, type=get_attr];
+"476 features_5_3_attn_lifted_tensor_27" [id=476, type=get_attr];
+"477 features_5_3_attn_lifted_tensor_28" [id=477, type=get_attr];
+"478 features_5_3_attn_lifted_tensor_29" [id=478, type=get_attr];
+"479 features_5_3_attn_lifted_tensor_30" [id=479, type=get_attr];
+"480 features_5_3_attn_lifted_tensor_31" [id=480, type=get_attr];
+"481 features_5_3_attn_lifted_tensor_32" [id=481, type=get_attr];
+"482 features_5_3_attn_lifted_tensor_33" [id=482, type=get_attr];
+"483 features_5_3_attn_lifted_tensor_34" [id=483, type=get_attr];
+"484 features_5_3_attn_lifted_tensor_35" [id=484, type=get_attr];
+"485 features_5_5_attn_lifted_tensor_36" [id=485, type=get_attr];
+"486 features_5_5_attn_lifted_tensor_37" [id=486, type=get_attr];
+"487 features_5_5_attn_lifted_tensor_38" [id=487, type=get_attr];
+"488 features_5_5_attn_lifted_tensor_39" [id=488, type=get_attr];
+"489 features_5_5_attn_lifted_tensor_40" [id=489, type=get_attr];
+"490 features_5_5_attn_lifted_tensor_41" [id=490, type=get_attr];
+"491 features_5_5_attn_lifted_tensor_42" [id=491, type=get_attr];
+"492 features_5_5_attn_lifted_tensor_43" [id=492, type=get_attr];
+"493 features_5_5_attn_lifted_tensor_44" [id=493, type=get_attr];
+"494 features_5_7_attn_lifted_tensor_45" [id=494, type=get_attr];
+"495 features_5_7_attn_lifted_tensor_46" [id=495, type=get_attr];
+"496 features_5_7_attn_lifted_tensor_47" [id=496, type=get_attr];
+"497 features_5_7_attn_lifted_tensor_48" [id=497, type=get_attr];
+"498 features_5_7_attn_lifted_tensor_49" [id=498, type=get_attr];
+"499 features_5_7_attn_lifted_tensor_50" [id=499, type=get_attr];
+"500 features_5_7_attn_lifted_tensor_51" [id=500, type=get_attr];
+"501 features_5_7_attn_lifted_tensor_52" [id=501, type=get_attr];
+"502 features_5_7_attn_lifted_tensor_53" [id=502, type=get_attr];
+"503 features_5_9_attn_lifted_tensor_54" [id=503, type=get_attr];
+"504 features_5_9_attn_lifted_tensor_55" [id=504, type=get_attr];
+"505 features_5_9_attn_lifted_tensor_56" [id=505, type=get_attr];
+"506 features_5_9_attn_lifted_tensor_57" [id=506, type=get_attr];
+"507 features_5_9_attn_lifted_tensor_58" [id=507, type=get_attr];
+"508 features_5_9_attn_lifted_tensor_59" [id=508, type=get_attr];
+"509 features_5_9_attn_lifted_tensor_60" [id=509, type=get_attr];
+"510 features_5_9_attn_lifted_tensor_61" [id=510, type=get_attr];
+"511 features_5_9_attn_lifted_tensor_62" [id=511, type=get_attr];
+"512 features_5_11_attn_lifted_tensor_63" [id=512, type=get_attr];
+"513 features_5_11_attn_lifted_tensor_64" [id=513, type=get_attr];
+"514 features_5_11_attn_lifted_tensor_65" [id=514, type=get_attr];
+"515 features_5_11_attn_lifted_tensor_66" [id=515, type=get_attr];
+"516 features_5_11_attn_lifted_tensor_67" [id=516, type=get_attr];
+"517 features_5_11_attn_lifted_tensor_68" [id=517, type=get_attr];
+"518 features_5_11_attn_lifted_tensor_69" [id=518, type=get_attr];
+"519 features_5_11_attn_lifted_tensor_70" [id=519, type=get_attr];
+"520 features_5_11_attn_lifted_tensor_71" [id=520, type=get_attr];
+"521 features_5_13_attn_lifted_tensor_72" [id=521, type=get_attr];
+"522 features_5_13_attn_lifted_tensor_73" [id=522, type=get_attr];
+"523 features_5_13_attn_lifted_tensor_74" [id=523, type=get_attr];
+"524 features_5_13_attn_lifted_tensor_75" [id=524, type=get_attr];
+"525 features_5_13_attn_lifted_tensor_76" [id=525, type=get_attr];
+"526 features_5_13_attn_lifted_tensor_77" [id=526, type=get_attr];
+"527 features_5_13_attn_lifted_tensor_78" [id=527, type=get_attr];
+"528 features_5_13_attn_lifted_tensor_79" [id=528, type=get_attr];
+"529 features_5_13_attn_lifted_tensor_80" [id=529, type=get_attr];
+"530 features_5_15_attn_lifted_tensor_81" [id=530, type=get_attr];
+"531 features_5_15_attn_lifted_tensor_82" [id=531, type=get_attr];
+"532 features_5_15_attn_lifted_tensor_83" [id=532, type=get_attr];
+"533 features_5_15_attn_lifted_tensor_84" [id=533, type=get_attr];
+"534 features_5_15_attn_lifted_tensor_85" [id=534, type=get_attr];
+"535 features_5_15_attn_lifted_tensor_86" [id=535, type=get_attr];
+"536 features_5_15_attn_lifted_tensor_87" [id=536, type=get_attr];
+"537 features_5_15_attn_lifted_tensor_88" [id=537, type=get_attr];
+"538 features_5_15_attn_lifted_tensor_89" [id=538, type=get_attr];
+"539 features_5_17_attn_lifted_tensor_90" [id=539, type=get_attr];
+"540 features_5_17_attn_lifted_tensor_91" [id=540, type=get_attr];
+"541 features_5_17_attn_lifted_tensor_92" [id=541, type=get_attr];
+"542 features_5_17_attn_lifted_tensor_93" [id=542, type=get_attr];
+"543 features_5_17_attn_lifted_tensor_94" [id=543, type=get_attr];
+"544 features_5_17_attn_lifted_tensor_95" [id=544, type=get_attr];
+"545 features_5_17_attn_lifted_tensor_96" [id=545, type=get_attr];
+"546 features_5_17_attn_lifted_tensor_97" [id=546, type=get_attr];
+"547 features_5_17_attn_lifted_tensor_98" [id=547, type=get_attr];
+"548 x" [id=548, type=input];
+"549 conv2d" [id=549, type=conv2d];
+"550 permute" [id=550, type=permute];
+"551 layer_norm" [id=551, type=layer_norm];
+"552 linear" [id=552, type=linear];
+"553 relu_" [id=553, type=relu_];
+"554 linear_1" [id=554, type=linear];
+"555 view" [id=555, type=view];
+"556 index" [id=556, type=index];
+"557 view_1" [id=557, type=view];
+"558 permute_1" [id=558, type=permute];
+"559 contiguous" [id=559, type=contiguous];
+"560 unsqueeze" [id=560, type=unsqueeze];
+"561 sigmoid" [id=561, type=sigmoid];
+"562 mul" [id=562, type=mul];
+"563 pad" [id=563, type=pad];
+"564 view_2" [id=564, type=view];
+"565 permute_2" [id=565, type=permute];
+"566 reshape" [id=566, type=reshape];
+"567 clone" [id=567, type=clone];
+"568 slice_1" [id=568, type=slice];
+"569 zero_" [id=569, type=zero_];
+"570 linear_2" [id=570, type=linear];
+"571 reshape_1" [id=571, type=reshape];
+"572 permute_3" [id=572, type=permute];
+"573 select" [id=573, type=select];
+"574 select_1" [id=574, type=select];
+"575 select_2" [id=575, type=select];
+"576 linalg_vector_norm" [id=576, type=linalg_vector_norm];
+"577 clamp_min" [id=577, type=clamp_min];
+"578 expand_as" [id=578, type=expand_as];
+"579 div" [id=579, type=div];
+"580 linalg_vector_norm_1" [id=580, type=linalg_vector_norm];
+"581 clamp_min_1" [id=581, type=clamp_min];
+"582 expand_as_1" [id=582, type=expand_as];
+"583 div_1" [id=583, type=div];
+"584 transpose" [id=584, type=transpose];
+"585 matmul" [id=585, type=matmul];
+"586 clamp" [id=586, type=clamp];
+"587 exp" [id=587, type=exp];
+"588 mul_1" [id=588, type=mul];
+"589 add" [id=589, type=add];
+"590 softmax" [id=590, type=softmax];
+"591 dropout" [id=591, type=dropout];
+"592 matmul_1" [id=592, type=matmul];
+"593 transpose_1" [id=593, type=transpose];
+"594 reshape_2" [id=594, type=reshape];
+"595 linear_3" [id=595, type=linear];
+"596 dropout_1" [id=596, type=dropout];
+"597 view_3" [id=597, type=view];
+"598 permute_4" [id=598, type=permute];
+"599 reshape_3" [id=599, type=reshape];
+"600 slice_2" [id=600, type=slice];
+"601 slice_3" [id=601, type=slice];
+"602 layer_norm_1" [id=602, type=layer_norm];
+"603 add_1" [id=603, type=add];
+"604 linear_4" [id=604, type=linear];
+"605 gelu" [id=605, type=gelu];
+"606 dropout_2" [id=606, type=dropout];
+"607 linear_5" [id=607, type=linear];
+"608 dropout_3" [id=608, type=dropout];
+"609 layer_norm_2" [id=609, type=layer_norm];
+"610 add_2" [id=610, type=add];
+"611 linear_6" [id=611, type=linear];
+"612 relu__1" [id=612, type=relu_];
+"613 linear_7" [id=613, type=linear];
+"614 view_4" [id=614, type=view];
+"615 index_1" [id=615, type=index];
+"616 view_5" [id=616, type=view];
+"617 permute_5" [id=617, type=permute];
+"618 contiguous_1" [id=618, type=contiguous];
+"619 unsqueeze_1" [id=619, type=unsqueeze];
+"620 sigmoid_1" [id=620, type=sigmoid];
+"621 mul_2" [id=621, type=mul];
+"622 pad_1" [id=622, type=pad];
+"623 roll" [id=623, type=roll];
+"624 view_6" [id=624, type=view];
+"625 permute_6" [id=625, type=permute];
+"626 reshape_4" [id=626, type=reshape];
+"627 clone_1" [id=627, type=clone];
+"628 slice_4" [id=628, type=slice];
+"629 zero__1" [id=629, type=zero_];
+"630 linear_8" [id=630, type=linear];
+"631 reshape_5" [id=631, type=reshape];
+"632 permute_7" [id=632, type=permute];
+"633 select_3" [id=633, type=select];
+"634 select_4" [id=634, type=select];
+"635 select_5" [id=635, type=select];
+"636 linalg_vector_norm_2" [id=636, type=linalg_vector_norm];
+"637 clamp_min_2" [id=637, type=clamp_min];
+"638 expand_as_2" [id=638, type=expand_as];
+"639 div_2" [id=639, type=div];
+"640 linalg_vector_norm_3" [id=640, type=linalg_vector_norm];
+"641 clamp_min_3" [id=641, type=clamp_min];
+"642 expand_as_3" [id=642, type=expand_as];
+"643 div_3" [id=643, type=div];
+"644 transpose_2" [id=644, type=transpose];
+"645 matmul_2" [id=645, type=matmul];
+"646 clamp_1" [id=646, type=clamp];
+"647 exp_1" [id=647, type=exp];
+"648 mul_3" [id=648, type=mul];
+"649 add_3" [id=649, type=add];
+"650 new_zeros" [id=650, type=new_zeros];
+"651 lift_fresh_copy" [id=651, type=lift_fresh_copy];
+"652 slice_5" [id=652, type=slice];
+"653 slice_6" [id=653, type=slice];
+"654 fill_" [id=654, type=fill_];
+"655 lift_fresh_copy_1" [id=655, type=lift_fresh_copy];
+"656 slice_7" [id=656, type=slice];
+"657 slice_8" [id=657, type=slice];
+"658 fill__1" [id=658, type=fill_];
+"659 lift_fresh_copy_2" [id=659, type=lift_fresh_copy];
+"660 slice_9" [id=660, type=slice];
+"661 slice_10" [id=661, type=slice];
+"662 fill__2" [id=662, type=fill_];
+"663 lift_fresh_copy_3" [id=663, type=lift_fresh_copy];
+"664 slice_11" [id=664, type=slice];
+"665 slice_12" [id=665, type=slice];
+"666 fill__3" [id=666, type=fill_];
+"667 lift_fresh_copy_4" [id=667, type=lift_fresh_copy];
+"668 slice_13" [id=668, type=slice];
+"669 slice_14" [id=669, type=slice];
+"670 fill__4" [id=670, type=fill_];
+"671 lift_fresh_copy_5" [id=671, type=lift_fresh_copy];
+"672 slice_15" [id=672, type=slice];
+"673 slice_16" [id=673, type=slice];
+"674 fill__5" [id=674, type=fill_];
+"675 lift_fresh_copy_6" [id=675, type=lift_fresh_copy];
+"676 slice_17" [id=676, type=slice];
+"677 slice_18" [id=677, type=slice];
+"678 fill__6" [id=678, type=fill_];
+"679 lift_fresh_copy_7" [id=679, type=lift_fresh_copy];
+"680 slice_19" [id=680, type=slice];
+"681 slice_20" [id=681, type=slice];
+"682 fill__7" [id=682, type=fill_];
+"683 lift_fresh_copy_8" [id=683, type=lift_fresh_copy];
+"684 slice_21" [id=684, type=slice];
+"685 slice_22" [id=685, type=slice];
+"686 fill__8" [id=686, type=fill_];
+"687 view_7" [id=687, type=view];
+"688 permute_8" [id=688, type=permute];
+"689 reshape_6" [id=689, type=reshape];
+"690 unsqueeze_2" [id=690, type=unsqueeze];
+"691 unsqueeze_3" [id=691, type=unsqueeze];
+"692 sub" [id=692, type=sub];
+"693 ne" [id=693, type=ne];
+"694 masked_fill" [id=694, type=masked_fill];
+"695 eq" [id=695, type=eq];
+"696 masked_fill_1" [id=696, type=masked_fill];
+"697 view_8" [id=697, type=view];
+"698 unsqueeze_4" [id=698, type=unsqueeze];
+"699 unsqueeze_5" [id=699, type=unsqueeze];
+"700 add_4" [id=700, type=add];
+"701 view_9" [id=701, type=view];
+"702 softmax_1" [id=702, type=softmax];
+"703 dropout_4" [id=703, type=dropout];
+"704 matmul_3" [id=704, type=matmul];
+"705 transpose_3" [id=705, type=transpose];
+"706 reshape_7" [id=706, type=reshape];
+"707 linear_9" [id=707, type=linear];
+"708 dropout_5" [id=708, type=dropout];
+"709 view_10" [id=709, type=view];
+"710 permute_9" [id=710, type=permute];
+"711 reshape_8" [id=711, type=reshape];
+"712 roll_1" [id=712, type=roll];
+"713 slice_23" [id=713, type=slice];
+"714 slice_24" [id=714, type=slice];
+"715 layer_norm_3" [id=715, type=layer_norm];
+"716 add_5" [id=716, type=add];
+"717 linear_10" [id=717, type=linear];
+"718 gelu_1" [id=718, type=gelu];
+"719 dropout_6" [id=719, type=dropout];
+"720 linear_11" [id=720, type=linear];
+"721 dropout_7" [id=721, type=dropout];
+"722 layer_norm_4" [id=722, type=layer_norm];
+"723 add_6" [id=723, type=add];
+"724 pad_2" [id=724, type=pad];
+"725 slice_25" [id=725, type=slice];
+"726 slice_26" [id=726, type=slice];
+"727 slice_27" [id=727, type=slice];
+"728 slice_28" [id=728, type=slice];
+"729 slice_29" [id=729, type=slice];
+"730 slice_30" [id=730, type=slice];
+"731 slice_31" [id=731, type=slice];
+"732 slice_32" [id=732, type=slice];
+"733 slice_33" [id=733, type=slice];
+"734 slice_34" [id=734, type=slice];
+"735 slice_35" [id=735, type=slice];
+"736 slice_36" [id=736, type=slice];
+"737 cat" [id=737, type=cat];
+"738 linear_12" [id=738, type=linear];
+"739 layer_norm_5" [id=739, type=layer_norm];
+"740 linear_13" [id=740, type=linear];
+"741 relu__2" [id=741, type=relu_];
+"742 linear_14" [id=742, type=linear];
+"743 view_11" [id=743, type=view];
+"744 index_2" [id=744, type=index];
+"745 view_12" [id=745, type=view];
+"746 permute_10" [id=746, type=permute];
+"747 contiguous_2" [id=747, type=contiguous];
+"748 unsqueeze_6" [id=748, type=unsqueeze];
+"749 sigmoid_2" [id=749, type=sigmoid];
+"750 mul_4" [id=750, type=mul];
+"751 pad_3" [id=751, type=pad];
+"752 view_13" [id=752, type=view];
+"753 permute_11" [id=753, type=permute];
+"754 reshape_9" [id=754, type=reshape];
+"755 clone_2" [id=755, type=clone];
+"756 slice_37" [id=756, type=slice];
+"757 zero__2" [id=757, type=zero_];
+"758 linear_15" [id=758, type=linear];
+"759 reshape_10" [id=759, type=reshape];
+"760 permute_12" [id=760, type=permute];
+"761 select_6" [id=761, type=select];
+"762 select_7" [id=762, type=select];
+"763 select_8" [id=763, type=select];
+"764 linalg_vector_norm_4" [id=764, type=linalg_vector_norm];
+"765 clamp_min_4" [id=765, type=clamp_min];
+"766 expand_as_4" [id=766, type=expand_as];
+"767 div_4" [id=767, type=div];
+"768 linalg_vector_norm_5" [id=768, type=linalg_vector_norm];
+"769 clamp_min_5" [id=769, type=clamp_min];
+"770 expand_as_5" [id=770, type=expand_as];
+"771 div_5" [id=771, type=div];
+"772 transpose_4" [id=772, type=transpose];
+"773 matmul_4" [id=773, type=matmul];
+"774 clamp_2" [id=774, type=clamp];
+"775 exp_2" [id=775, type=exp];
+"776 mul_5" [id=776, type=mul];
+"777 add_7" [id=777, type=add];
+"778 softmax_2" [id=778, type=softmax];
+"779 dropout_8" [id=779, type=dropout];
+"780 matmul_5" [id=780, type=matmul];
+"781 transpose_5" [id=781, type=transpose];
+"782 reshape_11" [id=782, type=reshape];
+"783 linear_16" [id=783, type=linear];
+"784 dropout_9" [id=784, type=dropout];
+"785 view_14" [id=785, type=view];
+"786 permute_13" [id=786, type=permute];
+"787 reshape_12" [id=787, type=reshape];
+"788 slice_38" [id=788, type=slice];
+"789 slice_39" [id=789, type=slice];
+"790 slice_40" [id=790, type=slice];
+"791 slice_41" [id=791, type=slice];
+"792 contiguous_3" [id=792, type=contiguous];
+"793 layer_norm_6" [id=793, type=layer_norm];
+"794 add_8" [id=794, type=add];
+"795 linear_17" [id=795, type=linear];
+"796 gelu_2" [id=796, type=gelu];
+"797 dropout_10" [id=797, type=dropout];
+"798 linear_18" [id=798, type=linear];
+"799 dropout_11" [id=799, type=dropout];
+"800 layer_norm_7" [id=800, type=layer_norm];
+"801 add_9" [id=801, type=add];
+"802 linear_19" [id=802, type=linear];
+"803 relu__3" [id=803, type=relu_];
+"804 linear_20" [id=804, type=linear];
+"805 view_15" [id=805, type=view];
+"806 index_3" [id=806, type=index];
+"807 view_16" [id=807, type=view];
+"808 permute_14" [id=808, type=permute];
+"809 contiguous_4" [id=809, type=contiguous];
+"810 unsqueeze_7" [id=810, type=unsqueeze];
+"811 sigmoid_3" [id=811, type=sigmoid];
+"812 mul_6" [id=812, type=mul];
+"813 pad_4" [id=813, type=pad];
+"814 roll_2" [id=814, type=roll];
+"815 view_17" [id=815, type=view];
+"816 permute_15" [id=816, type=permute];
+"817 reshape_13" [id=817, type=reshape];
+"818 clone_3" [id=818, type=clone];
+"819 slice_42" [id=819, type=slice];
+"820 zero__3" [id=820, type=zero_];
+"821 linear_21" [id=821, type=linear];
+"822 reshape_14" [id=822, type=reshape];
+"823 permute_16" [id=823, type=permute];
+"824 select_9" [id=824, type=select];
+"825 select_10" [id=825, type=select];
+"826 select_11" [id=826, type=select];
+"827 linalg_vector_norm_6" [id=827, type=linalg_vector_norm];
+"828 clamp_min_6" [id=828, type=clamp_min];
+"829 expand_as_6" [id=829, type=expand_as];
+"830 div_6" [id=830, type=div];
+"831 linalg_vector_norm_7" [id=831, type=linalg_vector_norm];
+"832 clamp_min_7" [id=832, type=clamp_min];
+"833 expand_as_7" [id=833, type=expand_as];
+"834 div_7" [id=834, type=div];
+"835 transpose_6" [id=835, type=transpose];
+"836 matmul_6" [id=836, type=matmul];
+"837 clamp_3" [id=837, type=clamp];
+"838 exp_3" [id=838, type=exp];
+"839 mul_7" [id=839, type=mul];
+"840 add_10" [id=840, type=add];
+"841 new_zeros_1" [id=841, type=new_zeros];
+"842 lift_fresh_copy_9" [id=842, type=lift_fresh_copy];
+"843 slice_43" [id=843, type=slice];
+"844 slice_44" [id=844, type=slice];
+"845 fill__9" [id=845, type=fill_];
+"846 lift_fresh_copy_10" [id=846, type=lift_fresh_copy];
+"847 slice_45" [id=847, type=slice];
+"848 slice_46" [id=848, type=slice];
+"849 fill__10" [id=849, type=fill_];
+"850 lift_fresh_copy_11" [id=850, type=lift_fresh_copy];
+"851 slice_47" [id=851, type=slice];
+"852 slice_48" [id=852, type=slice];
+"853 fill__11" [id=853, type=fill_];
+"854 lift_fresh_copy_12" [id=854, type=lift_fresh_copy];
+"855 slice_49" [id=855, type=slice];
+"856 slice_50" [id=856, type=slice];
+"857 fill__12" [id=857, type=fill_];
+"858 lift_fresh_copy_13" [id=858, type=lift_fresh_copy];
+"859 slice_51" [id=859, type=slice];
+"860 slice_52" [id=860, type=slice];
+"861 fill__13" [id=861, type=fill_];
+"862 lift_fresh_copy_14" [id=862, type=lift_fresh_copy];
+"863 slice_53" [id=863, type=slice];
+"864 slice_54" [id=864, type=slice];
+"865 fill__14" [id=865, type=fill_];
+"866 lift_fresh_copy_15" [id=866, type=lift_fresh_copy];
+"867 slice_55" [id=867, type=slice];
+"868 slice_56" [id=868, type=slice];
+"869 fill__15" [id=869, type=fill_];
+"870 lift_fresh_copy_16" [id=870, type=lift_fresh_copy];
+"871 slice_57" [id=871, type=slice];
+"872 slice_58" [id=872, type=slice];
+"873 fill__16" [id=873, type=fill_];
+"874 lift_fresh_copy_17" [id=874, type=lift_fresh_copy];
+"875 slice_59" [id=875, type=slice];
+"876 slice_60" [id=876, type=slice];
+"877 fill__17" [id=877, type=fill_];
+"878 view_18" [id=878, type=view];
+"879 permute_17" [id=879, type=permute];
+"880 reshape_15" [id=880, type=reshape];
+"881 unsqueeze_8" [id=881, type=unsqueeze];
+"882 unsqueeze_9" [id=882, type=unsqueeze];
+"883 sub_1" [id=883, type=sub];
+"884 ne_1" [id=884, type=ne];
+"885 masked_fill_2" [id=885, type=masked_fill];
+"886 eq_1" [id=886, type=eq];
+"887 masked_fill_3" [id=887, type=masked_fill];
+"888 view_19" [id=888, type=view];
+"889 unsqueeze_10" [id=889, type=unsqueeze];
+"890 unsqueeze_11" [id=890, type=unsqueeze];
+"891 add_11" [id=891, type=add];
+"892 view_20" [id=892, type=view];
+"893 softmax_3" [id=893, type=softmax];
+"894 dropout_12" [id=894, type=dropout];
+"895 matmul_7" [id=895, type=matmul];
+"896 transpose_7" [id=896, type=transpose];
+"897 reshape_16" [id=897, type=reshape];
+"898 linear_22" [id=898, type=linear];
+"899 dropout_13" [id=899, type=dropout];
+"900 view_21" [id=900, type=view];
+"901 permute_18" [id=901, type=permute];
+"902 reshape_17" [id=902, type=reshape];
+"903 roll_3" [id=903, type=roll];
+"904 slice_61" [id=904, type=slice];
+"905 slice_62" [id=905, type=slice];
+"906 slice_63" [id=906, type=slice];
+"907 slice_64" [id=907, type=slice];
+"908 contiguous_5" [id=908, type=contiguous];
+"909 layer_norm_8" [id=909, type=layer_norm];
+"910 add_12" [id=910, type=add];
+"911 linear_23" [id=911, type=linear];
+"912 gelu_3" [id=912, type=gelu];
+"913 dropout_14" [id=913, type=dropout];
+"914 linear_24" [id=914, type=linear];
+"915 dropout_15" [id=915, type=dropout];
+"916 layer_norm_9" [id=916, type=layer_norm];
+"917 add_13" [id=917, type=add];
+"918 pad_5" [id=918, type=pad];
+"919 slice_65" [id=919, type=slice];
+"920 slice_66" [id=920, type=slice];
+"921 slice_67" [id=921, type=slice];
+"922 slice_68" [id=922, type=slice];
+"923 slice_69" [id=923, type=slice];
+"924 slice_70" [id=924, type=slice];
+"925 slice_71" [id=925, type=slice];
+"926 slice_72" [id=926, type=slice];
+"927 slice_73" [id=927, type=slice];
+"928 slice_74" [id=928, type=slice];
+"929 slice_75" [id=929, type=slice];
+"930 slice_76" [id=930, type=slice];
+"931 cat_1" [id=931, type=cat];
+"932 linear_25" [id=932, type=linear];
+"933 layer_norm_10" [id=933, type=layer_norm];
+"934 linear_26" [id=934, type=linear];
+"935 relu__4" [id=935, type=relu_];
+"936 linear_27" [id=936, type=linear];
+"937 view_22" [id=937, type=view];
+"938 index_4" [id=938, type=index];
+"939 view_23" [id=939, type=view];
+"940 permute_19" [id=940, type=permute];
+"941 contiguous_6" [id=941, type=contiguous];
+"942 unsqueeze_12" [id=942, type=unsqueeze];
+"943 sigmoid_4" [id=943, type=sigmoid];
+"944 mul_8" [id=944, type=mul];
+"945 pad_6" [id=945, type=pad];
+"946 view_24" [id=946, type=view];
+"947 permute_20" [id=947, type=permute];
+"948 reshape_18" [id=948, type=reshape];
+"949 clone_4" [id=949, type=clone];
+"950 slice_77" [id=950, type=slice];
+"951 zero__4" [id=951, type=zero_];
+"952 linear_28" [id=952, type=linear];
+"953 reshape_19" [id=953, type=reshape];
+"954 permute_21" [id=954, type=permute];
+"955 select_12" [id=955, type=select];
+"956 select_13" [id=956, type=select];
+"957 select_14" [id=957, type=select];
+"958 linalg_vector_norm_8" [id=958, type=linalg_vector_norm];
+"959 clamp_min_8" [id=959, type=clamp_min];
+"960 expand_as_8" [id=960, type=expand_as];
+"961 div_8" [id=961, type=div];
+"962 linalg_vector_norm_9" [id=962, type=linalg_vector_norm];
+"963 clamp_min_9" [id=963, type=clamp_min];
+"964 expand_as_9" [id=964, type=expand_as];
+"965 div_9" [id=965, type=div];
+"966 transpose_8" [id=966, type=transpose];
+"967 matmul_8" [id=967, type=matmul];
+"968 clamp_4" [id=968, type=clamp];
+"969 exp_4" [id=969, type=exp];
+"970 mul_9" [id=970, type=mul];
+"971 add_14" [id=971, type=add];
+"972 softmax_4" [id=972, type=softmax];
+"973 dropout_16" [id=973, type=dropout];
+"974 matmul_9" [id=974, type=matmul];
+"975 transpose_9" [id=975, type=transpose];
+"976 reshape_20" [id=976, type=reshape];
+"977 linear_29" [id=977, type=linear];
+"978 dropout_17" [id=978, type=dropout];
+"979 view_25" [id=979, type=view];
+"980 permute_22" [id=980, type=permute];
+"981 reshape_21" [id=981, type=reshape];
+"982 slice_78" [id=982, type=slice];
+"983 slice_79" [id=983, type=slice];
+"984 slice_80" [id=984, type=slice];
+"985 slice_81" [id=985, type=slice];
+"986 contiguous_7" [id=986, type=contiguous];
+"987 layer_norm_11" [id=987, type=layer_norm];
+"988 add_15" [id=988, type=add];
+"989 linear_30" [id=989, type=linear];
+"990 gelu_4" [id=990, type=gelu];
+"991 dropout_18" [id=991, type=dropout];
+"992 linear_31" [id=992, type=linear];
+"993 dropout_19" [id=993, type=dropout];
+"994 layer_norm_12" [id=994, type=layer_norm];
+"995 add_16" [id=995, type=add];
+"996 linear_32" [id=996, type=linear];
+"997 relu__5" [id=997, type=relu_];
+"998 linear_33" [id=998, type=linear];
+"999 view_26" [id=999, type=view];
+"1000 index_5" [id=1000, type=index];
+"1001 view_27" [id=1001, type=view];
+"1002 permute_23" [id=1002, type=permute];
+"1003 contiguous_8" [id=1003, type=contiguous];
+"1004 unsqueeze_13" [id=1004, type=unsqueeze];
+"1005 sigmoid_5" [id=1005, type=sigmoid];
+"1006 mul_10" [id=1006, type=mul];
+"1007 pad_7" [id=1007, type=pad];
+"1008 roll_4" [id=1008, type=roll];
+"1009 view_28" [id=1009, type=view];
+"1010 permute_24" [id=1010, type=permute];
+"1011 reshape_22" [id=1011, type=reshape];
+"1012 clone_5" [id=1012, type=clone];
+"1013 slice_82" [id=1013, type=slice];
+"1014 zero__5" [id=1014, type=zero_];
+"1015 linear_34" [id=1015, type=linear];
+"1016 reshape_23" [id=1016, type=reshape];
+"1017 permute_25" [id=1017, type=permute];
+"1018 select_15" [id=1018, type=select];
+"1019 select_16" [id=1019, type=select];
+"1020 select_17" [id=1020, type=select];
+"1021 linalg_vector_norm_10" [id=1021, type=linalg_vector_norm];
+"1022 clamp_min_10" [id=1022, type=clamp_min];
+"1023 expand_as_10" [id=1023, type=expand_as];
+"1024 div_10" [id=1024, type=div];
+"1025 linalg_vector_norm_11" [id=1025, type=linalg_vector_norm];
+"1026 clamp_min_11" [id=1026, type=clamp_min];
+"1027 expand_as_11" [id=1027, type=expand_as];
+"1028 div_11" [id=1028, type=div];
+"1029 transpose_10" [id=1029, type=transpose];
+"1030 matmul_10" [id=1030, type=matmul];
+"1031 clamp_5" [id=1031, type=clamp];
+"1032 exp_5" [id=1032, type=exp];
+"1033 mul_11" [id=1033, type=mul];
+"1034 add_17" [id=1034, type=add];
+"1035 new_zeros_2" [id=1035, type=new_zeros];
+"1036 lift_fresh_copy_18" [id=1036, type=lift_fresh_copy];
+"1037 slice_83" [id=1037, type=slice];
+"1038 slice_84" [id=1038, type=slice];
+"1039 fill__18" [id=1039, type=fill_];
+"1040 lift_fresh_copy_19" [id=1040, type=lift_fresh_copy];
+"1041 slice_85" [id=1041, type=slice];
+"1042 slice_86" [id=1042, type=slice];
+"1043 fill__19" [id=1043, type=fill_];
+"1044 lift_fresh_copy_20" [id=1044, type=lift_fresh_copy];
+"1045 slice_87" [id=1045, type=slice];
+"1046 slice_88" [id=1046, type=slice];
+"1047 fill__20" [id=1047, type=fill_];
+"1048 lift_fresh_copy_21" [id=1048, type=lift_fresh_copy];
+"1049 slice_89" [id=1049, type=slice];
+"1050 slice_90" [id=1050, type=slice];
+"1051 fill__21" [id=1051, type=fill_];
+"1052 lift_fresh_copy_22" [id=1052, type=lift_fresh_copy];
+"1053 slice_91" [id=1053, type=slice];
+"1054 slice_92" [id=1054, type=slice];
+"1055 fill__22" [id=1055, type=fill_];
+"1056 lift_fresh_copy_23" [id=1056, type=lift_fresh_copy];
+"1057 slice_93" [id=1057, type=slice];
+"1058 slice_94" [id=1058, type=slice];
+"1059 fill__23" [id=1059, type=fill_];
+"1060 lift_fresh_copy_24" [id=1060, type=lift_fresh_copy];
+"1061 slice_95" [id=1061, type=slice];
+"1062 slice_96" [id=1062, type=slice];
+"1063 fill__24" [id=1063, type=fill_];
+"1064 lift_fresh_copy_25" [id=1064, type=lift_fresh_copy];
+"1065 slice_97" [id=1065, type=slice];
+"1066 slice_98" [id=1066, type=slice];
+"1067 fill__25" [id=1067, type=fill_];
+"1068 lift_fresh_copy_26" [id=1068, type=lift_fresh_copy];
+"1069 slice_99" [id=1069, type=slice];
+"1070 slice_100" [id=1070, type=slice];
+"1071 fill__26" [id=1071, type=fill_];
+"1072 view_29" [id=1072, type=view];
+"1073 permute_26" [id=1073, type=permute];
+"1074 reshape_24" [id=1074, type=reshape];
+"1075 unsqueeze_14" [id=1075, type=unsqueeze];
+"1076 unsqueeze_15" [id=1076, type=unsqueeze];
+"1077 sub_2" [id=1077, type=sub];
+"1078 ne_2" [id=1078, type=ne];
+"1079 masked_fill_4" [id=1079, type=masked_fill];
+"1080 eq_2" [id=1080, type=eq];
+"1081 masked_fill_5" [id=1081, type=masked_fill];
+"1082 view_30" [id=1082, type=view];
+"1083 unsqueeze_16" [id=1083, type=unsqueeze];
+"1084 unsqueeze_17" [id=1084, type=unsqueeze];
+"1085 add_18" [id=1085, type=add];
+"1086 view_31" [id=1086, type=view];
+"1087 softmax_5" [id=1087, type=softmax];
+"1088 dropout_20" [id=1088, type=dropout];
+"1089 matmul_11" [id=1089, type=matmul];
+"1090 transpose_11" [id=1090, type=transpose];
+"1091 reshape_25" [id=1091, type=reshape];
+"1092 linear_35" [id=1092, type=linear];
+"1093 dropout_21" [id=1093, type=dropout];
+"1094 view_32" [id=1094, type=view];
+"1095 permute_27" [id=1095, type=permute];
+"1096 reshape_26" [id=1096, type=reshape];
+"1097 roll_5" [id=1097, type=roll];
+"1098 slice_101" [id=1098, type=slice];
+"1099 slice_102" [id=1099, type=slice];
+"1100 slice_103" [id=1100, type=slice];
+"1101 slice_104" [id=1101, type=slice];
+"1102 contiguous_9" [id=1102, type=contiguous];
+"1103 layer_norm_13" [id=1103, type=layer_norm];
+"1104 add_19" [id=1104, type=add];
+"1105 linear_36" [id=1105, type=linear];
+"1106 gelu_5" [id=1106, type=gelu];
+"1107 dropout_22" [id=1107, type=dropout];
+"1108 linear_37" [id=1108, type=linear];
+"1109 dropout_23" [id=1109, type=dropout];
+"1110 layer_norm_14" [id=1110, type=layer_norm];
+"1111 add_20" [id=1111, type=add];
+"1112 linear_38" [id=1112, type=linear];
+"1113 relu__6" [id=1113, type=relu_];
+"1114 linear_39" [id=1114, type=linear];
+"1115 view_33" [id=1115, type=view];
+"1116 index_6" [id=1116, type=index];
+"1117 view_34" [id=1117, type=view];
+"1118 permute_28" [id=1118, type=permute];
+"1119 contiguous_10" [id=1119, type=contiguous];
+"1120 unsqueeze_18" [id=1120, type=unsqueeze];
+"1121 sigmoid_6" [id=1121, type=sigmoid];
+"1122 mul_12" [id=1122, type=mul];
+"1123 pad_8" [id=1123, type=pad];
+"1124 view_35" [id=1124, type=view];
+"1125 permute_29" [id=1125, type=permute];
+"1126 reshape_27" [id=1126, type=reshape];
+"1127 clone_6" [id=1127, type=clone];
+"1128 slice_105" [id=1128, type=slice];
+"1129 zero__6" [id=1129, type=zero_];
+"1130 linear_40" [id=1130, type=linear];
+"1131 reshape_28" [id=1131, type=reshape];
+"1132 permute_30" [id=1132, type=permute];
+"1133 select_18" [id=1133, type=select];
+"1134 select_19" [id=1134, type=select];
+"1135 select_20" [id=1135, type=select];
+"1136 linalg_vector_norm_12" [id=1136, type=linalg_vector_norm];
+"1137 clamp_min_12" [id=1137, type=clamp_min];
+"1138 expand_as_12" [id=1138, type=expand_as];
+"1139 div_12" [id=1139, type=div];
+"1140 linalg_vector_norm_13" [id=1140, type=linalg_vector_norm];
+"1141 clamp_min_13" [id=1141, type=clamp_min];
+"1142 expand_as_13" [id=1142, type=expand_as];
+"1143 div_13" [id=1143, type=div];
+"1144 transpose_12" [id=1144, type=transpose];
+"1145 matmul_12" [id=1145, type=matmul];
+"1146 clamp_6" [id=1146, type=clamp];
+"1147 exp_6" [id=1147, type=exp];
+"1148 mul_13" [id=1148, type=mul];
+"1149 add_21" [id=1149, type=add];
+"1150 softmax_6" [id=1150, type=softmax];
+"1151 dropout_24" [id=1151, type=dropout];
+"1152 matmul_13" [id=1152, type=matmul];
+"1153 transpose_13" [id=1153, type=transpose];
+"1154 reshape_29" [id=1154, type=reshape];
+"1155 linear_41" [id=1155, type=linear];
+"1156 dropout_25" [id=1156, type=dropout];
+"1157 view_36" [id=1157, type=view];
+"1158 permute_31" [id=1158, type=permute];
+"1159 reshape_30" [id=1159, type=reshape];
+"1160 slice_106" [id=1160, type=slice];
+"1161 slice_107" [id=1161, type=slice];
+"1162 slice_108" [id=1162, type=slice];
+"1163 slice_109" [id=1163, type=slice];
+"1164 contiguous_11" [id=1164, type=contiguous];
+"1165 layer_norm_15" [id=1165, type=layer_norm];
+"1166 add_22" [id=1166, type=add];
+"1167 linear_42" [id=1167, type=linear];
+"1168 gelu_6" [id=1168, type=gelu];
+"1169 dropout_26" [id=1169, type=dropout];
+"1170 linear_43" [id=1170, type=linear];
+"1171 dropout_27" [id=1171, type=dropout];
+"1172 layer_norm_16" [id=1172, type=layer_norm];
+"1173 add_23" [id=1173, type=add];
+"1174 linear_44" [id=1174, type=linear];
+"1175 relu__7" [id=1175, type=relu_];
+"1176 linear_45" [id=1176, type=linear];
+"1177 view_37" [id=1177, type=view];
+"1178 index_7" [id=1178, type=index];
+"1179 view_38" [id=1179, type=view];
+"1180 permute_32" [id=1180, type=permute];
+"1181 contiguous_12" [id=1181, type=contiguous];
+"1182 unsqueeze_19" [id=1182, type=unsqueeze];
+"1183 sigmoid_7" [id=1183, type=sigmoid];
+"1184 mul_14" [id=1184, type=mul];
+"1185 pad_9" [id=1185, type=pad];
+"1186 roll_6" [id=1186, type=roll];
+"1187 view_39" [id=1187, type=view];
+"1188 permute_33" [id=1188, type=permute];
+"1189 reshape_31" [id=1189, type=reshape];
+"1190 clone_7" [id=1190, type=clone];
+"1191 slice_110" [id=1191, type=slice];
+"1192 zero__7" [id=1192, type=zero_];
+"1193 linear_46" [id=1193, type=linear];
+"1194 reshape_32" [id=1194, type=reshape];
+"1195 permute_34" [id=1195, type=permute];
+"1196 select_21" [id=1196, type=select];
+"1197 select_22" [id=1197, type=select];
+"1198 select_23" [id=1198, type=select];
+"1199 linalg_vector_norm_14" [id=1199, type=linalg_vector_norm];
+"1200 clamp_min_14" [id=1200, type=clamp_min];
+"1201 expand_as_14" [id=1201, type=expand_as];
+"1202 div_14" [id=1202, type=div];
+"1203 linalg_vector_norm_15" [id=1203, type=linalg_vector_norm];
+"1204 clamp_min_15" [id=1204, type=clamp_min];
+"1205 expand_as_15" [id=1205, type=expand_as];
+"1206 div_15" [id=1206, type=div];
+"1207 transpose_14" [id=1207, type=transpose];
+"1208 matmul_14" [id=1208, type=matmul];
+"1209 clamp_7" [id=1209, type=clamp];
+"1210 exp_7" [id=1210, type=exp];
+"1211 mul_15" [id=1211, type=mul];
+"1212 add_24" [id=1212, type=add];
+"1213 new_zeros_3" [id=1213, type=new_zeros];
+"1214 lift_fresh_copy_27" [id=1214, type=lift_fresh_copy];
+"1215 slice_111" [id=1215, type=slice];
+"1216 slice_112" [id=1216, type=slice];
+"1217 fill__27" [id=1217, type=fill_];
+"1218 lift_fresh_copy_28" [id=1218, type=lift_fresh_copy];
+"1219 slice_113" [id=1219, type=slice];
+"1220 slice_114" [id=1220, type=slice];
+"1221 fill__28" [id=1221, type=fill_];
+"1222 lift_fresh_copy_29" [id=1222, type=lift_fresh_copy];
+"1223 slice_115" [id=1223, type=slice];
+"1224 slice_116" [id=1224, type=slice];
+"1225 fill__29" [id=1225, type=fill_];
+"1226 lift_fresh_copy_30" [id=1226, type=lift_fresh_copy];
+"1227 slice_117" [id=1227, type=slice];
+"1228 slice_118" [id=1228, type=slice];
+"1229 fill__30" [id=1229, type=fill_];
+"1230 lift_fresh_copy_31" [id=1230, type=lift_fresh_copy];
+"1231 slice_119" [id=1231, type=slice];
+"1232 slice_120" [id=1232, type=slice];
+"1233 fill__31" [id=1233, type=fill_];
+"1234 lift_fresh_copy_32" [id=1234, type=lift_fresh_copy];
+"1235 slice_121" [id=1235, type=slice];
+"1236 slice_122" [id=1236, type=slice];
+"1237 fill__32" [id=1237, type=fill_];
+"1238 lift_fresh_copy_33" [id=1238, type=lift_fresh_copy];
+"1239 slice_123" [id=1239, type=slice];
+"1240 slice_124" [id=1240, type=slice];
+"1241 fill__33" [id=1241, type=fill_];
+"1242 lift_fresh_copy_34" [id=1242, type=lift_fresh_copy];
+"1243 slice_125" [id=1243, type=slice];
+"1244 slice_126" [id=1244, type=slice];
+"1245 fill__34" [id=1245, type=fill_];
+"1246 lift_fresh_copy_35" [id=1246, type=lift_fresh_copy];
+"1247 slice_127" [id=1247, type=slice];
+"1248 slice_128" [id=1248, type=slice];
+"1249 fill__35" [id=1249, type=fill_];
+"1250 view_40" [id=1250, type=view];
+"1251 permute_35" [id=1251, type=permute];
+"1252 reshape_33" [id=1252, type=reshape];
+"1253 unsqueeze_20" [id=1253, type=unsqueeze];
+"1254 unsqueeze_21" [id=1254, type=unsqueeze];
+"1255 sub_3" [id=1255, type=sub];
+"1256 ne_3" [id=1256, type=ne];
+"1257 masked_fill_6" [id=1257, type=masked_fill];
+"1258 eq_3" [id=1258, type=eq];
+"1259 masked_fill_7" [id=1259, type=masked_fill];
+"1260 view_41" [id=1260, type=view];
+"1261 unsqueeze_22" [id=1261, type=unsqueeze];
+"1262 unsqueeze_23" [id=1262, type=unsqueeze];
+"1263 add_25" [id=1263, type=add];
+"1264 view_42" [id=1264, type=view];
+"1265 softmax_7" [id=1265, type=softmax];
+"1266 dropout_28" [id=1266, type=dropout];
+"1267 matmul_15" [id=1267, type=matmul];
+"1268 transpose_15" [id=1268, type=transpose];
+"1269 reshape_34" [id=1269, type=reshape];
+"1270 linear_47" [id=1270, type=linear];
+"1271 dropout_29" [id=1271, type=dropout];
+"1272 view_43" [id=1272, type=view];
+"1273 permute_36" [id=1273, type=permute];
+"1274 reshape_35" [id=1274, type=reshape];
+"1275 roll_7" [id=1275, type=roll];
+"1276 slice_129" [id=1276, type=slice];
+"1277 slice_130" [id=1277, type=slice];
+"1278 slice_131" [id=1278, type=slice];
+"1279 slice_132" [id=1279, type=slice];
+"1280 contiguous_13" [id=1280, type=contiguous];
+"1281 layer_norm_17" [id=1281, type=layer_norm];
+"1282 add_26" [id=1282, type=add];
+"1283 linear_48" [id=1283, type=linear];
+"1284 gelu_7" [id=1284, type=gelu];
+"1285 dropout_30" [id=1285, type=dropout];
+"1286 linear_49" [id=1286, type=linear];
+"1287 dropout_31" [id=1287, type=dropout];
+"1288 layer_norm_18" [id=1288, type=layer_norm];
+"1289 add_27" [id=1289, type=add];
+"1290 linear_50" [id=1290, type=linear];
+"1291 relu__8" [id=1291, type=relu_];
+"1292 linear_51" [id=1292, type=linear];
+"1293 view_44" [id=1293, type=view];
+"1294 index_8" [id=1294, type=index];
+"1295 view_45" [id=1295, type=view];
+"1296 permute_37" [id=1296, type=permute];
+"1297 contiguous_14" [id=1297, type=contiguous];
+"1298 unsqueeze_24" [id=1298, type=unsqueeze];
+"1299 sigmoid_8" [id=1299, type=sigmoid];
+"1300 mul_16" [id=1300, type=mul];
+"1301 pad_10" [id=1301, type=pad];
+"1302 view_46" [id=1302, type=view];
+"1303 permute_38" [id=1303, type=permute];
+"1304 reshape_36" [id=1304, type=reshape];
+"1305 clone_8" [id=1305, type=clone];
+"1306 slice_133" [id=1306, type=slice];
+"1307 zero__8" [id=1307, type=zero_];
+"1308 linear_52" [id=1308, type=linear];
+"1309 reshape_37" [id=1309, type=reshape];
+"1310 permute_39" [id=1310, type=permute];
+"1311 select_24" [id=1311, type=select];
+"1312 select_25" [id=1312, type=select];
+"1313 select_26" [id=1313, type=select];
+"1314 linalg_vector_norm_16" [id=1314, type=linalg_vector_norm];
+"1315 clamp_min_16" [id=1315, type=clamp_min];
+"1316 expand_as_16" [id=1316, type=expand_as];
+"1317 div_16" [id=1317, type=div];
+"1318 linalg_vector_norm_17" [id=1318, type=linalg_vector_norm];
+"1319 clamp_min_17" [id=1319, type=clamp_min];
+"1320 expand_as_17" [id=1320, type=expand_as];
+"1321 div_17" [id=1321, type=div];
+"1322 transpose_16" [id=1322, type=transpose];
+"1323 matmul_16" [id=1323, type=matmul];
+"1324 clamp_8" [id=1324, type=clamp];
+"1325 exp_8" [id=1325, type=exp];
+"1326 mul_17" [id=1326, type=mul];
+"1327 add_28" [id=1327, type=add];
+"1328 softmax_8" [id=1328, type=softmax];
+"1329 dropout_32" [id=1329, type=dropout];
+"1330 matmul_17" [id=1330, type=matmul];
+"1331 transpose_17" [id=1331, type=transpose];
+"1332 reshape_38" [id=1332, type=reshape];
+"1333 linear_53" [id=1333, type=linear];
+"1334 dropout_33" [id=1334, type=dropout];
+"1335 view_47" [id=1335, type=view];
+"1336 permute_40" [id=1336, type=permute];
+"1337 reshape_39" [id=1337, type=reshape];
+"1338 slice_134" [id=1338, type=slice];
+"1339 slice_135" [id=1339, type=slice];
+"1340 slice_136" [id=1340, type=slice];
+"1341 slice_137" [id=1341, type=slice];
+"1342 contiguous_15" [id=1342, type=contiguous];
+"1343 layer_norm_19" [id=1343, type=layer_norm];
+"1344 add_29" [id=1344, type=add];
+"1345 linear_54" [id=1345, type=linear];
+"1346 gelu_8" [id=1346, type=gelu];
+"1347 dropout_34" [id=1347, type=dropout];
+"1348 linear_55" [id=1348, type=linear];
+"1349 dropout_35" [id=1349, type=dropout];
+"1350 layer_norm_20" [id=1350, type=layer_norm];
+"1351 add_30" [id=1351, type=add];
+"1352 linear_56" [id=1352, type=linear];
+"1353 relu__9" [id=1353, type=relu_];
+"1354 linear_57" [id=1354, type=linear];
+"1355 view_48" [id=1355, type=view];
+"1356 index_9" [id=1356, type=index];
+"1357 view_49" [id=1357, type=view];
+"1358 permute_41" [id=1358, type=permute];
+"1359 contiguous_16" [id=1359, type=contiguous];
+"1360 unsqueeze_25" [id=1360, type=unsqueeze];
+"1361 sigmoid_9" [id=1361, type=sigmoid];
+"1362 mul_18" [id=1362, type=mul];
+"1363 pad_11" [id=1363, type=pad];
+"1364 roll_8" [id=1364, type=roll];
+"1365 view_50" [id=1365, type=view];
+"1366 permute_42" [id=1366, type=permute];
+"1367 reshape_40" [id=1367, type=reshape];
+"1368 clone_9" [id=1368, type=clone];
+"1369 slice_138" [id=1369, type=slice];
+"1370 zero__9" [id=1370, type=zero_];
+"1371 linear_58" [id=1371, type=linear];
+"1372 reshape_41" [id=1372, type=reshape];
+"1373 permute_43" [id=1373, type=permute];
+"1374 select_27" [id=1374, type=select];
+"1375 select_28" [id=1375, type=select];
+"1376 select_29" [id=1376, type=select];
+"1377 linalg_vector_norm_18" [id=1377, type=linalg_vector_norm];
+"1378 clamp_min_18" [id=1378, type=clamp_min];
+"1379 expand_as_18" [id=1379, type=expand_as];
+"1380 div_18" [id=1380, type=div];
+"1381 linalg_vector_norm_19" [id=1381, type=linalg_vector_norm];
+"1382 clamp_min_19" [id=1382, type=clamp_min];
+"1383 expand_as_19" [id=1383, type=expand_as];
+"1384 div_19" [id=1384, type=div];
+"1385 transpose_18" [id=1385, type=transpose];
+"1386 matmul_18" [id=1386, type=matmul];
+"1387 clamp_9" [id=1387, type=clamp];
+"1388 exp_9" [id=1388, type=exp];
+"1389 mul_19" [id=1389, type=mul];
+"1390 add_31" [id=1390, type=add];
+"1391 new_zeros_4" [id=1391, type=new_zeros];
+"1392 lift_fresh_copy_36" [id=1392, type=lift_fresh_copy];
+"1393 slice_139" [id=1393, type=slice];
+"1394 slice_140" [id=1394, type=slice];
+"1395 fill__36" [id=1395, type=fill_];
+"1396 lift_fresh_copy_37" [id=1396, type=lift_fresh_copy];
+"1397 slice_141" [id=1397, type=slice];
+"1398 slice_142" [id=1398, type=slice];
+"1399 fill__37" [id=1399, type=fill_];
+"1400 lift_fresh_copy_38" [id=1400, type=lift_fresh_copy];
+"1401 slice_143" [id=1401, type=slice];
+"1402 slice_144" [id=1402, type=slice];
+"1403 fill__38" [id=1403, type=fill_];
+"1404 lift_fresh_copy_39" [id=1404, type=lift_fresh_copy];
+"1405 slice_145" [id=1405, type=slice];
+"1406 slice_146" [id=1406, type=slice];
+"1407 fill__39" [id=1407, type=fill_];
+"1408 lift_fresh_copy_40" [id=1408, type=lift_fresh_copy];
+"1409 slice_147" [id=1409, type=slice];
+"1410 slice_148" [id=1410, type=slice];
+"1411 fill__40" [id=1411, type=fill_];
+"1412 lift_fresh_copy_41" [id=1412, type=lift_fresh_copy];
+"1413 slice_149" [id=1413, type=slice];
+"1414 slice_150" [id=1414, type=slice];
+"1415 fill__41" [id=1415, type=fill_];
+"1416 lift_fresh_copy_42" [id=1416, type=lift_fresh_copy];
+"1417 slice_151" [id=1417, type=slice];
+"1418 slice_152" [id=1418, type=slice];
+"1419 fill__42" [id=1419, type=fill_];
+"1420 lift_fresh_copy_43" [id=1420, type=lift_fresh_copy];
+"1421 slice_153" [id=1421, type=slice];
+"1422 slice_154" [id=1422, type=slice];
+"1423 fill__43" [id=1423, type=fill_];
+"1424 lift_fresh_copy_44" [id=1424, type=lift_fresh_copy];
+"1425 slice_155" [id=1425, type=slice];
+"1426 slice_156" [id=1426, type=slice];
+"1427 fill__44" [id=1427, type=fill_];
+"1428 view_51" [id=1428, type=view];
+"1429 permute_44" [id=1429, type=permute];
+"1430 reshape_42" [id=1430, type=reshape];
+"1431 unsqueeze_26" [id=1431, type=unsqueeze];
+"1432 unsqueeze_27" [id=1432, type=unsqueeze];
+"1433 sub_4" [id=1433, type=sub];
+"1434 ne_4" [id=1434, type=ne];
+"1435 masked_fill_8" [id=1435, type=masked_fill];
+"1436 eq_4" [id=1436, type=eq];
+"1437 masked_fill_9" [id=1437, type=masked_fill];
+"1438 view_52" [id=1438, type=view];
+"1439 unsqueeze_28" [id=1439, type=unsqueeze];
+"1440 unsqueeze_29" [id=1440, type=unsqueeze];
+"1441 add_32" [id=1441, type=add];
+"1442 view_53" [id=1442, type=view];
+"1443 softmax_9" [id=1443, type=softmax];
+"1444 dropout_36" [id=1444, type=dropout];
+"1445 matmul_19" [id=1445, type=matmul];
+"1446 transpose_19" [id=1446, type=transpose];
+"1447 reshape_43" [id=1447, type=reshape];
+"1448 linear_59" [id=1448, type=linear];
+"1449 dropout_37" [id=1449, type=dropout];
+"1450 view_54" [id=1450, type=view];
+"1451 permute_45" [id=1451, type=permute];
+"1452 reshape_44" [id=1452, type=reshape];
+"1453 roll_9" [id=1453, type=roll];
+"1454 slice_157" [id=1454, type=slice];
+"1455 slice_158" [id=1455, type=slice];
+"1456 slice_159" [id=1456, type=slice];
+"1457 slice_160" [id=1457, type=slice];
+"1458 contiguous_17" [id=1458, type=contiguous];
+"1459 layer_norm_21" [id=1459, type=layer_norm];
+"1460 add_33" [id=1460, type=add];
+"1461 linear_60" [id=1461, type=linear];
+"1462 gelu_9" [id=1462, type=gelu];
+"1463 dropout_38" [id=1463, type=dropout];
+"1464 linear_61" [id=1464, type=linear];
+"1465 dropout_39" [id=1465, type=dropout];
+"1466 layer_norm_22" [id=1466, type=layer_norm];
+"1467 add_34" [id=1467, type=add];
+"1468 linear_62" [id=1468, type=linear];
+"1469 relu__10" [id=1469, type=relu_];
+"1470 linear_63" [id=1470, type=linear];
+"1471 view_55" [id=1471, type=view];
+"1472 index_10" [id=1472, type=index];
+"1473 view_56" [id=1473, type=view];
+"1474 permute_46" [id=1474, type=permute];
+"1475 contiguous_18" [id=1475, type=contiguous];
+"1476 unsqueeze_30" [id=1476, type=unsqueeze];
+"1477 sigmoid_10" [id=1477, type=sigmoid];
+"1478 mul_20" [id=1478, type=mul];
+"1479 pad_12" [id=1479, type=pad];
+"1480 view_57" [id=1480, type=view];
+"1481 permute_47" [id=1481, type=permute];
+"1482 reshape_45" [id=1482, type=reshape];
+"1483 clone_10" [id=1483, type=clone];
+"1484 slice_161" [id=1484, type=slice];
+"1485 zero__10" [id=1485, type=zero_];
+"1486 linear_64" [id=1486, type=linear];
+"1487 reshape_46" [id=1487, type=reshape];
+"1488 permute_48" [id=1488, type=permute];
+"1489 select_30" [id=1489, type=select];
+"1490 select_31" [id=1490, type=select];
+"1491 select_32" [id=1491, type=select];
+"1492 linalg_vector_norm_20" [id=1492, type=linalg_vector_norm];
+"1493 clamp_min_20" [id=1493, type=clamp_min];
+"1494 expand_as_20" [id=1494, type=expand_as];
+"1495 div_20" [id=1495, type=div];
+"1496 linalg_vector_norm_21" [id=1496, type=linalg_vector_norm];
+"1497 clamp_min_21" [id=1497, type=clamp_min];
+"1498 expand_as_21" [id=1498, type=expand_as];
+"1499 div_21" [id=1499, type=div];
+"1500 transpose_20" [id=1500, type=transpose];
+"1501 matmul_20" [id=1501, type=matmul];
+"1502 clamp_10" [id=1502, type=clamp];
+"1503 exp_10" [id=1503, type=exp];
+"1504 mul_21" [id=1504, type=mul];
+"1505 add_35" [id=1505, type=add];
+"1506 softmax_10" [id=1506, type=softmax];
+"1507 dropout_40" [id=1507, type=dropout];
+"1508 matmul_21" [id=1508, type=matmul];
+"1509 transpose_21" [id=1509, type=transpose];
+"1510 reshape_47" [id=1510, type=reshape];
+"1511 linear_65" [id=1511, type=linear];
+"1512 dropout_41" [id=1512, type=dropout];
+"1513 view_58" [id=1513, type=view];
+"1514 permute_49" [id=1514, type=permute];
+"1515 reshape_48" [id=1515, type=reshape];
+"1516 slice_162" [id=1516, type=slice];
+"1517 slice_163" [id=1517, type=slice];
+"1518 slice_164" [id=1518, type=slice];
+"1519 slice_165" [id=1519, type=slice];
+"1520 contiguous_19" [id=1520, type=contiguous];
+"1521 layer_norm_23" [id=1521, type=layer_norm];
+"1522 add_36" [id=1522, type=add];
+"1523 linear_66" [id=1523, type=linear];
+"1524 gelu_10" [id=1524, type=gelu];
+"1525 dropout_42" [id=1525, type=dropout];
+"1526 linear_67" [id=1526, type=linear];
+"1527 dropout_43" [id=1527, type=dropout];
+"1528 layer_norm_24" [id=1528, type=layer_norm];
+"1529 add_37" [id=1529, type=add];
+"1530 linear_68" [id=1530, type=linear];
+"1531 relu__11" [id=1531, type=relu_];
+"1532 linear_69" [id=1532, type=linear];
+"1533 view_59" [id=1533, type=view];
+"1534 index_11" [id=1534, type=index];
+"1535 view_60" [id=1535, type=view];
+"1536 permute_50" [id=1536, type=permute];
+"1537 contiguous_20" [id=1537, type=contiguous];
+"1538 unsqueeze_31" [id=1538, type=unsqueeze];
+"1539 sigmoid_11" [id=1539, type=sigmoid];
+"1540 mul_22" [id=1540, type=mul];
+"1541 pad_13" [id=1541, type=pad];
+"1542 roll_10" [id=1542, type=roll];
+"1543 view_61" [id=1543, type=view];
+"1544 permute_51" [id=1544, type=permute];
+"1545 reshape_49" [id=1545, type=reshape];
+"1546 clone_11" [id=1546, type=clone];
+"1547 slice_166" [id=1547, type=slice];
+"1548 zero__11" [id=1548, type=zero_];
+"1549 linear_70" [id=1549, type=linear];
+"1550 reshape_50" [id=1550, type=reshape];
+"1551 permute_52" [id=1551, type=permute];
+"1552 select_33" [id=1552, type=select];
+"1553 select_34" [id=1553, type=select];
+"1554 select_35" [id=1554, type=select];
+"1555 linalg_vector_norm_22" [id=1555, type=linalg_vector_norm];
+"1556 clamp_min_22" [id=1556, type=clamp_min];
+"1557 expand_as_22" [id=1557, type=expand_as];
+"1558 div_22" [id=1558, type=div];
+"1559 linalg_vector_norm_23" [id=1559, type=linalg_vector_norm];
+"1560 clamp_min_23" [id=1560, type=clamp_min];
+"1561 expand_as_23" [id=1561, type=expand_as];
+"1562 div_23" [id=1562, type=div];
+"1563 transpose_22" [id=1563, type=transpose];
+"1564 matmul_22" [id=1564, type=matmul];
+"1565 clamp_11" [id=1565, type=clamp];
+"1566 exp_11" [id=1566, type=exp];
+"1567 mul_23" [id=1567, type=mul];
+"1568 add_38" [id=1568, type=add];
+"1569 new_zeros_5" [id=1569, type=new_zeros];
+"1570 lift_fresh_copy_45" [id=1570, type=lift_fresh_copy];
+"1571 slice_167" [id=1571, type=slice];
+"1572 slice_168" [id=1572, type=slice];
+"1573 fill__45" [id=1573, type=fill_];
+"1574 lift_fresh_copy_46" [id=1574, type=lift_fresh_copy];
+"1575 slice_169" [id=1575, type=slice];
+"1576 slice_170" [id=1576, type=slice];
+"1577 fill__46" [id=1577, type=fill_];
+"1578 lift_fresh_copy_47" [id=1578, type=lift_fresh_copy];
+"1579 slice_171" [id=1579, type=slice];
+"1580 slice_172" [id=1580, type=slice];
+"1581 fill__47" [id=1581, type=fill_];
+"1582 lift_fresh_copy_48" [id=1582, type=lift_fresh_copy];
+"1583 slice_173" [id=1583, type=slice];
+"1584 slice_174" [id=1584, type=slice];
+"1585 fill__48" [id=1585, type=fill_];
+"1586 lift_fresh_copy_49" [id=1586, type=lift_fresh_copy];
+"1587 slice_175" [id=1587, type=slice];
+"1588 slice_176" [id=1588, type=slice];
+"1589 fill__49" [id=1589, type=fill_];
+"1590 lift_fresh_copy_50" [id=1590, type=lift_fresh_copy];
+"1591 slice_177" [id=1591, type=slice];
+"1592 slice_178" [id=1592, type=slice];
+"1593 fill__50" [id=1593, type=fill_];
+"1594 lift_fresh_copy_51" [id=1594, type=lift_fresh_copy];
+"1595 slice_179" [id=1595, type=slice];
+"1596 slice_180" [id=1596, type=slice];
+"1597 fill__51" [id=1597, type=fill_];
+"1598 lift_fresh_copy_52" [id=1598, type=lift_fresh_copy];
+"1599 slice_181" [id=1599, type=slice];
+"1600 slice_182" [id=1600, type=slice];
+"1601 fill__52" [id=1601, type=fill_];
+"1602 lift_fresh_copy_53" [id=1602, type=lift_fresh_copy];
+"1603 slice_183" [id=1603, type=slice];
+"1604 slice_184" [id=1604, type=slice];
+"1605 fill__53" [id=1605, type=fill_];
+"1606 view_62" [id=1606, type=view];
+"1607 permute_53" [id=1607, type=permute];
+"1608 reshape_51" [id=1608, type=reshape];
+"1609 unsqueeze_32" [id=1609, type=unsqueeze];
+"1610 unsqueeze_33" [id=1610, type=unsqueeze];
+"1611 sub_5" [id=1611, type=sub];
+"1612 ne_5" [id=1612, type=ne];
+"1613 masked_fill_10" [id=1613, type=masked_fill];
+"1614 eq_5" [id=1614, type=eq];
+"1615 masked_fill_11" [id=1615, type=masked_fill];
+"1616 view_63" [id=1616, type=view];
+"1617 unsqueeze_34" [id=1617, type=unsqueeze];
+"1618 unsqueeze_35" [id=1618, type=unsqueeze];
+"1619 add_39" [id=1619, type=add];
+"1620 view_64" [id=1620, type=view];
+"1621 softmax_11" [id=1621, type=softmax];
+"1622 dropout_44" [id=1622, type=dropout];
+"1623 matmul_23" [id=1623, type=matmul];
+"1624 transpose_23" [id=1624, type=transpose];
+"1625 reshape_52" [id=1625, type=reshape];
+"1626 linear_71" [id=1626, type=linear];
+"1627 dropout_45" [id=1627, type=dropout];
+"1628 view_65" [id=1628, type=view];
+"1629 permute_54" [id=1629, type=permute];
+"1630 reshape_53" [id=1630, type=reshape];
+"1631 roll_11" [id=1631, type=roll];
+"1632 slice_185" [id=1632, type=slice];
+"1633 slice_186" [id=1633, type=slice];
+"1634 slice_187" [id=1634, type=slice];
+"1635 slice_188" [id=1635, type=slice];
+"1636 contiguous_21" [id=1636, type=contiguous];
+"1637 layer_norm_25" [id=1637, type=layer_norm];
+"1638 add_40" [id=1638, type=add];
+"1639 linear_72" [id=1639, type=linear];
+"1640 gelu_11" [id=1640, type=gelu];
+"1641 dropout_46" [id=1641, type=dropout];
+"1642 linear_73" [id=1642, type=linear];
+"1643 dropout_47" [id=1643, type=dropout];
+"1644 layer_norm_26" [id=1644, type=layer_norm];
+"1645 add_41" [id=1645, type=add];
+"1646 linear_74" [id=1646, type=linear];
+"1647 relu__12" [id=1647, type=relu_];
+"1648 linear_75" [id=1648, type=linear];
+"1649 view_66" [id=1649, type=view];
+"1650 index_12" [id=1650, type=index];
+"1651 view_67" [id=1651, type=view];
+"1652 permute_55" [id=1652, type=permute];
+"1653 contiguous_22" [id=1653, type=contiguous];
+"1654 unsqueeze_36" [id=1654, type=unsqueeze];
+"1655 sigmoid_12" [id=1655, type=sigmoid];
+"1656 mul_24" [id=1656, type=mul];
+"1657 pad_14" [id=1657, type=pad];
+"1658 view_68" [id=1658, type=view];
+"1659 permute_56" [id=1659, type=permute];
+"1660 reshape_54" [id=1660, type=reshape];
+"1661 clone_12" [id=1661, type=clone];
+"1662 slice_189" [id=1662, type=slice];
+"1663 zero__12" [id=1663, type=zero_];
+"1664 linear_76" [id=1664, type=linear];
+"1665 reshape_55" [id=1665, type=reshape];
+"1666 permute_57" [id=1666, type=permute];
+"1667 select_36" [id=1667, type=select];
+"1668 select_37" [id=1668, type=select];
+"1669 select_38" [id=1669, type=select];
+"1670 linalg_vector_norm_24" [id=1670, type=linalg_vector_norm];
+"1671 clamp_min_24" [id=1671, type=clamp_min];
+"1672 expand_as_24" [id=1672, type=expand_as];
+"1673 div_24" [id=1673, type=div];
+"1674 linalg_vector_norm_25" [id=1674, type=linalg_vector_norm];
+"1675 clamp_min_25" [id=1675, type=clamp_min];
+"1676 expand_as_25" [id=1676, type=expand_as];
+"1677 div_25" [id=1677, type=div];
+"1678 transpose_24" [id=1678, type=transpose];
+"1679 matmul_24" [id=1679, type=matmul];
+"1680 clamp_12" [id=1680, type=clamp];
+"1681 exp_12" [id=1681, type=exp];
+"1682 mul_25" [id=1682, type=mul];
+"1683 add_42" [id=1683, type=add];
+"1684 softmax_12" [id=1684, type=softmax];
+"1685 dropout_48" [id=1685, type=dropout];
+"1686 matmul_25" [id=1686, type=matmul];
+"1687 transpose_25" [id=1687, type=transpose];
+"1688 reshape_56" [id=1688, type=reshape];
+"1689 linear_77" [id=1689, type=linear];
+"1690 dropout_49" [id=1690, type=dropout];
+"1691 view_69" [id=1691, type=view];
+"1692 permute_58" [id=1692, type=permute];
+"1693 reshape_57" [id=1693, type=reshape];
+"1694 slice_190" [id=1694, type=slice];
+"1695 slice_191" [id=1695, type=slice];
+"1696 slice_192" [id=1696, type=slice];
+"1697 slice_193" [id=1697, type=slice];
+"1698 contiguous_23" [id=1698, type=contiguous];
+"1699 layer_norm_27" [id=1699, type=layer_norm];
+"1700 add_43" [id=1700, type=add];
+"1701 linear_78" [id=1701, type=linear];
+"1702 gelu_12" [id=1702, type=gelu];
+"1703 dropout_50" [id=1703, type=dropout];
+"1704 linear_79" [id=1704, type=linear];
+"1705 dropout_51" [id=1705, type=dropout];
+"1706 layer_norm_28" [id=1706, type=layer_norm];
+"1707 add_44" [id=1707, type=add];
+"1708 linear_80" [id=1708, type=linear];
+"1709 relu__13" [id=1709, type=relu_];
+"1710 linear_81" [id=1710, type=linear];
+"1711 view_70" [id=1711, type=view];
+"1712 index_13" [id=1712, type=index];
+"1713 view_71" [id=1713, type=view];
+"1714 permute_59" [id=1714, type=permute];
+"1715 contiguous_24" [id=1715, type=contiguous];
+"1716 unsqueeze_37" [id=1716, type=unsqueeze];
+"1717 sigmoid_13" [id=1717, type=sigmoid];
+"1718 mul_26" [id=1718, type=mul];
+"1719 pad_15" [id=1719, type=pad];
+"1720 roll_12" [id=1720, type=roll];
+"1721 view_72" [id=1721, type=view];
+"1722 permute_60" [id=1722, type=permute];
+"1723 reshape_58" [id=1723, type=reshape];
+"1724 clone_13" [id=1724, type=clone];
+"1725 slice_194" [id=1725, type=slice];
+"1726 zero__13" [id=1726, type=zero_];
+"1727 linear_82" [id=1727, type=linear];
+"1728 reshape_59" [id=1728, type=reshape];
+"1729 permute_61" [id=1729, type=permute];
+"1730 select_39" [id=1730, type=select];
+"1731 select_40" [id=1731, type=select];
+"1732 select_41" [id=1732, type=select];
+"1733 linalg_vector_norm_26" [id=1733, type=linalg_vector_norm];
+"1734 clamp_min_26" [id=1734, type=clamp_min];
+"1735 expand_as_26" [id=1735, type=expand_as];
+"1736 div_26" [id=1736, type=div];
+"1737 linalg_vector_norm_27" [id=1737, type=linalg_vector_norm];
+"1738 clamp_min_27" [id=1738, type=clamp_min];
+"1739 expand_as_27" [id=1739, type=expand_as];
+"1740 div_27" [id=1740, type=div];
+"1741 transpose_26" [id=1741, type=transpose];
+"1742 matmul_26" [id=1742, type=matmul];
+"1743 clamp_13" [id=1743, type=clamp];
+"1744 exp_13" [id=1744, type=exp];
+"1745 mul_27" [id=1745, type=mul];
+"1746 add_45" [id=1746, type=add];
+"1747 new_zeros_6" [id=1747, type=new_zeros];
+"1748 lift_fresh_copy_54" [id=1748, type=lift_fresh_copy];
+"1749 slice_195" [id=1749, type=slice];
+"1750 slice_196" [id=1750, type=slice];
+"1751 fill__54" [id=1751, type=fill_];
+"1752 lift_fresh_copy_55" [id=1752, type=lift_fresh_copy];
+"1753 slice_197" [id=1753, type=slice];
+"1754 slice_198" [id=1754, type=slice];
+"1755 fill__55" [id=1755, type=fill_];
+"1756 lift_fresh_copy_56" [id=1756, type=lift_fresh_copy];
+"1757 slice_199" [id=1757, type=slice];
+"1758 slice_200" [id=1758, type=slice];
+"1759 fill__56" [id=1759, type=fill_];
+"1760 lift_fresh_copy_57" [id=1760, type=lift_fresh_copy];
+"1761 slice_201" [id=1761, type=slice];
+"1762 slice_202" [id=1762, type=slice];
+"1763 fill__57" [id=1763, type=fill_];
+"1764 lift_fresh_copy_58" [id=1764, type=lift_fresh_copy];
+"1765 slice_203" [id=1765, type=slice];
+"1766 slice_204" [id=1766, type=slice];
+"1767 fill__58" [id=1767, type=fill_];
+"1768 lift_fresh_copy_59" [id=1768, type=lift_fresh_copy];
+"1769 slice_205" [id=1769, type=slice];
+"1770 slice_206" [id=1770, type=slice];
+"1771 fill__59" [id=1771, type=fill_];
+"1772 lift_fresh_copy_60" [id=1772, type=lift_fresh_copy];
+"1773 slice_207" [id=1773, type=slice];
+"1774 slice_208" [id=1774, type=slice];
+"1775 fill__60" [id=1775, type=fill_];
+"1776 lift_fresh_copy_61" [id=1776, type=lift_fresh_copy];
+"1777 slice_209" [id=1777, type=slice];
+"1778 slice_210" [id=1778, type=slice];
+"1779 fill__61" [id=1779, type=fill_];
+"1780 lift_fresh_copy_62" [id=1780, type=lift_fresh_copy];
+"1781 slice_211" [id=1781, type=slice];
+"1782 slice_212" [id=1782, type=slice];
+"1783 fill__62" [id=1783, type=fill_];
+"1784 view_73" [id=1784, type=view];
+"1785 permute_62" [id=1785, type=permute];
+"1786 reshape_60" [id=1786, type=reshape];
+"1787 unsqueeze_38" [id=1787, type=unsqueeze];
+"1788 unsqueeze_39" [id=1788, type=unsqueeze];
+"1789 sub_6" [id=1789, type=sub];
+"1790 ne_6" [id=1790, type=ne];
+"1791 masked_fill_12" [id=1791, type=masked_fill];
+"1792 eq_6" [id=1792, type=eq];
+"1793 masked_fill_13" [id=1793, type=masked_fill];
+"1794 view_74" [id=1794, type=view];
+"1795 unsqueeze_40" [id=1795, type=unsqueeze];
+"1796 unsqueeze_41" [id=1796, type=unsqueeze];
+"1797 add_46" [id=1797, type=add];
+"1798 view_75" [id=1798, type=view];
+"1799 softmax_13" [id=1799, type=softmax];
+"1800 dropout_52" [id=1800, type=dropout];
+"1801 matmul_27" [id=1801, type=matmul];
+"1802 transpose_27" [id=1802, type=transpose];
+"1803 reshape_61" [id=1803, type=reshape];
+"1804 linear_83" [id=1804, type=linear];
+"1805 dropout_53" [id=1805, type=dropout];
+"1806 view_76" [id=1806, type=view];
+"1807 permute_63" [id=1807, type=permute];
+"1808 reshape_62" [id=1808, type=reshape];
+"1809 roll_13" [id=1809, type=roll];
+"1810 slice_213" [id=1810, type=slice];
+"1811 slice_214" [id=1811, type=slice];
+"1812 slice_215" [id=1812, type=slice];
+"1813 slice_216" [id=1813, type=slice];
+"1814 contiguous_25" [id=1814, type=contiguous];
+"1815 layer_norm_29" [id=1815, type=layer_norm];
+"1816 add_47" [id=1816, type=add];
+"1817 linear_84" [id=1817, type=linear];
+"1818 gelu_13" [id=1818, type=gelu];
+"1819 dropout_54" [id=1819, type=dropout];
+"1820 linear_85" [id=1820, type=linear];
+"1821 dropout_55" [id=1821, type=dropout];
+"1822 layer_norm_30" [id=1822, type=layer_norm];
+"1823 add_48" [id=1823, type=add];
+"1824 linear_86" [id=1824, type=linear];
+"1825 relu__14" [id=1825, type=relu_];
+"1826 linear_87" [id=1826, type=linear];
+"1827 view_77" [id=1827, type=view];
+"1828 index_14" [id=1828, type=index];
+"1829 view_78" [id=1829, type=view];
+"1830 permute_64" [id=1830, type=permute];
+"1831 contiguous_26" [id=1831, type=contiguous];
+"1832 unsqueeze_42" [id=1832, type=unsqueeze];
+"1833 sigmoid_14" [id=1833, type=sigmoid];
+"1834 mul_28" [id=1834, type=mul];
+"1835 pad_16" [id=1835, type=pad];
+"1836 view_79" [id=1836, type=view];
+"1837 permute_65" [id=1837, type=permute];
+"1838 reshape_63" [id=1838, type=reshape];
+"1839 clone_14" [id=1839, type=clone];
+"1840 slice_217" [id=1840, type=slice];
+"1841 zero__14" [id=1841, type=zero_];
+"1842 linear_88" [id=1842, type=linear];
+"1843 reshape_64" [id=1843, type=reshape];
+"1844 permute_66" [id=1844, type=permute];
+"1845 select_42" [id=1845, type=select];
+"1846 select_43" [id=1846, type=select];
+"1847 select_44" [id=1847, type=select];
+"1848 linalg_vector_norm_28" [id=1848, type=linalg_vector_norm];
+"1849 clamp_min_28" [id=1849, type=clamp_min];
+"1850 expand_as_28" [id=1850, type=expand_as];
+"1851 div_28" [id=1851, type=div];
+"1852 linalg_vector_norm_29" [id=1852, type=linalg_vector_norm];
+"1853 clamp_min_29" [id=1853, type=clamp_min];
+"1854 expand_as_29" [id=1854, type=expand_as];
+"1855 div_29" [id=1855, type=div];
+"1856 transpose_28" [id=1856, type=transpose];
+"1857 matmul_28" [id=1857, type=matmul];
+"1858 clamp_14" [id=1858, type=clamp];
+"1859 exp_14" [id=1859, type=exp];
+"1860 mul_29" [id=1860, type=mul];
+"1861 add_49" [id=1861, type=add];
+"1862 softmax_14" [id=1862, type=softmax];
+"1863 dropout_56" [id=1863, type=dropout];
+"1864 matmul_29" [id=1864, type=matmul];
+"1865 transpose_29" [id=1865, type=transpose];
+"1866 reshape_65" [id=1866, type=reshape];
+"1867 linear_89" [id=1867, type=linear];
+"1868 dropout_57" [id=1868, type=dropout];
+"1869 view_80" [id=1869, type=view];
+"1870 permute_67" [id=1870, type=permute];
+"1871 reshape_66" [id=1871, type=reshape];
+"1872 slice_218" [id=1872, type=slice];
+"1873 slice_219" [id=1873, type=slice];
+"1874 slice_220" [id=1874, type=slice];
+"1875 slice_221" [id=1875, type=slice];
+"1876 contiguous_27" [id=1876, type=contiguous];
+"1877 layer_norm_31" [id=1877, type=layer_norm];
+"1878 add_50" [id=1878, type=add];
+"1879 linear_90" [id=1879, type=linear];
+"1880 gelu_14" [id=1880, type=gelu];
+"1881 dropout_58" [id=1881, type=dropout];
+"1882 linear_91" [id=1882, type=linear];
+"1883 dropout_59" [id=1883, type=dropout];
+"1884 layer_norm_32" [id=1884, type=layer_norm];
+"1885 add_51" [id=1885, type=add];
+"1886 linear_92" [id=1886, type=linear];
+"1887 relu__15" [id=1887, type=relu_];
+"1888 linear_93" [id=1888, type=linear];
+"1889 view_81" [id=1889, type=view];
+"1890 index_15" [id=1890, type=index];
+"1891 view_82" [id=1891, type=view];
+"1892 permute_68" [id=1892, type=permute];
+"1893 contiguous_28" [id=1893, type=contiguous];
+"1894 unsqueeze_43" [id=1894, type=unsqueeze];
+"1895 sigmoid_15" [id=1895, type=sigmoid];
+"1896 mul_30" [id=1896, type=mul];
+"1897 pad_17" [id=1897, type=pad];
+"1898 roll_14" [id=1898, type=roll];
+"1899 view_83" [id=1899, type=view];
+"1900 permute_69" [id=1900, type=permute];
+"1901 reshape_67" [id=1901, type=reshape];
+"1902 clone_15" [id=1902, type=clone];
+"1903 slice_222" [id=1903, type=slice];
+"1904 zero__15" [id=1904, type=zero_];
+"1905 linear_94" [id=1905, type=linear];
+"1906 reshape_68" [id=1906, type=reshape];
+"1907 permute_70" [id=1907, type=permute];
+"1908 select_45" [id=1908, type=select];
+"1909 select_46" [id=1909, type=select];
+"1910 select_47" [id=1910, type=select];
+"1911 linalg_vector_norm_30" [id=1911, type=linalg_vector_norm];
+"1912 clamp_min_30" [id=1912, type=clamp_min];
+"1913 expand_as_30" [id=1913, type=expand_as];
+"1914 div_30" [id=1914, type=div];
+"1915 linalg_vector_norm_31" [id=1915, type=linalg_vector_norm];
+"1916 clamp_min_31" [id=1916, type=clamp_min];
+"1917 expand_as_31" [id=1917, type=expand_as];
+"1918 div_31" [id=1918, type=div];
+"1919 transpose_30" [id=1919, type=transpose];
+"1920 matmul_30" [id=1920, type=matmul];
+"1921 clamp_15" [id=1921, type=clamp];
+"1922 exp_15" [id=1922, type=exp];
+"1923 mul_31" [id=1923, type=mul];
+"1924 add_52" [id=1924, type=add];
+"1925 new_zeros_7" [id=1925, type=new_zeros];
+"1926 lift_fresh_copy_63" [id=1926, type=lift_fresh_copy];
+"1927 slice_223" [id=1927, type=slice];
+"1928 slice_224" [id=1928, type=slice];
+"1929 fill__63" [id=1929, type=fill_];
+"1930 lift_fresh_copy_64" [id=1930, type=lift_fresh_copy];
+"1931 slice_225" [id=1931, type=slice];
+"1932 slice_226" [id=1932, type=slice];
+"1933 fill__64" [id=1933, type=fill_];
+"1934 lift_fresh_copy_65" [id=1934, type=lift_fresh_copy];
+"1935 slice_227" [id=1935, type=slice];
+"1936 slice_228" [id=1936, type=slice];
+"1937 fill__65" [id=1937, type=fill_];
+"1938 lift_fresh_copy_66" [id=1938, type=lift_fresh_copy];
+"1939 slice_229" [id=1939, type=slice];
+"1940 slice_230" [id=1940, type=slice];
+"1941 fill__66" [id=1941, type=fill_];
+"1942 lift_fresh_copy_67" [id=1942, type=lift_fresh_copy];
+"1943 slice_231" [id=1943, type=slice];
+"1944 slice_232" [id=1944, type=slice];
+"1945 fill__67" [id=1945, type=fill_];
+"1946 lift_fresh_copy_68" [id=1946, type=lift_fresh_copy];
+"1947 slice_233" [id=1947, type=slice];
+"1948 slice_234" [id=1948, type=slice];
+"1949 fill__68" [id=1949, type=fill_];
+"1950 lift_fresh_copy_69" [id=1950, type=lift_fresh_copy];
+"1951 slice_235" [id=1951, type=slice];
+"1952 slice_236" [id=1952, type=slice];
+"1953 fill__69" [id=1953, type=fill_];
+"1954 lift_fresh_copy_70" [id=1954, type=lift_fresh_copy];
+"1955 slice_237" [id=1955, type=slice];
+"1956 slice_238" [id=1956, type=slice];
+"1957 fill__70" [id=1957, type=fill_];
+"1958 lift_fresh_copy_71" [id=1958, type=lift_fresh_copy];
+"1959 slice_239" [id=1959, type=slice];
+"1960 slice_240" [id=1960, type=slice];
+"1961 fill__71" [id=1961, type=fill_];
+"1962 view_84" [id=1962, type=view];
+"1963 permute_71" [id=1963, type=permute];
+"1964 reshape_69" [id=1964, type=reshape];
+"1965 unsqueeze_44" [id=1965, type=unsqueeze];
+"1966 unsqueeze_45" [id=1966, type=unsqueeze];
+"1967 sub_7" [id=1967, type=sub];
+"1968 ne_7" [id=1968, type=ne];
+"1969 masked_fill_14" [id=1969, type=masked_fill];
+"1970 eq_7" [id=1970, type=eq];
+"1971 masked_fill_15" [id=1971, type=masked_fill];
+"1972 view_85" [id=1972, type=view];
+"1973 unsqueeze_46" [id=1973, type=unsqueeze];
+"1974 unsqueeze_47" [id=1974, type=unsqueeze];
+"1975 add_53" [id=1975, type=add];
+"1976 view_86" [id=1976, type=view];
+"1977 softmax_15" [id=1977, type=softmax];
+"1978 dropout_60" [id=1978, type=dropout];
+"1979 matmul_31" [id=1979, type=matmul];
+"1980 transpose_31" [id=1980, type=transpose];
+"1981 reshape_70" [id=1981, type=reshape];
+"1982 linear_95" [id=1982, type=linear];
+"1983 dropout_61" [id=1983, type=dropout];
+"1984 view_87" [id=1984, type=view];
+"1985 permute_72" [id=1985, type=permute];
+"1986 reshape_71" [id=1986, type=reshape];
+"1987 roll_15" [id=1987, type=roll];
+"1988 slice_241" [id=1988, type=slice];
+"1989 slice_242" [id=1989, type=slice];
+"1990 slice_243" [id=1990, type=slice];
+"1991 slice_244" [id=1991, type=slice];
+"1992 contiguous_29" [id=1992, type=contiguous];
+"1993 layer_norm_33" [id=1993, type=layer_norm];
+"1994 add_54" [id=1994, type=add];
+"1995 linear_96" [id=1995, type=linear];
+"1996 gelu_15" [id=1996, type=gelu];
+"1997 dropout_62" [id=1997, type=dropout];
+"1998 linear_97" [id=1998, type=linear];
+"1999 dropout_63" [id=1999, type=dropout];
+"2000 layer_norm_34" [id=2000, type=layer_norm];
+"2001 add_55" [id=2001, type=add];
+"2002 linear_98" [id=2002, type=linear];
+"2003 relu__16" [id=2003, type=relu_];
+"2004 linear_99" [id=2004, type=linear];
+"2005 view_88" [id=2005, type=view];
+"2006 index_16" [id=2006, type=index];
+"2007 view_89" [id=2007, type=view];
+"2008 permute_73" [id=2008, type=permute];
+"2009 contiguous_30" [id=2009, type=contiguous];
+"2010 unsqueeze_48" [id=2010, type=unsqueeze];
+"2011 sigmoid_16" [id=2011, type=sigmoid];
+"2012 mul_32" [id=2012, type=mul];
+"2013 pad_18" [id=2013, type=pad];
+"2014 view_90" [id=2014, type=view];
+"2015 permute_74" [id=2015, type=permute];
+"2016 reshape_72" [id=2016, type=reshape];
+"2017 clone_16" [id=2017, type=clone];
+"2018 slice_245" [id=2018, type=slice];
+"2019 zero__16" [id=2019, type=zero_];
+"2020 linear_100" [id=2020, type=linear];
+"2021 reshape_73" [id=2021, type=reshape];
+"2022 permute_75" [id=2022, type=permute];
+"2023 select_48" [id=2023, type=select];
+"2024 select_49" [id=2024, type=select];
+"2025 select_50" [id=2025, type=select];
+"2026 linalg_vector_norm_32" [id=2026, type=linalg_vector_norm];
+"2027 clamp_min_32" [id=2027, type=clamp_min];
+"2028 expand_as_32" [id=2028, type=expand_as];
+"2029 div_32" [id=2029, type=div];
+"2030 linalg_vector_norm_33" [id=2030, type=linalg_vector_norm];
+"2031 clamp_min_33" [id=2031, type=clamp_min];
+"2032 expand_as_33" [id=2032, type=expand_as];
+"2033 div_33" [id=2033, type=div];
+"2034 transpose_32" [id=2034, type=transpose];
+"2035 matmul_32" [id=2035, type=matmul];
+"2036 clamp_16" [id=2036, type=clamp];
+"2037 exp_16" [id=2037, type=exp];
+"2038 mul_33" [id=2038, type=mul];
+"2039 add_56" [id=2039, type=add];
+"2040 softmax_16" [id=2040, type=softmax];
+"2041 dropout_64" [id=2041, type=dropout];
+"2042 matmul_33" [id=2042, type=matmul];
+"2043 transpose_33" [id=2043, type=transpose];
+"2044 reshape_74" [id=2044, type=reshape];
+"2045 linear_101" [id=2045, type=linear];
+"2046 dropout_65" [id=2046, type=dropout];
+"2047 view_91" [id=2047, type=view];
+"2048 permute_76" [id=2048, type=permute];
+"2049 reshape_75" [id=2049, type=reshape];
+"2050 slice_246" [id=2050, type=slice];
+"2051 slice_247" [id=2051, type=slice];
+"2052 slice_248" [id=2052, type=slice];
+"2053 slice_249" [id=2053, type=slice];
+"2054 contiguous_31" [id=2054, type=contiguous];
+"2055 layer_norm_35" [id=2055, type=layer_norm];
+"2056 add_57" [id=2056, type=add];
+"2057 linear_102" [id=2057, type=linear];
+"2058 gelu_16" [id=2058, type=gelu];
+"2059 dropout_66" [id=2059, type=dropout];
+"2060 linear_103" [id=2060, type=linear];
+"2061 dropout_67" [id=2061, type=dropout];
+"2062 layer_norm_36" [id=2062, type=layer_norm];
+"2063 add_58" [id=2063, type=add];
+"2064 linear_104" [id=2064, type=linear];
+"2065 relu__17" [id=2065, type=relu_];
+"2066 linear_105" [id=2066, type=linear];
+"2067 view_92" [id=2067, type=view];
+"2068 index_17" [id=2068, type=index];
+"2069 view_93" [id=2069, type=view];
+"2070 permute_77" [id=2070, type=permute];
+"2071 contiguous_32" [id=2071, type=contiguous];
+"2072 unsqueeze_49" [id=2072, type=unsqueeze];
+"2073 sigmoid_17" [id=2073, type=sigmoid];
+"2074 mul_34" [id=2074, type=mul];
+"2075 pad_19" [id=2075, type=pad];
+"2076 roll_16" [id=2076, type=roll];
+"2077 view_94" [id=2077, type=view];
+"2078 permute_78" [id=2078, type=permute];
+"2079 reshape_76" [id=2079, type=reshape];
+"2080 clone_17" [id=2080, type=clone];
+"2081 slice_250" [id=2081, type=slice];
+"2082 zero__17" [id=2082, type=zero_];
+"2083 linear_106" [id=2083, type=linear];
+"2084 reshape_77" [id=2084, type=reshape];
+"2085 permute_79" [id=2085, type=permute];
+"2086 select_51" [id=2086, type=select];
+"2087 select_52" [id=2087, type=select];
+"2088 select_53" [id=2088, type=select];
+"2089 linalg_vector_norm_34" [id=2089, type=linalg_vector_norm];
+"2090 clamp_min_34" [id=2090, type=clamp_min];
+"2091 expand_as_34" [id=2091, type=expand_as];
+"2092 div_34" [id=2092, type=div];
+"2093 linalg_vector_norm_35" [id=2093, type=linalg_vector_norm];
+"2094 clamp_min_35" [id=2094, type=clamp_min];
+"2095 expand_as_35" [id=2095, type=expand_as];
+"2096 div_35" [id=2096, type=div];
+"2097 transpose_34" [id=2097, type=transpose];
+"2098 matmul_34" [id=2098, type=matmul];
+"2099 clamp_17" [id=2099, type=clamp];
+"2100 exp_17" [id=2100, type=exp];
+"2101 mul_35" [id=2101, type=mul];
+"2102 add_59" [id=2102, type=add];
+"2103 new_zeros_8" [id=2103, type=new_zeros];
+"2104 lift_fresh_copy_72" [id=2104, type=lift_fresh_copy];
+"2105 slice_251" [id=2105, type=slice];
+"2106 slice_252" [id=2106, type=slice];
+"2107 fill__72" [id=2107, type=fill_];
+"2108 lift_fresh_copy_73" [id=2108, type=lift_fresh_copy];
+"2109 slice_253" [id=2109, type=slice];
+"2110 slice_254" [id=2110, type=slice];
+"2111 fill__73" [id=2111, type=fill_];
+"2112 lift_fresh_copy_74" [id=2112, type=lift_fresh_copy];
+"2113 slice_255" [id=2113, type=slice];
+"2114 slice_256" [id=2114, type=slice];
+"2115 fill__74" [id=2115, type=fill_];
+"2116 lift_fresh_copy_75" [id=2116, type=lift_fresh_copy];
+"2117 slice_257" [id=2117, type=slice];
+"2118 slice_258" [id=2118, type=slice];
+"2119 fill__75" [id=2119, type=fill_];
+"2120 lift_fresh_copy_76" [id=2120, type=lift_fresh_copy];
+"2121 slice_259" [id=2121, type=slice];
+"2122 slice_260" [id=2122, type=slice];
+"2123 fill__76" [id=2123, type=fill_];
+"2124 lift_fresh_copy_77" [id=2124, type=lift_fresh_copy];
+"2125 slice_261" [id=2125, type=slice];
+"2126 slice_262" [id=2126, type=slice];
+"2127 fill__77" [id=2127, type=fill_];
+"2128 lift_fresh_copy_78" [id=2128, type=lift_fresh_copy];
+"2129 slice_263" [id=2129, type=slice];
+"2130 slice_264" [id=2130, type=slice];
+"2131 fill__78" [id=2131, type=fill_];
+"2132 lift_fresh_copy_79" [id=2132, type=lift_fresh_copy];
+"2133 slice_265" [id=2133, type=slice];
+"2134 slice_266" [id=2134, type=slice];
+"2135 fill__79" [id=2135, type=fill_];
+"2136 lift_fresh_copy_80" [id=2136, type=lift_fresh_copy];
+"2137 slice_267" [id=2137, type=slice];
+"2138 slice_268" [id=2138, type=slice];
+"2139 fill__80" [id=2139, type=fill_];
+"2140 view_95" [id=2140, type=view];
+"2141 permute_80" [id=2141, type=permute];
+"2142 reshape_78" [id=2142, type=reshape];
+"2143 unsqueeze_50" [id=2143, type=unsqueeze];
+"2144 unsqueeze_51" [id=2144, type=unsqueeze];
+"2145 sub_8" [id=2145, type=sub];
+"2146 ne_8" [id=2146, type=ne];
+"2147 masked_fill_16" [id=2147, type=masked_fill];
+"2148 eq_8" [id=2148, type=eq];
+"2149 masked_fill_17" [id=2149, type=masked_fill];
+"2150 view_96" [id=2150, type=view];
+"2151 unsqueeze_52" [id=2151, type=unsqueeze];
+"2152 unsqueeze_53" [id=2152, type=unsqueeze];
+"2153 add_60" [id=2153, type=add];
+"2154 view_97" [id=2154, type=view];
+"2155 softmax_17" [id=2155, type=softmax];
+"2156 dropout_68" [id=2156, type=dropout];
+"2157 matmul_35" [id=2157, type=matmul];
+"2158 transpose_35" [id=2158, type=transpose];
+"2159 reshape_79" [id=2159, type=reshape];
+"2160 linear_107" [id=2160, type=linear];
+"2161 dropout_69" [id=2161, type=dropout];
+"2162 view_98" [id=2162, type=view];
+"2163 permute_81" [id=2163, type=permute];
+"2164 reshape_80" [id=2164, type=reshape];
+"2165 roll_17" [id=2165, type=roll];
+"2166 slice_269" [id=2166, type=slice];
+"2167 slice_270" [id=2167, type=slice];
+"2168 slice_271" [id=2168, type=slice];
+"2169 slice_272" [id=2169, type=slice];
+"2170 contiguous_33" [id=2170, type=contiguous];
+"2171 layer_norm_37" [id=2171, type=layer_norm];
+"2172 add_61" [id=2172, type=add];
+"2173 linear_108" [id=2173, type=linear];
+"2174 gelu_17" [id=2174, type=gelu];
+"2175 dropout_70" [id=2175, type=dropout];
+"2176 linear_109" [id=2176, type=linear];
+"2177 dropout_71" [id=2177, type=dropout];
+"2178 layer_norm_38" [id=2178, type=layer_norm];
+"2179 add_62" [id=2179, type=add];
+"2180 linear_110" [id=2180, type=linear];
+"2181 relu__18" [id=2181, type=relu_];
+"2182 linear_111" [id=2182, type=linear];
+"2183 view_99" [id=2183, type=view];
+"2184 index_18" [id=2184, type=index];
+"2185 view_100" [id=2185, type=view];
+"2186 permute_82" [id=2186, type=permute];
+"2187 contiguous_34" [id=2187, type=contiguous];
+"2188 unsqueeze_54" [id=2188, type=unsqueeze];
+"2189 sigmoid_18" [id=2189, type=sigmoid];
+"2190 mul_36" [id=2190, type=mul];
+"2191 pad_20" [id=2191, type=pad];
+"2192 view_101" [id=2192, type=view];
+"2193 permute_83" [id=2193, type=permute];
+"2194 reshape_81" [id=2194, type=reshape];
+"2195 clone_18" [id=2195, type=clone];
+"2196 slice_273" [id=2196, type=slice];
+"2197 zero__18" [id=2197, type=zero_];
+"2198 linear_112" [id=2198, type=linear];
+"2199 reshape_82" [id=2199, type=reshape];
+"2200 permute_84" [id=2200, type=permute];
+"2201 select_54" [id=2201, type=select];
+"2202 select_55" [id=2202, type=select];
+"2203 select_56" [id=2203, type=select];
+"2204 linalg_vector_norm_36" [id=2204, type=linalg_vector_norm];
+"2205 clamp_min_36" [id=2205, type=clamp_min];
+"2206 expand_as_36" [id=2206, type=expand_as];
+"2207 div_36" [id=2207, type=div];
+"2208 linalg_vector_norm_37" [id=2208, type=linalg_vector_norm];
+"2209 clamp_min_37" [id=2209, type=clamp_min];
+"2210 expand_as_37" [id=2210, type=expand_as];
+"2211 div_37" [id=2211, type=div];
+"2212 transpose_36" [id=2212, type=transpose];
+"2213 matmul_36" [id=2213, type=matmul];
+"2214 clamp_18" [id=2214, type=clamp];
+"2215 exp_18" [id=2215, type=exp];
+"2216 mul_37" [id=2216, type=mul];
+"2217 add_63" [id=2217, type=add];
+"2218 softmax_18" [id=2218, type=softmax];
+"2219 dropout_72" [id=2219, type=dropout];
+"2220 matmul_37" [id=2220, type=matmul];
+"2221 transpose_37" [id=2221, type=transpose];
+"2222 reshape_83" [id=2222, type=reshape];
+"2223 linear_113" [id=2223, type=linear];
+"2224 dropout_73" [id=2224, type=dropout];
+"2225 view_102" [id=2225, type=view];
+"2226 permute_85" [id=2226, type=permute];
+"2227 reshape_84" [id=2227, type=reshape];
+"2228 slice_274" [id=2228, type=slice];
+"2229 slice_275" [id=2229, type=slice];
+"2230 slice_276" [id=2230, type=slice];
+"2231 slice_277" [id=2231, type=slice];
+"2232 contiguous_35" [id=2232, type=contiguous];
+"2233 layer_norm_39" [id=2233, type=layer_norm];
+"2234 add_64" [id=2234, type=add];
+"2235 linear_114" [id=2235, type=linear];
+"2236 gelu_18" [id=2236, type=gelu];
+"2237 dropout_74" [id=2237, type=dropout];
+"2238 linear_115" [id=2238, type=linear];
+"2239 dropout_75" [id=2239, type=dropout];
+"2240 layer_norm_40" [id=2240, type=layer_norm];
+"2241 add_65" [id=2241, type=add];
+"2242 linear_116" [id=2242, type=linear];
+"2243 relu__19" [id=2243, type=relu_];
+"2244 linear_117" [id=2244, type=linear];
+"2245 view_103" [id=2245, type=view];
+"2246 index_19" [id=2246, type=index];
+"2247 view_104" [id=2247, type=view];
+"2248 permute_86" [id=2248, type=permute];
+"2249 contiguous_36" [id=2249, type=contiguous];
+"2250 unsqueeze_55" [id=2250, type=unsqueeze];
+"2251 sigmoid_19" [id=2251, type=sigmoid];
+"2252 mul_38" [id=2252, type=mul];
+"2253 pad_21" [id=2253, type=pad];
+"2254 roll_18" [id=2254, type=roll];
+"2255 view_105" [id=2255, type=view];
+"2256 permute_87" [id=2256, type=permute];
+"2257 reshape_85" [id=2257, type=reshape];
+"2258 clone_19" [id=2258, type=clone];
+"2259 slice_278" [id=2259, type=slice];
+"2260 zero__19" [id=2260, type=zero_];
+"2261 linear_118" [id=2261, type=linear];
+"2262 reshape_86" [id=2262, type=reshape];
+"2263 permute_88" [id=2263, type=permute];
+"2264 select_57" [id=2264, type=select];
+"2265 select_58" [id=2265, type=select];
+"2266 select_59" [id=2266, type=select];
+"2267 linalg_vector_norm_38" [id=2267, type=linalg_vector_norm];
+"2268 clamp_min_38" [id=2268, type=clamp_min];
+"2269 expand_as_38" [id=2269, type=expand_as];
+"2270 div_38" [id=2270, type=div];
+"2271 linalg_vector_norm_39" [id=2271, type=linalg_vector_norm];
+"2272 clamp_min_39" [id=2272, type=clamp_min];
+"2273 expand_as_39" [id=2273, type=expand_as];
+"2274 div_39" [id=2274, type=div];
+"2275 transpose_38" [id=2275, type=transpose];
+"2276 matmul_38" [id=2276, type=matmul];
+"2277 clamp_19" [id=2277, type=clamp];
+"2278 exp_19" [id=2278, type=exp];
+"2279 mul_39" [id=2279, type=mul];
+"2280 add_66" [id=2280, type=add];
+"2281 new_zeros_9" [id=2281, type=new_zeros];
+"2282 lift_fresh_copy_81" [id=2282, type=lift_fresh_copy];
+"2283 slice_279" [id=2283, type=slice];
+"2284 slice_280" [id=2284, type=slice];
+"2285 fill__81" [id=2285, type=fill_];
+"2286 lift_fresh_copy_82" [id=2286, type=lift_fresh_copy];
+"2287 slice_281" [id=2287, type=slice];
+"2288 slice_282" [id=2288, type=slice];
+"2289 fill__82" [id=2289, type=fill_];
+"2290 lift_fresh_copy_83" [id=2290, type=lift_fresh_copy];
+"2291 slice_283" [id=2291, type=slice];
+"2292 slice_284" [id=2292, type=slice];
+"2293 fill__83" [id=2293, type=fill_];
+"2294 lift_fresh_copy_84" [id=2294, type=lift_fresh_copy];
+"2295 slice_285" [id=2295, type=slice];
+"2296 slice_286" [id=2296, type=slice];
+"2297 fill__84" [id=2297, type=fill_];
+"2298 lift_fresh_copy_85" [id=2298, type=lift_fresh_copy];
+"2299 slice_287" [id=2299, type=slice];
+"2300 slice_288" [id=2300, type=slice];
+"2301 fill__85" [id=2301, type=fill_];
+"2302 lift_fresh_copy_86" [id=2302, type=lift_fresh_copy];
+"2303 slice_289" [id=2303, type=slice];
+"2304 slice_290" [id=2304, type=slice];
+"2305 fill__86" [id=2305, type=fill_];
+"2306 lift_fresh_copy_87" [id=2306, type=lift_fresh_copy];
+"2307 slice_291" [id=2307, type=slice];
+"2308 slice_292" [id=2308, type=slice];
+"2309 fill__87" [id=2309, type=fill_];
+"2310 lift_fresh_copy_88" [id=2310, type=lift_fresh_copy];
+"2311 slice_293" [id=2311, type=slice];
+"2312 slice_294" [id=2312, type=slice];
+"2313 fill__88" [id=2313, type=fill_];
+"2314 lift_fresh_copy_89" [id=2314, type=lift_fresh_copy];
+"2315 slice_295" [id=2315, type=slice];
+"2316 slice_296" [id=2316, type=slice];
+"2317 fill__89" [id=2317, type=fill_];
+"2318 view_106" [id=2318, type=view];
+"2319 permute_89" [id=2319, type=permute];
+"2320 reshape_87" [id=2320, type=reshape];
+"2321 unsqueeze_56" [id=2321, type=unsqueeze];
+"2322 unsqueeze_57" [id=2322, type=unsqueeze];
+"2323 sub_9" [id=2323, type=sub];
+"2324 ne_9" [id=2324, type=ne];
+"2325 masked_fill_18" [id=2325, type=masked_fill];
+"2326 eq_9" [id=2326, type=eq];
+"2327 masked_fill_19" [id=2327, type=masked_fill];
+"2328 view_107" [id=2328, type=view];
+"2329 unsqueeze_58" [id=2329, type=unsqueeze];
+"2330 unsqueeze_59" [id=2330, type=unsqueeze];
+"2331 add_67" [id=2331, type=add];
+"2332 view_108" [id=2332, type=view];
+"2333 softmax_19" [id=2333, type=softmax];
+"2334 dropout_76" [id=2334, type=dropout];
+"2335 matmul_39" [id=2335, type=matmul];
+"2336 transpose_39" [id=2336, type=transpose];
+"2337 reshape_88" [id=2337, type=reshape];
+"2338 linear_119" [id=2338, type=linear];
+"2339 dropout_77" [id=2339, type=dropout];
+"2340 view_109" [id=2340, type=view];
+"2341 permute_90" [id=2341, type=permute];
+"2342 reshape_89" [id=2342, type=reshape];
+"2343 roll_19" [id=2343, type=roll];
+"2344 slice_297" [id=2344, type=slice];
+"2345 slice_298" [id=2345, type=slice];
+"2346 slice_299" [id=2346, type=slice];
+"2347 slice_300" [id=2347, type=slice];
+"2348 contiguous_37" [id=2348, type=contiguous];
+"2349 layer_norm_41" [id=2349, type=layer_norm];
+"2350 add_68" [id=2350, type=add];
+"2351 linear_120" [id=2351, type=linear];
+"2352 gelu_19" [id=2352, type=gelu];
+"2353 dropout_78" [id=2353, type=dropout];
+"2354 linear_121" [id=2354, type=linear];
+"2355 dropout_79" [id=2355, type=dropout];
+"2356 layer_norm_42" [id=2356, type=layer_norm];
+"2357 add_69" [id=2357, type=add];
+"2358 linear_122" [id=2358, type=linear];
+"2359 relu__20" [id=2359, type=relu_];
+"2360 linear_123" [id=2360, type=linear];
+"2361 view_110" [id=2361, type=view];
+"2362 index_20" [id=2362, type=index];
+"2363 view_111" [id=2363, type=view];
+"2364 permute_91" [id=2364, type=permute];
+"2365 contiguous_38" [id=2365, type=contiguous];
+"2366 unsqueeze_60" [id=2366, type=unsqueeze];
+"2367 sigmoid_20" [id=2367, type=sigmoid];
+"2368 mul_40" [id=2368, type=mul];
+"2369 pad_22" [id=2369, type=pad];
+"2370 view_112" [id=2370, type=view];
+"2371 permute_92" [id=2371, type=permute];
+"2372 reshape_90" [id=2372, type=reshape];
+"2373 clone_20" [id=2373, type=clone];
+"2374 slice_301" [id=2374, type=slice];
+"2375 zero__20" [id=2375, type=zero_];
+"2376 linear_124" [id=2376, type=linear];
+"2377 reshape_91" [id=2377, type=reshape];
+"2378 permute_93" [id=2378, type=permute];
+"2379 select_60" [id=2379, type=select];
+"2380 select_61" [id=2380, type=select];
+"2381 select_62" [id=2381, type=select];
+"2382 linalg_vector_norm_40" [id=2382, type=linalg_vector_norm];
+"2383 clamp_min_40" [id=2383, type=clamp_min];
+"2384 expand_as_40" [id=2384, type=expand_as];
+"2385 div_40" [id=2385, type=div];
+"2386 linalg_vector_norm_41" [id=2386, type=linalg_vector_norm];
+"2387 clamp_min_41" [id=2387, type=clamp_min];
+"2388 expand_as_41" [id=2388, type=expand_as];
+"2389 div_41" [id=2389, type=div];
+"2390 transpose_40" [id=2390, type=transpose];
+"2391 matmul_40" [id=2391, type=matmul];
+"2392 clamp_20" [id=2392, type=clamp];
+"2393 exp_20" [id=2393, type=exp];
+"2394 mul_41" [id=2394, type=mul];
+"2395 add_70" [id=2395, type=add];
+"2396 softmax_20" [id=2396, type=softmax];
+"2397 dropout_80" [id=2397, type=dropout];
+"2398 matmul_41" [id=2398, type=matmul];
+"2399 transpose_41" [id=2399, type=transpose];
+"2400 reshape_92" [id=2400, type=reshape];
+"2401 linear_125" [id=2401, type=linear];
+"2402 dropout_81" [id=2402, type=dropout];
+"2403 view_113" [id=2403, type=view];
+"2404 permute_94" [id=2404, type=permute];
+"2405 reshape_93" [id=2405, type=reshape];
+"2406 slice_302" [id=2406, type=slice];
+"2407 slice_303" [id=2407, type=slice];
+"2408 slice_304" [id=2408, type=slice];
+"2409 slice_305" [id=2409, type=slice];
+"2410 contiguous_39" [id=2410, type=contiguous];
+"2411 layer_norm_43" [id=2411, type=layer_norm];
+"2412 add_71" [id=2412, type=add];
+"2413 linear_126" [id=2413, type=linear];
+"2414 gelu_20" [id=2414, type=gelu];
+"2415 dropout_82" [id=2415, type=dropout];
+"2416 linear_127" [id=2416, type=linear];
+"2417 dropout_83" [id=2417, type=dropout];
+"2418 layer_norm_44" [id=2418, type=layer_norm];
+"2419 add_72" [id=2419, type=add];
+"2420 linear_128" [id=2420, type=linear];
+"2421 relu__21" [id=2421, type=relu_];
+"2422 linear_129" [id=2422, type=linear];
+"2423 view_114" [id=2423, type=view];
+"2424 index_21" [id=2424, type=index];
+"2425 view_115" [id=2425, type=view];
+"2426 permute_95" [id=2426, type=permute];
+"2427 contiguous_40" [id=2427, type=contiguous];
+"2428 unsqueeze_61" [id=2428, type=unsqueeze];
+"2429 sigmoid_21" [id=2429, type=sigmoid];
+"2430 mul_42" [id=2430, type=mul];
+"2431 pad_23" [id=2431, type=pad];
+"2432 roll_20" [id=2432, type=roll];
+"2433 view_116" [id=2433, type=view];
+"2434 permute_96" [id=2434, type=permute];
+"2435 reshape_94" [id=2435, type=reshape];
+"2436 clone_21" [id=2436, type=clone];
+"2437 slice_306" [id=2437, type=slice];
+"2438 zero__21" [id=2438, type=zero_];
+"2439 linear_130" [id=2439, type=linear];
+"2440 reshape_95" [id=2440, type=reshape];
+"2441 permute_97" [id=2441, type=permute];
+"2442 select_63" [id=2442, type=select];
+"2443 select_64" [id=2443, type=select];
+"2444 select_65" [id=2444, type=select];
+"2445 linalg_vector_norm_42" [id=2445, type=linalg_vector_norm];
+"2446 clamp_min_42" [id=2446, type=clamp_min];
+"2447 expand_as_42" [id=2447, type=expand_as];
+"2448 div_42" [id=2448, type=div];
+"2449 linalg_vector_norm_43" [id=2449, type=linalg_vector_norm];
+"2450 clamp_min_43" [id=2450, type=clamp_min];
+"2451 expand_as_43" [id=2451, type=expand_as];
+"2452 div_43" [id=2452, type=div];
+"2453 transpose_42" [id=2453, type=transpose];
+"2454 matmul_42" [id=2454, type=matmul];
+"2455 clamp_21" [id=2455, type=clamp];
+"2456 exp_21" [id=2456, type=exp];
+"2457 mul_43" [id=2457, type=mul];
+"2458 add_73" [id=2458, type=add];
+"2459 new_zeros_10" [id=2459, type=new_zeros];
+"2460 lift_fresh_copy_90" [id=2460, type=lift_fresh_copy];
+"2461 slice_307" [id=2461, type=slice];
+"2462 slice_308" [id=2462, type=slice];
+"2463 fill__90" [id=2463, type=fill_];
+"2464 lift_fresh_copy_91" [id=2464, type=lift_fresh_copy];
+"2465 slice_309" [id=2465, type=slice];
+"2466 slice_310" [id=2466, type=slice];
+"2467 fill__91" [id=2467, type=fill_];
+"2468 lift_fresh_copy_92" [id=2468, type=lift_fresh_copy];
+"2469 slice_311" [id=2469, type=slice];
+"2470 slice_312" [id=2470, type=slice];
+"2471 fill__92" [id=2471, type=fill_];
+"2472 lift_fresh_copy_93" [id=2472, type=lift_fresh_copy];
+"2473 slice_313" [id=2473, type=slice];
+"2474 slice_314" [id=2474, type=slice];
+"2475 fill__93" [id=2475, type=fill_];
+"2476 lift_fresh_copy_94" [id=2476, type=lift_fresh_copy];
+"2477 slice_315" [id=2477, type=slice];
+"2478 slice_316" [id=2478, type=slice];
+"2479 fill__94" [id=2479, type=fill_];
+"2480 lift_fresh_copy_95" [id=2480, type=lift_fresh_copy];
+"2481 slice_317" [id=2481, type=slice];
+"2482 slice_318" [id=2482, type=slice];
+"2483 fill__95" [id=2483, type=fill_];
+"2484 lift_fresh_copy_96" [id=2484, type=lift_fresh_copy];
+"2485 slice_319" [id=2485, type=slice];
+"2486 slice_320" [id=2486, type=slice];
+"2487 fill__96" [id=2487, type=fill_];
+"2488 lift_fresh_copy_97" [id=2488, type=lift_fresh_copy];
+"2489 slice_321" [id=2489, type=slice];
+"2490 slice_322" [id=2490, type=slice];
+"2491 fill__97" [id=2491, type=fill_];
+"2492 lift_fresh_copy_98" [id=2492, type=lift_fresh_copy];
+"2493 slice_323" [id=2493, type=slice];
+"2494 slice_324" [id=2494, type=slice];
+"2495 fill__98" [id=2495, type=fill_];
+"2496 view_117" [id=2496, type=view];
+"2497 permute_98" [id=2497, type=permute];
+"2498 reshape_96" [id=2498, type=reshape];
+"2499 unsqueeze_62" [id=2499, type=unsqueeze];
+"2500 unsqueeze_63" [id=2500, type=unsqueeze];
+"2501 sub_10" [id=2501, type=sub];
+"2502 ne_10" [id=2502, type=ne];
+"2503 masked_fill_20" [id=2503, type=masked_fill];
+"2504 eq_10" [id=2504, type=eq];
+"2505 masked_fill_21" [id=2505, type=masked_fill];
+"2506 view_118" [id=2506, type=view];
+"2507 unsqueeze_64" [id=2507, type=unsqueeze];
+"2508 unsqueeze_65" [id=2508, type=unsqueeze];
+"2509 add_74" [id=2509, type=add];
+"2510 view_119" [id=2510, type=view];
+"2511 softmax_21" [id=2511, type=softmax];
+"2512 dropout_84" [id=2512, type=dropout];
+"2513 matmul_43" [id=2513, type=matmul];
+"2514 transpose_43" [id=2514, type=transpose];
+"2515 reshape_97" [id=2515, type=reshape];
+"2516 linear_131" [id=2516, type=linear];
+"2517 dropout_85" [id=2517, type=dropout];
+"2518 view_120" [id=2518, type=view];
+"2519 permute_99" [id=2519, type=permute];
+"2520 reshape_98" [id=2520, type=reshape];
+"2521 roll_21" [id=2521, type=roll];
+"2522 slice_325" [id=2522, type=slice];
+"2523 slice_326" [id=2523, type=slice];
+"2524 slice_327" [id=2524, type=slice];
+"2525 slice_328" [id=2525, type=slice];
+"2526 contiguous_41" [id=2526, type=contiguous];
+"2527 layer_norm_45" [id=2527, type=layer_norm];
+"2528 add_75" [id=2528, type=add];
+"2529 linear_132" [id=2529, type=linear];
+"2530 gelu_21" [id=2530, type=gelu];
+"2531 dropout_86" [id=2531, type=dropout];
+"2532 linear_133" [id=2532, type=linear];
+"2533 dropout_87" [id=2533, type=dropout];
+"2534 layer_norm_46" [id=2534, type=layer_norm];
+"2535 add_76" [id=2535, type=add];
+"2536 pad_24" [id=2536, type=pad];
+"2537 slice_329" [id=2537, type=slice];
+"2538 slice_330" [id=2538, type=slice];
+"2539 slice_331" [id=2539, type=slice];
+"2540 slice_332" [id=2540, type=slice];
+"2541 slice_333" [id=2541, type=slice];
+"2542 slice_334" [id=2542, type=slice];
+"2543 slice_335" [id=2543, type=slice];
+"2544 slice_336" [id=2544, type=slice];
+"2545 slice_337" [id=2545, type=slice];
+"2546 slice_338" [id=2546, type=slice];
+"2547 slice_339" [id=2547, type=slice];
+"2548 slice_340" [id=2548, type=slice];
+"2549 cat_2" [id=2549, type=cat];
+"2550 linear_134" [id=2550, type=linear];
+"2551 layer_norm_47" [id=2551, type=layer_norm];
+"2552 linear_135" [id=2552, type=linear];
+"2553 relu__22" [id=2553, type=relu_];
+"2554 linear_136" [id=2554, type=linear];
+"2555 view_121" [id=2555, type=view];
+"2556 index_22" [id=2556, type=index];
+"2557 view_122" [id=2557, type=view];
+"2558 permute_100" [id=2558, type=permute];
+"2559 contiguous_42" [id=2559, type=contiguous];
+"2560 unsqueeze_66" [id=2560, type=unsqueeze];
+"2561 sigmoid_22" [id=2561, type=sigmoid];
+"2562 mul_44" [id=2562, type=mul];
+"2563 pad_25" [id=2563, type=pad];
+"2564 view_123" [id=2564, type=view];
+"2565 permute_101" [id=2565, type=permute];
+"2566 reshape_99" [id=2566, type=reshape];
+"2567 clone_22" [id=2567, type=clone];
+"2568 slice_341" [id=2568, type=slice];
+"2569 zero__22" [id=2569, type=zero_];
+"2570 linear_137" [id=2570, type=linear];
+"2571 reshape_100" [id=2571, type=reshape];
+"2572 permute_102" [id=2572, type=permute];
+"2573 select_66" [id=2573, type=select];
+"2574 select_67" [id=2574, type=select];
+"2575 select_68" [id=2575, type=select];
+"2576 linalg_vector_norm_44" [id=2576, type=linalg_vector_norm];
+"2577 clamp_min_44" [id=2577, type=clamp_min];
+"2578 expand_as_44" [id=2578, type=expand_as];
+"2579 div_44" [id=2579, type=div];
+"2580 linalg_vector_norm_45" [id=2580, type=linalg_vector_norm];
+"2581 clamp_min_45" [id=2581, type=clamp_min];
+"2582 expand_as_45" [id=2582, type=expand_as];
+"2583 div_45" [id=2583, type=div];
+"2584 transpose_44" [id=2584, type=transpose];
+"2585 matmul_44" [id=2585, type=matmul];
+"2586 clamp_22" [id=2586, type=clamp];
+"2587 exp_22" [id=2587, type=exp];
+"2588 mul_45" [id=2588, type=mul];
+"2589 add_77" [id=2589, type=add];
+"2590 softmax_22" [id=2590, type=softmax];
+"2591 dropout_88" [id=2591, type=dropout];
+"2592 matmul_45" [id=2592, type=matmul];
+"2593 transpose_45" [id=2593, type=transpose];
+"2594 reshape_101" [id=2594, type=reshape];
+"2595 linear_138" [id=2595, type=linear];
+"2596 dropout_89" [id=2596, type=dropout];
+"2597 view_124" [id=2597, type=view];
+"2598 permute_103" [id=2598, type=permute];
+"2599 reshape_102" [id=2599, type=reshape];
+"2600 slice_342" [id=2600, type=slice];
+"2601 slice_343" [id=2601, type=slice];
+"2602 slice_344" [id=2602, type=slice];
+"2603 slice_345" [id=2603, type=slice];
+"2604 contiguous_43" [id=2604, type=contiguous];
+"2605 layer_norm_48" [id=2605, type=layer_norm];
+"2606 add_78" [id=2606, type=add];
+"2607 linear_139" [id=2607, type=linear];
+"2608 gelu_22" [id=2608, type=gelu];
+"2609 dropout_90" [id=2609, type=dropout];
+"2610 linear_140" [id=2610, type=linear];
+"2611 dropout_91" [id=2611, type=dropout];
+"2612 layer_norm_49" [id=2612, type=layer_norm];
+"2613 add_79" [id=2613, type=add];
+"2614 linear_141" [id=2614, type=linear];
+"2615 relu__23" [id=2615, type=relu_];
+"2616 linear_142" [id=2616, type=linear];
+"2617 view_125" [id=2617, type=view];
+"2618 index_23" [id=2618, type=index];
+"2619 view_126" [id=2619, type=view];
+"2620 permute_104" [id=2620, type=permute];
+"2621 contiguous_44" [id=2621, type=contiguous];
+"2622 unsqueeze_67" [id=2622, type=unsqueeze];
+"2623 sigmoid_23" [id=2623, type=sigmoid];
+"2624 mul_46" [id=2624, type=mul];
+"2625 pad_26" [id=2625, type=pad];
+"2626 view_127" [id=2626, type=view];
+"2627 permute_105" [id=2627, type=permute];
+"2628 reshape_103" [id=2628, type=reshape];
+"2629 clone_23" [id=2629, type=clone];
+"2630 slice_346" [id=2630, type=slice];
+"2631 zero__23" [id=2631, type=zero_];
+"2632 linear_143" [id=2632, type=linear];
+"2633 reshape_104" [id=2633, type=reshape];
+"2634 permute_106" [id=2634, type=permute];
+"2635 select_69" [id=2635, type=select];
+"2636 select_70" [id=2636, type=select];
+"2637 select_71" [id=2637, type=select];
+"2638 linalg_vector_norm_46" [id=2638, type=linalg_vector_norm];
+"2639 clamp_min_46" [id=2639, type=clamp_min];
+"2640 expand_as_46" [id=2640, type=expand_as];
+"2641 div_46" [id=2641, type=div];
+"2642 linalg_vector_norm_47" [id=2642, type=linalg_vector_norm];
+"2643 clamp_min_47" [id=2643, type=clamp_min];
+"2644 expand_as_47" [id=2644, type=expand_as];
+"2645 div_47" [id=2645, type=div];
+"2646 transpose_46" [id=2646, type=transpose];
+"2647 matmul_46" [id=2647, type=matmul];
+"2648 clamp_23" [id=2648, type=clamp];
+"2649 exp_23" [id=2649, type=exp];
+"2650 mul_47" [id=2650, type=mul];
+"2651 add_80" [id=2651, type=add];
+"2652 softmax_23" [id=2652, type=softmax];
+"2653 dropout_92" [id=2653, type=dropout];
+"2654 matmul_47" [id=2654, type=matmul];
+"2655 transpose_47" [id=2655, type=transpose];
+"2656 reshape_105" [id=2656, type=reshape];
+"2657 linear_144" [id=2657, type=linear];
+"2658 dropout_93" [id=2658, type=dropout];
+"2659 view_128" [id=2659, type=view];
+"2660 permute_107" [id=2660, type=permute];
+"2661 reshape_106" [id=2661, type=reshape];
+"2662 slice_347" [id=2662, type=slice];
+"2663 slice_348" [id=2663, type=slice];
+"2664 slice_349" [id=2664, type=slice];
+"2665 slice_350" [id=2665, type=slice];
+"2666 contiguous_45" [id=2666, type=contiguous];
+"2667 layer_norm_50" [id=2667, type=layer_norm];
+"2668 add_81" [id=2668, type=add];
+"2669 linear_145" [id=2669, type=linear];
+"2670 gelu_23" [id=2670, type=gelu];
+"2671 dropout_94" [id=2671, type=dropout];
+"2672 linear_146" [id=2672, type=linear];
+"2673 dropout_95" [id=2673, type=dropout];
+"2674 layer_norm_51" [id=2674, type=layer_norm];
+"2675 add_82" [id=2675, type=add];
+"2676 layer_norm_52" [id=2676, type=layer_norm];
+"2677 permute_108" [id=2677, type=permute];
+"2678 adaptive_avg_pool2d" [id=2678, type=adaptive_avg_pool2d];
+"2679 flatten" [id=2679, type=flatten];
 "2680 linear_147" [id=2680, type=linear];
-"2681 output" [id=2681, type=output];
-"0 arg0_1" -> "3 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"1 _param_constant0" -> "3 conv2d"  [label="(96, 3, 4, 4)", style=solid];
-"2 _param_constant1" -> "3 conv2d"  [label="(96,)", style=solid];
-"3 conv2d" -> "4 permute"  [label="(1, 96, 56, 56)", style=solid];
-"4 permute" -> "7 layer_norm"  [label="(1, 56, 56, 96)", style=solid];
-"5 _param_constant2" -> "7 layer_norm"  [label="(96,)", style=solid];
-"6 _param_constant3" -> "7 layer_norm"  [label="(96,)", style=solid];
-"7 layer_norm" -> "24 pad"  [label="(1, 56, 56, 96)", style=solid];
-"7 layer_norm" -> "71 add_1"  [label="(1, 56, 56, 96)", style=solid];
-"8 _tensor_constant0" -> "11 linear"  [label="(1, 15, 15, 2)", style=solid];
-"9 _param_constant4" -> "11 linear"  [label="(512, 2)", style=solid];
-"10 _param_constant5" -> "11 linear"  [label="(512,)", style=solid];
-"11 linear" -> "12 relu_"  [label="(1, 15, 15, 512)", style=solid];
-"12 relu_" -> "14 linear_1"  [label="(1, 15, 15, 512)", style=solid];
-"13 _param_constant6" -> "14 linear_1"  [label="(3, 512)", style=solid];
-"14 linear_1" -> "15 view"  [label="(1, 15, 15, 3)", style=solid];
-"15 view" -> "17 index"  [label="(225, 3)", style=solid];
-"16 _tensor_constant1" -> "17 index"  [label="(4096,)", style=solid];
-"17 index" -> "18 view_1"  [label="(4096, 3)", style=solid];
-"18 view_1" -> "19 permute_1"  [label="(64, 64, 3)", style=solid];
-"19 permute_1" -> "20 contiguous"  [label="(3, 64, 64)", style=solid];
-"20 contiguous" -> "21 unsqueeze"  [label="(3, 64, 64)", style=solid];
-"21 unsqueeze" -> "22 sigmoid"  [label="(1, 3, 64, 64)", style=solid];
-"22 sigmoid" -> "23 mul"  [label="(1, 3, 64, 64)", style=solid];
-"23 mul" -> "53 add"  [label="(1, 3, 64, 64)", style=solid];
-"24 pad" -> "25 view_2"  [label="(1, 56, 56, 96)", style=solid];
-"25 view_2" -> "26 permute_2"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"26 permute_2" -> "27 reshape"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"27 reshape" -> "33 linear_2"  [label="(49, 64, 96)", style=solid];
-"28 _param_constant7" -> "29 clone"  [label="(288,)", style=solid];
-"29 clone" -> "30 slice_1"  [label="(288,)", style=solid];
-"29 clone" -> "33 linear_2"  [label="(288,)", style=solid];
-"30 slice_1" -> "31 zero_"  [label="(96,)", style=solid];
-"32 _param_constant8" -> "33 linear_2"  [label="(288, 96)", style=solid];
-"33 linear_2" -> "34 reshape_1"  [label="(49, 64, 288)", style=solid];
-"34 reshape_1" -> "35 permute_3"  [label="(49, 64, 3, 3, 32)", style=solid];
-"35 permute_3" -> "36 select"  [label="(3, 49, 3, 64, 32)", style=solid];
-"35 permute_3" -> "37 select_1"  [label="(3, 49, 3, 64, 32)", style=solid];
-"35 permute_3" -> "38 select_2"  [label="(3, 49, 3, 64, 32)", style=solid];
-"36 select" -> "39 linalg_vector_norm"  [label="(49, 3, 64, 32)", style=solid];
-"36 select" -> "41 expand_as"  [label="(49, 3, 64, 32)", style=solid];
-"36 select" -> "42 div"  [label="(49, 3, 64, 32)", style=solid];
-"37 select_1" -> "43 linalg_vector_norm_1"  [label="(49, 3, 64, 32)", style=solid];
-"37 select_1" -> "45 expand_as_1"  [label="(49, 3, 64, 32)", style=solid];
-"37 select_1" -> "46 div_1"  [label="(49, 3, 64, 32)", style=solid];
-"38 select_2" -> "56 matmul_1"  [label="(49, 3, 64, 32)", style=solid];
-"39 linalg_vector_norm" -> "40 clamp_min"  [label="(49, 3, 64, 1)", style=solid];
-"40 clamp_min" -> "41 expand_as"  [label="(49, 3, 64, 1)", style=solid];
-"41 expand_as" -> "42 div"  [label="(49, 3, 64, 32)", style=solid];
-"42 div" -> "48 matmul"  [label="(49, 3, 64, 32)", style=solid];
-"43 linalg_vector_norm_1" -> "44 clamp_min_1"  [label="(49, 3, 64, 1)", style=solid];
-"44 clamp_min_1" -> "45 expand_as_1"  [label="(49, 3, 64, 1)", style=solid];
-"45 expand_as_1" -> "46 div_1"  [label="(49, 3, 64, 32)", style=solid];
-"46 div_1" -> "47 transpose"  [label="(49, 3, 64, 32)", style=solid];
-"47 transpose" -> "48 matmul"  [label="(49, 3, 32, 64)", style=solid];
-"48 matmul" -> "52 mul_1"  [label="(49, 3, 64, 64)", style=solid];
-"49 _param_constant9" -> "50 clamp"  [label="(3, 1, 1)", style=solid];
-"50 clamp" -> "51 exp"  [label="(3, 1, 1)", style=solid];
-"51 exp" -> "52 mul_1"  [label="(3, 1, 1)", style=solid];
-"52 mul_1" -> "53 add"  [label="(49, 3, 64, 64)", style=solid];
-"53 add" -> "54 softmax"  [label="(49, 3, 64, 64)", style=solid];
-"54 softmax" -> "55 dropout"  [label="(49, 3, 64, 64)", style=solid];
-"55 dropout" -> "56 matmul_1"  [label="(49, 3, 64, 64)", style=solid];
-"56 matmul_1" -> "57 transpose_1"  [label="(49, 3, 64, 32)", style=solid];
-"57 transpose_1" -> "58 reshape_2"  [label="(49, 64, 3, 32)", style=solid];
-"58 reshape_2" -> "61 linear_3"  [label="(49, 64, 96)", style=solid];
-"59 _param_constant10" -> "61 linear_3"  [label="(96, 96)", style=solid];
-"60 _param_constant11" -> "61 linear_3"  [label="(96,)", style=solid];
-"61 linear_3" -> "62 dropout_1"  [label="(49, 64, 96)", style=solid];
-"62 dropout_1" -> "63 view_3"  [label="(49, 64, 96)", style=solid];
-"63 view_3" -> "64 permute_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"64 permute_4" -> "65 reshape_3"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"65 reshape_3" -> "66 slice_2"  [label="(1, 56, 56, 96)", style=solid];
-"66 slice_2" -> "67 slice_3"  [label="(1, 56, 56, 96)", style=solid];
-"67 slice_3" -> "70 layer_norm_1"  [label="(1, 56, 56, 96)", style=solid];
-"68 _param_constant12" -> "70 layer_norm_1"  [label="(96,)", style=solid];
-"69 _param_constant13" -> "70 layer_norm_1"  [label="(96,)", style=solid];
-"70 layer_norm_1" -> "71 add_1"  [label="(1, 56, 56, 96)", style=solid];
-"71 add_1" -> "74 linear_4"  [label="(1, 56, 56, 96)", style=solid];
-"71 add_1" -> "84 add_2"  [label="(1, 56, 56, 96)", style=solid];
-"72 _param_constant14" -> "74 linear_4"  [label="(384, 96)", style=solid];
-"73 _param_constant15" -> "74 linear_4"  [label="(384,)", style=solid];
-"74 linear_4" -> "75 gelu"  [label="(1, 56, 56, 384)", style=solid];
-"75 gelu" -> "76 dropout_2"  [label="(1, 56, 56, 384)", style=solid];
-"76 dropout_2" -> "79 linear_5"  [label="(1, 56, 56, 384)", style=solid];
-"77 _param_constant16" -> "79 linear_5"  [label="(96, 384)", style=solid];
-"78 _param_constant17" -> "79 linear_5"  [label="(96,)", style=solid];
-"79 linear_5" -> "80 dropout_3"  [label="(1, 56, 56, 96)", style=solid];
-"80 dropout_3" -> "83 layer_norm_2"  [label="(1, 56, 56, 96)", style=solid];
-"81 _param_constant18" -> "83 layer_norm_2"  [label="(96,)", style=solid];
-"82 _param_constant19" -> "83 layer_norm_2"  [label="(96,)", style=solid];
-"83 layer_norm_2" -> "84 add_2"  [label="(1, 56, 56, 96)", style=solid];
-"84 add_2" -> "101 pad_1"  [label="(1, 56, 56, 96)", style=solid];
-"84 add_2" -> "211 add_5"  [label="(1, 56, 56, 96)", style=solid];
-"85 _tensor_constant2" -> "88 linear_6"  [label="(1, 15, 15, 2)", style=solid];
-"86 _param_constant20" -> "88 linear_6"  [label="(512, 2)", style=solid];
-"87 _param_constant21" -> "88 linear_6"  [label="(512,)", style=solid];
-"88 linear_6" -> "89 relu__1"  [label="(1, 15, 15, 512)", style=solid];
-"89 relu__1" -> "91 linear_7"  [label="(1, 15, 15, 512)", style=solid];
-"90 _param_constant22" -> "91 linear_7"  [label="(3, 512)", style=solid];
-"91 linear_7" -> "92 view_4"  [label="(1, 15, 15, 3)", style=solid];
-"92 view_4" -> "94 index_1"  [label="(225, 3)", style=solid];
-"93 _tensor_constant3" -> "94 index_1"  [label="(4096,)", style=solid];
-"94 index_1" -> "95 view_5"  [label="(4096, 3)", style=solid];
-"95 view_5" -> "96 permute_5"  [label="(64, 64, 3)", style=solid];
-"96 permute_5" -> "97 contiguous_1"  [label="(3, 64, 64)", style=solid];
-"97 contiguous_1" -> "98 unsqueeze_1"  [label="(3, 64, 64)", style=solid];
-"98 unsqueeze_1" -> "99 sigmoid_1"  [label="(1, 3, 64, 64)", style=solid];
-"99 sigmoid_1" -> "100 mul_2"  [label="(1, 3, 64, 64)", style=solid];
-"100 mul_2" -> "131 add_3"  [label="(1, 3, 64, 64)", style=solid];
-"101 pad_1" -> "102 roll"  [label="(1, 56, 56, 96)", style=solid];
-"102 roll" -> "103 view_6"  [label="(1, 56, 56, 96)", style=solid];
-"103 view_6" -> "104 permute_6"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"104 permute_6" -> "105 reshape_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"105 reshape_4" -> "111 linear_8"  [label="(49, 64, 96)", style=solid];
-"105 reshape_4" -> "132 new_zeros"  [label="(49, 64, 96)", style=solid];
-"106 _param_constant23" -> "107 clone_1"  [label="(288,)", style=solid];
-"107 clone_1" -> "108 slice_4"  [label="(288,)", style=solid];
-"107 clone_1" -> "111 linear_8"  [label="(288,)", style=solid];
-"108 slice_4" -> "109 zero__1"  [label="(96,)", style=solid];
-"110 _param_constant24" -> "111 linear_8"  [label="(288, 96)", style=solid];
-"111 linear_8" -> "112 reshape_5"  [label="(49, 64, 288)", style=solid];
-"112 reshape_5" -> "113 permute_7"  [label="(49, 64, 3, 3, 32)", style=solid];
-"113 permute_7" -> "114 select_3"  [label="(3, 49, 3, 64, 32)", style=solid];
-"113 permute_7" -> "115 select_4"  [label="(3, 49, 3, 64, 32)", style=solid];
-"113 permute_7" -> "116 select_5"  [label="(3, 49, 3, 64, 32)", style=solid];
-"114 select_3" -> "117 linalg_vector_norm_2"  [label="(49, 3, 64, 32)", style=solid];
-"114 select_3" -> "119 expand_as_2"  [label="(49, 3, 64, 32)", style=solid];
-"114 select_3" -> "120 div_2"  [label="(49, 3, 64, 32)", style=solid];
-"115 select_4" -> "121 linalg_vector_norm_3"  [label="(49, 3, 64, 32)", style=solid];
-"115 select_4" -> "123 expand_as_3"  [label="(49, 3, 64, 32)", style=solid];
-"115 select_4" -> "124 div_3"  [label="(49, 3, 64, 32)", style=solid];
-"116 select_5" -> "195 matmul_3"  [label="(49, 3, 64, 32)", style=solid];
-"117 linalg_vector_norm_2" -> "118 clamp_min_2"  [label="(49, 3, 64, 1)", style=solid];
-"118 clamp_min_2" -> "119 expand_as_2"  [label="(49, 3, 64, 1)", style=solid];
-"119 expand_as_2" -> "120 div_2"  [label="(49, 3, 64, 32)", style=solid];
-"120 div_2" -> "126 matmul_2"  [label="(49, 3, 64, 32)", style=solid];
-"121 linalg_vector_norm_3" -> "122 clamp_min_3"  [label="(49, 3, 64, 1)", style=solid];
-"122 clamp_min_3" -> "123 expand_as_3"  [label="(49, 3, 64, 1)", style=solid];
-"123 expand_as_3" -> "124 div_3"  [label="(49, 3, 64, 32)", style=solid];
-"124 div_3" -> "125 transpose_2"  [label="(49, 3, 64, 32)", style=solid];
-"125 transpose_2" -> "126 matmul_2"  [label="(49, 3, 32, 64)", style=solid];
-"126 matmul_2" -> "130 mul_3"  [label="(49, 3, 64, 64)", style=solid];
-"127 _param_constant25" -> "128 clamp_1"  [label="(3, 1, 1)", style=solid];
-"128 clamp_1" -> "129 exp_1"  [label="(3, 1, 1)", style=solid];
-"129 exp_1" -> "130 mul_3"  [label="(3, 1, 1)", style=solid];
-"130 mul_3" -> "131 add_3"  [label="(49, 3, 64, 64)", style=solid];
-"131 add_3" -> "188 view_8"  [label="(49, 3, 64, 64)", style=solid];
-"132 new_zeros" -> "135 slice_5"  [label="(56, 56)", style=solid];
-"132 new_zeros" -> "140 slice_7"  [label="(56, 56)", style=solid];
-"132 new_zeros" -> "145 slice_9"  [label="(56, 56)", style=solid];
-"132 new_zeros" -> "150 slice_11"  [label="(56, 56)", style=solid];
-"132 new_zeros" -> "155 slice_13"  [label="(56, 56)", style=solid];
-"132 new_zeros" -> "160 slice_15"  [label="(56, 56)", style=solid];
-"132 new_zeros" -> "165 slice_17"  [label="(56, 56)", style=solid];
-"132 new_zeros" -> "170 slice_19"  [label="(56, 56)", style=solid];
-"132 new_zeros" -> "175 slice_21"  [label="(56, 56)", style=solid];
-"132 new_zeros" -> "178 view_7"  [label="(56, 56)", style=solid];
-"133 _tensor_constant4" -> "134 lift_fresh_copy"  [label="()", style=solid];
-"134 lift_fresh_copy" -> "137 fill_"  [label="()", style=solid];
-"135 slice_5" -> "136 slice_6"  [label="(48, 56)", style=solid];
-"136 slice_6" -> "137 fill_"  [label="(48, 48)", style=solid];
-"138 _tensor_constant5" -> "139 lift_fresh_copy_1"  [label="()", style=solid];
-"139 lift_fresh_copy_1" -> "142 fill__1"  [label="()", style=solid];
-"140 slice_7" -> "141 slice_8"  [label="(48, 56)", style=solid];
-"141 slice_8" -> "142 fill__1"  [label="(48, 4)", style=solid];
-"143 _tensor_constant6" -> "144 lift_fresh_copy_2"  [label="()", style=solid];
-"144 lift_fresh_copy_2" -> "147 fill__2"  [label="()", style=solid];
-"145 slice_9" -> "146 slice_10"  [label="(48, 56)", style=solid];
-"146 slice_10" -> "147 fill__2"  [label="(48, 4)", style=solid];
-"148 _tensor_constant7" -> "149 lift_fresh_copy_3"  [label="()", style=solid];
-"149 lift_fresh_copy_3" -> "152 fill__3"  [label="()", style=solid];
-"150 slice_11" -> "151 slice_12"  [label="(4, 56)", style=solid];
-"151 slice_12" -> "152 fill__3"  [label="(4, 48)", style=solid];
-"153 _tensor_constant8" -> "154 lift_fresh_copy_4"  [label="()", style=solid];
-"154 lift_fresh_copy_4" -> "157 fill__4"  [label="()", style=solid];
-"155 slice_13" -> "156 slice_14"  [label="(4, 56)", style=solid];
-"156 slice_14" -> "157 fill__4"  [label="(4, 4)", style=solid];
-"158 _tensor_constant9" -> "159 lift_fresh_copy_5"  [label="()", style=solid];
-"159 lift_fresh_copy_5" -> "162 fill__5"  [label="()", style=solid];
-"160 slice_15" -> "161 slice_16"  [label="(4, 56)", style=solid];
-"161 slice_16" -> "162 fill__5"  [label="(4, 4)", style=solid];
-"163 _tensor_constant10" -> "164 lift_fresh_copy_6"  [label="()", style=solid];
-"164 lift_fresh_copy_6" -> "167 fill__6"  [label="()", style=solid];
-"165 slice_17" -> "166 slice_18"  [label="(4, 56)", style=solid];
-"166 slice_18" -> "167 fill__6"  [label="(4, 48)", style=solid];
-"168 _tensor_constant11" -> "169 lift_fresh_copy_7"  [label="()", style=solid];
-"169 lift_fresh_copy_7" -> "172 fill__7"  [label="()", style=solid];
-"170 slice_19" -> "171 slice_20"  [label="(4, 56)", style=solid];
-"171 slice_20" -> "172 fill__7"  [label="(4, 4)", style=solid];
-"173 _tensor_constant12" -> "174 lift_fresh_copy_8"  [label="()", style=solid];
-"174 lift_fresh_copy_8" -> "177 fill__8"  [label="()", style=solid];
-"175 slice_21" -> "176 slice_22"  [label="(4, 56)", style=solid];
-"176 slice_22" -> "177 fill__8"  [label="(4, 4)", style=solid];
-"178 view_7" -> "179 permute_8"  [label="(7, 8, 7, 8)", style=solid];
-"179 permute_8" -> "180 reshape_6"  [label="(7, 7, 8, 8)", style=solid];
-"180 reshape_6" -> "181 unsqueeze_2"  [label="(49, 64)", style=solid];
-"180 reshape_6" -> "182 unsqueeze_3"  [label="(49, 64)", style=solid];
-"181 unsqueeze_2" -> "183 sub"  [label="(49, 1, 64)", style=solid];
-"182 unsqueeze_3" -> "183 sub"  [label="(49, 64, 1)", style=solid];
-"183 sub" -> "184 ne"  [label="(49, 64, 64)", style=solid];
-"183 sub" -> "185 masked_fill"  [label="(49, 64, 64)", style=solid];
-"183 sub" -> "186 eq"  [label="(49, 64, 64)", style=solid];
-"184 ne" -> "185 masked_fill"  [label="(49, 64, 64)", style=solid];
-"185 masked_fill" -> "187 masked_fill_1"  [label="(49, 64, 64)", style=solid];
-"186 eq" -> "187 masked_fill_1"  [label="(49, 64, 64)", style=solid];
-"187 masked_fill_1" -> "189 unsqueeze_4"  [label="(49, 64, 64)", style=solid];
-"188 view_8" -> "191 add_4"  [label="(1, 49, 3, 64, 64)", style=solid];
-"189 unsqueeze_4" -> "190 unsqueeze_5"  [label="(49, 1, 64, 64)", style=solid];
-"190 unsqueeze_5" -> "191 add_4"  [label="(1, 49, 1, 64, 64)", style=solid];
-"191 add_4" -> "192 view_9"  [label="(1, 49, 3, 64, 64)", style=solid];
-"192 view_9" -> "193 softmax_1"  [label="(49, 3, 64, 64)", style=solid];
-"193 softmax_1" -> "194 dropout_4"  [label="(49, 3, 64, 64)", style=solid];
-"194 dropout_4" -> "195 matmul_3"  [label="(49, 3, 64, 64)", style=solid];
-"195 matmul_3" -> "196 transpose_3"  [label="(49, 3, 64, 32)", style=solid];
-"196 transpose_3" -> "197 reshape_7"  [label="(49, 64, 3, 32)", style=solid];
-"197 reshape_7" -> "200 linear_9"  [label="(49, 64, 96)", style=solid];
-"198 _param_constant26" -> "200 linear_9"  [label="(96, 96)", style=solid];
-"199 _param_constant27" -> "200 linear_9"  [label="(96,)", style=solid];
-"200 linear_9" -> "201 dropout_5"  [label="(49, 64, 96)", style=solid];
-"201 dropout_5" -> "202 view_10"  [label="(49, 64, 96)", style=solid];
-"202 view_10" -> "203 permute_9"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
-"203 permute_9" -> "204 reshape_8"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
-"204 reshape_8" -> "205 roll_1"  [label="(1, 56, 56, 96)", style=solid];
-"205 roll_1" -> "206 slice_23"  [label="(1, 56, 56, 96)", style=solid];
-"206 slice_23" -> "207 slice_24"  [label="(1, 56, 56, 96)", style=solid];
-"207 slice_24" -> "210 layer_norm_3"  [label="(1, 56, 56, 96)", style=solid];
-"208 _param_constant28" -> "210 layer_norm_3"  [label="(96,)", style=solid];
-"209 _param_constant29" -> "210 layer_norm_3"  [label="(96,)", style=solid];
-"210 layer_norm_3" -> "211 add_5"  [label="(1, 56, 56, 96)", style=solid];
-"211 add_5" -> "214 linear_10"  [label="(1, 56, 56, 96)", style=solid];
-"211 add_5" -> "224 add_6"  [label="(1, 56, 56, 96)", style=solid];
-"212 _param_constant30" -> "214 linear_10"  [label="(384, 96)", style=solid];
-"213 _param_constant31" -> "214 linear_10"  [label="(384,)", style=solid];
-"214 linear_10" -> "215 gelu_1"  [label="(1, 56, 56, 384)", style=solid];
-"215 gelu_1" -> "216 dropout_6"  [label="(1, 56, 56, 384)", style=solid];
-"216 dropout_6" -> "219 linear_11"  [label="(1, 56, 56, 384)", style=solid];
-"217 _param_constant32" -> "219 linear_11"  [label="(96, 384)", style=solid];
-"218 _param_constant33" -> "219 linear_11"  [label="(96,)", style=solid];
-"219 linear_11" -> "220 dropout_7"  [label="(1, 56, 56, 96)", style=solid];
-"220 dropout_7" -> "223 layer_norm_4"  [label="(1, 56, 56, 96)", style=solid];
-"221 _param_constant34" -> "223 layer_norm_4"  [label="(96,)", style=solid];
-"222 _param_constant35" -> "223 layer_norm_4"  [label="(96,)", style=solid];
-"223 layer_norm_4" -> "224 add_6"  [label="(1, 56, 56, 96)", style=solid];
-"224 add_6" -> "225 pad_2"  [label="(1, 56, 56, 96)", style=solid];
-"225 pad_2" -> "226 slice_25"  [label="(1, 56, 56, 96)", style=solid];
-"225 pad_2" -> "229 slice_28"  [label="(1, 56, 56, 96)", style=solid];
-"225 pad_2" -> "232 slice_31"  [label="(1, 56, 56, 96)", style=solid];
-"225 pad_2" -> "235 slice_34"  [label="(1, 56, 56, 96)", style=solid];
-"226 slice_25" -> "227 slice_26"  [label="(1, 28, 56, 96)", style=solid];
-"227 slice_26" -> "228 slice_27"  [label="(1, 28, 28, 96)", style=solid];
-"228 slice_27" -> "238 cat"  [label="(1, 28, 28, 96)", style=solid];
-"229 slice_28" -> "230 slice_29"  [label="(1, 28, 56, 96)", style=solid];
-"230 slice_29" -> "231 slice_30"  [label="(1, 28, 28, 96)", style=solid];
-"231 slice_30" -> "238 cat"  [label="(1, 28, 28, 96)", style=solid];
-"232 slice_31" -> "233 slice_32"  [label="(1, 28, 56, 96)", style=solid];
-"233 slice_32" -> "234 slice_33"  [label="(1, 28, 28, 96)", style=solid];
-"234 slice_33" -> "238 cat"  [label="(1, 28, 28, 96)", style=solid];
-"235 slice_34" -> "236 slice_35"  [label="(1, 28, 56, 96)", style=solid];
-"236 slice_35" -> "237 slice_36"  [label="(1, 28, 28, 96)", style=solid];
-"237 slice_36" -> "238 cat"  [label="(1, 28, 28, 96)", style=solid];
-"238 cat" -> "240 linear_12"  [label="(1, 28, 28, 384)", style=solid];
-"239 _param_constant36" -> "240 linear_12"  [label="(192, 384)", style=solid];
-"240 linear_12" -> "243 layer_norm_5"  [label="(1, 28, 28, 192)", style=solid];
-"241 _param_constant37" -> "243 layer_norm_5"  [label="(192,)", style=solid];
-"242 _param_constant38" -> "243 layer_norm_5"  [label="(192,)", style=solid];
-"243 layer_norm_5" -> "260 pad_3"  [label="(1, 28, 28, 192)", style=solid];
-"243 layer_norm_5" -> "310 add_8"  [label="(1, 28, 28, 192)", style=solid];
-"244 _tensor_constant13" -> "247 linear_13"  [label="(1, 15, 15, 2)", style=solid];
-"245 _param_constant39" -> "247 linear_13"  [label="(512, 2)", style=solid];
-"246 _param_constant40" -> "247 linear_13"  [label="(512,)", style=solid];
-"247 linear_13" -> "248 relu__2"  [label="(1, 15, 15, 512)", style=solid];
-"248 relu__2" -> "250 linear_14"  [label="(1, 15, 15, 512)", style=solid];
-"249 _param_constant41" -> "250 linear_14"  [label="(6, 512)", style=solid];
-"250 linear_14" -> "251 view_11"  [label="(1, 15, 15, 6)", style=solid];
-"251 view_11" -> "253 index_2"  [label="(225, 6)", style=solid];
-"252 _tensor_constant14" -> "253 index_2"  [label="(4096,)", style=solid];
-"253 index_2" -> "254 view_12"  [label="(4096, 6)", style=solid];
-"254 view_12" -> "255 permute_10"  [label="(64, 64, 6)", style=solid];
-"255 permute_10" -> "256 contiguous_2"  [label="(6, 64, 64)", style=solid];
-"256 contiguous_2" -> "257 unsqueeze_6"  [label="(6, 64, 64)", style=solid];
-"257 unsqueeze_6" -> "258 sigmoid_2"  [label="(1, 6, 64, 64)", style=solid];
-"258 sigmoid_2" -> "259 mul_4"  [label="(1, 6, 64, 64)", style=solid];
-"259 mul_4" -> "289 add_7"  [label="(1, 6, 64, 64)", style=solid];
-"260 pad_3" -> "261 view_13"  [label="(1, 32, 32, 192)", style=solid];
-"261 view_13" -> "262 permute_11"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"262 permute_11" -> "263 reshape_9"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"263 reshape_9" -> "269 linear_15"  [label="(16, 64, 192)", style=solid];
-"264 _param_constant42" -> "265 clone_2"  [label="(576,)", style=solid];
-"265 clone_2" -> "266 slice_37"  [label="(576,)", style=solid];
-"265 clone_2" -> "269 linear_15"  [label="(576,)", style=solid];
-"266 slice_37" -> "267 zero__2"  [label="(192,)", style=solid];
-"268 _param_constant43" -> "269 linear_15"  [label="(576, 192)", style=solid];
-"269 linear_15" -> "270 reshape_10"  [label="(16, 64, 576)", style=solid];
-"270 reshape_10" -> "271 permute_12"  [label="(16, 64, 3, 6, 32)", style=solid];
-"271 permute_12" -> "272 select_6"  [label="(3, 16, 6, 64, 32)", style=solid];
-"271 permute_12" -> "273 select_7"  [label="(3, 16, 6, 64, 32)", style=solid];
-"271 permute_12" -> "274 select_8"  [label="(3, 16, 6, 64, 32)", style=solid];
-"272 select_6" -> "275 linalg_vector_norm_4"  [label="(16, 6, 64, 32)", style=solid];
-"272 select_6" -> "277 expand_as_4"  [label="(16, 6, 64, 32)", style=solid];
-"272 select_6" -> "278 div_4"  [label="(16, 6, 64, 32)", style=solid];
-"273 select_7" -> "279 linalg_vector_norm_5"  [label="(16, 6, 64, 32)", style=solid];
-"273 select_7" -> "281 expand_as_5"  [label="(16, 6, 64, 32)", style=solid];
-"273 select_7" -> "282 div_5"  [label="(16, 6, 64, 32)", style=solid];
-"274 select_8" -> "292 matmul_5"  [label="(16, 6, 64, 32)", style=solid];
-"275 linalg_vector_norm_4" -> "276 clamp_min_4"  [label="(16, 6, 64, 1)", style=solid];
-"276 clamp_min_4" -> "277 expand_as_4"  [label="(16, 6, 64, 1)", style=solid];
-"277 expand_as_4" -> "278 div_4"  [label="(16, 6, 64, 32)", style=solid];
-"278 div_4" -> "284 matmul_4"  [label="(16, 6, 64, 32)", style=solid];
-"279 linalg_vector_norm_5" -> "280 clamp_min_5"  [label="(16, 6, 64, 1)", style=solid];
-"280 clamp_min_5" -> "281 expand_as_5"  [label="(16, 6, 64, 1)", style=solid];
-"281 expand_as_5" -> "282 div_5"  [label="(16, 6, 64, 32)", style=solid];
-"282 div_5" -> "283 transpose_4"  [label="(16, 6, 64, 32)", style=solid];
-"283 transpose_4" -> "284 matmul_4"  [label="(16, 6, 32, 64)", style=solid];
-"284 matmul_4" -> "288 mul_5"  [label="(16, 6, 64, 64)", style=solid];
-"285 _param_constant44" -> "286 clamp_2"  [label="(6, 1, 1)", style=solid];
-"286 clamp_2" -> "287 exp_2"  [label="(6, 1, 1)", style=solid];
-"287 exp_2" -> "288 mul_5"  [label="(6, 1, 1)", style=solid];
-"288 mul_5" -> "289 add_7"  [label="(16, 6, 64, 64)", style=solid];
-"289 add_7" -> "290 softmax_2"  [label="(16, 6, 64, 64)", style=solid];
-"290 softmax_2" -> "291 dropout_8"  [label="(16, 6, 64, 64)", style=solid];
-"291 dropout_8" -> "292 matmul_5"  [label="(16, 6, 64, 64)", style=solid];
-"292 matmul_5" -> "293 transpose_5"  [label="(16, 6, 64, 32)", style=solid];
-"293 transpose_5" -> "294 reshape_11"  [label="(16, 64, 6, 32)", style=solid];
-"294 reshape_11" -> "297 linear_16"  [label="(16, 64, 192)", style=solid];
-"295 _param_constant45" -> "297 linear_16"  [label="(192, 192)", style=solid];
-"296 _param_constant46" -> "297 linear_16"  [label="(192,)", style=solid];
-"297 linear_16" -> "298 dropout_9"  [label="(16, 64, 192)", style=solid];
-"298 dropout_9" -> "299 view_14"  [label="(16, 64, 192)", style=solid];
-"299 view_14" -> "300 permute_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"300 permute_13" -> "301 reshape_12"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"301 reshape_12" -> "302 slice_38"  [label="(1, 32, 32, 192)", style=solid];
-"302 slice_38" -> "303 slice_39"  [label="(1, 32, 32, 192)", style=solid];
-"303 slice_39" -> "304 slice_40"  [label="(1, 28, 32, 192)", style=solid];
-"304 slice_40" -> "305 slice_41"  [label="(1, 28, 28, 192)", style=solid];
-"305 slice_41" -> "306 contiguous_3"  [label="(1, 28, 28, 192)", style=solid];
-"306 contiguous_3" -> "309 layer_norm_6"  [label="(1, 28, 28, 192)", style=solid];
-"307 _param_constant47" -> "309 layer_norm_6"  [label="(192,)", style=solid];
-"308 _param_constant48" -> "309 layer_norm_6"  [label="(192,)", style=solid];
-"309 layer_norm_6" -> "310 add_8"  [label="(1, 28, 28, 192)", style=solid];
-"310 add_8" -> "313 linear_17"  [label="(1, 28, 28, 192)", style=solid];
-"310 add_8" -> "323 add_9"  [label="(1, 28, 28, 192)", style=solid];
-"311 _param_constant49" -> "313 linear_17"  [label="(768, 192)", style=solid];
-"312 _param_constant50" -> "313 linear_17"  [label="(768,)", style=solid];
-"313 linear_17" -> "314 gelu_2"  [label="(1, 28, 28, 768)", style=solid];
-"314 gelu_2" -> "315 dropout_10"  [label="(1, 28, 28, 768)", style=solid];
-"315 dropout_10" -> "318 linear_18"  [label="(1, 28, 28, 768)", style=solid];
-"316 _param_constant51" -> "318 linear_18"  [label="(192, 768)", style=solid];
-"317 _param_constant52" -> "318 linear_18"  [label="(192,)", style=solid];
-"318 linear_18" -> "319 dropout_11"  [label="(1, 28, 28, 192)", style=solid];
-"319 dropout_11" -> "322 layer_norm_7"  [label="(1, 28, 28, 192)", style=solid];
-"320 _param_constant53" -> "322 layer_norm_7"  [label="(192,)", style=solid];
-"321 _param_constant54" -> "322 layer_norm_7"  [label="(192,)", style=solid];
-"322 layer_norm_7" -> "323 add_9"  [label="(1, 28, 28, 192)", style=solid];
-"323 add_9" -> "340 pad_4"  [label="(1, 28, 28, 192)", style=solid];
-"323 add_9" -> "453 add_12"  [label="(1, 28, 28, 192)", style=solid];
-"324 _tensor_constant15" -> "327 linear_19"  [label="(1, 15, 15, 2)", style=solid];
-"325 _param_constant55" -> "327 linear_19"  [label="(512, 2)", style=solid];
-"326 _param_constant56" -> "327 linear_19"  [label="(512,)", style=solid];
-"327 linear_19" -> "328 relu__3"  [label="(1, 15, 15, 512)", style=solid];
-"328 relu__3" -> "330 linear_20"  [label="(1, 15, 15, 512)", style=solid];
-"329 _param_constant57" -> "330 linear_20"  [label="(6, 512)", style=solid];
-"330 linear_20" -> "331 view_15"  [label="(1, 15, 15, 6)", style=solid];
-"331 view_15" -> "333 index_3"  [label="(225, 6)", style=solid];
-"332 _tensor_constant16" -> "333 index_3"  [label="(4096,)", style=solid];
-"333 index_3" -> "334 view_16"  [label="(4096, 6)", style=solid];
-"334 view_16" -> "335 permute_14"  [label="(64, 64, 6)", style=solid];
-"335 permute_14" -> "336 contiguous_4"  [label="(6, 64, 64)", style=solid];
-"336 contiguous_4" -> "337 unsqueeze_7"  [label="(6, 64, 64)", style=solid];
-"337 unsqueeze_7" -> "338 sigmoid_3"  [label="(1, 6, 64, 64)", style=solid];
-"338 sigmoid_3" -> "339 mul_6"  [label="(1, 6, 64, 64)", style=solid];
-"339 mul_6" -> "370 add_10"  [label="(1, 6, 64, 64)", style=solid];
-"340 pad_4" -> "341 roll_2"  [label="(1, 32, 32, 192)", style=solid];
-"341 roll_2" -> "342 view_17"  [label="(1, 32, 32, 192)", style=solid];
-"342 view_17" -> "343 permute_15"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"343 permute_15" -> "344 reshape_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"344 reshape_13" -> "350 linear_21"  [label="(16, 64, 192)", style=solid];
-"344 reshape_13" -> "371 new_zeros_1"  [label="(16, 64, 192)", style=solid];
-"345 _param_constant58" -> "346 clone_3"  [label="(576,)", style=solid];
-"346 clone_3" -> "347 slice_42"  [label="(576,)", style=solid];
-"346 clone_3" -> "350 linear_21"  [label="(576,)", style=solid];
-"347 slice_42" -> "348 zero__3"  [label="(192,)", style=solid];
-"349 _param_constant59" -> "350 linear_21"  [label="(576, 192)", style=solid];
-"350 linear_21" -> "351 reshape_14"  [label="(16, 64, 576)", style=solid];
-"351 reshape_14" -> "352 permute_16"  [label="(16, 64, 3, 6, 32)", style=solid];
-"352 permute_16" -> "353 select_9"  [label="(3, 16, 6, 64, 32)", style=solid];
-"352 permute_16" -> "354 select_10"  [label="(3, 16, 6, 64, 32)", style=solid];
-"352 permute_16" -> "355 select_11"  [label="(3, 16, 6, 64, 32)", style=solid];
-"353 select_9" -> "356 linalg_vector_norm_6"  [label="(16, 6, 64, 32)", style=solid];
-"353 select_9" -> "358 expand_as_6"  [label="(16, 6, 64, 32)", style=solid];
-"353 select_9" -> "359 div_6"  [label="(16, 6, 64, 32)", style=solid];
-"354 select_10" -> "360 linalg_vector_norm_7"  [label="(16, 6, 64, 32)", style=solid];
-"354 select_10" -> "362 expand_as_7"  [label="(16, 6, 64, 32)", style=solid];
-"354 select_10" -> "363 div_7"  [label="(16, 6, 64, 32)", style=solid];
-"355 select_11" -> "434 matmul_7"  [label="(16, 6, 64, 32)", style=solid];
-"356 linalg_vector_norm_6" -> "357 clamp_min_6"  [label="(16, 6, 64, 1)", style=solid];
-"357 clamp_min_6" -> "358 expand_as_6"  [label="(16, 6, 64, 1)", style=solid];
-"358 expand_as_6" -> "359 div_6"  [label="(16, 6, 64, 32)", style=solid];
-"359 div_6" -> "365 matmul_6"  [label="(16, 6, 64, 32)", style=solid];
-"360 linalg_vector_norm_7" -> "361 clamp_min_7"  [label="(16, 6, 64, 1)", style=solid];
-"361 clamp_min_7" -> "362 expand_as_7"  [label="(16, 6, 64, 1)", style=solid];
-"362 expand_as_7" -> "363 div_7"  [label="(16, 6, 64, 32)", style=solid];
-"363 div_7" -> "364 transpose_6"  [label="(16, 6, 64, 32)", style=solid];
-"364 transpose_6" -> "365 matmul_6"  [label="(16, 6, 32, 64)", style=solid];
-"365 matmul_6" -> "369 mul_7"  [label="(16, 6, 64, 64)", style=solid];
-"366 _param_constant60" -> "367 clamp_3"  [label="(6, 1, 1)", style=solid];
-"367 clamp_3" -> "368 exp_3"  [label="(6, 1, 1)", style=solid];
-"368 exp_3" -> "369 mul_7"  [label="(6, 1, 1)", style=solid];
-"369 mul_7" -> "370 add_10"  [label="(16, 6, 64, 64)", style=solid];
-"370 add_10" -> "427 view_19"  [label="(16, 6, 64, 64)", style=solid];
-"371 new_zeros_1" -> "374 slice_43"  [label="(32, 32)", style=solid];
-"371 new_zeros_1" -> "379 slice_45"  [label="(32, 32)", style=solid];
-"371 new_zeros_1" -> "384 slice_47"  [label="(32, 32)", style=solid];
-"371 new_zeros_1" -> "389 slice_49"  [label="(32, 32)", style=solid];
-"371 new_zeros_1" -> "394 slice_51"  [label="(32, 32)", style=solid];
-"371 new_zeros_1" -> "399 slice_53"  [label="(32, 32)", style=solid];
-"371 new_zeros_1" -> "404 slice_55"  [label="(32, 32)", style=solid];
-"371 new_zeros_1" -> "409 slice_57"  [label="(32, 32)", style=solid];
-"371 new_zeros_1" -> "414 slice_59"  [label="(32, 32)", style=solid];
-"371 new_zeros_1" -> "417 view_18"  [label="(32, 32)", style=solid];
-"372 _tensor_constant17" -> "373 lift_fresh_copy_9"  [label="()", style=solid];
-"373 lift_fresh_copy_9" -> "376 fill__9"  [label="()", style=solid];
-"374 slice_43" -> "375 slice_44"  [label="(24, 32)", style=solid];
-"375 slice_44" -> "376 fill__9"  [label="(24, 24)", style=solid];
-"377 _tensor_constant18" -> "378 lift_fresh_copy_10"  [label="()", style=solid];
-"378 lift_fresh_copy_10" -> "381 fill__10"  [label="()", style=solid];
-"379 slice_45" -> "380 slice_46"  [label="(24, 32)", style=solid];
-"380 slice_46" -> "381 fill__10"  [label="(24, 4)", style=solid];
-"382 _tensor_constant19" -> "383 lift_fresh_copy_11"  [label="()", style=solid];
-"383 lift_fresh_copy_11" -> "386 fill__11"  [label="()", style=solid];
-"384 slice_47" -> "385 slice_48"  [label="(24, 32)", style=solid];
-"385 slice_48" -> "386 fill__11"  [label="(24, 4)", style=solid];
-"387 _tensor_constant20" -> "388 lift_fresh_copy_12"  [label="()", style=solid];
-"388 lift_fresh_copy_12" -> "391 fill__12"  [label="()", style=solid];
-"389 slice_49" -> "390 slice_50"  [label="(4, 32)", style=solid];
-"390 slice_50" -> "391 fill__12"  [label="(4, 24)", style=solid];
-"392 _tensor_constant21" -> "393 lift_fresh_copy_13"  [label="()", style=solid];
-"393 lift_fresh_copy_13" -> "396 fill__13"  [label="()", style=solid];
-"394 slice_51" -> "395 slice_52"  [label="(4, 32)", style=solid];
-"395 slice_52" -> "396 fill__13"  [label="(4, 4)", style=solid];
-"397 _tensor_constant22" -> "398 lift_fresh_copy_14"  [label="()", style=solid];
-"398 lift_fresh_copy_14" -> "401 fill__14"  [label="()", style=solid];
-"399 slice_53" -> "400 slice_54"  [label="(4, 32)", style=solid];
-"400 slice_54" -> "401 fill__14"  [label="(4, 4)", style=solid];
-"402 _tensor_constant23" -> "403 lift_fresh_copy_15"  [label="()", style=solid];
-"403 lift_fresh_copy_15" -> "406 fill__15"  [label="()", style=solid];
-"404 slice_55" -> "405 slice_56"  [label="(4, 32)", style=solid];
-"405 slice_56" -> "406 fill__15"  [label="(4, 24)", style=solid];
-"407 _tensor_constant24" -> "408 lift_fresh_copy_16"  [label="()", style=solid];
-"408 lift_fresh_copy_16" -> "411 fill__16"  [label="()", style=solid];
-"409 slice_57" -> "410 slice_58"  [label="(4, 32)", style=solid];
-"410 slice_58" -> "411 fill__16"  [label="(4, 4)", style=solid];
-"412 _tensor_constant25" -> "413 lift_fresh_copy_17"  [label="()", style=solid];
-"413 lift_fresh_copy_17" -> "416 fill__17"  [label="()", style=solid];
-"414 slice_59" -> "415 slice_60"  [label="(4, 32)", style=solid];
-"415 slice_60" -> "416 fill__17"  [label="(4, 4)", style=solid];
-"417 view_18" -> "418 permute_17"  [label="(4, 8, 4, 8)", style=solid];
-"418 permute_17" -> "419 reshape_15"  [label="(4, 4, 8, 8)", style=solid];
-"419 reshape_15" -> "420 unsqueeze_8"  [label="(16, 64)", style=solid];
-"419 reshape_15" -> "421 unsqueeze_9"  [label="(16, 64)", style=solid];
-"420 unsqueeze_8" -> "422 sub_1"  [label="(16, 1, 64)", style=solid];
-"421 unsqueeze_9" -> "422 sub_1"  [label="(16, 64, 1)", style=solid];
-"422 sub_1" -> "423 ne_1"  [label="(16, 64, 64)", style=solid];
-"422 sub_1" -> "424 masked_fill_2"  [label="(16, 64, 64)", style=solid];
-"422 sub_1" -> "425 eq_1"  [label="(16, 64, 64)", style=solid];
-"423 ne_1" -> "424 masked_fill_2"  [label="(16, 64, 64)", style=solid];
-"424 masked_fill_2" -> "426 masked_fill_3"  [label="(16, 64, 64)", style=solid];
-"425 eq_1" -> "426 masked_fill_3"  [label="(16, 64, 64)", style=solid];
-"426 masked_fill_3" -> "428 unsqueeze_10"  [label="(16, 64, 64)", style=solid];
-"427 view_19" -> "430 add_11"  [label="(1, 16, 6, 64, 64)", style=solid];
-"428 unsqueeze_10" -> "429 unsqueeze_11"  [label="(16, 1, 64, 64)", style=solid];
-"429 unsqueeze_11" -> "430 add_11"  [label="(1, 16, 1, 64, 64)", style=solid];
-"430 add_11" -> "431 view_20"  [label="(1, 16, 6, 64, 64)", style=solid];
-"431 view_20" -> "432 softmax_3"  [label="(16, 6, 64, 64)", style=solid];
-"432 softmax_3" -> "433 dropout_12"  [label="(16, 6, 64, 64)", style=solid];
-"433 dropout_12" -> "434 matmul_7"  [label="(16, 6, 64, 64)", style=solid];
-"434 matmul_7" -> "435 transpose_7"  [label="(16, 6, 64, 32)", style=solid];
-"435 transpose_7" -> "436 reshape_16"  [label="(16, 64, 6, 32)", style=solid];
-"436 reshape_16" -> "439 linear_22"  [label="(16, 64, 192)", style=solid];
-"437 _param_constant61" -> "439 linear_22"  [label="(192, 192)", style=solid];
-"438 _param_constant62" -> "439 linear_22"  [label="(192,)", style=solid];
-"439 linear_22" -> "440 dropout_13"  [label="(16, 64, 192)", style=solid];
-"440 dropout_13" -> "441 view_21"  [label="(16, 64, 192)", style=solid];
-"441 view_21" -> "442 permute_18"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
-"442 permute_18" -> "443 reshape_17"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
-"443 reshape_17" -> "444 roll_3"  [label="(1, 32, 32, 192)", style=solid];
-"444 roll_3" -> "445 slice_61"  [label="(1, 32, 32, 192)", style=solid];
-"445 slice_61" -> "446 slice_62"  [label="(1, 32, 32, 192)", style=solid];
-"446 slice_62" -> "447 slice_63"  [label="(1, 28, 32, 192)", style=solid];
-"447 slice_63" -> "448 slice_64"  [label="(1, 28, 28, 192)", style=solid];
-"448 slice_64" -> "449 contiguous_5"  [label="(1, 28, 28, 192)", style=solid];
-"449 contiguous_5" -> "452 layer_norm_8"  [label="(1, 28, 28, 192)", style=solid];
-"450 _param_constant63" -> "452 layer_norm_8"  [label="(192,)", style=solid];
-"451 _param_constant64" -> "452 layer_norm_8"  [label="(192,)", style=solid];
-"452 layer_norm_8" -> "453 add_12"  [label="(1, 28, 28, 192)", style=solid];
-"453 add_12" -> "456 linear_23"  [label="(1, 28, 28, 192)", style=solid];
-"453 add_12" -> "466 add_13"  [label="(1, 28, 28, 192)", style=solid];
-"454 _param_constant65" -> "456 linear_23"  [label="(768, 192)", style=solid];
-"455 _param_constant66" -> "456 linear_23"  [label="(768,)", style=solid];
-"456 linear_23" -> "457 gelu_3"  [label="(1, 28, 28, 768)", style=solid];
-"457 gelu_3" -> "458 dropout_14"  [label="(1, 28, 28, 768)", style=solid];
-"458 dropout_14" -> "461 linear_24"  [label="(1, 28, 28, 768)", style=solid];
-"459 _param_constant67" -> "461 linear_24"  [label="(192, 768)", style=solid];
-"460 _param_constant68" -> "461 linear_24"  [label="(192,)", style=solid];
-"461 linear_24" -> "462 dropout_15"  [label="(1, 28, 28, 192)", style=solid];
-"462 dropout_15" -> "465 layer_norm_9"  [label="(1, 28, 28, 192)", style=solid];
-"463 _param_constant69" -> "465 layer_norm_9"  [label="(192,)", style=solid];
-"464 _param_constant70" -> "465 layer_norm_9"  [label="(192,)", style=solid];
-"465 layer_norm_9" -> "466 add_13"  [label="(1, 28, 28, 192)", style=solid];
-"466 add_13" -> "467 pad_5"  [label="(1, 28, 28, 192)", style=solid];
-"467 pad_5" -> "468 slice_65"  [label="(1, 28, 28, 192)", style=solid];
-"467 pad_5" -> "471 slice_68"  [label="(1, 28, 28, 192)", style=solid];
-"467 pad_5" -> "474 slice_71"  [label="(1, 28, 28, 192)", style=solid];
-"467 pad_5" -> "477 slice_74"  [label="(1, 28, 28, 192)", style=solid];
-"468 slice_65" -> "469 slice_66"  [label="(1, 14, 28, 192)", style=solid];
-"469 slice_66" -> "470 slice_67"  [label="(1, 14, 14, 192)", style=solid];
-"470 slice_67" -> "480 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"471 slice_68" -> "472 slice_69"  [label="(1, 14, 28, 192)", style=solid];
-"472 slice_69" -> "473 slice_70"  [label="(1, 14, 14, 192)", style=solid];
-"473 slice_70" -> "480 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"474 slice_71" -> "475 slice_72"  [label="(1, 14, 28, 192)", style=solid];
-"475 slice_72" -> "476 slice_73"  [label="(1, 14, 14, 192)", style=solid];
-"476 slice_73" -> "480 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"477 slice_74" -> "478 slice_75"  [label="(1, 14, 28, 192)", style=solid];
-"478 slice_75" -> "479 slice_76"  [label="(1, 14, 14, 192)", style=solid];
-"479 slice_76" -> "480 cat_1"  [label="(1, 14, 14, 192)", style=solid];
-"480 cat_1" -> "482 linear_25"  [label="(1, 14, 14, 768)", style=solid];
-"481 _param_constant71" -> "482 linear_25"  [label="(384, 768)", style=solid];
-"482 linear_25" -> "485 layer_norm_10"  [label="(1, 14, 14, 384)", style=solid];
-"483 _param_constant72" -> "485 layer_norm_10"  [label="(384,)", style=solid];
-"484 _param_constant73" -> "485 layer_norm_10"  [label="(384,)", style=solid];
-"485 layer_norm_10" -> "502 pad_6"  [label="(1, 14, 14, 384)", style=solid];
-"485 layer_norm_10" -> "552 add_15"  [label="(1, 14, 14, 384)", style=solid];
-"486 _tensor_constant26" -> "489 linear_26"  [label="(1, 15, 15, 2)", style=solid];
-"487 _param_constant74" -> "489 linear_26"  [label="(512, 2)", style=solid];
-"488 _param_constant75" -> "489 linear_26"  [label="(512,)", style=solid];
-"489 linear_26" -> "490 relu__4"  [label="(1, 15, 15, 512)", style=solid];
-"490 relu__4" -> "492 linear_27"  [label="(1, 15, 15, 512)", style=solid];
-"491 _param_constant76" -> "492 linear_27"  [label="(12, 512)", style=solid];
-"492 linear_27" -> "493 view_22"  [label="(1, 15, 15, 12)", style=solid];
-"493 view_22" -> "495 index_4"  [label="(225, 12)", style=solid];
-"494 _tensor_constant27" -> "495 index_4"  [label="(4096,)", style=solid];
-"495 index_4" -> "496 view_23"  [label="(4096, 12)", style=solid];
-"496 view_23" -> "497 permute_19"  [label="(64, 64, 12)", style=solid];
-"497 permute_19" -> "498 contiguous_6"  [label="(12, 64, 64)", style=solid];
-"498 contiguous_6" -> "499 unsqueeze_12"  [label="(12, 64, 64)", style=solid];
-"499 unsqueeze_12" -> "500 sigmoid_4"  [label="(1, 12, 64, 64)", style=solid];
-"500 sigmoid_4" -> "501 mul_8"  [label="(1, 12, 64, 64)", style=solid];
-"501 mul_8" -> "531 add_14"  [label="(1, 12, 64, 64)", style=solid];
-"502 pad_6" -> "503 view_24"  [label="(1, 16, 16, 384)", style=solid];
-"503 view_24" -> "504 permute_20"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"504 permute_20" -> "505 reshape_18"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"505 reshape_18" -> "511 linear_28"  [label="(4, 64, 384)", style=solid];
-"506 _param_constant77" -> "507 clone_4"  [label="(1152,)", style=solid];
-"507 clone_4" -> "508 slice_77"  [label="(1152,)", style=solid];
-"507 clone_4" -> "511 linear_28"  [label="(1152,)", style=solid];
-"508 slice_77" -> "509 zero__4"  [label="(384,)", style=solid];
-"510 _param_constant78" -> "511 linear_28"  [label="(1152, 384)", style=solid];
-"511 linear_28" -> "512 reshape_19"  [label="(4, 64, 1152)", style=solid];
-"512 reshape_19" -> "513 permute_21"  [label="(4, 64, 3, 12, 32)", style=solid];
-"513 permute_21" -> "514 select_12"  [label="(3, 4, 12, 64, 32)", style=solid];
-"513 permute_21" -> "515 select_13"  [label="(3, 4, 12, 64, 32)", style=solid];
-"513 permute_21" -> "516 select_14"  [label="(3, 4, 12, 64, 32)", style=solid];
-"514 select_12" -> "517 linalg_vector_norm_8"  [label="(4, 12, 64, 32)", style=solid];
-"514 select_12" -> "519 expand_as_8"  [label="(4, 12, 64, 32)", style=solid];
-"514 select_12" -> "520 div_8"  [label="(4, 12, 64, 32)", style=solid];
-"515 select_13" -> "521 linalg_vector_norm_9"  [label="(4, 12, 64, 32)", style=solid];
-"515 select_13" -> "523 expand_as_9"  [label="(4, 12, 64, 32)", style=solid];
-"515 select_13" -> "524 div_9"  [label="(4, 12, 64, 32)", style=solid];
-"516 select_14" -> "534 matmul_9"  [label="(4, 12, 64, 32)", style=solid];
-"517 linalg_vector_norm_8" -> "518 clamp_min_8"  [label="(4, 12, 64, 1)", style=solid];
-"518 clamp_min_8" -> "519 expand_as_8"  [label="(4, 12, 64, 1)", style=solid];
-"519 expand_as_8" -> "520 div_8"  [label="(4, 12, 64, 32)", style=solid];
-"520 div_8" -> "526 matmul_8"  [label="(4, 12, 64, 32)", style=solid];
-"521 linalg_vector_norm_9" -> "522 clamp_min_9"  [label="(4, 12, 64, 1)", style=solid];
-"522 clamp_min_9" -> "523 expand_as_9"  [label="(4, 12, 64, 1)", style=solid];
-"523 expand_as_9" -> "524 div_9"  [label="(4, 12, 64, 32)", style=solid];
-"524 div_9" -> "525 transpose_8"  [label="(4, 12, 64, 32)", style=solid];
-"525 transpose_8" -> "526 matmul_8"  [label="(4, 12, 32, 64)", style=solid];
-"526 matmul_8" -> "530 mul_9"  [label="(4, 12, 64, 64)", style=solid];
-"527 _param_constant79" -> "528 clamp_4"  [label="(12, 1, 1)", style=solid];
-"528 clamp_4" -> "529 exp_4"  [label="(12, 1, 1)", style=solid];
-"529 exp_4" -> "530 mul_9"  [label="(12, 1, 1)", style=solid];
-"530 mul_9" -> "531 add_14"  [label="(4, 12, 64, 64)", style=solid];
-"531 add_14" -> "532 softmax_4"  [label="(4, 12, 64, 64)", style=solid];
-"532 softmax_4" -> "533 dropout_16"  [label="(4, 12, 64, 64)", style=solid];
-"533 dropout_16" -> "534 matmul_9"  [label="(4, 12, 64, 64)", style=solid];
-"534 matmul_9" -> "535 transpose_9"  [label="(4, 12, 64, 32)", style=solid];
-"535 transpose_9" -> "536 reshape_20"  [label="(4, 64, 12, 32)", style=solid];
-"536 reshape_20" -> "539 linear_29"  [label="(4, 64, 384)", style=solid];
-"537 _param_constant80" -> "539 linear_29"  [label="(384, 384)", style=solid];
-"538 _param_constant81" -> "539 linear_29"  [label="(384,)", style=solid];
-"539 linear_29" -> "540 dropout_17"  [label="(4, 64, 384)", style=solid];
-"540 dropout_17" -> "541 view_25"  [label="(4, 64, 384)", style=solid];
-"541 view_25" -> "542 permute_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"542 permute_22" -> "543 reshape_21"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"543 reshape_21" -> "544 slice_78"  [label="(1, 16, 16, 384)", style=solid];
-"544 slice_78" -> "545 slice_79"  [label="(1, 16, 16, 384)", style=solid];
-"545 slice_79" -> "546 slice_80"  [label="(1, 14, 16, 384)", style=solid];
-"546 slice_80" -> "547 slice_81"  [label="(1, 14, 14, 384)", style=solid];
-"547 slice_81" -> "548 contiguous_7"  [label="(1, 14, 14, 384)", style=solid];
-"548 contiguous_7" -> "551 layer_norm_11"  [label="(1, 14, 14, 384)", style=solid];
-"549 _param_constant82" -> "551 layer_norm_11"  [label="(384,)", style=solid];
-"550 _param_constant83" -> "551 layer_norm_11"  [label="(384,)", style=solid];
-"551 layer_norm_11" -> "552 add_15"  [label="(1, 14, 14, 384)", style=solid];
-"552 add_15" -> "555 linear_30"  [label="(1, 14, 14, 384)", style=solid];
-"552 add_15" -> "565 add_16"  [label="(1, 14, 14, 384)", style=solid];
-"553 _param_constant84" -> "555 linear_30"  [label="(1536, 384)", style=solid];
-"554 _param_constant85" -> "555 linear_30"  [label="(1536,)", style=solid];
-"555 linear_30" -> "556 gelu_4"  [label="(1, 14, 14, 1536)", style=solid];
-"556 gelu_4" -> "557 dropout_18"  [label="(1, 14, 14, 1536)", style=solid];
-"557 dropout_18" -> "560 linear_31"  [label="(1, 14, 14, 1536)", style=solid];
-"558 _param_constant86" -> "560 linear_31"  [label="(384, 1536)", style=solid];
-"559 _param_constant87" -> "560 linear_31"  [label="(384,)", style=solid];
-"560 linear_31" -> "561 dropout_19"  [label="(1, 14, 14, 384)", style=solid];
-"561 dropout_19" -> "564 layer_norm_12"  [label="(1, 14, 14, 384)", style=solid];
-"562 _param_constant88" -> "564 layer_norm_12"  [label="(384,)", style=solid];
-"563 _param_constant89" -> "564 layer_norm_12"  [label="(384,)", style=solid];
-"564 layer_norm_12" -> "565 add_16"  [label="(1, 14, 14, 384)", style=solid];
-"565 add_16" -> "582 pad_7"  [label="(1, 14, 14, 384)", style=solid];
-"565 add_16" -> "695 add_19"  [label="(1, 14, 14, 384)", style=solid];
-"566 _tensor_constant28" -> "569 linear_32"  [label="(1, 15, 15, 2)", style=solid];
-"567 _param_constant90" -> "569 linear_32"  [label="(512, 2)", style=solid];
-"568 _param_constant91" -> "569 linear_32"  [label="(512,)", style=solid];
-"569 linear_32" -> "570 relu__5"  [label="(1, 15, 15, 512)", style=solid];
-"570 relu__5" -> "572 linear_33"  [label="(1, 15, 15, 512)", style=solid];
-"571 _param_constant92" -> "572 linear_33"  [label="(12, 512)", style=solid];
-"572 linear_33" -> "573 view_26"  [label="(1, 15, 15, 12)", style=solid];
-"573 view_26" -> "575 index_5"  [label="(225, 12)", style=solid];
-"574 _tensor_constant29" -> "575 index_5"  [label="(4096,)", style=solid];
-"575 index_5" -> "576 view_27"  [label="(4096, 12)", style=solid];
-"576 view_27" -> "577 permute_23"  [label="(64, 64, 12)", style=solid];
-"577 permute_23" -> "578 contiguous_8"  [label="(12, 64, 64)", style=solid];
-"578 contiguous_8" -> "579 unsqueeze_13"  [label="(12, 64, 64)", style=solid];
-"579 unsqueeze_13" -> "580 sigmoid_5"  [label="(1, 12, 64, 64)", style=solid];
-"580 sigmoid_5" -> "581 mul_10"  [label="(1, 12, 64, 64)", style=solid];
-"581 mul_10" -> "612 add_17"  [label="(1, 12, 64, 64)", style=solid];
-"582 pad_7" -> "583 roll_4"  [label="(1, 16, 16, 384)", style=solid];
-"583 roll_4" -> "584 view_28"  [label="(1, 16, 16, 384)", style=solid];
-"584 view_28" -> "585 permute_24"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"585 permute_24" -> "586 reshape_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"586 reshape_22" -> "592 linear_34"  [label="(4, 64, 384)", style=solid];
-"586 reshape_22" -> "613 new_zeros_2"  [label="(4, 64, 384)", style=solid];
-"587 _param_constant93" -> "588 clone_5"  [label="(1152,)", style=solid];
-"588 clone_5" -> "589 slice_82"  [label="(1152,)", style=solid];
-"588 clone_5" -> "592 linear_34"  [label="(1152,)", style=solid];
-"589 slice_82" -> "590 zero__5"  [label="(384,)", style=solid];
-"591 _param_constant94" -> "592 linear_34"  [label="(1152, 384)", style=solid];
-"592 linear_34" -> "593 reshape_23"  [label="(4, 64, 1152)", style=solid];
-"593 reshape_23" -> "594 permute_25"  [label="(4, 64, 3, 12, 32)", style=solid];
-"594 permute_25" -> "595 select_15"  [label="(3, 4, 12, 64, 32)", style=solid];
-"594 permute_25" -> "596 select_16"  [label="(3, 4, 12, 64, 32)", style=solid];
-"594 permute_25" -> "597 select_17"  [label="(3, 4, 12, 64, 32)", style=solid];
-"595 select_15" -> "598 linalg_vector_norm_10"  [label="(4, 12, 64, 32)", style=solid];
-"595 select_15" -> "600 expand_as_10"  [label="(4, 12, 64, 32)", style=solid];
-"595 select_15" -> "601 div_10"  [label="(4, 12, 64, 32)", style=solid];
-"596 select_16" -> "602 linalg_vector_norm_11"  [label="(4, 12, 64, 32)", style=solid];
-"596 select_16" -> "604 expand_as_11"  [label="(4, 12, 64, 32)", style=solid];
-"596 select_16" -> "605 div_11"  [label="(4, 12, 64, 32)", style=solid];
-"597 select_17" -> "676 matmul_11"  [label="(4, 12, 64, 32)", style=solid];
-"598 linalg_vector_norm_10" -> "599 clamp_min_10"  [label="(4, 12, 64, 1)", style=solid];
-"599 clamp_min_10" -> "600 expand_as_10"  [label="(4, 12, 64, 1)", style=solid];
-"600 expand_as_10" -> "601 div_10"  [label="(4, 12, 64, 32)", style=solid];
-"601 div_10" -> "607 matmul_10"  [label="(4, 12, 64, 32)", style=solid];
-"602 linalg_vector_norm_11" -> "603 clamp_min_11"  [label="(4, 12, 64, 1)", style=solid];
-"603 clamp_min_11" -> "604 expand_as_11"  [label="(4, 12, 64, 1)", style=solid];
-"604 expand_as_11" -> "605 div_11"  [label="(4, 12, 64, 32)", style=solid];
-"605 div_11" -> "606 transpose_10"  [label="(4, 12, 64, 32)", style=solid];
-"606 transpose_10" -> "607 matmul_10"  [label="(4, 12, 32, 64)", style=solid];
-"607 matmul_10" -> "611 mul_11"  [label="(4, 12, 64, 64)", style=solid];
-"608 _param_constant95" -> "609 clamp_5"  [label="(12, 1, 1)", style=solid];
-"609 clamp_5" -> "610 exp_5"  [label="(12, 1, 1)", style=solid];
-"610 exp_5" -> "611 mul_11"  [label="(12, 1, 1)", style=solid];
-"611 mul_11" -> "612 add_17"  [label="(4, 12, 64, 64)", style=solid];
-"612 add_17" -> "669 view_30"  [label="(4, 12, 64, 64)", style=solid];
-"613 new_zeros_2" -> "616 slice_83"  [label="(16, 16)", style=solid];
-"613 new_zeros_2" -> "621 slice_85"  [label="(16, 16)", style=solid];
-"613 new_zeros_2" -> "626 slice_87"  [label="(16, 16)", style=solid];
-"613 new_zeros_2" -> "631 slice_89"  [label="(16, 16)", style=solid];
-"613 new_zeros_2" -> "636 slice_91"  [label="(16, 16)", style=solid];
-"613 new_zeros_2" -> "641 slice_93"  [label="(16, 16)", style=solid];
-"613 new_zeros_2" -> "646 slice_95"  [label="(16, 16)", style=solid];
-"613 new_zeros_2" -> "651 slice_97"  [label="(16, 16)", style=solid];
-"613 new_zeros_2" -> "656 slice_99"  [label="(16, 16)", style=solid];
-"613 new_zeros_2" -> "659 view_29"  [label="(16, 16)", style=solid];
-"614 _tensor_constant30" -> "615 lift_fresh_copy_18"  [label="()", style=solid];
-"615 lift_fresh_copy_18" -> "618 fill__18"  [label="()", style=solid];
-"616 slice_83" -> "617 slice_84"  [label="(8, 16)", style=solid];
-"617 slice_84" -> "618 fill__18"  [label="(8, 8)", style=solid];
-"619 _tensor_constant31" -> "620 lift_fresh_copy_19"  [label="()", style=solid];
-"620 lift_fresh_copy_19" -> "623 fill__19"  [label="()", style=solid];
-"621 slice_85" -> "622 slice_86"  [label="(8, 16)", style=solid];
-"622 slice_86" -> "623 fill__19"  [label="(8, 4)", style=solid];
-"624 _tensor_constant32" -> "625 lift_fresh_copy_20"  [label="()", style=solid];
-"625 lift_fresh_copy_20" -> "628 fill__20"  [label="()", style=solid];
-"626 slice_87" -> "627 slice_88"  [label="(8, 16)", style=solid];
-"627 slice_88" -> "628 fill__20"  [label="(8, 4)", style=solid];
-"629 _tensor_constant33" -> "630 lift_fresh_copy_21"  [label="()", style=solid];
-"630 lift_fresh_copy_21" -> "633 fill__21"  [label="()", style=solid];
-"631 slice_89" -> "632 slice_90"  [label="(4, 16)", style=solid];
-"632 slice_90" -> "633 fill__21"  [label="(4, 8)", style=solid];
-"634 _tensor_constant34" -> "635 lift_fresh_copy_22"  [label="()", style=solid];
-"635 lift_fresh_copy_22" -> "638 fill__22"  [label="()", style=solid];
-"636 slice_91" -> "637 slice_92"  [label="(4, 16)", style=solid];
-"637 slice_92" -> "638 fill__22"  [label="(4, 4)", style=solid];
-"639 _tensor_constant35" -> "640 lift_fresh_copy_23"  [label="()", style=solid];
-"640 lift_fresh_copy_23" -> "643 fill__23"  [label="()", style=solid];
-"641 slice_93" -> "642 slice_94"  [label="(4, 16)", style=solid];
-"642 slice_94" -> "643 fill__23"  [label="(4, 4)", style=solid];
-"644 _tensor_constant36" -> "645 lift_fresh_copy_24"  [label="()", style=solid];
-"645 lift_fresh_copy_24" -> "648 fill__24"  [label="()", style=solid];
-"646 slice_95" -> "647 slice_96"  [label="(4, 16)", style=solid];
-"647 slice_96" -> "648 fill__24"  [label="(4, 8)", style=solid];
-"649 _tensor_constant37" -> "650 lift_fresh_copy_25"  [label="()", style=solid];
-"650 lift_fresh_copy_25" -> "653 fill__25"  [label="()", style=solid];
-"651 slice_97" -> "652 slice_98"  [label="(4, 16)", style=solid];
-"652 slice_98" -> "653 fill__25"  [label="(4, 4)", style=solid];
-"654 _tensor_constant38" -> "655 lift_fresh_copy_26"  [label="()", style=solid];
-"655 lift_fresh_copy_26" -> "658 fill__26"  [label="()", style=solid];
-"656 slice_99" -> "657 slice_100"  [label="(4, 16)", style=solid];
-"657 slice_100" -> "658 fill__26"  [label="(4, 4)", style=solid];
-"659 view_29" -> "660 permute_26"  [label="(2, 8, 2, 8)", style=solid];
-"660 permute_26" -> "661 reshape_24"  [label="(2, 2, 8, 8)", style=solid];
-"661 reshape_24" -> "662 unsqueeze_14"  [label="(4, 64)", style=solid];
-"661 reshape_24" -> "663 unsqueeze_15"  [label="(4, 64)", style=solid];
-"662 unsqueeze_14" -> "664 sub_2"  [label="(4, 1, 64)", style=solid];
-"663 unsqueeze_15" -> "664 sub_2"  [label="(4, 64, 1)", style=solid];
-"664 sub_2" -> "665 ne_2"  [label="(4, 64, 64)", style=solid];
-"664 sub_2" -> "666 masked_fill_4"  [label="(4, 64, 64)", style=solid];
-"664 sub_2" -> "667 eq_2"  [label="(4, 64, 64)", style=solid];
-"665 ne_2" -> "666 masked_fill_4"  [label="(4, 64, 64)", style=solid];
-"666 masked_fill_4" -> "668 masked_fill_5"  [label="(4, 64, 64)", style=solid];
-"667 eq_2" -> "668 masked_fill_5"  [label="(4, 64, 64)", style=solid];
-"668 masked_fill_5" -> "670 unsqueeze_16"  [label="(4, 64, 64)", style=solid];
-"669 view_30" -> "672 add_18"  [label="(1, 4, 12, 64, 64)", style=solid];
-"670 unsqueeze_16" -> "671 unsqueeze_17"  [label="(4, 1, 64, 64)", style=solid];
-"671 unsqueeze_17" -> "672 add_18"  [label="(1, 4, 1, 64, 64)", style=solid];
-"672 add_18" -> "673 view_31"  [label="(1, 4, 12, 64, 64)", style=solid];
-"673 view_31" -> "674 softmax_5"  [label="(4, 12, 64, 64)", style=solid];
-"674 softmax_5" -> "675 dropout_20"  [label="(4, 12, 64, 64)", style=solid];
-"675 dropout_20" -> "676 matmul_11"  [label="(4, 12, 64, 64)", style=solid];
-"676 matmul_11" -> "677 transpose_11"  [label="(4, 12, 64, 32)", style=solid];
-"677 transpose_11" -> "678 reshape_25"  [label="(4, 64, 12, 32)", style=solid];
-"678 reshape_25" -> "681 linear_35"  [label="(4, 64, 384)", style=solid];
-"679 _param_constant96" -> "681 linear_35"  [label="(384, 384)", style=solid];
-"680 _param_constant97" -> "681 linear_35"  [label="(384,)", style=solid];
-"681 linear_35" -> "682 dropout_21"  [label="(4, 64, 384)", style=solid];
-"682 dropout_21" -> "683 view_32"  [label="(4, 64, 384)", style=solid];
-"683 view_32" -> "684 permute_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"684 permute_27" -> "685 reshape_26"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"685 reshape_26" -> "686 roll_5"  [label="(1, 16, 16, 384)", style=solid];
-"686 roll_5" -> "687 slice_101"  [label="(1, 16, 16, 384)", style=solid];
-"687 slice_101" -> "688 slice_102"  [label="(1, 16, 16, 384)", style=solid];
-"688 slice_102" -> "689 slice_103"  [label="(1, 14, 16, 384)", style=solid];
-"689 slice_103" -> "690 slice_104"  [label="(1, 14, 14, 384)", style=solid];
-"690 slice_104" -> "691 contiguous_9"  [label="(1, 14, 14, 384)", style=solid];
-"691 contiguous_9" -> "694 layer_norm_13"  [label="(1, 14, 14, 384)", style=solid];
-"692 _param_constant98" -> "694 layer_norm_13"  [label="(384,)", style=solid];
-"693 _param_constant99" -> "694 layer_norm_13"  [label="(384,)", style=solid];
-"694 layer_norm_13" -> "695 add_19"  [label="(1, 14, 14, 384)", style=solid];
-"695 add_19" -> "698 linear_36"  [label="(1, 14, 14, 384)", style=solid];
-"695 add_19" -> "708 add_20"  [label="(1, 14, 14, 384)", style=solid];
-"696 _param_constant100" -> "698 linear_36"  [label="(1536, 384)", style=solid];
-"697 _param_constant101" -> "698 linear_36"  [label="(1536,)", style=solid];
-"698 linear_36" -> "699 gelu_5"  [label="(1, 14, 14, 1536)", style=solid];
-"699 gelu_5" -> "700 dropout_22"  [label="(1, 14, 14, 1536)", style=solid];
-"700 dropout_22" -> "703 linear_37"  [label="(1, 14, 14, 1536)", style=solid];
-"701 _param_constant102" -> "703 linear_37"  [label="(384, 1536)", style=solid];
-"702 _param_constant103" -> "703 linear_37"  [label="(384,)", style=solid];
-"703 linear_37" -> "704 dropout_23"  [label="(1, 14, 14, 384)", style=solid];
-"704 dropout_23" -> "707 layer_norm_14"  [label="(1, 14, 14, 384)", style=solid];
-"705 _param_constant104" -> "707 layer_norm_14"  [label="(384,)", style=solid];
-"706 _param_constant105" -> "707 layer_norm_14"  [label="(384,)", style=solid];
-"707 layer_norm_14" -> "708 add_20"  [label="(1, 14, 14, 384)", style=solid];
-"708 add_20" -> "725 pad_8"  [label="(1, 14, 14, 384)", style=solid];
-"708 add_20" -> "775 add_22"  [label="(1, 14, 14, 384)", style=solid];
-"709 _tensor_constant39" -> "712 linear_38"  [label="(1, 15, 15, 2)", style=solid];
-"710 _param_constant106" -> "712 linear_38"  [label="(512, 2)", style=solid];
-"711 _param_constant107" -> "712 linear_38"  [label="(512,)", style=solid];
-"712 linear_38" -> "713 relu__6"  [label="(1, 15, 15, 512)", style=solid];
-"713 relu__6" -> "715 linear_39"  [label="(1, 15, 15, 512)", style=solid];
-"714 _param_constant108" -> "715 linear_39"  [label="(12, 512)", style=solid];
-"715 linear_39" -> "716 view_33"  [label="(1, 15, 15, 12)", style=solid];
-"716 view_33" -> "718 index_6"  [label="(225, 12)", style=solid];
-"717 _tensor_constant40" -> "718 index_6"  [label="(4096,)", style=solid];
-"718 index_6" -> "719 view_34"  [label="(4096, 12)", style=solid];
-"719 view_34" -> "720 permute_28"  [label="(64, 64, 12)", style=solid];
-"720 permute_28" -> "721 contiguous_10"  [label="(12, 64, 64)", style=solid];
-"721 contiguous_10" -> "722 unsqueeze_18"  [label="(12, 64, 64)", style=solid];
-"722 unsqueeze_18" -> "723 sigmoid_6"  [label="(1, 12, 64, 64)", style=solid];
-"723 sigmoid_6" -> "724 mul_12"  [label="(1, 12, 64, 64)", style=solid];
-"724 mul_12" -> "754 add_21"  [label="(1, 12, 64, 64)", style=solid];
-"725 pad_8" -> "726 view_35"  [label="(1, 16, 16, 384)", style=solid];
-"726 view_35" -> "727 permute_29"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"727 permute_29" -> "728 reshape_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"728 reshape_27" -> "734 linear_40"  [label="(4, 64, 384)", style=solid];
-"729 _param_constant109" -> "730 clone_6"  [label="(1152,)", style=solid];
-"730 clone_6" -> "731 slice_105"  [label="(1152,)", style=solid];
-"730 clone_6" -> "734 linear_40"  [label="(1152,)", style=solid];
-"731 slice_105" -> "732 zero__6"  [label="(384,)", style=solid];
-"733 _param_constant110" -> "734 linear_40"  [label="(1152, 384)", style=solid];
-"734 linear_40" -> "735 reshape_28"  [label="(4, 64, 1152)", style=solid];
-"735 reshape_28" -> "736 permute_30"  [label="(4, 64, 3, 12, 32)", style=solid];
-"736 permute_30" -> "737 select_18"  [label="(3, 4, 12, 64, 32)", style=solid];
-"736 permute_30" -> "738 select_19"  [label="(3, 4, 12, 64, 32)", style=solid];
-"736 permute_30" -> "739 select_20"  [label="(3, 4, 12, 64, 32)", style=solid];
-"737 select_18" -> "740 linalg_vector_norm_12"  [label="(4, 12, 64, 32)", style=solid];
-"737 select_18" -> "742 expand_as_12"  [label="(4, 12, 64, 32)", style=solid];
-"737 select_18" -> "743 div_12"  [label="(4, 12, 64, 32)", style=solid];
-"738 select_19" -> "744 linalg_vector_norm_13"  [label="(4, 12, 64, 32)", style=solid];
-"738 select_19" -> "746 expand_as_13"  [label="(4, 12, 64, 32)", style=solid];
-"738 select_19" -> "747 div_13"  [label="(4, 12, 64, 32)", style=solid];
-"739 select_20" -> "757 matmul_13"  [label="(4, 12, 64, 32)", style=solid];
-"740 linalg_vector_norm_12" -> "741 clamp_min_12"  [label="(4, 12, 64, 1)", style=solid];
-"741 clamp_min_12" -> "742 expand_as_12"  [label="(4, 12, 64, 1)", style=solid];
-"742 expand_as_12" -> "743 div_12"  [label="(4, 12, 64, 32)", style=solid];
-"743 div_12" -> "749 matmul_12"  [label="(4, 12, 64, 32)", style=solid];
-"744 linalg_vector_norm_13" -> "745 clamp_min_13"  [label="(4, 12, 64, 1)", style=solid];
-"745 clamp_min_13" -> "746 expand_as_13"  [label="(4, 12, 64, 1)", style=solid];
-"746 expand_as_13" -> "747 div_13"  [label="(4, 12, 64, 32)", style=solid];
-"747 div_13" -> "748 transpose_12"  [label="(4, 12, 64, 32)", style=solid];
-"748 transpose_12" -> "749 matmul_12"  [label="(4, 12, 32, 64)", style=solid];
-"749 matmul_12" -> "753 mul_13"  [label="(4, 12, 64, 64)", style=solid];
-"750 _param_constant111" -> "751 clamp_6"  [label="(12, 1, 1)", style=solid];
-"751 clamp_6" -> "752 exp_6"  [label="(12, 1, 1)", style=solid];
-"752 exp_6" -> "753 mul_13"  [label="(12, 1, 1)", style=solid];
-"753 mul_13" -> "754 add_21"  [label="(4, 12, 64, 64)", style=solid];
-"754 add_21" -> "755 softmax_6"  [label="(4, 12, 64, 64)", style=solid];
-"755 softmax_6" -> "756 dropout_24"  [label="(4, 12, 64, 64)", style=solid];
-"756 dropout_24" -> "757 matmul_13"  [label="(4, 12, 64, 64)", style=solid];
-"757 matmul_13" -> "758 transpose_13"  [label="(4, 12, 64, 32)", style=solid];
-"758 transpose_13" -> "759 reshape_29"  [label="(4, 64, 12, 32)", style=solid];
-"759 reshape_29" -> "762 linear_41"  [label="(4, 64, 384)", style=solid];
-"760 _param_constant112" -> "762 linear_41"  [label="(384, 384)", style=solid];
-"761 _param_constant113" -> "762 linear_41"  [label="(384,)", style=solid];
-"762 linear_41" -> "763 dropout_25"  [label="(4, 64, 384)", style=solid];
-"763 dropout_25" -> "764 view_36"  [label="(4, 64, 384)", style=solid];
-"764 view_36" -> "765 permute_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"765 permute_31" -> "766 reshape_30"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"766 reshape_30" -> "767 slice_106"  [label="(1, 16, 16, 384)", style=solid];
-"767 slice_106" -> "768 slice_107"  [label="(1, 16, 16, 384)", style=solid];
-"768 slice_107" -> "769 slice_108"  [label="(1, 14, 16, 384)", style=solid];
-"769 slice_108" -> "770 slice_109"  [label="(1, 14, 14, 384)", style=solid];
-"770 slice_109" -> "771 contiguous_11"  [label="(1, 14, 14, 384)", style=solid];
-"771 contiguous_11" -> "774 layer_norm_15"  [label="(1, 14, 14, 384)", style=solid];
-"772 _param_constant114" -> "774 layer_norm_15"  [label="(384,)", style=solid];
-"773 _param_constant115" -> "774 layer_norm_15"  [label="(384,)", style=solid];
-"774 layer_norm_15" -> "775 add_22"  [label="(1, 14, 14, 384)", style=solid];
-"775 add_22" -> "778 linear_42"  [label="(1, 14, 14, 384)", style=solid];
-"775 add_22" -> "788 add_23"  [label="(1, 14, 14, 384)", style=solid];
-"776 _param_constant116" -> "778 linear_42"  [label="(1536, 384)", style=solid];
-"777 _param_constant117" -> "778 linear_42"  [label="(1536,)", style=solid];
-"778 linear_42" -> "779 gelu_6"  [label="(1, 14, 14, 1536)", style=solid];
-"779 gelu_6" -> "780 dropout_26"  [label="(1, 14, 14, 1536)", style=solid];
-"780 dropout_26" -> "783 linear_43"  [label="(1, 14, 14, 1536)", style=solid];
-"781 _param_constant118" -> "783 linear_43"  [label="(384, 1536)", style=solid];
-"782 _param_constant119" -> "783 linear_43"  [label="(384,)", style=solid];
-"783 linear_43" -> "784 dropout_27"  [label="(1, 14, 14, 384)", style=solid];
-"784 dropout_27" -> "787 layer_norm_16"  [label="(1, 14, 14, 384)", style=solid];
-"785 _param_constant120" -> "787 layer_norm_16"  [label="(384,)", style=solid];
-"786 _param_constant121" -> "787 layer_norm_16"  [label="(384,)", style=solid];
-"787 layer_norm_16" -> "788 add_23"  [label="(1, 14, 14, 384)", style=solid];
-"788 add_23" -> "805 pad_9"  [label="(1, 14, 14, 384)", style=solid];
-"788 add_23" -> "918 add_26"  [label="(1, 14, 14, 384)", style=solid];
-"789 _tensor_constant41" -> "792 linear_44"  [label="(1, 15, 15, 2)", style=solid];
-"790 _param_constant122" -> "792 linear_44"  [label="(512, 2)", style=solid];
-"791 _param_constant123" -> "792 linear_44"  [label="(512,)", style=solid];
-"792 linear_44" -> "793 relu__7"  [label="(1, 15, 15, 512)", style=solid];
-"793 relu__7" -> "795 linear_45"  [label="(1, 15, 15, 512)", style=solid];
-"794 _param_constant124" -> "795 linear_45"  [label="(12, 512)", style=solid];
-"795 linear_45" -> "796 view_37"  [label="(1, 15, 15, 12)", style=solid];
-"796 view_37" -> "798 index_7"  [label="(225, 12)", style=solid];
-"797 _tensor_constant42" -> "798 index_7"  [label="(4096,)", style=solid];
-"798 index_7" -> "799 view_38"  [label="(4096, 12)", style=solid];
-"799 view_38" -> "800 permute_32"  [label="(64, 64, 12)", style=solid];
-"800 permute_32" -> "801 contiguous_12"  [label="(12, 64, 64)", style=solid];
-"801 contiguous_12" -> "802 unsqueeze_19"  [label="(12, 64, 64)", style=solid];
-"802 unsqueeze_19" -> "803 sigmoid_7"  [label="(1, 12, 64, 64)", style=solid];
-"803 sigmoid_7" -> "804 mul_14"  [label="(1, 12, 64, 64)", style=solid];
-"804 mul_14" -> "835 add_24"  [label="(1, 12, 64, 64)", style=solid];
-"805 pad_9" -> "806 roll_6"  [label="(1, 16, 16, 384)", style=solid];
-"806 roll_6" -> "807 view_39"  [label="(1, 16, 16, 384)", style=solid];
-"807 view_39" -> "808 permute_33"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"808 permute_33" -> "809 reshape_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"809 reshape_31" -> "815 linear_46"  [label="(4, 64, 384)", style=solid];
-"809 reshape_31" -> "836 new_zeros_3"  [label="(4, 64, 384)", style=solid];
-"810 _param_constant125" -> "811 clone_7"  [label="(1152,)", style=solid];
-"811 clone_7" -> "812 slice_110"  [label="(1152,)", style=solid];
-"811 clone_7" -> "815 linear_46"  [label="(1152,)", style=solid];
-"812 slice_110" -> "813 zero__7"  [label="(384,)", style=solid];
-"814 _param_constant126" -> "815 linear_46"  [label="(1152, 384)", style=solid];
-"815 linear_46" -> "816 reshape_32"  [label="(4, 64, 1152)", style=solid];
-"816 reshape_32" -> "817 permute_34"  [label="(4, 64, 3, 12, 32)", style=solid];
-"817 permute_34" -> "818 select_21"  [label="(3, 4, 12, 64, 32)", style=solid];
-"817 permute_34" -> "819 select_22"  [label="(3, 4, 12, 64, 32)", style=solid];
-"817 permute_34" -> "820 select_23"  [label="(3, 4, 12, 64, 32)", style=solid];
-"818 select_21" -> "821 linalg_vector_norm_14"  [label="(4, 12, 64, 32)", style=solid];
-"818 select_21" -> "823 expand_as_14"  [label="(4, 12, 64, 32)", style=solid];
-"818 select_21" -> "824 div_14"  [label="(4, 12, 64, 32)", style=solid];
-"819 select_22" -> "825 linalg_vector_norm_15"  [label="(4, 12, 64, 32)", style=solid];
-"819 select_22" -> "827 expand_as_15"  [label="(4, 12, 64, 32)", style=solid];
-"819 select_22" -> "828 div_15"  [label="(4, 12, 64, 32)", style=solid];
-"820 select_23" -> "899 matmul_15"  [label="(4, 12, 64, 32)", style=solid];
-"821 linalg_vector_norm_14" -> "822 clamp_min_14"  [label="(4, 12, 64, 1)", style=solid];
-"822 clamp_min_14" -> "823 expand_as_14"  [label="(4, 12, 64, 1)", style=solid];
-"823 expand_as_14" -> "824 div_14"  [label="(4, 12, 64, 32)", style=solid];
-"824 div_14" -> "830 matmul_14"  [label="(4, 12, 64, 32)", style=solid];
-"825 linalg_vector_norm_15" -> "826 clamp_min_15"  [label="(4, 12, 64, 1)", style=solid];
-"826 clamp_min_15" -> "827 expand_as_15"  [label="(4, 12, 64, 1)", style=solid];
-"827 expand_as_15" -> "828 div_15"  [label="(4, 12, 64, 32)", style=solid];
-"828 div_15" -> "829 transpose_14"  [label="(4, 12, 64, 32)", style=solid];
-"829 transpose_14" -> "830 matmul_14"  [label="(4, 12, 32, 64)", style=solid];
-"830 matmul_14" -> "834 mul_15"  [label="(4, 12, 64, 64)", style=solid];
-"831 _param_constant127" -> "832 clamp_7"  [label="(12, 1, 1)", style=solid];
-"832 clamp_7" -> "833 exp_7"  [label="(12, 1, 1)", style=solid];
-"833 exp_7" -> "834 mul_15"  [label="(12, 1, 1)", style=solid];
-"834 mul_15" -> "835 add_24"  [label="(4, 12, 64, 64)", style=solid];
-"835 add_24" -> "892 view_41"  [label="(4, 12, 64, 64)", style=solid];
-"836 new_zeros_3" -> "839 slice_111"  [label="(16, 16)", style=solid];
-"836 new_zeros_3" -> "844 slice_113"  [label="(16, 16)", style=solid];
-"836 new_zeros_3" -> "849 slice_115"  [label="(16, 16)", style=solid];
-"836 new_zeros_3" -> "854 slice_117"  [label="(16, 16)", style=solid];
-"836 new_zeros_3" -> "859 slice_119"  [label="(16, 16)", style=solid];
-"836 new_zeros_3" -> "864 slice_121"  [label="(16, 16)", style=solid];
-"836 new_zeros_3" -> "869 slice_123"  [label="(16, 16)", style=solid];
-"836 new_zeros_3" -> "874 slice_125"  [label="(16, 16)", style=solid];
-"836 new_zeros_3" -> "879 slice_127"  [label="(16, 16)", style=solid];
-"836 new_zeros_3" -> "882 view_40"  [label="(16, 16)", style=solid];
-"837 _tensor_constant43" -> "838 lift_fresh_copy_27"  [label="()", style=solid];
-"838 lift_fresh_copy_27" -> "841 fill__27"  [label="()", style=solid];
-"839 slice_111" -> "840 slice_112"  [label="(8, 16)", style=solid];
-"840 slice_112" -> "841 fill__27"  [label="(8, 8)", style=solid];
-"842 _tensor_constant44" -> "843 lift_fresh_copy_28"  [label="()", style=solid];
-"843 lift_fresh_copy_28" -> "846 fill__28"  [label="()", style=solid];
-"844 slice_113" -> "845 slice_114"  [label="(8, 16)", style=solid];
-"845 slice_114" -> "846 fill__28"  [label="(8, 4)", style=solid];
-"847 _tensor_constant45" -> "848 lift_fresh_copy_29"  [label="()", style=solid];
-"848 lift_fresh_copy_29" -> "851 fill__29"  [label="()", style=solid];
-"849 slice_115" -> "850 slice_116"  [label="(8, 16)", style=solid];
-"850 slice_116" -> "851 fill__29"  [label="(8, 4)", style=solid];
-"852 _tensor_constant46" -> "853 lift_fresh_copy_30"  [label="()", style=solid];
-"853 lift_fresh_copy_30" -> "856 fill__30"  [label="()", style=solid];
-"854 slice_117" -> "855 slice_118"  [label="(4, 16)", style=solid];
-"855 slice_118" -> "856 fill__30"  [label="(4, 8)", style=solid];
-"857 _tensor_constant47" -> "858 lift_fresh_copy_31"  [label="()", style=solid];
-"858 lift_fresh_copy_31" -> "861 fill__31"  [label="()", style=solid];
-"859 slice_119" -> "860 slice_120"  [label="(4, 16)", style=solid];
-"860 slice_120" -> "861 fill__31"  [label="(4, 4)", style=solid];
-"862 _tensor_constant48" -> "863 lift_fresh_copy_32"  [label="()", style=solid];
-"863 lift_fresh_copy_32" -> "866 fill__32"  [label="()", style=solid];
-"864 slice_121" -> "865 slice_122"  [label="(4, 16)", style=solid];
-"865 slice_122" -> "866 fill__32"  [label="(4, 4)", style=solid];
-"867 _tensor_constant49" -> "868 lift_fresh_copy_33"  [label="()", style=solid];
-"868 lift_fresh_copy_33" -> "871 fill__33"  [label="()", style=solid];
-"869 slice_123" -> "870 slice_124"  [label="(4, 16)", style=solid];
-"870 slice_124" -> "871 fill__33"  [label="(4, 8)", style=solid];
-"872 _tensor_constant50" -> "873 lift_fresh_copy_34"  [label="()", style=solid];
-"873 lift_fresh_copy_34" -> "876 fill__34"  [label="()", style=solid];
-"874 slice_125" -> "875 slice_126"  [label="(4, 16)", style=solid];
-"875 slice_126" -> "876 fill__34"  [label="(4, 4)", style=solid];
-"877 _tensor_constant51" -> "878 lift_fresh_copy_35"  [label="()", style=solid];
-"878 lift_fresh_copy_35" -> "881 fill__35"  [label="()", style=solid];
-"879 slice_127" -> "880 slice_128"  [label="(4, 16)", style=solid];
-"880 slice_128" -> "881 fill__35"  [label="(4, 4)", style=solid];
-"882 view_40" -> "883 permute_35"  [label="(2, 8, 2, 8)", style=solid];
-"883 permute_35" -> "884 reshape_33"  [label="(2, 2, 8, 8)", style=solid];
-"884 reshape_33" -> "885 unsqueeze_20"  [label="(4, 64)", style=solid];
-"884 reshape_33" -> "886 unsqueeze_21"  [label="(4, 64)", style=solid];
-"885 unsqueeze_20" -> "887 sub_3"  [label="(4, 1, 64)", style=solid];
-"886 unsqueeze_21" -> "887 sub_3"  [label="(4, 64, 1)", style=solid];
-"887 sub_3" -> "888 ne_3"  [label="(4, 64, 64)", style=solid];
-"887 sub_3" -> "889 masked_fill_6"  [label="(4, 64, 64)", style=solid];
-"887 sub_3" -> "890 eq_3"  [label="(4, 64, 64)", style=solid];
-"888 ne_3" -> "889 masked_fill_6"  [label="(4, 64, 64)", style=solid];
-"889 masked_fill_6" -> "891 masked_fill_7"  [label="(4, 64, 64)", style=solid];
-"890 eq_3" -> "891 masked_fill_7"  [label="(4, 64, 64)", style=solid];
-"891 masked_fill_7" -> "893 unsqueeze_22"  [label="(4, 64, 64)", style=solid];
-"892 view_41" -> "895 add_25"  [label="(1, 4, 12, 64, 64)", style=solid];
-"893 unsqueeze_22" -> "894 unsqueeze_23"  [label="(4, 1, 64, 64)", style=solid];
-"894 unsqueeze_23" -> "895 add_25"  [label="(1, 4, 1, 64, 64)", style=solid];
-"895 add_25" -> "896 view_42"  [label="(1, 4, 12, 64, 64)", style=solid];
-"896 view_42" -> "897 softmax_7"  [label="(4, 12, 64, 64)", style=solid];
-"897 softmax_7" -> "898 dropout_28"  [label="(4, 12, 64, 64)", style=solid];
-"898 dropout_28" -> "899 matmul_15"  [label="(4, 12, 64, 64)", style=solid];
-"899 matmul_15" -> "900 transpose_15"  [label="(4, 12, 64, 32)", style=solid];
-"900 transpose_15" -> "901 reshape_34"  [label="(4, 64, 12, 32)", style=solid];
-"901 reshape_34" -> "904 linear_47"  [label="(4, 64, 384)", style=solid];
-"902 _param_constant128" -> "904 linear_47"  [label="(384, 384)", style=solid];
-"903 _param_constant129" -> "904 linear_47"  [label="(384,)", style=solid];
-"904 linear_47" -> "905 dropout_29"  [label="(4, 64, 384)", style=solid];
-"905 dropout_29" -> "906 view_43"  [label="(4, 64, 384)", style=solid];
-"906 view_43" -> "907 permute_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"907 permute_36" -> "908 reshape_35"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"908 reshape_35" -> "909 roll_7"  [label="(1, 16, 16, 384)", style=solid];
-"909 roll_7" -> "910 slice_129"  [label="(1, 16, 16, 384)", style=solid];
-"910 slice_129" -> "911 slice_130"  [label="(1, 16, 16, 384)", style=solid];
-"911 slice_130" -> "912 slice_131"  [label="(1, 14, 16, 384)", style=solid];
-"912 slice_131" -> "913 slice_132"  [label="(1, 14, 14, 384)", style=solid];
-"913 slice_132" -> "914 contiguous_13"  [label="(1, 14, 14, 384)", style=solid];
-"914 contiguous_13" -> "917 layer_norm_17"  [label="(1, 14, 14, 384)", style=solid];
-"915 _param_constant130" -> "917 layer_norm_17"  [label="(384,)", style=solid];
-"916 _param_constant131" -> "917 layer_norm_17"  [label="(384,)", style=solid];
-"917 layer_norm_17" -> "918 add_26"  [label="(1, 14, 14, 384)", style=solid];
-"918 add_26" -> "921 linear_48"  [label="(1, 14, 14, 384)", style=solid];
-"918 add_26" -> "931 add_27"  [label="(1, 14, 14, 384)", style=solid];
-"919 _param_constant132" -> "921 linear_48"  [label="(1536, 384)", style=solid];
-"920 _param_constant133" -> "921 linear_48"  [label="(1536,)", style=solid];
-"921 linear_48" -> "922 gelu_7"  [label="(1, 14, 14, 1536)", style=solid];
-"922 gelu_7" -> "923 dropout_30"  [label="(1, 14, 14, 1536)", style=solid];
-"923 dropout_30" -> "926 linear_49"  [label="(1, 14, 14, 1536)", style=solid];
-"924 _param_constant134" -> "926 linear_49"  [label="(384, 1536)", style=solid];
-"925 _param_constant135" -> "926 linear_49"  [label="(384,)", style=solid];
-"926 linear_49" -> "927 dropout_31"  [label="(1, 14, 14, 384)", style=solid];
-"927 dropout_31" -> "930 layer_norm_18"  [label="(1, 14, 14, 384)", style=solid];
-"928 _param_constant136" -> "930 layer_norm_18"  [label="(384,)", style=solid];
-"929 _param_constant137" -> "930 layer_norm_18"  [label="(384,)", style=solid];
-"930 layer_norm_18" -> "931 add_27"  [label="(1, 14, 14, 384)", style=solid];
-"931 add_27" -> "948 pad_10"  [label="(1, 14, 14, 384)", style=solid];
-"931 add_27" -> "998 add_29"  [label="(1, 14, 14, 384)", style=solid];
-"932 _tensor_constant52" -> "935 linear_50"  [label="(1, 15, 15, 2)", style=solid];
-"933 _param_constant138" -> "935 linear_50"  [label="(512, 2)", style=solid];
-"934 _param_constant139" -> "935 linear_50"  [label="(512,)", style=solid];
-"935 linear_50" -> "936 relu__8"  [label="(1, 15, 15, 512)", style=solid];
-"936 relu__8" -> "938 linear_51"  [label="(1, 15, 15, 512)", style=solid];
-"937 _param_constant140" -> "938 linear_51"  [label="(12, 512)", style=solid];
-"938 linear_51" -> "939 view_44"  [label="(1, 15, 15, 12)", style=solid];
-"939 view_44" -> "941 index_8"  [label="(225, 12)", style=solid];
-"940 _tensor_constant53" -> "941 index_8"  [label="(4096,)", style=solid];
-"941 index_8" -> "942 view_45"  [label="(4096, 12)", style=solid];
-"942 view_45" -> "943 permute_37"  [label="(64, 64, 12)", style=solid];
-"943 permute_37" -> "944 contiguous_14"  [label="(12, 64, 64)", style=solid];
-"944 contiguous_14" -> "945 unsqueeze_24"  [label="(12, 64, 64)", style=solid];
-"945 unsqueeze_24" -> "946 sigmoid_8"  [label="(1, 12, 64, 64)", style=solid];
-"946 sigmoid_8" -> "947 mul_16"  [label="(1, 12, 64, 64)", style=solid];
-"947 mul_16" -> "977 add_28"  [label="(1, 12, 64, 64)", style=solid];
-"948 pad_10" -> "949 view_46"  [label="(1, 16, 16, 384)", style=solid];
-"949 view_46" -> "950 permute_38"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"950 permute_38" -> "951 reshape_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"951 reshape_36" -> "957 linear_52"  [label="(4, 64, 384)", style=solid];
-"952 _param_constant141" -> "953 clone_8"  [label="(1152,)", style=solid];
-"953 clone_8" -> "954 slice_133"  [label="(1152,)", style=solid];
-"953 clone_8" -> "957 linear_52"  [label="(1152,)", style=solid];
-"954 slice_133" -> "955 zero__8"  [label="(384,)", style=solid];
-"956 _param_constant142" -> "957 linear_52"  [label="(1152, 384)", style=solid];
-"957 linear_52" -> "958 reshape_37"  [label="(4, 64, 1152)", style=solid];
-"958 reshape_37" -> "959 permute_39"  [label="(4, 64, 3, 12, 32)", style=solid];
-"959 permute_39" -> "960 select_24"  [label="(3, 4, 12, 64, 32)", style=solid];
-"959 permute_39" -> "961 select_25"  [label="(3, 4, 12, 64, 32)", style=solid];
-"959 permute_39" -> "962 select_26"  [label="(3, 4, 12, 64, 32)", style=solid];
-"960 select_24" -> "963 linalg_vector_norm_16"  [label="(4, 12, 64, 32)", style=solid];
-"960 select_24" -> "965 expand_as_16"  [label="(4, 12, 64, 32)", style=solid];
-"960 select_24" -> "966 div_16"  [label="(4, 12, 64, 32)", style=solid];
-"961 select_25" -> "967 linalg_vector_norm_17"  [label="(4, 12, 64, 32)", style=solid];
-"961 select_25" -> "969 expand_as_17"  [label="(4, 12, 64, 32)", style=solid];
-"961 select_25" -> "970 div_17"  [label="(4, 12, 64, 32)", style=solid];
-"962 select_26" -> "980 matmul_17"  [label="(4, 12, 64, 32)", style=solid];
-"963 linalg_vector_norm_16" -> "964 clamp_min_16"  [label="(4, 12, 64, 1)", style=solid];
-"964 clamp_min_16" -> "965 expand_as_16"  [label="(4, 12, 64, 1)", style=solid];
-"965 expand_as_16" -> "966 div_16"  [label="(4, 12, 64, 32)", style=solid];
-"966 div_16" -> "972 matmul_16"  [label="(4, 12, 64, 32)", style=solid];
-"967 linalg_vector_norm_17" -> "968 clamp_min_17"  [label="(4, 12, 64, 1)", style=solid];
-"968 clamp_min_17" -> "969 expand_as_17"  [label="(4, 12, 64, 1)", style=solid];
-"969 expand_as_17" -> "970 div_17"  [label="(4, 12, 64, 32)", style=solid];
-"970 div_17" -> "971 transpose_16"  [label="(4, 12, 64, 32)", style=solid];
-"971 transpose_16" -> "972 matmul_16"  [label="(4, 12, 32, 64)", style=solid];
-"972 matmul_16" -> "976 mul_17"  [label="(4, 12, 64, 64)", style=solid];
-"973 _param_constant143" -> "974 clamp_8"  [label="(12, 1, 1)", style=solid];
-"974 clamp_8" -> "975 exp_8"  [label="(12, 1, 1)", style=solid];
-"975 exp_8" -> "976 mul_17"  [label="(12, 1, 1)", style=solid];
-"976 mul_17" -> "977 add_28"  [label="(4, 12, 64, 64)", style=solid];
-"977 add_28" -> "978 softmax_8"  [label="(4, 12, 64, 64)", style=solid];
-"978 softmax_8" -> "979 dropout_32"  [label="(4, 12, 64, 64)", style=solid];
-"979 dropout_32" -> "980 matmul_17"  [label="(4, 12, 64, 64)", style=solid];
-"980 matmul_17" -> "981 transpose_17"  [label="(4, 12, 64, 32)", style=solid];
-"981 transpose_17" -> "982 reshape_38"  [label="(4, 64, 12, 32)", style=solid];
-"982 reshape_38" -> "985 linear_53"  [label="(4, 64, 384)", style=solid];
-"983 _param_constant144" -> "985 linear_53"  [label="(384, 384)", style=solid];
-"984 _param_constant145" -> "985 linear_53"  [label="(384,)", style=solid];
-"985 linear_53" -> "986 dropout_33"  [label="(4, 64, 384)", style=solid];
-"986 dropout_33" -> "987 view_47"  [label="(4, 64, 384)", style=solid];
-"987 view_47" -> "988 permute_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"988 permute_40" -> "989 reshape_39"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"989 reshape_39" -> "990 slice_134"  [label="(1, 16, 16, 384)", style=solid];
-"990 slice_134" -> "991 slice_135"  [label="(1, 16, 16, 384)", style=solid];
-"991 slice_135" -> "992 slice_136"  [label="(1, 14, 16, 384)", style=solid];
-"992 slice_136" -> "993 slice_137"  [label="(1, 14, 14, 384)", style=solid];
-"993 slice_137" -> "994 contiguous_15"  [label="(1, 14, 14, 384)", style=solid];
-"994 contiguous_15" -> "997 layer_norm_19"  [label="(1, 14, 14, 384)", style=solid];
-"995 _param_constant146" -> "997 layer_norm_19"  [label="(384,)", style=solid];
-"996 _param_constant147" -> "997 layer_norm_19"  [label="(384,)", style=solid];
-"997 layer_norm_19" -> "998 add_29"  [label="(1, 14, 14, 384)", style=solid];
-"998 add_29" -> "1001 linear_54"  [label="(1, 14, 14, 384)", style=solid];
-"998 add_29" -> "1011 add_30"  [label="(1, 14, 14, 384)", style=solid];
-"999 _param_constant148" -> "1001 linear_54"  [label="(1536, 384)", style=solid];
-"1000 _param_constant149" -> "1001 linear_54"  [label="(1536,)", style=solid];
-"1001 linear_54" -> "1002 gelu_8"  [label="(1, 14, 14, 1536)", style=solid];
-"1002 gelu_8" -> "1003 dropout_34"  [label="(1, 14, 14, 1536)", style=solid];
-"1003 dropout_34" -> "1006 linear_55"  [label="(1, 14, 14, 1536)", style=solid];
-"1004 _param_constant150" -> "1006 linear_55"  [label="(384, 1536)", style=solid];
-"1005 _param_constant151" -> "1006 linear_55"  [label="(384,)", style=solid];
-"1006 linear_55" -> "1007 dropout_35"  [label="(1, 14, 14, 384)", style=solid];
-"1007 dropout_35" -> "1010 layer_norm_20"  [label="(1, 14, 14, 384)", style=solid];
-"1008 _param_constant152" -> "1010 layer_norm_20"  [label="(384,)", style=solid];
-"1009 _param_constant153" -> "1010 layer_norm_20"  [label="(384,)", style=solid];
-"1010 layer_norm_20" -> "1011 add_30"  [label="(1, 14, 14, 384)", style=solid];
-"1011 add_30" -> "1028 pad_11"  [label="(1, 14, 14, 384)", style=solid];
-"1011 add_30" -> "1141 add_33"  [label="(1, 14, 14, 384)", style=solid];
-"1012 _tensor_constant54" -> "1015 linear_56"  [label="(1, 15, 15, 2)", style=solid];
-"1013 _param_constant154" -> "1015 linear_56"  [label="(512, 2)", style=solid];
-"1014 _param_constant155" -> "1015 linear_56"  [label="(512,)", style=solid];
-"1015 linear_56" -> "1016 relu__9"  [label="(1, 15, 15, 512)", style=solid];
-"1016 relu__9" -> "1018 linear_57"  [label="(1, 15, 15, 512)", style=solid];
-"1017 _param_constant156" -> "1018 linear_57"  [label="(12, 512)", style=solid];
-"1018 linear_57" -> "1019 view_48"  [label="(1, 15, 15, 12)", style=solid];
-"1019 view_48" -> "1021 index_9"  [label="(225, 12)", style=solid];
-"1020 _tensor_constant55" -> "1021 index_9"  [label="(4096,)", style=solid];
-"1021 index_9" -> "1022 view_49"  [label="(4096, 12)", style=solid];
-"1022 view_49" -> "1023 permute_41"  [label="(64, 64, 12)", style=solid];
-"1023 permute_41" -> "1024 contiguous_16"  [label="(12, 64, 64)", style=solid];
-"1024 contiguous_16" -> "1025 unsqueeze_25"  [label="(12, 64, 64)", style=solid];
-"1025 unsqueeze_25" -> "1026 sigmoid_9"  [label="(1, 12, 64, 64)", style=solid];
-"1026 sigmoid_9" -> "1027 mul_18"  [label="(1, 12, 64, 64)", style=solid];
-"1027 mul_18" -> "1058 add_31"  [label="(1, 12, 64, 64)", style=solid];
-"1028 pad_11" -> "1029 roll_8"  [label="(1, 16, 16, 384)", style=solid];
-"1029 roll_8" -> "1030 view_50"  [label="(1, 16, 16, 384)", style=solid];
-"1030 view_50" -> "1031 permute_42"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1031 permute_42" -> "1032 reshape_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1032 reshape_40" -> "1038 linear_58"  [label="(4, 64, 384)", style=solid];
-"1032 reshape_40" -> "1059 new_zeros_4"  [label="(4, 64, 384)", style=solid];
-"1033 _param_constant157" -> "1034 clone_9"  [label="(1152,)", style=solid];
-"1034 clone_9" -> "1035 slice_138"  [label="(1152,)", style=solid];
-"1034 clone_9" -> "1038 linear_58"  [label="(1152,)", style=solid];
-"1035 slice_138" -> "1036 zero__9"  [label="(384,)", style=solid];
-"1037 _param_constant158" -> "1038 linear_58"  [label="(1152, 384)", style=solid];
-"1038 linear_58" -> "1039 reshape_41"  [label="(4, 64, 1152)", style=solid];
-"1039 reshape_41" -> "1040 permute_43"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1040 permute_43" -> "1041 select_27"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1040 permute_43" -> "1042 select_28"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1040 permute_43" -> "1043 select_29"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1041 select_27" -> "1044 linalg_vector_norm_18"  [label="(4, 12, 64, 32)", style=solid];
-"1041 select_27" -> "1046 expand_as_18"  [label="(4, 12, 64, 32)", style=solid];
-"1041 select_27" -> "1047 div_18"  [label="(4, 12, 64, 32)", style=solid];
-"1042 select_28" -> "1048 linalg_vector_norm_19"  [label="(4, 12, 64, 32)", style=solid];
-"1042 select_28" -> "1050 expand_as_19"  [label="(4, 12, 64, 32)", style=solid];
-"1042 select_28" -> "1051 div_19"  [label="(4, 12, 64, 32)", style=solid];
-"1043 select_29" -> "1122 matmul_19"  [label="(4, 12, 64, 32)", style=solid];
-"1044 linalg_vector_norm_18" -> "1045 clamp_min_18"  [label="(4, 12, 64, 1)", style=solid];
-"1045 clamp_min_18" -> "1046 expand_as_18"  [label="(4, 12, 64, 1)", style=solid];
-"1046 expand_as_18" -> "1047 div_18"  [label="(4, 12, 64, 32)", style=solid];
-"1047 div_18" -> "1053 matmul_18"  [label="(4, 12, 64, 32)", style=solid];
-"1048 linalg_vector_norm_19" -> "1049 clamp_min_19"  [label="(4, 12, 64, 1)", style=solid];
-"1049 clamp_min_19" -> "1050 expand_as_19"  [label="(4, 12, 64, 1)", style=solid];
-"1050 expand_as_19" -> "1051 div_19"  [label="(4, 12, 64, 32)", style=solid];
-"1051 div_19" -> "1052 transpose_18"  [label="(4, 12, 64, 32)", style=solid];
-"1052 transpose_18" -> "1053 matmul_18"  [label="(4, 12, 32, 64)", style=solid];
-"1053 matmul_18" -> "1057 mul_19"  [label="(4, 12, 64, 64)", style=solid];
-"1054 _param_constant159" -> "1055 clamp_9"  [label="(12, 1, 1)", style=solid];
-"1055 clamp_9" -> "1056 exp_9"  [label="(12, 1, 1)", style=solid];
-"1056 exp_9" -> "1057 mul_19"  [label="(12, 1, 1)", style=solid];
-"1057 mul_19" -> "1058 add_31"  [label="(4, 12, 64, 64)", style=solid];
-"1058 add_31" -> "1115 view_52"  [label="(4, 12, 64, 64)", style=solid];
-"1059 new_zeros_4" -> "1062 slice_139"  [label="(16, 16)", style=solid];
-"1059 new_zeros_4" -> "1067 slice_141"  [label="(16, 16)", style=solid];
-"1059 new_zeros_4" -> "1072 slice_143"  [label="(16, 16)", style=solid];
-"1059 new_zeros_4" -> "1077 slice_145"  [label="(16, 16)", style=solid];
-"1059 new_zeros_4" -> "1082 slice_147"  [label="(16, 16)", style=solid];
-"1059 new_zeros_4" -> "1087 slice_149"  [label="(16, 16)", style=solid];
-"1059 new_zeros_4" -> "1092 slice_151"  [label="(16, 16)", style=solid];
-"1059 new_zeros_4" -> "1097 slice_153"  [label="(16, 16)", style=solid];
-"1059 new_zeros_4" -> "1102 slice_155"  [label="(16, 16)", style=solid];
-"1059 new_zeros_4" -> "1105 view_51"  [label="(16, 16)", style=solid];
-"1060 _tensor_constant56" -> "1061 lift_fresh_copy_36"  [label="()", style=solid];
-"1061 lift_fresh_copy_36" -> "1064 fill__36"  [label="()", style=solid];
-"1062 slice_139" -> "1063 slice_140"  [label="(8, 16)", style=solid];
-"1063 slice_140" -> "1064 fill__36"  [label="(8, 8)", style=solid];
-"1065 _tensor_constant57" -> "1066 lift_fresh_copy_37"  [label="()", style=solid];
-"1066 lift_fresh_copy_37" -> "1069 fill__37"  [label="()", style=solid];
-"1067 slice_141" -> "1068 slice_142"  [label="(8, 16)", style=solid];
-"1068 slice_142" -> "1069 fill__37"  [label="(8, 4)", style=solid];
-"1070 _tensor_constant58" -> "1071 lift_fresh_copy_38"  [label="()", style=solid];
-"1071 lift_fresh_copy_38" -> "1074 fill__38"  [label="()", style=solid];
-"1072 slice_143" -> "1073 slice_144"  [label="(8, 16)", style=solid];
-"1073 slice_144" -> "1074 fill__38"  [label="(8, 4)", style=solid];
-"1075 _tensor_constant59" -> "1076 lift_fresh_copy_39"  [label="()", style=solid];
-"1076 lift_fresh_copy_39" -> "1079 fill__39"  [label="()", style=solid];
-"1077 slice_145" -> "1078 slice_146"  [label="(4, 16)", style=solid];
-"1078 slice_146" -> "1079 fill__39"  [label="(4, 8)", style=solid];
-"1080 _tensor_constant60" -> "1081 lift_fresh_copy_40"  [label="()", style=solid];
-"1081 lift_fresh_copy_40" -> "1084 fill__40"  [label="()", style=solid];
-"1082 slice_147" -> "1083 slice_148"  [label="(4, 16)", style=solid];
-"1083 slice_148" -> "1084 fill__40"  [label="(4, 4)", style=solid];
-"1085 _tensor_constant61" -> "1086 lift_fresh_copy_41"  [label="()", style=solid];
-"1086 lift_fresh_copy_41" -> "1089 fill__41"  [label="()", style=solid];
-"1087 slice_149" -> "1088 slice_150"  [label="(4, 16)", style=solid];
-"1088 slice_150" -> "1089 fill__41"  [label="(4, 4)", style=solid];
-"1090 _tensor_constant62" -> "1091 lift_fresh_copy_42"  [label="()", style=solid];
-"1091 lift_fresh_copy_42" -> "1094 fill__42"  [label="()", style=solid];
-"1092 slice_151" -> "1093 slice_152"  [label="(4, 16)", style=solid];
-"1093 slice_152" -> "1094 fill__42"  [label="(4, 8)", style=solid];
-"1095 _tensor_constant63" -> "1096 lift_fresh_copy_43"  [label="()", style=solid];
-"1096 lift_fresh_copy_43" -> "1099 fill__43"  [label="()", style=solid];
-"1097 slice_153" -> "1098 slice_154"  [label="(4, 16)", style=solid];
-"1098 slice_154" -> "1099 fill__43"  [label="(4, 4)", style=solid];
-"1100 _tensor_constant64" -> "1101 lift_fresh_copy_44"  [label="()", style=solid];
-"1101 lift_fresh_copy_44" -> "1104 fill__44"  [label="()", style=solid];
-"1102 slice_155" -> "1103 slice_156"  [label="(4, 16)", style=solid];
-"1103 slice_156" -> "1104 fill__44"  [label="(4, 4)", style=solid];
-"1105 view_51" -> "1106 permute_44"  [label="(2, 8, 2, 8)", style=solid];
-"1106 permute_44" -> "1107 reshape_42"  [label="(2, 2, 8, 8)", style=solid];
-"1107 reshape_42" -> "1108 unsqueeze_26"  [label="(4, 64)", style=solid];
-"1107 reshape_42" -> "1109 unsqueeze_27"  [label="(4, 64)", style=solid];
-"1108 unsqueeze_26" -> "1110 sub_4"  [label="(4, 1, 64)", style=solid];
-"1109 unsqueeze_27" -> "1110 sub_4"  [label="(4, 64, 1)", style=solid];
-"1110 sub_4" -> "1111 ne_4"  [label="(4, 64, 64)", style=solid];
-"1110 sub_4" -> "1112 masked_fill_8"  [label="(4, 64, 64)", style=solid];
-"1110 sub_4" -> "1113 eq_4"  [label="(4, 64, 64)", style=solid];
-"1111 ne_4" -> "1112 masked_fill_8"  [label="(4, 64, 64)", style=solid];
-"1112 masked_fill_8" -> "1114 masked_fill_9"  [label="(4, 64, 64)", style=solid];
-"1113 eq_4" -> "1114 masked_fill_9"  [label="(4, 64, 64)", style=solid];
-"1114 masked_fill_9" -> "1116 unsqueeze_28"  [label="(4, 64, 64)", style=solid];
-"1115 view_52" -> "1118 add_32"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1116 unsqueeze_28" -> "1117 unsqueeze_29"  [label="(4, 1, 64, 64)", style=solid];
-"1117 unsqueeze_29" -> "1118 add_32"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1118 add_32" -> "1119 view_53"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1119 view_53" -> "1120 softmax_9"  [label="(4, 12, 64, 64)", style=solid];
-"1120 softmax_9" -> "1121 dropout_36"  [label="(4, 12, 64, 64)", style=solid];
-"1121 dropout_36" -> "1122 matmul_19"  [label="(4, 12, 64, 64)", style=solid];
-"1122 matmul_19" -> "1123 transpose_19"  [label="(4, 12, 64, 32)", style=solid];
-"1123 transpose_19" -> "1124 reshape_43"  [label="(4, 64, 12, 32)", style=solid];
-"1124 reshape_43" -> "1127 linear_59"  [label="(4, 64, 384)", style=solid];
-"1125 _param_constant160" -> "1127 linear_59"  [label="(384, 384)", style=solid];
-"1126 _param_constant161" -> "1127 linear_59"  [label="(384,)", style=solid];
-"1127 linear_59" -> "1128 dropout_37"  [label="(4, 64, 384)", style=solid];
-"1128 dropout_37" -> "1129 view_54"  [label="(4, 64, 384)", style=solid];
-"1129 view_54" -> "1130 permute_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1130 permute_45" -> "1131 reshape_44"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1131 reshape_44" -> "1132 roll_9"  [label="(1, 16, 16, 384)", style=solid];
-"1132 roll_9" -> "1133 slice_157"  [label="(1, 16, 16, 384)", style=solid];
-"1133 slice_157" -> "1134 slice_158"  [label="(1, 16, 16, 384)", style=solid];
-"1134 slice_158" -> "1135 slice_159"  [label="(1, 14, 16, 384)", style=solid];
-"1135 slice_159" -> "1136 slice_160"  [label="(1, 14, 14, 384)", style=solid];
-"1136 slice_160" -> "1137 contiguous_17"  [label="(1, 14, 14, 384)", style=solid];
-"1137 contiguous_17" -> "1140 layer_norm_21"  [label="(1, 14, 14, 384)", style=solid];
-"1138 _param_constant162" -> "1140 layer_norm_21"  [label="(384,)", style=solid];
-"1139 _param_constant163" -> "1140 layer_norm_21"  [label="(384,)", style=solid];
-"1140 layer_norm_21" -> "1141 add_33"  [label="(1, 14, 14, 384)", style=solid];
-"1141 add_33" -> "1144 linear_60"  [label="(1, 14, 14, 384)", style=solid];
-"1141 add_33" -> "1154 add_34"  [label="(1, 14, 14, 384)", style=solid];
-"1142 _param_constant164" -> "1144 linear_60"  [label="(1536, 384)", style=solid];
-"1143 _param_constant165" -> "1144 linear_60"  [label="(1536,)", style=solid];
-"1144 linear_60" -> "1145 gelu_9"  [label="(1, 14, 14, 1536)", style=solid];
-"1145 gelu_9" -> "1146 dropout_38"  [label="(1, 14, 14, 1536)", style=solid];
-"1146 dropout_38" -> "1149 linear_61"  [label="(1, 14, 14, 1536)", style=solid];
-"1147 _param_constant166" -> "1149 linear_61"  [label="(384, 1536)", style=solid];
-"1148 _param_constant167" -> "1149 linear_61"  [label="(384,)", style=solid];
-"1149 linear_61" -> "1150 dropout_39"  [label="(1, 14, 14, 384)", style=solid];
-"1150 dropout_39" -> "1153 layer_norm_22"  [label="(1, 14, 14, 384)", style=solid];
-"1151 _param_constant168" -> "1153 layer_norm_22"  [label="(384,)", style=solid];
-"1152 _param_constant169" -> "1153 layer_norm_22"  [label="(384,)", style=solid];
-"1153 layer_norm_22" -> "1154 add_34"  [label="(1, 14, 14, 384)", style=solid];
-"1154 add_34" -> "1171 pad_12"  [label="(1, 14, 14, 384)", style=solid];
-"1154 add_34" -> "1221 add_36"  [label="(1, 14, 14, 384)", style=solid];
-"1155 _tensor_constant65" -> "1158 linear_62"  [label="(1, 15, 15, 2)", style=solid];
-"1156 _param_constant170" -> "1158 linear_62"  [label="(512, 2)", style=solid];
-"1157 _param_constant171" -> "1158 linear_62"  [label="(512,)", style=solid];
-"1158 linear_62" -> "1159 relu__10"  [label="(1, 15, 15, 512)", style=solid];
-"1159 relu__10" -> "1161 linear_63"  [label="(1, 15, 15, 512)", style=solid];
-"1160 _param_constant172" -> "1161 linear_63"  [label="(12, 512)", style=solid];
-"1161 linear_63" -> "1162 view_55"  [label="(1, 15, 15, 12)", style=solid];
-"1162 view_55" -> "1164 index_10"  [label="(225, 12)", style=solid];
-"1163 _tensor_constant66" -> "1164 index_10"  [label="(4096,)", style=solid];
-"1164 index_10" -> "1165 view_56"  [label="(4096, 12)", style=solid];
-"1165 view_56" -> "1166 permute_46"  [label="(64, 64, 12)", style=solid];
-"1166 permute_46" -> "1167 contiguous_18"  [label="(12, 64, 64)", style=solid];
-"1167 contiguous_18" -> "1168 unsqueeze_30"  [label="(12, 64, 64)", style=solid];
-"1168 unsqueeze_30" -> "1169 sigmoid_10"  [label="(1, 12, 64, 64)", style=solid];
-"1169 sigmoid_10" -> "1170 mul_20"  [label="(1, 12, 64, 64)", style=solid];
-"1170 mul_20" -> "1200 add_35"  [label="(1, 12, 64, 64)", style=solid];
-"1171 pad_12" -> "1172 view_57"  [label="(1, 16, 16, 384)", style=solid];
-"1172 view_57" -> "1173 permute_47"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1173 permute_47" -> "1174 reshape_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1174 reshape_45" -> "1180 linear_64"  [label="(4, 64, 384)", style=solid];
-"1175 _param_constant173" -> "1176 clone_10"  [label="(1152,)", style=solid];
-"1176 clone_10" -> "1177 slice_161"  [label="(1152,)", style=solid];
-"1176 clone_10" -> "1180 linear_64"  [label="(1152,)", style=solid];
-"1177 slice_161" -> "1178 zero__10"  [label="(384,)", style=solid];
-"1179 _param_constant174" -> "1180 linear_64"  [label="(1152, 384)", style=solid];
-"1180 linear_64" -> "1181 reshape_46"  [label="(4, 64, 1152)", style=solid];
-"1181 reshape_46" -> "1182 permute_48"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1182 permute_48" -> "1183 select_30"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1182 permute_48" -> "1184 select_31"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1182 permute_48" -> "1185 select_32"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1183 select_30" -> "1186 linalg_vector_norm_20"  [label="(4, 12, 64, 32)", style=solid];
-"1183 select_30" -> "1188 expand_as_20"  [label="(4, 12, 64, 32)", style=solid];
-"1183 select_30" -> "1189 div_20"  [label="(4, 12, 64, 32)", style=solid];
-"1184 select_31" -> "1190 linalg_vector_norm_21"  [label="(4, 12, 64, 32)", style=solid];
-"1184 select_31" -> "1192 expand_as_21"  [label="(4, 12, 64, 32)", style=solid];
-"1184 select_31" -> "1193 div_21"  [label="(4, 12, 64, 32)", style=solid];
-"1185 select_32" -> "1203 matmul_21"  [label="(4, 12, 64, 32)", style=solid];
-"1186 linalg_vector_norm_20" -> "1187 clamp_min_20"  [label="(4, 12, 64, 1)", style=solid];
-"1187 clamp_min_20" -> "1188 expand_as_20"  [label="(4, 12, 64, 1)", style=solid];
-"1188 expand_as_20" -> "1189 div_20"  [label="(4, 12, 64, 32)", style=solid];
-"1189 div_20" -> "1195 matmul_20"  [label="(4, 12, 64, 32)", style=solid];
-"1190 linalg_vector_norm_21" -> "1191 clamp_min_21"  [label="(4, 12, 64, 1)", style=solid];
-"1191 clamp_min_21" -> "1192 expand_as_21"  [label="(4, 12, 64, 1)", style=solid];
-"1192 expand_as_21" -> "1193 div_21"  [label="(4, 12, 64, 32)", style=solid];
-"1193 div_21" -> "1194 transpose_20"  [label="(4, 12, 64, 32)", style=solid];
-"1194 transpose_20" -> "1195 matmul_20"  [label="(4, 12, 32, 64)", style=solid];
-"1195 matmul_20" -> "1199 mul_21"  [label="(4, 12, 64, 64)", style=solid];
-"1196 _param_constant175" -> "1197 clamp_10"  [label="(12, 1, 1)", style=solid];
-"1197 clamp_10" -> "1198 exp_10"  [label="(12, 1, 1)", style=solid];
-"1198 exp_10" -> "1199 mul_21"  [label="(12, 1, 1)", style=solid];
-"1199 mul_21" -> "1200 add_35"  [label="(4, 12, 64, 64)", style=solid];
-"1200 add_35" -> "1201 softmax_10"  [label="(4, 12, 64, 64)", style=solid];
-"1201 softmax_10" -> "1202 dropout_40"  [label="(4, 12, 64, 64)", style=solid];
-"1202 dropout_40" -> "1203 matmul_21"  [label="(4, 12, 64, 64)", style=solid];
-"1203 matmul_21" -> "1204 transpose_21"  [label="(4, 12, 64, 32)", style=solid];
-"1204 transpose_21" -> "1205 reshape_47"  [label="(4, 64, 12, 32)", style=solid];
-"1205 reshape_47" -> "1208 linear_65"  [label="(4, 64, 384)", style=solid];
-"1206 _param_constant176" -> "1208 linear_65"  [label="(384, 384)", style=solid];
-"1207 _param_constant177" -> "1208 linear_65"  [label="(384,)", style=solid];
-"1208 linear_65" -> "1209 dropout_41"  [label="(4, 64, 384)", style=solid];
-"1209 dropout_41" -> "1210 view_58"  [label="(4, 64, 384)", style=solid];
-"1210 view_58" -> "1211 permute_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1211 permute_49" -> "1212 reshape_48"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1212 reshape_48" -> "1213 slice_162"  [label="(1, 16, 16, 384)", style=solid];
-"1213 slice_162" -> "1214 slice_163"  [label="(1, 16, 16, 384)", style=solid];
-"1214 slice_163" -> "1215 slice_164"  [label="(1, 14, 16, 384)", style=solid];
-"1215 slice_164" -> "1216 slice_165"  [label="(1, 14, 14, 384)", style=solid];
-"1216 slice_165" -> "1217 contiguous_19"  [label="(1, 14, 14, 384)", style=solid];
-"1217 contiguous_19" -> "1220 layer_norm_23"  [label="(1, 14, 14, 384)", style=solid];
-"1218 _param_constant178" -> "1220 layer_norm_23"  [label="(384,)", style=solid];
-"1219 _param_constant179" -> "1220 layer_norm_23"  [label="(384,)", style=solid];
-"1220 layer_norm_23" -> "1221 add_36"  [label="(1, 14, 14, 384)", style=solid];
-"1221 add_36" -> "1224 linear_66"  [label="(1, 14, 14, 384)", style=solid];
-"1221 add_36" -> "1234 add_37"  [label="(1, 14, 14, 384)", style=solid];
-"1222 _param_constant180" -> "1224 linear_66"  [label="(1536, 384)", style=solid];
-"1223 _param_constant181" -> "1224 linear_66"  [label="(1536,)", style=solid];
-"1224 linear_66" -> "1225 gelu_10"  [label="(1, 14, 14, 1536)", style=solid];
-"1225 gelu_10" -> "1226 dropout_42"  [label="(1, 14, 14, 1536)", style=solid];
-"1226 dropout_42" -> "1229 linear_67"  [label="(1, 14, 14, 1536)", style=solid];
-"1227 _param_constant182" -> "1229 linear_67"  [label="(384, 1536)", style=solid];
-"1228 _param_constant183" -> "1229 linear_67"  [label="(384,)", style=solid];
-"1229 linear_67" -> "1230 dropout_43"  [label="(1, 14, 14, 384)", style=solid];
-"1230 dropout_43" -> "1233 layer_norm_24"  [label="(1, 14, 14, 384)", style=solid];
-"1231 _param_constant184" -> "1233 layer_norm_24"  [label="(384,)", style=solid];
-"1232 _param_constant185" -> "1233 layer_norm_24"  [label="(384,)", style=solid];
-"1233 layer_norm_24" -> "1234 add_37"  [label="(1, 14, 14, 384)", style=solid];
-"1234 add_37" -> "1251 pad_13"  [label="(1, 14, 14, 384)", style=solid];
-"1234 add_37" -> "1364 add_40"  [label="(1, 14, 14, 384)", style=solid];
-"1235 _tensor_constant67" -> "1238 linear_68"  [label="(1, 15, 15, 2)", style=solid];
-"1236 _param_constant186" -> "1238 linear_68"  [label="(512, 2)", style=solid];
-"1237 _param_constant187" -> "1238 linear_68"  [label="(512,)", style=solid];
-"1238 linear_68" -> "1239 relu__11"  [label="(1, 15, 15, 512)", style=solid];
-"1239 relu__11" -> "1241 linear_69"  [label="(1, 15, 15, 512)", style=solid];
-"1240 _param_constant188" -> "1241 linear_69"  [label="(12, 512)", style=solid];
-"1241 linear_69" -> "1242 view_59"  [label="(1, 15, 15, 12)", style=solid];
-"1242 view_59" -> "1244 index_11"  [label="(225, 12)", style=solid];
-"1243 _tensor_constant68" -> "1244 index_11"  [label="(4096,)", style=solid];
-"1244 index_11" -> "1245 view_60"  [label="(4096, 12)", style=solid];
-"1245 view_60" -> "1246 permute_50"  [label="(64, 64, 12)", style=solid];
-"1246 permute_50" -> "1247 contiguous_20"  [label="(12, 64, 64)", style=solid];
-"1247 contiguous_20" -> "1248 unsqueeze_31"  [label="(12, 64, 64)", style=solid];
-"1248 unsqueeze_31" -> "1249 sigmoid_11"  [label="(1, 12, 64, 64)", style=solid];
-"1249 sigmoid_11" -> "1250 mul_22"  [label="(1, 12, 64, 64)", style=solid];
-"1250 mul_22" -> "1281 add_38"  [label="(1, 12, 64, 64)", style=solid];
-"1251 pad_13" -> "1252 roll_10"  [label="(1, 16, 16, 384)", style=solid];
-"1252 roll_10" -> "1253 view_61"  [label="(1, 16, 16, 384)", style=solid];
-"1253 view_61" -> "1254 permute_51"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1254 permute_51" -> "1255 reshape_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1255 reshape_49" -> "1261 linear_70"  [label="(4, 64, 384)", style=solid];
-"1255 reshape_49" -> "1282 new_zeros_5"  [label="(4, 64, 384)", style=solid];
-"1256 _param_constant189" -> "1257 clone_11"  [label="(1152,)", style=solid];
-"1257 clone_11" -> "1258 slice_166"  [label="(1152,)", style=solid];
-"1257 clone_11" -> "1261 linear_70"  [label="(1152,)", style=solid];
-"1258 slice_166" -> "1259 zero__11"  [label="(384,)", style=solid];
-"1260 _param_constant190" -> "1261 linear_70"  [label="(1152, 384)", style=solid];
-"1261 linear_70" -> "1262 reshape_50"  [label="(4, 64, 1152)", style=solid];
-"1262 reshape_50" -> "1263 permute_52"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1263 permute_52" -> "1264 select_33"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1263 permute_52" -> "1265 select_34"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1263 permute_52" -> "1266 select_35"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1264 select_33" -> "1267 linalg_vector_norm_22"  [label="(4, 12, 64, 32)", style=solid];
-"1264 select_33" -> "1269 expand_as_22"  [label="(4, 12, 64, 32)", style=solid];
-"1264 select_33" -> "1270 div_22"  [label="(4, 12, 64, 32)", style=solid];
-"1265 select_34" -> "1271 linalg_vector_norm_23"  [label="(4, 12, 64, 32)", style=solid];
-"1265 select_34" -> "1273 expand_as_23"  [label="(4, 12, 64, 32)", style=solid];
-"1265 select_34" -> "1274 div_23"  [label="(4, 12, 64, 32)", style=solid];
-"1266 select_35" -> "1345 matmul_23"  [label="(4, 12, 64, 32)", style=solid];
-"1267 linalg_vector_norm_22" -> "1268 clamp_min_22"  [label="(4, 12, 64, 1)", style=solid];
-"1268 clamp_min_22" -> "1269 expand_as_22"  [label="(4, 12, 64, 1)", style=solid];
-"1269 expand_as_22" -> "1270 div_22"  [label="(4, 12, 64, 32)", style=solid];
-"1270 div_22" -> "1276 matmul_22"  [label="(4, 12, 64, 32)", style=solid];
-"1271 linalg_vector_norm_23" -> "1272 clamp_min_23"  [label="(4, 12, 64, 1)", style=solid];
-"1272 clamp_min_23" -> "1273 expand_as_23"  [label="(4, 12, 64, 1)", style=solid];
-"1273 expand_as_23" -> "1274 div_23"  [label="(4, 12, 64, 32)", style=solid];
-"1274 div_23" -> "1275 transpose_22"  [label="(4, 12, 64, 32)", style=solid];
-"1275 transpose_22" -> "1276 matmul_22"  [label="(4, 12, 32, 64)", style=solid];
-"1276 matmul_22" -> "1280 mul_23"  [label="(4, 12, 64, 64)", style=solid];
-"1277 _param_constant191" -> "1278 clamp_11"  [label="(12, 1, 1)", style=solid];
-"1278 clamp_11" -> "1279 exp_11"  [label="(12, 1, 1)", style=solid];
-"1279 exp_11" -> "1280 mul_23"  [label="(12, 1, 1)", style=solid];
-"1280 mul_23" -> "1281 add_38"  [label="(4, 12, 64, 64)", style=solid];
-"1281 add_38" -> "1338 view_63"  [label="(4, 12, 64, 64)", style=solid];
-"1282 new_zeros_5" -> "1285 slice_167"  [label="(16, 16)", style=solid];
-"1282 new_zeros_5" -> "1290 slice_169"  [label="(16, 16)", style=solid];
-"1282 new_zeros_5" -> "1295 slice_171"  [label="(16, 16)", style=solid];
-"1282 new_zeros_5" -> "1300 slice_173"  [label="(16, 16)", style=solid];
-"1282 new_zeros_5" -> "1305 slice_175"  [label="(16, 16)", style=solid];
-"1282 new_zeros_5" -> "1310 slice_177"  [label="(16, 16)", style=solid];
-"1282 new_zeros_5" -> "1315 slice_179"  [label="(16, 16)", style=solid];
-"1282 new_zeros_5" -> "1320 slice_181"  [label="(16, 16)", style=solid];
-"1282 new_zeros_5" -> "1325 slice_183"  [label="(16, 16)", style=solid];
-"1282 new_zeros_5" -> "1328 view_62"  [label="(16, 16)", style=solid];
-"1283 _tensor_constant69" -> "1284 lift_fresh_copy_45"  [label="()", style=solid];
-"1284 lift_fresh_copy_45" -> "1287 fill__45"  [label="()", style=solid];
-"1285 slice_167" -> "1286 slice_168"  [label="(8, 16)", style=solid];
-"1286 slice_168" -> "1287 fill__45"  [label="(8, 8)", style=solid];
-"1288 _tensor_constant70" -> "1289 lift_fresh_copy_46"  [label="()", style=solid];
-"1289 lift_fresh_copy_46" -> "1292 fill__46"  [label="()", style=solid];
-"1290 slice_169" -> "1291 slice_170"  [label="(8, 16)", style=solid];
-"1291 slice_170" -> "1292 fill__46"  [label="(8, 4)", style=solid];
-"1293 _tensor_constant71" -> "1294 lift_fresh_copy_47"  [label="()", style=solid];
-"1294 lift_fresh_copy_47" -> "1297 fill__47"  [label="()", style=solid];
-"1295 slice_171" -> "1296 slice_172"  [label="(8, 16)", style=solid];
-"1296 slice_172" -> "1297 fill__47"  [label="(8, 4)", style=solid];
-"1298 _tensor_constant72" -> "1299 lift_fresh_copy_48"  [label="()", style=solid];
-"1299 lift_fresh_copy_48" -> "1302 fill__48"  [label="()", style=solid];
-"1300 slice_173" -> "1301 slice_174"  [label="(4, 16)", style=solid];
-"1301 slice_174" -> "1302 fill__48"  [label="(4, 8)", style=solid];
-"1303 _tensor_constant73" -> "1304 lift_fresh_copy_49"  [label="()", style=solid];
-"1304 lift_fresh_copy_49" -> "1307 fill__49"  [label="()", style=solid];
-"1305 slice_175" -> "1306 slice_176"  [label="(4, 16)", style=solid];
-"1306 slice_176" -> "1307 fill__49"  [label="(4, 4)", style=solid];
-"1308 _tensor_constant74" -> "1309 lift_fresh_copy_50"  [label="()", style=solid];
-"1309 lift_fresh_copy_50" -> "1312 fill__50"  [label="()", style=solid];
-"1310 slice_177" -> "1311 slice_178"  [label="(4, 16)", style=solid];
-"1311 slice_178" -> "1312 fill__50"  [label="(4, 4)", style=solid];
-"1313 _tensor_constant75" -> "1314 lift_fresh_copy_51"  [label="()", style=solid];
-"1314 lift_fresh_copy_51" -> "1317 fill__51"  [label="()", style=solid];
-"1315 slice_179" -> "1316 slice_180"  [label="(4, 16)", style=solid];
-"1316 slice_180" -> "1317 fill__51"  [label="(4, 8)", style=solid];
-"1318 _tensor_constant76" -> "1319 lift_fresh_copy_52"  [label="()", style=solid];
-"1319 lift_fresh_copy_52" -> "1322 fill__52"  [label="()", style=solid];
-"1320 slice_181" -> "1321 slice_182"  [label="(4, 16)", style=solid];
-"1321 slice_182" -> "1322 fill__52"  [label="(4, 4)", style=solid];
-"1323 _tensor_constant77" -> "1324 lift_fresh_copy_53"  [label="()", style=solid];
-"1324 lift_fresh_copy_53" -> "1327 fill__53"  [label="()", style=solid];
-"1325 slice_183" -> "1326 slice_184"  [label="(4, 16)", style=solid];
-"1326 slice_184" -> "1327 fill__53"  [label="(4, 4)", style=solid];
-"1328 view_62" -> "1329 permute_53"  [label="(2, 8, 2, 8)", style=solid];
-"1329 permute_53" -> "1330 reshape_51"  [label="(2, 2, 8, 8)", style=solid];
-"1330 reshape_51" -> "1331 unsqueeze_32"  [label="(4, 64)", style=solid];
-"1330 reshape_51" -> "1332 unsqueeze_33"  [label="(4, 64)", style=solid];
-"1331 unsqueeze_32" -> "1333 sub_5"  [label="(4, 1, 64)", style=solid];
-"1332 unsqueeze_33" -> "1333 sub_5"  [label="(4, 64, 1)", style=solid];
-"1333 sub_5" -> "1334 ne_5"  [label="(4, 64, 64)", style=solid];
-"1333 sub_5" -> "1335 masked_fill_10"  [label="(4, 64, 64)", style=solid];
-"1333 sub_5" -> "1336 eq_5"  [label="(4, 64, 64)", style=solid];
-"1334 ne_5" -> "1335 masked_fill_10"  [label="(4, 64, 64)", style=solid];
-"1335 masked_fill_10" -> "1337 masked_fill_11"  [label="(4, 64, 64)", style=solid];
-"1336 eq_5" -> "1337 masked_fill_11"  [label="(4, 64, 64)", style=solid];
-"1337 masked_fill_11" -> "1339 unsqueeze_34"  [label="(4, 64, 64)", style=solid];
-"1338 view_63" -> "1341 add_39"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1339 unsqueeze_34" -> "1340 unsqueeze_35"  [label="(4, 1, 64, 64)", style=solid];
-"1340 unsqueeze_35" -> "1341 add_39"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1341 add_39" -> "1342 view_64"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1342 view_64" -> "1343 softmax_11"  [label="(4, 12, 64, 64)", style=solid];
-"1343 softmax_11" -> "1344 dropout_44"  [label="(4, 12, 64, 64)", style=solid];
-"1344 dropout_44" -> "1345 matmul_23"  [label="(4, 12, 64, 64)", style=solid];
-"1345 matmul_23" -> "1346 transpose_23"  [label="(4, 12, 64, 32)", style=solid];
-"1346 transpose_23" -> "1347 reshape_52"  [label="(4, 64, 12, 32)", style=solid];
-"1347 reshape_52" -> "1350 linear_71"  [label="(4, 64, 384)", style=solid];
-"1348 _param_constant192" -> "1350 linear_71"  [label="(384, 384)", style=solid];
-"1349 _param_constant193" -> "1350 linear_71"  [label="(384,)", style=solid];
-"1350 linear_71" -> "1351 dropout_45"  [label="(4, 64, 384)", style=solid];
-"1351 dropout_45" -> "1352 view_65"  [label="(4, 64, 384)", style=solid];
-"1352 view_65" -> "1353 permute_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1353 permute_54" -> "1354 reshape_53"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1354 reshape_53" -> "1355 roll_11"  [label="(1, 16, 16, 384)", style=solid];
-"1355 roll_11" -> "1356 slice_185"  [label="(1, 16, 16, 384)", style=solid];
-"1356 slice_185" -> "1357 slice_186"  [label="(1, 16, 16, 384)", style=solid];
-"1357 slice_186" -> "1358 slice_187"  [label="(1, 14, 16, 384)", style=solid];
-"1358 slice_187" -> "1359 slice_188"  [label="(1, 14, 14, 384)", style=solid];
-"1359 slice_188" -> "1360 contiguous_21"  [label="(1, 14, 14, 384)", style=solid];
-"1360 contiguous_21" -> "1363 layer_norm_25"  [label="(1, 14, 14, 384)", style=solid];
-"1361 _param_constant194" -> "1363 layer_norm_25"  [label="(384,)", style=solid];
-"1362 _param_constant195" -> "1363 layer_norm_25"  [label="(384,)", style=solid];
-"1363 layer_norm_25" -> "1364 add_40"  [label="(1, 14, 14, 384)", style=solid];
-"1364 add_40" -> "1367 linear_72"  [label="(1, 14, 14, 384)", style=solid];
-"1364 add_40" -> "1377 add_41"  [label="(1, 14, 14, 384)", style=solid];
-"1365 _param_constant196" -> "1367 linear_72"  [label="(1536, 384)", style=solid];
-"1366 _param_constant197" -> "1367 linear_72"  [label="(1536,)", style=solid];
-"1367 linear_72" -> "1368 gelu_11"  [label="(1, 14, 14, 1536)", style=solid];
-"1368 gelu_11" -> "1369 dropout_46"  [label="(1, 14, 14, 1536)", style=solid];
-"1369 dropout_46" -> "1372 linear_73"  [label="(1, 14, 14, 1536)", style=solid];
-"1370 _param_constant198" -> "1372 linear_73"  [label="(384, 1536)", style=solid];
-"1371 _param_constant199" -> "1372 linear_73"  [label="(384,)", style=solid];
-"1372 linear_73" -> "1373 dropout_47"  [label="(1, 14, 14, 384)", style=solid];
-"1373 dropout_47" -> "1376 layer_norm_26"  [label="(1, 14, 14, 384)", style=solid];
-"1374 _param_constant200" -> "1376 layer_norm_26"  [label="(384,)", style=solid];
-"1375 _param_constant201" -> "1376 layer_norm_26"  [label="(384,)", style=solid];
-"1376 layer_norm_26" -> "1377 add_41"  [label="(1, 14, 14, 384)", style=solid];
-"1377 add_41" -> "1394 pad_14"  [label="(1, 14, 14, 384)", style=solid];
-"1377 add_41" -> "1444 add_43"  [label="(1, 14, 14, 384)", style=solid];
-"1378 _tensor_constant78" -> "1381 linear_74"  [label="(1, 15, 15, 2)", style=solid];
-"1379 _param_constant202" -> "1381 linear_74"  [label="(512, 2)", style=solid];
-"1380 _param_constant203" -> "1381 linear_74"  [label="(512,)", style=solid];
-"1381 linear_74" -> "1382 relu__12"  [label="(1, 15, 15, 512)", style=solid];
-"1382 relu__12" -> "1384 linear_75"  [label="(1, 15, 15, 512)", style=solid];
-"1383 _param_constant204" -> "1384 linear_75"  [label="(12, 512)", style=solid];
-"1384 linear_75" -> "1385 view_66"  [label="(1, 15, 15, 12)", style=solid];
-"1385 view_66" -> "1387 index_12"  [label="(225, 12)", style=solid];
-"1386 _tensor_constant79" -> "1387 index_12"  [label="(4096,)", style=solid];
-"1387 index_12" -> "1388 view_67"  [label="(4096, 12)", style=solid];
-"1388 view_67" -> "1389 permute_55"  [label="(64, 64, 12)", style=solid];
-"1389 permute_55" -> "1390 contiguous_22"  [label="(12, 64, 64)", style=solid];
-"1390 contiguous_22" -> "1391 unsqueeze_36"  [label="(12, 64, 64)", style=solid];
-"1391 unsqueeze_36" -> "1392 sigmoid_12"  [label="(1, 12, 64, 64)", style=solid];
-"1392 sigmoid_12" -> "1393 mul_24"  [label="(1, 12, 64, 64)", style=solid];
-"1393 mul_24" -> "1423 add_42"  [label="(1, 12, 64, 64)", style=solid];
-"1394 pad_14" -> "1395 view_68"  [label="(1, 16, 16, 384)", style=solid];
-"1395 view_68" -> "1396 permute_56"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1396 permute_56" -> "1397 reshape_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1397 reshape_54" -> "1403 linear_76"  [label="(4, 64, 384)", style=solid];
-"1398 _param_constant205" -> "1399 clone_12"  [label="(1152,)", style=solid];
-"1399 clone_12" -> "1400 slice_189"  [label="(1152,)", style=solid];
-"1399 clone_12" -> "1403 linear_76"  [label="(1152,)", style=solid];
-"1400 slice_189" -> "1401 zero__12"  [label="(384,)", style=solid];
-"1402 _param_constant206" -> "1403 linear_76"  [label="(1152, 384)", style=solid];
-"1403 linear_76" -> "1404 reshape_55"  [label="(4, 64, 1152)", style=solid];
-"1404 reshape_55" -> "1405 permute_57"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1405 permute_57" -> "1406 select_36"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1405 permute_57" -> "1407 select_37"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1405 permute_57" -> "1408 select_38"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1406 select_36" -> "1409 linalg_vector_norm_24"  [label="(4, 12, 64, 32)", style=solid];
-"1406 select_36" -> "1411 expand_as_24"  [label="(4, 12, 64, 32)", style=solid];
-"1406 select_36" -> "1412 div_24"  [label="(4, 12, 64, 32)", style=solid];
-"1407 select_37" -> "1413 linalg_vector_norm_25"  [label="(4, 12, 64, 32)", style=solid];
-"1407 select_37" -> "1415 expand_as_25"  [label="(4, 12, 64, 32)", style=solid];
-"1407 select_37" -> "1416 div_25"  [label="(4, 12, 64, 32)", style=solid];
-"1408 select_38" -> "1426 matmul_25"  [label="(4, 12, 64, 32)", style=solid];
-"1409 linalg_vector_norm_24" -> "1410 clamp_min_24"  [label="(4, 12, 64, 1)", style=solid];
-"1410 clamp_min_24" -> "1411 expand_as_24"  [label="(4, 12, 64, 1)", style=solid];
-"1411 expand_as_24" -> "1412 div_24"  [label="(4, 12, 64, 32)", style=solid];
-"1412 div_24" -> "1418 matmul_24"  [label="(4, 12, 64, 32)", style=solid];
-"1413 linalg_vector_norm_25" -> "1414 clamp_min_25"  [label="(4, 12, 64, 1)", style=solid];
-"1414 clamp_min_25" -> "1415 expand_as_25"  [label="(4, 12, 64, 1)", style=solid];
-"1415 expand_as_25" -> "1416 div_25"  [label="(4, 12, 64, 32)", style=solid];
-"1416 div_25" -> "1417 transpose_24"  [label="(4, 12, 64, 32)", style=solid];
-"1417 transpose_24" -> "1418 matmul_24"  [label="(4, 12, 32, 64)", style=solid];
-"1418 matmul_24" -> "1422 mul_25"  [label="(4, 12, 64, 64)", style=solid];
-"1419 _param_constant207" -> "1420 clamp_12"  [label="(12, 1, 1)", style=solid];
-"1420 clamp_12" -> "1421 exp_12"  [label="(12, 1, 1)", style=solid];
-"1421 exp_12" -> "1422 mul_25"  [label="(12, 1, 1)", style=solid];
-"1422 mul_25" -> "1423 add_42"  [label="(4, 12, 64, 64)", style=solid];
-"1423 add_42" -> "1424 softmax_12"  [label="(4, 12, 64, 64)", style=solid];
-"1424 softmax_12" -> "1425 dropout_48"  [label="(4, 12, 64, 64)", style=solid];
-"1425 dropout_48" -> "1426 matmul_25"  [label="(4, 12, 64, 64)", style=solid];
-"1426 matmul_25" -> "1427 transpose_25"  [label="(4, 12, 64, 32)", style=solid];
-"1427 transpose_25" -> "1428 reshape_56"  [label="(4, 64, 12, 32)", style=solid];
-"1428 reshape_56" -> "1431 linear_77"  [label="(4, 64, 384)", style=solid];
-"1429 _param_constant208" -> "1431 linear_77"  [label="(384, 384)", style=solid];
-"1430 _param_constant209" -> "1431 linear_77"  [label="(384,)", style=solid];
-"1431 linear_77" -> "1432 dropout_49"  [label="(4, 64, 384)", style=solid];
-"1432 dropout_49" -> "1433 view_69"  [label="(4, 64, 384)", style=solid];
-"1433 view_69" -> "1434 permute_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1434 permute_58" -> "1435 reshape_57"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1435 reshape_57" -> "1436 slice_190"  [label="(1, 16, 16, 384)", style=solid];
-"1436 slice_190" -> "1437 slice_191"  [label="(1, 16, 16, 384)", style=solid];
-"1437 slice_191" -> "1438 slice_192"  [label="(1, 14, 16, 384)", style=solid];
-"1438 slice_192" -> "1439 slice_193"  [label="(1, 14, 14, 384)", style=solid];
-"1439 slice_193" -> "1440 contiguous_23"  [label="(1, 14, 14, 384)", style=solid];
-"1440 contiguous_23" -> "1443 layer_norm_27"  [label="(1, 14, 14, 384)", style=solid];
-"1441 _param_constant210" -> "1443 layer_norm_27"  [label="(384,)", style=solid];
-"1442 _param_constant211" -> "1443 layer_norm_27"  [label="(384,)", style=solid];
-"1443 layer_norm_27" -> "1444 add_43"  [label="(1, 14, 14, 384)", style=solid];
-"1444 add_43" -> "1447 linear_78"  [label="(1, 14, 14, 384)", style=solid];
-"1444 add_43" -> "1457 add_44"  [label="(1, 14, 14, 384)", style=solid];
-"1445 _param_constant212" -> "1447 linear_78"  [label="(1536, 384)", style=solid];
-"1446 _param_constant213" -> "1447 linear_78"  [label="(1536,)", style=solid];
-"1447 linear_78" -> "1448 gelu_12"  [label="(1, 14, 14, 1536)", style=solid];
-"1448 gelu_12" -> "1449 dropout_50"  [label="(1, 14, 14, 1536)", style=solid];
-"1449 dropout_50" -> "1452 linear_79"  [label="(1, 14, 14, 1536)", style=solid];
-"1450 _param_constant214" -> "1452 linear_79"  [label="(384, 1536)", style=solid];
-"1451 _param_constant215" -> "1452 linear_79"  [label="(384,)", style=solid];
-"1452 linear_79" -> "1453 dropout_51"  [label="(1, 14, 14, 384)", style=solid];
-"1453 dropout_51" -> "1456 layer_norm_28"  [label="(1, 14, 14, 384)", style=solid];
-"1454 _param_constant216" -> "1456 layer_norm_28"  [label="(384,)", style=solid];
-"1455 _param_constant217" -> "1456 layer_norm_28"  [label="(384,)", style=solid];
-"1456 layer_norm_28" -> "1457 add_44"  [label="(1, 14, 14, 384)", style=solid];
-"1457 add_44" -> "1474 pad_15"  [label="(1, 14, 14, 384)", style=solid];
-"1457 add_44" -> "1587 add_47"  [label="(1, 14, 14, 384)", style=solid];
-"1458 _tensor_constant80" -> "1461 linear_80"  [label="(1, 15, 15, 2)", style=solid];
-"1459 _param_constant218" -> "1461 linear_80"  [label="(512, 2)", style=solid];
-"1460 _param_constant219" -> "1461 linear_80"  [label="(512,)", style=solid];
-"1461 linear_80" -> "1462 relu__13"  [label="(1, 15, 15, 512)", style=solid];
-"1462 relu__13" -> "1464 linear_81"  [label="(1, 15, 15, 512)", style=solid];
-"1463 _param_constant220" -> "1464 linear_81"  [label="(12, 512)", style=solid];
-"1464 linear_81" -> "1465 view_70"  [label="(1, 15, 15, 12)", style=solid];
-"1465 view_70" -> "1467 index_13"  [label="(225, 12)", style=solid];
-"1466 _tensor_constant81" -> "1467 index_13"  [label="(4096,)", style=solid];
-"1467 index_13" -> "1468 view_71"  [label="(4096, 12)", style=solid];
-"1468 view_71" -> "1469 permute_59"  [label="(64, 64, 12)", style=solid];
-"1469 permute_59" -> "1470 contiguous_24"  [label="(12, 64, 64)", style=solid];
-"1470 contiguous_24" -> "1471 unsqueeze_37"  [label="(12, 64, 64)", style=solid];
-"1471 unsqueeze_37" -> "1472 sigmoid_13"  [label="(1, 12, 64, 64)", style=solid];
-"1472 sigmoid_13" -> "1473 mul_26"  [label="(1, 12, 64, 64)", style=solid];
-"1473 mul_26" -> "1504 add_45"  [label="(1, 12, 64, 64)", style=solid];
-"1474 pad_15" -> "1475 roll_12"  [label="(1, 16, 16, 384)", style=solid];
-"1475 roll_12" -> "1476 view_72"  [label="(1, 16, 16, 384)", style=solid];
-"1476 view_72" -> "1477 permute_60"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1477 permute_60" -> "1478 reshape_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1478 reshape_58" -> "1484 linear_82"  [label="(4, 64, 384)", style=solid];
-"1478 reshape_58" -> "1505 new_zeros_6"  [label="(4, 64, 384)", style=solid];
-"1479 _param_constant221" -> "1480 clone_13"  [label="(1152,)", style=solid];
-"1480 clone_13" -> "1481 slice_194"  [label="(1152,)", style=solid];
-"1480 clone_13" -> "1484 linear_82"  [label="(1152,)", style=solid];
-"1481 slice_194" -> "1482 zero__13"  [label="(384,)", style=solid];
-"1483 _param_constant222" -> "1484 linear_82"  [label="(1152, 384)", style=solid];
-"1484 linear_82" -> "1485 reshape_59"  [label="(4, 64, 1152)", style=solid];
-"1485 reshape_59" -> "1486 permute_61"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1486 permute_61" -> "1487 select_39"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1486 permute_61" -> "1488 select_40"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1486 permute_61" -> "1489 select_41"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1487 select_39" -> "1490 linalg_vector_norm_26"  [label="(4, 12, 64, 32)", style=solid];
-"1487 select_39" -> "1492 expand_as_26"  [label="(4, 12, 64, 32)", style=solid];
-"1487 select_39" -> "1493 div_26"  [label="(4, 12, 64, 32)", style=solid];
-"1488 select_40" -> "1494 linalg_vector_norm_27"  [label="(4, 12, 64, 32)", style=solid];
-"1488 select_40" -> "1496 expand_as_27"  [label="(4, 12, 64, 32)", style=solid];
-"1488 select_40" -> "1497 div_27"  [label="(4, 12, 64, 32)", style=solid];
-"1489 select_41" -> "1568 matmul_27"  [label="(4, 12, 64, 32)", style=solid];
-"1490 linalg_vector_norm_26" -> "1491 clamp_min_26"  [label="(4, 12, 64, 1)", style=solid];
-"1491 clamp_min_26" -> "1492 expand_as_26"  [label="(4, 12, 64, 1)", style=solid];
-"1492 expand_as_26" -> "1493 div_26"  [label="(4, 12, 64, 32)", style=solid];
-"1493 div_26" -> "1499 matmul_26"  [label="(4, 12, 64, 32)", style=solid];
-"1494 linalg_vector_norm_27" -> "1495 clamp_min_27"  [label="(4, 12, 64, 1)", style=solid];
-"1495 clamp_min_27" -> "1496 expand_as_27"  [label="(4, 12, 64, 1)", style=solid];
-"1496 expand_as_27" -> "1497 div_27"  [label="(4, 12, 64, 32)", style=solid];
-"1497 div_27" -> "1498 transpose_26"  [label="(4, 12, 64, 32)", style=solid];
-"1498 transpose_26" -> "1499 matmul_26"  [label="(4, 12, 32, 64)", style=solid];
-"1499 matmul_26" -> "1503 mul_27"  [label="(4, 12, 64, 64)", style=solid];
-"1500 _param_constant223" -> "1501 clamp_13"  [label="(12, 1, 1)", style=solid];
-"1501 clamp_13" -> "1502 exp_13"  [label="(12, 1, 1)", style=solid];
-"1502 exp_13" -> "1503 mul_27"  [label="(12, 1, 1)", style=solid];
-"1503 mul_27" -> "1504 add_45"  [label="(4, 12, 64, 64)", style=solid];
-"1504 add_45" -> "1561 view_74"  [label="(4, 12, 64, 64)", style=solid];
-"1505 new_zeros_6" -> "1508 slice_195"  [label="(16, 16)", style=solid];
-"1505 new_zeros_6" -> "1513 slice_197"  [label="(16, 16)", style=solid];
-"1505 new_zeros_6" -> "1518 slice_199"  [label="(16, 16)", style=solid];
-"1505 new_zeros_6" -> "1523 slice_201"  [label="(16, 16)", style=solid];
-"1505 new_zeros_6" -> "1528 slice_203"  [label="(16, 16)", style=solid];
-"1505 new_zeros_6" -> "1533 slice_205"  [label="(16, 16)", style=solid];
-"1505 new_zeros_6" -> "1538 slice_207"  [label="(16, 16)", style=solid];
-"1505 new_zeros_6" -> "1543 slice_209"  [label="(16, 16)", style=solid];
-"1505 new_zeros_6" -> "1548 slice_211"  [label="(16, 16)", style=solid];
-"1505 new_zeros_6" -> "1551 view_73"  [label="(16, 16)", style=solid];
-"1506 _tensor_constant82" -> "1507 lift_fresh_copy_54"  [label="()", style=solid];
-"1507 lift_fresh_copy_54" -> "1510 fill__54"  [label="()", style=solid];
-"1508 slice_195" -> "1509 slice_196"  [label="(8, 16)", style=solid];
-"1509 slice_196" -> "1510 fill__54"  [label="(8, 8)", style=solid];
-"1511 _tensor_constant83" -> "1512 lift_fresh_copy_55"  [label="()", style=solid];
-"1512 lift_fresh_copy_55" -> "1515 fill__55"  [label="()", style=solid];
-"1513 slice_197" -> "1514 slice_198"  [label="(8, 16)", style=solid];
-"1514 slice_198" -> "1515 fill__55"  [label="(8, 4)", style=solid];
-"1516 _tensor_constant84" -> "1517 lift_fresh_copy_56"  [label="()", style=solid];
-"1517 lift_fresh_copy_56" -> "1520 fill__56"  [label="()", style=solid];
-"1518 slice_199" -> "1519 slice_200"  [label="(8, 16)", style=solid];
-"1519 slice_200" -> "1520 fill__56"  [label="(8, 4)", style=solid];
-"1521 _tensor_constant85" -> "1522 lift_fresh_copy_57"  [label="()", style=solid];
-"1522 lift_fresh_copy_57" -> "1525 fill__57"  [label="()", style=solid];
-"1523 slice_201" -> "1524 slice_202"  [label="(4, 16)", style=solid];
-"1524 slice_202" -> "1525 fill__57"  [label="(4, 8)", style=solid];
-"1526 _tensor_constant86" -> "1527 lift_fresh_copy_58"  [label="()", style=solid];
-"1527 lift_fresh_copy_58" -> "1530 fill__58"  [label="()", style=solid];
-"1528 slice_203" -> "1529 slice_204"  [label="(4, 16)", style=solid];
-"1529 slice_204" -> "1530 fill__58"  [label="(4, 4)", style=solid];
-"1531 _tensor_constant87" -> "1532 lift_fresh_copy_59"  [label="()", style=solid];
-"1532 lift_fresh_copy_59" -> "1535 fill__59"  [label="()", style=solid];
-"1533 slice_205" -> "1534 slice_206"  [label="(4, 16)", style=solid];
-"1534 slice_206" -> "1535 fill__59"  [label="(4, 4)", style=solid];
-"1536 _tensor_constant88" -> "1537 lift_fresh_copy_60"  [label="()", style=solid];
-"1537 lift_fresh_copy_60" -> "1540 fill__60"  [label="()", style=solid];
-"1538 slice_207" -> "1539 slice_208"  [label="(4, 16)", style=solid];
-"1539 slice_208" -> "1540 fill__60"  [label="(4, 8)", style=solid];
-"1541 _tensor_constant89" -> "1542 lift_fresh_copy_61"  [label="()", style=solid];
-"1542 lift_fresh_copy_61" -> "1545 fill__61"  [label="()", style=solid];
-"1543 slice_209" -> "1544 slice_210"  [label="(4, 16)", style=solid];
-"1544 slice_210" -> "1545 fill__61"  [label="(4, 4)", style=solid];
-"1546 _tensor_constant90" -> "1547 lift_fresh_copy_62"  [label="()", style=solid];
-"1547 lift_fresh_copy_62" -> "1550 fill__62"  [label="()", style=solid];
-"1548 slice_211" -> "1549 slice_212"  [label="(4, 16)", style=solid];
-"1549 slice_212" -> "1550 fill__62"  [label="(4, 4)", style=solid];
-"1551 view_73" -> "1552 permute_62"  [label="(2, 8, 2, 8)", style=solid];
-"1552 permute_62" -> "1553 reshape_60"  [label="(2, 2, 8, 8)", style=solid];
-"1553 reshape_60" -> "1554 unsqueeze_38"  [label="(4, 64)", style=solid];
-"1553 reshape_60" -> "1555 unsqueeze_39"  [label="(4, 64)", style=solid];
-"1554 unsqueeze_38" -> "1556 sub_6"  [label="(4, 1, 64)", style=solid];
-"1555 unsqueeze_39" -> "1556 sub_6"  [label="(4, 64, 1)", style=solid];
-"1556 sub_6" -> "1557 ne_6"  [label="(4, 64, 64)", style=solid];
-"1556 sub_6" -> "1558 masked_fill_12"  [label="(4, 64, 64)", style=solid];
-"1556 sub_6" -> "1559 eq_6"  [label="(4, 64, 64)", style=solid];
-"1557 ne_6" -> "1558 masked_fill_12"  [label="(4, 64, 64)", style=solid];
-"1558 masked_fill_12" -> "1560 masked_fill_13"  [label="(4, 64, 64)", style=solid];
-"1559 eq_6" -> "1560 masked_fill_13"  [label="(4, 64, 64)", style=solid];
-"1560 masked_fill_13" -> "1562 unsqueeze_40"  [label="(4, 64, 64)", style=solid];
-"1561 view_74" -> "1564 add_46"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1562 unsqueeze_40" -> "1563 unsqueeze_41"  [label="(4, 1, 64, 64)", style=solid];
-"1563 unsqueeze_41" -> "1564 add_46"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1564 add_46" -> "1565 view_75"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1565 view_75" -> "1566 softmax_13"  [label="(4, 12, 64, 64)", style=solid];
-"1566 softmax_13" -> "1567 dropout_52"  [label="(4, 12, 64, 64)", style=solid];
-"1567 dropout_52" -> "1568 matmul_27"  [label="(4, 12, 64, 64)", style=solid];
-"1568 matmul_27" -> "1569 transpose_27"  [label="(4, 12, 64, 32)", style=solid];
-"1569 transpose_27" -> "1570 reshape_61"  [label="(4, 64, 12, 32)", style=solid];
-"1570 reshape_61" -> "1573 linear_83"  [label="(4, 64, 384)", style=solid];
-"1571 _param_constant224" -> "1573 linear_83"  [label="(384, 384)", style=solid];
-"1572 _param_constant225" -> "1573 linear_83"  [label="(384,)", style=solid];
-"1573 linear_83" -> "1574 dropout_53"  [label="(4, 64, 384)", style=solid];
-"1574 dropout_53" -> "1575 view_76"  [label="(4, 64, 384)", style=solid];
-"1575 view_76" -> "1576 permute_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1576 permute_63" -> "1577 reshape_62"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1577 reshape_62" -> "1578 roll_13"  [label="(1, 16, 16, 384)", style=solid];
-"1578 roll_13" -> "1579 slice_213"  [label="(1, 16, 16, 384)", style=solid];
-"1579 slice_213" -> "1580 slice_214"  [label="(1, 16, 16, 384)", style=solid];
-"1580 slice_214" -> "1581 slice_215"  [label="(1, 14, 16, 384)", style=solid];
-"1581 slice_215" -> "1582 slice_216"  [label="(1, 14, 14, 384)", style=solid];
-"1582 slice_216" -> "1583 contiguous_25"  [label="(1, 14, 14, 384)", style=solid];
-"1583 contiguous_25" -> "1586 layer_norm_29"  [label="(1, 14, 14, 384)", style=solid];
-"1584 _param_constant226" -> "1586 layer_norm_29"  [label="(384,)", style=solid];
-"1585 _param_constant227" -> "1586 layer_norm_29"  [label="(384,)", style=solid];
-"1586 layer_norm_29" -> "1587 add_47"  [label="(1, 14, 14, 384)", style=solid];
-"1587 add_47" -> "1590 linear_84"  [label="(1, 14, 14, 384)", style=solid];
-"1587 add_47" -> "1600 add_48"  [label="(1, 14, 14, 384)", style=solid];
-"1588 _param_constant228" -> "1590 linear_84"  [label="(1536, 384)", style=solid];
-"1589 _param_constant229" -> "1590 linear_84"  [label="(1536,)", style=solid];
-"1590 linear_84" -> "1591 gelu_13"  [label="(1, 14, 14, 1536)", style=solid];
-"1591 gelu_13" -> "1592 dropout_54"  [label="(1, 14, 14, 1536)", style=solid];
-"1592 dropout_54" -> "1595 linear_85"  [label="(1, 14, 14, 1536)", style=solid];
-"1593 _param_constant230" -> "1595 linear_85"  [label="(384, 1536)", style=solid];
-"1594 _param_constant231" -> "1595 linear_85"  [label="(384,)", style=solid];
-"1595 linear_85" -> "1596 dropout_55"  [label="(1, 14, 14, 384)", style=solid];
-"1596 dropout_55" -> "1599 layer_norm_30"  [label="(1, 14, 14, 384)", style=solid];
-"1597 _param_constant232" -> "1599 layer_norm_30"  [label="(384,)", style=solid];
-"1598 _param_constant233" -> "1599 layer_norm_30"  [label="(384,)", style=solid];
-"1599 layer_norm_30" -> "1600 add_48"  [label="(1, 14, 14, 384)", style=solid];
-"1600 add_48" -> "1617 pad_16"  [label="(1, 14, 14, 384)", style=solid];
-"1600 add_48" -> "1667 add_50"  [label="(1, 14, 14, 384)", style=solid];
-"1601 _tensor_constant91" -> "1604 linear_86"  [label="(1, 15, 15, 2)", style=solid];
-"1602 _param_constant234" -> "1604 linear_86"  [label="(512, 2)", style=solid];
-"1603 _param_constant235" -> "1604 linear_86"  [label="(512,)", style=solid];
-"1604 linear_86" -> "1605 relu__14"  [label="(1, 15, 15, 512)", style=solid];
-"1605 relu__14" -> "1607 linear_87"  [label="(1, 15, 15, 512)", style=solid];
-"1606 _param_constant236" -> "1607 linear_87"  [label="(12, 512)", style=solid];
-"1607 linear_87" -> "1608 view_77"  [label="(1, 15, 15, 12)", style=solid];
-"1608 view_77" -> "1610 index_14"  [label="(225, 12)", style=solid];
-"1609 _tensor_constant92" -> "1610 index_14"  [label="(4096,)", style=solid];
-"1610 index_14" -> "1611 view_78"  [label="(4096, 12)", style=solid];
-"1611 view_78" -> "1612 permute_64"  [label="(64, 64, 12)", style=solid];
-"1612 permute_64" -> "1613 contiguous_26"  [label="(12, 64, 64)", style=solid];
-"1613 contiguous_26" -> "1614 unsqueeze_42"  [label="(12, 64, 64)", style=solid];
-"1614 unsqueeze_42" -> "1615 sigmoid_14"  [label="(1, 12, 64, 64)", style=solid];
-"1615 sigmoid_14" -> "1616 mul_28"  [label="(1, 12, 64, 64)", style=solid];
-"1616 mul_28" -> "1646 add_49"  [label="(1, 12, 64, 64)", style=solid];
-"1617 pad_16" -> "1618 view_79"  [label="(1, 16, 16, 384)", style=solid];
-"1618 view_79" -> "1619 permute_65"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1619 permute_65" -> "1620 reshape_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1620 reshape_63" -> "1626 linear_88"  [label="(4, 64, 384)", style=solid];
-"1621 _param_constant237" -> "1622 clone_14"  [label="(1152,)", style=solid];
-"1622 clone_14" -> "1623 slice_217"  [label="(1152,)", style=solid];
-"1622 clone_14" -> "1626 linear_88"  [label="(1152,)", style=solid];
-"1623 slice_217" -> "1624 zero__14"  [label="(384,)", style=solid];
-"1625 _param_constant238" -> "1626 linear_88"  [label="(1152, 384)", style=solid];
-"1626 linear_88" -> "1627 reshape_64"  [label="(4, 64, 1152)", style=solid];
-"1627 reshape_64" -> "1628 permute_66"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1628 permute_66" -> "1629 select_42"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1628 permute_66" -> "1630 select_43"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1628 permute_66" -> "1631 select_44"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1629 select_42" -> "1632 linalg_vector_norm_28"  [label="(4, 12, 64, 32)", style=solid];
-"1629 select_42" -> "1634 expand_as_28"  [label="(4, 12, 64, 32)", style=solid];
-"1629 select_42" -> "1635 div_28"  [label="(4, 12, 64, 32)", style=solid];
-"1630 select_43" -> "1636 linalg_vector_norm_29"  [label="(4, 12, 64, 32)", style=solid];
-"1630 select_43" -> "1638 expand_as_29"  [label="(4, 12, 64, 32)", style=solid];
-"1630 select_43" -> "1639 div_29"  [label="(4, 12, 64, 32)", style=solid];
-"1631 select_44" -> "1649 matmul_29"  [label="(4, 12, 64, 32)", style=solid];
-"1632 linalg_vector_norm_28" -> "1633 clamp_min_28"  [label="(4, 12, 64, 1)", style=solid];
-"1633 clamp_min_28" -> "1634 expand_as_28"  [label="(4, 12, 64, 1)", style=solid];
-"1634 expand_as_28" -> "1635 div_28"  [label="(4, 12, 64, 32)", style=solid];
-"1635 div_28" -> "1641 matmul_28"  [label="(4, 12, 64, 32)", style=solid];
-"1636 linalg_vector_norm_29" -> "1637 clamp_min_29"  [label="(4, 12, 64, 1)", style=solid];
-"1637 clamp_min_29" -> "1638 expand_as_29"  [label="(4, 12, 64, 1)", style=solid];
-"1638 expand_as_29" -> "1639 div_29"  [label="(4, 12, 64, 32)", style=solid];
-"1639 div_29" -> "1640 transpose_28"  [label="(4, 12, 64, 32)", style=solid];
-"1640 transpose_28" -> "1641 matmul_28"  [label="(4, 12, 32, 64)", style=solid];
-"1641 matmul_28" -> "1645 mul_29"  [label="(4, 12, 64, 64)", style=solid];
-"1642 _param_constant239" -> "1643 clamp_14"  [label="(12, 1, 1)", style=solid];
-"1643 clamp_14" -> "1644 exp_14"  [label="(12, 1, 1)", style=solid];
-"1644 exp_14" -> "1645 mul_29"  [label="(12, 1, 1)", style=solid];
-"1645 mul_29" -> "1646 add_49"  [label="(4, 12, 64, 64)", style=solid];
-"1646 add_49" -> "1647 softmax_14"  [label="(4, 12, 64, 64)", style=solid];
-"1647 softmax_14" -> "1648 dropout_56"  [label="(4, 12, 64, 64)", style=solid];
-"1648 dropout_56" -> "1649 matmul_29"  [label="(4, 12, 64, 64)", style=solid];
-"1649 matmul_29" -> "1650 transpose_29"  [label="(4, 12, 64, 32)", style=solid];
-"1650 transpose_29" -> "1651 reshape_65"  [label="(4, 64, 12, 32)", style=solid];
-"1651 reshape_65" -> "1654 linear_89"  [label="(4, 64, 384)", style=solid];
-"1652 _param_constant240" -> "1654 linear_89"  [label="(384, 384)", style=solid];
-"1653 _param_constant241" -> "1654 linear_89"  [label="(384,)", style=solid];
-"1654 linear_89" -> "1655 dropout_57"  [label="(4, 64, 384)", style=solid];
-"1655 dropout_57" -> "1656 view_80"  [label="(4, 64, 384)", style=solid];
-"1656 view_80" -> "1657 permute_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1657 permute_67" -> "1658 reshape_66"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1658 reshape_66" -> "1659 slice_218"  [label="(1, 16, 16, 384)", style=solid];
-"1659 slice_218" -> "1660 slice_219"  [label="(1, 16, 16, 384)", style=solid];
-"1660 slice_219" -> "1661 slice_220"  [label="(1, 14, 16, 384)", style=solid];
-"1661 slice_220" -> "1662 slice_221"  [label="(1, 14, 14, 384)", style=solid];
-"1662 slice_221" -> "1663 contiguous_27"  [label="(1, 14, 14, 384)", style=solid];
-"1663 contiguous_27" -> "1666 layer_norm_31"  [label="(1, 14, 14, 384)", style=solid];
-"1664 _param_constant242" -> "1666 layer_norm_31"  [label="(384,)", style=solid];
-"1665 _param_constant243" -> "1666 layer_norm_31"  [label="(384,)", style=solid];
-"1666 layer_norm_31" -> "1667 add_50"  [label="(1, 14, 14, 384)", style=solid];
-"1667 add_50" -> "1670 linear_90"  [label="(1, 14, 14, 384)", style=solid];
-"1667 add_50" -> "1680 add_51"  [label="(1, 14, 14, 384)", style=solid];
-"1668 _param_constant244" -> "1670 linear_90"  [label="(1536, 384)", style=solid];
-"1669 _param_constant245" -> "1670 linear_90"  [label="(1536,)", style=solid];
-"1670 linear_90" -> "1671 gelu_14"  [label="(1, 14, 14, 1536)", style=solid];
-"1671 gelu_14" -> "1672 dropout_58"  [label="(1, 14, 14, 1536)", style=solid];
-"1672 dropout_58" -> "1675 linear_91"  [label="(1, 14, 14, 1536)", style=solid];
-"1673 _param_constant246" -> "1675 linear_91"  [label="(384, 1536)", style=solid];
-"1674 _param_constant247" -> "1675 linear_91"  [label="(384,)", style=solid];
-"1675 linear_91" -> "1676 dropout_59"  [label="(1, 14, 14, 384)", style=solid];
-"1676 dropout_59" -> "1679 layer_norm_32"  [label="(1, 14, 14, 384)", style=solid];
-"1677 _param_constant248" -> "1679 layer_norm_32"  [label="(384,)", style=solid];
-"1678 _param_constant249" -> "1679 layer_norm_32"  [label="(384,)", style=solid];
-"1679 layer_norm_32" -> "1680 add_51"  [label="(1, 14, 14, 384)", style=solid];
-"1680 add_51" -> "1697 pad_17"  [label="(1, 14, 14, 384)", style=solid];
-"1680 add_51" -> "1810 add_54"  [label="(1, 14, 14, 384)", style=solid];
-"1681 _tensor_constant93" -> "1684 linear_92"  [label="(1, 15, 15, 2)", style=solid];
-"1682 _param_constant250" -> "1684 linear_92"  [label="(512, 2)", style=solid];
-"1683 _param_constant251" -> "1684 linear_92"  [label="(512,)", style=solid];
-"1684 linear_92" -> "1685 relu__15"  [label="(1, 15, 15, 512)", style=solid];
-"1685 relu__15" -> "1687 linear_93"  [label="(1, 15, 15, 512)", style=solid];
-"1686 _param_constant252" -> "1687 linear_93"  [label="(12, 512)", style=solid];
-"1687 linear_93" -> "1688 view_81"  [label="(1, 15, 15, 12)", style=solid];
-"1688 view_81" -> "1690 index_15"  [label="(225, 12)", style=solid];
-"1689 _tensor_constant94" -> "1690 index_15"  [label="(4096,)", style=solid];
-"1690 index_15" -> "1691 view_82"  [label="(4096, 12)", style=solid];
-"1691 view_82" -> "1692 permute_68"  [label="(64, 64, 12)", style=solid];
-"1692 permute_68" -> "1693 contiguous_28"  [label="(12, 64, 64)", style=solid];
-"1693 contiguous_28" -> "1694 unsqueeze_43"  [label="(12, 64, 64)", style=solid];
-"1694 unsqueeze_43" -> "1695 sigmoid_15"  [label="(1, 12, 64, 64)", style=solid];
-"1695 sigmoid_15" -> "1696 mul_30"  [label="(1, 12, 64, 64)", style=solid];
-"1696 mul_30" -> "1727 add_52"  [label="(1, 12, 64, 64)", style=solid];
-"1697 pad_17" -> "1698 roll_14"  [label="(1, 16, 16, 384)", style=solid];
-"1698 roll_14" -> "1699 view_83"  [label="(1, 16, 16, 384)", style=solid];
-"1699 view_83" -> "1700 permute_69"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1700 permute_69" -> "1701 reshape_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1701 reshape_67" -> "1707 linear_94"  [label="(4, 64, 384)", style=solid];
-"1701 reshape_67" -> "1728 new_zeros_7"  [label="(4, 64, 384)", style=solid];
-"1702 _param_constant253" -> "1703 clone_15"  [label="(1152,)", style=solid];
-"1703 clone_15" -> "1704 slice_222"  [label="(1152,)", style=solid];
-"1703 clone_15" -> "1707 linear_94"  [label="(1152,)", style=solid];
-"1704 slice_222" -> "1705 zero__15"  [label="(384,)", style=solid];
-"1706 _param_constant254" -> "1707 linear_94"  [label="(1152, 384)", style=solid];
-"1707 linear_94" -> "1708 reshape_68"  [label="(4, 64, 1152)", style=solid];
-"1708 reshape_68" -> "1709 permute_70"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1709 permute_70" -> "1710 select_45"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1709 permute_70" -> "1711 select_46"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1709 permute_70" -> "1712 select_47"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1710 select_45" -> "1713 linalg_vector_norm_30"  [label="(4, 12, 64, 32)", style=solid];
-"1710 select_45" -> "1715 expand_as_30"  [label="(4, 12, 64, 32)", style=solid];
-"1710 select_45" -> "1716 div_30"  [label="(4, 12, 64, 32)", style=solid];
-"1711 select_46" -> "1717 linalg_vector_norm_31"  [label="(4, 12, 64, 32)", style=solid];
-"1711 select_46" -> "1719 expand_as_31"  [label="(4, 12, 64, 32)", style=solid];
-"1711 select_46" -> "1720 div_31"  [label="(4, 12, 64, 32)", style=solid];
-"1712 select_47" -> "1791 matmul_31"  [label="(4, 12, 64, 32)", style=solid];
-"1713 linalg_vector_norm_30" -> "1714 clamp_min_30"  [label="(4, 12, 64, 1)", style=solid];
-"1714 clamp_min_30" -> "1715 expand_as_30"  [label="(4, 12, 64, 1)", style=solid];
-"1715 expand_as_30" -> "1716 div_30"  [label="(4, 12, 64, 32)", style=solid];
-"1716 div_30" -> "1722 matmul_30"  [label="(4, 12, 64, 32)", style=solid];
-"1717 linalg_vector_norm_31" -> "1718 clamp_min_31"  [label="(4, 12, 64, 1)", style=solid];
-"1718 clamp_min_31" -> "1719 expand_as_31"  [label="(4, 12, 64, 1)", style=solid];
-"1719 expand_as_31" -> "1720 div_31"  [label="(4, 12, 64, 32)", style=solid];
-"1720 div_31" -> "1721 transpose_30"  [label="(4, 12, 64, 32)", style=solid];
-"1721 transpose_30" -> "1722 matmul_30"  [label="(4, 12, 32, 64)", style=solid];
-"1722 matmul_30" -> "1726 mul_31"  [label="(4, 12, 64, 64)", style=solid];
-"1723 _param_constant255" -> "1724 clamp_15"  [label="(12, 1, 1)", style=solid];
-"1724 clamp_15" -> "1725 exp_15"  [label="(12, 1, 1)", style=solid];
-"1725 exp_15" -> "1726 mul_31"  [label="(12, 1, 1)", style=solid];
-"1726 mul_31" -> "1727 add_52"  [label="(4, 12, 64, 64)", style=solid];
-"1727 add_52" -> "1784 view_85"  [label="(4, 12, 64, 64)", style=solid];
-"1728 new_zeros_7" -> "1731 slice_223"  [label="(16, 16)", style=solid];
-"1728 new_zeros_7" -> "1736 slice_225"  [label="(16, 16)", style=solid];
-"1728 new_zeros_7" -> "1741 slice_227"  [label="(16, 16)", style=solid];
-"1728 new_zeros_7" -> "1746 slice_229"  [label="(16, 16)", style=solid];
-"1728 new_zeros_7" -> "1751 slice_231"  [label="(16, 16)", style=solid];
-"1728 new_zeros_7" -> "1756 slice_233"  [label="(16, 16)", style=solid];
-"1728 new_zeros_7" -> "1761 slice_235"  [label="(16, 16)", style=solid];
-"1728 new_zeros_7" -> "1766 slice_237"  [label="(16, 16)", style=solid];
-"1728 new_zeros_7" -> "1771 slice_239"  [label="(16, 16)", style=solid];
-"1728 new_zeros_7" -> "1774 view_84"  [label="(16, 16)", style=solid];
-"1729 _tensor_constant95" -> "1730 lift_fresh_copy_63"  [label="()", style=solid];
-"1730 lift_fresh_copy_63" -> "1733 fill__63"  [label="()", style=solid];
-"1731 slice_223" -> "1732 slice_224"  [label="(8, 16)", style=solid];
-"1732 slice_224" -> "1733 fill__63"  [label="(8, 8)", style=solid];
-"1734 _tensor_constant96" -> "1735 lift_fresh_copy_64"  [label="()", style=solid];
-"1735 lift_fresh_copy_64" -> "1738 fill__64"  [label="()", style=solid];
-"1736 slice_225" -> "1737 slice_226"  [label="(8, 16)", style=solid];
-"1737 slice_226" -> "1738 fill__64"  [label="(8, 4)", style=solid];
-"1739 _tensor_constant97" -> "1740 lift_fresh_copy_65"  [label="()", style=solid];
-"1740 lift_fresh_copy_65" -> "1743 fill__65"  [label="()", style=solid];
-"1741 slice_227" -> "1742 slice_228"  [label="(8, 16)", style=solid];
-"1742 slice_228" -> "1743 fill__65"  [label="(8, 4)", style=solid];
-"1744 _tensor_constant98" -> "1745 lift_fresh_copy_66"  [label="()", style=solid];
-"1745 lift_fresh_copy_66" -> "1748 fill__66"  [label="()", style=solid];
-"1746 slice_229" -> "1747 slice_230"  [label="(4, 16)", style=solid];
-"1747 slice_230" -> "1748 fill__66"  [label="(4, 8)", style=solid];
-"1749 _tensor_constant99" -> "1750 lift_fresh_copy_67"  [label="()", style=solid];
-"1750 lift_fresh_copy_67" -> "1753 fill__67"  [label="()", style=solid];
-"1751 slice_231" -> "1752 slice_232"  [label="(4, 16)", style=solid];
-"1752 slice_232" -> "1753 fill__67"  [label="(4, 4)", style=solid];
-"1754 _tensor_constant100" -> "1755 lift_fresh_copy_68"  [label="()", style=solid];
-"1755 lift_fresh_copy_68" -> "1758 fill__68"  [label="()", style=solid];
-"1756 slice_233" -> "1757 slice_234"  [label="(4, 16)", style=solid];
-"1757 slice_234" -> "1758 fill__68"  [label="(4, 4)", style=solid];
-"1759 _tensor_constant101" -> "1760 lift_fresh_copy_69"  [label="()", style=solid];
-"1760 lift_fresh_copy_69" -> "1763 fill__69"  [label="()", style=solid];
-"1761 slice_235" -> "1762 slice_236"  [label="(4, 16)", style=solid];
-"1762 slice_236" -> "1763 fill__69"  [label="(4, 8)", style=solid];
-"1764 _tensor_constant102" -> "1765 lift_fresh_copy_70"  [label="()", style=solid];
-"1765 lift_fresh_copy_70" -> "1768 fill__70"  [label="()", style=solid];
-"1766 slice_237" -> "1767 slice_238"  [label="(4, 16)", style=solid];
-"1767 slice_238" -> "1768 fill__70"  [label="(4, 4)", style=solid];
-"1769 _tensor_constant103" -> "1770 lift_fresh_copy_71"  [label="()", style=solid];
-"1770 lift_fresh_copy_71" -> "1773 fill__71"  [label="()", style=solid];
-"1771 slice_239" -> "1772 slice_240"  [label="(4, 16)", style=solid];
-"1772 slice_240" -> "1773 fill__71"  [label="(4, 4)", style=solid];
-"1774 view_84" -> "1775 permute_71"  [label="(2, 8, 2, 8)", style=solid];
-"1775 permute_71" -> "1776 reshape_69"  [label="(2, 2, 8, 8)", style=solid];
-"1776 reshape_69" -> "1777 unsqueeze_44"  [label="(4, 64)", style=solid];
-"1776 reshape_69" -> "1778 unsqueeze_45"  [label="(4, 64)", style=solid];
-"1777 unsqueeze_44" -> "1779 sub_7"  [label="(4, 1, 64)", style=solid];
-"1778 unsqueeze_45" -> "1779 sub_7"  [label="(4, 64, 1)", style=solid];
-"1779 sub_7" -> "1780 ne_7"  [label="(4, 64, 64)", style=solid];
-"1779 sub_7" -> "1781 masked_fill_14"  [label="(4, 64, 64)", style=solid];
-"1779 sub_7" -> "1782 eq_7"  [label="(4, 64, 64)", style=solid];
-"1780 ne_7" -> "1781 masked_fill_14"  [label="(4, 64, 64)", style=solid];
-"1781 masked_fill_14" -> "1783 masked_fill_15"  [label="(4, 64, 64)", style=solid];
-"1782 eq_7" -> "1783 masked_fill_15"  [label="(4, 64, 64)", style=solid];
-"1783 masked_fill_15" -> "1785 unsqueeze_46"  [label="(4, 64, 64)", style=solid];
-"1784 view_85" -> "1787 add_53"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1785 unsqueeze_46" -> "1786 unsqueeze_47"  [label="(4, 1, 64, 64)", style=solid];
-"1786 unsqueeze_47" -> "1787 add_53"  [label="(1, 4, 1, 64, 64)", style=solid];
-"1787 add_53" -> "1788 view_86"  [label="(1, 4, 12, 64, 64)", style=solid];
-"1788 view_86" -> "1789 softmax_15"  [label="(4, 12, 64, 64)", style=solid];
-"1789 softmax_15" -> "1790 dropout_60"  [label="(4, 12, 64, 64)", style=solid];
-"1790 dropout_60" -> "1791 matmul_31"  [label="(4, 12, 64, 64)", style=solid];
-"1791 matmul_31" -> "1792 transpose_31"  [label="(4, 12, 64, 32)", style=solid];
-"1792 transpose_31" -> "1793 reshape_70"  [label="(4, 64, 12, 32)", style=solid];
-"1793 reshape_70" -> "1796 linear_95"  [label="(4, 64, 384)", style=solid];
-"1794 _param_constant256" -> "1796 linear_95"  [label="(384, 384)", style=solid];
-"1795 _param_constant257" -> "1796 linear_95"  [label="(384,)", style=solid];
-"1796 linear_95" -> "1797 dropout_61"  [label="(4, 64, 384)", style=solid];
-"1797 dropout_61" -> "1798 view_87"  [label="(4, 64, 384)", style=solid];
-"1798 view_87" -> "1799 permute_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1799 permute_72" -> "1800 reshape_71"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1800 reshape_71" -> "1801 roll_15"  [label="(1, 16, 16, 384)", style=solid];
-"1801 roll_15" -> "1802 slice_241"  [label="(1, 16, 16, 384)", style=solid];
-"1802 slice_241" -> "1803 slice_242"  [label="(1, 16, 16, 384)", style=solid];
-"1803 slice_242" -> "1804 slice_243"  [label="(1, 14, 16, 384)", style=solid];
-"1804 slice_243" -> "1805 slice_244"  [label="(1, 14, 14, 384)", style=solid];
-"1805 slice_244" -> "1806 contiguous_29"  [label="(1, 14, 14, 384)", style=solid];
-"1806 contiguous_29" -> "1809 layer_norm_33"  [label="(1, 14, 14, 384)", style=solid];
-"1807 _param_constant258" -> "1809 layer_norm_33"  [label="(384,)", style=solid];
-"1808 _param_constant259" -> "1809 layer_norm_33"  [label="(384,)", style=solid];
-"1809 layer_norm_33" -> "1810 add_54"  [label="(1, 14, 14, 384)", style=solid];
-"1810 add_54" -> "1813 linear_96"  [label="(1, 14, 14, 384)", style=solid];
-"1810 add_54" -> "1823 add_55"  [label="(1, 14, 14, 384)", style=solid];
-"1811 _param_constant260" -> "1813 linear_96"  [label="(1536, 384)", style=solid];
-"1812 _param_constant261" -> "1813 linear_96"  [label="(1536,)", style=solid];
-"1813 linear_96" -> "1814 gelu_15"  [label="(1, 14, 14, 1536)", style=solid];
-"1814 gelu_15" -> "1815 dropout_62"  [label="(1, 14, 14, 1536)", style=solid];
-"1815 dropout_62" -> "1818 linear_97"  [label="(1, 14, 14, 1536)", style=solid];
-"1816 _param_constant262" -> "1818 linear_97"  [label="(384, 1536)", style=solid];
-"1817 _param_constant263" -> "1818 linear_97"  [label="(384,)", style=solid];
-"1818 linear_97" -> "1819 dropout_63"  [label="(1, 14, 14, 384)", style=solid];
-"1819 dropout_63" -> "1822 layer_norm_34"  [label="(1, 14, 14, 384)", style=solid];
-"1820 _param_constant264" -> "1822 layer_norm_34"  [label="(384,)", style=solid];
-"1821 _param_constant265" -> "1822 layer_norm_34"  [label="(384,)", style=solid];
-"1822 layer_norm_34" -> "1823 add_55"  [label="(1, 14, 14, 384)", style=solid];
-"1823 add_55" -> "1840 pad_18"  [label="(1, 14, 14, 384)", style=solid];
-"1823 add_55" -> "1890 add_57"  [label="(1, 14, 14, 384)", style=solid];
-"1824 _tensor_constant104" -> "1827 linear_98"  [label="(1, 15, 15, 2)", style=solid];
-"1825 _param_constant266" -> "1827 linear_98"  [label="(512, 2)", style=solid];
-"1826 _param_constant267" -> "1827 linear_98"  [label="(512,)", style=solid];
-"1827 linear_98" -> "1828 relu__16"  [label="(1, 15, 15, 512)", style=solid];
-"1828 relu__16" -> "1830 linear_99"  [label="(1, 15, 15, 512)", style=solid];
-"1829 _param_constant268" -> "1830 linear_99"  [label="(12, 512)", style=solid];
-"1830 linear_99" -> "1831 view_88"  [label="(1, 15, 15, 12)", style=solid];
-"1831 view_88" -> "1833 index_16"  [label="(225, 12)", style=solid];
-"1832 _tensor_constant105" -> "1833 index_16"  [label="(4096,)", style=solid];
-"1833 index_16" -> "1834 view_89"  [label="(4096, 12)", style=solid];
-"1834 view_89" -> "1835 permute_73"  [label="(64, 64, 12)", style=solid];
-"1835 permute_73" -> "1836 contiguous_30"  [label="(12, 64, 64)", style=solid];
-"1836 contiguous_30" -> "1837 unsqueeze_48"  [label="(12, 64, 64)", style=solid];
-"1837 unsqueeze_48" -> "1838 sigmoid_16"  [label="(1, 12, 64, 64)", style=solid];
-"1838 sigmoid_16" -> "1839 mul_32"  [label="(1, 12, 64, 64)", style=solid];
-"1839 mul_32" -> "1869 add_56"  [label="(1, 12, 64, 64)", style=solid];
-"1840 pad_18" -> "1841 view_90"  [label="(1, 16, 16, 384)", style=solid];
-"1841 view_90" -> "1842 permute_74"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1842 permute_74" -> "1843 reshape_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1843 reshape_72" -> "1849 linear_100"  [label="(4, 64, 384)", style=solid];
-"1844 _param_constant269" -> "1845 clone_16"  [label="(1152,)", style=solid];
-"1845 clone_16" -> "1846 slice_245"  [label="(1152,)", style=solid];
-"1845 clone_16" -> "1849 linear_100"  [label="(1152,)", style=solid];
-"1846 slice_245" -> "1847 zero__16"  [label="(384,)", style=solid];
-"1848 _param_constant270" -> "1849 linear_100"  [label="(1152, 384)", style=solid];
-"1849 linear_100" -> "1850 reshape_73"  [label="(4, 64, 1152)", style=solid];
-"1850 reshape_73" -> "1851 permute_75"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1851 permute_75" -> "1852 select_48"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1851 permute_75" -> "1853 select_49"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1851 permute_75" -> "1854 select_50"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1852 select_48" -> "1855 linalg_vector_norm_32"  [label="(4, 12, 64, 32)", style=solid];
-"1852 select_48" -> "1857 expand_as_32"  [label="(4, 12, 64, 32)", style=solid];
-"1852 select_48" -> "1858 div_32"  [label="(4, 12, 64, 32)", style=solid];
-"1853 select_49" -> "1859 linalg_vector_norm_33"  [label="(4, 12, 64, 32)", style=solid];
-"1853 select_49" -> "1861 expand_as_33"  [label="(4, 12, 64, 32)", style=solid];
-"1853 select_49" -> "1862 div_33"  [label="(4, 12, 64, 32)", style=solid];
-"1854 select_50" -> "1872 matmul_33"  [label="(4, 12, 64, 32)", style=solid];
-"1855 linalg_vector_norm_32" -> "1856 clamp_min_32"  [label="(4, 12, 64, 1)", style=solid];
-"1856 clamp_min_32" -> "1857 expand_as_32"  [label="(4, 12, 64, 1)", style=solid];
-"1857 expand_as_32" -> "1858 div_32"  [label="(4, 12, 64, 32)", style=solid];
-"1858 div_32" -> "1864 matmul_32"  [label="(4, 12, 64, 32)", style=solid];
-"1859 linalg_vector_norm_33" -> "1860 clamp_min_33"  [label="(4, 12, 64, 1)", style=solid];
-"1860 clamp_min_33" -> "1861 expand_as_33"  [label="(4, 12, 64, 1)", style=solid];
-"1861 expand_as_33" -> "1862 div_33"  [label="(4, 12, 64, 32)", style=solid];
-"1862 div_33" -> "1863 transpose_32"  [label="(4, 12, 64, 32)", style=solid];
-"1863 transpose_32" -> "1864 matmul_32"  [label="(4, 12, 32, 64)", style=solid];
-"1864 matmul_32" -> "1868 mul_33"  [label="(4, 12, 64, 64)", style=solid];
-"1865 _param_constant271" -> "1866 clamp_16"  [label="(12, 1, 1)", style=solid];
-"1866 clamp_16" -> "1867 exp_16"  [label="(12, 1, 1)", style=solid];
-"1867 exp_16" -> "1868 mul_33"  [label="(12, 1, 1)", style=solid];
-"1868 mul_33" -> "1869 add_56"  [label="(4, 12, 64, 64)", style=solid];
-"1869 add_56" -> "1870 softmax_16"  [label="(4, 12, 64, 64)", style=solid];
-"1870 softmax_16" -> "1871 dropout_64"  [label="(4, 12, 64, 64)", style=solid];
-"1871 dropout_64" -> "1872 matmul_33"  [label="(4, 12, 64, 64)", style=solid];
-"1872 matmul_33" -> "1873 transpose_33"  [label="(4, 12, 64, 32)", style=solid];
-"1873 transpose_33" -> "1874 reshape_74"  [label="(4, 64, 12, 32)", style=solid];
-"1874 reshape_74" -> "1877 linear_101"  [label="(4, 64, 384)", style=solid];
-"1875 _param_constant272" -> "1877 linear_101"  [label="(384, 384)", style=solid];
-"1876 _param_constant273" -> "1877 linear_101"  [label="(384,)", style=solid];
-"1877 linear_101" -> "1878 dropout_65"  [label="(4, 64, 384)", style=solid];
-"1878 dropout_65" -> "1879 view_91"  [label="(4, 64, 384)", style=solid];
-"1879 view_91" -> "1880 permute_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1880 permute_76" -> "1881 reshape_75"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1881 reshape_75" -> "1882 slice_246"  [label="(1, 16, 16, 384)", style=solid];
-"1882 slice_246" -> "1883 slice_247"  [label="(1, 16, 16, 384)", style=solid];
-"1883 slice_247" -> "1884 slice_248"  [label="(1, 14, 16, 384)", style=solid];
-"1884 slice_248" -> "1885 slice_249"  [label="(1, 14, 14, 384)", style=solid];
-"1885 slice_249" -> "1886 contiguous_31"  [label="(1, 14, 14, 384)", style=solid];
-"1886 contiguous_31" -> "1889 layer_norm_35"  [label="(1, 14, 14, 384)", style=solid];
-"1887 _param_constant274" -> "1889 layer_norm_35"  [label="(384,)", style=solid];
-"1888 _param_constant275" -> "1889 layer_norm_35"  [label="(384,)", style=solid];
-"1889 layer_norm_35" -> "1890 add_57"  [label="(1, 14, 14, 384)", style=solid];
-"1890 add_57" -> "1893 linear_102"  [label="(1, 14, 14, 384)", style=solid];
-"1890 add_57" -> "1903 add_58"  [label="(1, 14, 14, 384)", style=solid];
-"1891 _param_constant276" -> "1893 linear_102"  [label="(1536, 384)", style=solid];
-"1892 _param_constant277" -> "1893 linear_102"  [label="(1536,)", style=solid];
-"1893 linear_102" -> "1894 gelu_16"  [label="(1, 14, 14, 1536)", style=solid];
-"1894 gelu_16" -> "1895 dropout_66"  [label="(1, 14, 14, 1536)", style=solid];
-"1895 dropout_66" -> "1898 linear_103"  [label="(1, 14, 14, 1536)", style=solid];
-"1896 _param_constant278" -> "1898 linear_103"  [label="(384, 1536)", style=solid];
-"1897 _param_constant279" -> "1898 linear_103"  [label="(384,)", style=solid];
-"1898 linear_103" -> "1899 dropout_67"  [label="(1, 14, 14, 384)", style=solid];
-"1899 dropout_67" -> "1902 layer_norm_36"  [label="(1, 14, 14, 384)", style=solid];
-"1900 _param_constant280" -> "1902 layer_norm_36"  [label="(384,)", style=solid];
-"1901 _param_constant281" -> "1902 layer_norm_36"  [label="(384,)", style=solid];
-"1902 layer_norm_36" -> "1903 add_58"  [label="(1, 14, 14, 384)", style=solid];
-"1903 add_58" -> "1920 pad_19"  [label="(1, 14, 14, 384)", style=solid];
-"1903 add_58" -> "2033 add_61"  [label="(1, 14, 14, 384)", style=solid];
-"1904 _tensor_constant106" -> "1907 linear_104"  [label="(1, 15, 15, 2)", style=solid];
-"1905 _param_constant282" -> "1907 linear_104"  [label="(512, 2)", style=solid];
-"1906 _param_constant283" -> "1907 linear_104"  [label="(512,)", style=solid];
-"1907 linear_104" -> "1908 relu__17"  [label="(1, 15, 15, 512)", style=solid];
-"1908 relu__17" -> "1910 linear_105"  [label="(1, 15, 15, 512)", style=solid];
-"1909 _param_constant284" -> "1910 linear_105"  [label="(12, 512)", style=solid];
-"1910 linear_105" -> "1911 view_92"  [label="(1, 15, 15, 12)", style=solid];
-"1911 view_92" -> "1913 index_17"  [label="(225, 12)", style=solid];
-"1912 _tensor_constant107" -> "1913 index_17"  [label="(4096,)", style=solid];
-"1913 index_17" -> "1914 view_93"  [label="(4096, 12)", style=solid];
-"1914 view_93" -> "1915 permute_77"  [label="(64, 64, 12)", style=solid];
-"1915 permute_77" -> "1916 contiguous_32"  [label="(12, 64, 64)", style=solid];
-"1916 contiguous_32" -> "1917 unsqueeze_49"  [label="(12, 64, 64)", style=solid];
-"1917 unsqueeze_49" -> "1918 sigmoid_17"  [label="(1, 12, 64, 64)", style=solid];
-"1918 sigmoid_17" -> "1919 mul_34"  [label="(1, 12, 64, 64)", style=solid];
-"1919 mul_34" -> "1950 add_59"  [label="(1, 12, 64, 64)", style=solid];
-"1920 pad_19" -> "1921 roll_16"  [label="(1, 16, 16, 384)", style=solid];
-"1921 roll_16" -> "1922 view_94"  [label="(1, 16, 16, 384)", style=solid];
-"1922 view_94" -> "1923 permute_78"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"1923 permute_78" -> "1924 reshape_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"1924 reshape_76" -> "1930 linear_106"  [label="(4, 64, 384)", style=solid];
-"1924 reshape_76" -> "1951 new_zeros_8"  [label="(4, 64, 384)", style=solid];
-"1925 _param_constant285" -> "1926 clone_17"  [label="(1152,)", style=solid];
-"1926 clone_17" -> "1927 slice_250"  [label="(1152,)", style=solid];
-"1926 clone_17" -> "1930 linear_106"  [label="(1152,)", style=solid];
-"1927 slice_250" -> "1928 zero__17"  [label="(384,)", style=solid];
-"1929 _param_constant286" -> "1930 linear_106"  [label="(1152, 384)", style=solid];
-"1930 linear_106" -> "1931 reshape_77"  [label="(4, 64, 1152)", style=solid];
-"1931 reshape_77" -> "1932 permute_79"  [label="(4, 64, 3, 12, 32)", style=solid];
-"1932 permute_79" -> "1933 select_51"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1932 permute_79" -> "1934 select_52"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1932 permute_79" -> "1935 select_53"  [label="(3, 4, 12, 64, 32)", style=solid];
-"1933 select_51" -> "1936 linalg_vector_norm_34"  [label="(4, 12, 64, 32)", style=solid];
-"1933 select_51" -> "1938 expand_as_34"  [label="(4, 12, 64, 32)", style=solid];
-"1933 select_51" -> "1939 div_34"  [label="(4, 12, 64, 32)", style=solid];
-"1934 select_52" -> "1940 linalg_vector_norm_35"  [label="(4, 12, 64, 32)", style=solid];
-"1934 select_52" -> "1942 expand_as_35"  [label="(4, 12, 64, 32)", style=solid];
-"1934 select_52" -> "1943 div_35"  [label="(4, 12, 64, 32)", style=solid];
-"1935 select_53" -> "2014 matmul_35"  [label="(4, 12, 64, 32)", style=solid];
-"1936 linalg_vector_norm_34" -> "1937 clamp_min_34"  [label="(4, 12, 64, 1)", style=solid];
-"1937 clamp_min_34" -> "1938 expand_as_34"  [label="(4, 12, 64, 1)", style=solid];
-"1938 expand_as_34" -> "1939 div_34"  [label="(4, 12, 64, 32)", style=solid];
-"1939 div_34" -> "1945 matmul_34"  [label="(4, 12, 64, 32)", style=solid];
-"1940 linalg_vector_norm_35" -> "1941 clamp_min_35"  [label="(4, 12, 64, 1)", style=solid];
-"1941 clamp_min_35" -> "1942 expand_as_35"  [label="(4, 12, 64, 1)", style=solid];
-"1942 expand_as_35" -> "1943 div_35"  [label="(4, 12, 64, 32)", style=solid];
-"1943 div_35" -> "1944 transpose_34"  [label="(4, 12, 64, 32)", style=solid];
-"1944 transpose_34" -> "1945 matmul_34"  [label="(4, 12, 32, 64)", style=solid];
-"1945 matmul_34" -> "1949 mul_35"  [label="(4, 12, 64, 64)", style=solid];
-"1946 _param_constant287" -> "1947 clamp_17"  [label="(12, 1, 1)", style=solid];
-"1947 clamp_17" -> "1948 exp_17"  [label="(12, 1, 1)", style=solid];
-"1948 exp_17" -> "1949 mul_35"  [label="(12, 1, 1)", style=solid];
-"1949 mul_35" -> "1950 add_59"  [label="(4, 12, 64, 64)", style=solid];
-"1950 add_59" -> "2007 view_96"  [label="(4, 12, 64, 64)", style=solid];
-"1951 new_zeros_8" -> "1954 slice_251"  [label="(16, 16)", style=solid];
-"1951 new_zeros_8" -> "1959 slice_253"  [label="(16, 16)", style=solid];
-"1951 new_zeros_8" -> "1964 slice_255"  [label="(16, 16)", style=solid];
-"1951 new_zeros_8" -> "1969 slice_257"  [label="(16, 16)", style=solid];
-"1951 new_zeros_8" -> "1974 slice_259"  [label="(16, 16)", style=solid];
-"1951 new_zeros_8" -> "1979 slice_261"  [label="(16, 16)", style=solid];
-"1951 new_zeros_8" -> "1984 slice_263"  [label="(16, 16)", style=solid];
-"1951 new_zeros_8" -> "1989 slice_265"  [label="(16, 16)", style=solid];
-"1951 new_zeros_8" -> "1994 slice_267"  [label="(16, 16)", style=solid];
-"1951 new_zeros_8" -> "1997 view_95"  [label="(16, 16)", style=solid];
-"1952 _tensor_constant108" -> "1953 lift_fresh_copy_72"  [label="()", style=solid];
-"1953 lift_fresh_copy_72" -> "1956 fill__72"  [label="()", style=solid];
-"1954 slice_251" -> "1955 slice_252"  [label="(8, 16)", style=solid];
-"1955 slice_252" -> "1956 fill__72"  [label="(8, 8)", style=solid];
-"1957 _tensor_constant109" -> "1958 lift_fresh_copy_73"  [label="()", style=solid];
-"1958 lift_fresh_copy_73" -> "1961 fill__73"  [label="()", style=solid];
-"1959 slice_253" -> "1960 slice_254"  [label="(8, 16)", style=solid];
-"1960 slice_254" -> "1961 fill__73"  [label="(8, 4)", style=solid];
-"1962 _tensor_constant110" -> "1963 lift_fresh_copy_74"  [label="()", style=solid];
-"1963 lift_fresh_copy_74" -> "1966 fill__74"  [label="()", style=solid];
-"1964 slice_255" -> "1965 slice_256"  [label="(8, 16)", style=solid];
-"1965 slice_256" -> "1966 fill__74"  [label="(8, 4)", style=solid];
-"1967 _tensor_constant111" -> "1968 lift_fresh_copy_75"  [label="()", style=solid];
-"1968 lift_fresh_copy_75" -> "1971 fill__75"  [label="()", style=solid];
-"1969 slice_257" -> "1970 slice_258"  [label="(4, 16)", style=solid];
-"1970 slice_258" -> "1971 fill__75"  [label="(4, 8)", style=solid];
-"1972 _tensor_constant112" -> "1973 lift_fresh_copy_76"  [label="()", style=solid];
-"1973 lift_fresh_copy_76" -> "1976 fill__76"  [label="()", style=solid];
-"1974 slice_259" -> "1975 slice_260"  [label="(4, 16)", style=solid];
-"1975 slice_260" -> "1976 fill__76"  [label="(4, 4)", style=solid];
-"1977 _tensor_constant113" -> "1978 lift_fresh_copy_77"  [label="()", style=solid];
-"1978 lift_fresh_copy_77" -> "1981 fill__77"  [label="()", style=solid];
-"1979 slice_261" -> "1980 slice_262"  [label="(4, 16)", style=solid];
-"1980 slice_262" -> "1981 fill__77"  [label="(4, 4)", style=solid];
-"1982 _tensor_constant114" -> "1983 lift_fresh_copy_78"  [label="()", style=solid];
-"1983 lift_fresh_copy_78" -> "1986 fill__78"  [label="()", style=solid];
-"1984 slice_263" -> "1985 slice_264"  [label="(4, 16)", style=solid];
-"1985 slice_264" -> "1986 fill__78"  [label="(4, 8)", style=solid];
-"1987 _tensor_constant115" -> "1988 lift_fresh_copy_79"  [label="()", style=solid];
-"1988 lift_fresh_copy_79" -> "1991 fill__79"  [label="()", style=solid];
-"1989 slice_265" -> "1990 slice_266"  [label="(4, 16)", style=solid];
-"1990 slice_266" -> "1991 fill__79"  [label="(4, 4)", style=solid];
-"1992 _tensor_constant116" -> "1993 lift_fresh_copy_80"  [label="()", style=solid];
-"1993 lift_fresh_copy_80" -> "1996 fill__80"  [label="()", style=solid];
-"1994 slice_267" -> "1995 slice_268"  [label="(4, 16)", style=solid];
-"1995 slice_268" -> "1996 fill__80"  [label="(4, 4)", style=solid];
-"1997 view_95" -> "1998 permute_80"  [label="(2, 8, 2, 8)", style=solid];
-"1998 permute_80" -> "1999 reshape_78"  [label="(2, 2, 8, 8)", style=solid];
-"1999 reshape_78" -> "2000 unsqueeze_50"  [label="(4, 64)", style=solid];
-"1999 reshape_78" -> "2001 unsqueeze_51"  [label="(4, 64)", style=solid];
-"2000 unsqueeze_50" -> "2002 sub_8"  [label="(4, 1, 64)", style=solid];
-"2001 unsqueeze_51" -> "2002 sub_8"  [label="(4, 64, 1)", style=solid];
-"2002 sub_8" -> "2003 ne_8"  [label="(4, 64, 64)", style=solid];
-"2002 sub_8" -> "2004 masked_fill_16"  [label="(4, 64, 64)", style=solid];
-"2002 sub_8" -> "2005 eq_8"  [label="(4, 64, 64)", style=solid];
-"2003 ne_8" -> "2004 masked_fill_16"  [label="(4, 64, 64)", style=solid];
-"2004 masked_fill_16" -> "2006 masked_fill_17"  [label="(4, 64, 64)", style=solid];
-"2005 eq_8" -> "2006 masked_fill_17"  [label="(4, 64, 64)", style=solid];
-"2006 masked_fill_17" -> "2008 unsqueeze_52"  [label="(4, 64, 64)", style=solid];
-"2007 view_96" -> "2010 add_60"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2008 unsqueeze_52" -> "2009 unsqueeze_53"  [label="(4, 1, 64, 64)", style=solid];
-"2009 unsqueeze_53" -> "2010 add_60"  [label="(1, 4, 1, 64, 64)", style=solid];
-"2010 add_60" -> "2011 view_97"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2011 view_97" -> "2012 softmax_17"  [label="(4, 12, 64, 64)", style=solid];
-"2012 softmax_17" -> "2013 dropout_68"  [label="(4, 12, 64, 64)", style=solid];
-"2013 dropout_68" -> "2014 matmul_35"  [label="(4, 12, 64, 64)", style=solid];
-"2014 matmul_35" -> "2015 transpose_35"  [label="(4, 12, 64, 32)", style=solid];
-"2015 transpose_35" -> "2016 reshape_79"  [label="(4, 64, 12, 32)", style=solid];
-"2016 reshape_79" -> "2019 linear_107"  [label="(4, 64, 384)", style=solid];
-"2017 _param_constant288" -> "2019 linear_107"  [label="(384, 384)", style=solid];
-"2018 _param_constant289" -> "2019 linear_107"  [label="(384,)", style=solid];
-"2019 linear_107" -> "2020 dropout_69"  [label="(4, 64, 384)", style=solid];
-"2020 dropout_69" -> "2021 view_98"  [label="(4, 64, 384)", style=solid];
-"2021 view_98" -> "2022 permute_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2022 permute_81" -> "2023 reshape_80"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2023 reshape_80" -> "2024 roll_17"  [label="(1, 16, 16, 384)", style=solid];
-"2024 roll_17" -> "2025 slice_269"  [label="(1, 16, 16, 384)", style=solid];
-"2025 slice_269" -> "2026 slice_270"  [label="(1, 16, 16, 384)", style=solid];
-"2026 slice_270" -> "2027 slice_271"  [label="(1, 14, 16, 384)", style=solid];
-"2027 slice_271" -> "2028 slice_272"  [label="(1, 14, 14, 384)", style=solid];
-"2028 slice_272" -> "2029 contiguous_33"  [label="(1, 14, 14, 384)", style=solid];
-"2029 contiguous_33" -> "2032 layer_norm_37"  [label="(1, 14, 14, 384)", style=solid];
-"2030 _param_constant290" -> "2032 layer_norm_37"  [label="(384,)", style=solid];
-"2031 _param_constant291" -> "2032 layer_norm_37"  [label="(384,)", style=solid];
-"2032 layer_norm_37" -> "2033 add_61"  [label="(1, 14, 14, 384)", style=solid];
-"2033 add_61" -> "2036 linear_108"  [label="(1, 14, 14, 384)", style=solid];
-"2033 add_61" -> "2046 add_62"  [label="(1, 14, 14, 384)", style=solid];
-"2034 _param_constant292" -> "2036 linear_108"  [label="(1536, 384)", style=solid];
-"2035 _param_constant293" -> "2036 linear_108"  [label="(1536,)", style=solid];
-"2036 linear_108" -> "2037 gelu_17"  [label="(1, 14, 14, 1536)", style=solid];
-"2037 gelu_17" -> "2038 dropout_70"  [label="(1, 14, 14, 1536)", style=solid];
-"2038 dropout_70" -> "2041 linear_109"  [label="(1, 14, 14, 1536)", style=solid];
-"2039 _param_constant294" -> "2041 linear_109"  [label="(384, 1536)", style=solid];
-"2040 _param_constant295" -> "2041 linear_109"  [label="(384,)", style=solid];
-"2041 linear_109" -> "2042 dropout_71"  [label="(1, 14, 14, 384)", style=solid];
-"2042 dropout_71" -> "2045 layer_norm_38"  [label="(1, 14, 14, 384)", style=solid];
-"2043 _param_constant296" -> "2045 layer_norm_38"  [label="(384,)", style=solid];
-"2044 _param_constant297" -> "2045 layer_norm_38"  [label="(384,)", style=solid];
-"2045 layer_norm_38" -> "2046 add_62"  [label="(1, 14, 14, 384)", style=solid];
-"2046 add_62" -> "2063 pad_20"  [label="(1, 14, 14, 384)", style=solid];
-"2046 add_62" -> "2113 add_64"  [label="(1, 14, 14, 384)", style=solid];
-"2047 _tensor_constant117" -> "2050 linear_110"  [label="(1, 15, 15, 2)", style=solid];
-"2048 _param_constant298" -> "2050 linear_110"  [label="(512, 2)", style=solid];
-"2049 _param_constant299" -> "2050 linear_110"  [label="(512,)", style=solid];
-"2050 linear_110" -> "2051 relu__18"  [label="(1, 15, 15, 512)", style=solid];
-"2051 relu__18" -> "2053 linear_111"  [label="(1, 15, 15, 512)", style=solid];
-"2052 _param_constant300" -> "2053 linear_111"  [label="(12, 512)", style=solid];
-"2053 linear_111" -> "2054 view_99"  [label="(1, 15, 15, 12)", style=solid];
-"2054 view_99" -> "2056 index_18"  [label="(225, 12)", style=solid];
-"2055 _tensor_constant118" -> "2056 index_18"  [label="(4096,)", style=solid];
-"2056 index_18" -> "2057 view_100"  [label="(4096, 12)", style=solid];
-"2057 view_100" -> "2058 permute_82"  [label="(64, 64, 12)", style=solid];
-"2058 permute_82" -> "2059 contiguous_34"  [label="(12, 64, 64)", style=solid];
-"2059 contiguous_34" -> "2060 unsqueeze_54"  [label="(12, 64, 64)", style=solid];
-"2060 unsqueeze_54" -> "2061 sigmoid_18"  [label="(1, 12, 64, 64)", style=solid];
-"2061 sigmoid_18" -> "2062 mul_36"  [label="(1, 12, 64, 64)", style=solid];
-"2062 mul_36" -> "2092 add_63"  [label="(1, 12, 64, 64)", style=solid];
-"2063 pad_20" -> "2064 view_101"  [label="(1, 16, 16, 384)", style=solid];
-"2064 view_101" -> "2065 permute_83"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2065 permute_83" -> "2066 reshape_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2066 reshape_81" -> "2072 linear_112"  [label="(4, 64, 384)", style=solid];
-"2067 _param_constant301" -> "2068 clone_18"  [label="(1152,)", style=solid];
-"2068 clone_18" -> "2069 slice_273"  [label="(1152,)", style=solid];
-"2068 clone_18" -> "2072 linear_112"  [label="(1152,)", style=solid];
-"2069 slice_273" -> "2070 zero__18"  [label="(384,)", style=solid];
-"2071 _param_constant302" -> "2072 linear_112"  [label="(1152, 384)", style=solid];
-"2072 linear_112" -> "2073 reshape_82"  [label="(4, 64, 1152)", style=solid];
-"2073 reshape_82" -> "2074 permute_84"  [label="(4, 64, 3, 12, 32)", style=solid];
-"2074 permute_84" -> "2075 select_54"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2074 permute_84" -> "2076 select_55"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2074 permute_84" -> "2077 select_56"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2075 select_54" -> "2078 linalg_vector_norm_36"  [label="(4, 12, 64, 32)", style=solid];
-"2075 select_54" -> "2080 expand_as_36"  [label="(4, 12, 64, 32)", style=solid];
-"2075 select_54" -> "2081 div_36"  [label="(4, 12, 64, 32)", style=solid];
-"2076 select_55" -> "2082 linalg_vector_norm_37"  [label="(4, 12, 64, 32)", style=solid];
-"2076 select_55" -> "2084 expand_as_37"  [label="(4, 12, 64, 32)", style=solid];
-"2076 select_55" -> "2085 div_37"  [label="(4, 12, 64, 32)", style=solid];
-"2077 select_56" -> "2095 matmul_37"  [label="(4, 12, 64, 32)", style=solid];
-"2078 linalg_vector_norm_36" -> "2079 clamp_min_36"  [label="(4, 12, 64, 1)", style=solid];
-"2079 clamp_min_36" -> "2080 expand_as_36"  [label="(4, 12, 64, 1)", style=solid];
-"2080 expand_as_36" -> "2081 div_36"  [label="(4, 12, 64, 32)", style=solid];
-"2081 div_36" -> "2087 matmul_36"  [label="(4, 12, 64, 32)", style=solid];
-"2082 linalg_vector_norm_37" -> "2083 clamp_min_37"  [label="(4, 12, 64, 1)", style=solid];
-"2083 clamp_min_37" -> "2084 expand_as_37"  [label="(4, 12, 64, 1)", style=solid];
-"2084 expand_as_37" -> "2085 div_37"  [label="(4, 12, 64, 32)", style=solid];
-"2085 div_37" -> "2086 transpose_36"  [label="(4, 12, 64, 32)", style=solid];
-"2086 transpose_36" -> "2087 matmul_36"  [label="(4, 12, 32, 64)", style=solid];
-"2087 matmul_36" -> "2091 mul_37"  [label="(4, 12, 64, 64)", style=solid];
-"2088 _param_constant303" -> "2089 clamp_18"  [label="(12, 1, 1)", style=solid];
-"2089 clamp_18" -> "2090 exp_18"  [label="(12, 1, 1)", style=solid];
-"2090 exp_18" -> "2091 mul_37"  [label="(12, 1, 1)", style=solid];
-"2091 mul_37" -> "2092 add_63"  [label="(4, 12, 64, 64)", style=solid];
-"2092 add_63" -> "2093 softmax_18"  [label="(4, 12, 64, 64)", style=solid];
-"2093 softmax_18" -> "2094 dropout_72"  [label="(4, 12, 64, 64)", style=solid];
-"2094 dropout_72" -> "2095 matmul_37"  [label="(4, 12, 64, 64)", style=solid];
-"2095 matmul_37" -> "2096 transpose_37"  [label="(4, 12, 64, 32)", style=solid];
-"2096 transpose_37" -> "2097 reshape_83"  [label="(4, 64, 12, 32)", style=solid];
-"2097 reshape_83" -> "2100 linear_113"  [label="(4, 64, 384)", style=solid];
-"2098 _param_constant304" -> "2100 linear_113"  [label="(384, 384)", style=solid];
-"2099 _param_constant305" -> "2100 linear_113"  [label="(384,)", style=solid];
-"2100 linear_113" -> "2101 dropout_73"  [label="(4, 64, 384)", style=solid];
-"2101 dropout_73" -> "2102 view_102"  [label="(4, 64, 384)", style=solid];
-"2102 view_102" -> "2103 permute_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2103 permute_85" -> "2104 reshape_84"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2104 reshape_84" -> "2105 slice_274"  [label="(1, 16, 16, 384)", style=solid];
-"2105 slice_274" -> "2106 slice_275"  [label="(1, 16, 16, 384)", style=solid];
-"2106 slice_275" -> "2107 slice_276"  [label="(1, 14, 16, 384)", style=solid];
-"2107 slice_276" -> "2108 slice_277"  [label="(1, 14, 14, 384)", style=solid];
-"2108 slice_277" -> "2109 contiguous_35"  [label="(1, 14, 14, 384)", style=solid];
-"2109 contiguous_35" -> "2112 layer_norm_39"  [label="(1, 14, 14, 384)", style=solid];
-"2110 _param_constant306" -> "2112 layer_norm_39"  [label="(384,)", style=solid];
-"2111 _param_constant307" -> "2112 layer_norm_39"  [label="(384,)", style=solid];
-"2112 layer_norm_39" -> "2113 add_64"  [label="(1, 14, 14, 384)", style=solid];
-"2113 add_64" -> "2116 linear_114"  [label="(1, 14, 14, 384)", style=solid];
-"2113 add_64" -> "2126 add_65"  [label="(1, 14, 14, 384)", style=solid];
-"2114 _param_constant308" -> "2116 linear_114"  [label="(1536, 384)", style=solid];
-"2115 _param_constant309" -> "2116 linear_114"  [label="(1536,)", style=solid];
-"2116 linear_114" -> "2117 gelu_18"  [label="(1, 14, 14, 1536)", style=solid];
-"2117 gelu_18" -> "2118 dropout_74"  [label="(1, 14, 14, 1536)", style=solid];
-"2118 dropout_74" -> "2121 linear_115"  [label="(1, 14, 14, 1536)", style=solid];
-"2119 _param_constant310" -> "2121 linear_115"  [label="(384, 1536)", style=solid];
-"2120 _param_constant311" -> "2121 linear_115"  [label="(384,)", style=solid];
-"2121 linear_115" -> "2122 dropout_75"  [label="(1, 14, 14, 384)", style=solid];
-"2122 dropout_75" -> "2125 layer_norm_40"  [label="(1, 14, 14, 384)", style=solid];
-"2123 _param_constant312" -> "2125 layer_norm_40"  [label="(384,)", style=solid];
-"2124 _param_constant313" -> "2125 layer_norm_40"  [label="(384,)", style=solid];
-"2125 layer_norm_40" -> "2126 add_65"  [label="(1, 14, 14, 384)", style=solid];
-"2126 add_65" -> "2143 pad_21"  [label="(1, 14, 14, 384)", style=solid];
-"2126 add_65" -> "2256 add_68"  [label="(1, 14, 14, 384)", style=solid];
-"2127 _tensor_constant119" -> "2130 linear_116"  [label="(1, 15, 15, 2)", style=solid];
-"2128 _param_constant314" -> "2130 linear_116"  [label="(512, 2)", style=solid];
-"2129 _param_constant315" -> "2130 linear_116"  [label="(512,)", style=solid];
-"2130 linear_116" -> "2131 relu__19"  [label="(1, 15, 15, 512)", style=solid];
-"2131 relu__19" -> "2133 linear_117"  [label="(1, 15, 15, 512)", style=solid];
-"2132 _param_constant316" -> "2133 linear_117"  [label="(12, 512)", style=solid];
-"2133 linear_117" -> "2134 view_103"  [label="(1, 15, 15, 12)", style=solid];
-"2134 view_103" -> "2136 index_19"  [label="(225, 12)", style=solid];
-"2135 _tensor_constant120" -> "2136 index_19"  [label="(4096,)", style=solid];
-"2136 index_19" -> "2137 view_104"  [label="(4096, 12)", style=solid];
-"2137 view_104" -> "2138 permute_86"  [label="(64, 64, 12)", style=solid];
-"2138 permute_86" -> "2139 contiguous_36"  [label="(12, 64, 64)", style=solid];
-"2139 contiguous_36" -> "2140 unsqueeze_55"  [label="(12, 64, 64)", style=solid];
-"2140 unsqueeze_55" -> "2141 sigmoid_19"  [label="(1, 12, 64, 64)", style=solid];
-"2141 sigmoid_19" -> "2142 mul_38"  [label="(1, 12, 64, 64)", style=solid];
-"2142 mul_38" -> "2173 add_66"  [label="(1, 12, 64, 64)", style=solid];
-"2143 pad_21" -> "2144 roll_18"  [label="(1, 16, 16, 384)", style=solid];
-"2144 roll_18" -> "2145 view_105"  [label="(1, 16, 16, 384)", style=solid];
-"2145 view_105" -> "2146 permute_87"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2146 permute_87" -> "2147 reshape_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2147 reshape_85" -> "2153 linear_118"  [label="(4, 64, 384)", style=solid];
-"2147 reshape_85" -> "2174 new_zeros_9"  [label="(4, 64, 384)", style=solid];
-"2148 _param_constant317" -> "2149 clone_19"  [label="(1152,)", style=solid];
-"2149 clone_19" -> "2150 slice_278"  [label="(1152,)", style=solid];
-"2149 clone_19" -> "2153 linear_118"  [label="(1152,)", style=solid];
-"2150 slice_278" -> "2151 zero__19"  [label="(384,)", style=solid];
-"2152 _param_constant318" -> "2153 linear_118"  [label="(1152, 384)", style=solid];
-"2153 linear_118" -> "2154 reshape_86"  [label="(4, 64, 1152)", style=solid];
-"2154 reshape_86" -> "2155 permute_88"  [label="(4, 64, 3, 12, 32)", style=solid];
-"2155 permute_88" -> "2156 select_57"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2155 permute_88" -> "2157 select_58"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2155 permute_88" -> "2158 select_59"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2156 select_57" -> "2159 linalg_vector_norm_38"  [label="(4, 12, 64, 32)", style=solid];
-"2156 select_57" -> "2161 expand_as_38"  [label="(4, 12, 64, 32)", style=solid];
-"2156 select_57" -> "2162 div_38"  [label="(4, 12, 64, 32)", style=solid];
-"2157 select_58" -> "2163 linalg_vector_norm_39"  [label="(4, 12, 64, 32)", style=solid];
-"2157 select_58" -> "2165 expand_as_39"  [label="(4, 12, 64, 32)", style=solid];
-"2157 select_58" -> "2166 div_39"  [label="(4, 12, 64, 32)", style=solid];
-"2158 select_59" -> "2237 matmul_39"  [label="(4, 12, 64, 32)", style=solid];
-"2159 linalg_vector_norm_38" -> "2160 clamp_min_38"  [label="(4, 12, 64, 1)", style=solid];
-"2160 clamp_min_38" -> "2161 expand_as_38"  [label="(4, 12, 64, 1)", style=solid];
-"2161 expand_as_38" -> "2162 div_38"  [label="(4, 12, 64, 32)", style=solid];
-"2162 div_38" -> "2168 matmul_38"  [label="(4, 12, 64, 32)", style=solid];
-"2163 linalg_vector_norm_39" -> "2164 clamp_min_39"  [label="(4, 12, 64, 1)", style=solid];
-"2164 clamp_min_39" -> "2165 expand_as_39"  [label="(4, 12, 64, 1)", style=solid];
-"2165 expand_as_39" -> "2166 div_39"  [label="(4, 12, 64, 32)", style=solid];
-"2166 div_39" -> "2167 transpose_38"  [label="(4, 12, 64, 32)", style=solid];
-"2167 transpose_38" -> "2168 matmul_38"  [label="(4, 12, 32, 64)", style=solid];
-"2168 matmul_38" -> "2172 mul_39"  [label="(4, 12, 64, 64)", style=solid];
-"2169 _param_constant319" -> "2170 clamp_19"  [label="(12, 1, 1)", style=solid];
-"2170 clamp_19" -> "2171 exp_19"  [label="(12, 1, 1)", style=solid];
-"2171 exp_19" -> "2172 mul_39"  [label="(12, 1, 1)", style=solid];
-"2172 mul_39" -> "2173 add_66"  [label="(4, 12, 64, 64)", style=solid];
-"2173 add_66" -> "2230 view_107"  [label="(4, 12, 64, 64)", style=solid];
-"2174 new_zeros_9" -> "2177 slice_279"  [label="(16, 16)", style=solid];
-"2174 new_zeros_9" -> "2182 slice_281"  [label="(16, 16)", style=solid];
-"2174 new_zeros_9" -> "2187 slice_283"  [label="(16, 16)", style=solid];
-"2174 new_zeros_9" -> "2192 slice_285"  [label="(16, 16)", style=solid];
-"2174 new_zeros_9" -> "2197 slice_287"  [label="(16, 16)", style=solid];
-"2174 new_zeros_9" -> "2202 slice_289"  [label="(16, 16)", style=solid];
-"2174 new_zeros_9" -> "2207 slice_291"  [label="(16, 16)", style=solid];
-"2174 new_zeros_9" -> "2212 slice_293"  [label="(16, 16)", style=solid];
-"2174 new_zeros_9" -> "2217 slice_295"  [label="(16, 16)", style=solid];
-"2174 new_zeros_9" -> "2220 view_106"  [label="(16, 16)", style=solid];
-"2175 _tensor_constant121" -> "2176 lift_fresh_copy_81"  [label="()", style=solid];
-"2176 lift_fresh_copy_81" -> "2179 fill__81"  [label="()", style=solid];
-"2177 slice_279" -> "2178 slice_280"  [label="(8, 16)", style=solid];
-"2178 slice_280" -> "2179 fill__81"  [label="(8, 8)", style=solid];
-"2180 _tensor_constant122" -> "2181 lift_fresh_copy_82"  [label="()", style=solid];
-"2181 lift_fresh_copy_82" -> "2184 fill__82"  [label="()", style=solid];
-"2182 slice_281" -> "2183 slice_282"  [label="(8, 16)", style=solid];
-"2183 slice_282" -> "2184 fill__82"  [label="(8, 4)", style=solid];
-"2185 _tensor_constant123" -> "2186 lift_fresh_copy_83"  [label="()", style=solid];
-"2186 lift_fresh_copy_83" -> "2189 fill__83"  [label="()", style=solid];
-"2187 slice_283" -> "2188 slice_284"  [label="(8, 16)", style=solid];
-"2188 slice_284" -> "2189 fill__83"  [label="(8, 4)", style=solid];
-"2190 _tensor_constant124" -> "2191 lift_fresh_copy_84"  [label="()", style=solid];
-"2191 lift_fresh_copy_84" -> "2194 fill__84"  [label="()", style=solid];
-"2192 slice_285" -> "2193 slice_286"  [label="(4, 16)", style=solid];
-"2193 slice_286" -> "2194 fill__84"  [label="(4, 8)", style=solid];
-"2195 _tensor_constant125" -> "2196 lift_fresh_copy_85"  [label="()", style=solid];
-"2196 lift_fresh_copy_85" -> "2199 fill__85"  [label="()", style=solid];
-"2197 slice_287" -> "2198 slice_288"  [label="(4, 16)", style=solid];
-"2198 slice_288" -> "2199 fill__85"  [label="(4, 4)", style=solid];
-"2200 _tensor_constant126" -> "2201 lift_fresh_copy_86"  [label="()", style=solid];
-"2201 lift_fresh_copy_86" -> "2204 fill__86"  [label="()", style=solid];
-"2202 slice_289" -> "2203 slice_290"  [label="(4, 16)", style=solid];
-"2203 slice_290" -> "2204 fill__86"  [label="(4, 4)", style=solid];
-"2205 _tensor_constant127" -> "2206 lift_fresh_copy_87"  [label="()", style=solid];
-"2206 lift_fresh_copy_87" -> "2209 fill__87"  [label="()", style=solid];
-"2207 slice_291" -> "2208 slice_292"  [label="(4, 16)", style=solid];
-"2208 slice_292" -> "2209 fill__87"  [label="(4, 8)", style=solid];
-"2210 _tensor_constant128" -> "2211 lift_fresh_copy_88"  [label="()", style=solid];
-"2211 lift_fresh_copy_88" -> "2214 fill__88"  [label="()", style=solid];
-"2212 slice_293" -> "2213 slice_294"  [label="(4, 16)", style=solid];
-"2213 slice_294" -> "2214 fill__88"  [label="(4, 4)", style=solid];
-"2215 _tensor_constant129" -> "2216 lift_fresh_copy_89"  [label="()", style=solid];
-"2216 lift_fresh_copy_89" -> "2219 fill__89"  [label="()", style=solid];
-"2217 slice_295" -> "2218 slice_296"  [label="(4, 16)", style=solid];
-"2218 slice_296" -> "2219 fill__89"  [label="(4, 4)", style=solid];
-"2220 view_106" -> "2221 permute_89"  [label="(2, 8, 2, 8)", style=solid];
-"2221 permute_89" -> "2222 reshape_87"  [label="(2, 2, 8, 8)", style=solid];
-"2222 reshape_87" -> "2223 unsqueeze_56"  [label="(4, 64)", style=solid];
-"2222 reshape_87" -> "2224 unsqueeze_57"  [label="(4, 64)", style=solid];
-"2223 unsqueeze_56" -> "2225 sub_9"  [label="(4, 1, 64)", style=solid];
-"2224 unsqueeze_57" -> "2225 sub_9"  [label="(4, 64, 1)", style=solid];
-"2225 sub_9" -> "2226 ne_9"  [label="(4, 64, 64)", style=solid];
-"2225 sub_9" -> "2227 masked_fill_18"  [label="(4, 64, 64)", style=solid];
-"2225 sub_9" -> "2228 eq_9"  [label="(4, 64, 64)", style=solid];
-"2226 ne_9" -> "2227 masked_fill_18"  [label="(4, 64, 64)", style=solid];
-"2227 masked_fill_18" -> "2229 masked_fill_19"  [label="(4, 64, 64)", style=solid];
-"2228 eq_9" -> "2229 masked_fill_19"  [label="(4, 64, 64)", style=solid];
-"2229 masked_fill_19" -> "2231 unsqueeze_58"  [label="(4, 64, 64)", style=solid];
-"2230 view_107" -> "2233 add_67"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2231 unsqueeze_58" -> "2232 unsqueeze_59"  [label="(4, 1, 64, 64)", style=solid];
-"2232 unsqueeze_59" -> "2233 add_67"  [label="(1, 4, 1, 64, 64)", style=solid];
-"2233 add_67" -> "2234 view_108"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2234 view_108" -> "2235 softmax_19"  [label="(4, 12, 64, 64)", style=solid];
-"2235 softmax_19" -> "2236 dropout_76"  [label="(4, 12, 64, 64)", style=solid];
-"2236 dropout_76" -> "2237 matmul_39"  [label="(4, 12, 64, 64)", style=solid];
-"2237 matmul_39" -> "2238 transpose_39"  [label="(4, 12, 64, 32)", style=solid];
-"2238 transpose_39" -> "2239 reshape_88"  [label="(4, 64, 12, 32)", style=solid];
-"2239 reshape_88" -> "2242 linear_119"  [label="(4, 64, 384)", style=solid];
-"2240 _param_constant320" -> "2242 linear_119"  [label="(384, 384)", style=solid];
-"2241 _param_constant321" -> "2242 linear_119"  [label="(384,)", style=solid];
-"2242 linear_119" -> "2243 dropout_77"  [label="(4, 64, 384)", style=solid];
-"2243 dropout_77" -> "2244 view_109"  [label="(4, 64, 384)", style=solid];
-"2244 view_109" -> "2245 permute_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2245 permute_90" -> "2246 reshape_89"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2246 reshape_89" -> "2247 roll_19"  [label="(1, 16, 16, 384)", style=solid];
-"2247 roll_19" -> "2248 slice_297"  [label="(1, 16, 16, 384)", style=solid];
-"2248 slice_297" -> "2249 slice_298"  [label="(1, 16, 16, 384)", style=solid];
-"2249 slice_298" -> "2250 slice_299"  [label="(1, 14, 16, 384)", style=solid];
-"2250 slice_299" -> "2251 slice_300"  [label="(1, 14, 14, 384)", style=solid];
-"2251 slice_300" -> "2252 contiguous_37"  [label="(1, 14, 14, 384)", style=solid];
-"2252 contiguous_37" -> "2255 layer_norm_41"  [label="(1, 14, 14, 384)", style=solid];
-"2253 _param_constant322" -> "2255 layer_norm_41"  [label="(384,)", style=solid];
-"2254 _param_constant323" -> "2255 layer_norm_41"  [label="(384,)", style=solid];
-"2255 layer_norm_41" -> "2256 add_68"  [label="(1, 14, 14, 384)", style=solid];
-"2256 add_68" -> "2259 linear_120"  [label="(1, 14, 14, 384)", style=solid];
-"2256 add_68" -> "2269 add_69"  [label="(1, 14, 14, 384)", style=solid];
-"2257 _param_constant324" -> "2259 linear_120"  [label="(1536, 384)", style=solid];
-"2258 _param_constant325" -> "2259 linear_120"  [label="(1536,)", style=solid];
-"2259 linear_120" -> "2260 gelu_19"  [label="(1, 14, 14, 1536)", style=solid];
-"2260 gelu_19" -> "2261 dropout_78"  [label="(1, 14, 14, 1536)", style=solid];
-"2261 dropout_78" -> "2264 linear_121"  [label="(1, 14, 14, 1536)", style=solid];
-"2262 _param_constant326" -> "2264 linear_121"  [label="(384, 1536)", style=solid];
-"2263 _param_constant327" -> "2264 linear_121"  [label="(384,)", style=solid];
-"2264 linear_121" -> "2265 dropout_79"  [label="(1, 14, 14, 384)", style=solid];
-"2265 dropout_79" -> "2268 layer_norm_42"  [label="(1, 14, 14, 384)", style=solid];
-"2266 _param_constant328" -> "2268 layer_norm_42"  [label="(384,)", style=solid];
-"2267 _param_constant329" -> "2268 layer_norm_42"  [label="(384,)", style=solid];
-"2268 layer_norm_42" -> "2269 add_69"  [label="(1, 14, 14, 384)", style=solid];
-"2269 add_69" -> "2286 pad_22"  [label="(1, 14, 14, 384)", style=solid];
-"2269 add_69" -> "2336 add_71"  [label="(1, 14, 14, 384)", style=solid];
-"2270 _tensor_constant130" -> "2273 linear_122"  [label="(1, 15, 15, 2)", style=solid];
-"2271 _param_constant330" -> "2273 linear_122"  [label="(512, 2)", style=solid];
-"2272 _param_constant331" -> "2273 linear_122"  [label="(512,)", style=solid];
-"2273 linear_122" -> "2274 relu__20"  [label="(1, 15, 15, 512)", style=solid];
-"2274 relu__20" -> "2276 linear_123"  [label="(1, 15, 15, 512)", style=solid];
-"2275 _param_constant332" -> "2276 linear_123"  [label="(12, 512)", style=solid];
-"2276 linear_123" -> "2277 view_110"  [label="(1, 15, 15, 12)", style=solid];
-"2277 view_110" -> "2279 index_20"  [label="(225, 12)", style=solid];
-"2278 _tensor_constant131" -> "2279 index_20"  [label="(4096,)", style=solid];
-"2279 index_20" -> "2280 view_111"  [label="(4096, 12)", style=solid];
-"2280 view_111" -> "2281 permute_91"  [label="(64, 64, 12)", style=solid];
-"2281 permute_91" -> "2282 contiguous_38"  [label="(12, 64, 64)", style=solid];
-"2282 contiguous_38" -> "2283 unsqueeze_60"  [label="(12, 64, 64)", style=solid];
-"2283 unsqueeze_60" -> "2284 sigmoid_20"  [label="(1, 12, 64, 64)", style=solid];
-"2284 sigmoid_20" -> "2285 mul_40"  [label="(1, 12, 64, 64)", style=solid];
-"2285 mul_40" -> "2315 add_70"  [label="(1, 12, 64, 64)", style=solid];
-"2286 pad_22" -> "2287 view_112"  [label="(1, 16, 16, 384)", style=solid];
-"2287 view_112" -> "2288 permute_92"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2288 permute_92" -> "2289 reshape_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2289 reshape_90" -> "2295 linear_124"  [label="(4, 64, 384)", style=solid];
-"2290 _param_constant333" -> "2291 clone_20"  [label="(1152,)", style=solid];
-"2291 clone_20" -> "2292 slice_301"  [label="(1152,)", style=solid];
-"2291 clone_20" -> "2295 linear_124"  [label="(1152,)", style=solid];
-"2292 slice_301" -> "2293 zero__20"  [label="(384,)", style=solid];
-"2294 _param_constant334" -> "2295 linear_124"  [label="(1152, 384)", style=solid];
-"2295 linear_124" -> "2296 reshape_91"  [label="(4, 64, 1152)", style=solid];
-"2296 reshape_91" -> "2297 permute_93"  [label="(4, 64, 3, 12, 32)", style=solid];
-"2297 permute_93" -> "2298 select_60"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2297 permute_93" -> "2299 select_61"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2297 permute_93" -> "2300 select_62"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2298 select_60" -> "2301 linalg_vector_norm_40"  [label="(4, 12, 64, 32)", style=solid];
-"2298 select_60" -> "2303 expand_as_40"  [label="(4, 12, 64, 32)", style=solid];
-"2298 select_60" -> "2304 div_40"  [label="(4, 12, 64, 32)", style=solid];
-"2299 select_61" -> "2305 linalg_vector_norm_41"  [label="(4, 12, 64, 32)", style=solid];
-"2299 select_61" -> "2307 expand_as_41"  [label="(4, 12, 64, 32)", style=solid];
-"2299 select_61" -> "2308 div_41"  [label="(4, 12, 64, 32)", style=solid];
-"2300 select_62" -> "2318 matmul_41"  [label="(4, 12, 64, 32)", style=solid];
-"2301 linalg_vector_norm_40" -> "2302 clamp_min_40"  [label="(4, 12, 64, 1)", style=solid];
-"2302 clamp_min_40" -> "2303 expand_as_40"  [label="(4, 12, 64, 1)", style=solid];
-"2303 expand_as_40" -> "2304 div_40"  [label="(4, 12, 64, 32)", style=solid];
-"2304 div_40" -> "2310 matmul_40"  [label="(4, 12, 64, 32)", style=solid];
-"2305 linalg_vector_norm_41" -> "2306 clamp_min_41"  [label="(4, 12, 64, 1)", style=solid];
-"2306 clamp_min_41" -> "2307 expand_as_41"  [label="(4, 12, 64, 1)", style=solid];
-"2307 expand_as_41" -> "2308 div_41"  [label="(4, 12, 64, 32)", style=solid];
-"2308 div_41" -> "2309 transpose_40"  [label="(4, 12, 64, 32)", style=solid];
-"2309 transpose_40" -> "2310 matmul_40"  [label="(4, 12, 32, 64)", style=solid];
-"2310 matmul_40" -> "2314 mul_41"  [label="(4, 12, 64, 64)", style=solid];
-"2311 _param_constant335" -> "2312 clamp_20"  [label="(12, 1, 1)", style=solid];
-"2312 clamp_20" -> "2313 exp_20"  [label="(12, 1, 1)", style=solid];
-"2313 exp_20" -> "2314 mul_41"  [label="(12, 1, 1)", style=solid];
-"2314 mul_41" -> "2315 add_70"  [label="(4, 12, 64, 64)", style=solid];
-"2315 add_70" -> "2316 softmax_20"  [label="(4, 12, 64, 64)", style=solid];
-"2316 softmax_20" -> "2317 dropout_80"  [label="(4, 12, 64, 64)", style=solid];
-"2317 dropout_80" -> "2318 matmul_41"  [label="(4, 12, 64, 64)", style=solid];
-"2318 matmul_41" -> "2319 transpose_41"  [label="(4, 12, 64, 32)", style=solid];
-"2319 transpose_41" -> "2320 reshape_92"  [label="(4, 64, 12, 32)", style=solid];
-"2320 reshape_92" -> "2323 linear_125"  [label="(4, 64, 384)", style=solid];
-"2321 _param_constant336" -> "2323 linear_125"  [label="(384, 384)", style=solid];
-"2322 _param_constant337" -> "2323 linear_125"  [label="(384,)", style=solid];
-"2323 linear_125" -> "2324 dropout_81"  [label="(4, 64, 384)", style=solid];
-"2324 dropout_81" -> "2325 view_113"  [label="(4, 64, 384)", style=solid];
-"2325 view_113" -> "2326 permute_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2326 permute_94" -> "2327 reshape_93"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2327 reshape_93" -> "2328 slice_302"  [label="(1, 16, 16, 384)", style=solid];
-"2328 slice_302" -> "2329 slice_303"  [label="(1, 16, 16, 384)", style=solid];
-"2329 slice_303" -> "2330 slice_304"  [label="(1, 14, 16, 384)", style=solid];
-"2330 slice_304" -> "2331 slice_305"  [label="(1, 14, 14, 384)", style=solid];
-"2331 slice_305" -> "2332 contiguous_39"  [label="(1, 14, 14, 384)", style=solid];
-"2332 contiguous_39" -> "2335 layer_norm_43"  [label="(1, 14, 14, 384)", style=solid];
-"2333 _param_constant338" -> "2335 layer_norm_43"  [label="(384,)", style=solid];
-"2334 _param_constant339" -> "2335 layer_norm_43"  [label="(384,)", style=solid];
-"2335 layer_norm_43" -> "2336 add_71"  [label="(1, 14, 14, 384)", style=solid];
-"2336 add_71" -> "2339 linear_126"  [label="(1, 14, 14, 384)", style=solid];
-"2336 add_71" -> "2349 add_72"  [label="(1, 14, 14, 384)", style=solid];
-"2337 _param_constant340" -> "2339 linear_126"  [label="(1536, 384)", style=solid];
-"2338 _param_constant341" -> "2339 linear_126"  [label="(1536,)", style=solid];
-"2339 linear_126" -> "2340 gelu_20"  [label="(1, 14, 14, 1536)", style=solid];
-"2340 gelu_20" -> "2341 dropout_82"  [label="(1, 14, 14, 1536)", style=solid];
-"2341 dropout_82" -> "2344 linear_127"  [label="(1, 14, 14, 1536)", style=solid];
-"2342 _param_constant342" -> "2344 linear_127"  [label="(384, 1536)", style=solid];
-"2343 _param_constant343" -> "2344 linear_127"  [label="(384,)", style=solid];
-"2344 linear_127" -> "2345 dropout_83"  [label="(1, 14, 14, 384)", style=solid];
-"2345 dropout_83" -> "2348 layer_norm_44"  [label="(1, 14, 14, 384)", style=solid];
-"2346 _param_constant344" -> "2348 layer_norm_44"  [label="(384,)", style=solid];
-"2347 _param_constant345" -> "2348 layer_norm_44"  [label="(384,)", style=solid];
-"2348 layer_norm_44" -> "2349 add_72"  [label="(1, 14, 14, 384)", style=solid];
-"2349 add_72" -> "2366 pad_23"  [label="(1, 14, 14, 384)", style=solid];
-"2349 add_72" -> "2479 add_75"  [label="(1, 14, 14, 384)", style=solid];
-"2350 _tensor_constant132" -> "2353 linear_128"  [label="(1, 15, 15, 2)", style=solid];
-"2351 _param_constant346" -> "2353 linear_128"  [label="(512, 2)", style=solid];
-"2352 _param_constant347" -> "2353 linear_128"  [label="(512,)", style=solid];
-"2353 linear_128" -> "2354 relu__21"  [label="(1, 15, 15, 512)", style=solid];
-"2354 relu__21" -> "2356 linear_129"  [label="(1, 15, 15, 512)", style=solid];
-"2355 _param_constant348" -> "2356 linear_129"  [label="(12, 512)", style=solid];
-"2356 linear_129" -> "2357 view_114"  [label="(1, 15, 15, 12)", style=solid];
-"2357 view_114" -> "2359 index_21"  [label="(225, 12)", style=solid];
-"2358 _tensor_constant133" -> "2359 index_21"  [label="(4096,)", style=solid];
-"2359 index_21" -> "2360 view_115"  [label="(4096, 12)", style=solid];
-"2360 view_115" -> "2361 permute_95"  [label="(64, 64, 12)", style=solid];
-"2361 permute_95" -> "2362 contiguous_40"  [label="(12, 64, 64)", style=solid];
-"2362 contiguous_40" -> "2363 unsqueeze_61"  [label="(12, 64, 64)", style=solid];
-"2363 unsqueeze_61" -> "2364 sigmoid_21"  [label="(1, 12, 64, 64)", style=solid];
-"2364 sigmoid_21" -> "2365 mul_42"  [label="(1, 12, 64, 64)", style=solid];
-"2365 mul_42" -> "2396 add_73"  [label="(1, 12, 64, 64)", style=solid];
-"2366 pad_23" -> "2367 roll_20"  [label="(1, 16, 16, 384)", style=solid];
-"2367 roll_20" -> "2368 view_116"  [label="(1, 16, 16, 384)", style=solid];
-"2368 view_116" -> "2369 permute_96"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2369 permute_96" -> "2370 reshape_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2370 reshape_94" -> "2376 linear_130"  [label="(4, 64, 384)", style=solid];
-"2370 reshape_94" -> "2397 new_zeros_10"  [label="(4, 64, 384)", style=solid];
-"2371 _param_constant349" -> "2372 clone_21"  [label="(1152,)", style=solid];
-"2372 clone_21" -> "2373 slice_306"  [label="(1152,)", style=solid];
-"2372 clone_21" -> "2376 linear_130"  [label="(1152,)", style=solid];
-"2373 slice_306" -> "2374 zero__21"  [label="(384,)", style=solid];
-"2375 _param_constant350" -> "2376 linear_130"  [label="(1152, 384)", style=solid];
-"2376 linear_130" -> "2377 reshape_95"  [label="(4, 64, 1152)", style=solid];
-"2377 reshape_95" -> "2378 permute_97"  [label="(4, 64, 3, 12, 32)", style=solid];
-"2378 permute_97" -> "2379 select_63"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2378 permute_97" -> "2380 select_64"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2378 permute_97" -> "2381 select_65"  [label="(3, 4, 12, 64, 32)", style=solid];
-"2379 select_63" -> "2382 linalg_vector_norm_42"  [label="(4, 12, 64, 32)", style=solid];
-"2379 select_63" -> "2384 expand_as_42"  [label="(4, 12, 64, 32)", style=solid];
-"2379 select_63" -> "2385 div_42"  [label="(4, 12, 64, 32)", style=solid];
-"2380 select_64" -> "2386 linalg_vector_norm_43"  [label="(4, 12, 64, 32)", style=solid];
-"2380 select_64" -> "2388 expand_as_43"  [label="(4, 12, 64, 32)", style=solid];
-"2380 select_64" -> "2389 div_43"  [label="(4, 12, 64, 32)", style=solid];
-"2381 select_65" -> "2460 matmul_43"  [label="(4, 12, 64, 32)", style=solid];
-"2382 linalg_vector_norm_42" -> "2383 clamp_min_42"  [label="(4, 12, 64, 1)", style=solid];
-"2383 clamp_min_42" -> "2384 expand_as_42"  [label="(4, 12, 64, 1)", style=solid];
-"2384 expand_as_42" -> "2385 div_42"  [label="(4, 12, 64, 32)", style=solid];
-"2385 div_42" -> "2391 matmul_42"  [label="(4, 12, 64, 32)", style=solid];
-"2386 linalg_vector_norm_43" -> "2387 clamp_min_43"  [label="(4, 12, 64, 1)", style=solid];
-"2387 clamp_min_43" -> "2388 expand_as_43"  [label="(4, 12, 64, 1)", style=solid];
-"2388 expand_as_43" -> "2389 div_43"  [label="(4, 12, 64, 32)", style=solid];
-"2389 div_43" -> "2390 transpose_42"  [label="(4, 12, 64, 32)", style=solid];
-"2390 transpose_42" -> "2391 matmul_42"  [label="(4, 12, 32, 64)", style=solid];
-"2391 matmul_42" -> "2395 mul_43"  [label="(4, 12, 64, 64)", style=solid];
-"2392 _param_constant351" -> "2393 clamp_21"  [label="(12, 1, 1)", style=solid];
-"2393 clamp_21" -> "2394 exp_21"  [label="(12, 1, 1)", style=solid];
-"2394 exp_21" -> "2395 mul_43"  [label="(12, 1, 1)", style=solid];
-"2395 mul_43" -> "2396 add_73"  [label="(4, 12, 64, 64)", style=solid];
-"2396 add_73" -> "2453 view_118"  [label="(4, 12, 64, 64)", style=solid];
-"2397 new_zeros_10" -> "2400 slice_307"  [label="(16, 16)", style=solid];
-"2397 new_zeros_10" -> "2405 slice_309"  [label="(16, 16)", style=solid];
-"2397 new_zeros_10" -> "2410 slice_311"  [label="(16, 16)", style=solid];
-"2397 new_zeros_10" -> "2415 slice_313"  [label="(16, 16)", style=solid];
-"2397 new_zeros_10" -> "2420 slice_315"  [label="(16, 16)", style=solid];
-"2397 new_zeros_10" -> "2425 slice_317"  [label="(16, 16)", style=solid];
-"2397 new_zeros_10" -> "2430 slice_319"  [label="(16, 16)", style=solid];
-"2397 new_zeros_10" -> "2435 slice_321"  [label="(16, 16)", style=solid];
-"2397 new_zeros_10" -> "2440 slice_323"  [label="(16, 16)", style=solid];
-"2397 new_zeros_10" -> "2443 view_117"  [label="(16, 16)", style=solid];
-"2398 _tensor_constant134" -> "2399 lift_fresh_copy_90"  [label="()", style=solid];
-"2399 lift_fresh_copy_90" -> "2402 fill__90"  [label="()", style=solid];
-"2400 slice_307" -> "2401 slice_308"  [label="(8, 16)", style=solid];
-"2401 slice_308" -> "2402 fill__90"  [label="(8, 8)", style=solid];
-"2403 _tensor_constant135" -> "2404 lift_fresh_copy_91"  [label="()", style=solid];
-"2404 lift_fresh_copy_91" -> "2407 fill__91"  [label="()", style=solid];
-"2405 slice_309" -> "2406 slice_310"  [label="(8, 16)", style=solid];
-"2406 slice_310" -> "2407 fill__91"  [label="(8, 4)", style=solid];
-"2408 _tensor_constant136" -> "2409 lift_fresh_copy_92"  [label="()", style=solid];
-"2409 lift_fresh_copy_92" -> "2412 fill__92"  [label="()", style=solid];
-"2410 slice_311" -> "2411 slice_312"  [label="(8, 16)", style=solid];
-"2411 slice_312" -> "2412 fill__92"  [label="(8, 4)", style=solid];
-"2413 _tensor_constant137" -> "2414 lift_fresh_copy_93"  [label="()", style=solid];
-"2414 lift_fresh_copy_93" -> "2417 fill__93"  [label="()", style=solid];
-"2415 slice_313" -> "2416 slice_314"  [label="(4, 16)", style=solid];
-"2416 slice_314" -> "2417 fill__93"  [label="(4, 8)", style=solid];
-"2418 _tensor_constant138" -> "2419 lift_fresh_copy_94"  [label="()", style=solid];
-"2419 lift_fresh_copy_94" -> "2422 fill__94"  [label="()", style=solid];
-"2420 slice_315" -> "2421 slice_316"  [label="(4, 16)", style=solid];
-"2421 slice_316" -> "2422 fill__94"  [label="(4, 4)", style=solid];
-"2423 _tensor_constant139" -> "2424 lift_fresh_copy_95"  [label="()", style=solid];
-"2424 lift_fresh_copy_95" -> "2427 fill__95"  [label="()", style=solid];
-"2425 slice_317" -> "2426 slice_318"  [label="(4, 16)", style=solid];
-"2426 slice_318" -> "2427 fill__95"  [label="(4, 4)", style=solid];
-"2428 _tensor_constant140" -> "2429 lift_fresh_copy_96"  [label="()", style=solid];
-"2429 lift_fresh_copy_96" -> "2432 fill__96"  [label="()", style=solid];
-"2430 slice_319" -> "2431 slice_320"  [label="(4, 16)", style=solid];
-"2431 slice_320" -> "2432 fill__96"  [label="(4, 8)", style=solid];
-"2433 _tensor_constant141" -> "2434 lift_fresh_copy_97"  [label="()", style=solid];
-"2434 lift_fresh_copy_97" -> "2437 fill__97"  [label="()", style=solid];
-"2435 slice_321" -> "2436 slice_322"  [label="(4, 16)", style=solid];
-"2436 slice_322" -> "2437 fill__97"  [label="(4, 4)", style=solid];
-"2438 _tensor_constant142" -> "2439 lift_fresh_copy_98"  [label="()", style=solid];
-"2439 lift_fresh_copy_98" -> "2442 fill__98"  [label="()", style=solid];
-"2440 slice_323" -> "2441 slice_324"  [label="(4, 16)", style=solid];
-"2441 slice_324" -> "2442 fill__98"  [label="(4, 4)", style=solid];
-"2443 view_117" -> "2444 permute_98"  [label="(2, 8, 2, 8)", style=solid];
-"2444 permute_98" -> "2445 reshape_96"  [label="(2, 2, 8, 8)", style=solid];
-"2445 reshape_96" -> "2446 unsqueeze_62"  [label="(4, 64)", style=solid];
-"2445 reshape_96" -> "2447 unsqueeze_63"  [label="(4, 64)", style=solid];
-"2446 unsqueeze_62" -> "2448 sub_10"  [label="(4, 1, 64)", style=solid];
-"2447 unsqueeze_63" -> "2448 sub_10"  [label="(4, 64, 1)", style=solid];
-"2448 sub_10" -> "2449 ne_10"  [label="(4, 64, 64)", style=solid];
-"2448 sub_10" -> "2450 masked_fill_20"  [label="(4, 64, 64)", style=solid];
-"2448 sub_10" -> "2451 eq_10"  [label="(4, 64, 64)", style=solid];
-"2449 ne_10" -> "2450 masked_fill_20"  [label="(4, 64, 64)", style=solid];
-"2450 masked_fill_20" -> "2452 masked_fill_21"  [label="(4, 64, 64)", style=solid];
-"2451 eq_10" -> "2452 masked_fill_21"  [label="(4, 64, 64)", style=solid];
-"2452 masked_fill_21" -> "2454 unsqueeze_64"  [label="(4, 64, 64)", style=solid];
-"2453 view_118" -> "2456 add_74"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2454 unsqueeze_64" -> "2455 unsqueeze_65"  [label="(4, 1, 64, 64)", style=solid];
-"2455 unsqueeze_65" -> "2456 add_74"  [label="(1, 4, 1, 64, 64)", style=solid];
-"2456 add_74" -> "2457 view_119"  [label="(1, 4, 12, 64, 64)", style=solid];
-"2457 view_119" -> "2458 softmax_21"  [label="(4, 12, 64, 64)", style=solid];
-"2458 softmax_21" -> "2459 dropout_84"  [label="(4, 12, 64, 64)", style=solid];
-"2459 dropout_84" -> "2460 matmul_43"  [label="(4, 12, 64, 64)", style=solid];
-"2460 matmul_43" -> "2461 transpose_43"  [label="(4, 12, 64, 32)", style=solid];
-"2461 transpose_43" -> "2462 reshape_97"  [label="(4, 64, 12, 32)", style=solid];
-"2462 reshape_97" -> "2465 linear_131"  [label="(4, 64, 384)", style=solid];
-"2463 _param_constant352" -> "2465 linear_131"  [label="(384, 384)", style=solid];
-"2464 _param_constant353" -> "2465 linear_131"  [label="(384,)", style=solid];
-"2465 linear_131" -> "2466 dropout_85"  [label="(4, 64, 384)", style=solid];
-"2466 dropout_85" -> "2467 view_120"  [label="(4, 64, 384)", style=solid];
-"2467 view_120" -> "2468 permute_99"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
-"2468 permute_99" -> "2469 reshape_98"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
-"2469 reshape_98" -> "2470 roll_21"  [label="(1, 16, 16, 384)", style=solid];
-"2470 roll_21" -> "2471 slice_325"  [label="(1, 16, 16, 384)", style=solid];
-"2471 slice_325" -> "2472 slice_326"  [label="(1, 16, 16, 384)", style=solid];
-"2472 slice_326" -> "2473 slice_327"  [label="(1, 14, 16, 384)", style=solid];
-"2473 slice_327" -> "2474 slice_328"  [label="(1, 14, 14, 384)", style=solid];
-"2474 slice_328" -> "2475 contiguous_41"  [label="(1, 14, 14, 384)", style=solid];
-"2475 contiguous_41" -> "2478 layer_norm_45"  [label="(1, 14, 14, 384)", style=solid];
-"2476 _param_constant354" -> "2478 layer_norm_45"  [label="(384,)", style=solid];
-"2477 _param_constant355" -> "2478 layer_norm_45"  [label="(384,)", style=solid];
-"2478 layer_norm_45" -> "2479 add_75"  [label="(1, 14, 14, 384)", style=solid];
-"2479 add_75" -> "2482 linear_132"  [label="(1, 14, 14, 384)", style=solid];
-"2479 add_75" -> "2492 add_76"  [label="(1, 14, 14, 384)", style=solid];
-"2480 _param_constant356" -> "2482 linear_132"  [label="(1536, 384)", style=solid];
-"2481 _param_constant357" -> "2482 linear_132"  [label="(1536,)", style=solid];
-"2482 linear_132" -> "2483 gelu_21"  [label="(1, 14, 14, 1536)", style=solid];
-"2483 gelu_21" -> "2484 dropout_86"  [label="(1, 14, 14, 1536)", style=solid];
-"2484 dropout_86" -> "2487 linear_133"  [label="(1, 14, 14, 1536)", style=solid];
-"2485 _param_constant358" -> "2487 linear_133"  [label="(384, 1536)", style=solid];
-"2486 _param_constant359" -> "2487 linear_133"  [label="(384,)", style=solid];
-"2487 linear_133" -> "2488 dropout_87"  [label="(1, 14, 14, 384)", style=solid];
-"2488 dropout_87" -> "2491 layer_norm_46"  [label="(1, 14, 14, 384)", style=solid];
-"2489 _param_constant360" -> "2491 layer_norm_46"  [label="(384,)", style=solid];
-"2490 _param_constant361" -> "2491 layer_norm_46"  [label="(384,)", style=solid];
-"2491 layer_norm_46" -> "2492 add_76"  [label="(1, 14, 14, 384)", style=solid];
-"2492 add_76" -> "2493 pad_24"  [label="(1, 14, 14, 384)", style=solid];
-"2493 pad_24" -> "2494 slice_329"  [label="(1, 14, 14, 384)", style=solid];
-"2493 pad_24" -> "2497 slice_332"  [label="(1, 14, 14, 384)", style=solid];
-"2493 pad_24" -> "2500 slice_335"  [label="(1, 14, 14, 384)", style=solid];
-"2493 pad_24" -> "2503 slice_338"  [label="(1, 14, 14, 384)", style=solid];
-"2494 slice_329" -> "2495 slice_330"  [label="(1, 7, 14, 384)", style=solid];
-"2495 slice_330" -> "2496 slice_331"  [label="(1, 7, 7, 384)", style=solid];
-"2496 slice_331" -> "2506 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2497 slice_332" -> "2498 slice_333"  [label="(1, 7, 14, 384)", style=solid];
-"2498 slice_333" -> "2499 slice_334"  [label="(1, 7, 7, 384)", style=solid];
-"2499 slice_334" -> "2506 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2500 slice_335" -> "2501 slice_336"  [label="(1, 7, 14, 384)", style=solid];
-"2501 slice_336" -> "2502 slice_337"  [label="(1, 7, 7, 384)", style=solid];
-"2502 slice_337" -> "2506 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2503 slice_338" -> "2504 slice_339"  [label="(1, 7, 14, 384)", style=solid];
-"2504 slice_339" -> "2505 slice_340"  [label="(1, 7, 7, 384)", style=solid];
-"2505 slice_340" -> "2506 cat_2"  [label="(1, 7, 7, 384)", style=solid];
-"2506 cat_2" -> "2508 linear_134"  [label="(1, 7, 7, 1536)", style=solid];
-"2507 _param_constant362" -> "2508 linear_134"  [label="(768, 1536)", style=solid];
-"2508 linear_134" -> "2511 layer_norm_47"  [label="(1, 7, 7, 768)", style=solid];
-"2509 _param_constant363" -> "2511 layer_norm_47"  [label="(768,)", style=solid];
-"2510 _param_constant364" -> "2511 layer_norm_47"  [label="(768,)", style=solid];
-"2511 layer_norm_47" -> "2528 pad_25"  [label="(1, 7, 7, 768)", style=solid];
-"2511 layer_norm_47" -> "2578 add_78"  [label="(1, 7, 7, 768)", style=solid];
-"2512 _tensor_constant143" -> "2515 linear_135"  [label="(1, 15, 15, 2)", style=solid];
-"2513 _param_constant365" -> "2515 linear_135"  [label="(512, 2)", style=solid];
-"2514 _param_constant366" -> "2515 linear_135"  [label="(512,)", style=solid];
-"2515 linear_135" -> "2516 relu__22"  [label="(1, 15, 15, 512)", style=solid];
-"2516 relu__22" -> "2518 linear_136"  [label="(1, 15, 15, 512)", style=solid];
-"2517 _param_constant367" -> "2518 linear_136"  [label="(24, 512)", style=solid];
-"2518 linear_136" -> "2519 view_121"  [label="(1, 15, 15, 24)", style=solid];
-"2519 view_121" -> "2521 index_22"  [label="(225, 24)", style=solid];
-"2520 _tensor_constant144" -> "2521 index_22"  [label="(4096,)", style=solid];
-"2521 index_22" -> "2522 view_122"  [label="(4096, 24)", style=solid];
-"2522 view_122" -> "2523 permute_100"  [label="(64, 64, 24)", style=solid];
-"2523 permute_100" -> "2524 contiguous_42"  [label="(24, 64, 64)", style=solid];
-"2524 contiguous_42" -> "2525 unsqueeze_66"  [label="(24, 64, 64)", style=solid];
-"2525 unsqueeze_66" -> "2526 sigmoid_22"  [label="(1, 24, 64, 64)", style=solid];
-"2526 sigmoid_22" -> "2527 mul_44"  [label="(1, 24, 64, 64)", style=solid];
-"2527 mul_44" -> "2557 add_77"  [label="(1, 24, 64, 64)", style=solid];
-"2528 pad_25" -> "2529 view_123"  [label="(1, 8, 8, 768)", style=solid];
-"2529 view_123" -> "2530 permute_101"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2530 permute_101" -> "2531 reshape_99"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2531 reshape_99" -> "2537 linear_137"  [label="(1, 64, 768)", style=solid];
-"2532 _param_constant368" -> "2533 clone_22"  [label="(2304,)", style=solid];
-"2533 clone_22" -> "2534 slice_341"  [label="(2304,)", style=solid];
-"2533 clone_22" -> "2537 linear_137"  [label="(2304,)", style=solid];
-"2534 slice_341" -> "2535 zero__22"  [label="(768,)", style=solid];
-"2536 _param_constant369" -> "2537 linear_137"  [label="(2304, 768)", style=solid];
-"2537 linear_137" -> "2538 reshape_100"  [label="(1, 64, 2304)", style=solid];
-"2538 reshape_100" -> "2539 permute_102"  [label="(1, 64, 3, 24, 32)", style=solid];
-"2539 permute_102" -> "2540 select_66"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2539 permute_102" -> "2541 select_67"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2539 permute_102" -> "2542 select_68"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2540 select_66" -> "2543 linalg_vector_norm_44"  [label="(1, 24, 64, 32)", style=solid];
-"2540 select_66" -> "2545 expand_as_44"  [label="(1, 24, 64, 32)", style=solid];
-"2540 select_66" -> "2546 div_44"  [label="(1, 24, 64, 32)", style=solid];
-"2541 select_67" -> "2547 linalg_vector_norm_45"  [label="(1, 24, 64, 32)", style=solid];
-"2541 select_67" -> "2549 expand_as_45"  [label="(1, 24, 64, 32)", style=solid];
-"2541 select_67" -> "2550 div_45"  [label="(1, 24, 64, 32)", style=solid];
-"2542 select_68" -> "2560 matmul_45"  [label="(1, 24, 64, 32)", style=solid];
-"2543 linalg_vector_norm_44" -> "2544 clamp_min_44"  [label="(1, 24, 64, 1)", style=solid];
-"2544 clamp_min_44" -> "2545 expand_as_44"  [label="(1, 24, 64, 1)", style=solid];
-"2545 expand_as_44" -> "2546 div_44"  [label="(1, 24, 64, 32)", style=solid];
-"2546 div_44" -> "2552 matmul_44"  [label="(1, 24, 64, 32)", style=solid];
-"2547 linalg_vector_norm_45" -> "2548 clamp_min_45"  [label="(1, 24, 64, 1)", style=solid];
-"2548 clamp_min_45" -> "2549 expand_as_45"  [label="(1, 24, 64, 1)", style=solid];
-"2549 expand_as_45" -> "2550 div_45"  [label="(1, 24, 64, 32)", style=solid];
-"2550 div_45" -> "2551 transpose_44"  [label="(1, 24, 64, 32)", style=solid];
-"2551 transpose_44" -> "2552 matmul_44"  [label="(1, 24, 32, 64)", style=solid];
-"2552 matmul_44" -> "2556 mul_45"  [label="(1, 24, 64, 64)", style=solid];
-"2553 _param_constant370" -> "2554 clamp_22"  [label="(24, 1, 1)", style=solid];
-"2554 clamp_22" -> "2555 exp_22"  [label="(24, 1, 1)", style=solid];
-"2555 exp_22" -> "2556 mul_45"  [label="(24, 1, 1)", style=solid];
-"2556 mul_45" -> "2557 add_77"  [label="(1, 24, 64, 64)", style=solid];
-"2557 add_77" -> "2558 softmax_22"  [label="(1, 24, 64, 64)", style=solid];
-"2558 softmax_22" -> "2559 dropout_88"  [label="(1, 24, 64, 64)", style=solid];
-"2559 dropout_88" -> "2560 matmul_45"  [label="(1, 24, 64, 64)", style=solid];
-"2560 matmul_45" -> "2561 transpose_45"  [label="(1, 24, 64, 32)", style=solid];
-"2561 transpose_45" -> "2562 reshape_101"  [label="(1, 64, 24, 32)", style=solid];
-"2562 reshape_101" -> "2565 linear_138"  [label="(1, 64, 768)", style=solid];
-"2563 _param_constant371" -> "2565 linear_138"  [label="(768, 768)", style=solid];
-"2564 _param_constant372" -> "2565 linear_138"  [label="(768,)", style=solid];
-"2565 linear_138" -> "2566 dropout_89"  [label="(1, 64, 768)", style=solid];
-"2566 dropout_89" -> "2567 view_124"  [label="(1, 64, 768)", style=solid];
-"2567 view_124" -> "2568 permute_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2568 permute_103" -> "2569 reshape_102"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2569 reshape_102" -> "2570 slice_342"  [label="(1, 8, 8, 768)", style=solid];
-"2570 slice_342" -> "2571 slice_343"  [label="(1, 8, 8, 768)", style=solid];
-"2571 slice_343" -> "2572 slice_344"  [label="(1, 7, 8, 768)", style=solid];
-"2572 slice_344" -> "2573 slice_345"  [label="(1, 7, 7, 768)", style=solid];
-"2573 slice_345" -> "2574 contiguous_43"  [label="(1, 7, 7, 768)", style=solid];
-"2574 contiguous_43" -> "2577 layer_norm_48"  [label="(1, 7, 7, 768)", style=solid];
-"2575 _param_constant373" -> "2577 layer_norm_48"  [label="(768,)", style=solid];
-"2576 _param_constant374" -> "2577 layer_norm_48"  [label="(768,)", style=solid];
-"2577 layer_norm_48" -> "2578 add_78"  [label="(1, 7, 7, 768)", style=solid];
-"2578 add_78" -> "2581 linear_139"  [label="(1, 7, 7, 768)", style=solid];
-"2578 add_78" -> "2591 add_79"  [label="(1, 7, 7, 768)", style=solid];
-"2579 _param_constant375" -> "2581 linear_139"  [label="(3072, 768)", style=solid];
-"2580 _param_constant376" -> "2581 linear_139"  [label="(3072,)", style=solid];
-"2581 linear_139" -> "2582 gelu_22"  [label="(1, 7, 7, 3072)", style=solid];
-"2582 gelu_22" -> "2583 dropout_90"  [label="(1, 7, 7, 3072)", style=solid];
-"2583 dropout_90" -> "2586 linear_140"  [label="(1, 7, 7, 3072)", style=solid];
-"2584 _param_constant377" -> "2586 linear_140"  [label="(768, 3072)", style=solid];
-"2585 _param_constant378" -> "2586 linear_140"  [label="(768,)", style=solid];
-"2586 linear_140" -> "2587 dropout_91"  [label="(1, 7, 7, 768)", style=solid];
-"2587 dropout_91" -> "2590 layer_norm_49"  [label="(1, 7, 7, 768)", style=solid];
-"2588 _param_constant379" -> "2590 layer_norm_49"  [label="(768,)", style=solid];
-"2589 _param_constant380" -> "2590 layer_norm_49"  [label="(768,)", style=solid];
-"2590 layer_norm_49" -> "2591 add_79"  [label="(1, 7, 7, 768)", style=solid];
-"2591 add_79" -> "2608 pad_26"  [label="(1, 7, 7, 768)", style=solid];
-"2591 add_79" -> "2658 add_81"  [label="(1, 7, 7, 768)", style=solid];
-"2592 _tensor_constant145" -> "2595 linear_141"  [label="(1, 15, 15, 2)", style=solid];
-"2593 _param_constant381" -> "2595 linear_141"  [label="(512, 2)", style=solid];
-"2594 _param_constant382" -> "2595 linear_141"  [label="(512,)", style=solid];
-"2595 linear_141" -> "2596 relu__23"  [label="(1, 15, 15, 512)", style=solid];
-"2596 relu__23" -> "2598 linear_142"  [label="(1, 15, 15, 512)", style=solid];
-"2597 _param_constant383" -> "2598 linear_142"  [label="(24, 512)", style=solid];
-"2598 linear_142" -> "2599 view_125"  [label="(1, 15, 15, 24)", style=solid];
-"2599 view_125" -> "2601 index_23"  [label="(225, 24)", style=solid];
-"2600 _tensor_constant146" -> "2601 index_23"  [label="(4096,)", style=solid];
-"2601 index_23" -> "2602 view_126"  [label="(4096, 24)", style=solid];
-"2602 view_126" -> "2603 permute_104"  [label="(64, 64, 24)", style=solid];
-"2603 permute_104" -> "2604 contiguous_44"  [label="(24, 64, 64)", style=solid];
-"2604 contiguous_44" -> "2605 unsqueeze_67"  [label="(24, 64, 64)", style=solid];
-"2605 unsqueeze_67" -> "2606 sigmoid_23"  [label="(1, 24, 64, 64)", style=solid];
-"2606 sigmoid_23" -> "2607 mul_46"  [label="(1, 24, 64, 64)", style=solid];
-"2607 mul_46" -> "2637 add_80"  [label="(1, 24, 64, 64)", style=solid];
-"2608 pad_26" -> "2609 view_127"  [label="(1, 8, 8, 768)", style=solid];
-"2609 view_127" -> "2610 permute_105"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2610 permute_105" -> "2611 reshape_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2611 reshape_103" -> "2617 linear_143"  [label="(1, 64, 768)", style=solid];
-"2612 _param_constant384" -> "2613 clone_23"  [label="(2304,)", style=solid];
-"2613 clone_23" -> "2614 slice_346"  [label="(2304,)", style=solid];
-"2613 clone_23" -> "2617 linear_143"  [label="(2304,)", style=solid];
-"2614 slice_346" -> "2615 zero__23"  [label="(768,)", style=solid];
-"2616 _param_constant385" -> "2617 linear_143"  [label="(2304, 768)", style=solid];
-"2617 linear_143" -> "2618 reshape_104"  [label="(1, 64, 2304)", style=solid];
-"2618 reshape_104" -> "2619 permute_106"  [label="(1, 64, 3, 24, 32)", style=solid];
-"2619 permute_106" -> "2620 select_69"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2619 permute_106" -> "2621 select_70"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2619 permute_106" -> "2622 select_71"  [label="(3, 1, 24, 64, 32)", style=solid];
-"2620 select_69" -> "2623 linalg_vector_norm_46"  [label="(1, 24, 64, 32)", style=solid];
-"2620 select_69" -> "2625 expand_as_46"  [label="(1, 24, 64, 32)", style=solid];
-"2620 select_69" -> "2626 div_46"  [label="(1, 24, 64, 32)", style=solid];
-"2621 select_70" -> "2627 linalg_vector_norm_47"  [label="(1, 24, 64, 32)", style=solid];
-"2621 select_70" -> "2629 expand_as_47"  [label="(1, 24, 64, 32)", style=solid];
-"2621 select_70" -> "2630 div_47"  [label="(1, 24, 64, 32)", style=solid];
-"2622 select_71" -> "2640 matmul_47"  [label="(1, 24, 64, 32)", style=solid];
-"2623 linalg_vector_norm_46" -> "2624 clamp_min_46"  [label="(1, 24, 64, 1)", style=solid];
-"2624 clamp_min_46" -> "2625 expand_as_46"  [label="(1, 24, 64, 1)", style=solid];
-"2625 expand_as_46" -> "2626 div_46"  [label="(1, 24, 64, 32)", style=solid];
-"2626 div_46" -> "2632 matmul_46"  [label="(1, 24, 64, 32)", style=solid];
-"2627 linalg_vector_norm_47" -> "2628 clamp_min_47"  [label="(1, 24, 64, 1)", style=solid];
-"2628 clamp_min_47" -> "2629 expand_as_47"  [label="(1, 24, 64, 1)", style=solid];
-"2629 expand_as_47" -> "2630 div_47"  [label="(1, 24, 64, 32)", style=solid];
-"2630 div_47" -> "2631 transpose_46"  [label="(1, 24, 64, 32)", style=solid];
-"2631 transpose_46" -> "2632 matmul_46"  [label="(1, 24, 32, 64)", style=solid];
-"2632 matmul_46" -> "2636 mul_47"  [label="(1, 24, 64, 64)", style=solid];
-"2633 _param_constant386" -> "2634 clamp_23"  [label="(24, 1, 1)", style=solid];
-"2634 clamp_23" -> "2635 exp_23"  [label="(24, 1, 1)", style=solid];
-"2635 exp_23" -> "2636 mul_47"  [label="(24, 1, 1)", style=solid];
-"2636 mul_47" -> "2637 add_80"  [label="(1, 24, 64, 64)", style=solid];
-"2637 add_80" -> "2638 softmax_23"  [label="(1, 24, 64, 64)", style=solid];
-"2638 softmax_23" -> "2639 dropout_92"  [label="(1, 24, 64, 64)", style=solid];
-"2639 dropout_92" -> "2640 matmul_47"  [label="(1, 24, 64, 64)", style=solid];
-"2640 matmul_47" -> "2641 transpose_47"  [label="(1, 24, 64, 32)", style=solid];
-"2641 transpose_47" -> "2642 reshape_105"  [label="(1, 64, 24, 32)", style=solid];
-"2642 reshape_105" -> "2645 linear_144"  [label="(1, 64, 768)", style=solid];
-"2643 _param_constant387" -> "2645 linear_144"  [label="(768, 768)", style=solid];
-"2644 _param_constant388" -> "2645 linear_144"  [label="(768,)", style=solid];
-"2645 linear_144" -> "2646 dropout_93"  [label="(1, 64, 768)", style=solid];
-"2646 dropout_93" -> "2647 view_128"  [label="(1, 64, 768)", style=solid];
-"2647 view_128" -> "2648 permute_107"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
-"2648 permute_107" -> "2649 reshape_106"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
-"2649 reshape_106" -> "2650 slice_347"  [label="(1, 8, 8, 768)", style=solid];
-"2650 slice_347" -> "2651 slice_348"  [label="(1, 8, 8, 768)", style=solid];
-"2651 slice_348" -> "2652 slice_349"  [label="(1, 7, 8, 768)", style=solid];
-"2652 slice_349" -> "2653 slice_350"  [label="(1, 7, 7, 768)", style=solid];
-"2653 slice_350" -> "2654 contiguous_45"  [label="(1, 7, 7, 768)", style=solid];
-"2654 contiguous_45" -> "2657 layer_norm_50"  [label="(1, 7, 7, 768)", style=solid];
-"2655 _param_constant389" -> "2657 layer_norm_50"  [label="(768,)", style=solid];
-"2656 _param_constant390" -> "2657 layer_norm_50"  [label="(768,)", style=solid];
-"2657 layer_norm_50" -> "2658 add_81"  [label="(1, 7, 7, 768)", style=solid];
-"2658 add_81" -> "2661 linear_145"  [label="(1, 7, 7, 768)", style=solid];
-"2658 add_81" -> "2671 add_82"  [label="(1, 7, 7, 768)", style=solid];
-"2659 _param_constant391" -> "2661 linear_145"  [label="(3072, 768)", style=solid];
-"2660 _param_constant392" -> "2661 linear_145"  [label="(3072,)", style=solid];
-"2661 linear_145" -> "2662 gelu_23"  [label="(1, 7, 7, 3072)", style=solid];
-"2662 gelu_23" -> "2663 dropout_94"  [label="(1, 7, 7, 3072)", style=solid];
-"2663 dropout_94" -> "2666 linear_146"  [label="(1, 7, 7, 3072)", style=solid];
-"2664 _param_constant393" -> "2666 linear_146"  [label="(768, 3072)", style=solid];
-"2665 _param_constant394" -> "2666 linear_146"  [label="(768,)", style=solid];
-"2666 linear_146" -> "2667 dropout_95"  [label="(1, 7, 7, 768)", style=solid];
-"2667 dropout_95" -> "2670 layer_norm_51"  [label="(1, 7, 7, 768)", style=solid];
-"2668 _param_constant395" -> "2670 layer_norm_51"  [label="(768,)", style=solid];
-"2669 _param_constant396" -> "2670 layer_norm_51"  [label="(768,)", style=solid];
-"2670 layer_norm_51" -> "2671 add_82"  [label="(1, 7, 7, 768)", style=solid];
-"2671 add_82" -> "2674 layer_norm_52"  [label="(1, 7, 7, 768)", style=solid];
-"2672 _param_constant397" -> "2674 layer_norm_52"  [label="(768,)", style=solid];
-"2673 _param_constant398" -> "2674 layer_norm_52"  [label="(768,)", style=solid];
-"2674 layer_norm_52" -> "2675 permute_108"  [label="(1, 7, 7, 768)", style=solid];
-"2675 permute_108" -> "2676 adaptive_avg_pool2d"  [label="(1, 768, 7, 7)", style=solid];
-"2676 adaptive_avg_pool2d" -> "2677 flatten"  [label="(1, 768, 1, 1)", style=solid];
-"2677 flatten" -> "2680 linear_147"  [label="(1, 768)", style=solid];
-"2678 _param_constant399" -> "2680 linear_147"  [label="(1000, 768)", style=solid];
-"2679 _param_constant400" -> "2680 linear_147"  [label="(1000,)", style=solid];
-"2680 linear_147" -> "2681 output"  [label="(1, 1000)", style=solid];
+"2681 output_1" [id=2681, type=output];
+"0 features_1_0_attn_qkv_weight" -> "570 linear_2"  [label="(288, 96)", style=solid];
+"1 features_1_0_attn_proj_weight" -> "595 linear_3"  [label="(96, 96)", style=solid];
+"2 features_1_0_attn_qkv_bias" -> "567 clone"  [label="(288,)", style=solid];
+"3 features_1_0_attn_proj_bias" -> "595 linear_3"  [label="(96,)", style=solid];
+"4 features_1_0_attn_logit_scale" -> "586 clamp"  [label="(3, 1, 1)", style=solid];
+"5 features_1_1_attn_qkv_weight" -> "630 linear_8"  [label="(288, 96)", style=solid];
+"6 features_1_1_attn_proj_weight" -> "707 linear_9"  [label="(96, 96)", style=solid];
+"7 features_1_1_attn_qkv_bias" -> "627 clone_1"  [label="(288,)", style=solid];
+"8 features_1_1_attn_proj_bias" -> "707 linear_9"  [label="(96,)", style=solid];
+"9 features_1_1_attn_logit_scale" -> "646 clamp_1"  [label="(3, 1, 1)", style=solid];
+"10 features_3_0_attn_qkv_weight" -> "758 linear_15"  [label="(576, 192)", style=solid];
+"11 features_3_0_attn_proj_weight" -> "783 linear_16"  [label="(192, 192)", style=solid];
+"12 features_3_0_attn_qkv_bias" -> "755 clone_2"  [label="(576,)", style=solid];
+"13 features_3_0_attn_proj_bias" -> "783 linear_16"  [label="(192,)", style=solid];
+"14 features_3_0_attn_logit_scale" -> "774 clamp_2"  [label="(6, 1, 1)", style=solid];
+"15 features_3_1_attn_qkv_weight" -> "821 linear_21"  [label="(576, 192)", style=solid];
+"16 features_3_1_attn_proj_weight" -> "898 linear_22"  [label="(192, 192)", style=solid];
+"17 features_3_1_attn_qkv_bias" -> "818 clone_3"  [label="(576,)", style=solid];
+"18 features_3_1_attn_proj_bias" -> "898 linear_22"  [label="(192,)", style=solid];
+"19 features_3_1_attn_logit_scale" -> "837 clamp_3"  [label="(6, 1, 1)", style=solid];
+"20 features_5_0_attn_qkv_weight" -> "952 linear_28"  [label="(1152, 384)", style=solid];
+"21 features_5_0_attn_proj_weight" -> "977 linear_29"  [label="(384, 384)", style=solid];
+"22 features_5_0_attn_qkv_bias" -> "949 clone_4"  [label="(1152,)", style=solid];
+"23 features_5_0_attn_proj_bias" -> "977 linear_29"  [label="(384,)", style=solid];
+"24 features_5_0_attn_logit_scale" -> "968 clamp_4"  [label="(12, 1, 1)", style=solid];
+"25 features_5_1_attn_qkv_weight" -> "1015 linear_34"  [label="(1152, 384)", style=solid];
+"26 features_5_1_attn_proj_weight" -> "1092 linear_35"  [label="(384, 384)", style=solid];
+"27 features_5_1_attn_qkv_bias" -> "1012 clone_5"  [label="(1152,)", style=solid];
+"28 features_5_1_attn_proj_bias" -> "1092 linear_35"  [label="(384,)", style=solid];
+"29 features_5_1_attn_logit_scale" -> "1031 clamp_5"  [label="(12, 1, 1)", style=solid];
+"30 features_5_2_attn_qkv_weight" -> "1130 linear_40"  [label="(1152, 384)", style=solid];
+"31 features_5_2_attn_proj_weight" -> "1155 linear_41"  [label="(384, 384)", style=solid];
+"32 features_5_2_attn_qkv_bias" -> "1127 clone_6"  [label="(1152,)", style=solid];
+"33 features_5_2_attn_proj_bias" -> "1155 linear_41"  [label="(384,)", style=solid];
+"34 features_5_2_attn_logit_scale" -> "1146 clamp_6"  [label="(12, 1, 1)", style=solid];
+"35 features_5_3_attn_qkv_weight" -> "1193 linear_46"  [label="(1152, 384)", style=solid];
+"36 features_5_3_attn_proj_weight" -> "1270 linear_47"  [label="(384, 384)", style=solid];
+"37 features_5_3_attn_qkv_bias" -> "1190 clone_7"  [label="(1152,)", style=solid];
+"38 features_5_3_attn_proj_bias" -> "1270 linear_47"  [label="(384,)", style=solid];
+"39 features_5_3_attn_logit_scale" -> "1209 clamp_7"  [label="(12, 1, 1)", style=solid];
+"40 features_5_4_attn_qkv_weight" -> "1308 linear_52"  [label="(1152, 384)", style=solid];
+"41 features_5_4_attn_proj_weight" -> "1333 linear_53"  [label="(384, 384)", style=solid];
+"42 features_5_4_attn_qkv_bias" -> "1305 clone_8"  [label="(1152,)", style=solid];
+"43 features_5_4_attn_proj_bias" -> "1333 linear_53"  [label="(384,)", style=solid];
+"44 features_5_4_attn_logit_scale" -> "1324 clamp_8"  [label="(12, 1, 1)", style=solid];
+"45 features_5_5_attn_qkv_weight" -> "1371 linear_58"  [label="(1152, 384)", style=solid];
+"46 features_5_5_attn_proj_weight" -> "1448 linear_59"  [label="(384, 384)", style=solid];
+"47 features_5_5_attn_qkv_bias" -> "1368 clone_9"  [label="(1152,)", style=solid];
+"48 features_5_5_attn_proj_bias" -> "1448 linear_59"  [label="(384,)", style=solid];
+"49 features_5_5_attn_logit_scale" -> "1387 clamp_9"  [label="(12, 1, 1)", style=solid];
+"50 features_5_6_attn_qkv_weight" -> "1486 linear_64"  [label="(1152, 384)", style=solid];
+"51 features_5_6_attn_proj_weight" -> "1511 linear_65"  [label="(384, 384)", style=solid];
+"52 features_5_6_attn_qkv_bias" -> "1483 clone_10"  [label="(1152,)", style=solid];
+"53 features_5_6_attn_proj_bias" -> "1511 linear_65"  [label="(384,)", style=solid];
+"54 features_5_6_attn_logit_scale" -> "1502 clamp_10"  [label="(12, 1, 1)", style=solid];
+"55 features_5_7_attn_qkv_weight" -> "1549 linear_70"  [label="(1152, 384)", style=solid];
+"56 features_5_7_attn_proj_weight" -> "1626 linear_71"  [label="(384, 384)", style=solid];
+"57 features_5_7_attn_qkv_bias" -> "1546 clone_11"  [label="(1152,)", style=solid];
+"58 features_5_7_attn_proj_bias" -> "1626 linear_71"  [label="(384,)", style=solid];
+"59 features_5_7_attn_logit_scale" -> "1565 clamp_11"  [label="(12, 1, 1)", style=solid];
+"60 features_5_8_attn_qkv_weight" -> "1664 linear_76"  [label="(1152, 384)", style=solid];
+"61 features_5_8_attn_proj_weight" -> "1689 linear_77"  [label="(384, 384)", style=solid];
+"62 features_5_8_attn_qkv_bias" -> "1661 clone_12"  [label="(1152,)", style=solid];
+"63 features_5_8_attn_proj_bias" -> "1689 linear_77"  [label="(384,)", style=solid];
+"64 features_5_8_attn_logit_scale" -> "1680 clamp_12"  [label="(12, 1, 1)", style=solid];
+"65 features_5_9_attn_qkv_weight" -> "1727 linear_82"  [label="(1152, 384)", style=solid];
+"66 features_5_9_attn_proj_weight" -> "1804 linear_83"  [label="(384, 384)", style=solid];
+"67 features_5_9_attn_qkv_bias" -> "1724 clone_13"  [label="(1152,)", style=solid];
+"68 features_5_9_attn_proj_bias" -> "1804 linear_83"  [label="(384,)", style=solid];
+"69 features_5_9_attn_logit_scale" -> "1743 clamp_13"  [label="(12, 1, 1)", style=solid];
+"70 features_5_10_attn_qkv_weight" -> "1842 linear_88"  [label="(1152, 384)", style=solid];
+"71 features_5_10_attn_proj_weight" -> "1867 linear_89"  [label="(384, 384)", style=solid];
+"72 features_5_10_attn_qkv_bias" -> "1839 clone_14"  [label="(1152,)", style=solid];
+"73 features_5_10_attn_proj_bias" -> "1867 linear_89"  [label="(384,)", style=solid];
+"74 features_5_10_attn_logit_scale" -> "1858 clamp_14"  [label="(12, 1, 1)", style=solid];
+"75 features_5_11_attn_qkv_weight" -> "1905 linear_94"  [label="(1152, 384)", style=solid];
+"76 features_5_11_attn_proj_weight" -> "1982 linear_95"  [label="(384, 384)", style=solid];
+"77 features_5_11_attn_qkv_bias" -> "1902 clone_15"  [label="(1152,)", style=solid];
+"78 features_5_11_attn_proj_bias" -> "1982 linear_95"  [label="(384,)", style=solid];
+"79 features_5_11_attn_logit_scale" -> "1921 clamp_15"  [label="(12, 1, 1)", style=solid];
+"80 features_5_12_attn_qkv_weight" -> "2020 linear_100"  [label="(1152, 384)", style=solid];
+"81 features_5_12_attn_proj_weight" -> "2045 linear_101"  [label="(384, 384)", style=solid];
+"82 features_5_12_attn_qkv_bias" -> "2017 clone_16"  [label="(1152,)", style=solid];
+"83 features_5_12_attn_proj_bias" -> "2045 linear_101"  [label="(384,)", style=solid];
+"84 features_5_12_attn_logit_scale" -> "2036 clamp_16"  [label="(12, 1, 1)", style=solid];
+"85 features_5_13_attn_qkv_weight" -> "2083 linear_106"  [label="(1152, 384)", style=solid];
+"86 features_5_13_attn_proj_weight" -> "2160 linear_107"  [label="(384, 384)", style=solid];
+"87 features_5_13_attn_qkv_bias" -> "2080 clone_17"  [label="(1152,)", style=solid];
+"88 features_5_13_attn_proj_bias" -> "2160 linear_107"  [label="(384,)", style=solid];
+"89 features_5_13_attn_logit_scale" -> "2099 clamp_17"  [label="(12, 1, 1)", style=solid];
+"90 features_5_14_attn_qkv_weight" -> "2198 linear_112"  [label="(1152, 384)", style=solid];
+"91 features_5_14_attn_proj_weight" -> "2223 linear_113"  [label="(384, 384)", style=solid];
+"92 features_5_14_attn_qkv_bias" -> "2195 clone_18"  [label="(1152,)", style=solid];
+"93 features_5_14_attn_proj_bias" -> "2223 linear_113"  [label="(384,)", style=solid];
+"94 features_5_14_attn_logit_scale" -> "2214 clamp_18"  [label="(12, 1, 1)", style=solid];
+"95 features_5_15_attn_qkv_weight" -> "2261 linear_118"  [label="(1152, 384)", style=solid];
+"96 features_5_15_attn_proj_weight" -> "2338 linear_119"  [label="(384, 384)", style=solid];
+"97 features_5_15_attn_qkv_bias" -> "2258 clone_19"  [label="(1152,)", style=solid];
+"98 features_5_15_attn_proj_bias" -> "2338 linear_119"  [label="(384,)", style=solid];
+"99 features_5_15_attn_logit_scale" -> "2277 clamp_19"  [label="(12, 1, 1)", style=solid];
+"100 features_5_16_attn_qkv_weight" -> "2376 linear_124"  [label="(1152, 384)", style=solid];
+"101 features_5_16_attn_proj_weight" -> "2401 linear_125"  [label="(384, 384)", style=solid];
+"102 features_5_16_attn_qkv_bias" -> "2373 clone_20"  [label="(1152,)", style=solid];
+"103 features_5_16_attn_proj_bias" -> "2401 linear_125"  [label="(384,)", style=solid];
+"104 features_5_16_attn_logit_scale" -> "2392 clamp_20"  [label="(12, 1, 1)", style=solid];
+"105 features_5_17_attn_qkv_weight" -> "2439 linear_130"  [label="(1152, 384)", style=solid];
+"106 features_5_17_attn_proj_weight" -> "2516 linear_131"  [label="(384, 384)", style=solid];
+"107 features_5_17_attn_qkv_bias" -> "2436 clone_21"  [label="(1152,)", style=solid];
+"108 features_5_17_attn_proj_bias" -> "2516 linear_131"  [label="(384,)", style=solid];
+"109 features_5_17_attn_logit_scale" -> "2455 clamp_21"  [label="(12, 1, 1)", style=solid];
+"110 features_7_0_attn_qkv_weight" -> "2570 linear_137"  [label="(2304, 768)", style=solid];
+"111 features_7_0_attn_proj_weight" -> "2595 linear_138"  [label="(768, 768)", style=solid];
+"112 features_7_0_attn_qkv_bias" -> "2567 clone_22"  [label="(2304,)", style=solid];
+"113 features_7_0_attn_proj_bias" -> "2595 linear_138"  [label="(768,)", style=solid];
+"114 features_7_0_attn_logit_scale" -> "2586 clamp_22"  [label="(24, 1, 1)", style=solid];
+"115 features_7_1_attn_qkv_weight" -> "2632 linear_143"  [label="(2304, 768)", style=solid];
+"116 features_7_1_attn_proj_weight" -> "2657 linear_144"  [label="(768, 768)", style=solid];
+"117 features_7_1_attn_qkv_bias" -> "2629 clone_23"  [label="(2304,)", style=solid];
+"118 features_7_1_attn_proj_bias" -> "2657 linear_144"  [label="(768,)", style=solid];
+"119 features_7_1_attn_logit_scale" -> "2648 clamp_23"  [label="(24, 1, 1)", style=solid];
+"120 features_0_0_weight" -> "549 conv2d"  [label="(96, 3, 4, 4)", style=solid];
+"121 features_0_0_bias" -> "549 conv2d"  [label="(96,)", style=solid];
+"122 features_0_2_weight" -> "551 layer_norm"  [label="(96,)", style=solid];
+"123 features_0_2_bias" -> "551 layer_norm"  [label="(96,)", style=solid];
+"124 features_1_0_attn_cpb_mlp_0_weight" -> "552 linear"  [label="(512, 2)", style=solid];
+"125 features_1_0_attn_cpb_mlp_0_bias" -> "552 linear"  [label="(512,)", style=solid];
+"126 features_1_0_attn_cpb_mlp_2_weight" -> "554 linear_1"  [label="(3, 512)", style=solid];
+"127 features_1_0_norm1_weight" -> "602 layer_norm_1"  [label="(96,)", style=solid];
+"128 features_1_0_norm1_bias" -> "602 layer_norm_1"  [label="(96,)", style=solid];
+"129 features_1_0_mlp_0_weight" -> "604 linear_4"  [label="(384, 96)", style=solid];
+"130 features_1_0_mlp_0_bias" -> "604 linear_4"  [label="(384,)", style=solid];
+"131 features_1_0_mlp_3_weight" -> "607 linear_5"  [label="(96, 384)", style=solid];
+"132 features_1_0_mlp_3_bias" -> "607 linear_5"  [label="(96,)", style=solid];
+"133 features_1_0_norm2_weight" -> "609 layer_norm_2"  [label="(96,)", style=solid];
+"134 features_1_0_norm2_bias" -> "609 layer_norm_2"  [label="(96,)", style=solid];
+"135 features_1_1_attn_cpb_mlp_0_weight" -> "611 linear_6"  [label="(512, 2)", style=solid];
+"136 features_1_1_attn_cpb_mlp_0_bias" -> "611 linear_6"  [label="(512,)", style=solid];
+"137 features_1_1_attn_cpb_mlp_2_weight" -> "613 linear_7"  [label="(3, 512)", style=solid];
+"138 features_1_1_norm1_weight" -> "715 layer_norm_3"  [label="(96,)", style=solid];
+"139 features_1_1_norm1_bias" -> "715 layer_norm_3"  [label="(96,)", style=solid];
+"140 features_1_1_mlp_0_weight" -> "717 linear_10"  [label="(384, 96)", style=solid];
+"141 features_1_1_mlp_0_bias" -> "717 linear_10"  [label="(384,)", style=solid];
+"142 features_1_1_mlp_3_weight" -> "720 linear_11"  [label="(96, 384)", style=solid];
+"143 features_1_1_mlp_3_bias" -> "720 linear_11"  [label="(96,)", style=solid];
+"144 features_1_1_norm2_weight" -> "722 layer_norm_4"  [label="(96,)", style=solid];
+"145 features_1_1_norm2_bias" -> "722 layer_norm_4"  [label="(96,)", style=solid];
+"146 features_2_reduction_weight" -> "738 linear_12"  [label="(192, 384)", style=solid];
+"147 features_2_norm_weight" -> "739 layer_norm_5"  [label="(192,)", style=solid];
+"148 features_2_norm_bias" -> "739 layer_norm_5"  [label="(192,)", style=solid];
+"149 features_3_0_attn_cpb_mlp_0_weight" -> "740 linear_13"  [label="(512, 2)", style=solid];
+"150 features_3_0_attn_cpb_mlp_0_bias" -> "740 linear_13"  [label="(512,)", style=solid];
+"151 features_3_0_attn_cpb_mlp_2_weight" -> "742 linear_14"  [label="(6, 512)", style=solid];
+"152 features_3_0_norm1_weight" -> "793 layer_norm_6"  [label="(192,)", style=solid];
+"153 features_3_0_norm1_bias" -> "793 layer_norm_6"  [label="(192,)", style=solid];
+"154 features_3_0_mlp_0_weight" -> "795 linear_17"  [label="(768, 192)", style=solid];
+"155 features_3_0_mlp_0_bias" -> "795 linear_17"  [label="(768,)", style=solid];
+"156 features_3_0_mlp_3_weight" -> "798 linear_18"  [label="(192, 768)", style=solid];
+"157 features_3_0_mlp_3_bias" -> "798 linear_18"  [label="(192,)", style=solid];
+"158 features_3_0_norm2_weight" -> "800 layer_norm_7"  [label="(192,)", style=solid];
+"159 features_3_0_norm2_bias" -> "800 layer_norm_7"  [label="(192,)", style=solid];
+"160 features_3_1_attn_cpb_mlp_0_weight" -> "802 linear_19"  [label="(512, 2)", style=solid];
+"161 features_3_1_attn_cpb_mlp_0_bias" -> "802 linear_19"  [label="(512,)", style=solid];
+"162 features_3_1_attn_cpb_mlp_2_weight" -> "804 linear_20"  [label="(6, 512)", style=solid];
+"163 features_3_1_norm1_weight" -> "909 layer_norm_8"  [label="(192,)", style=solid];
+"164 features_3_1_norm1_bias" -> "909 layer_norm_8"  [label="(192,)", style=solid];
+"165 features_3_1_mlp_0_weight" -> "911 linear_23"  [label="(768, 192)", style=solid];
+"166 features_3_1_mlp_0_bias" -> "911 linear_23"  [label="(768,)", style=solid];
+"167 features_3_1_mlp_3_weight" -> "914 linear_24"  [label="(192, 768)", style=solid];
+"168 features_3_1_mlp_3_bias" -> "914 linear_24"  [label="(192,)", style=solid];
+"169 features_3_1_norm2_weight" -> "916 layer_norm_9"  [label="(192,)", style=solid];
+"170 features_3_1_norm2_bias" -> "916 layer_norm_9"  [label="(192,)", style=solid];
+"171 features_4_reduction_weight" -> "932 linear_25"  [label="(384, 768)", style=solid];
+"172 features_4_norm_weight" -> "933 layer_norm_10"  [label="(384,)", style=solid];
+"173 features_4_norm_bias" -> "933 layer_norm_10"  [label="(384,)", style=solid];
+"174 features_5_0_attn_cpb_mlp_0_weight" -> "934 linear_26"  [label="(512, 2)", style=solid];
+"175 features_5_0_attn_cpb_mlp_0_bias" -> "934 linear_26"  [label="(512,)", style=solid];
+"176 features_5_0_attn_cpb_mlp_2_weight" -> "936 linear_27"  [label="(12, 512)", style=solid];
+"177 features_5_0_norm1_weight" -> "987 layer_norm_11"  [label="(384,)", style=solid];
+"178 features_5_0_norm1_bias" -> "987 layer_norm_11"  [label="(384,)", style=solid];
+"179 features_5_0_mlp_0_weight" -> "989 linear_30"  [label="(1536, 384)", style=solid];
+"180 features_5_0_mlp_0_bias" -> "989 linear_30"  [label="(1536,)", style=solid];
+"181 features_5_0_mlp_3_weight" -> "992 linear_31"  [label="(384, 1536)", style=solid];
+"182 features_5_0_mlp_3_bias" -> "992 linear_31"  [label="(384,)", style=solid];
+"183 features_5_0_norm2_weight" -> "994 layer_norm_12"  [label="(384,)", style=solid];
+"184 features_5_0_norm2_bias" -> "994 layer_norm_12"  [label="(384,)", style=solid];
+"185 features_5_1_attn_cpb_mlp_0_weight" -> "996 linear_32"  [label="(512, 2)", style=solid];
+"186 features_5_1_attn_cpb_mlp_0_bias" -> "996 linear_32"  [label="(512,)", style=solid];
+"187 features_5_1_attn_cpb_mlp_2_weight" -> "998 linear_33"  [label="(12, 512)", style=solid];
+"188 features_5_1_norm1_weight" -> "1103 layer_norm_13"  [label="(384,)", style=solid];
+"189 features_5_1_norm1_bias" -> "1103 layer_norm_13"  [label="(384,)", style=solid];
+"190 features_5_1_mlp_0_weight" -> "1105 linear_36"  [label="(1536, 384)", style=solid];
+"191 features_5_1_mlp_0_bias" -> "1105 linear_36"  [label="(1536,)", style=solid];
+"192 features_5_1_mlp_3_weight" -> "1108 linear_37"  [label="(384, 1536)", style=solid];
+"193 features_5_1_mlp_3_bias" -> "1108 linear_37"  [label="(384,)", style=solid];
+"194 features_5_1_norm2_weight" -> "1110 layer_norm_14"  [label="(384,)", style=solid];
+"195 features_5_1_norm2_bias" -> "1110 layer_norm_14"  [label="(384,)", style=solid];
+"196 features_5_2_attn_cpb_mlp_0_weight" -> "1112 linear_38"  [label="(512, 2)", style=solid];
+"197 features_5_2_attn_cpb_mlp_0_bias" -> "1112 linear_38"  [label="(512,)", style=solid];
+"198 features_5_2_attn_cpb_mlp_2_weight" -> "1114 linear_39"  [label="(12, 512)", style=solid];
+"199 features_5_2_norm1_weight" -> "1165 layer_norm_15"  [label="(384,)", style=solid];
+"200 features_5_2_norm1_bias" -> "1165 layer_norm_15"  [label="(384,)", style=solid];
+"201 features_5_2_mlp_0_weight" -> "1167 linear_42"  [label="(1536, 384)", style=solid];
+"202 features_5_2_mlp_0_bias" -> "1167 linear_42"  [label="(1536,)", style=solid];
+"203 features_5_2_mlp_3_weight" -> "1170 linear_43"  [label="(384, 1536)", style=solid];
+"204 features_5_2_mlp_3_bias" -> "1170 linear_43"  [label="(384,)", style=solid];
+"205 features_5_2_norm2_weight" -> "1172 layer_norm_16"  [label="(384,)", style=solid];
+"206 features_5_2_norm2_bias" -> "1172 layer_norm_16"  [label="(384,)", style=solid];
+"207 features_5_3_attn_cpb_mlp_0_weight" -> "1174 linear_44"  [label="(512, 2)", style=solid];
+"208 features_5_3_attn_cpb_mlp_0_bias" -> "1174 linear_44"  [label="(512,)", style=solid];
+"209 features_5_3_attn_cpb_mlp_2_weight" -> "1176 linear_45"  [label="(12, 512)", style=solid];
+"210 features_5_3_norm1_weight" -> "1281 layer_norm_17"  [label="(384,)", style=solid];
+"211 features_5_3_norm1_bias" -> "1281 layer_norm_17"  [label="(384,)", style=solid];
+"212 features_5_3_mlp_0_weight" -> "1283 linear_48"  [label="(1536, 384)", style=solid];
+"213 features_5_3_mlp_0_bias" -> "1283 linear_48"  [label="(1536,)", style=solid];
+"214 features_5_3_mlp_3_weight" -> "1286 linear_49"  [label="(384, 1536)", style=solid];
+"215 features_5_3_mlp_3_bias" -> "1286 linear_49"  [label="(384,)", style=solid];
+"216 features_5_3_norm2_weight" -> "1288 layer_norm_18"  [label="(384,)", style=solid];
+"217 features_5_3_norm2_bias" -> "1288 layer_norm_18"  [label="(384,)", style=solid];
+"218 features_5_4_attn_cpb_mlp_0_weight" -> "1290 linear_50"  [label="(512, 2)", style=solid];
+"219 features_5_4_attn_cpb_mlp_0_bias" -> "1290 linear_50"  [label="(512,)", style=solid];
+"220 features_5_4_attn_cpb_mlp_2_weight" -> "1292 linear_51"  [label="(12, 512)", style=solid];
+"221 features_5_4_norm1_weight" -> "1343 layer_norm_19"  [label="(384,)", style=solid];
+"222 features_5_4_norm1_bias" -> "1343 layer_norm_19"  [label="(384,)", style=solid];
+"223 features_5_4_mlp_0_weight" -> "1345 linear_54"  [label="(1536, 384)", style=solid];
+"224 features_5_4_mlp_0_bias" -> "1345 linear_54"  [label="(1536,)", style=solid];
+"225 features_5_4_mlp_3_weight" -> "1348 linear_55"  [label="(384, 1536)", style=solid];
+"226 features_5_4_mlp_3_bias" -> "1348 linear_55"  [label="(384,)", style=solid];
+"227 features_5_4_norm2_weight" -> "1350 layer_norm_20"  [label="(384,)", style=solid];
+"228 features_5_4_norm2_bias" -> "1350 layer_norm_20"  [label="(384,)", style=solid];
+"229 features_5_5_attn_cpb_mlp_0_weight" -> "1352 linear_56"  [label="(512, 2)", style=solid];
+"230 features_5_5_attn_cpb_mlp_0_bias" -> "1352 linear_56"  [label="(512,)", style=solid];
+"231 features_5_5_attn_cpb_mlp_2_weight" -> "1354 linear_57"  [label="(12, 512)", style=solid];
+"232 features_5_5_norm1_weight" -> "1459 layer_norm_21"  [label="(384,)", style=solid];
+"233 features_5_5_norm1_bias" -> "1459 layer_norm_21"  [label="(384,)", style=solid];
+"234 features_5_5_mlp_0_weight" -> "1461 linear_60"  [label="(1536, 384)", style=solid];
+"235 features_5_5_mlp_0_bias" -> "1461 linear_60"  [label="(1536,)", style=solid];
+"236 features_5_5_mlp_3_weight" -> "1464 linear_61"  [label="(384, 1536)", style=solid];
+"237 features_5_5_mlp_3_bias" -> "1464 linear_61"  [label="(384,)", style=solid];
+"238 features_5_5_norm2_weight" -> "1466 layer_norm_22"  [label="(384,)", style=solid];
+"239 features_5_5_norm2_bias" -> "1466 layer_norm_22"  [label="(384,)", style=solid];
+"240 features_5_6_attn_cpb_mlp_0_weight" -> "1468 linear_62"  [label="(512, 2)", style=solid];
+"241 features_5_6_attn_cpb_mlp_0_bias" -> "1468 linear_62"  [label="(512,)", style=solid];
+"242 features_5_6_attn_cpb_mlp_2_weight" -> "1470 linear_63"  [label="(12, 512)", style=solid];
+"243 features_5_6_norm1_weight" -> "1521 layer_norm_23"  [label="(384,)", style=solid];
+"244 features_5_6_norm1_bias" -> "1521 layer_norm_23"  [label="(384,)", style=solid];
+"245 features_5_6_mlp_0_weight" -> "1523 linear_66"  [label="(1536, 384)", style=solid];
+"246 features_5_6_mlp_0_bias" -> "1523 linear_66"  [label="(1536,)", style=solid];
+"247 features_5_6_mlp_3_weight" -> "1526 linear_67"  [label="(384, 1536)", style=solid];
+"248 features_5_6_mlp_3_bias" -> "1526 linear_67"  [label="(384,)", style=solid];
+"249 features_5_6_norm2_weight" -> "1528 layer_norm_24"  [label="(384,)", style=solid];
+"250 features_5_6_norm2_bias" -> "1528 layer_norm_24"  [label="(384,)", style=solid];
+"251 features_5_7_attn_cpb_mlp_0_weight" -> "1530 linear_68"  [label="(512, 2)", style=solid];
+"252 features_5_7_attn_cpb_mlp_0_bias" -> "1530 linear_68"  [label="(512,)", style=solid];
+"253 features_5_7_attn_cpb_mlp_2_weight" -> "1532 linear_69"  [label="(12, 512)", style=solid];
+"254 features_5_7_norm1_weight" -> "1637 layer_norm_25"  [label="(384,)", style=solid];
+"255 features_5_7_norm1_bias" -> "1637 layer_norm_25"  [label="(384,)", style=solid];
+"256 features_5_7_mlp_0_weight" -> "1639 linear_72"  [label="(1536, 384)", style=solid];
+"257 features_5_7_mlp_0_bias" -> "1639 linear_72"  [label="(1536,)", style=solid];
+"258 features_5_7_mlp_3_weight" -> "1642 linear_73"  [label="(384, 1536)", style=solid];
+"259 features_5_7_mlp_3_bias" -> "1642 linear_73"  [label="(384,)", style=solid];
+"260 features_5_7_norm2_weight" -> "1644 layer_norm_26"  [label="(384,)", style=solid];
+"261 features_5_7_norm2_bias" -> "1644 layer_norm_26"  [label="(384,)", style=solid];
+"262 features_5_8_attn_cpb_mlp_0_weight" -> "1646 linear_74"  [label="(512, 2)", style=solid];
+"263 features_5_8_attn_cpb_mlp_0_bias" -> "1646 linear_74"  [label="(512,)", style=solid];
+"264 features_5_8_attn_cpb_mlp_2_weight" -> "1648 linear_75"  [label="(12, 512)", style=solid];
+"265 features_5_8_norm1_weight" -> "1699 layer_norm_27"  [label="(384,)", style=solid];
+"266 features_5_8_norm1_bias" -> "1699 layer_norm_27"  [label="(384,)", style=solid];
+"267 features_5_8_mlp_0_weight" -> "1701 linear_78"  [label="(1536, 384)", style=solid];
+"268 features_5_8_mlp_0_bias" -> "1701 linear_78"  [label="(1536,)", style=solid];
+"269 features_5_8_mlp_3_weight" -> "1704 linear_79"  [label="(384, 1536)", style=solid];
+"270 features_5_8_mlp_3_bias" -> "1704 linear_79"  [label="(384,)", style=solid];
+"271 features_5_8_norm2_weight" -> "1706 layer_norm_28"  [label="(384,)", style=solid];
+"272 features_5_8_norm2_bias" -> "1706 layer_norm_28"  [label="(384,)", style=solid];
+"273 features_5_9_attn_cpb_mlp_0_weight" -> "1708 linear_80"  [label="(512, 2)", style=solid];
+"274 features_5_9_attn_cpb_mlp_0_bias" -> "1708 linear_80"  [label="(512,)", style=solid];
+"275 features_5_9_attn_cpb_mlp_2_weight" -> "1710 linear_81"  [label="(12, 512)", style=solid];
+"276 features_5_9_norm1_weight" -> "1815 layer_norm_29"  [label="(384,)", style=solid];
+"277 features_5_9_norm1_bias" -> "1815 layer_norm_29"  [label="(384,)", style=solid];
+"278 features_5_9_mlp_0_weight" -> "1817 linear_84"  [label="(1536, 384)", style=solid];
+"279 features_5_9_mlp_0_bias" -> "1817 linear_84"  [label="(1536,)", style=solid];
+"280 features_5_9_mlp_3_weight" -> "1820 linear_85"  [label="(384, 1536)", style=solid];
+"281 features_5_9_mlp_3_bias" -> "1820 linear_85"  [label="(384,)", style=solid];
+"282 features_5_9_norm2_weight" -> "1822 layer_norm_30"  [label="(384,)", style=solid];
+"283 features_5_9_norm2_bias" -> "1822 layer_norm_30"  [label="(384,)", style=solid];
+"284 features_5_10_attn_cpb_mlp_0_weight" -> "1824 linear_86"  [label="(512, 2)", style=solid];
+"285 features_5_10_attn_cpb_mlp_0_bias" -> "1824 linear_86"  [label="(512,)", style=solid];
+"286 features_5_10_attn_cpb_mlp_2_weight" -> "1826 linear_87"  [label="(12, 512)", style=solid];
+"287 features_5_10_norm1_weight" -> "1877 layer_norm_31"  [label="(384,)", style=solid];
+"288 features_5_10_norm1_bias" -> "1877 layer_norm_31"  [label="(384,)", style=solid];
+"289 features_5_10_mlp_0_weight" -> "1879 linear_90"  [label="(1536, 384)", style=solid];
+"290 features_5_10_mlp_0_bias" -> "1879 linear_90"  [label="(1536,)", style=solid];
+"291 features_5_10_mlp_3_weight" -> "1882 linear_91"  [label="(384, 1536)", style=solid];
+"292 features_5_10_mlp_3_bias" -> "1882 linear_91"  [label="(384,)", style=solid];
+"293 features_5_10_norm2_weight" -> "1884 layer_norm_32"  [label="(384,)", style=solid];
+"294 features_5_10_norm2_bias" -> "1884 layer_norm_32"  [label="(384,)", style=solid];
+"295 features_5_11_attn_cpb_mlp_0_weight" -> "1886 linear_92"  [label="(512, 2)", style=solid];
+"296 features_5_11_attn_cpb_mlp_0_bias" -> "1886 linear_92"  [label="(512,)", style=solid];
+"297 features_5_11_attn_cpb_mlp_2_weight" -> "1888 linear_93"  [label="(12, 512)", style=solid];
+"298 features_5_11_norm1_weight" -> "1993 layer_norm_33"  [label="(384,)", style=solid];
+"299 features_5_11_norm1_bias" -> "1993 layer_norm_33"  [label="(384,)", style=solid];
+"300 features_5_11_mlp_0_weight" -> "1995 linear_96"  [label="(1536, 384)", style=solid];
+"301 features_5_11_mlp_0_bias" -> "1995 linear_96"  [label="(1536,)", style=solid];
+"302 features_5_11_mlp_3_weight" -> "1998 linear_97"  [label="(384, 1536)", style=solid];
+"303 features_5_11_mlp_3_bias" -> "1998 linear_97"  [label="(384,)", style=solid];
+"304 features_5_11_norm2_weight" -> "2000 layer_norm_34"  [label="(384,)", style=solid];
+"305 features_5_11_norm2_bias" -> "2000 layer_norm_34"  [label="(384,)", style=solid];
+"306 features_5_12_attn_cpb_mlp_0_weight" -> "2002 linear_98"  [label="(512, 2)", style=solid];
+"307 features_5_12_attn_cpb_mlp_0_bias" -> "2002 linear_98"  [label="(512,)", style=solid];
+"308 features_5_12_attn_cpb_mlp_2_weight" -> "2004 linear_99"  [label="(12, 512)", style=solid];
+"309 features_5_12_norm1_weight" -> "2055 layer_norm_35"  [label="(384,)", style=solid];
+"310 features_5_12_norm1_bias" -> "2055 layer_norm_35"  [label="(384,)", style=solid];
+"311 features_5_12_mlp_0_weight" -> "2057 linear_102"  [label="(1536, 384)", style=solid];
+"312 features_5_12_mlp_0_bias" -> "2057 linear_102"  [label="(1536,)", style=solid];
+"313 features_5_12_mlp_3_weight" -> "2060 linear_103"  [label="(384, 1536)", style=solid];
+"314 features_5_12_mlp_3_bias" -> "2060 linear_103"  [label="(384,)", style=solid];
+"315 features_5_12_norm2_weight" -> "2062 layer_norm_36"  [label="(384,)", style=solid];
+"316 features_5_12_norm2_bias" -> "2062 layer_norm_36"  [label="(384,)", style=solid];
+"317 features_5_13_attn_cpb_mlp_0_weight" -> "2064 linear_104"  [label="(512, 2)", style=solid];
+"318 features_5_13_attn_cpb_mlp_0_bias" -> "2064 linear_104"  [label="(512,)", style=solid];
+"319 features_5_13_attn_cpb_mlp_2_weight" -> "2066 linear_105"  [label="(12, 512)", style=solid];
+"320 features_5_13_norm1_weight" -> "2171 layer_norm_37"  [label="(384,)", style=solid];
+"321 features_5_13_norm1_bias" -> "2171 layer_norm_37"  [label="(384,)", style=solid];
+"322 features_5_13_mlp_0_weight" -> "2173 linear_108"  [label="(1536, 384)", style=solid];
+"323 features_5_13_mlp_0_bias" -> "2173 linear_108"  [label="(1536,)", style=solid];
+"324 features_5_13_mlp_3_weight" -> "2176 linear_109"  [label="(384, 1536)", style=solid];
+"325 features_5_13_mlp_3_bias" -> "2176 linear_109"  [label="(384,)", style=solid];
+"326 features_5_13_norm2_weight" -> "2178 layer_norm_38"  [label="(384,)", style=solid];
+"327 features_5_13_norm2_bias" -> "2178 layer_norm_38"  [label="(384,)", style=solid];
+"328 features_5_14_attn_cpb_mlp_0_weight" -> "2180 linear_110"  [label="(512, 2)", style=solid];
+"329 features_5_14_attn_cpb_mlp_0_bias" -> "2180 linear_110"  [label="(512,)", style=solid];
+"330 features_5_14_attn_cpb_mlp_2_weight" -> "2182 linear_111"  [label="(12, 512)", style=solid];
+"331 features_5_14_norm1_weight" -> "2233 layer_norm_39"  [label="(384,)", style=solid];
+"332 features_5_14_norm1_bias" -> "2233 layer_norm_39"  [label="(384,)", style=solid];
+"333 features_5_14_mlp_0_weight" -> "2235 linear_114"  [label="(1536, 384)", style=solid];
+"334 features_5_14_mlp_0_bias" -> "2235 linear_114"  [label="(1536,)", style=solid];
+"335 features_5_14_mlp_3_weight" -> "2238 linear_115"  [label="(384, 1536)", style=solid];
+"336 features_5_14_mlp_3_bias" -> "2238 linear_115"  [label="(384,)", style=solid];
+"337 features_5_14_norm2_weight" -> "2240 layer_norm_40"  [label="(384,)", style=solid];
+"338 features_5_14_norm2_bias" -> "2240 layer_norm_40"  [label="(384,)", style=solid];
+"339 features_5_15_attn_cpb_mlp_0_weight" -> "2242 linear_116"  [label="(512, 2)", style=solid];
+"340 features_5_15_attn_cpb_mlp_0_bias" -> "2242 linear_116"  [label="(512,)", style=solid];
+"341 features_5_15_attn_cpb_mlp_2_weight" -> "2244 linear_117"  [label="(12, 512)", style=solid];
+"342 features_5_15_norm1_weight" -> "2349 layer_norm_41"  [label="(384,)", style=solid];
+"343 features_5_15_norm1_bias" -> "2349 layer_norm_41"  [label="(384,)", style=solid];
+"344 features_5_15_mlp_0_weight" -> "2351 linear_120"  [label="(1536, 384)", style=solid];
+"345 features_5_15_mlp_0_bias" -> "2351 linear_120"  [label="(1536,)", style=solid];
+"346 features_5_15_mlp_3_weight" -> "2354 linear_121"  [label="(384, 1536)", style=solid];
+"347 features_5_15_mlp_3_bias" -> "2354 linear_121"  [label="(384,)", style=solid];
+"348 features_5_15_norm2_weight" -> "2356 layer_norm_42"  [label="(384,)", style=solid];
+"349 features_5_15_norm2_bias" -> "2356 layer_norm_42"  [label="(384,)", style=solid];
+"350 features_5_16_attn_cpb_mlp_0_weight" -> "2358 linear_122"  [label="(512, 2)", style=solid];
+"351 features_5_16_attn_cpb_mlp_0_bias" -> "2358 linear_122"  [label="(512,)", style=solid];
+"352 features_5_16_attn_cpb_mlp_2_weight" -> "2360 linear_123"  [label="(12, 512)", style=solid];
+"353 features_5_16_norm1_weight" -> "2411 layer_norm_43"  [label="(384,)", style=solid];
+"354 features_5_16_norm1_bias" -> "2411 layer_norm_43"  [label="(384,)", style=solid];
+"355 features_5_16_mlp_0_weight" -> "2413 linear_126"  [label="(1536, 384)", style=solid];
+"356 features_5_16_mlp_0_bias" -> "2413 linear_126"  [label="(1536,)", style=solid];
+"357 features_5_16_mlp_3_weight" -> "2416 linear_127"  [label="(384, 1536)", style=solid];
+"358 features_5_16_mlp_3_bias" -> "2416 linear_127"  [label="(384,)", style=solid];
+"359 features_5_16_norm2_weight" -> "2418 layer_norm_44"  [label="(384,)", style=solid];
+"360 features_5_16_norm2_bias" -> "2418 layer_norm_44"  [label="(384,)", style=solid];
+"361 features_5_17_attn_cpb_mlp_0_weight" -> "2420 linear_128"  [label="(512, 2)", style=solid];
+"362 features_5_17_attn_cpb_mlp_0_bias" -> "2420 linear_128"  [label="(512,)", style=solid];
+"363 features_5_17_attn_cpb_mlp_2_weight" -> "2422 linear_129"  [label="(12, 512)", style=solid];
+"364 features_5_17_norm1_weight" -> "2527 layer_norm_45"  [label="(384,)", style=solid];
+"365 features_5_17_norm1_bias" -> "2527 layer_norm_45"  [label="(384,)", style=solid];
+"366 features_5_17_mlp_0_weight" -> "2529 linear_132"  [label="(1536, 384)", style=solid];
+"367 features_5_17_mlp_0_bias" -> "2529 linear_132"  [label="(1536,)", style=solid];
+"368 features_5_17_mlp_3_weight" -> "2532 linear_133"  [label="(384, 1536)", style=solid];
+"369 features_5_17_mlp_3_bias" -> "2532 linear_133"  [label="(384,)", style=solid];
+"370 features_5_17_norm2_weight" -> "2534 layer_norm_46"  [label="(384,)", style=solid];
+"371 features_5_17_norm2_bias" -> "2534 layer_norm_46"  [label="(384,)", style=solid];
+"372 features_6_reduction_weight" -> "2550 linear_134"  [label="(768, 1536)", style=solid];
+"373 features_6_norm_weight" -> "2551 layer_norm_47"  [label="(768,)", style=solid];
+"374 features_6_norm_bias" -> "2551 layer_norm_47"  [label="(768,)", style=solid];
+"375 features_7_0_attn_cpb_mlp_0_weight" -> "2552 linear_135"  [label="(512, 2)", style=solid];
+"376 features_7_0_attn_cpb_mlp_0_bias" -> "2552 linear_135"  [label="(512,)", style=solid];
+"377 features_7_0_attn_cpb_mlp_2_weight" -> "2554 linear_136"  [label="(24, 512)", style=solid];
+"378 features_7_0_norm1_weight" -> "2605 layer_norm_48"  [label="(768,)", style=solid];
+"379 features_7_0_norm1_bias" -> "2605 layer_norm_48"  [label="(768,)", style=solid];
+"380 features_7_0_mlp_0_weight" -> "2607 linear_139"  [label="(3072, 768)", style=solid];
+"381 features_7_0_mlp_0_bias" -> "2607 linear_139"  [label="(3072,)", style=solid];
+"382 features_7_0_mlp_3_weight" -> "2610 linear_140"  [label="(768, 3072)", style=solid];
+"383 features_7_0_mlp_3_bias" -> "2610 linear_140"  [label="(768,)", style=solid];
+"384 features_7_0_norm2_weight" -> "2612 layer_norm_49"  [label="(768,)", style=solid];
+"385 features_7_0_norm2_bias" -> "2612 layer_norm_49"  [label="(768,)", style=solid];
+"386 features_7_1_attn_cpb_mlp_0_weight" -> "2614 linear_141"  [label="(512, 2)", style=solid];
+"387 features_7_1_attn_cpb_mlp_0_bias" -> "2614 linear_141"  [label="(512,)", style=solid];
+"388 features_7_1_attn_cpb_mlp_2_weight" -> "2616 linear_142"  [label="(24, 512)", style=solid];
+"389 features_7_1_norm1_weight" -> "2667 layer_norm_50"  [label="(768,)", style=solid];
+"390 features_7_1_norm1_bias" -> "2667 layer_norm_50"  [label="(768,)", style=solid];
+"391 features_7_1_mlp_0_weight" -> "2669 linear_145"  [label="(3072, 768)", style=solid];
+"392 features_7_1_mlp_0_bias" -> "2669 linear_145"  [label="(3072,)", style=solid];
+"393 features_7_1_mlp_3_weight" -> "2672 linear_146"  [label="(768, 3072)", style=solid];
+"394 features_7_1_mlp_3_bias" -> "2672 linear_146"  [label="(768,)", style=solid];
+"395 features_7_1_norm2_weight" -> "2674 layer_norm_51"  [label="(768,)", style=solid];
+"396 features_7_1_norm2_bias" -> "2674 layer_norm_51"  [label="(768,)", style=solid];
+"397 norm_weight" -> "2676 layer_norm_52"  [label="(768,)", style=solid];
+"398 norm_bias" -> "2676 layer_norm_52"  [label="(768,)", style=solid];
+"399 head_weight" -> "2680 linear_147"  [label="(1000, 768)", style=solid];
+"400 head_bias" -> "2680 linear_147"  [label="(1000,)", style=solid];
+"401 features_1_0_attn_relative_coords_table" -> "552 linear"  [label="(1, 15, 15, 2)", style=solid];
+"402 features_1_0_attn_relative_position_index" -> "556 index"  [label="(4096,)", style=solid];
+"403 features_1_1_attn_relative_coords_table" -> "611 linear_6"  [label="(1, 15, 15, 2)", style=solid];
+"404 features_1_1_attn_relative_position_index" -> "615 index_1"  [label="(4096,)", style=solid];
+"405 features_3_0_attn_relative_coords_table" -> "740 linear_13"  [label="(1, 15, 15, 2)", style=solid];
+"406 features_3_0_attn_relative_position_index" -> "744 index_2"  [label="(4096,)", style=solid];
+"407 features_3_1_attn_relative_coords_table" -> "802 linear_19"  [label="(1, 15, 15, 2)", style=solid];
+"408 features_3_1_attn_relative_position_index" -> "806 index_3"  [label="(4096,)", style=solid];
+"409 features_5_0_attn_relative_coords_table" -> "934 linear_26"  [label="(1, 15, 15, 2)", style=solid];
+"410 features_5_0_attn_relative_position_index" -> "938 index_4"  [label="(4096,)", style=solid];
+"411 features_5_1_attn_relative_coords_table" -> "996 linear_32"  [label="(1, 15, 15, 2)", style=solid];
+"412 features_5_1_attn_relative_position_index" -> "1000 index_5"  [label="(4096,)", style=solid];
+"413 features_5_2_attn_relative_coords_table" -> "1112 linear_38"  [label="(1, 15, 15, 2)", style=solid];
+"414 features_5_2_attn_relative_position_index" -> "1116 index_6"  [label="(4096,)", style=solid];
+"415 features_5_3_attn_relative_coords_table" -> "1174 linear_44"  [label="(1, 15, 15, 2)", style=solid];
+"416 features_5_3_attn_relative_position_index" -> "1178 index_7"  [label="(4096,)", style=solid];
+"417 features_5_4_attn_relative_coords_table" -> "1290 linear_50"  [label="(1, 15, 15, 2)", style=solid];
+"418 features_5_4_attn_relative_position_index" -> "1294 index_8"  [label="(4096,)", style=solid];
+"419 features_5_5_attn_relative_coords_table" -> "1352 linear_56"  [label="(1, 15, 15, 2)", style=solid];
+"420 features_5_5_attn_relative_position_index" -> "1356 index_9"  [label="(4096,)", style=solid];
+"421 features_5_6_attn_relative_coords_table" -> "1468 linear_62"  [label="(1, 15, 15, 2)", style=solid];
+"422 features_5_6_attn_relative_position_index" -> "1472 index_10"  [label="(4096,)", style=solid];
+"423 features_5_7_attn_relative_coords_table" -> "1530 linear_68"  [label="(1, 15, 15, 2)", style=solid];
+"424 features_5_7_attn_relative_position_index" -> "1534 index_11"  [label="(4096,)", style=solid];
+"425 features_5_8_attn_relative_coords_table" -> "1646 linear_74"  [label="(1, 15, 15, 2)", style=solid];
+"426 features_5_8_attn_relative_position_index" -> "1650 index_12"  [label="(4096,)", style=solid];
+"427 features_5_9_attn_relative_coords_table" -> "1708 linear_80"  [label="(1, 15, 15, 2)", style=solid];
+"428 features_5_9_attn_relative_position_index" -> "1712 index_13"  [label="(4096,)", style=solid];
+"429 features_5_10_attn_relative_coords_table" -> "1824 linear_86"  [label="(1, 15, 15, 2)", style=solid];
+"430 features_5_10_attn_relative_position_index" -> "1828 index_14"  [label="(4096,)", style=solid];
+"431 features_5_11_attn_relative_coords_table" -> "1886 linear_92"  [label="(1, 15, 15, 2)", style=solid];
+"432 features_5_11_attn_relative_position_index" -> "1890 index_15"  [label="(4096,)", style=solid];
+"433 features_5_12_attn_relative_coords_table" -> "2002 linear_98"  [label="(1, 15, 15, 2)", style=solid];
+"434 features_5_12_attn_relative_position_index" -> "2006 index_16"  [label="(4096,)", style=solid];
+"435 features_5_13_attn_relative_coords_table" -> "2064 linear_104"  [label="(1, 15, 15, 2)", style=solid];
+"436 features_5_13_attn_relative_position_index" -> "2068 index_17"  [label="(4096,)", style=solid];
+"437 features_5_14_attn_relative_coords_table" -> "2180 linear_110"  [label="(1, 15, 15, 2)", style=solid];
+"438 features_5_14_attn_relative_position_index" -> "2184 index_18"  [label="(4096,)", style=solid];
+"439 features_5_15_attn_relative_coords_table" -> "2242 linear_116"  [label="(1, 15, 15, 2)", style=solid];
+"440 features_5_15_attn_relative_position_index" -> "2246 index_19"  [label="(4096,)", style=solid];
+"441 features_5_16_attn_relative_coords_table" -> "2358 linear_122"  [label="(1, 15, 15, 2)", style=solid];
+"442 features_5_16_attn_relative_position_index" -> "2362 index_20"  [label="(4096,)", style=solid];
+"443 features_5_17_attn_relative_coords_table" -> "2420 linear_128"  [label="(1, 15, 15, 2)", style=solid];
+"444 features_5_17_attn_relative_position_index" -> "2424 index_21"  [label="(4096,)", style=solid];
+"445 features_7_0_attn_relative_coords_table" -> "2552 linear_135"  [label="(1, 15, 15, 2)", style=solid];
+"446 features_7_0_attn_relative_position_index" -> "2556 index_22"  [label="(4096,)", style=solid];
+"447 features_7_1_attn_relative_coords_table" -> "2614 linear_141"  [label="(1, 15, 15, 2)", style=solid];
+"448 features_7_1_attn_relative_position_index" -> "2618 index_23"  [label="(4096,)", style=solid];
+"449 features_1_1_attn_lifted_tensor_0" -> "651 lift_fresh_copy"  [label="()", style=solid];
+"450 features_1_1_attn_lifted_tensor_1" -> "655 lift_fresh_copy_1"  [label="()", style=solid];
+"451 features_1_1_attn_lifted_tensor_2" -> "659 lift_fresh_copy_2"  [label="()", style=solid];
+"452 features_1_1_attn_lifted_tensor_3" -> "663 lift_fresh_copy_3"  [label="()", style=solid];
+"453 features_1_1_attn_lifted_tensor_4" -> "667 lift_fresh_copy_4"  [label="()", style=solid];
+"454 features_1_1_attn_lifted_tensor_5" -> "671 lift_fresh_copy_5"  [label="()", style=solid];
+"455 features_1_1_attn_lifted_tensor_6" -> "675 lift_fresh_copy_6"  [label="()", style=solid];
+"456 features_1_1_attn_lifted_tensor_7" -> "679 lift_fresh_copy_7"  [label="()", style=solid];
+"457 features_1_1_attn_lifted_tensor_8" -> "683 lift_fresh_copy_8"  [label="()", style=solid];
+"458 features_3_1_attn_lifted_tensor_9" -> "842 lift_fresh_copy_9"  [label="()", style=solid];
+"459 features_3_1_attn_lifted_tensor_10" -> "846 lift_fresh_copy_10"  [label="()", style=solid];
+"460 features_3_1_attn_lifted_tensor_11" -> "850 lift_fresh_copy_11"  [label="()", style=solid];
+"461 features_3_1_attn_lifted_tensor_12" -> "854 lift_fresh_copy_12"  [label="()", style=solid];
+"462 features_3_1_attn_lifted_tensor_13" -> "858 lift_fresh_copy_13"  [label="()", style=solid];
+"463 features_3_1_attn_lifted_tensor_14" -> "862 lift_fresh_copy_14"  [label="()", style=solid];
+"464 features_3_1_attn_lifted_tensor_15" -> "866 lift_fresh_copy_15"  [label="()", style=solid];
+"465 features_3_1_attn_lifted_tensor_16" -> "870 lift_fresh_copy_16"  [label="()", style=solid];
+"466 features_3_1_attn_lifted_tensor_17" -> "874 lift_fresh_copy_17"  [label="()", style=solid];
+"467 features_5_1_attn_lifted_tensor_18" -> "1036 lift_fresh_copy_18"  [label="()", style=solid];
+"468 features_5_1_attn_lifted_tensor_19" -> "1040 lift_fresh_copy_19"  [label="()", style=solid];
+"469 features_5_1_attn_lifted_tensor_20" -> "1044 lift_fresh_copy_20"  [label="()", style=solid];
+"470 features_5_1_attn_lifted_tensor_21" -> "1048 lift_fresh_copy_21"  [label="()", style=solid];
+"471 features_5_1_attn_lifted_tensor_22" -> "1052 lift_fresh_copy_22"  [label="()", style=solid];
+"472 features_5_1_attn_lifted_tensor_23" -> "1056 lift_fresh_copy_23"  [label="()", style=solid];
+"473 features_5_1_attn_lifted_tensor_24" -> "1060 lift_fresh_copy_24"  [label="()", style=solid];
+"474 features_5_1_attn_lifted_tensor_25" -> "1064 lift_fresh_copy_25"  [label="()", style=solid];
+"475 features_5_1_attn_lifted_tensor_26" -> "1068 lift_fresh_copy_26"  [label="()", style=solid];
+"476 features_5_3_attn_lifted_tensor_27" -> "1214 lift_fresh_copy_27"  [label="()", style=solid];
+"477 features_5_3_attn_lifted_tensor_28" -> "1218 lift_fresh_copy_28"  [label="()", style=solid];
+"478 features_5_3_attn_lifted_tensor_29" -> "1222 lift_fresh_copy_29"  [label="()", style=solid];
+"479 features_5_3_attn_lifted_tensor_30" -> "1226 lift_fresh_copy_30"  [label="()", style=solid];
+"480 features_5_3_attn_lifted_tensor_31" -> "1230 lift_fresh_copy_31"  [label="()", style=solid];
+"481 features_5_3_attn_lifted_tensor_32" -> "1234 lift_fresh_copy_32"  [label="()", style=solid];
+"482 features_5_3_attn_lifted_tensor_33" -> "1238 lift_fresh_copy_33"  [label="()", style=solid];
+"483 features_5_3_attn_lifted_tensor_34" -> "1242 lift_fresh_copy_34"  [label="()", style=solid];
+"484 features_5_3_attn_lifted_tensor_35" -> "1246 lift_fresh_copy_35"  [label="()", style=solid];
+"485 features_5_5_attn_lifted_tensor_36" -> "1392 lift_fresh_copy_36"  [label="()", style=solid];
+"486 features_5_5_attn_lifted_tensor_37" -> "1396 lift_fresh_copy_37"  [label="()", style=solid];
+"487 features_5_5_attn_lifted_tensor_38" -> "1400 lift_fresh_copy_38"  [label="()", style=solid];
+"488 features_5_5_attn_lifted_tensor_39" -> "1404 lift_fresh_copy_39"  [label="()", style=solid];
+"489 features_5_5_attn_lifted_tensor_40" -> "1408 lift_fresh_copy_40"  [label="()", style=solid];
+"490 features_5_5_attn_lifted_tensor_41" -> "1412 lift_fresh_copy_41"  [label="()", style=solid];
+"491 features_5_5_attn_lifted_tensor_42" -> "1416 lift_fresh_copy_42"  [label="()", style=solid];
+"492 features_5_5_attn_lifted_tensor_43" -> "1420 lift_fresh_copy_43"  [label="()", style=solid];
+"493 features_5_5_attn_lifted_tensor_44" -> "1424 lift_fresh_copy_44"  [label="()", style=solid];
+"494 features_5_7_attn_lifted_tensor_45" -> "1570 lift_fresh_copy_45"  [label="()", style=solid];
+"495 features_5_7_attn_lifted_tensor_46" -> "1574 lift_fresh_copy_46"  [label="()", style=solid];
+"496 features_5_7_attn_lifted_tensor_47" -> "1578 lift_fresh_copy_47"  [label="()", style=solid];
+"497 features_5_7_attn_lifted_tensor_48" -> "1582 lift_fresh_copy_48"  [label="()", style=solid];
+"498 features_5_7_attn_lifted_tensor_49" -> "1586 lift_fresh_copy_49"  [label="()", style=solid];
+"499 features_5_7_attn_lifted_tensor_50" -> "1590 lift_fresh_copy_50"  [label="()", style=solid];
+"500 features_5_7_attn_lifted_tensor_51" -> "1594 lift_fresh_copy_51"  [label="()", style=solid];
+"501 features_5_7_attn_lifted_tensor_52" -> "1598 lift_fresh_copy_52"  [label="()", style=solid];
+"502 features_5_7_attn_lifted_tensor_53" -> "1602 lift_fresh_copy_53"  [label="()", style=solid];
+"503 features_5_9_attn_lifted_tensor_54" -> "1748 lift_fresh_copy_54"  [label="()", style=solid];
+"504 features_5_9_attn_lifted_tensor_55" -> "1752 lift_fresh_copy_55"  [label="()", style=solid];
+"505 features_5_9_attn_lifted_tensor_56" -> "1756 lift_fresh_copy_56"  [label="()", style=solid];
+"506 features_5_9_attn_lifted_tensor_57" -> "1760 lift_fresh_copy_57"  [label="()", style=solid];
+"507 features_5_9_attn_lifted_tensor_58" -> "1764 lift_fresh_copy_58"  [label="()", style=solid];
+"508 features_5_9_attn_lifted_tensor_59" -> "1768 lift_fresh_copy_59"  [label="()", style=solid];
+"509 features_5_9_attn_lifted_tensor_60" -> "1772 lift_fresh_copy_60"  [label="()", style=solid];
+"510 features_5_9_attn_lifted_tensor_61" -> "1776 lift_fresh_copy_61"  [label="()", style=solid];
+"511 features_5_9_attn_lifted_tensor_62" -> "1780 lift_fresh_copy_62"  [label="()", style=solid];
+"512 features_5_11_attn_lifted_tensor_63" -> "1926 lift_fresh_copy_63"  [label="()", style=solid];
+"513 features_5_11_attn_lifted_tensor_64" -> "1930 lift_fresh_copy_64"  [label="()", style=solid];
+"514 features_5_11_attn_lifted_tensor_65" -> "1934 lift_fresh_copy_65"  [label="()", style=solid];
+"515 features_5_11_attn_lifted_tensor_66" -> "1938 lift_fresh_copy_66"  [label="()", style=solid];
+"516 features_5_11_attn_lifted_tensor_67" -> "1942 lift_fresh_copy_67"  [label="()", style=solid];
+"517 features_5_11_attn_lifted_tensor_68" -> "1946 lift_fresh_copy_68"  [label="()", style=solid];
+"518 features_5_11_attn_lifted_tensor_69" -> "1950 lift_fresh_copy_69"  [label="()", style=solid];
+"519 features_5_11_attn_lifted_tensor_70" -> "1954 lift_fresh_copy_70"  [label="()", style=solid];
+"520 features_5_11_attn_lifted_tensor_71" -> "1958 lift_fresh_copy_71"  [label="()", style=solid];
+"521 features_5_13_attn_lifted_tensor_72" -> "2104 lift_fresh_copy_72"  [label="()", style=solid];
+"522 features_5_13_attn_lifted_tensor_73" -> "2108 lift_fresh_copy_73"  [label="()", style=solid];
+"523 features_5_13_attn_lifted_tensor_74" -> "2112 lift_fresh_copy_74"  [label="()", style=solid];
+"524 features_5_13_attn_lifted_tensor_75" -> "2116 lift_fresh_copy_75"  [label="()", style=solid];
+"525 features_5_13_attn_lifted_tensor_76" -> "2120 lift_fresh_copy_76"  [label="()", style=solid];
+"526 features_5_13_attn_lifted_tensor_77" -> "2124 lift_fresh_copy_77"  [label="()", style=solid];
+"527 features_5_13_attn_lifted_tensor_78" -> "2128 lift_fresh_copy_78"  [label="()", style=solid];
+"528 features_5_13_attn_lifted_tensor_79" -> "2132 lift_fresh_copy_79"  [label="()", style=solid];
+"529 features_5_13_attn_lifted_tensor_80" -> "2136 lift_fresh_copy_80"  [label="()", style=solid];
+"530 features_5_15_attn_lifted_tensor_81" -> "2282 lift_fresh_copy_81"  [label="()", style=solid];
+"531 features_5_15_attn_lifted_tensor_82" -> "2286 lift_fresh_copy_82"  [label="()", style=solid];
+"532 features_5_15_attn_lifted_tensor_83" -> "2290 lift_fresh_copy_83"  [label="()", style=solid];
+"533 features_5_15_attn_lifted_tensor_84" -> "2294 lift_fresh_copy_84"  [label="()", style=solid];
+"534 features_5_15_attn_lifted_tensor_85" -> "2298 lift_fresh_copy_85"  [label="()", style=solid];
+"535 features_5_15_attn_lifted_tensor_86" -> "2302 lift_fresh_copy_86"  [label="()", style=solid];
+"536 features_5_15_attn_lifted_tensor_87" -> "2306 lift_fresh_copy_87"  [label="()", style=solid];
+"537 features_5_15_attn_lifted_tensor_88" -> "2310 lift_fresh_copy_88"  [label="()", style=solid];
+"538 features_5_15_attn_lifted_tensor_89" -> "2314 lift_fresh_copy_89"  [label="()", style=solid];
+"539 features_5_17_attn_lifted_tensor_90" -> "2460 lift_fresh_copy_90"  [label="()", style=solid];
+"540 features_5_17_attn_lifted_tensor_91" -> "2464 lift_fresh_copy_91"  [label="()", style=solid];
+"541 features_5_17_attn_lifted_tensor_92" -> "2468 lift_fresh_copy_92"  [label="()", style=solid];
+"542 features_5_17_attn_lifted_tensor_93" -> "2472 lift_fresh_copy_93"  [label="()", style=solid];
+"543 features_5_17_attn_lifted_tensor_94" -> "2476 lift_fresh_copy_94"  [label="()", style=solid];
+"544 features_5_17_attn_lifted_tensor_95" -> "2480 lift_fresh_copy_95"  [label="()", style=solid];
+"545 features_5_17_attn_lifted_tensor_96" -> "2484 lift_fresh_copy_96"  [label="()", style=solid];
+"546 features_5_17_attn_lifted_tensor_97" -> "2488 lift_fresh_copy_97"  [label="()", style=solid];
+"547 features_5_17_attn_lifted_tensor_98" -> "2492 lift_fresh_copy_98"  [label="()", style=solid];
+"548 x" -> "549 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"549 conv2d" -> "550 permute"  [label="(1, 96, 56, 56)", style=solid];
+"550 permute" -> "551 layer_norm"  [label="(1, 56, 56, 96)", style=solid];
+"551 layer_norm" -> "563 pad"  [label="(1, 56, 56, 96)", style=solid];
+"551 layer_norm" -> "603 add_1"  [label="(1, 56, 56, 96)", style=solid];
+"552 linear" -> "553 relu_"  [label="(1, 15, 15, 512)", style=solid];
+"553 relu_" -> "554 linear_1"  [label="(1, 15, 15, 512)", style=solid];
+"554 linear_1" -> "555 view"  [label="(1, 15, 15, 3)", style=solid];
+"555 view" -> "556 index"  [label="(225, 3)", style=solid];
+"556 index" -> "557 view_1"  [label="(4096, 3)", style=solid];
+"557 view_1" -> "558 permute_1"  [label="(64, 64, 3)", style=solid];
+"558 permute_1" -> "559 contiguous"  [label="(3, 64, 64)", style=solid];
+"559 contiguous" -> "560 unsqueeze"  [label="(3, 64, 64)", style=solid];
+"560 unsqueeze" -> "561 sigmoid"  [label="(1, 3, 64, 64)", style=solid];
+"561 sigmoid" -> "562 mul"  [label="(1, 3, 64, 64)", style=solid];
+"562 mul" -> "589 add"  [label="(1, 3, 64, 64)", style=solid];
+"563 pad" -> "564 view_2"  [label="(1, 56, 56, 96)", style=solid];
+"564 view_2" -> "565 permute_2"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"565 permute_2" -> "566 reshape"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"566 reshape" -> "570 linear_2"  [label="(49, 64, 96)", style=solid];
+"567 clone" -> "568 slice_1"  [label="(288,)", style=solid];
+"567 clone" -> "570 linear_2"  [label="(288,)", style=solid];
+"568 slice_1" -> "569 zero_"  [label="(96,)", style=solid];
+"570 linear_2" -> "571 reshape_1"  [label="(49, 64, 288)", style=solid];
+"571 reshape_1" -> "572 permute_3"  [label="(49, 64, 3, 3, 32)", style=solid];
+"572 permute_3" -> "573 select"  [label="(3, 49, 3, 64, 32)", style=solid];
+"572 permute_3" -> "574 select_1"  [label="(3, 49, 3, 64, 32)", style=solid];
+"572 permute_3" -> "575 select_2"  [label="(3, 49, 3, 64, 32)", style=solid];
+"573 select" -> "576 linalg_vector_norm"  [label="(49, 3, 64, 32)", style=solid];
+"573 select" -> "578 expand_as"  [label="(49, 3, 64, 32)", style=solid];
+"573 select" -> "579 div"  [label="(49, 3, 64, 32)", style=solid];
+"574 select_1" -> "580 linalg_vector_norm_1"  [label="(49, 3, 64, 32)", style=solid];
+"574 select_1" -> "582 expand_as_1"  [label="(49, 3, 64, 32)", style=solid];
+"574 select_1" -> "583 div_1"  [label="(49, 3, 64, 32)", style=solid];
+"575 select_2" -> "592 matmul_1"  [label="(49, 3, 64, 32)", style=solid];
+"576 linalg_vector_norm" -> "577 clamp_min"  [label="(49, 3, 64, 1)", style=solid];
+"577 clamp_min" -> "578 expand_as"  [label="(49, 3, 64, 1)", style=solid];
+"578 expand_as" -> "579 div"  [label="(49, 3, 64, 32)", style=solid];
+"579 div" -> "585 matmul"  [label="(49, 3, 64, 32)", style=solid];
+"580 linalg_vector_norm_1" -> "581 clamp_min_1"  [label="(49, 3, 64, 1)", style=solid];
+"581 clamp_min_1" -> "582 expand_as_1"  [label="(49, 3, 64, 1)", style=solid];
+"582 expand_as_1" -> "583 div_1"  [label="(49, 3, 64, 32)", style=solid];
+"583 div_1" -> "584 transpose"  [label="(49, 3, 64, 32)", style=solid];
+"584 transpose" -> "585 matmul"  [label="(49, 3, 32, 64)", style=solid];
+"585 matmul" -> "588 mul_1"  [label="(49, 3, 64, 64)", style=solid];
+"586 clamp" -> "587 exp"  [label="(3, 1, 1)", style=solid];
+"587 exp" -> "588 mul_1"  [label="(3, 1, 1)", style=solid];
+"588 mul_1" -> "589 add"  [label="(49, 3, 64, 64)", style=solid];
+"589 add" -> "590 softmax"  [label="(49, 3, 64, 64)", style=solid];
+"590 softmax" -> "591 dropout"  [label="(49, 3, 64, 64)", style=solid];
+"591 dropout" -> "592 matmul_1"  [label="(49, 3, 64, 64)", style=solid];
+"592 matmul_1" -> "593 transpose_1"  [label="(49, 3, 64, 32)", style=solid];
+"593 transpose_1" -> "594 reshape_2"  [label="(49, 64, 3, 32)", style=solid];
+"594 reshape_2" -> "595 linear_3"  [label="(49, 64, 96)", style=solid];
+"595 linear_3" -> "596 dropout_1"  [label="(49, 64, 96)", style=solid];
+"596 dropout_1" -> "597 view_3"  [label="(49, 64, 96)", style=solid];
+"597 view_3" -> "598 permute_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"598 permute_4" -> "599 reshape_3"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"599 reshape_3" -> "600 slice_2"  [label="(1, 56, 56, 96)", style=solid];
+"600 slice_2" -> "601 slice_3"  [label="(1, 56, 56, 96)", style=solid];
+"601 slice_3" -> "602 layer_norm_1"  [label="(1, 56, 56, 96)", style=solid];
+"602 layer_norm_1" -> "603 add_1"  [label="(1, 56, 56, 96)", style=solid];
+"603 add_1" -> "604 linear_4"  [label="(1, 56, 56, 96)", style=solid];
+"603 add_1" -> "610 add_2"  [label="(1, 56, 56, 96)", style=solid];
+"604 linear_4" -> "605 gelu"  [label="(1, 56, 56, 384)", style=solid];
+"605 gelu" -> "606 dropout_2"  [label="(1, 56, 56, 384)", style=solid];
+"606 dropout_2" -> "607 linear_5"  [label="(1, 56, 56, 384)", style=solid];
+"607 linear_5" -> "608 dropout_3"  [label="(1, 56, 56, 96)", style=solid];
+"608 dropout_3" -> "609 layer_norm_2"  [label="(1, 56, 56, 96)", style=solid];
+"609 layer_norm_2" -> "610 add_2"  [label="(1, 56, 56, 96)", style=solid];
+"610 add_2" -> "622 pad_1"  [label="(1, 56, 56, 96)", style=solid];
+"610 add_2" -> "716 add_5"  [label="(1, 56, 56, 96)", style=solid];
+"611 linear_6" -> "612 relu__1"  [label="(1, 15, 15, 512)", style=solid];
+"612 relu__1" -> "613 linear_7"  [label="(1, 15, 15, 512)", style=solid];
+"613 linear_7" -> "614 view_4"  [label="(1, 15, 15, 3)", style=solid];
+"614 view_4" -> "615 index_1"  [label="(225, 3)", style=solid];
+"615 index_1" -> "616 view_5"  [label="(4096, 3)", style=solid];
+"616 view_5" -> "617 permute_5"  [label="(64, 64, 3)", style=solid];
+"617 permute_5" -> "618 contiguous_1"  [label="(3, 64, 64)", style=solid];
+"618 contiguous_1" -> "619 unsqueeze_1"  [label="(3, 64, 64)", style=solid];
+"619 unsqueeze_1" -> "620 sigmoid_1"  [label="(1, 3, 64, 64)", style=solid];
+"620 sigmoid_1" -> "621 mul_2"  [label="(1, 3, 64, 64)", style=solid];
+"621 mul_2" -> "649 add_3"  [label="(1, 3, 64, 64)", style=solid];
+"622 pad_1" -> "623 roll"  [label="(1, 56, 56, 96)", style=solid];
+"623 roll" -> "624 view_6"  [label="(1, 56, 56, 96)", style=solid];
+"624 view_6" -> "625 permute_6"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"625 permute_6" -> "626 reshape_4"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"626 reshape_4" -> "630 linear_8"  [label="(49, 64, 96)", style=solid];
+"626 reshape_4" -> "650 new_zeros"  [label="(49, 64, 96)", style=solid];
+"627 clone_1" -> "628 slice_4"  [label="(288,)", style=solid];
+"627 clone_1" -> "630 linear_8"  [label="(288,)", style=solid];
+"628 slice_4" -> "629 zero__1"  [label="(96,)", style=solid];
+"630 linear_8" -> "631 reshape_5"  [label="(49, 64, 288)", style=solid];
+"631 reshape_5" -> "632 permute_7"  [label="(49, 64, 3, 3, 32)", style=solid];
+"632 permute_7" -> "633 select_3"  [label="(3, 49, 3, 64, 32)", style=solid];
+"632 permute_7" -> "634 select_4"  [label="(3, 49, 3, 64, 32)", style=solid];
+"632 permute_7" -> "635 select_5"  [label="(3, 49, 3, 64, 32)", style=solid];
+"633 select_3" -> "636 linalg_vector_norm_2"  [label="(49, 3, 64, 32)", style=solid];
+"633 select_3" -> "638 expand_as_2"  [label="(49, 3, 64, 32)", style=solid];
+"633 select_3" -> "639 div_2"  [label="(49, 3, 64, 32)", style=solid];
+"634 select_4" -> "640 linalg_vector_norm_3"  [label="(49, 3, 64, 32)", style=solid];
+"634 select_4" -> "642 expand_as_3"  [label="(49, 3, 64, 32)", style=solid];
+"634 select_4" -> "643 div_3"  [label="(49, 3, 64, 32)", style=solid];
+"635 select_5" -> "704 matmul_3"  [label="(49, 3, 64, 32)", style=solid];
+"636 linalg_vector_norm_2" -> "637 clamp_min_2"  [label="(49, 3, 64, 1)", style=solid];
+"637 clamp_min_2" -> "638 expand_as_2"  [label="(49, 3, 64, 1)", style=solid];
+"638 expand_as_2" -> "639 div_2"  [label="(49, 3, 64, 32)", style=solid];
+"639 div_2" -> "645 matmul_2"  [label="(49, 3, 64, 32)", style=solid];
+"640 linalg_vector_norm_3" -> "641 clamp_min_3"  [label="(49, 3, 64, 1)", style=solid];
+"641 clamp_min_3" -> "642 expand_as_3"  [label="(49, 3, 64, 1)", style=solid];
+"642 expand_as_3" -> "643 div_3"  [label="(49, 3, 64, 32)", style=solid];
+"643 div_3" -> "644 transpose_2"  [label="(49, 3, 64, 32)", style=solid];
+"644 transpose_2" -> "645 matmul_2"  [label="(49, 3, 32, 64)", style=solid];
+"645 matmul_2" -> "648 mul_3"  [label="(49, 3, 64, 64)", style=solid];
+"646 clamp_1" -> "647 exp_1"  [label="(3, 1, 1)", style=solid];
+"647 exp_1" -> "648 mul_3"  [label="(3, 1, 1)", style=solid];
+"648 mul_3" -> "649 add_3"  [label="(49, 3, 64, 64)", style=solid];
+"649 add_3" -> "697 view_8"  [label="(49, 3, 64, 64)", style=solid];
+"650 new_zeros" -> "652 slice_5"  [label="(56, 56)", style=solid];
+"650 new_zeros" -> "656 slice_7"  [label="(56, 56)", style=solid];
+"650 new_zeros" -> "660 slice_9"  [label="(56, 56)", style=solid];
+"650 new_zeros" -> "664 slice_11"  [label="(56, 56)", style=solid];
+"650 new_zeros" -> "668 slice_13"  [label="(56, 56)", style=solid];
+"650 new_zeros" -> "672 slice_15"  [label="(56, 56)", style=solid];
+"650 new_zeros" -> "676 slice_17"  [label="(56, 56)", style=solid];
+"650 new_zeros" -> "680 slice_19"  [label="(56, 56)", style=solid];
+"650 new_zeros" -> "684 slice_21"  [label="(56, 56)", style=solid];
+"650 new_zeros" -> "687 view_7"  [label="(56, 56)", style=solid];
+"651 lift_fresh_copy" -> "654 fill_"  [label="()", style=solid];
+"652 slice_5" -> "653 slice_6"  [label="(48, 56)", style=solid];
+"653 slice_6" -> "654 fill_"  [label="(48, 48)", style=solid];
+"655 lift_fresh_copy_1" -> "658 fill__1"  [label="()", style=solid];
+"656 slice_7" -> "657 slice_8"  [label="(48, 56)", style=solid];
+"657 slice_8" -> "658 fill__1"  [label="(48, 4)", style=solid];
+"659 lift_fresh_copy_2" -> "662 fill__2"  [label="()", style=solid];
+"660 slice_9" -> "661 slice_10"  [label="(48, 56)", style=solid];
+"661 slice_10" -> "662 fill__2"  [label="(48, 4)", style=solid];
+"663 lift_fresh_copy_3" -> "666 fill__3"  [label="()", style=solid];
+"664 slice_11" -> "665 slice_12"  [label="(4, 56)", style=solid];
+"665 slice_12" -> "666 fill__3"  [label="(4, 48)", style=solid];
+"667 lift_fresh_copy_4" -> "670 fill__4"  [label="()", style=solid];
+"668 slice_13" -> "669 slice_14"  [label="(4, 56)", style=solid];
+"669 slice_14" -> "670 fill__4"  [label="(4, 4)", style=solid];
+"671 lift_fresh_copy_5" -> "674 fill__5"  [label="()", style=solid];
+"672 slice_15" -> "673 slice_16"  [label="(4, 56)", style=solid];
+"673 slice_16" -> "674 fill__5"  [label="(4, 4)", style=solid];
+"675 lift_fresh_copy_6" -> "678 fill__6"  [label="()", style=solid];
+"676 slice_17" -> "677 slice_18"  [label="(4, 56)", style=solid];
+"677 slice_18" -> "678 fill__6"  [label="(4, 48)", style=solid];
+"679 lift_fresh_copy_7" -> "682 fill__7"  [label="()", style=solid];
+"680 slice_19" -> "681 slice_20"  [label="(4, 56)", style=solid];
+"681 slice_20" -> "682 fill__7"  [label="(4, 4)", style=solid];
+"683 lift_fresh_copy_8" -> "686 fill__8"  [label="()", style=solid];
+"684 slice_21" -> "685 slice_22"  [label="(4, 56)", style=solid];
+"685 slice_22" -> "686 fill__8"  [label="(4, 4)", style=solid];
+"687 view_7" -> "688 permute_8"  [label="(7, 8, 7, 8)", style=solid];
+"688 permute_8" -> "689 reshape_6"  [label="(7, 7, 8, 8)", style=solid];
+"689 reshape_6" -> "690 unsqueeze_2"  [label="(49, 64)", style=solid];
+"689 reshape_6" -> "691 unsqueeze_3"  [label="(49, 64)", style=solid];
+"690 unsqueeze_2" -> "692 sub"  [label="(49, 1, 64)", style=solid];
+"691 unsqueeze_3" -> "692 sub"  [label="(49, 64, 1)", style=solid];
+"692 sub" -> "693 ne"  [label="(49, 64, 64)", style=solid];
+"692 sub" -> "694 masked_fill"  [label="(49, 64, 64)", style=solid];
+"692 sub" -> "695 eq"  [label="(49, 64, 64)", style=solid];
+"693 ne" -> "694 masked_fill"  [label="(49, 64, 64)", style=solid];
+"694 masked_fill" -> "696 masked_fill_1"  [label="(49, 64, 64)", style=solid];
+"695 eq" -> "696 masked_fill_1"  [label="(49, 64, 64)", style=solid];
+"696 masked_fill_1" -> "698 unsqueeze_4"  [label="(49, 64, 64)", style=solid];
+"697 view_8" -> "700 add_4"  [label="(1, 49, 3, 64, 64)", style=solid];
+"698 unsqueeze_4" -> "699 unsqueeze_5"  [label="(49, 1, 64, 64)", style=solid];
+"699 unsqueeze_5" -> "700 add_4"  [label="(1, 49, 1, 64, 64)", style=solid];
+"700 add_4" -> "701 view_9"  [label="(1, 49, 3, 64, 64)", style=solid];
+"701 view_9" -> "702 softmax_1"  [label="(49, 3, 64, 64)", style=solid];
+"702 softmax_1" -> "703 dropout_4"  [label="(49, 3, 64, 64)", style=solid];
+"703 dropout_4" -> "704 matmul_3"  [label="(49, 3, 64, 64)", style=solid];
+"704 matmul_3" -> "705 transpose_3"  [label="(49, 3, 64, 32)", style=solid];
+"705 transpose_3" -> "706 reshape_7"  [label="(49, 64, 3, 32)", style=solid];
+"706 reshape_7" -> "707 linear_9"  [label="(49, 64, 96)", style=solid];
+"707 linear_9" -> "708 dropout_5"  [label="(49, 64, 96)", style=solid];
+"708 dropout_5" -> "709 view_10"  [label="(49, 64, 96)", style=solid];
+"709 view_10" -> "710 permute_9"  [label="(1, 7, 7, 8, 8, 96)", style=solid];
+"710 permute_9" -> "711 reshape_8"  [label="(1, 7, 8, 7, 8, 96)", style=solid];
+"711 reshape_8" -> "712 roll_1"  [label="(1, 56, 56, 96)", style=solid];
+"712 roll_1" -> "713 slice_23"  [label="(1, 56, 56, 96)", style=solid];
+"713 slice_23" -> "714 slice_24"  [label="(1, 56, 56, 96)", style=solid];
+"714 slice_24" -> "715 layer_norm_3"  [label="(1, 56, 56, 96)", style=solid];
+"715 layer_norm_3" -> "716 add_5"  [label="(1, 56, 56, 96)", style=solid];
+"716 add_5" -> "717 linear_10"  [label="(1, 56, 56, 96)", style=solid];
+"716 add_5" -> "723 add_6"  [label="(1, 56, 56, 96)", style=solid];
+"717 linear_10" -> "718 gelu_1"  [label="(1, 56, 56, 384)", style=solid];
+"718 gelu_1" -> "719 dropout_6"  [label="(1, 56, 56, 384)", style=solid];
+"719 dropout_6" -> "720 linear_11"  [label="(1, 56, 56, 384)", style=solid];
+"720 linear_11" -> "721 dropout_7"  [label="(1, 56, 56, 96)", style=solid];
+"721 dropout_7" -> "722 layer_norm_4"  [label="(1, 56, 56, 96)", style=solid];
+"722 layer_norm_4" -> "723 add_6"  [label="(1, 56, 56, 96)", style=solid];
+"723 add_6" -> "724 pad_2"  [label="(1, 56, 56, 96)", style=solid];
+"724 pad_2" -> "725 slice_25"  [label="(1, 56, 56, 96)", style=solid];
+"724 pad_2" -> "728 slice_28"  [label="(1, 56, 56, 96)", style=solid];
+"724 pad_2" -> "731 slice_31"  [label="(1, 56, 56, 96)", style=solid];
+"724 pad_2" -> "734 slice_34"  [label="(1, 56, 56, 96)", style=solid];
+"725 slice_25" -> "726 slice_26"  [label="(1, 28, 56, 96)", style=solid];
+"726 slice_26" -> "727 slice_27"  [label="(1, 28, 28, 96)", style=solid];
+"727 slice_27" -> "737 cat"  [label="(1, 28, 28, 96)", style=solid];
+"728 slice_28" -> "729 slice_29"  [label="(1, 28, 56, 96)", style=solid];
+"729 slice_29" -> "730 slice_30"  [label="(1, 28, 28, 96)", style=solid];
+"730 slice_30" -> "737 cat"  [label="(1, 28, 28, 96)", style=solid];
+"731 slice_31" -> "732 slice_32"  [label="(1, 28, 56, 96)", style=solid];
+"732 slice_32" -> "733 slice_33"  [label="(1, 28, 28, 96)", style=solid];
+"733 slice_33" -> "737 cat"  [label="(1, 28, 28, 96)", style=solid];
+"734 slice_34" -> "735 slice_35"  [label="(1, 28, 56, 96)", style=solid];
+"735 slice_35" -> "736 slice_36"  [label="(1, 28, 28, 96)", style=solid];
+"736 slice_36" -> "737 cat"  [label="(1, 28, 28, 96)", style=solid];
+"737 cat" -> "738 linear_12"  [label="(1, 28, 28, 384)", style=solid];
+"738 linear_12" -> "739 layer_norm_5"  [label="(1, 28, 28, 192)", style=solid];
+"739 layer_norm_5" -> "751 pad_3"  [label="(1, 28, 28, 192)", style=solid];
+"739 layer_norm_5" -> "794 add_8"  [label="(1, 28, 28, 192)", style=solid];
+"740 linear_13" -> "741 relu__2"  [label="(1, 15, 15, 512)", style=solid];
+"741 relu__2" -> "742 linear_14"  [label="(1, 15, 15, 512)", style=solid];
+"742 linear_14" -> "743 view_11"  [label="(1, 15, 15, 6)", style=solid];
+"743 view_11" -> "744 index_2"  [label="(225, 6)", style=solid];
+"744 index_2" -> "745 view_12"  [label="(4096, 6)", style=solid];
+"745 view_12" -> "746 permute_10"  [label="(64, 64, 6)", style=solid];
+"746 permute_10" -> "747 contiguous_2"  [label="(6, 64, 64)", style=solid];
+"747 contiguous_2" -> "748 unsqueeze_6"  [label="(6, 64, 64)", style=solid];
+"748 unsqueeze_6" -> "749 sigmoid_2"  [label="(1, 6, 64, 64)", style=solid];
+"749 sigmoid_2" -> "750 mul_4"  [label="(1, 6, 64, 64)", style=solid];
+"750 mul_4" -> "777 add_7"  [label="(1, 6, 64, 64)", style=solid];
+"751 pad_3" -> "752 view_13"  [label="(1, 32, 32, 192)", style=solid];
+"752 view_13" -> "753 permute_11"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"753 permute_11" -> "754 reshape_9"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"754 reshape_9" -> "758 linear_15"  [label="(16, 64, 192)", style=solid];
+"755 clone_2" -> "756 slice_37"  [label="(576,)", style=solid];
+"755 clone_2" -> "758 linear_15"  [label="(576,)", style=solid];
+"756 slice_37" -> "757 zero__2"  [label="(192,)", style=solid];
+"758 linear_15" -> "759 reshape_10"  [label="(16, 64, 576)", style=solid];
+"759 reshape_10" -> "760 permute_12"  [label="(16, 64, 3, 6, 32)", style=solid];
+"760 permute_12" -> "761 select_6"  [label="(3, 16, 6, 64, 32)", style=solid];
+"760 permute_12" -> "762 select_7"  [label="(3, 16, 6, 64, 32)", style=solid];
+"760 permute_12" -> "763 select_8"  [label="(3, 16, 6, 64, 32)", style=solid];
+"761 select_6" -> "764 linalg_vector_norm_4"  [label="(16, 6, 64, 32)", style=solid];
+"761 select_6" -> "766 expand_as_4"  [label="(16, 6, 64, 32)", style=solid];
+"761 select_6" -> "767 div_4"  [label="(16, 6, 64, 32)", style=solid];
+"762 select_7" -> "768 linalg_vector_norm_5"  [label="(16, 6, 64, 32)", style=solid];
+"762 select_7" -> "770 expand_as_5"  [label="(16, 6, 64, 32)", style=solid];
+"762 select_7" -> "771 div_5"  [label="(16, 6, 64, 32)", style=solid];
+"763 select_8" -> "780 matmul_5"  [label="(16, 6, 64, 32)", style=solid];
+"764 linalg_vector_norm_4" -> "765 clamp_min_4"  [label="(16, 6, 64, 1)", style=solid];
+"765 clamp_min_4" -> "766 expand_as_4"  [label="(16, 6, 64, 1)", style=solid];
+"766 expand_as_4" -> "767 div_4"  [label="(16, 6, 64, 32)", style=solid];
+"767 div_4" -> "773 matmul_4"  [label="(16, 6, 64, 32)", style=solid];
+"768 linalg_vector_norm_5" -> "769 clamp_min_5"  [label="(16, 6, 64, 1)", style=solid];
+"769 clamp_min_5" -> "770 expand_as_5"  [label="(16, 6, 64, 1)", style=solid];
+"770 expand_as_5" -> "771 div_5"  [label="(16, 6, 64, 32)", style=solid];
+"771 div_5" -> "772 transpose_4"  [label="(16, 6, 64, 32)", style=solid];
+"772 transpose_4" -> "773 matmul_4"  [label="(16, 6, 32, 64)", style=solid];
+"773 matmul_4" -> "776 mul_5"  [label="(16, 6, 64, 64)", style=solid];
+"774 clamp_2" -> "775 exp_2"  [label="(6, 1, 1)", style=solid];
+"775 exp_2" -> "776 mul_5"  [label="(6, 1, 1)", style=solid];
+"776 mul_5" -> "777 add_7"  [label="(16, 6, 64, 64)", style=solid];
+"777 add_7" -> "778 softmax_2"  [label="(16, 6, 64, 64)", style=solid];
+"778 softmax_2" -> "779 dropout_8"  [label="(16, 6, 64, 64)", style=solid];
+"779 dropout_8" -> "780 matmul_5"  [label="(16, 6, 64, 64)", style=solid];
+"780 matmul_5" -> "781 transpose_5"  [label="(16, 6, 64, 32)", style=solid];
+"781 transpose_5" -> "782 reshape_11"  [label="(16, 64, 6, 32)", style=solid];
+"782 reshape_11" -> "783 linear_16"  [label="(16, 64, 192)", style=solid];
+"783 linear_16" -> "784 dropout_9"  [label="(16, 64, 192)", style=solid];
+"784 dropout_9" -> "785 view_14"  [label="(16, 64, 192)", style=solid];
+"785 view_14" -> "786 permute_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"786 permute_13" -> "787 reshape_12"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"787 reshape_12" -> "788 slice_38"  [label="(1, 32, 32, 192)", style=solid];
+"788 slice_38" -> "789 slice_39"  [label="(1, 32, 32, 192)", style=solid];
+"789 slice_39" -> "790 slice_40"  [label="(1, 28, 32, 192)", style=solid];
+"790 slice_40" -> "791 slice_41"  [label="(1, 28, 28, 192)", style=solid];
+"791 slice_41" -> "792 contiguous_3"  [label="(1, 28, 28, 192)", style=solid];
+"792 contiguous_3" -> "793 layer_norm_6"  [label="(1, 28, 28, 192)", style=solid];
+"793 layer_norm_6" -> "794 add_8"  [label="(1, 28, 28, 192)", style=solid];
+"794 add_8" -> "795 linear_17"  [label="(1, 28, 28, 192)", style=solid];
+"794 add_8" -> "801 add_9"  [label="(1, 28, 28, 192)", style=solid];
+"795 linear_17" -> "796 gelu_2"  [label="(1, 28, 28, 768)", style=solid];
+"796 gelu_2" -> "797 dropout_10"  [label="(1, 28, 28, 768)", style=solid];
+"797 dropout_10" -> "798 linear_18"  [label="(1, 28, 28, 768)", style=solid];
+"798 linear_18" -> "799 dropout_11"  [label="(1, 28, 28, 192)", style=solid];
+"799 dropout_11" -> "800 layer_norm_7"  [label="(1, 28, 28, 192)", style=solid];
+"800 layer_norm_7" -> "801 add_9"  [label="(1, 28, 28, 192)", style=solid];
+"801 add_9" -> "813 pad_4"  [label="(1, 28, 28, 192)", style=solid];
+"801 add_9" -> "910 add_12"  [label="(1, 28, 28, 192)", style=solid];
+"802 linear_19" -> "803 relu__3"  [label="(1, 15, 15, 512)", style=solid];
+"803 relu__3" -> "804 linear_20"  [label="(1, 15, 15, 512)", style=solid];
+"804 linear_20" -> "805 view_15"  [label="(1, 15, 15, 6)", style=solid];
+"805 view_15" -> "806 index_3"  [label="(225, 6)", style=solid];
+"806 index_3" -> "807 view_16"  [label="(4096, 6)", style=solid];
+"807 view_16" -> "808 permute_14"  [label="(64, 64, 6)", style=solid];
+"808 permute_14" -> "809 contiguous_4"  [label="(6, 64, 64)", style=solid];
+"809 contiguous_4" -> "810 unsqueeze_7"  [label="(6, 64, 64)", style=solid];
+"810 unsqueeze_7" -> "811 sigmoid_3"  [label="(1, 6, 64, 64)", style=solid];
+"811 sigmoid_3" -> "812 mul_6"  [label="(1, 6, 64, 64)", style=solid];
+"812 mul_6" -> "840 add_10"  [label="(1, 6, 64, 64)", style=solid];
+"813 pad_4" -> "814 roll_2"  [label="(1, 32, 32, 192)", style=solid];
+"814 roll_2" -> "815 view_17"  [label="(1, 32, 32, 192)", style=solid];
+"815 view_17" -> "816 permute_15"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"816 permute_15" -> "817 reshape_13"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"817 reshape_13" -> "821 linear_21"  [label="(16, 64, 192)", style=solid];
+"817 reshape_13" -> "841 new_zeros_1"  [label="(16, 64, 192)", style=solid];
+"818 clone_3" -> "819 slice_42"  [label="(576,)", style=solid];
+"818 clone_3" -> "821 linear_21"  [label="(576,)", style=solid];
+"819 slice_42" -> "820 zero__3"  [label="(192,)", style=solid];
+"821 linear_21" -> "822 reshape_14"  [label="(16, 64, 576)", style=solid];
+"822 reshape_14" -> "823 permute_16"  [label="(16, 64, 3, 6, 32)", style=solid];
+"823 permute_16" -> "824 select_9"  [label="(3, 16, 6, 64, 32)", style=solid];
+"823 permute_16" -> "825 select_10"  [label="(3, 16, 6, 64, 32)", style=solid];
+"823 permute_16" -> "826 select_11"  [label="(3, 16, 6, 64, 32)", style=solid];
+"824 select_9" -> "827 linalg_vector_norm_6"  [label="(16, 6, 64, 32)", style=solid];
+"824 select_9" -> "829 expand_as_6"  [label="(16, 6, 64, 32)", style=solid];
+"824 select_9" -> "830 div_6"  [label="(16, 6, 64, 32)", style=solid];
+"825 select_10" -> "831 linalg_vector_norm_7"  [label="(16, 6, 64, 32)", style=solid];
+"825 select_10" -> "833 expand_as_7"  [label="(16, 6, 64, 32)", style=solid];
+"825 select_10" -> "834 div_7"  [label="(16, 6, 64, 32)", style=solid];
+"826 select_11" -> "895 matmul_7"  [label="(16, 6, 64, 32)", style=solid];
+"827 linalg_vector_norm_6" -> "828 clamp_min_6"  [label="(16, 6, 64, 1)", style=solid];
+"828 clamp_min_6" -> "829 expand_as_6"  [label="(16, 6, 64, 1)", style=solid];
+"829 expand_as_6" -> "830 div_6"  [label="(16, 6, 64, 32)", style=solid];
+"830 div_6" -> "836 matmul_6"  [label="(16, 6, 64, 32)", style=solid];
+"831 linalg_vector_norm_7" -> "832 clamp_min_7"  [label="(16, 6, 64, 1)", style=solid];
+"832 clamp_min_7" -> "833 expand_as_7"  [label="(16, 6, 64, 1)", style=solid];
+"833 expand_as_7" -> "834 div_7"  [label="(16, 6, 64, 32)", style=solid];
+"834 div_7" -> "835 transpose_6"  [label="(16, 6, 64, 32)", style=solid];
+"835 transpose_6" -> "836 matmul_6"  [label="(16, 6, 32, 64)", style=solid];
+"836 matmul_6" -> "839 mul_7"  [label="(16, 6, 64, 64)", style=solid];
+"837 clamp_3" -> "838 exp_3"  [label="(6, 1, 1)", style=solid];
+"838 exp_3" -> "839 mul_7"  [label="(6, 1, 1)", style=solid];
+"839 mul_7" -> "840 add_10"  [label="(16, 6, 64, 64)", style=solid];
+"840 add_10" -> "888 view_19"  [label="(16, 6, 64, 64)", style=solid];
+"841 new_zeros_1" -> "843 slice_43"  [label="(32, 32)", style=solid];
+"841 new_zeros_1" -> "847 slice_45"  [label="(32, 32)", style=solid];
+"841 new_zeros_1" -> "851 slice_47"  [label="(32, 32)", style=solid];
+"841 new_zeros_1" -> "855 slice_49"  [label="(32, 32)", style=solid];
+"841 new_zeros_1" -> "859 slice_51"  [label="(32, 32)", style=solid];
+"841 new_zeros_1" -> "863 slice_53"  [label="(32, 32)", style=solid];
+"841 new_zeros_1" -> "867 slice_55"  [label="(32, 32)", style=solid];
+"841 new_zeros_1" -> "871 slice_57"  [label="(32, 32)", style=solid];
+"841 new_zeros_1" -> "875 slice_59"  [label="(32, 32)", style=solid];
+"841 new_zeros_1" -> "878 view_18"  [label="(32, 32)", style=solid];
+"842 lift_fresh_copy_9" -> "845 fill__9"  [label="()", style=solid];
+"843 slice_43" -> "844 slice_44"  [label="(24, 32)", style=solid];
+"844 slice_44" -> "845 fill__9"  [label="(24, 24)", style=solid];
+"846 lift_fresh_copy_10" -> "849 fill__10"  [label="()", style=solid];
+"847 slice_45" -> "848 slice_46"  [label="(24, 32)", style=solid];
+"848 slice_46" -> "849 fill__10"  [label="(24, 4)", style=solid];
+"850 lift_fresh_copy_11" -> "853 fill__11"  [label="()", style=solid];
+"851 slice_47" -> "852 slice_48"  [label="(24, 32)", style=solid];
+"852 slice_48" -> "853 fill__11"  [label="(24, 4)", style=solid];
+"854 lift_fresh_copy_12" -> "857 fill__12"  [label="()", style=solid];
+"855 slice_49" -> "856 slice_50"  [label="(4, 32)", style=solid];
+"856 slice_50" -> "857 fill__12"  [label="(4, 24)", style=solid];
+"858 lift_fresh_copy_13" -> "861 fill__13"  [label="()", style=solid];
+"859 slice_51" -> "860 slice_52"  [label="(4, 32)", style=solid];
+"860 slice_52" -> "861 fill__13"  [label="(4, 4)", style=solid];
+"862 lift_fresh_copy_14" -> "865 fill__14"  [label="()", style=solid];
+"863 slice_53" -> "864 slice_54"  [label="(4, 32)", style=solid];
+"864 slice_54" -> "865 fill__14"  [label="(4, 4)", style=solid];
+"866 lift_fresh_copy_15" -> "869 fill__15"  [label="()", style=solid];
+"867 slice_55" -> "868 slice_56"  [label="(4, 32)", style=solid];
+"868 slice_56" -> "869 fill__15"  [label="(4, 24)", style=solid];
+"870 lift_fresh_copy_16" -> "873 fill__16"  [label="()", style=solid];
+"871 slice_57" -> "872 slice_58"  [label="(4, 32)", style=solid];
+"872 slice_58" -> "873 fill__16"  [label="(4, 4)", style=solid];
+"874 lift_fresh_copy_17" -> "877 fill__17"  [label="()", style=solid];
+"875 slice_59" -> "876 slice_60"  [label="(4, 32)", style=solid];
+"876 slice_60" -> "877 fill__17"  [label="(4, 4)", style=solid];
+"878 view_18" -> "879 permute_17"  [label="(4, 8, 4, 8)", style=solid];
+"879 permute_17" -> "880 reshape_15"  [label="(4, 4, 8, 8)", style=solid];
+"880 reshape_15" -> "881 unsqueeze_8"  [label="(16, 64)", style=solid];
+"880 reshape_15" -> "882 unsqueeze_9"  [label="(16, 64)", style=solid];
+"881 unsqueeze_8" -> "883 sub_1"  [label="(16, 1, 64)", style=solid];
+"882 unsqueeze_9" -> "883 sub_1"  [label="(16, 64, 1)", style=solid];
+"883 sub_1" -> "884 ne_1"  [label="(16, 64, 64)", style=solid];
+"883 sub_1" -> "885 masked_fill_2"  [label="(16, 64, 64)", style=solid];
+"883 sub_1" -> "886 eq_1"  [label="(16, 64, 64)", style=solid];
+"884 ne_1" -> "885 masked_fill_2"  [label="(16, 64, 64)", style=solid];
+"885 masked_fill_2" -> "887 masked_fill_3"  [label="(16, 64, 64)", style=solid];
+"886 eq_1" -> "887 masked_fill_3"  [label="(16, 64, 64)", style=solid];
+"887 masked_fill_3" -> "889 unsqueeze_10"  [label="(16, 64, 64)", style=solid];
+"888 view_19" -> "891 add_11"  [label="(1, 16, 6, 64, 64)", style=solid];
+"889 unsqueeze_10" -> "890 unsqueeze_11"  [label="(16, 1, 64, 64)", style=solid];
+"890 unsqueeze_11" -> "891 add_11"  [label="(1, 16, 1, 64, 64)", style=solid];
+"891 add_11" -> "892 view_20"  [label="(1, 16, 6, 64, 64)", style=solid];
+"892 view_20" -> "893 softmax_3"  [label="(16, 6, 64, 64)", style=solid];
+"893 softmax_3" -> "894 dropout_12"  [label="(16, 6, 64, 64)", style=solid];
+"894 dropout_12" -> "895 matmul_7"  [label="(16, 6, 64, 64)", style=solid];
+"895 matmul_7" -> "896 transpose_7"  [label="(16, 6, 64, 32)", style=solid];
+"896 transpose_7" -> "897 reshape_16"  [label="(16, 64, 6, 32)", style=solid];
+"897 reshape_16" -> "898 linear_22"  [label="(16, 64, 192)", style=solid];
+"898 linear_22" -> "899 dropout_13"  [label="(16, 64, 192)", style=solid];
+"899 dropout_13" -> "900 view_21"  [label="(16, 64, 192)", style=solid];
+"900 view_21" -> "901 permute_18"  [label="(1, 4, 4, 8, 8, 192)", style=solid];
+"901 permute_18" -> "902 reshape_17"  [label="(1, 4, 8, 4, 8, 192)", style=solid];
+"902 reshape_17" -> "903 roll_3"  [label="(1, 32, 32, 192)", style=solid];
+"903 roll_3" -> "904 slice_61"  [label="(1, 32, 32, 192)", style=solid];
+"904 slice_61" -> "905 slice_62"  [label="(1, 32, 32, 192)", style=solid];
+"905 slice_62" -> "906 slice_63"  [label="(1, 28, 32, 192)", style=solid];
+"906 slice_63" -> "907 slice_64"  [label="(1, 28, 28, 192)", style=solid];
+"907 slice_64" -> "908 contiguous_5"  [label="(1, 28, 28, 192)", style=solid];
+"908 contiguous_5" -> "909 layer_norm_8"  [label="(1, 28, 28, 192)", style=solid];
+"909 layer_norm_8" -> "910 add_12"  [label="(1, 28, 28, 192)", style=solid];
+"910 add_12" -> "911 linear_23"  [label="(1, 28, 28, 192)", style=solid];
+"910 add_12" -> "917 add_13"  [label="(1, 28, 28, 192)", style=solid];
+"911 linear_23" -> "912 gelu_3"  [label="(1, 28, 28, 768)", style=solid];
+"912 gelu_3" -> "913 dropout_14"  [label="(1, 28, 28, 768)", style=solid];
+"913 dropout_14" -> "914 linear_24"  [label="(1, 28, 28, 768)", style=solid];
+"914 linear_24" -> "915 dropout_15"  [label="(1, 28, 28, 192)", style=solid];
+"915 dropout_15" -> "916 layer_norm_9"  [label="(1, 28, 28, 192)", style=solid];
+"916 layer_norm_9" -> "917 add_13"  [label="(1, 28, 28, 192)", style=solid];
+"917 add_13" -> "918 pad_5"  [label="(1, 28, 28, 192)", style=solid];
+"918 pad_5" -> "919 slice_65"  [label="(1, 28, 28, 192)", style=solid];
+"918 pad_5" -> "922 slice_68"  [label="(1, 28, 28, 192)", style=solid];
+"918 pad_5" -> "925 slice_71"  [label="(1, 28, 28, 192)", style=solid];
+"918 pad_5" -> "928 slice_74"  [label="(1, 28, 28, 192)", style=solid];
+"919 slice_65" -> "920 slice_66"  [label="(1, 14, 28, 192)", style=solid];
+"920 slice_66" -> "921 slice_67"  [label="(1, 14, 14, 192)", style=solid];
+"921 slice_67" -> "931 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"922 slice_68" -> "923 slice_69"  [label="(1, 14, 28, 192)", style=solid];
+"923 slice_69" -> "924 slice_70"  [label="(1, 14, 14, 192)", style=solid];
+"924 slice_70" -> "931 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"925 slice_71" -> "926 slice_72"  [label="(1, 14, 28, 192)", style=solid];
+"926 slice_72" -> "927 slice_73"  [label="(1, 14, 14, 192)", style=solid];
+"927 slice_73" -> "931 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"928 slice_74" -> "929 slice_75"  [label="(1, 14, 28, 192)", style=solid];
+"929 slice_75" -> "930 slice_76"  [label="(1, 14, 14, 192)", style=solid];
+"930 slice_76" -> "931 cat_1"  [label="(1, 14, 14, 192)", style=solid];
+"931 cat_1" -> "932 linear_25"  [label="(1, 14, 14, 768)", style=solid];
+"932 linear_25" -> "933 layer_norm_10"  [label="(1, 14, 14, 384)", style=solid];
+"933 layer_norm_10" -> "945 pad_6"  [label="(1, 14, 14, 384)", style=solid];
+"933 layer_norm_10" -> "988 add_15"  [label="(1, 14, 14, 384)", style=solid];
+"934 linear_26" -> "935 relu__4"  [label="(1, 15, 15, 512)", style=solid];
+"935 relu__4" -> "936 linear_27"  [label="(1, 15, 15, 512)", style=solid];
+"936 linear_27" -> "937 view_22"  [label="(1, 15, 15, 12)", style=solid];
+"937 view_22" -> "938 index_4"  [label="(225, 12)", style=solid];
+"938 index_4" -> "939 view_23"  [label="(4096, 12)", style=solid];
+"939 view_23" -> "940 permute_19"  [label="(64, 64, 12)", style=solid];
+"940 permute_19" -> "941 contiguous_6"  [label="(12, 64, 64)", style=solid];
+"941 contiguous_6" -> "942 unsqueeze_12"  [label="(12, 64, 64)", style=solid];
+"942 unsqueeze_12" -> "943 sigmoid_4"  [label="(1, 12, 64, 64)", style=solid];
+"943 sigmoid_4" -> "944 mul_8"  [label="(1, 12, 64, 64)", style=solid];
+"944 mul_8" -> "971 add_14"  [label="(1, 12, 64, 64)", style=solid];
+"945 pad_6" -> "946 view_24"  [label="(1, 16, 16, 384)", style=solid];
+"946 view_24" -> "947 permute_20"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"947 permute_20" -> "948 reshape_18"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"948 reshape_18" -> "952 linear_28"  [label="(4, 64, 384)", style=solid];
+"949 clone_4" -> "950 slice_77"  [label="(1152,)", style=solid];
+"949 clone_4" -> "952 linear_28"  [label="(1152,)", style=solid];
+"950 slice_77" -> "951 zero__4"  [label="(384,)", style=solid];
+"952 linear_28" -> "953 reshape_19"  [label="(4, 64, 1152)", style=solid];
+"953 reshape_19" -> "954 permute_21"  [label="(4, 64, 3, 12, 32)", style=solid];
+"954 permute_21" -> "955 select_12"  [label="(3, 4, 12, 64, 32)", style=solid];
+"954 permute_21" -> "956 select_13"  [label="(3, 4, 12, 64, 32)", style=solid];
+"954 permute_21" -> "957 select_14"  [label="(3, 4, 12, 64, 32)", style=solid];
+"955 select_12" -> "958 linalg_vector_norm_8"  [label="(4, 12, 64, 32)", style=solid];
+"955 select_12" -> "960 expand_as_8"  [label="(4, 12, 64, 32)", style=solid];
+"955 select_12" -> "961 div_8"  [label="(4, 12, 64, 32)", style=solid];
+"956 select_13" -> "962 linalg_vector_norm_9"  [label="(4, 12, 64, 32)", style=solid];
+"956 select_13" -> "964 expand_as_9"  [label="(4, 12, 64, 32)", style=solid];
+"956 select_13" -> "965 div_9"  [label="(4, 12, 64, 32)", style=solid];
+"957 select_14" -> "974 matmul_9"  [label="(4, 12, 64, 32)", style=solid];
+"958 linalg_vector_norm_8" -> "959 clamp_min_8"  [label="(4, 12, 64, 1)", style=solid];
+"959 clamp_min_8" -> "960 expand_as_8"  [label="(4, 12, 64, 1)", style=solid];
+"960 expand_as_8" -> "961 div_8"  [label="(4, 12, 64, 32)", style=solid];
+"961 div_8" -> "967 matmul_8"  [label="(4, 12, 64, 32)", style=solid];
+"962 linalg_vector_norm_9" -> "963 clamp_min_9"  [label="(4, 12, 64, 1)", style=solid];
+"963 clamp_min_9" -> "964 expand_as_9"  [label="(4, 12, 64, 1)", style=solid];
+"964 expand_as_9" -> "965 div_9"  [label="(4, 12, 64, 32)", style=solid];
+"965 div_9" -> "966 transpose_8"  [label="(4, 12, 64, 32)", style=solid];
+"966 transpose_8" -> "967 matmul_8"  [label="(4, 12, 32, 64)", style=solid];
+"967 matmul_8" -> "970 mul_9"  [label="(4, 12, 64, 64)", style=solid];
+"968 clamp_4" -> "969 exp_4"  [label="(12, 1, 1)", style=solid];
+"969 exp_4" -> "970 mul_9"  [label="(12, 1, 1)", style=solid];
+"970 mul_9" -> "971 add_14"  [label="(4, 12, 64, 64)", style=solid];
+"971 add_14" -> "972 softmax_4"  [label="(4, 12, 64, 64)", style=solid];
+"972 softmax_4" -> "973 dropout_16"  [label="(4, 12, 64, 64)", style=solid];
+"973 dropout_16" -> "974 matmul_9"  [label="(4, 12, 64, 64)", style=solid];
+"974 matmul_9" -> "975 transpose_9"  [label="(4, 12, 64, 32)", style=solid];
+"975 transpose_9" -> "976 reshape_20"  [label="(4, 64, 12, 32)", style=solid];
+"976 reshape_20" -> "977 linear_29"  [label="(4, 64, 384)", style=solid];
+"977 linear_29" -> "978 dropout_17"  [label="(4, 64, 384)", style=solid];
+"978 dropout_17" -> "979 view_25"  [label="(4, 64, 384)", style=solid];
+"979 view_25" -> "980 permute_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"980 permute_22" -> "981 reshape_21"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"981 reshape_21" -> "982 slice_78"  [label="(1, 16, 16, 384)", style=solid];
+"982 slice_78" -> "983 slice_79"  [label="(1, 16, 16, 384)", style=solid];
+"983 slice_79" -> "984 slice_80"  [label="(1, 14, 16, 384)", style=solid];
+"984 slice_80" -> "985 slice_81"  [label="(1, 14, 14, 384)", style=solid];
+"985 slice_81" -> "986 contiguous_7"  [label="(1, 14, 14, 384)", style=solid];
+"986 contiguous_7" -> "987 layer_norm_11"  [label="(1, 14, 14, 384)", style=solid];
+"987 layer_norm_11" -> "988 add_15"  [label="(1, 14, 14, 384)", style=solid];
+"988 add_15" -> "989 linear_30"  [label="(1, 14, 14, 384)", style=solid];
+"988 add_15" -> "995 add_16"  [label="(1, 14, 14, 384)", style=solid];
+"989 linear_30" -> "990 gelu_4"  [label="(1, 14, 14, 1536)", style=solid];
+"990 gelu_4" -> "991 dropout_18"  [label="(1, 14, 14, 1536)", style=solid];
+"991 dropout_18" -> "992 linear_31"  [label="(1, 14, 14, 1536)", style=solid];
+"992 linear_31" -> "993 dropout_19"  [label="(1, 14, 14, 384)", style=solid];
+"993 dropout_19" -> "994 layer_norm_12"  [label="(1, 14, 14, 384)", style=solid];
+"994 layer_norm_12" -> "995 add_16"  [label="(1, 14, 14, 384)", style=solid];
+"995 add_16" -> "1007 pad_7"  [label="(1, 14, 14, 384)", style=solid];
+"995 add_16" -> "1104 add_19"  [label="(1, 14, 14, 384)", style=solid];
+"996 linear_32" -> "997 relu__5"  [label="(1, 15, 15, 512)", style=solid];
+"997 relu__5" -> "998 linear_33"  [label="(1, 15, 15, 512)", style=solid];
+"998 linear_33" -> "999 view_26"  [label="(1, 15, 15, 12)", style=solid];
+"999 view_26" -> "1000 index_5"  [label="(225, 12)", style=solid];
+"1000 index_5" -> "1001 view_27"  [label="(4096, 12)", style=solid];
+"1001 view_27" -> "1002 permute_23"  [label="(64, 64, 12)", style=solid];
+"1002 permute_23" -> "1003 contiguous_8"  [label="(12, 64, 64)", style=solid];
+"1003 contiguous_8" -> "1004 unsqueeze_13"  [label="(12, 64, 64)", style=solid];
+"1004 unsqueeze_13" -> "1005 sigmoid_5"  [label="(1, 12, 64, 64)", style=solid];
+"1005 sigmoid_5" -> "1006 mul_10"  [label="(1, 12, 64, 64)", style=solid];
+"1006 mul_10" -> "1034 add_17"  [label="(1, 12, 64, 64)", style=solid];
+"1007 pad_7" -> "1008 roll_4"  [label="(1, 16, 16, 384)", style=solid];
+"1008 roll_4" -> "1009 view_28"  [label="(1, 16, 16, 384)", style=solid];
+"1009 view_28" -> "1010 permute_24"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1010 permute_24" -> "1011 reshape_22"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1011 reshape_22" -> "1015 linear_34"  [label="(4, 64, 384)", style=solid];
+"1011 reshape_22" -> "1035 new_zeros_2"  [label="(4, 64, 384)", style=solid];
+"1012 clone_5" -> "1013 slice_82"  [label="(1152,)", style=solid];
+"1012 clone_5" -> "1015 linear_34"  [label="(1152,)", style=solid];
+"1013 slice_82" -> "1014 zero__5"  [label="(384,)", style=solid];
+"1015 linear_34" -> "1016 reshape_23"  [label="(4, 64, 1152)", style=solid];
+"1016 reshape_23" -> "1017 permute_25"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1017 permute_25" -> "1018 select_15"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1017 permute_25" -> "1019 select_16"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1017 permute_25" -> "1020 select_17"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1018 select_15" -> "1021 linalg_vector_norm_10"  [label="(4, 12, 64, 32)", style=solid];
+"1018 select_15" -> "1023 expand_as_10"  [label="(4, 12, 64, 32)", style=solid];
+"1018 select_15" -> "1024 div_10"  [label="(4, 12, 64, 32)", style=solid];
+"1019 select_16" -> "1025 linalg_vector_norm_11"  [label="(4, 12, 64, 32)", style=solid];
+"1019 select_16" -> "1027 expand_as_11"  [label="(4, 12, 64, 32)", style=solid];
+"1019 select_16" -> "1028 div_11"  [label="(4, 12, 64, 32)", style=solid];
+"1020 select_17" -> "1089 matmul_11"  [label="(4, 12, 64, 32)", style=solid];
+"1021 linalg_vector_norm_10" -> "1022 clamp_min_10"  [label="(4, 12, 64, 1)", style=solid];
+"1022 clamp_min_10" -> "1023 expand_as_10"  [label="(4, 12, 64, 1)", style=solid];
+"1023 expand_as_10" -> "1024 div_10"  [label="(4, 12, 64, 32)", style=solid];
+"1024 div_10" -> "1030 matmul_10"  [label="(4, 12, 64, 32)", style=solid];
+"1025 linalg_vector_norm_11" -> "1026 clamp_min_11"  [label="(4, 12, 64, 1)", style=solid];
+"1026 clamp_min_11" -> "1027 expand_as_11"  [label="(4, 12, 64, 1)", style=solid];
+"1027 expand_as_11" -> "1028 div_11"  [label="(4, 12, 64, 32)", style=solid];
+"1028 div_11" -> "1029 transpose_10"  [label="(4, 12, 64, 32)", style=solid];
+"1029 transpose_10" -> "1030 matmul_10"  [label="(4, 12, 32, 64)", style=solid];
+"1030 matmul_10" -> "1033 mul_11"  [label="(4, 12, 64, 64)", style=solid];
+"1031 clamp_5" -> "1032 exp_5"  [label="(12, 1, 1)", style=solid];
+"1032 exp_5" -> "1033 mul_11"  [label="(12, 1, 1)", style=solid];
+"1033 mul_11" -> "1034 add_17"  [label="(4, 12, 64, 64)", style=solid];
+"1034 add_17" -> "1082 view_30"  [label="(4, 12, 64, 64)", style=solid];
+"1035 new_zeros_2" -> "1037 slice_83"  [label="(16, 16)", style=solid];
+"1035 new_zeros_2" -> "1041 slice_85"  [label="(16, 16)", style=solid];
+"1035 new_zeros_2" -> "1045 slice_87"  [label="(16, 16)", style=solid];
+"1035 new_zeros_2" -> "1049 slice_89"  [label="(16, 16)", style=solid];
+"1035 new_zeros_2" -> "1053 slice_91"  [label="(16, 16)", style=solid];
+"1035 new_zeros_2" -> "1057 slice_93"  [label="(16, 16)", style=solid];
+"1035 new_zeros_2" -> "1061 slice_95"  [label="(16, 16)", style=solid];
+"1035 new_zeros_2" -> "1065 slice_97"  [label="(16, 16)", style=solid];
+"1035 new_zeros_2" -> "1069 slice_99"  [label="(16, 16)", style=solid];
+"1035 new_zeros_2" -> "1072 view_29"  [label="(16, 16)", style=solid];
+"1036 lift_fresh_copy_18" -> "1039 fill__18"  [label="()", style=solid];
+"1037 slice_83" -> "1038 slice_84"  [label="(8, 16)", style=solid];
+"1038 slice_84" -> "1039 fill__18"  [label="(8, 8)", style=solid];
+"1040 lift_fresh_copy_19" -> "1043 fill__19"  [label="()", style=solid];
+"1041 slice_85" -> "1042 slice_86"  [label="(8, 16)", style=solid];
+"1042 slice_86" -> "1043 fill__19"  [label="(8, 4)", style=solid];
+"1044 lift_fresh_copy_20" -> "1047 fill__20"  [label="()", style=solid];
+"1045 slice_87" -> "1046 slice_88"  [label="(8, 16)", style=solid];
+"1046 slice_88" -> "1047 fill__20"  [label="(8, 4)", style=solid];
+"1048 lift_fresh_copy_21" -> "1051 fill__21"  [label="()", style=solid];
+"1049 slice_89" -> "1050 slice_90"  [label="(4, 16)", style=solid];
+"1050 slice_90" -> "1051 fill__21"  [label="(4, 8)", style=solid];
+"1052 lift_fresh_copy_22" -> "1055 fill__22"  [label="()", style=solid];
+"1053 slice_91" -> "1054 slice_92"  [label="(4, 16)", style=solid];
+"1054 slice_92" -> "1055 fill__22"  [label="(4, 4)", style=solid];
+"1056 lift_fresh_copy_23" -> "1059 fill__23"  [label="()", style=solid];
+"1057 slice_93" -> "1058 slice_94"  [label="(4, 16)", style=solid];
+"1058 slice_94" -> "1059 fill__23"  [label="(4, 4)", style=solid];
+"1060 lift_fresh_copy_24" -> "1063 fill__24"  [label="()", style=solid];
+"1061 slice_95" -> "1062 slice_96"  [label="(4, 16)", style=solid];
+"1062 slice_96" -> "1063 fill__24"  [label="(4, 8)", style=solid];
+"1064 lift_fresh_copy_25" -> "1067 fill__25"  [label="()", style=solid];
+"1065 slice_97" -> "1066 slice_98"  [label="(4, 16)", style=solid];
+"1066 slice_98" -> "1067 fill__25"  [label="(4, 4)", style=solid];
+"1068 lift_fresh_copy_26" -> "1071 fill__26"  [label="()", style=solid];
+"1069 slice_99" -> "1070 slice_100"  [label="(4, 16)", style=solid];
+"1070 slice_100" -> "1071 fill__26"  [label="(4, 4)", style=solid];
+"1072 view_29" -> "1073 permute_26"  [label="(2, 8, 2, 8)", style=solid];
+"1073 permute_26" -> "1074 reshape_24"  [label="(2, 2, 8, 8)", style=solid];
+"1074 reshape_24" -> "1075 unsqueeze_14"  [label="(4, 64)", style=solid];
+"1074 reshape_24" -> "1076 unsqueeze_15"  [label="(4, 64)", style=solid];
+"1075 unsqueeze_14" -> "1077 sub_2"  [label="(4, 1, 64)", style=solid];
+"1076 unsqueeze_15" -> "1077 sub_2"  [label="(4, 64, 1)", style=solid];
+"1077 sub_2" -> "1078 ne_2"  [label="(4, 64, 64)", style=solid];
+"1077 sub_2" -> "1079 masked_fill_4"  [label="(4, 64, 64)", style=solid];
+"1077 sub_2" -> "1080 eq_2"  [label="(4, 64, 64)", style=solid];
+"1078 ne_2" -> "1079 masked_fill_4"  [label="(4, 64, 64)", style=solid];
+"1079 masked_fill_4" -> "1081 masked_fill_5"  [label="(4, 64, 64)", style=solid];
+"1080 eq_2" -> "1081 masked_fill_5"  [label="(4, 64, 64)", style=solid];
+"1081 masked_fill_5" -> "1083 unsqueeze_16"  [label="(4, 64, 64)", style=solid];
+"1082 view_30" -> "1085 add_18"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1083 unsqueeze_16" -> "1084 unsqueeze_17"  [label="(4, 1, 64, 64)", style=solid];
+"1084 unsqueeze_17" -> "1085 add_18"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1085 add_18" -> "1086 view_31"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1086 view_31" -> "1087 softmax_5"  [label="(4, 12, 64, 64)", style=solid];
+"1087 softmax_5" -> "1088 dropout_20"  [label="(4, 12, 64, 64)", style=solid];
+"1088 dropout_20" -> "1089 matmul_11"  [label="(4, 12, 64, 64)", style=solid];
+"1089 matmul_11" -> "1090 transpose_11"  [label="(4, 12, 64, 32)", style=solid];
+"1090 transpose_11" -> "1091 reshape_25"  [label="(4, 64, 12, 32)", style=solid];
+"1091 reshape_25" -> "1092 linear_35"  [label="(4, 64, 384)", style=solid];
+"1092 linear_35" -> "1093 dropout_21"  [label="(4, 64, 384)", style=solid];
+"1093 dropout_21" -> "1094 view_32"  [label="(4, 64, 384)", style=solid];
+"1094 view_32" -> "1095 permute_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1095 permute_27" -> "1096 reshape_26"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1096 reshape_26" -> "1097 roll_5"  [label="(1, 16, 16, 384)", style=solid];
+"1097 roll_5" -> "1098 slice_101"  [label="(1, 16, 16, 384)", style=solid];
+"1098 slice_101" -> "1099 slice_102"  [label="(1, 16, 16, 384)", style=solid];
+"1099 slice_102" -> "1100 slice_103"  [label="(1, 14, 16, 384)", style=solid];
+"1100 slice_103" -> "1101 slice_104"  [label="(1, 14, 14, 384)", style=solid];
+"1101 slice_104" -> "1102 contiguous_9"  [label="(1, 14, 14, 384)", style=solid];
+"1102 contiguous_9" -> "1103 layer_norm_13"  [label="(1, 14, 14, 384)", style=solid];
+"1103 layer_norm_13" -> "1104 add_19"  [label="(1, 14, 14, 384)", style=solid];
+"1104 add_19" -> "1105 linear_36"  [label="(1, 14, 14, 384)", style=solid];
+"1104 add_19" -> "1111 add_20"  [label="(1, 14, 14, 384)", style=solid];
+"1105 linear_36" -> "1106 gelu_5"  [label="(1, 14, 14, 1536)", style=solid];
+"1106 gelu_5" -> "1107 dropout_22"  [label="(1, 14, 14, 1536)", style=solid];
+"1107 dropout_22" -> "1108 linear_37"  [label="(1, 14, 14, 1536)", style=solid];
+"1108 linear_37" -> "1109 dropout_23"  [label="(1, 14, 14, 384)", style=solid];
+"1109 dropout_23" -> "1110 layer_norm_14"  [label="(1, 14, 14, 384)", style=solid];
+"1110 layer_norm_14" -> "1111 add_20"  [label="(1, 14, 14, 384)", style=solid];
+"1111 add_20" -> "1123 pad_8"  [label="(1, 14, 14, 384)", style=solid];
+"1111 add_20" -> "1166 add_22"  [label="(1, 14, 14, 384)", style=solid];
+"1112 linear_38" -> "1113 relu__6"  [label="(1, 15, 15, 512)", style=solid];
+"1113 relu__6" -> "1114 linear_39"  [label="(1, 15, 15, 512)", style=solid];
+"1114 linear_39" -> "1115 view_33"  [label="(1, 15, 15, 12)", style=solid];
+"1115 view_33" -> "1116 index_6"  [label="(225, 12)", style=solid];
+"1116 index_6" -> "1117 view_34"  [label="(4096, 12)", style=solid];
+"1117 view_34" -> "1118 permute_28"  [label="(64, 64, 12)", style=solid];
+"1118 permute_28" -> "1119 contiguous_10"  [label="(12, 64, 64)", style=solid];
+"1119 contiguous_10" -> "1120 unsqueeze_18"  [label="(12, 64, 64)", style=solid];
+"1120 unsqueeze_18" -> "1121 sigmoid_6"  [label="(1, 12, 64, 64)", style=solid];
+"1121 sigmoid_6" -> "1122 mul_12"  [label="(1, 12, 64, 64)", style=solid];
+"1122 mul_12" -> "1149 add_21"  [label="(1, 12, 64, 64)", style=solid];
+"1123 pad_8" -> "1124 view_35"  [label="(1, 16, 16, 384)", style=solid];
+"1124 view_35" -> "1125 permute_29"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1125 permute_29" -> "1126 reshape_27"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1126 reshape_27" -> "1130 linear_40"  [label="(4, 64, 384)", style=solid];
+"1127 clone_6" -> "1128 slice_105"  [label="(1152,)", style=solid];
+"1127 clone_6" -> "1130 linear_40"  [label="(1152,)", style=solid];
+"1128 slice_105" -> "1129 zero__6"  [label="(384,)", style=solid];
+"1130 linear_40" -> "1131 reshape_28"  [label="(4, 64, 1152)", style=solid];
+"1131 reshape_28" -> "1132 permute_30"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1132 permute_30" -> "1133 select_18"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1132 permute_30" -> "1134 select_19"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1132 permute_30" -> "1135 select_20"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1133 select_18" -> "1136 linalg_vector_norm_12"  [label="(4, 12, 64, 32)", style=solid];
+"1133 select_18" -> "1138 expand_as_12"  [label="(4, 12, 64, 32)", style=solid];
+"1133 select_18" -> "1139 div_12"  [label="(4, 12, 64, 32)", style=solid];
+"1134 select_19" -> "1140 linalg_vector_norm_13"  [label="(4, 12, 64, 32)", style=solid];
+"1134 select_19" -> "1142 expand_as_13"  [label="(4, 12, 64, 32)", style=solid];
+"1134 select_19" -> "1143 div_13"  [label="(4, 12, 64, 32)", style=solid];
+"1135 select_20" -> "1152 matmul_13"  [label="(4, 12, 64, 32)", style=solid];
+"1136 linalg_vector_norm_12" -> "1137 clamp_min_12"  [label="(4, 12, 64, 1)", style=solid];
+"1137 clamp_min_12" -> "1138 expand_as_12"  [label="(4, 12, 64, 1)", style=solid];
+"1138 expand_as_12" -> "1139 div_12"  [label="(4, 12, 64, 32)", style=solid];
+"1139 div_12" -> "1145 matmul_12"  [label="(4, 12, 64, 32)", style=solid];
+"1140 linalg_vector_norm_13" -> "1141 clamp_min_13"  [label="(4, 12, 64, 1)", style=solid];
+"1141 clamp_min_13" -> "1142 expand_as_13"  [label="(4, 12, 64, 1)", style=solid];
+"1142 expand_as_13" -> "1143 div_13"  [label="(4, 12, 64, 32)", style=solid];
+"1143 div_13" -> "1144 transpose_12"  [label="(4, 12, 64, 32)", style=solid];
+"1144 transpose_12" -> "1145 matmul_12"  [label="(4, 12, 32, 64)", style=solid];
+"1145 matmul_12" -> "1148 mul_13"  [label="(4, 12, 64, 64)", style=solid];
+"1146 clamp_6" -> "1147 exp_6"  [label="(12, 1, 1)", style=solid];
+"1147 exp_6" -> "1148 mul_13"  [label="(12, 1, 1)", style=solid];
+"1148 mul_13" -> "1149 add_21"  [label="(4, 12, 64, 64)", style=solid];
+"1149 add_21" -> "1150 softmax_6"  [label="(4, 12, 64, 64)", style=solid];
+"1150 softmax_6" -> "1151 dropout_24"  [label="(4, 12, 64, 64)", style=solid];
+"1151 dropout_24" -> "1152 matmul_13"  [label="(4, 12, 64, 64)", style=solid];
+"1152 matmul_13" -> "1153 transpose_13"  [label="(4, 12, 64, 32)", style=solid];
+"1153 transpose_13" -> "1154 reshape_29"  [label="(4, 64, 12, 32)", style=solid];
+"1154 reshape_29" -> "1155 linear_41"  [label="(4, 64, 384)", style=solid];
+"1155 linear_41" -> "1156 dropout_25"  [label="(4, 64, 384)", style=solid];
+"1156 dropout_25" -> "1157 view_36"  [label="(4, 64, 384)", style=solid];
+"1157 view_36" -> "1158 permute_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1158 permute_31" -> "1159 reshape_30"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1159 reshape_30" -> "1160 slice_106"  [label="(1, 16, 16, 384)", style=solid];
+"1160 slice_106" -> "1161 slice_107"  [label="(1, 16, 16, 384)", style=solid];
+"1161 slice_107" -> "1162 slice_108"  [label="(1, 14, 16, 384)", style=solid];
+"1162 slice_108" -> "1163 slice_109"  [label="(1, 14, 14, 384)", style=solid];
+"1163 slice_109" -> "1164 contiguous_11"  [label="(1, 14, 14, 384)", style=solid];
+"1164 contiguous_11" -> "1165 layer_norm_15"  [label="(1, 14, 14, 384)", style=solid];
+"1165 layer_norm_15" -> "1166 add_22"  [label="(1, 14, 14, 384)", style=solid];
+"1166 add_22" -> "1167 linear_42"  [label="(1, 14, 14, 384)", style=solid];
+"1166 add_22" -> "1173 add_23"  [label="(1, 14, 14, 384)", style=solid];
+"1167 linear_42" -> "1168 gelu_6"  [label="(1, 14, 14, 1536)", style=solid];
+"1168 gelu_6" -> "1169 dropout_26"  [label="(1, 14, 14, 1536)", style=solid];
+"1169 dropout_26" -> "1170 linear_43"  [label="(1, 14, 14, 1536)", style=solid];
+"1170 linear_43" -> "1171 dropout_27"  [label="(1, 14, 14, 384)", style=solid];
+"1171 dropout_27" -> "1172 layer_norm_16"  [label="(1, 14, 14, 384)", style=solid];
+"1172 layer_norm_16" -> "1173 add_23"  [label="(1, 14, 14, 384)", style=solid];
+"1173 add_23" -> "1185 pad_9"  [label="(1, 14, 14, 384)", style=solid];
+"1173 add_23" -> "1282 add_26"  [label="(1, 14, 14, 384)", style=solid];
+"1174 linear_44" -> "1175 relu__7"  [label="(1, 15, 15, 512)", style=solid];
+"1175 relu__7" -> "1176 linear_45"  [label="(1, 15, 15, 512)", style=solid];
+"1176 linear_45" -> "1177 view_37"  [label="(1, 15, 15, 12)", style=solid];
+"1177 view_37" -> "1178 index_7"  [label="(225, 12)", style=solid];
+"1178 index_7" -> "1179 view_38"  [label="(4096, 12)", style=solid];
+"1179 view_38" -> "1180 permute_32"  [label="(64, 64, 12)", style=solid];
+"1180 permute_32" -> "1181 contiguous_12"  [label="(12, 64, 64)", style=solid];
+"1181 contiguous_12" -> "1182 unsqueeze_19"  [label="(12, 64, 64)", style=solid];
+"1182 unsqueeze_19" -> "1183 sigmoid_7"  [label="(1, 12, 64, 64)", style=solid];
+"1183 sigmoid_7" -> "1184 mul_14"  [label="(1, 12, 64, 64)", style=solid];
+"1184 mul_14" -> "1212 add_24"  [label="(1, 12, 64, 64)", style=solid];
+"1185 pad_9" -> "1186 roll_6"  [label="(1, 16, 16, 384)", style=solid];
+"1186 roll_6" -> "1187 view_39"  [label="(1, 16, 16, 384)", style=solid];
+"1187 view_39" -> "1188 permute_33"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1188 permute_33" -> "1189 reshape_31"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1189 reshape_31" -> "1193 linear_46"  [label="(4, 64, 384)", style=solid];
+"1189 reshape_31" -> "1213 new_zeros_3"  [label="(4, 64, 384)", style=solid];
+"1190 clone_7" -> "1191 slice_110"  [label="(1152,)", style=solid];
+"1190 clone_7" -> "1193 linear_46"  [label="(1152,)", style=solid];
+"1191 slice_110" -> "1192 zero__7"  [label="(384,)", style=solid];
+"1193 linear_46" -> "1194 reshape_32"  [label="(4, 64, 1152)", style=solid];
+"1194 reshape_32" -> "1195 permute_34"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1195 permute_34" -> "1196 select_21"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1195 permute_34" -> "1197 select_22"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1195 permute_34" -> "1198 select_23"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1196 select_21" -> "1199 linalg_vector_norm_14"  [label="(4, 12, 64, 32)", style=solid];
+"1196 select_21" -> "1201 expand_as_14"  [label="(4, 12, 64, 32)", style=solid];
+"1196 select_21" -> "1202 div_14"  [label="(4, 12, 64, 32)", style=solid];
+"1197 select_22" -> "1203 linalg_vector_norm_15"  [label="(4, 12, 64, 32)", style=solid];
+"1197 select_22" -> "1205 expand_as_15"  [label="(4, 12, 64, 32)", style=solid];
+"1197 select_22" -> "1206 div_15"  [label="(4, 12, 64, 32)", style=solid];
+"1198 select_23" -> "1267 matmul_15"  [label="(4, 12, 64, 32)", style=solid];
+"1199 linalg_vector_norm_14" -> "1200 clamp_min_14"  [label="(4, 12, 64, 1)", style=solid];
+"1200 clamp_min_14" -> "1201 expand_as_14"  [label="(4, 12, 64, 1)", style=solid];
+"1201 expand_as_14" -> "1202 div_14"  [label="(4, 12, 64, 32)", style=solid];
+"1202 div_14" -> "1208 matmul_14"  [label="(4, 12, 64, 32)", style=solid];
+"1203 linalg_vector_norm_15" -> "1204 clamp_min_15"  [label="(4, 12, 64, 1)", style=solid];
+"1204 clamp_min_15" -> "1205 expand_as_15"  [label="(4, 12, 64, 1)", style=solid];
+"1205 expand_as_15" -> "1206 div_15"  [label="(4, 12, 64, 32)", style=solid];
+"1206 div_15" -> "1207 transpose_14"  [label="(4, 12, 64, 32)", style=solid];
+"1207 transpose_14" -> "1208 matmul_14"  [label="(4, 12, 32, 64)", style=solid];
+"1208 matmul_14" -> "1211 mul_15"  [label="(4, 12, 64, 64)", style=solid];
+"1209 clamp_7" -> "1210 exp_7"  [label="(12, 1, 1)", style=solid];
+"1210 exp_7" -> "1211 mul_15"  [label="(12, 1, 1)", style=solid];
+"1211 mul_15" -> "1212 add_24"  [label="(4, 12, 64, 64)", style=solid];
+"1212 add_24" -> "1260 view_41"  [label="(4, 12, 64, 64)", style=solid];
+"1213 new_zeros_3" -> "1215 slice_111"  [label="(16, 16)", style=solid];
+"1213 new_zeros_3" -> "1219 slice_113"  [label="(16, 16)", style=solid];
+"1213 new_zeros_3" -> "1223 slice_115"  [label="(16, 16)", style=solid];
+"1213 new_zeros_3" -> "1227 slice_117"  [label="(16, 16)", style=solid];
+"1213 new_zeros_3" -> "1231 slice_119"  [label="(16, 16)", style=solid];
+"1213 new_zeros_3" -> "1235 slice_121"  [label="(16, 16)", style=solid];
+"1213 new_zeros_3" -> "1239 slice_123"  [label="(16, 16)", style=solid];
+"1213 new_zeros_3" -> "1243 slice_125"  [label="(16, 16)", style=solid];
+"1213 new_zeros_3" -> "1247 slice_127"  [label="(16, 16)", style=solid];
+"1213 new_zeros_3" -> "1250 view_40"  [label="(16, 16)", style=solid];
+"1214 lift_fresh_copy_27" -> "1217 fill__27"  [label="()", style=solid];
+"1215 slice_111" -> "1216 slice_112"  [label="(8, 16)", style=solid];
+"1216 slice_112" -> "1217 fill__27"  [label="(8, 8)", style=solid];
+"1218 lift_fresh_copy_28" -> "1221 fill__28"  [label="()", style=solid];
+"1219 slice_113" -> "1220 slice_114"  [label="(8, 16)", style=solid];
+"1220 slice_114" -> "1221 fill__28"  [label="(8, 4)", style=solid];
+"1222 lift_fresh_copy_29" -> "1225 fill__29"  [label="()", style=solid];
+"1223 slice_115" -> "1224 slice_116"  [label="(8, 16)", style=solid];
+"1224 slice_116" -> "1225 fill__29"  [label="(8, 4)", style=solid];
+"1226 lift_fresh_copy_30" -> "1229 fill__30"  [label="()", style=solid];
+"1227 slice_117" -> "1228 slice_118"  [label="(4, 16)", style=solid];
+"1228 slice_118" -> "1229 fill__30"  [label="(4, 8)", style=solid];
+"1230 lift_fresh_copy_31" -> "1233 fill__31"  [label="()", style=solid];
+"1231 slice_119" -> "1232 slice_120"  [label="(4, 16)", style=solid];
+"1232 slice_120" -> "1233 fill__31"  [label="(4, 4)", style=solid];
+"1234 lift_fresh_copy_32" -> "1237 fill__32"  [label="()", style=solid];
+"1235 slice_121" -> "1236 slice_122"  [label="(4, 16)", style=solid];
+"1236 slice_122" -> "1237 fill__32"  [label="(4, 4)", style=solid];
+"1238 lift_fresh_copy_33" -> "1241 fill__33"  [label="()", style=solid];
+"1239 slice_123" -> "1240 slice_124"  [label="(4, 16)", style=solid];
+"1240 slice_124" -> "1241 fill__33"  [label="(4, 8)", style=solid];
+"1242 lift_fresh_copy_34" -> "1245 fill__34"  [label="()", style=solid];
+"1243 slice_125" -> "1244 slice_126"  [label="(4, 16)", style=solid];
+"1244 slice_126" -> "1245 fill__34"  [label="(4, 4)", style=solid];
+"1246 lift_fresh_copy_35" -> "1249 fill__35"  [label="()", style=solid];
+"1247 slice_127" -> "1248 slice_128"  [label="(4, 16)", style=solid];
+"1248 slice_128" -> "1249 fill__35"  [label="(4, 4)", style=solid];
+"1250 view_40" -> "1251 permute_35"  [label="(2, 8, 2, 8)", style=solid];
+"1251 permute_35" -> "1252 reshape_33"  [label="(2, 2, 8, 8)", style=solid];
+"1252 reshape_33" -> "1253 unsqueeze_20"  [label="(4, 64)", style=solid];
+"1252 reshape_33" -> "1254 unsqueeze_21"  [label="(4, 64)", style=solid];
+"1253 unsqueeze_20" -> "1255 sub_3"  [label="(4, 1, 64)", style=solid];
+"1254 unsqueeze_21" -> "1255 sub_3"  [label="(4, 64, 1)", style=solid];
+"1255 sub_3" -> "1256 ne_3"  [label="(4, 64, 64)", style=solid];
+"1255 sub_3" -> "1257 masked_fill_6"  [label="(4, 64, 64)", style=solid];
+"1255 sub_3" -> "1258 eq_3"  [label="(4, 64, 64)", style=solid];
+"1256 ne_3" -> "1257 masked_fill_6"  [label="(4, 64, 64)", style=solid];
+"1257 masked_fill_6" -> "1259 masked_fill_7"  [label="(4, 64, 64)", style=solid];
+"1258 eq_3" -> "1259 masked_fill_7"  [label="(4, 64, 64)", style=solid];
+"1259 masked_fill_7" -> "1261 unsqueeze_22"  [label="(4, 64, 64)", style=solid];
+"1260 view_41" -> "1263 add_25"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1261 unsqueeze_22" -> "1262 unsqueeze_23"  [label="(4, 1, 64, 64)", style=solid];
+"1262 unsqueeze_23" -> "1263 add_25"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1263 add_25" -> "1264 view_42"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1264 view_42" -> "1265 softmax_7"  [label="(4, 12, 64, 64)", style=solid];
+"1265 softmax_7" -> "1266 dropout_28"  [label="(4, 12, 64, 64)", style=solid];
+"1266 dropout_28" -> "1267 matmul_15"  [label="(4, 12, 64, 64)", style=solid];
+"1267 matmul_15" -> "1268 transpose_15"  [label="(4, 12, 64, 32)", style=solid];
+"1268 transpose_15" -> "1269 reshape_34"  [label="(4, 64, 12, 32)", style=solid];
+"1269 reshape_34" -> "1270 linear_47"  [label="(4, 64, 384)", style=solid];
+"1270 linear_47" -> "1271 dropout_29"  [label="(4, 64, 384)", style=solid];
+"1271 dropout_29" -> "1272 view_43"  [label="(4, 64, 384)", style=solid];
+"1272 view_43" -> "1273 permute_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1273 permute_36" -> "1274 reshape_35"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1274 reshape_35" -> "1275 roll_7"  [label="(1, 16, 16, 384)", style=solid];
+"1275 roll_7" -> "1276 slice_129"  [label="(1, 16, 16, 384)", style=solid];
+"1276 slice_129" -> "1277 slice_130"  [label="(1, 16, 16, 384)", style=solid];
+"1277 slice_130" -> "1278 slice_131"  [label="(1, 14, 16, 384)", style=solid];
+"1278 slice_131" -> "1279 slice_132"  [label="(1, 14, 14, 384)", style=solid];
+"1279 slice_132" -> "1280 contiguous_13"  [label="(1, 14, 14, 384)", style=solid];
+"1280 contiguous_13" -> "1281 layer_norm_17"  [label="(1, 14, 14, 384)", style=solid];
+"1281 layer_norm_17" -> "1282 add_26"  [label="(1, 14, 14, 384)", style=solid];
+"1282 add_26" -> "1283 linear_48"  [label="(1, 14, 14, 384)", style=solid];
+"1282 add_26" -> "1289 add_27"  [label="(1, 14, 14, 384)", style=solid];
+"1283 linear_48" -> "1284 gelu_7"  [label="(1, 14, 14, 1536)", style=solid];
+"1284 gelu_7" -> "1285 dropout_30"  [label="(1, 14, 14, 1536)", style=solid];
+"1285 dropout_30" -> "1286 linear_49"  [label="(1, 14, 14, 1536)", style=solid];
+"1286 linear_49" -> "1287 dropout_31"  [label="(1, 14, 14, 384)", style=solid];
+"1287 dropout_31" -> "1288 layer_norm_18"  [label="(1, 14, 14, 384)", style=solid];
+"1288 layer_norm_18" -> "1289 add_27"  [label="(1, 14, 14, 384)", style=solid];
+"1289 add_27" -> "1301 pad_10"  [label="(1, 14, 14, 384)", style=solid];
+"1289 add_27" -> "1344 add_29"  [label="(1, 14, 14, 384)", style=solid];
+"1290 linear_50" -> "1291 relu__8"  [label="(1, 15, 15, 512)", style=solid];
+"1291 relu__8" -> "1292 linear_51"  [label="(1, 15, 15, 512)", style=solid];
+"1292 linear_51" -> "1293 view_44"  [label="(1, 15, 15, 12)", style=solid];
+"1293 view_44" -> "1294 index_8"  [label="(225, 12)", style=solid];
+"1294 index_8" -> "1295 view_45"  [label="(4096, 12)", style=solid];
+"1295 view_45" -> "1296 permute_37"  [label="(64, 64, 12)", style=solid];
+"1296 permute_37" -> "1297 contiguous_14"  [label="(12, 64, 64)", style=solid];
+"1297 contiguous_14" -> "1298 unsqueeze_24"  [label="(12, 64, 64)", style=solid];
+"1298 unsqueeze_24" -> "1299 sigmoid_8"  [label="(1, 12, 64, 64)", style=solid];
+"1299 sigmoid_8" -> "1300 mul_16"  [label="(1, 12, 64, 64)", style=solid];
+"1300 mul_16" -> "1327 add_28"  [label="(1, 12, 64, 64)", style=solid];
+"1301 pad_10" -> "1302 view_46"  [label="(1, 16, 16, 384)", style=solid];
+"1302 view_46" -> "1303 permute_38"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1303 permute_38" -> "1304 reshape_36"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1304 reshape_36" -> "1308 linear_52"  [label="(4, 64, 384)", style=solid];
+"1305 clone_8" -> "1306 slice_133"  [label="(1152,)", style=solid];
+"1305 clone_8" -> "1308 linear_52"  [label="(1152,)", style=solid];
+"1306 slice_133" -> "1307 zero__8"  [label="(384,)", style=solid];
+"1308 linear_52" -> "1309 reshape_37"  [label="(4, 64, 1152)", style=solid];
+"1309 reshape_37" -> "1310 permute_39"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1310 permute_39" -> "1311 select_24"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1310 permute_39" -> "1312 select_25"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1310 permute_39" -> "1313 select_26"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1311 select_24" -> "1314 linalg_vector_norm_16"  [label="(4, 12, 64, 32)", style=solid];
+"1311 select_24" -> "1316 expand_as_16"  [label="(4, 12, 64, 32)", style=solid];
+"1311 select_24" -> "1317 div_16"  [label="(4, 12, 64, 32)", style=solid];
+"1312 select_25" -> "1318 linalg_vector_norm_17"  [label="(4, 12, 64, 32)", style=solid];
+"1312 select_25" -> "1320 expand_as_17"  [label="(4, 12, 64, 32)", style=solid];
+"1312 select_25" -> "1321 div_17"  [label="(4, 12, 64, 32)", style=solid];
+"1313 select_26" -> "1330 matmul_17"  [label="(4, 12, 64, 32)", style=solid];
+"1314 linalg_vector_norm_16" -> "1315 clamp_min_16"  [label="(4, 12, 64, 1)", style=solid];
+"1315 clamp_min_16" -> "1316 expand_as_16"  [label="(4, 12, 64, 1)", style=solid];
+"1316 expand_as_16" -> "1317 div_16"  [label="(4, 12, 64, 32)", style=solid];
+"1317 div_16" -> "1323 matmul_16"  [label="(4, 12, 64, 32)", style=solid];
+"1318 linalg_vector_norm_17" -> "1319 clamp_min_17"  [label="(4, 12, 64, 1)", style=solid];
+"1319 clamp_min_17" -> "1320 expand_as_17"  [label="(4, 12, 64, 1)", style=solid];
+"1320 expand_as_17" -> "1321 div_17"  [label="(4, 12, 64, 32)", style=solid];
+"1321 div_17" -> "1322 transpose_16"  [label="(4, 12, 64, 32)", style=solid];
+"1322 transpose_16" -> "1323 matmul_16"  [label="(4, 12, 32, 64)", style=solid];
+"1323 matmul_16" -> "1326 mul_17"  [label="(4, 12, 64, 64)", style=solid];
+"1324 clamp_8" -> "1325 exp_8"  [label="(12, 1, 1)", style=solid];
+"1325 exp_8" -> "1326 mul_17"  [label="(12, 1, 1)", style=solid];
+"1326 mul_17" -> "1327 add_28"  [label="(4, 12, 64, 64)", style=solid];
+"1327 add_28" -> "1328 softmax_8"  [label="(4, 12, 64, 64)", style=solid];
+"1328 softmax_8" -> "1329 dropout_32"  [label="(4, 12, 64, 64)", style=solid];
+"1329 dropout_32" -> "1330 matmul_17"  [label="(4, 12, 64, 64)", style=solid];
+"1330 matmul_17" -> "1331 transpose_17"  [label="(4, 12, 64, 32)", style=solid];
+"1331 transpose_17" -> "1332 reshape_38"  [label="(4, 64, 12, 32)", style=solid];
+"1332 reshape_38" -> "1333 linear_53"  [label="(4, 64, 384)", style=solid];
+"1333 linear_53" -> "1334 dropout_33"  [label="(4, 64, 384)", style=solid];
+"1334 dropout_33" -> "1335 view_47"  [label="(4, 64, 384)", style=solid];
+"1335 view_47" -> "1336 permute_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1336 permute_40" -> "1337 reshape_39"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1337 reshape_39" -> "1338 slice_134"  [label="(1, 16, 16, 384)", style=solid];
+"1338 slice_134" -> "1339 slice_135"  [label="(1, 16, 16, 384)", style=solid];
+"1339 slice_135" -> "1340 slice_136"  [label="(1, 14, 16, 384)", style=solid];
+"1340 slice_136" -> "1341 slice_137"  [label="(1, 14, 14, 384)", style=solid];
+"1341 slice_137" -> "1342 contiguous_15"  [label="(1, 14, 14, 384)", style=solid];
+"1342 contiguous_15" -> "1343 layer_norm_19"  [label="(1, 14, 14, 384)", style=solid];
+"1343 layer_norm_19" -> "1344 add_29"  [label="(1, 14, 14, 384)", style=solid];
+"1344 add_29" -> "1345 linear_54"  [label="(1, 14, 14, 384)", style=solid];
+"1344 add_29" -> "1351 add_30"  [label="(1, 14, 14, 384)", style=solid];
+"1345 linear_54" -> "1346 gelu_8"  [label="(1, 14, 14, 1536)", style=solid];
+"1346 gelu_8" -> "1347 dropout_34"  [label="(1, 14, 14, 1536)", style=solid];
+"1347 dropout_34" -> "1348 linear_55"  [label="(1, 14, 14, 1536)", style=solid];
+"1348 linear_55" -> "1349 dropout_35"  [label="(1, 14, 14, 384)", style=solid];
+"1349 dropout_35" -> "1350 layer_norm_20"  [label="(1, 14, 14, 384)", style=solid];
+"1350 layer_norm_20" -> "1351 add_30"  [label="(1, 14, 14, 384)", style=solid];
+"1351 add_30" -> "1363 pad_11"  [label="(1, 14, 14, 384)", style=solid];
+"1351 add_30" -> "1460 add_33"  [label="(1, 14, 14, 384)", style=solid];
+"1352 linear_56" -> "1353 relu__9"  [label="(1, 15, 15, 512)", style=solid];
+"1353 relu__9" -> "1354 linear_57"  [label="(1, 15, 15, 512)", style=solid];
+"1354 linear_57" -> "1355 view_48"  [label="(1, 15, 15, 12)", style=solid];
+"1355 view_48" -> "1356 index_9"  [label="(225, 12)", style=solid];
+"1356 index_9" -> "1357 view_49"  [label="(4096, 12)", style=solid];
+"1357 view_49" -> "1358 permute_41"  [label="(64, 64, 12)", style=solid];
+"1358 permute_41" -> "1359 contiguous_16"  [label="(12, 64, 64)", style=solid];
+"1359 contiguous_16" -> "1360 unsqueeze_25"  [label="(12, 64, 64)", style=solid];
+"1360 unsqueeze_25" -> "1361 sigmoid_9"  [label="(1, 12, 64, 64)", style=solid];
+"1361 sigmoid_9" -> "1362 mul_18"  [label="(1, 12, 64, 64)", style=solid];
+"1362 mul_18" -> "1390 add_31"  [label="(1, 12, 64, 64)", style=solid];
+"1363 pad_11" -> "1364 roll_8"  [label="(1, 16, 16, 384)", style=solid];
+"1364 roll_8" -> "1365 view_50"  [label="(1, 16, 16, 384)", style=solid];
+"1365 view_50" -> "1366 permute_42"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1366 permute_42" -> "1367 reshape_40"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1367 reshape_40" -> "1371 linear_58"  [label="(4, 64, 384)", style=solid];
+"1367 reshape_40" -> "1391 new_zeros_4"  [label="(4, 64, 384)", style=solid];
+"1368 clone_9" -> "1369 slice_138"  [label="(1152,)", style=solid];
+"1368 clone_9" -> "1371 linear_58"  [label="(1152,)", style=solid];
+"1369 slice_138" -> "1370 zero__9"  [label="(384,)", style=solid];
+"1371 linear_58" -> "1372 reshape_41"  [label="(4, 64, 1152)", style=solid];
+"1372 reshape_41" -> "1373 permute_43"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1373 permute_43" -> "1374 select_27"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1373 permute_43" -> "1375 select_28"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1373 permute_43" -> "1376 select_29"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1374 select_27" -> "1377 linalg_vector_norm_18"  [label="(4, 12, 64, 32)", style=solid];
+"1374 select_27" -> "1379 expand_as_18"  [label="(4, 12, 64, 32)", style=solid];
+"1374 select_27" -> "1380 div_18"  [label="(4, 12, 64, 32)", style=solid];
+"1375 select_28" -> "1381 linalg_vector_norm_19"  [label="(4, 12, 64, 32)", style=solid];
+"1375 select_28" -> "1383 expand_as_19"  [label="(4, 12, 64, 32)", style=solid];
+"1375 select_28" -> "1384 div_19"  [label="(4, 12, 64, 32)", style=solid];
+"1376 select_29" -> "1445 matmul_19"  [label="(4, 12, 64, 32)", style=solid];
+"1377 linalg_vector_norm_18" -> "1378 clamp_min_18"  [label="(4, 12, 64, 1)", style=solid];
+"1378 clamp_min_18" -> "1379 expand_as_18"  [label="(4, 12, 64, 1)", style=solid];
+"1379 expand_as_18" -> "1380 div_18"  [label="(4, 12, 64, 32)", style=solid];
+"1380 div_18" -> "1386 matmul_18"  [label="(4, 12, 64, 32)", style=solid];
+"1381 linalg_vector_norm_19" -> "1382 clamp_min_19"  [label="(4, 12, 64, 1)", style=solid];
+"1382 clamp_min_19" -> "1383 expand_as_19"  [label="(4, 12, 64, 1)", style=solid];
+"1383 expand_as_19" -> "1384 div_19"  [label="(4, 12, 64, 32)", style=solid];
+"1384 div_19" -> "1385 transpose_18"  [label="(4, 12, 64, 32)", style=solid];
+"1385 transpose_18" -> "1386 matmul_18"  [label="(4, 12, 32, 64)", style=solid];
+"1386 matmul_18" -> "1389 mul_19"  [label="(4, 12, 64, 64)", style=solid];
+"1387 clamp_9" -> "1388 exp_9"  [label="(12, 1, 1)", style=solid];
+"1388 exp_9" -> "1389 mul_19"  [label="(12, 1, 1)", style=solid];
+"1389 mul_19" -> "1390 add_31"  [label="(4, 12, 64, 64)", style=solid];
+"1390 add_31" -> "1438 view_52"  [label="(4, 12, 64, 64)", style=solid];
+"1391 new_zeros_4" -> "1393 slice_139"  [label="(16, 16)", style=solid];
+"1391 new_zeros_4" -> "1397 slice_141"  [label="(16, 16)", style=solid];
+"1391 new_zeros_4" -> "1401 slice_143"  [label="(16, 16)", style=solid];
+"1391 new_zeros_4" -> "1405 slice_145"  [label="(16, 16)", style=solid];
+"1391 new_zeros_4" -> "1409 slice_147"  [label="(16, 16)", style=solid];
+"1391 new_zeros_4" -> "1413 slice_149"  [label="(16, 16)", style=solid];
+"1391 new_zeros_4" -> "1417 slice_151"  [label="(16, 16)", style=solid];
+"1391 new_zeros_4" -> "1421 slice_153"  [label="(16, 16)", style=solid];
+"1391 new_zeros_4" -> "1425 slice_155"  [label="(16, 16)", style=solid];
+"1391 new_zeros_4" -> "1428 view_51"  [label="(16, 16)", style=solid];
+"1392 lift_fresh_copy_36" -> "1395 fill__36"  [label="()", style=solid];
+"1393 slice_139" -> "1394 slice_140"  [label="(8, 16)", style=solid];
+"1394 slice_140" -> "1395 fill__36"  [label="(8, 8)", style=solid];
+"1396 lift_fresh_copy_37" -> "1399 fill__37"  [label="()", style=solid];
+"1397 slice_141" -> "1398 slice_142"  [label="(8, 16)", style=solid];
+"1398 slice_142" -> "1399 fill__37"  [label="(8, 4)", style=solid];
+"1400 lift_fresh_copy_38" -> "1403 fill__38"  [label="()", style=solid];
+"1401 slice_143" -> "1402 slice_144"  [label="(8, 16)", style=solid];
+"1402 slice_144" -> "1403 fill__38"  [label="(8, 4)", style=solid];
+"1404 lift_fresh_copy_39" -> "1407 fill__39"  [label="()", style=solid];
+"1405 slice_145" -> "1406 slice_146"  [label="(4, 16)", style=solid];
+"1406 slice_146" -> "1407 fill__39"  [label="(4, 8)", style=solid];
+"1408 lift_fresh_copy_40" -> "1411 fill__40"  [label="()", style=solid];
+"1409 slice_147" -> "1410 slice_148"  [label="(4, 16)", style=solid];
+"1410 slice_148" -> "1411 fill__40"  [label="(4, 4)", style=solid];
+"1412 lift_fresh_copy_41" -> "1415 fill__41"  [label="()", style=solid];
+"1413 slice_149" -> "1414 slice_150"  [label="(4, 16)", style=solid];
+"1414 slice_150" -> "1415 fill__41"  [label="(4, 4)", style=solid];
+"1416 lift_fresh_copy_42" -> "1419 fill__42"  [label="()", style=solid];
+"1417 slice_151" -> "1418 slice_152"  [label="(4, 16)", style=solid];
+"1418 slice_152" -> "1419 fill__42"  [label="(4, 8)", style=solid];
+"1420 lift_fresh_copy_43" -> "1423 fill__43"  [label="()", style=solid];
+"1421 slice_153" -> "1422 slice_154"  [label="(4, 16)", style=solid];
+"1422 slice_154" -> "1423 fill__43"  [label="(4, 4)", style=solid];
+"1424 lift_fresh_copy_44" -> "1427 fill__44"  [label="()", style=solid];
+"1425 slice_155" -> "1426 slice_156"  [label="(4, 16)", style=solid];
+"1426 slice_156" -> "1427 fill__44"  [label="(4, 4)", style=solid];
+"1428 view_51" -> "1429 permute_44"  [label="(2, 8, 2, 8)", style=solid];
+"1429 permute_44" -> "1430 reshape_42"  [label="(2, 2, 8, 8)", style=solid];
+"1430 reshape_42" -> "1431 unsqueeze_26"  [label="(4, 64)", style=solid];
+"1430 reshape_42" -> "1432 unsqueeze_27"  [label="(4, 64)", style=solid];
+"1431 unsqueeze_26" -> "1433 sub_4"  [label="(4, 1, 64)", style=solid];
+"1432 unsqueeze_27" -> "1433 sub_4"  [label="(4, 64, 1)", style=solid];
+"1433 sub_4" -> "1434 ne_4"  [label="(4, 64, 64)", style=solid];
+"1433 sub_4" -> "1435 masked_fill_8"  [label="(4, 64, 64)", style=solid];
+"1433 sub_4" -> "1436 eq_4"  [label="(4, 64, 64)", style=solid];
+"1434 ne_4" -> "1435 masked_fill_8"  [label="(4, 64, 64)", style=solid];
+"1435 masked_fill_8" -> "1437 masked_fill_9"  [label="(4, 64, 64)", style=solid];
+"1436 eq_4" -> "1437 masked_fill_9"  [label="(4, 64, 64)", style=solid];
+"1437 masked_fill_9" -> "1439 unsqueeze_28"  [label="(4, 64, 64)", style=solid];
+"1438 view_52" -> "1441 add_32"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1439 unsqueeze_28" -> "1440 unsqueeze_29"  [label="(4, 1, 64, 64)", style=solid];
+"1440 unsqueeze_29" -> "1441 add_32"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1441 add_32" -> "1442 view_53"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1442 view_53" -> "1443 softmax_9"  [label="(4, 12, 64, 64)", style=solid];
+"1443 softmax_9" -> "1444 dropout_36"  [label="(4, 12, 64, 64)", style=solid];
+"1444 dropout_36" -> "1445 matmul_19"  [label="(4, 12, 64, 64)", style=solid];
+"1445 matmul_19" -> "1446 transpose_19"  [label="(4, 12, 64, 32)", style=solid];
+"1446 transpose_19" -> "1447 reshape_43"  [label="(4, 64, 12, 32)", style=solid];
+"1447 reshape_43" -> "1448 linear_59"  [label="(4, 64, 384)", style=solid];
+"1448 linear_59" -> "1449 dropout_37"  [label="(4, 64, 384)", style=solid];
+"1449 dropout_37" -> "1450 view_54"  [label="(4, 64, 384)", style=solid];
+"1450 view_54" -> "1451 permute_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1451 permute_45" -> "1452 reshape_44"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1452 reshape_44" -> "1453 roll_9"  [label="(1, 16, 16, 384)", style=solid];
+"1453 roll_9" -> "1454 slice_157"  [label="(1, 16, 16, 384)", style=solid];
+"1454 slice_157" -> "1455 slice_158"  [label="(1, 16, 16, 384)", style=solid];
+"1455 slice_158" -> "1456 slice_159"  [label="(1, 14, 16, 384)", style=solid];
+"1456 slice_159" -> "1457 slice_160"  [label="(1, 14, 14, 384)", style=solid];
+"1457 slice_160" -> "1458 contiguous_17"  [label="(1, 14, 14, 384)", style=solid];
+"1458 contiguous_17" -> "1459 layer_norm_21"  [label="(1, 14, 14, 384)", style=solid];
+"1459 layer_norm_21" -> "1460 add_33"  [label="(1, 14, 14, 384)", style=solid];
+"1460 add_33" -> "1461 linear_60"  [label="(1, 14, 14, 384)", style=solid];
+"1460 add_33" -> "1467 add_34"  [label="(1, 14, 14, 384)", style=solid];
+"1461 linear_60" -> "1462 gelu_9"  [label="(1, 14, 14, 1536)", style=solid];
+"1462 gelu_9" -> "1463 dropout_38"  [label="(1, 14, 14, 1536)", style=solid];
+"1463 dropout_38" -> "1464 linear_61"  [label="(1, 14, 14, 1536)", style=solid];
+"1464 linear_61" -> "1465 dropout_39"  [label="(1, 14, 14, 384)", style=solid];
+"1465 dropout_39" -> "1466 layer_norm_22"  [label="(1, 14, 14, 384)", style=solid];
+"1466 layer_norm_22" -> "1467 add_34"  [label="(1, 14, 14, 384)", style=solid];
+"1467 add_34" -> "1479 pad_12"  [label="(1, 14, 14, 384)", style=solid];
+"1467 add_34" -> "1522 add_36"  [label="(1, 14, 14, 384)", style=solid];
+"1468 linear_62" -> "1469 relu__10"  [label="(1, 15, 15, 512)", style=solid];
+"1469 relu__10" -> "1470 linear_63"  [label="(1, 15, 15, 512)", style=solid];
+"1470 linear_63" -> "1471 view_55"  [label="(1, 15, 15, 12)", style=solid];
+"1471 view_55" -> "1472 index_10"  [label="(225, 12)", style=solid];
+"1472 index_10" -> "1473 view_56"  [label="(4096, 12)", style=solid];
+"1473 view_56" -> "1474 permute_46"  [label="(64, 64, 12)", style=solid];
+"1474 permute_46" -> "1475 contiguous_18"  [label="(12, 64, 64)", style=solid];
+"1475 contiguous_18" -> "1476 unsqueeze_30"  [label="(12, 64, 64)", style=solid];
+"1476 unsqueeze_30" -> "1477 sigmoid_10"  [label="(1, 12, 64, 64)", style=solid];
+"1477 sigmoid_10" -> "1478 mul_20"  [label="(1, 12, 64, 64)", style=solid];
+"1478 mul_20" -> "1505 add_35"  [label="(1, 12, 64, 64)", style=solid];
+"1479 pad_12" -> "1480 view_57"  [label="(1, 16, 16, 384)", style=solid];
+"1480 view_57" -> "1481 permute_47"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1481 permute_47" -> "1482 reshape_45"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1482 reshape_45" -> "1486 linear_64"  [label="(4, 64, 384)", style=solid];
+"1483 clone_10" -> "1484 slice_161"  [label="(1152,)", style=solid];
+"1483 clone_10" -> "1486 linear_64"  [label="(1152,)", style=solid];
+"1484 slice_161" -> "1485 zero__10"  [label="(384,)", style=solid];
+"1486 linear_64" -> "1487 reshape_46"  [label="(4, 64, 1152)", style=solid];
+"1487 reshape_46" -> "1488 permute_48"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1488 permute_48" -> "1489 select_30"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1488 permute_48" -> "1490 select_31"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1488 permute_48" -> "1491 select_32"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1489 select_30" -> "1492 linalg_vector_norm_20"  [label="(4, 12, 64, 32)", style=solid];
+"1489 select_30" -> "1494 expand_as_20"  [label="(4, 12, 64, 32)", style=solid];
+"1489 select_30" -> "1495 div_20"  [label="(4, 12, 64, 32)", style=solid];
+"1490 select_31" -> "1496 linalg_vector_norm_21"  [label="(4, 12, 64, 32)", style=solid];
+"1490 select_31" -> "1498 expand_as_21"  [label="(4, 12, 64, 32)", style=solid];
+"1490 select_31" -> "1499 div_21"  [label="(4, 12, 64, 32)", style=solid];
+"1491 select_32" -> "1508 matmul_21"  [label="(4, 12, 64, 32)", style=solid];
+"1492 linalg_vector_norm_20" -> "1493 clamp_min_20"  [label="(4, 12, 64, 1)", style=solid];
+"1493 clamp_min_20" -> "1494 expand_as_20"  [label="(4, 12, 64, 1)", style=solid];
+"1494 expand_as_20" -> "1495 div_20"  [label="(4, 12, 64, 32)", style=solid];
+"1495 div_20" -> "1501 matmul_20"  [label="(4, 12, 64, 32)", style=solid];
+"1496 linalg_vector_norm_21" -> "1497 clamp_min_21"  [label="(4, 12, 64, 1)", style=solid];
+"1497 clamp_min_21" -> "1498 expand_as_21"  [label="(4, 12, 64, 1)", style=solid];
+"1498 expand_as_21" -> "1499 div_21"  [label="(4, 12, 64, 32)", style=solid];
+"1499 div_21" -> "1500 transpose_20"  [label="(4, 12, 64, 32)", style=solid];
+"1500 transpose_20" -> "1501 matmul_20"  [label="(4, 12, 32, 64)", style=solid];
+"1501 matmul_20" -> "1504 mul_21"  [label="(4, 12, 64, 64)", style=solid];
+"1502 clamp_10" -> "1503 exp_10"  [label="(12, 1, 1)", style=solid];
+"1503 exp_10" -> "1504 mul_21"  [label="(12, 1, 1)", style=solid];
+"1504 mul_21" -> "1505 add_35"  [label="(4, 12, 64, 64)", style=solid];
+"1505 add_35" -> "1506 softmax_10"  [label="(4, 12, 64, 64)", style=solid];
+"1506 softmax_10" -> "1507 dropout_40"  [label="(4, 12, 64, 64)", style=solid];
+"1507 dropout_40" -> "1508 matmul_21"  [label="(4, 12, 64, 64)", style=solid];
+"1508 matmul_21" -> "1509 transpose_21"  [label="(4, 12, 64, 32)", style=solid];
+"1509 transpose_21" -> "1510 reshape_47"  [label="(4, 64, 12, 32)", style=solid];
+"1510 reshape_47" -> "1511 linear_65"  [label="(4, 64, 384)", style=solid];
+"1511 linear_65" -> "1512 dropout_41"  [label="(4, 64, 384)", style=solid];
+"1512 dropout_41" -> "1513 view_58"  [label="(4, 64, 384)", style=solid];
+"1513 view_58" -> "1514 permute_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1514 permute_49" -> "1515 reshape_48"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1515 reshape_48" -> "1516 slice_162"  [label="(1, 16, 16, 384)", style=solid];
+"1516 slice_162" -> "1517 slice_163"  [label="(1, 16, 16, 384)", style=solid];
+"1517 slice_163" -> "1518 slice_164"  [label="(1, 14, 16, 384)", style=solid];
+"1518 slice_164" -> "1519 slice_165"  [label="(1, 14, 14, 384)", style=solid];
+"1519 slice_165" -> "1520 contiguous_19"  [label="(1, 14, 14, 384)", style=solid];
+"1520 contiguous_19" -> "1521 layer_norm_23"  [label="(1, 14, 14, 384)", style=solid];
+"1521 layer_norm_23" -> "1522 add_36"  [label="(1, 14, 14, 384)", style=solid];
+"1522 add_36" -> "1523 linear_66"  [label="(1, 14, 14, 384)", style=solid];
+"1522 add_36" -> "1529 add_37"  [label="(1, 14, 14, 384)", style=solid];
+"1523 linear_66" -> "1524 gelu_10"  [label="(1, 14, 14, 1536)", style=solid];
+"1524 gelu_10" -> "1525 dropout_42"  [label="(1, 14, 14, 1536)", style=solid];
+"1525 dropout_42" -> "1526 linear_67"  [label="(1, 14, 14, 1536)", style=solid];
+"1526 linear_67" -> "1527 dropout_43"  [label="(1, 14, 14, 384)", style=solid];
+"1527 dropout_43" -> "1528 layer_norm_24"  [label="(1, 14, 14, 384)", style=solid];
+"1528 layer_norm_24" -> "1529 add_37"  [label="(1, 14, 14, 384)", style=solid];
+"1529 add_37" -> "1541 pad_13"  [label="(1, 14, 14, 384)", style=solid];
+"1529 add_37" -> "1638 add_40"  [label="(1, 14, 14, 384)", style=solid];
+"1530 linear_68" -> "1531 relu__11"  [label="(1, 15, 15, 512)", style=solid];
+"1531 relu__11" -> "1532 linear_69"  [label="(1, 15, 15, 512)", style=solid];
+"1532 linear_69" -> "1533 view_59"  [label="(1, 15, 15, 12)", style=solid];
+"1533 view_59" -> "1534 index_11"  [label="(225, 12)", style=solid];
+"1534 index_11" -> "1535 view_60"  [label="(4096, 12)", style=solid];
+"1535 view_60" -> "1536 permute_50"  [label="(64, 64, 12)", style=solid];
+"1536 permute_50" -> "1537 contiguous_20"  [label="(12, 64, 64)", style=solid];
+"1537 contiguous_20" -> "1538 unsqueeze_31"  [label="(12, 64, 64)", style=solid];
+"1538 unsqueeze_31" -> "1539 sigmoid_11"  [label="(1, 12, 64, 64)", style=solid];
+"1539 sigmoid_11" -> "1540 mul_22"  [label="(1, 12, 64, 64)", style=solid];
+"1540 mul_22" -> "1568 add_38"  [label="(1, 12, 64, 64)", style=solid];
+"1541 pad_13" -> "1542 roll_10"  [label="(1, 16, 16, 384)", style=solid];
+"1542 roll_10" -> "1543 view_61"  [label="(1, 16, 16, 384)", style=solid];
+"1543 view_61" -> "1544 permute_51"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1544 permute_51" -> "1545 reshape_49"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1545 reshape_49" -> "1549 linear_70"  [label="(4, 64, 384)", style=solid];
+"1545 reshape_49" -> "1569 new_zeros_5"  [label="(4, 64, 384)", style=solid];
+"1546 clone_11" -> "1547 slice_166"  [label="(1152,)", style=solid];
+"1546 clone_11" -> "1549 linear_70"  [label="(1152,)", style=solid];
+"1547 slice_166" -> "1548 zero__11"  [label="(384,)", style=solid];
+"1549 linear_70" -> "1550 reshape_50"  [label="(4, 64, 1152)", style=solid];
+"1550 reshape_50" -> "1551 permute_52"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1551 permute_52" -> "1552 select_33"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1551 permute_52" -> "1553 select_34"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1551 permute_52" -> "1554 select_35"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1552 select_33" -> "1555 linalg_vector_norm_22"  [label="(4, 12, 64, 32)", style=solid];
+"1552 select_33" -> "1557 expand_as_22"  [label="(4, 12, 64, 32)", style=solid];
+"1552 select_33" -> "1558 div_22"  [label="(4, 12, 64, 32)", style=solid];
+"1553 select_34" -> "1559 linalg_vector_norm_23"  [label="(4, 12, 64, 32)", style=solid];
+"1553 select_34" -> "1561 expand_as_23"  [label="(4, 12, 64, 32)", style=solid];
+"1553 select_34" -> "1562 div_23"  [label="(4, 12, 64, 32)", style=solid];
+"1554 select_35" -> "1623 matmul_23"  [label="(4, 12, 64, 32)", style=solid];
+"1555 linalg_vector_norm_22" -> "1556 clamp_min_22"  [label="(4, 12, 64, 1)", style=solid];
+"1556 clamp_min_22" -> "1557 expand_as_22"  [label="(4, 12, 64, 1)", style=solid];
+"1557 expand_as_22" -> "1558 div_22"  [label="(4, 12, 64, 32)", style=solid];
+"1558 div_22" -> "1564 matmul_22"  [label="(4, 12, 64, 32)", style=solid];
+"1559 linalg_vector_norm_23" -> "1560 clamp_min_23"  [label="(4, 12, 64, 1)", style=solid];
+"1560 clamp_min_23" -> "1561 expand_as_23"  [label="(4, 12, 64, 1)", style=solid];
+"1561 expand_as_23" -> "1562 div_23"  [label="(4, 12, 64, 32)", style=solid];
+"1562 div_23" -> "1563 transpose_22"  [label="(4, 12, 64, 32)", style=solid];
+"1563 transpose_22" -> "1564 matmul_22"  [label="(4, 12, 32, 64)", style=solid];
+"1564 matmul_22" -> "1567 mul_23"  [label="(4, 12, 64, 64)", style=solid];
+"1565 clamp_11" -> "1566 exp_11"  [label="(12, 1, 1)", style=solid];
+"1566 exp_11" -> "1567 mul_23"  [label="(12, 1, 1)", style=solid];
+"1567 mul_23" -> "1568 add_38"  [label="(4, 12, 64, 64)", style=solid];
+"1568 add_38" -> "1616 view_63"  [label="(4, 12, 64, 64)", style=solid];
+"1569 new_zeros_5" -> "1571 slice_167"  [label="(16, 16)", style=solid];
+"1569 new_zeros_5" -> "1575 slice_169"  [label="(16, 16)", style=solid];
+"1569 new_zeros_5" -> "1579 slice_171"  [label="(16, 16)", style=solid];
+"1569 new_zeros_5" -> "1583 slice_173"  [label="(16, 16)", style=solid];
+"1569 new_zeros_5" -> "1587 slice_175"  [label="(16, 16)", style=solid];
+"1569 new_zeros_5" -> "1591 slice_177"  [label="(16, 16)", style=solid];
+"1569 new_zeros_5" -> "1595 slice_179"  [label="(16, 16)", style=solid];
+"1569 new_zeros_5" -> "1599 slice_181"  [label="(16, 16)", style=solid];
+"1569 new_zeros_5" -> "1603 slice_183"  [label="(16, 16)", style=solid];
+"1569 new_zeros_5" -> "1606 view_62"  [label="(16, 16)", style=solid];
+"1570 lift_fresh_copy_45" -> "1573 fill__45"  [label="()", style=solid];
+"1571 slice_167" -> "1572 slice_168"  [label="(8, 16)", style=solid];
+"1572 slice_168" -> "1573 fill__45"  [label="(8, 8)", style=solid];
+"1574 lift_fresh_copy_46" -> "1577 fill__46"  [label="()", style=solid];
+"1575 slice_169" -> "1576 slice_170"  [label="(8, 16)", style=solid];
+"1576 slice_170" -> "1577 fill__46"  [label="(8, 4)", style=solid];
+"1578 lift_fresh_copy_47" -> "1581 fill__47"  [label="()", style=solid];
+"1579 slice_171" -> "1580 slice_172"  [label="(8, 16)", style=solid];
+"1580 slice_172" -> "1581 fill__47"  [label="(8, 4)", style=solid];
+"1582 lift_fresh_copy_48" -> "1585 fill__48"  [label="()", style=solid];
+"1583 slice_173" -> "1584 slice_174"  [label="(4, 16)", style=solid];
+"1584 slice_174" -> "1585 fill__48"  [label="(4, 8)", style=solid];
+"1586 lift_fresh_copy_49" -> "1589 fill__49"  [label="()", style=solid];
+"1587 slice_175" -> "1588 slice_176"  [label="(4, 16)", style=solid];
+"1588 slice_176" -> "1589 fill__49"  [label="(4, 4)", style=solid];
+"1590 lift_fresh_copy_50" -> "1593 fill__50"  [label="()", style=solid];
+"1591 slice_177" -> "1592 slice_178"  [label="(4, 16)", style=solid];
+"1592 slice_178" -> "1593 fill__50"  [label="(4, 4)", style=solid];
+"1594 lift_fresh_copy_51" -> "1597 fill__51"  [label="()", style=solid];
+"1595 slice_179" -> "1596 slice_180"  [label="(4, 16)", style=solid];
+"1596 slice_180" -> "1597 fill__51"  [label="(4, 8)", style=solid];
+"1598 lift_fresh_copy_52" -> "1601 fill__52"  [label="()", style=solid];
+"1599 slice_181" -> "1600 slice_182"  [label="(4, 16)", style=solid];
+"1600 slice_182" -> "1601 fill__52"  [label="(4, 4)", style=solid];
+"1602 lift_fresh_copy_53" -> "1605 fill__53"  [label="()", style=solid];
+"1603 slice_183" -> "1604 slice_184"  [label="(4, 16)", style=solid];
+"1604 slice_184" -> "1605 fill__53"  [label="(4, 4)", style=solid];
+"1606 view_62" -> "1607 permute_53"  [label="(2, 8, 2, 8)", style=solid];
+"1607 permute_53" -> "1608 reshape_51"  [label="(2, 2, 8, 8)", style=solid];
+"1608 reshape_51" -> "1609 unsqueeze_32"  [label="(4, 64)", style=solid];
+"1608 reshape_51" -> "1610 unsqueeze_33"  [label="(4, 64)", style=solid];
+"1609 unsqueeze_32" -> "1611 sub_5"  [label="(4, 1, 64)", style=solid];
+"1610 unsqueeze_33" -> "1611 sub_5"  [label="(4, 64, 1)", style=solid];
+"1611 sub_5" -> "1612 ne_5"  [label="(4, 64, 64)", style=solid];
+"1611 sub_5" -> "1613 masked_fill_10"  [label="(4, 64, 64)", style=solid];
+"1611 sub_5" -> "1614 eq_5"  [label="(4, 64, 64)", style=solid];
+"1612 ne_5" -> "1613 masked_fill_10"  [label="(4, 64, 64)", style=solid];
+"1613 masked_fill_10" -> "1615 masked_fill_11"  [label="(4, 64, 64)", style=solid];
+"1614 eq_5" -> "1615 masked_fill_11"  [label="(4, 64, 64)", style=solid];
+"1615 masked_fill_11" -> "1617 unsqueeze_34"  [label="(4, 64, 64)", style=solid];
+"1616 view_63" -> "1619 add_39"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1617 unsqueeze_34" -> "1618 unsqueeze_35"  [label="(4, 1, 64, 64)", style=solid];
+"1618 unsqueeze_35" -> "1619 add_39"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1619 add_39" -> "1620 view_64"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1620 view_64" -> "1621 softmax_11"  [label="(4, 12, 64, 64)", style=solid];
+"1621 softmax_11" -> "1622 dropout_44"  [label="(4, 12, 64, 64)", style=solid];
+"1622 dropout_44" -> "1623 matmul_23"  [label="(4, 12, 64, 64)", style=solid];
+"1623 matmul_23" -> "1624 transpose_23"  [label="(4, 12, 64, 32)", style=solid];
+"1624 transpose_23" -> "1625 reshape_52"  [label="(4, 64, 12, 32)", style=solid];
+"1625 reshape_52" -> "1626 linear_71"  [label="(4, 64, 384)", style=solid];
+"1626 linear_71" -> "1627 dropout_45"  [label="(4, 64, 384)", style=solid];
+"1627 dropout_45" -> "1628 view_65"  [label="(4, 64, 384)", style=solid];
+"1628 view_65" -> "1629 permute_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1629 permute_54" -> "1630 reshape_53"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1630 reshape_53" -> "1631 roll_11"  [label="(1, 16, 16, 384)", style=solid];
+"1631 roll_11" -> "1632 slice_185"  [label="(1, 16, 16, 384)", style=solid];
+"1632 slice_185" -> "1633 slice_186"  [label="(1, 16, 16, 384)", style=solid];
+"1633 slice_186" -> "1634 slice_187"  [label="(1, 14, 16, 384)", style=solid];
+"1634 slice_187" -> "1635 slice_188"  [label="(1, 14, 14, 384)", style=solid];
+"1635 slice_188" -> "1636 contiguous_21"  [label="(1, 14, 14, 384)", style=solid];
+"1636 contiguous_21" -> "1637 layer_norm_25"  [label="(1, 14, 14, 384)", style=solid];
+"1637 layer_norm_25" -> "1638 add_40"  [label="(1, 14, 14, 384)", style=solid];
+"1638 add_40" -> "1639 linear_72"  [label="(1, 14, 14, 384)", style=solid];
+"1638 add_40" -> "1645 add_41"  [label="(1, 14, 14, 384)", style=solid];
+"1639 linear_72" -> "1640 gelu_11"  [label="(1, 14, 14, 1536)", style=solid];
+"1640 gelu_11" -> "1641 dropout_46"  [label="(1, 14, 14, 1536)", style=solid];
+"1641 dropout_46" -> "1642 linear_73"  [label="(1, 14, 14, 1536)", style=solid];
+"1642 linear_73" -> "1643 dropout_47"  [label="(1, 14, 14, 384)", style=solid];
+"1643 dropout_47" -> "1644 layer_norm_26"  [label="(1, 14, 14, 384)", style=solid];
+"1644 layer_norm_26" -> "1645 add_41"  [label="(1, 14, 14, 384)", style=solid];
+"1645 add_41" -> "1657 pad_14"  [label="(1, 14, 14, 384)", style=solid];
+"1645 add_41" -> "1700 add_43"  [label="(1, 14, 14, 384)", style=solid];
+"1646 linear_74" -> "1647 relu__12"  [label="(1, 15, 15, 512)", style=solid];
+"1647 relu__12" -> "1648 linear_75"  [label="(1, 15, 15, 512)", style=solid];
+"1648 linear_75" -> "1649 view_66"  [label="(1, 15, 15, 12)", style=solid];
+"1649 view_66" -> "1650 index_12"  [label="(225, 12)", style=solid];
+"1650 index_12" -> "1651 view_67"  [label="(4096, 12)", style=solid];
+"1651 view_67" -> "1652 permute_55"  [label="(64, 64, 12)", style=solid];
+"1652 permute_55" -> "1653 contiguous_22"  [label="(12, 64, 64)", style=solid];
+"1653 contiguous_22" -> "1654 unsqueeze_36"  [label="(12, 64, 64)", style=solid];
+"1654 unsqueeze_36" -> "1655 sigmoid_12"  [label="(1, 12, 64, 64)", style=solid];
+"1655 sigmoid_12" -> "1656 mul_24"  [label="(1, 12, 64, 64)", style=solid];
+"1656 mul_24" -> "1683 add_42"  [label="(1, 12, 64, 64)", style=solid];
+"1657 pad_14" -> "1658 view_68"  [label="(1, 16, 16, 384)", style=solid];
+"1658 view_68" -> "1659 permute_56"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1659 permute_56" -> "1660 reshape_54"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1660 reshape_54" -> "1664 linear_76"  [label="(4, 64, 384)", style=solid];
+"1661 clone_12" -> "1662 slice_189"  [label="(1152,)", style=solid];
+"1661 clone_12" -> "1664 linear_76"  [label="(1152,)", style=solid];
+"1662 slice_189" -> "1663 zero__12"  [label="(384,)", style=solid];
+"1664 linear_76" -> "1665 reshape_55"  [label="(4, 64, 1152)", style=solid];
+"1665 reshape_55" -> "1666 permute_57"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1666 permute_57" -> "1667 select_36"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1666 permute_57" -> "1668 select_37"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1666 permute_57" -> "1669 select_38"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1667 select_36" -> "1670 linalg_vector_norm_24"  [label="(4, 12, 64, 32)", style=solid];
+"1667 select_36" -> "1672 expand_as_24"  [label="(4, 12, 64, 32)", style=solid];
+"1667 select_36" -> "1673 div_24"  [label="(4, 12, 64, 32)", style=solid];
+"1668 select_37" -> "1674 linalg_vector_norm_25"  [label="(4, 12, 64, 32)", style=solid];
+"1668 select_37" -> "1676 expand_as_25"  [label="(4, 12, 64, 32)", style=solid];
+"1668 select_37" -> "1677 div_25"  [label="(4, 12, 64, 32)", style=solid];
+"1669 select_38" -> "1686 matmul_25"  [label="(4, 12, 64, 32)", style=solid];
+"1670 linalg_vector_norm_24" -> "1671 clamp_min_24"  [label="(4, 12, 64, 1)", style=solid];
+"1671 clamp_min_24" -> "1672 expand_as_24"  [label="(4, 12, 64, 1)", style=solid];
+"1672 expand_as_24" -> "1673 div_24"  [label="(4, 12, 64, 32)", style=solid];
+"1673 div_24" -> "1679 matmul_24"  [label="(4, 12, 64, 32)", style=solid];
+"1674 linalg_vector_norm_25" -> "1675 clamp_min_25"  [label="(4, 12, 64, 1)", style=solid];
+"1675 clamp_min_25" -> "1676 expand_as_25"  [label="(4, 12, 64, 1)", style=solid];
+"1676 expand_as_25" -> "1677 div_25"  [label="(4, 12, 64, 32)", style=solid];
+"1677 div_25" -> "1678 transpose_24"  [label="(4, 12, 64, 32)", style=solid];
+"1678 transpose_24" -> "1679 matmul_24"  [label="(4, 12, 32, 64)", style=solid];
+"1679 matmul_24" -> "1682 mul_25"  [label="(4, 12, 64, 64)", style=solid];
+"1680 clamp_12" -> "1681 exp_12"  [label="(12, 1, 1)", style=solid];
+"1681 exp_12" -> "1682 mul_25"  [label="(12, 1, 1)", style=solid];
+"1682 mul_25" -> "1683 add_42"  [label="(4, 12, 64, 64)", style=solid];
+"1683 add_42" -> "1684 softmax_12"  [label="(4, 12, 64, 64)", style=solid];
+"1684 softmax_12" -> "1685 dropout_48"  [label="(4, 12, 64, 64)", style=solid];
+"1685 dropout_48" -> "1686 matmul_25"  [label="(4, 12, 64, 64)", style=solid];
+"1686 matmul_25" -> "1687 transpose_25"  [label="(4, 12, 64, 32)", style=solid];
+"1687 transpose_25" -> "1688 reshape_56"  [label="(4, 64, 12, 32)", style=solid];
+"1688 reshape_56" -> "1689 linear_77"  [label="(4, 64, 384)", style=solid];
+"1689 linear_77" -> "1690 dropout_49"  [label="(4, 64, 384)", style=solid];
+"1690 dropout_49" -> "1691 view_69"  [label="(4, 64, 384)", style=solid];
+"1691 view_69" -> "1692 permute_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1692 permute_58" -> "1693 reshape_57"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1693 reshape_57" -> "1694 slice_190"  [label="(1, 16, 16, 384)", style=solid];
+"1694 slice_190" -> "1695 slice_191"  [label="(1, 16, 16, 384)", style=solid];
+"1695 slice_191" -> "1696 slice_192"  [label="(1, 14, 16, 384)", style=solid];
+"1696 slice_192" -> "1697 slice_193"  [label="(1, 14, 14, 384)", style=solid];
+"1697 slice_193" -> "1698 contiguous_23"  [label="(1, 14, 14, 384)", style=solid];
+"1698 contiguous_23" -> "1699 layer_norm_27"  [label="(1, 14, 14, 384)", style=solid];
+"1699 layer_norm_27" -> "1700 add_43"  [label="(1, 14, 14, 384)", style=solid];
+"1700 add_43" -> "1701 linear_78"  [label="(1, 14, 14, 384)", style=solid];
+"1700 add_43" -> "1707 add_44"  [label="(1, 14, 14, 384)", style=solid];
+"1701 linear_78" -> "1702 gelu_12"  [label="(1, 14, 14, 1536)", style=solid];
+"1702 gelu_12" -> "1703 dropout_50"  [label="(1, 14, 14, 1536)", style=solid];
+"1703 dropout_50" -> "1704 linear_79"  [label="(1, 14, 14, 1536)", style=solid];
+"1704 linear_79" -> "1705 dropout_51"  [label="(1, 14, 14, 384)", style=solid];
+"1705 dropout_51" -> "1706 layer_norm_28"  [label="(1, 14, 14, 384)", style=solid];
+"1706 layer_norm_28" -> "1707 add_44"  [label="(1, 14, 14, 384)", style=solid];
+"1707 add_44" -> "1719 pad_15"  [label="(1, 14, 14, 384)", style=solid];
+"1707 add_44" -> "1816 add_47"  [label="(1, 14, 14, 384)", style=solid];
+"1708 linear_80" -> "1709 relu__13"  [label="(1, 15, 15, 512)", style=solid];
+"1709 relu__13" -> "1710 linear_81"  [label="(1, 15, 15, 512)", style=solid];
+"1710 linear_81" -> "1711 view_70"  [label="(1, 15, 15, 12)", style=solid];
+"1711 view_70" -> "1712 index_13"  [label="(225, 12)", style=solid];
+"1712 index_13" -> "1713 view_71"  [label="(4096, 12)", style=solid];
+"1713 view_71" -> "1714 permute_59"  [label="(64, 64, 12)", style=solid];
+"1714 permute_59" -> "1715 contiguous_24"  [label="(12, 64, 64)", style=solid];
+"1715 contiguous_24" -> "1716 unsqueeze_37"  [label="(12, 64, 64)", style=solid];
+"1716 unsqueeze_37" -> "1717 sigmoid_13"  [label="(1, 12, 64, 64)", style=solid];
+"1717 sigmoid_13" -> "1718 mul_26"  [label="(1, 12, 64, 64)", style=solid];
+"1718 mul_26" -> "1746 add_45"  [label="(1, 12, 64, 64)", style=solid];
+"1719 pad_15" -> "1720 roll_12"  [label="(1, 16, 16, 384)", style=solid];
+"1720 roll_12" -> "1721 view_72"  [label="(1, 16, 16, 384)", style=solid];
+"1721 view_72" -> "1722 permute_60"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1722 permute_60" -> "1723 reshape_58"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1723 reshape_58" -> "1727 linear_82"  [label="(4, 64, 384)", style=solid];
+"1723 reshape_58" -> "1747 new_zeros_6"  [label="(4, 64, 384)", style=solid];
+"1724 clone_13" -> "1725 slice_194"  [label="(1152,)", style=solid];
+"1724 clone_13" -> "1727 linear_82"  [label="(1152,)", style=solid];
+"1725 slice_194" -> "1726 zero__13"  [label="(384,)", style=solid];
+"1727 linear_82" -> "1728 reshape_59"  [label="(4, 64, 1152)", style=solid];
+"1728 reshape_59" -> "1729 permute_61"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1729 permute_61" -> "1730 select_39"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1729 permute_61" -> "1731 select_40"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1729 permute_61" -> "1732 select_41"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1730 select_39" -> "1733 linalg_vector_norm_26"  [label="(4, 12, 64, 32)", style=solid];
+"1730 select_39" -> "1735 expand_as_26"  [label="(4, 12, 64, 32)", style=solid];
+"1730 select_39" -> "1736 div_26"  [label="(4, 12, 64, 32)", style=solid];
+"1731 select_40" -> "1737 linalg_vector_norm_27"  [label="(4, 12, 64, 32)", style=solid];
+"1731 select_40" -> "1739 expand_as_27"  [label="(4, 12, 64, 32)", style=solid];
+"1731 select_40" -> "1740 div_27"  [label="(4, 12, 64, 32)", style=solid];
+"1732 select_41" -> "1801 matmul_27"  [label="(4, 12, 64, 32)", style=solid];
+"1733 linalg_vector_norm_26" -> "1734 clamp_min_26"  [label="(4, 12, 64, 1)", style=solid];
+"1734 clamp_min_26" -> "1735 expand_as_26"  [label="(4, 12, 64, 1)", style=solid];
+"1735 expand_as_26" -> "1736 div_26"  [label="(4, 12, 64, 32)", style=solid];
+"1736 div_26" -> "1742 matmul_26"  [label="(4, 12, 64, 32)", style=solid];
+"1737 linalg_vector_norm_27" -> "1738 clamp_min_27"  [label="(4, 12, 64, 1)", style=solid];
+"1738 clamp_min_27" -> "1739 expand_as_27"  [label="(4, 12, 64, 1)", style=solid];
+"1739 expand_as_27" -> "1740 div_27"  [label="(4, 12, 64, 32)", style=solid];
+"1740 div_27" -> "1741 transpose_26"  [label="(4, 12, 64, 32)", style=solid];
+"1741 transpose_26" -> "1742 matmul_26"  [label="(4, 12, 32, 64)", style=solid];
+"1742 matmul_26" -> "1745 mul_27"  [label="(4, 12, 64, 64)", style=solid];
+"1743 clamp_13" -> "1744 exp_13"  [label="(12, 1, 1)", style=solid];
+"1744 exp_13" -> "1745 mul_27"  [label="(12, 1, 1)", style=solid];
+"1745 mul_27" -> "1746 add_45"  [label="(4, 12, 64, 64)", style=solid];
+"1746 add_45" -> "1794 view_74"  [label="(4, 12, 64, 64)", style=solid];
+"1747 new_zeros_6" -> "1749 slice_195"  [label="(16, 16)", style=solid];
+"1747 new_zeros_6" -> "1753 slice_197"  [label="(16, 16)", style=solid];
+"1747 new_zeros_6" -> "1757 slice_199"  [label="(16, 16)", style=solid];
+"1747 new_zeros_6" -> "1761 slice_201"  [label="(16, 16)", style=solid];
+"1747 new_zeros_6" -> "1765 slice_203"  [label="(16, 16)", style=solid];
+"1747 new_zeros_6" -> "1769 slice_205"  [label="(16, 16)", style=solid];
+"1747 new_zeros_6" -> "1773 slice_207"  [label="(16, 16)", style=solid];
+"1747 new_zeros_6" -> "1777 slice_209"  [label="(16, 16)", style=solid];
+"1747 new_zeros_6" -> "1781 slice_211"  [label="(16, 16)", style=solid];
+"1747 new_zeros_6" -> "1784 view_73"  [label="(16, 16)", style=solid];
+"1748 lift_fresh_copy_54" -> "1751 fill__54"  [label="()", style=solid];
+"1749 slice_195" -> "1750 slice_196"  [label="(8, 16)", style=solid];
+"1750 slice_196" -> "1751 fill__54"  [label="(8, 8)", style=solid];
+"1752 lift_fresh_copy_55" -> "1755 fill__55"  [label="()", style=solid];
+"1753 slice_197" -> "1754 slice_198"  [label="(8, 16)", style=solid];
+"1754 slice_198" -> "1755 fill__55"  [label="(8, 4)", style=solid];
+"1756 lift_fresh_copy_56" -> "1759 fill__56"  [label="()", style=solid];
+"1757 slice_199" -> "1758 slice_200"  [label="(8, 16)", style=solid];
+"1758 slice_200" -> "1759 fill__56"  [label="(8, 4)", style=solid];
+"1760 lift_fresh_copy_57" -> "1763 fill__57"  [label="()", style=solid];
+"1761 slice_201" -> "1762 slice_202"  [label="(4, 16)", style=solid];
+"1762 slice_202" -> "1763 fill__57"  [label="(4, 8)", style=solid];
+"1764 lift_fresh_copy_58" -> "1767 fill__58"  [label="()", style=solid];
+"1765 slice_203" -> "1766 slice_204"  [label="(4, 16)", style=solid];
+"1766 slice_204" -> "1767 fill__58"  [label="(4, 4)", style=solid];
+"1768 lift_fresh_copy_59" -> "1771 fill__59"  [label="()", style=solid];
+"1769 slice_205" -> "1770 slice_206"  [label="(4, 16)", style=solid];
+"1770 slice_206" -> "1771 fill__59"  [label="(4, 4)", style=solid];
+"1772 lift_fresh_copy_60" -> "1775 fill__60"  [label="()", style=solid];
+"1773 slice_207" -> "1774 slice_208"  [label="(4, 16)", style=solid];
+"1774 slice_208" -> "1775 fill__60"  [label="(4, 8)", style=solid];
+"1776 lift_fresh_copy_61" -> "1779 fill__61"  [label="()", style=solid];
+"1777 slice_209" -> "1778 slice_210"  [label="(4, 16)", style=solid];
+"1778 slice_210" -> "1779 fill__61"  [label="(4, 4)", style=solid];
+"1780 lift_fresh_copy_62" -> "1783 fill__62"  [label="()", style=solid];
+"1781 slice_211" -> "1782 slice_212"  [label="(4, 16)", style=solid];
+"1782 slice_212" -> "1783 fill__62"  [label="(4, 4)", style=solid];
+"1784 view_73" -> "1785 permute_62"  [label="(2, 8, 2, 8)", style=solid];
+"1785 permute_62" -> "1786 reshape_60"  [label="(2, 2, 8, 8)", style=solid];
+"1786 reshape_60" -> "1787 unsqueeze_38"  [label="(4, 64)", style=solid];
+"1786 reshape_60" -> "1788 unsqueeze_39"  [label="(4, 64)", style=solid];
+"1787 unsqueeze_38" -> "1789 sub_6"  [label="(4, 1, 64)", style=solid];
+"1788 unsqueeze_39" -> "1789 sub_6"  [label="(4, 64, 1)", style=solid];
+"1789 sub_6" -> "1790 ne_6"  [label="(4, 64, 64)", style=solid];
+"1789 sub_6" -> "1791 masked_fill_12"  [label="(4, 64, 64)", style=solid];
+"1789 sub_6" -> "1792 eq_6"  [label="(4, 64, 64)", style=solid];
+"1790 ne_6" -> "1791 masked_fill_12"  [label="(4, 64, 64)", style=solid];
+"1791 masked_fill_12" -> "1793 masked_fill_13"  [label="(4, 64, 64)", style=solid];
+"1792 eq_6" -> "1793 masked_fill_13"  [label="(4, 64, 64)", style=solid];
+"1793 masked_fill_13" -> "1795 unsqueeze_40"  [label="(4, 64, 64)", style=solid];
+"1794 view_74" -> "1797 add_46"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1795 unsqueeze_40" -> "1796 unsqueeze_41"  [label="(4, 1, 64, 64)", style=solid];
+"1796 unsqueeze_41" -> "1797 add_46"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1797 add_46" -> "1798 view_75"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1798 view_75" -> "1799 softmax_13"  [label="(4, 12, 64, 64)", style=solid];
+"1799 softmax_13" -> "1800 dropout_52"  [label="(4, 12, 64, 64)", style=solid];
+"1800 dropout_52" -> "1801 matmul_27"  [label="(4, 12, 64, 64)", style=solid];
+"1801 matmul_27" -> "1802 transpose_27"  [label="(4, 12, 64, 32)", style=solid];
+"1802 transpose_27" -> "1803 reshape_61"  [label="(4, 64, 12, 32)", style=solid];
+"1803 reshape_61" -> "1804 linear_83"  [label="(4, 64, 384)", style=solid];
+"1804 linear_83" -> "1805 dropout_53"  [label="(4, 64, 384)", style=solid];
+"1805 dropout_53" -> "1806 view_76"  [label="(4, 64, 384)", style=solid];
+"1806 view_76" -> "1807 permute_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1807 permute_63" -> "1808 reshape_62"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1808 reshape_62" -> "1809 roll_13"  [label="(1, 16, 16, 384)", style=solid];
+"1809 roll_13" -> "1810 slice_213"  [label="(1, 16, 16, 384)", style=solid];
+"1810 slice_213" -> "1811 slice_214"  [label="(1, 16, 16, 384)", style=solid];
+"1811 slice_214" -> "1812 slice_215"  [label="(1, 14, 16, 384)", style=solid];
+"1812 slice_215" -> "1813 slice_216"  [label="(1, 14, 14, 384)", style=solid];
+"1813 slice_216" -> "1814 contiguous_25"  [label="(1, 14, 14, 384)", style=solid];
+"1814 contiguous_25" -> "1815 layer_norm_29"  [label="(1, 14, 14, 384)", style=solid];
+"1815 layer_norm_29" -> "1816 add_47"  [label="(1, 14, 14, 384)", style=solid];
+"1816 add_47" -> "1817 linear_84"  [label="(1, 14, 14, 384)", style=solid];
+"1816 add_47" -> "1823 add_48"  [label="(1, 14, 14, 384)", style=solid];
+"1817 linear_84" -> "1818 gelu_13"  [label="(1, 14, 14, 1536)", style=solid];
+"1818 gelu_13" -> "1819 dropout_54"  [label="(1, 14, 14, 1536)", style=solid];
+"1819 dropout_54" -> "1820 linear_85"  [label="(1, 14, 14, 1536)", style=solid];
+"1820 linear_85" -> "1821 dropout_55"  [label="(1, 14, 14, 384)", style=solid];
+"1821 dropout_55" -> "1822 layer_norm_30"  [label="(1, 14, 14, 384)", style=solid];
+"1822 layer_norm_30" -> "1823 add_48"  [label="(1, 14, 14, 384)", style=solid];
+"1823 add_48" -> "1835 pad_16"  [label="(1, 14, 14, 384)", style=solid];
+"1823 add_48" -> "1878 add_50"  [label="(1, 14, 14, 384)", style=solid];
+"1824 linear_86" -> "1825 relu__14"  [label="(1, 15, 15, 512)", style=solid];
+"1825 relu__14" -> "1826 linear_87"  [label="(1, 15, 15, 512)", style=solid];
+"1826 linear_87" -> "1827 view_77"  [label="(1, 15, 15, 12)", style=solid];
+"1827 view_77" -> "1828 index_14"  [label="(225, 12)", style=solid];
+"1828 index_14" -> "1829 view_78"  [label="(4096, 12)", style=solid];
+"1829 view_78" -> "1830 permute_64"  [label="(64, 64, 12)", style=solid];
+"1830 permute_64" -> "1831 contiguous_26"  [label="(12, 64, 64)", style=solid];
+"1831 contiguous_26" -> "1832 unsqueeze_42"  [label="(12, 64, 64)", style=solid];
+"1832 unsqueeze_42" -> "1833 sigmoid_14"  [label="(1, 12, 64, 64)", style=solid];
+"1833 sigmoid_14" -> "1834 mul_28"  [label="(1, 12, 64, 64)", style=solid];
+"1834 mul_28" -> "1861 add_49"  [label="(1, 12, 64, 64)", style=solid];
+"1835 pad_16" -> "1836 view_79"  [label="(1, 16, 16, 384)", style=solid];
+"1836 view_79" -> "1837 permute_65"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1837 permute_65" -> "1838 reshape_63"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1838 reshape_63" -> "1842 linear_88"  [label="(4, 64, 384)", style=solid];
+"1839 clone_14" -> "1840 slice_217"  [label="(1152,)", style=solid];
+"1839 clone_14" -> "1842 linear_88"  [label="(1152,)", style=solid];
+"1840 slice_217" -> "1841 zero__14"  [label="(384,)", style=solid];
+"1842 linear_88" -> "1843 reshape_64"  [label="(4, 64, 1152)", style=solid];
+"1843 reshape_64" -> "1844 permute_66"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1844 permute_66" -> "1845 select_42"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1844 permute_66" -> "1846 select_43"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1844 permute_66" -> "1847 select_44"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1845 select_42" -> "1848 linalg_vector_norm_28"  [label="(4, 12, 64, 32)", style=solid];
+"1845 select_42" -> "1850 expand_as_28"  [label="(4, 12, 64, 32)", style=solid];
+"1845 select_42" -> "1851 div_28"  [label="(4, 12, 64, 32)", style=solid];
+"1846 select_43" -> "1852 linalg_vector_norm_29"  [label="(4, 12, 64, 32)", style=solid];
+"1846 select_43" -> "1854 expand_as_29"  [label="(4, 12, 64, 32)", style=solid];
+"1846 select_43" -> "1855 div_29"  [label="(4, 12, 64, 32)", style=solid];
+"1847 select_44" -> "1864 matmul_29"  [label="(4, 12, 64, 32)", style=solid];
+"1848 linalg_vector_norm_28" -> "1849 clamp_min_28"  [label="(4, 12, 64, 1)", style=solid];
+"1849 clamp_min_28" -> "1850 expand_as_28"  [label="(4, 12, 64, 1)", style=solid];
+"1850 expand_as_28" -> "1851 div_28"  [label="(4, 12, 64, 32)", style=solid];
+"1851 div_28" -> "1857 matmul_28"  [label="(4, 12, 64, 32)", style=solid];
+"1852 linalg_vector_norm_29" -> "1853 clamp_min_29"  [label="(4, 12, 64, 1)", style=solid];
+"1853 clamp_min_29" -> "1854 expand_as_29"  [label="(4, 12, 64, 1)", style=solid];
+"1854 expand_as_29" -> "1855 div_29"  [label="(4, 12, 64, 32)", style=solid];
+"1855 div_29" -> "1856 transpose_28"  [label="(4, 12, 64, 32)", style=solid];
+"1856 transpose_28" -> "1857 matmul_28"  [label="(4, 12, 32, 64)", style=solid];
+"1857 matmul_28" -> "1860 mul_29"  [label="(4, 12, 64, 64)", style=solid];
+"1858 clamp_14" -> "1859 exp_14"  [label="(12, 1, 1)", style=solid];
+"1859 exp_14" -> "1860 mul_29"  [label="(12, 1, 1)", style=solid];
+"1860 mul_29" -> "1861 add_49"  [label="(4, 12, 64, 64)", style=solid];
+"1861 add_49" -> "1862 softmax_14"  [label="(4, 12, 64, 64)", style=solid];
+"1862 softmax_14" -> "1863 dropout_56"  [label="(4, 12, 64, 64)", style=solid];
+"1863 dropout_56" -> "1864 matmul_29"  [label="(4, 12, 64, 64)", style=solid];
+"1864 matmul_29" -> "1865 transpose_29"  [label="(4, 12, 64, 32)", style=solid];
+"1865 transpose_29" -> "1866 reshape_65"  [label="(4, 64, 12, 32)", style=solid];
+"1866 reshape_65" -> "1867 linear_89"  [label="(4, 64, 384)", style=solid];
+"1867 linear_89" -> "1868 dropout_57"  [label="(4, 64, 384)", style=solid];
+"1868 dropout_57" -> "1869 view_80"  [label="(4, 64, 384)", style=solid];
+"1869 view_80" -> "1870 permute_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1870 permute_67" -> "1871 reshape_66"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1871 reshape_66" -> "1872 slice_218"  [label="(1, 16, 16, 384)", style=solid];
+"1872 slice_218" -> "1873 slice_219"  [label="(1, 16, 16, 384)", style=solid];
+"1873 slice_219" -> "1874 slice_220"  [label="(1, 14, 16, 384)", style=solid];
+"1874 slice_220" -> "1875 slice_221"  [label="(1, 14, 14, 384)", style=solid];
+"1875 slice_221" -> "1876 contiguous_27"  [label="(1, 14, 14, 384)", style=solid];
+"1876 contiguous_27" -> "1877 layer_norm_31"  [label="(1, 14, 14, 384)", style=solid];
+"1877 layer_norm_31" -> "1878 add_50"  [label="(1, 14, 14, 384)", style=solid];
+"1878 add_50" -> "1879 linear_90"  [label="(1, 14, 14, 384)", style=solid];
+"1878 add_50" -> "1885 add_51"  [label="(1, 14, 14, 384)", style=solid];
+"1879 linear_90" -> "1880 gelu_14"  [label="(1, 14, 14, 1536)", style=solid];
+"1880 gelu_14" -> "1881 dropout_58"  [label="(1, 14, 14, 1536)", style=solid];
+"1881 dropout_58" -> "1882 linear_91"  [label="(1, 14, 14, 1536)", style=solid];
+"1882 linear_91" -> "1883 dropout_59"  [label="(1, 14, 14, 384)", style=solid];
+"1883 dropout_59" -> "1884 layer_norm_32"  [label="(1, 14, 14, 384)", style=solid];
+"1884 layer_norm_32" -> "1885 add_51"  [label="(1, 14, 14, 384)", style=solid];
+"1885 add_51" -> "1897 pad_17"  [label="(1, 14, 14, 384)", style=solid];
+"1885 add_51" -> "1994 add_54"  [label="(1, 14, 14, 384)", style=solid];
+"1886 linear_92" -> "1887 relu__15"  [label="(1, 15, 15, 512)", style=solid];
+"1887 relu__15" -> "1888 linear_93"  [label="(1, 15, 15, 512)", style=solid];
+"1888 linear_93" -> "1889 view_81"  [label="(1, 15, 15, 12)", style=solid];
+"1889 view_81" -> "1890 index_15"  [label="(225, 12)", style=solid];
+"1890 index_15" -> "1891 view_82"  [label="(4096, 12)", style=solid];
+"1891 view_82" -> "1892 permute_68"  [label="(64, 64, 12)", style=solid];
+"1892 permute_68" -> "1893 contiguous_28"  [label="(12, 64, 64)", style=solid];
+"1893 contiguous_28" -> "1894 unsqueeze_43"  [label="(12, 64, 64)", style=solid];
+"1894 unsqueeze_43" -> "1895 sigmoid_15"  [label="(1, 12, 64, 64)", style=solid];
+"1895 sigmoid_15" -> "1896 mul_30"  [label="(1, 12, 64, 64)", style=solid];
+"1896 mul_30" -> "1924 add_52"  [label="(1, 12, 64, 64)", style=solid];
+"1897 pad_17" -> "1898 roll_14"  [label="(1, 16, 16, 384)", style=solid];
+"1898 roll_14" -> "1899 view_83"  [label="(1, 16, 16, 384)", style=solid];
+"1899 view_83" -> "1900 permute_69"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1900 permute_69" -> "1901 reshape_67"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1901 reshape_67" -> "1905 linear_94"  [label="(4, 64, 384)", style=solid];
+"1901 reshape_67" -> "1925 new_zeros_7"  [label="(4, 64, 384)", style=solid];
+"1902 clone_15" -> "1903 slice_222"  [label="(1152,)", style=solid];
+"1902 clone_15" -> "1905 linear_94"  [label="(1152,)", style=solid];
+"1903 slice_222" -> "1904 zero__15"  [label="(384,)", style=solid];
+"1905 linear_94" -> "1906 reshape_68"  [label="(4, 64, 1152)", style=solid];
+"1906 reshape_68" -> "1907 permute_70"  [label="(4, 64, 3, 12, 32)", style=solid];
+"1907 permute_70" -> "1908 select_45"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1907 permute_70" -> "1909 select_46"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1907 permute_70" -> "1910 select_47"  [label="(3, 4, 12, 64, 32)", style=solid];
+"1908 select_45" -> "1911 linalg_vector_norm_30"  [label="(4, 12, 64, 32)", style=solid];
+"1908 select_45" -> "1913 expand_as_30"  [label="(4, 12, 64, 32)", style=solid];
+"1908 select_45" -> "1914 div_30"  [label="(4, 12, 64, 32)", style=solid];
+"1909 select_46" -> "1915 linalg_vector_norm_31"  [label="(4, 12, 64, 32)", style=solid];
+"1909 select_46" -> "1917 expand_as_31"  [label="(4, 12, 64, 32)", style=solid];
+"1909 select_46" -> "1918 div_31"  [label="(4, 12, 64, 32)", style=solid];
+"1910 select_47" -> "1979 matmul_31"  [label="(4, 12, 64, 32)", style=solid];
+"1911 linalg_vector_norm_30" -> "1912 clamp_min_30"  [label="(4, 12, 64, 1)", style=solid];
+"1912 clamp_min_30" -> "1913 expand_as_30"  [label="(4, 12, 64, 1)", style=solid];
+"1913 expand_as_30" -> "1914 div_30"  [label="(4, 12, 64, 32)", style=solid];
+"1914 div_30" -> "1920 matmul_30"  [label="(4, 12, 64, 32)", style=solid];
+"1915 linalg_vector_norm_31" -> "1916 clamp_min_31"  [label="(4, 12, 64, 1)", style=solid];
+"1916 clamp_min_31" -> "1917 expand_as_31"  [label="(4, 12, 64, 1)", style=solid];
+"1917 expand_as_31" -> "1918 div_31"  [label="(4, 12, 64, 32)", style=solid];
+"1918 div_31" -> "1919 transpose_30"  [label="(4, 12, 64, 32)", style=solid];
+"1919 transpose_30" -> "1920 matmul_30"  [label="(4, 12, 32, 64)", style=solid];
+"1920 matmul_30" -> "1923 mul_31"  [label="(4, 12, 64, 64)", style=solid];
+"1921 clamp_15" -> "1922 exp_15"  [label="(12, 1, 1)", style=solid];
+"1922 exp_15" -> "1923 mul_31"  [label="(12, 1, 1)", style=solid];
+"1923 mul_31" -> "1924 add_52"  [label="(4, 12, 64, 64)", style=solid];
+"1924 add_52" -> "1972 view_85"  [label="(4, 12, 64, 64)", style=solid];
+"1925 new_zeros_7" -> "1927 slice_223"  [label="(16, 16)", style=solid];
+"1925 new_zeros_7" -> "1931 slice_225"  [label="(16, 16)", style=solid];
+"1925 new_zeros_7" -> "1935 slice_227"  [label="(16, 16)", style=solid];
+"1925 new_zeros_7" -> "1939 slice_229"  [label="(16, 16)", style=solid];
+"1925 new_zeros_7" -> "1943 slice_231"  [label="(16, 16)", style=solid];
+"1925 new_zeros_7" -> "1947 slice_233"  [label="(16, 16)", style=solid];
+"1925 new_zeros_7" -> "1951 slice_235"  [label="(16, 16)", style=solid];
+"1925 new_zeros_7" -> "1955 slice_237"  [label="(16, 16)", style=solid];
+"1925 new_zeros_7" -> "1959 slice_239"  [label="(16, 16)", style=solid];
+"1925 new_zeros_7" -> "1962 view_84"  [label="(16, 16)", style=solid];
+"1926 lift_fresh_copy_63" -> "1929 fill__63"  [label="()", style=solid];
+"1927 slice_223" -> "1928 slice_224"  [label="(8, 16)", style=solid];
+"1928 slice_224" -> "1929 fill__63"  [label="(8, 8)", style=solid];
+"1930 lift_fresh_copy_64" -> "1933 fill__64"  [label="()", style=solid];
+"1931 slice_225" -> "1932 slice_226"  [label="(8, 16)", style=solid];
+"1932 slice_226" -> "1933 fill__64"  [label="(8, 4)", style=solid];
+"1934 lift_fresh_copy_65" -> "1937 fill__65"  [label="()", style=solid];
+"1935 slice_227" -> "1936 slice_228"  [label="(8, 16)", style=solid];
+"1936 slice_228" -> "1937 fill__65"  [label="(8, 4)", style=solid];
+"1938 lift_fresh_copy_66" -> "1941 fill__66"  [label="()", style=solid];
+"1939 slice_229" -> "1940 slice_230"  [label="(4, 16)", style=solid];
+"1940 slice_230" -> "1941 fill__66"  [label="(4, 8)", style=solid];
+"1942 lift_fresh_copy_67" -> "1945 fill__67"  [label="()", style=solid];
+"1943 slice_231" -> "1944 slice_232"  [label="(4, 16)", style=solid];
+"1944 slice_232" -> "1945 fill__67"  [label="(4, 4)", style=solid];
+"1946 lift_fresh_copy_68" -> "1949 fill__68"  [label="()", style=solid];
+"1947 slice_233" -> "1948 slice_234"  [label="(4, 16)", style=solid];
+"1948 slice_234" -> "1949 fill__68"  [label="(4, 4)", style=solid];
+"1950 lift_fresh_copy_69" -> "1953 fill__69"  [label="()", style=solid];
+"1951 slice_235" -> "1952 slice_236"  [label="(4, 16)", style=solid];
+"1952 slice_236" -> "1953 fill__69"  [label="(4, 8)", style=solid];
+"1954 lift_fresh_copy_70" -> "1957 fill__70"  [label="()", style=solid];
+"1955 slice_237" -> "1956 slice_238"  [label="(4, 16)", style=solid];
+"1956 slice_238" -> "1957 fill__70"  [label="(4, 4)", style=solid];
+"1958 lift_fresh_copy_71" -> "1961 fill__71"  [label="()", style=solid];
+"1959 slice_239" -> "1960 slice_240"  [label="(4, 16)", style=solid];
+"1960 slice_240" -> "1961 fill__71"  [label="(4, 4)", style=solid];
+"1962 view_84" -> "1963 permute_71"  [label="(2, 8, 2, 8)", style=solid];
+"1963 permute_71" -> "1964 reshape_69"  [label="(2, 2, 8, 8)", style=solid];
+"1964 reshape_69" -> "1965 unsqueeze_44"  [label="(4, 64)", style=solid];
+"1964 reshape_69" -> "1966 unsqueeze_45"  [label="(4, 64)", style=solid];
+"1965 unsqueeze_44" -> "1967 sub_7"  [label="(4, 1, 64)", style=solid];
+"1966 unsqueeze_45" -> "1967 sub_7"  [label="(4, 64, 1)", style=solid];
+"1967 sub_7" -> "1968 ne_7"  [label="(4, 64, 64)", style=solid];
+"1967 sub_7" -> "1969 masked_fill_14"  [label="(4, 64, 64)", style=solid];
+"1967 sub_7" -> "1970 eq_7"  [label="(4, 64, 64)", style=solid];
+"1968 ne_7" -> "1969 masked_fill_14"  [label="(4, 64, 64)", style=solid];
+"1969 masked_fill_14" -> "1971 masked_fill_15"  [label="(4, 64, 64)", style=solid];
+"1970 eq_7" -> "1971 masked_fill_15"  [label="(4, 64, 64)", style=solid];
+"1971 masked_fill_15" -> "1973 unsqueeze_46"  [label="(4, 64, 64)", style=solid];
+"1972 view_85" -> "1975 add_53"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1973 unsqueeze_46" -> "1974 unsqueeze_47"  [label="(4, 1, 64, 64)", style=solid];
+"1974 unsqueeze_47" -> "1975 add_53"  [label="(1, 4, 1, 64, 64)", style=solid];
+"1975 add_53" -> "1976 view_86"  [label="(1, 4, 12, 64, 64)", style=solid];
+"1976 view_86" -> "1977 softmax_15"  [label="(4, 12, 64, 64)", style=solid];
+"1977 softmax_15" -> "1978 dropout_60"  [label="(4, 12, 64, 64)", style=solid];
+"1978 dropout_60" -> "1979 matmul_31"  [label="(4, 12, 64, 64)", style=solid];
+"1979 matmul_31" -> "1980 transpose_31"  [label="(4, 12, 64, 32)", style=solid];
+"1980 transpose_31" -> "1981 reshape_70"  [label="(4, 64, 12, 32)", style=solid];
+"1981 reshape_70" -> "1982 linear_95"  [label="(4, 64, 384)", style=solid];
+"1982 linear_95" -> "1983 dropout_61"  [label="(4, 64, 384)", style=solid];
+"1983 dropout_61" -> "1984 view_87"  [label="(4, 64, 384)", style=solid];
+"1984 view_87" -> "1985 permute_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"1985 permute_72" -> "1986 reshape_71"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"1986 reshape_71" -> "1987 roll_15"  [label="(1, 16, 16, 384)", style=solid];
+"1987 roll_15" -> "1988 slice_241"  [label="(1, 16, 16, 384)", style=solid];
+"1988 slice_241" -> "1989 slice_242"  [label="(1, 16, 16, 384)", style=solid];
+"1989 slice_242" -> "1990 slice_243"  [label="(1, 14, 16, 384)", style=solid];
+"1990 slice_243" -> "1991 slice_244"  [label="(1, 14, 14, 384)", style=solid];
+"1991 slice_244" -> "1992 contiguous_29"  [label="(1, 14, 14, 384)", style=solid];
+"1992 contiguous_29" -> "1993 layer_norm_33"  [label="(1, 14, 14, 384)", style=solid];
+"1993 layer_norm_33" -> "1994 add_54"  [label="(1, 14, 14, 384)", style=solid];
+"1994 add_54" -> "1995 linear_96"  [label="(1, 14, 14, 384)", style=solid];
+"1994 add_54" -> "2001 add_55"  [label="(1, 14, 14, 384)", style=solid];
+"1995 linear_96" -> "1996 gelu_15"  [label="(1, 14, 14, 1536)", style=solid];
+"1996 gelu_15" -> "1997 dropout_62"  [label="(1, 14, 14, 1536)", style=solid];
+"1997 dropout_62" -> "1998 linear_97"  [label="(1, 14, 14, 1536)", style=solid];
+"1998 linear_97" -> "1999 dropout_63"  [label="(1, 14, 14, 384)", style=solid];
+"1999 dropout_63" -> "2000 layer_norm_34"  [label="(1, 14, 14, 384)", style=solid];
+"2000 layer_norm_34" -> "2001 add_55"  [label="(1, 14, 14, 384)", style=solid];
+"2001 add_55" -> "2013 pad_18"  [label="(1, 14, 14, 384)", style=solid];
+"2001 add_55" -> "2056 add_57"  [label="(1, 14, 14, 384)", style=solid];
+"2002 linear_98" -> "2003 relu__16"  [label="(1, 15, 15, 512)", style=solid];
+"2003 relu__16" -> "2004 linear_99"  [label="(1, 15, 15, 512)", style=solid];
+"2004 linear_99" -> "2005 view_88"  [label="(1, 15, 15, 12)", style=solid];
+"2005 view_88" -> "2006 index_16"  [label="(225, 12)", style=solid];
+"2006 index_16" -> "2007 view_89"  [label="(4096, 12)", style=solid];
+"2007 view_89" -> "2008 permute_73"  [label="(64, 64, 12)", style=solid];
+"2008 permute_73" -> "2009 contiguous_30"  [label="(12, 64, 64)", style=solid];
+"2009 contiguous_30" -> "2010 unsqueeze_48"  [label="(12, 64, 64)", style=solid];
+"2010 unsqueeze_48" -> "2011 sigmoid_16"  [label="(1, 12, 64, 64)", style=solid];
+"2011 sigmoid_16" -> "2012 mul_32"  [label="(1, 12, 64, 64)", style=solid];
+"2012 mul_32" -> "2039 add_56"  [label="(1, 12, 64, 64)", style=solid];
+"2013 pad_18" -> "2014 view_90"  [label="(1, 16, 16, 384)", style=solid];
+"2014 view_90" -> "2015 permute_74"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2015 permute_74" -> "2016 reshape_72"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2016 reshape_72" -> "2020 linear_100"  [label="(4, 64, 384)", style=solid];
+"2017 clone_16" -> "2018 slice_245"  [label="(1152,)", style=solid];
+"2017 clone_16" -> "2020 linear_100"  [label="(1152,)", style=solid];
+"2018 slice_245" -> "2019 zero__16"  [label="(384,)", style=solid];
+"2020 linear_100" -> "2021 reshape_73"  [label="(4, 64, 1152)", style=solid];
+"2021 reshape_73" -> "2022 permute_75"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2022 permute_75" -> "2023 select_48"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2022 permute_75" -> "2024 select_49"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2022 permute_75" -> "2025 select_50"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2023 select_48" -> "2026 linalg_vector_norm_32"  [label="(4, 12, 64, 32)", style=solid];
+"2023 select_48" -> "2028 expand_as_32"  [label="(4, 12, 64, 32)", style=solid];
+"2023 select_48" -> "2029 div_32"  [label="(4, 12, 64, 32)", style=solid];
+"2024 select_49" -> "2030 linalg_vector_norm_33"  [label="(4, 12, 64, 32)", style=solid];
+"2024 select_49" -> "2032 expand_as_33"  [label="(4, 12, 64, 32)", style=solid];
+"2024 select_49" -> "2033 div_33"  [label="(4, 12, 64, 32)", style=solid];
+"2025 select_50" -> "2042 matmul_33"  [label="(4, 12, 64, 32)", style=solid];
+"2026 linalg_vector_norm_32" -> "2027 clamp_min_32"  [label="(4, 12, 64, 1)", style=solid];
+"2027 clamp_min_32" -> "2028 expand_as_32"  [label="(4, 12, 64, 1)", style=solid];
+"2028 expand_as_32" -> "2029 div_32"  [label="(4, 12, 64, 32)", style=solid];
+"2029 div_32" -> "2035 matmul_32"  [label="(4, 12, 64, 32)", style=solid];
+"2030 linalg_vector_norm_33" -> "2031 clamp_min_33"  [label="(4, 12, 64, 1)", style=solid];
+"2031 clamp_min_33" -> "2032 expand_as_33"  [label="(4, 12, 64, 1)", style=solid];
+"2032 expand_as_33" -> "2033 div_33"  [label="(4, 12, 64, 32)", style=solid];
+"2033 div_33" -> "2034 transpose_32"  [label="(4, 12, 64, 32)", style=solid];
+"2034 transpose_32" -> "2035 matmul_32"  [label="(4, 12, 32, 64)", style=solid];
+"2035 matmul_32" -> "2038 mul_33"  [label="(4, 12, 64, 64)", style=solid];
+"2036 clamp_16" -> "2037 exp_16"  [label="(12, 1, 1)", style=solid];
+"2037 exp_16" -> "2038 mul_33"  [label="(12, 1, 1)", style=solid];
+"2038 mul_33" -> "2039 add_56"  [label="(4, 12, 64, 64)", style=solid];
+"2039 add_56" -> "2040 softmax_16"  [label="(4, 12, 64, 64)", style=solid];
+"2040 softmax_16" -> "2041 dropout_64"  [label="(4, 12, 64, 64)", style=solid];
+"2041 dropout_64" -> "2042 matmul_33"  [label="(4, 12, 64, 64)", style=solid];
+"2042 matmul_33" -> "2043 transpose_33"  [label="(4, 12, 64, 32)", style=solid];
+"2043 transpose_33" -> "2044 reshape_74"  [label="(4, 64, 12, 32)", style=solid];
+"2044 reshape_74" -> "2045 linear_101"  [label="(4, 64, 384)", style=solid];
+"2045 linear_101" -> "2046 dropout_65"  [label="(4, 64, 384)", style=solid];
+"2046 dropout_65" -> "2047 view_91"  [label="(4, 64, 384)", style=solid];
+"2047 view_91" -> "2048 permute_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2048 permute_76" -> "2049 reshape_75"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2049 reshape_75" -> "2050 slice_246"  [label="(1, 16, 16, 384)", style=solid];
+"2050 slice_246" -> "2051 slice_247"  [label="(1, 16, 16, 384)", style=solid];
+"2051 slice_247" -> "2052 slice_248"  [label="(1, 14, 16, 384)", style=solid];
+"2052 slice_248" -> "2053 slice_249"  [label="(1, 14, 14, 384)", style=solid];
+"2053 slice_249" -> "2054 contiguous_31"  [label="(1, 14, 14, 384)", style=solid];
+"2054 contiguous_31" -> "2055 layer_norm_35"  [label="(1, 14, 14, 384)", style=solid];
+"2055 layer_norm_35" -> "2056 add_57"  [label="(1, 14, 14, 384)", style=solid];
+"2056 add_57" -> "2057 linear_102"  [label="(1, 14, 14, 384)", style=solid];
+"2056 add_57" -> "2063 add_58"  [label="(1, 14, 14, 384)", style=solid];
+"2057 linear_102" -> "2058 gelu_16"  [label="(1, 14, 14, 1536)", style=solid];
+"2058 gelu_16" -> "2059 dropout_66"  [label="(1, 14, 14, 1536)", style=solid];
+"2059 dropout_66" -> "2060 linear_103"  [label="(1, 14, 14, 1536)", style=solid];
+"2060 linear_103" -> "2061 dropout_67"  [label="(1, 14, 14, 384)", style=solid];
+"2061 dropout_67" -> "2062 layer_norm_36"  [label="(1, 14, 14, 384)", style=solid];
+"2062 layer_norm_36" -> "2063 add_58"  [label="(1, 14, 14, 384)", style=solid];
+"2063 add_58" -> "2075 pad_19"  [label="(1, 14, 14, 384)", style=solid];
+"2063 add_58" -> "2172 add_61"  [label="(1, 14, 14, 384)", style=solid];
+"2064 linear_104" -> "2065 relu__17"  [label="(1, 15, 15, 512)", style=solid];
+"2065 relu__17" -> "2066 linear_105"  [label="(1, 15, 15, 512)", style=solid];
+"2066 linear_105" -> "2067 view_92"  [label="(1, 15, 15, 12)", style=solid];
+"2067 view_92" -> "2068 index_17"  [label="(225, 12)", style=solid];
+"2068 index_17" -> "2069 view_93"  [label="(4096, 12)", style=solid];
+"2069 view_93" -> "2070 permute_77"  [label="(64, 64, 12)", style=solid];
+"2070 permute_77" -> "2071 contiguous_32"  [label="(12, 64, 64)", style=solid];
+"2071 contiguous_32" -> "2072 unsqueeze_49"  [label="(12, 64, 64)", style=solid];
+"2072 unsqueeze_49" -> "2073 sigmoid_17"  [label="(1, 12, 64, 64)", style=solid];
+"2073 sigmoid_17" -> "2074 mul_34"  [label="(1, 12, 64, 64)", style=solid];
+"2074 mul_34" -> "2102 add_59"  [label="(1, 12, 64, 64)", style=solid];
+"2075 pad_19" -> "2076 roll_16"  [label="(1, 16, 16, 384)", style=solid];
+"2076 roll_16" -> "2077 view_94"  [label="(1, 16, 16, 384)", style=solid];
+"2077 view_94" -> "2078 permute_78"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2078 permute_78" -> "2079 reshape_76"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2079 reshape_76" -> "2083 linear_106"  [label="(4, 64, 384)", style=solid];
+"2079 reshape_76" -> "2103 new_zeros_8"  [label="(4, 64, 384)", style=solid];
+"2080 clone_17" -> "2081 slice_250"  [label="(1152,)", style=solid];
+"2080 clone_17" -> "2083 linear_106"  [label="(1152,)", style=solid];
+"2081 slice_250" -> "2082 zero__17"  [label="(384,)", style=solid];
+"2083 linear_106" -> "2084 reshape_77"  [label="(4, 64, 1152)", style=solid];
+"2084 reshape_77" -> "2085 permute_79"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2085 permute_79" -> "2086 select_51"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2085 permute_79" -> "2087 select_52"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2085 permute_79" -> "2088 select_53"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2086 select_51" -> "2089 linalg_vector_norm_34"  [label="(4, 12, 64, 32)", style=solid];
+"2086 select_51" -> "2091 expand_as_34"  [label="(4, 12, 64, 32)", style=solid];
+"2086 select_51" -> "2092 div_34"  [label="(4, 12, 64, 32)", style=solid];
+"2087 select_52" -> "2093 linalg_vector_norm_35"  [label="(4, 12, 64, 32)", style=solid];
+"2087 select_52" -> "2095 expand_as_35"  [label="(4, 12, 64, 32)", style=solid];
+"2087 select_52" -> "2096 div_35"  [label="(4, 12, 64, 32)", style=solid];
+"2088 select_53" -> "2157 matmul_35"  [label="(4, 12, 64, 32)", style=solid];
+"2089 linalg_vector_norm_34" -> "2090 clamp_min_34"  [label="(4, 12, 64, 1)", style=solid];
+"2090 clamp_min_34" -> "2091 expand_as_34"  [label="(4, 12, 64, 1)", style=solid];
+"2091 expand_as_34" -> "2092 div_34"  [label="(4, 12, 64, 32)", style=solid];
+"2092 div_34" -> "2098 matmul_34"  [label="(4, 12, 64, 32)", style=solid];
+"2093 linalg_vector_norm_35" -> "2094 clamp_min_35"  [label="(4, 12, 64, 1)", style=solid];
+"2094 clamp_min_35" -> "2095 expand_as_35"  [label="(4, 12, 64, 1)", style=solid];
+"2095 expand_as_35" -> "2096 div_35"  [label="(4, 12, 64, 32)", style=solid];
+"2096 div_35" -> "2097 transpose_34"  [label="(4, 12, 64, 32)", style=solid];
+"2097 transpose_34" -> "2098 matmul_34"  [label="(4, 12, 32, 64)", style=solid];
+"2098 matmul_34" -> "2101 mul_35"  [label="(4, 12, 64, 64)", style=solid];
+"2099 clamp_17" -> "2100 exp_17"  [label="(12, 1, 1)", style=solid];
+"2100 exp_17" -> "2101 mul_35"  [label="(12, 1, 1)", style=solid];
+"2101 mul_35" -> "2102 add_59"  [label="(4, 12, 64, 64)", style=solid];
+"2102 add_59" -> "2150 view_96"  [label="(4, 12, 64, 64)", style=solid];
+"2103 new_zeros_8" -> "2105 slice_251"  [label="(16, 16)", style=solid];
+"2103 new_zeros_8" -> "2109 slice_253"  [label="(16, 16)", style=solid];
+"2103 new_zeros_8" -> "2113 slice_255"  [label="(16, 16)", style=solid];
+"2103 new_zeros_8" -> "2117 slice_257"  [label="(16, 16)", style=solid];
+"2103 new_zeros_8" -> "2121 slice_259"  [label="(16, 16)", style=solid];
+"2103 new_zeros_8" -> "2125 slice_261"  [label="(16, 16)", style=solid];
+"2103 new_zeros_8" -> "2129 slice_263"  [label="(16, 16)", style=solid];
+"2103 new_zeros_8" -> "2133 slice_265"  [label="(16, 16)", style=solid];
+"2103 new_zeros_8" -> "2137 slice_267"  [label="(16, 16)", style=solid];
+"2103 new_zeros_8" -> "2140 view_95"  [label="(16, 16)", style=solid];
+"2104 lift_fresh_copy_72" -> "2107 fill__72"  [label="()", style=solid];
+"2105 slice_251" -> "2106 slice_252"  [label="(8, 16)", style=solid];
+"2106 slice_252" -> "2107 fill__72"  [label="(8, 8)", style=solid];
+"2108 lift_fresh_copy_73" -> "2111 fill__73"  [label="()", style=solid];
+"2109 slice_253" -> "2110 slice_254"  [label="(8, 16)", style=solid];
+"2110 slice_254" -> "2111 fill__73"  [label="(8, 4)", style=solid];
+"2112 lift_fresh_copy_74" -> "2115 fill__74"  [label="()", style=solid];
+"2113 slice_255" -> "2114 slice_256"  [label="(8, 16)", style=solid];
+"2114 slice_256" -> "2115 fill__74"  [label="(8, 4)", style=solid];
+"2116 lift_fresh_copy_75" -> "2119 fill__75"  [label="()", style=solid];
+"2117 slice_257" -> "2118 slice_258"  [label="(4, 16)", style=solid];
+"2118 slice_258" -> "2119 fill__75"  [label="(4, 8)", style=solid];
+"2120 lift_fresh_copy_76" -> "2123 fill__76"  [label="()", style=solid];
+"2121 slice_259" -> "2122 slice_260"  [label="(4, 16)", style=solid];
+"2122 slice_260" -> "2123 fill__76"  [label="(4, 4)", style=solid];
+"2124 lift_fresh_copy_77" -> "2127 fill__77"  [label="()", style=solid];
+"2125 slice_261" -> "2126 slice_262"  [label="(4, 16)", style=solid];
+"2126 slice_262" -> "2127 fill__77"  [label="(4, 4)", style=solid];
+"2128 lift_fresh_copy_78" -> "2131 fill__78"  [label="()", style=solid];
+"2129 slice_263" -> "2130 slice_264"  [label="(4, 16)", style=solid];
+"2130 slice_264" -> "2131 fill__78"  [label="(4, 8)", style=solid];
+"2132 lift_fresh_copy_79" -> "2135 fill__79"  [label="()", style=solid];
+"2133 slice_265" -> "2134 slice_266"  [label="(4, 16)", style=solid];
+"2134 slice_266" -> "2135 fill__79"  [label="(4, 4)", style=solid];
+"2136 lift_fresh_copy_80" -> "2139 fill__80"  [label="()", style=solid];
+"2137 slice_267" -> "2138 slice_268"  [label="(4, 16)", style=solid];
+"2138 slice_268" -> "2139 fill__80"  [label="(4, 4)", style=solid];
+"2140 view_95" -> "2141 permute_80"  [label="(2, 8, 2, 8)", style=solid];
+"2141 permute_80" -> "2142 reshape_78"  [label="(2, 2, 8, 8)", style=solid];
+"2142 reshape_78" -> "2143 unsqueeze_50"  [label="(4, 64)", style=solid];
+"2142 reshape_78" -> "2144 unsqueeze_51"  [label="(4, 64)", style=solid];
+"2143 unsqueeze_50" -> "2145 sub_8"  [label="(4, 1, 64)", style=solid];
+"2144 unsqueeze_51" -> "2145 sub_8"  [label="(4, 64, 1)", style=solid];
+"2145 sub_8" -> "2146 ne_8"  [label="(4, 64, 64)", style=solid];
+"2145 sub_8" -> "2147 masked_fill_16"  [label="(4, 64, 64)", style=solid];
+"2145 sub_8" -> "2148 eq_8"  [label="(4, 64, 64)", style=solid];
+"2146 ne_8" -> "2147 masked_fill_16"  [label="(4, 64, 64)", style=solid];
+"2147 masked_fill_16" -> "2149 masked_fill_17"  [label="(4, 64, 64)", style=solid];
+"2148 eq_8" -> "2149 masked_fill_17"  [label="(4, 64, 64)", style=solid];
+"2149 masked_fill_17" -> "2151 unsqueeze_52"  [label="(4, 64, 64)", style=solid];
+"2150 view_96" -> "2153 add_60"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2151 unsqueeze_52" -> "2152 unsqueeze_53"  [label="(4, 1, 64, 64)", style=solid];
+"2152 unsqueeze_53" -> "2153 add_60"  [label="(1, 4, 1, 64, 64)", style=solid];
+"2153 add_60" -> "2154 view_97"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2154 view_97" -> "2155 softmax_17"  [label="(4, 12, 64, 64)", style=solid];
+"2155 softmax_17" -> "2156 dropout_68"  [label="(4, 12, 64, 64)", style=solid];
+"2156 dropout_68" -> "2157 matmul_35"  [label="(4, 12, 64, 64)", style=solid];
+"2157 matmul_35" -> "2158 transpose_35"  [label="(4, 12, 64, 32)", style=solid];
+"2158 transpose_35" -> "2159 reshape_79"  [label="(4, 64, 12, 32)", style=solid];
+"2159 reshape_79" -> "2160 linear_107"  [label="(4, 64, 384)", style=solid];
+"2160 linear_107" -> "2161 dropout_69"  [label="(4, 64, 384)", style=solid];
+"2161 dropout_69" -> "2162 view_98"  [label="(4, 64, 384)", style=solid];
+"2162 view_98" -> "2163 permute_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2163 permute_81" -> "2164 reshape_80"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2164 reshape_80" -> "2165 roll_17"  [label="(1, 16, 16, 384)", style=solid];
+"2165 roll_17" -> "2166 slice_269"  [label="(1, 16, 16, 384)", style=solid];
+"2166 slice_269" -> "2167 slice_270"  [label="(1, 16, 16, 384)", style=solid];
+"2167 slice_270" -> "2168 slice_271"  [label="(1, 14, 16, 384)", style=solid];
+"2168 slice_271" -> "2169 slice_272"  [label="(1, 14, 14, 384)", style=solid];
+"2169 slice_272" -> "2170 contiguous_33"  [label="(1, 14, 14, 384)", style=solid];
+"2170 contiguous_33" -> "2171 layer_norm_37"  [label="(1, 14, 14, 384)", style=solid];
+"2171 layer_norm_37" -> "2172 add_61"  [label="(1, 14, 14, 384)", style=solid];
+"2172 add_61" -> "2173 linear_108"  [label="(1, 14, 14, 384)", style=solid];
+"2172 add_61" -> "2179 add_62"  [label="(1, 14, 14, 384)", style=solid];
+"2173 linear_108" -> "2174 gelu_17"  [label="(1, 14, 14, 1536)", style=solid];
+"2174 gelu_17" -> "2175 dropout_70"  [label="(1, 14, 14, 1536)", style=solid];
+"2175 dropout_70" -> "2176 linear_109"  [label="(1, 14, 14, 1536)", style=solid];
+"2176 linear_109" -> "2177 dropout_71"  [label="(1, 14, 14, 384)", style=solid];
+"2177 dropout_71" -> "2178 layer_norm_38"  [label="(1, 14, 14, 384)", style=solid];
+"2178 layer_norm_38" -> "2179 add_62"  [label="(1, 14, 14, 384)", style=solid];
+"2179 add_62" -> "2191 pad_20"  [label="(1, 14, 14, 384)", style=solid];
+"2179 add_62" -> "2234 add_64"  [label="(1, 14, 14, 384)", style=solid];
+"2180 linear_110" -> "2181 relu__18"  [label="(1, 15, 15, 512)", style=solid];
+"2181 relu__18" -> "2182 linear_111"  [label="(1, 15, 15, 512)", style=solid];
+"2182 linear_111" -> "2183 view_99"  [label="(1, 15, 15, 12)", style=solid];
+"2183 view_99" -> "2184 index_18"  [label="(225, 12)", style=solid];
+"2184 index_18" -> "2185 view_100"  [label="(4096, 12)", style=solid];
+"2185 view_100" -> "2186 permute_82"  [label="(64, 64, 12)", style=solid];
+"2186 permute_82" -> "2187 contiguous_34"  [label="(12, 64, 64)", style=solid];
+"2187 contiguous_34" -> "2188 unsqueeze_54"  [label="(12, 64, 64)", style=solid];
+"2188 unsqueeze_54" -> "2189 sigmoid_18"  [label="(1, 12, 64, 64)", style=solid];
+"2189 sigmoid_18" -> "2190 mul_36"  [label="(1, 12, 64, 64)", style=solid];
+"2190 mul_36" -> "2217 add_63"  [label="(1, 12, 64, 64)", style=solid];
+"2191 pad_20" -> "2192 view_101"  [label="(1, 16, 16, 384)", style=solid];
+"2192 view_101" -> "2193 permute_83"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2193 permute_83" -> "2194 reshape_81"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2194 reshape_81" -> "2198 linear_112"  [label="(4, 64, 384)", style=solid];
+"2195 clone_18" -> "2196 slice_273"  [label="(1152,)", style=solid];
+"2195 clone_18" -> "2198 linear_112"  [label="(1152,)", style=solid];
+"2196 slice_273" -> "2197 zero__18"  [label="(384,)", style=solid];
+"2198 linear_112" -> "2199 reshape_82"  [label="(4, 64, 1152)", style=solid];
+"2199 reshape_82" -> "2200 permute_84"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2200 permute_84" -> "2201 select_54"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2200 permute_84" -> "2202 select_55"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2200 permute_84" -> "2203 select_56"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2201 select_54" -> "2204 linalg_vector_norm_36"  [label="(4, 12, 64, 32)", style=solid];
+"2201 select_54" -> "2206 expand_as_36"  [label="(4, 12, 64, 32)", style=solid];
+"2201 select_54" -> "2207 div_36"  [label="(4, 12, 64, 32)", style=solid];
+"2202 select_55" -> "2208 linalg_vector_norm_37"  [label="(4, 12, 64, 32)", style=solid];
+"2202 select_55" -> "2210 expand_as_37"  [label="(4, 12, 64, 32)", style=solid];
+"2202 select_55" -> "2211 div_37"  [label="(4, 12, 64, 32)", style=solid];
+"2203 select_56" -> "2220 matmul_37"  [label="(4, 12, 64, 32)", style=solid];
+"2204 linalg_vector_norm_36" -> "2205 clamp_min_36"  [label="(4, 12, 64, 1)", style=solid];
+"2205 clamp_min_36" -> "2206 expand_as_36"  [label="(4, 12, 64, 1)", style=solid];
+"2206 expand_as_36" -> "2207 div_36"  [label="(4, 12, 64, 32)", style=solid];
+"2207 div_36" -> "2213 matmul_36"  [label="(4, 12, 64, 32)", style=solid];
+"2208 linalg_vector_norm_37" -> "2209 clamp_min_37"  [label="(4, 12, 64, 1)", style=solid];
+"2209 clamp_min_37" -> "2210 expand_as_37"  [label="(4, 12, 64, 1)", style=solid];
+"2210 expand_as_37" -> "2211 div_37"  [label="(4, 12, 64, 32)", style=solid];
+"2211 div_37" -> "2212 transpose_36"  [label="(4, 12, 64, 32)", style=solid];
+"2212 transpose_36" -> "2213 matmul_36"  [label="(4, 12, 32, 64)", style=solid];
+"2213 matmul_36" -> "2216 mul_37"  [label="(4, 12, 64, 64)", style=solid];
+"2214 clamp_18" -> "2215 exp_18"  [label="(12, 1, 1)", style=solid];
+"2215 exp_18" -> "2216 mul_37"  [label="(12, 1, 1)", style=solid];
+"2216 mul_37" -> "2217 add_63"  [label="(4, 12, 64, 64)", style=solid];
+"2217 add_63" -> "2218 softmax_18"  [label="(4, 12, 64, 64)", style=solid];
+"2218 softmax_18" -> "2219 dropout_72"  [label="(4, 12, 64, 64)", style=solid];
+"2219 dropout_72" -> "2220 matmul_37"  [label="(4, 12, 64, 64)", style=solid];
+"2220 matmul_37" -> "2221 transpose_37"  [label="(4, 12, 64, 32)", style=solid];
+"2221 transpose_37" -> "2222 reshape_83"  [label="(4, 64, 12, 32)", style=solid];
+"2222 reshape_83" -> "2223 linear_113"  [label="(4, 64, 384)", style=solid];
+"2223 linear_113" -> "2224 dropout_73"  [label="(4, 64, 384)", style=solid];
+"2224 dropout_73" -> "2225 view_102"  [label="(4, 64, 384)", style=solid];
+"2225 view_102" -> "2226 permute_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2226 permute_85" -> "2227 reshape_84"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2227 reshape_84" -> "2228 slice_274"  [label="(1, 16, 16, 384)", style=solid];
+"2228 slice_274" -> "2229 slice_275"  [label="(1, 16, 16, 384)", style=solid];
+"2229 slice_275" -> "2230 slice_276"  [label="(1, 14, 16, 384)", style=solid];
+"2230 slice_276" -> "2231 slice_277"  [label="(1, 14, 14, 384)", style=solid];
+"2231 slice_277" -> "2232 contiguous_35"  [label="(1, 14, 14, 384)", style=solid];
+"2232 contiguous_35" -> "2233 layer_norm_39"  [label="(1, 14, 14, 384)", style=solid];
+"2233 layer_norm_39" -> "2234 add_64"  [label="(1, 14, 14, 384)", style=solid];
+"2234 add_64" -> "2235 linear_114"  [label="(1, 14, 14, 384)", style=solid];
+"2234 add_64" -> "2241 add_65"  [label="(1, 14, 14, 384)", style=solid];
+"2235 linear_114" -> "2236 gelu_18"  [label="(1, 14, 14, 1536)", style=solid];
+"2236 gelu_18" -> "2237 dropout_74"  [label="(1, 14, 14, 1536)", style=solid];
+"2237 dropout_74" -> "2238 linear_115"  [label="(1, 14, 14, 1536)", style=solid];
+"2238 linear_115" -> "2239 dropout_75"  [label="(1, 14, 14, 384)", style=solid];
+"2239 dropout_75" -> "2240 layer_norm_40"  [label="(1, 14, 14, 384)", style=solid];
+"2240 layer_norm_40" -> "2241 add_65"  [label="(1, 14, 14, 384)", style=solid];
+"2241 add_65" -> "2253 pad_21"  [label="(1, 14, 14, 384)", style=solid];
+"2241 add_65" -> "2350 add_68"  [label="(1, 14, 14, 384)", style=solid];
+"2242 linear_116" -> "2243 relu__19"  [label="(1, 15, 15, 512)", style=solid];
+"2243 relu__19" -> "2244 linear_117"  [label="(1, 15, 15, 512)", style=solid];
+"2244 linear_117" -> "2245 view_103"  [label="(1, 15, 15, 12)", style=solid];
+"2245 view_103" -> "2246 index_19"  [label="(225, 12)", style=solid];
+"2246 index_19" -> "2247 view_104"  [label="(4096, 12)", style=solid];
+"2247 view_104" -> "2248 permute_86"  [label="(64, 64, 12)", style=solid];
+"2248 permute_86" -> "2249 contiguous_36"  [label="(12, 64, 64)", style=solid];
+"2249 contiguous_36" -> "2250 unsqueeze_55"  [label="(12, 64, 64)", style=solid];
+"2250 unsqueeze_55" -> "2251 sigmoid_19"  [label="(1, 12, 64, 64)", style=solid];
+"2251 sigmoid_19" -> "2252 mul_38"  [label="(1, 12, 64, 64)", style=solid];
+"2252 mul_38" -> "2280 add_66"  [label="(1, 12, 64, 64)", style=solid];
+"2253 pad_21" -> "2254 roll_18"  [label="(1, 16, 16, 384)", style=solid];
+"2254 roll_18" -> "2255 view_105"  [label="(1, 16, 16, 384)", style=solid];
+"2255 view_105" -> "2256 permute_87"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2256 permute_87" -> "2257 reshape_85"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2257 reshape_85" -> "2261 linear_118"  [label="(4, 64, 384)", style=solid];
+"2257 reshape_85" -> "2281 new_zeros_9"  [label="(4, 64, 384)", style=solid];
+"2258 clone_19" -> "2259 slice_278"  [label="(1152,)", style=solid];
+"2258 clone_19" -> "2261 linear_118"  [label="(1152,)", style=solid];
+"2259 slice_278" -> "2260 zero__19"  [label="(384,)", style=solid];
+"2261 linear_118" -> "2262 reshape_86"  [label="(4, 64, 1152)", style=solid];
+"2262 reshape_86" -> "2263 permute_88"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2263 permute_88" -> "2264 select_57"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2263 permute_88" -> "2265 select_58"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2263 permute_88" -> "2266 select_59"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2264 select_57" -> "2267 linalg_vector_norm_38"  [label="(4, 12, 64, 32)", style=solid];
+"2264 select_57" -> "2269 expand_as_38"  [label="(4, 12, 64, 32)", style=solid];
+"2264 select_57" -> "2270 div_38"  [label="(4, 12, 64, 32)", style=solid];
+"2265 select_58" -> "2271 linalg_vector_norm_39"  [label="(4, 12, 64, 32)", style=solid];
+"2265 select_58" -> "2273 expand_as_39"  [label="(4, 12, 64, 32)", style=solid];
+"2265 select_58" -> "2274 div_39"  [label="(4, 12, 64, 32)", style=solid];
+"2266 select_59" -> "2335 matmul_39"  [label="(4, 12, 64, 32)", style=solid];
+"2267 linalg_vector_norm_38" -> "2268 clamp_min_38"  [label="(4, 12, 64, 1)", style=solid];
+"2268 clamp_min_38" -> "2269 expand_as_38"  [label="(4, 12, 64, 1)", style=solid];
+"2269 expand_as_38" -> "2270 div_38"  [label="(4, 12, 64, 32)", style=solid];
+"2270 div_38" -> "2276 matmul_38"  [label="(4, 12, 64, 32)", style=solid];
+"2271 linalg_vector_norm_39" -> "2272 clamp_min_39"  [label="(4, 12, 64, 1)", style=solid];
+"2272 clamp_min_39" -> "2273 expand_as_39"  [label="(4, 12, 64, 1)", style=solid];
+"2273 expand_as_39" -> "2274 div_39"  [label="(4, 12, 64, 32)", style=solid];
+"2274 div_39" -> "2275 transpose_38"  [label="(4, 12, 64, 32)", style=solid];
+"2275 transpose_38" -> "2276 matmul_38"  [label="(4, 12, 32, 64)", style=solid];
+"2276 matmul_38" -> "2279 mul_39"  [label="(4, 12, 64, 64)", style=solid];
+"2277 clamp_19" -> "2278 exp_19"  [label="(12, 1, 1)", style=solid];
+"2278 exp_19" -> "2279 mul_39"  [label="(12, 1, 1)", style=solid];
+"2279 mul_39" -> "2280 add_66"  [label="(4, 12, 64, 64)", style=solid];
+"2280 add_66" -> "2328 view_107"  [label="(4, 12, 64, 64)", style=solid];
+"2281 new_zeros_9" -> "2283 slice_279"  [label="(16, 16)", style=solid];
+"2281 new_zeros_9" -> "2287 slice_281"  [label="(16, 16)", style=solid];
+"2281 new_zeros_9" -> "2291 slice_283"  [label="(16, 16)", style=solid];
+"2281 new_zeros_9" -> "2295 slice_285"  [label="(16, 16)", style=solid];
+"2281 new_zeros_9" -> "2299 slice_287"  [label="(16, 16)", style=solid];
+"2281 new_zeros_9" -> "2303 slice_289"  [label="(16, 16)", style=solid];
+"2281 new_zeros_9" -> "2307 slice_291"  [label="(16, 16)", style=solid];
+"2281 new_zeros_9" -> "2311 slice_293"  [label="(16, 16)", style=solid];
+"2281 new_zeros_9" -> "2315 slice_295"  [label="(16, 16)", style=solid];
+"2281 new_zeros_9" -> "2318 view_106"  [label="(16, 16)", style=solid];
+"2282 lift_fresh_copy_81" -> "2285 fill__81"  [label="()", style=solid];
+"2283 slice_279" -> "2284 slice_280"  [label="(8, 16)", style=solid];
+"2284 slice_280" -> "2285 fill__81"  [label="(8, 8)", style=solid];
+"2286 lift_fresh_copy_82" -> "2289 fill__82"  [label="()", style=solid];
+"2287 slice_281" -> "2288 slice_282"  [label="(8, 16)", style=solid];
+"2288 slice_282" -> "2289 fill__82"  [label="(8, 4)", style=solid];
+"2290 lift_fresh_copy_83" -> "2293 fill__83"  [label="()", style=solid];
+"2291 slice_283" -> "2292 slice_284"  [label="(8, 16)", style=solid];
+"2292 slice_284" -> "2293 fill__83"  [label="(8, 4)", style=solid];
+"2294 lift_fresh_copy_84" -> "2297 fill__84"  [label="()", style=solid];
+"2295 slice_285" -> "2296 slice_286"  [label="(4, 16)", style=solid];
+"2296 slice_286" -> "2297 fill__84"  [label="(4, 8)", style=solid];
+"2298 lift_fresh_copy_85" -> "2301 fill__85"  [label="()", style=solid];
+"2299 slice_287" -> "2300 slice_288"  [label="(4, 16)", style=solid];
+"2300 slice_288" -> "2301 fill__85"  [label="(4, 4)", style=solid];
+"2302 lift_fresh_copy_86" -> "2305 fill__86"  [label="()", style=solid];
+"2303 slice_289" -> "2304 slice_290"  [label="(4, 16)", style=solid];
+"2304 slice_290" -> "2305 fill__86"  [label="(4, 4)", style=solid];
+"2306 lift_fresh_copy_87" -> "2309 fill__87"  [label="()", style=solid];
+"2307 slice_291" -> "2308 slice_292"  [label="(4, 16)", style=solid];
+"2308 slice_292" -> "2309 fill__87"  [label="(4, 8)", style=solid];
+"2310 lift_fresh_copy_88" -> "2313 fill__88"  [label="()", style=solid];
+"2311 slice_293" -> "2312 slice_294"  [label="(4, 16)", style=solid];
+"2312 slice_294" -> "2313 fill__88"  [label="(4, 4)", style=solid];
+"2314 lift_fresh_copy_89" -> "2317 fill__89"  [label="()", style=solid];
+"2315 slice_295" -> "2316 slice_296"  [label="(4, 16)", style=solid];
+"2316 slice_296" -> "2317 fill__89"  [label="(4, 4)", style=solid];
+"2318 view_106" -> "2319 permute_89"  [label="(2, 8, 2, 8)", style=solid];
+"2319 permute_89" -> "2320 reshape_87"  [label="(2, 2, 8, 8)", style=solid];
+"2320 reshape_87" -> "2321 unsqueeze_56"  [label="(4, 64)", style=solid];
+"2320 reshape_87" -> "2322 unsqueeze_57"  [label="(4, 64)", style=solid];
+"2321 unsqueeze_56" -> "2323 sub_9"  [label="(4, 1, 64)", style=solid];
+"2322 unsqueeze_57" -> "2323 sub_9"  [label="(4, 64, 1)", style=solid];
+"2323 sub_9" -> "2324 ne_9"  [label="(4, 64, 64)", style=solid];
+"2323 sub_9" -> "2325 masked_fill_18"  [label="(4, 64, 64)", style=solid];
+"2323 sub_9" -> "2326 eq_9"  [label="(4, 64, 64)", style=solid];
+"2324 ne_9" -> "2325 masked_fill_18"  [label="(4, 64, 64)", style=solid];
+"2325 masked_fill_18" -> "2327 masked_fill_19"  [label="(4, 64, 64)", style=solid];
+"2326 eq_9" -> "2327 masked_fill_19"  [label="(4, 64, 64)", style=solid];
+"2327 masked_fill_19" -> "2329 unsqueeze_58"  [label="(4, 64, 64)", style=solid];
+"2328 view_107" -> "2331 add_67"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2329 unsqueeze_58" -> "2330 unsqueeze_59"  [label="(4, 1, 64, 64)", style=solid];
+"2330 unsqueeze_59" -> "2331 add_67"  [label="(1, 4, 1, 64, 64)", style=solid];
+"2331 add_67" -> "2332 view_108"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2332 view_108" -> "2333 softmax_19"  [label="(4, 12, 64, 64)", style=solid];
+"2333 softmax_19" -> "2334 dropout_76"  [label="(4, 12, 64, 64)", style=solid];
+"2334 dropout_76" -> "2335 matmul_39"  [label="(4, 12, 64, 64)", style=solid];
+"2335 matmul_39" -> "2336 transpose_39"  [label="(4, 12, 64, 32)", style=solid];
+"2336 transpose_39" -> "2337 reshape_88"  [label="(4, 64, 12, 32)", style=solid];
+"2337 reshape_88" -> "2338 linear_119"  [label="(4, 64, 384)", style=solid];
+"2338 linear_119" -> "2339 dropout_77"  [label="(4, 64, 384)", style=solid];
+"2339 dropout_77" -> "2340 view_109"  [label="(4, 64, 384)", style=solid];
+"2340 view_109" -> "2341 permute_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2341 permute_90" -> "2342 reshape_89"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2342 reshape_89" -> "2343 roll_19"  [label="(1, 16, 16, 384)", style=solid];
+"2343 roll_19" -> "2344 slice_297"  [label="(1, 16, 16, 384)", style=solid];
+"2344 slice_297" -> "2345 slice_298"  [label="(1, 16, 16, 384)", style=solid];
+"2345 slice_298" -> "2346 slice_299"  [label="(1, 14, 16, 384)", style=solid];
+"2346 slice_299" -> "2347 slice_300"  [label="(1, 14, 14, 384)", style=solid];
+"2347 slice_300" -> "2348 contiguous_37"  [label="(1, 14, 14, 384)", style=solid];
+"2348 contiguous_37" -> "2349 layer_norm_41"  [label="(1, 14, 14, 384)", style=solid];
+"2349 layer_norm_41" -> "2350 add_68"  [label="(1, 14, 14, 384)", style=solid];
+"2350 add_68" -> "2351 linear_120"  [label="(1, 14, 14, 384)", style=solid];
+"2350 add_68" -> "2357 add_69"  [label="(1, 14, 14, 384)", style=solid];
+"2351 linear_120" -> "2352 gelu_19"  [label="(1, 14, 14, 1536)", style=solid];
+"2352 gelu_19" -> "2353 dropout_78"  [label="(1, 14, 14, 1536)", style=solid];
+"2353 dropout_78" -> "2354 linear_121"  [label="(1, 14, 14, 1536)", style=solid];
+"2354 linear_121" -> "2355 dropout_79"  [label="(1, 14, 14, 384)", style=solid];
+"2355 dropout_79" -> "2356 layer_norm_42"  [label="(1, 14, 14, 384)", style=solid];
+"2356 layer_norm_42" -> "2357 add_69"  [label="(1, 14, 14, 384)", style=solid];
+"2357 add_69" -> "2369 pad_22"  [label="(1, 14, 14, 384)", style=solid];
+"2357 add_69" -> "2412 add_71"  [label="(1, 14, 14, 384)", style=solid];
+"2358 linear_122" -> "2359 relu__20"  [label="(1, 15, 15, 512)", style=solid];
+"2359 relu__20" -> "2360 linear_123"  [label="(1, 15, 15, 512)", style=solid];
+"2360 linear_123" -> "2361 view_110"  [label="(1, 15, 15, 12)", style=solid];
+"2361 view_110" -> "2362 index_20"  [label="(225, 12)", style=solid];
+"2362 index_20" -> "2363 view_111"  [label="(4096, 12)", style=solid];
+"2363 view_111" -> "2364 permute_91"  [label="(64, 64, 12)", style=solid];
+"2364 permute_91" -> "2365 contiguous_38"  [label="(12, 64, 64)", style=solid];
+"2365 contiguous_38" -> "2366 unsqueeze_60"  [label="(12, 64, 64)", style=solid];
+"2366 unsqueeze_60" -> "2367 sigmoid_20"  [label="(1, 12, 64, 64)", style=solid];
+"2367 sigmoid_20" -> "2368 mul_40"  [label="(1, 12, 64, 64)", style=solid];
+"2368 mul_40" -> "2395 add_70"  [label="(1, 12, 64, 64)", style=solid];
+"2369 pad_22" -> "2370 view_112"  [label="(1, 16, 16, 384)", style=solid];
+"2370 view_112" -> "2371 permute_92"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2371 permute_92" -> "2372 reshape_90"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2372 reshape_90" -> "2376 linear_124"  [label="(4, 64, 384)", style=solid];
+"2373 clone_20" -> "2374 slice_301"  [label="(1152,)", style=solid];
+"2373 clone_20" -> "2376 linear_124"  [label="(1152,)", style=solid];
+"2374 slice_301" -> "2375 zero__20"  [label="(384,)", style=solid];
+"2376 linear_124" -> "2377 reshape_91"  [label="(4, 64, 1152)", style=solid];
+"2377 reshape_91" -> "2378 permute_93"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2378 permute_93" -> "2379 select_60"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2378 permute_93" -> "2380 select_61"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2378 permute_93" -> "2381 select_62"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2379 select_60" -> "2382 linalg_vector_norm_40"  [label="(4, 12, 64, 32)", style=solid];
+"2379 select_60" -> "2384 expand_as_40"  [label="(4, 12, 64, 32)", style=solid];
+"2379 select_60" -> "2385 div_40"  [label="(4, 12, 64, 32)", style=solid];
+"2380 select_61" -> "2386 linalg_vector_norm_41"  [label="(4, 12, 64, 32)", style=solid];
+"2380 select_61" -> "2388 expand_as_41"  [label="(4, 12, 64, 32)", style=solid];
+"2380 select_61" -> "2389 div_41"  [label="(4, 12, 64, 32)", style=solid];
+"2381 select_62" -> "2398 matmul_41"  [label="(4, 12, 64, 32)", style=solid];
+"2382 linalg_vector_norm_40" -> "2383 clamp_min_40"  [label="(4, 12, 64, 1)", style=solid];
+"2383 clamp_min_40" -> "2384 expand_as_40"  [label="(4, 12, 64, 1)", style=solid];
+"2384 expand_as_40" -> "2385 div_40"  [label="(4, 12, 64, 32)", style=solid];
+"2385 div_40" -> "2391 matmul_40"  [label="(4, 12, 64, 32)", style=solid];
+"2386 linalg_vector_norm_41" -> "2387 clamp_min_41"  [label="(4, 12, 64, 1)", style=solid];
+"2387 clamp_min_41" -> "2388 expand_as_41"  [label="(4, 12, 64, 1)", style=solid];
+"2388 expand_as_41" -> "2389 div_41"  [label="(4, 12, 64, 32)", style=solid];
+"2389 div_41" -> "2390 transpose_40"  [label="(4, 12, 64, 32)", style=solid];
+"2390 transpose_40" -> "2391 matmul_40"  [label="(4, 12, 32, 64)", style=solid];
+"2391 matmul_40" -> "2394 mul_41"  [label="(4, 12, 64, 64)", style=solid];
+"2392 clamp_20" -> "2393 exp_20"  [label="(12, 1, 1)", style=solid];
+"2393 exp_20" -> "2394 mul_41"  [label="(12, 1, 1)", style=solid];
+"2394 mul_41" -> "2395 add_70"  [label="(4, 12, 64, 64)", style=solid];
+"2395 add_70" -> "2396 softmax_20"  [label="(4, 12, 64, 64)", style=solid];
+"2396 softmax_20" -> "2397 dropout_80"  [label="(4, 12, 64, 64)", style=solid];
+"2397 dropout_80" -> "2398 matmul_41"  [label="(4, 12, 64, 64)", style=solid];
+"2398 matmul_41" -> "2399 transpose_41"  [label="(4, 12, 64, 32)", style=solid];
+"2399 transpose_41" -> "2400 reshape_92"  [label="(4, 64, 12, 32)", style=solid];
+"2400 reshape_92" -> "2401 linear_125"  [label="(4, 64, 384)", style=solid];
+"2401 linear_125" -> "2402 dropout_81"  [label="(4, 64, 384)", style=solid];
+"2402 dropout_81" -> "2403 view_113"  [label="(4, 64, 384)", style=solid];
+"2403 view_113" -> "2404 permute_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2404 permute_94" -> "2405 reshape_93"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2405 reshape_93" -> "2406 slice_302"  [label="(1, 16, 16, 384)", style=solid];
+"2406 slice_302" -> "2407 slice_303"  [label="(1, 16, 16, 384)", style=solid];
+"2407 slice_303" -> "2408 slice_304"  [label="(1, 14, 16, 384)", style=solid];
+"2408 slice_304" -> "2409 slice_305"  [label="(1, 14, 14, 384)", style=solid];
+"2409 slice_305" -> "2410 contiguous_39"  [label="(1, 14, 14, 384)", style=solid];
+"2410 contiguous_39" -> "2411 layer_norm_43"  [label="(1, 14, 14, 384)", style=solid];
+"2411 layer_norm_43" -> "2412 add_71"  [label="(1, 14, 14, 384)", style=solid];
+"2412 add_71" -> "2413 linear_126"  [label="(1, 14, 14, 384)", style=solid];
+"2412 add_71" -> "2419 add_72"  [label="(1, 14, 14, 384)", style=solid];
+"2413 linear_126" -> "2414 gelu_20"  [label="(1, 14, 14, 1536)", style=solid];
+"2414 gelu_20" -> "2415 dropout_82"  [label="(1, 14, 14, 1536)", style=solid];
+"2415 dropout_82" -> "2416 linear_127"  [label="(1, 14, 14, 1536)", style=solid];
+"2416 linear_127" -> "2417 dropout_83"  [label="(1, 14, 14, 384)", style=solid];
+"2417 dropout_83" -> "2418 layer_norm_44"  [label="(1, 14, 14, 384)", style=solid];
+"2418 layer_norm_44" -> "2419 add_72"  [label="(1, 14, 14, 384)", style=solid];
+"2419 add_72" -> "2431 pad_23"  [label="(1, 14, 14, 384)", style=solid];
+"2419 add_72" -> "2528 add_75"  [label="(1, 14, 14, 384)", style=solid];
+"2420 linear_128" -> "2421 relu__21"  [label="(1, 15, 15, 512)", style=solid];
+"2421 relu__21" -> "2422 linear_129"  [label="(1, 15, 15, 512)", style=solid];
+"2422 linear_129" -> "2423 view_114"  [label="(1, 15, 15, 12)", style=solid];
+"2423 view_114" -> "2424 index_21"  [label="(225, 12)", style=solid];
+"2424 index_21" -> "2425 view_115"  [label="(4096, 12)", style=solid];
+"2425 view_115" -> "2426 permute_95"  [label="(64, 64, 12)", style=solid];
+"2426 permute_95" -> "2427 contiguous_40"  [label="(12, 64, 64)", style=solid];
+"2427 contiguous_40" -> "2428 unsqueeze_61"  [label="(12, 64, 64)", style=solid];
+"2428 unsqueeze_61" -> "2429 sigmoid_21"  [label="(1, 12, 64, 64)", style=solid];
+"2429 sigmoid_21" -> "2430 mul_42"  [label="(1, 12, 64, 64)", style=solid];
+"2430 mul_42" -> "2458 add_73"  [label="(1, 12, 64, 64)", style=solid];
+"2431 pad_23" -> "2432 roll_20"  [label="(1, 16, 16, 384)", style=solid];
+"2432 roll_20" -> "2433 view_116"  [label="(1, 16, 16, 384)", style=solid];
+"2433 view_116" -> "2434 permute_96"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2434 permute_96" -> "2435 reshape_94"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2435 reshape_94" -> "2439 linear_130"  [label="(4, 64, 384)", style=solid];
+"2435 reshape_94" -> "2459 new_zeros_10"  [label="(4, 64, 384)", style=solid];
+"2436 clone_21" -> "2437 slice_306"  [label="(1152,)", style=solid];
+"2436 clone_21" -> "2439 linear_130"  [label="(1152,)", style=solid];
+"2437 slice_306" -> "2438 zero__21"  [label="(384,)", style=solid];
+"2439 linear_130" -> "2440 reshape_95"  [label="(4, 64, 1152)", style=solid];
+"2440 reshape_95" -> "2441 permute_97"  [label="(4, 64, 3, 12, 32)", style=solid];
+"2441 permute_97" -> "2442 select_63"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2441 permute_97" -> "2443 select_64"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2441 permute_97" -> "2444 select_65"  [label="(3, 4, 12, 64, 32)", style=solid];
+"2442 select_63" -> "2445 linalg_vector_norm_42"  [label="(4, 12, 64, 32)", style=solid];
+"2442 select_63" -> "2447 expand_as_42"  [label="(4, 12, 64, 32)", style=solid];
+"2442 select_63" -> "2448 div_42"  [label="(4, 12, 64, 32)", style=solid];
+"2443 select_64" -> "2449 linalg_vector_norm_43"  [label="(4, 12, 64, 32)", style=solid];
+"2443 select_64" -> "2451 expand_as_43"  [label="(4, 12, 64, 32)", style=solid];
+"2443 select_64" -> "2452 div_43"  [label="(4, 12, 64, 32)", style=solid];
+"2444 select_65" -> "2513 matmul_43"  [label="(4, 12, 64, 32)", style=solid];
+"2445 linalg_vector_norm_42" -> "2446 clamp_min_42"  [label="(4, 12, 64, 1)", style=solid];
+"2446 clamp_min_42" -> "2447 expand_as_42"  [label="(4, 12, 64, 1)", style=solid];
+"2447 expand_as_42" -> "2448 div_42"  [label="(4, 12, 64, 32)", style=solid];
+"2448 div_42" -> "2454 matmul_42"  [label="(4, 12, 64, 32)", style=solid];
+"2449 linalg_vector_norm_43" -> "2450 clamp_min_43"  [label="(4, 12, 64, 1)", style=solid];
+"2450 clamp_min_43" -> "2451 expand_as_43"  [label="(4, 12, 64, 1)", style=solid];
+"2451 expand_as_43" -> "2452 div_43"  [label="(4, 12, 64, 32)", style=solid];
+"2452 div_43" -> "2453 transpose_42"  [label="(4, 12, 64, 32)", style=solid];
+"2453 transpose_42" -> "2454 matmul_42"  [label="(4, 12, 32, 64)", style=solid];
+"2454 matmul_42" -> "2457 mul_43"  [label="(4, 12, 64, 64)", style=solid];
+"2455 clamp_21" -> "2456 exp_21"  [label="(12, 1, 1)", style=solid];
+"2456 exp_21" -> "2457 mul_43"  [label="(12, 1, 1)", style=solid];
+"2457 mul_43" -> "2458 add_73"  [label="(4, 12, 64, 64)", style=solid];
+"2458 add_73" -> "2506 view_118"  [label="(4, 12, 64, 64)", style=solid];
+"2459 new_zeros_10" -> "2461 slice_307"  [label="(16, 16)", style=solid];
+"2459 new_zeros_10" -> "2465 slice_309"  [label="(16, 16)", style=solid];
+"2459 new_zeros_10" -> "2469 slice_311"  [label="(16, 16)", style=solid];
+"2459 new_zeros_10" -> "2473 slice_313"  [label="(16, 16)", style=solid];
+"2459 new_zeros_10" -> "2477 slice_315"  [label="(16, 16)", style=solid];
+"2459 new_zeros_10" -> "2481 slice_317"  [label="(16, 16)", style=solid];
+"2459 new_zeros_10" -> "2485 slice_319"  [label="(16, 16)", style=solid];
+"2459 new_zeros_10" -> "2489 slice_321"  [label="(16, 16)", style=solid];
+"2459 new_zeros_10" -> "2493 slice_323"  [label="(16, 16)", style=solid];
+"2459 new_zeros_10" -> "2496 view_117"  [label="(16, 16)", style=solid];
+"2460 lift_fresh_copy_90" -> "2463 fill__90"  [label="()", style=solid];
+"2461 slice_307" -> "2462 slice_308"  [label="(8, 16)", style=solid];
+"2462 slice_308" -> "2463 fill__90"  [label="(8, 8)", style=solid];
+"2464 lift_fresh_copy_91" -> "2467 fill__91"  [label="()", style=solid];
+"2465 slice_309" -> "2466 slice_310"  [label="(8, 16)", style=solid];
+"2466 slice_310" -> "2467 fill__91"  [label="(8, 4)", style=solid];
+"2468 lift_fresh_copy_92" -> "2471 fill__92"  [label="()", style=solid];
+"2469 slice_311" -> "2470 slice_312"  [label="(8, 16)", style=solid];
+"2470 slice_312" -> "2471 fill__92"  [label="(8, 4)", style=solid];
+"2472 lift_fresh_copy_93" -> "2475 fill__93"  [label="()", style=solid];
+"2473 slice_313" -> "2474 slice_314"  [label="(4, 16)", style=solid];
+"2474 slice_314" -> "2475 fill__93"  [label="(4, 8)", style=solid];
+"2476 lift_fresh_copy_94" -> "2479 fill__94"  [label="()", style=solid];
+"2477 slice_315" -> "2478 slice_316"  [label="(4, 16)", style=solid];
+"2478 slice_316" -> "2479 fill__94"  [label="(4, 4)", style=solid];
+"2480 lift_fresh_copy_95" -> "2483 fill__95"  [label="()", style=solid];
+"2481 slice_317" -> "2482 slice_318"  [label="(4, 16)", style=solid];
+"2482 slice_318" -> "2483 fill__95"  [label="(4, 4)", style=solid];
+"2484 lift_fresh_copy_96" -> "2487 fill__96"  [label="()", style=solid];
+"2485 slice_319" -> "2486 slice_320"  [label="(4, 16)", style=solid];
+"2486 slice_320" -> "2487 fill__96"  [label="(4, 8)", style=solid];
+"2488 lift_fresh_copy_97" -> "2491 fill__97"  [label="()", style=solid];
+"2489 slice_321" -> "2490 slice_322"  [label="(4, 16)", style=solid];
+"2490 slice_322" -> "2491 fill__97"  [label="(4, 4)", style=solid];
+"2492 lift_fresh_copy_98" -> "2495 fill__98"  [label="()", style=solid];
+"2493 slice_323" -> "2494 slice_324"  [label="(4, 16)", style=solid];
+"2494 slice_324" -> "2495 fill__98"  [label="(4, 4)", style=solid];
+"2496 view_117" -> "2497 permute_98"  [label="(2, 8, 2, 8)", style=solid];
+"2497 permute_98" -> "2498 reshape_96"  [label="(2, 2, 8, 8)", style=solid];
+"2498 reshape_96" -> "2499 unsqueeze_62"  [label="(4, 64)", style=solid];
+"2498 reshape_96" -> "2500 unsqueeze_63"  [label="(4, 64)", style=solid];
+"2499 unsqueeze_62" -> "2501 sub_10"  [label="(4, 1, 64)", style=solid];
+"2500 unsqueeze_63" -> "2501 sub_10"  [label="(4, 64, 1)", style=solid];
+"2501 sub_10" -> "2502 ne_10"  [label="(4, 64, 64)", style=solid];
+"2501 sub_10" -> "2503 masked_fill_20"  [label="(4, 64, 64)", style=solid];
+"2501 sub_10" -> "2504 eq_10"  [label="(4, 64, 64)", style=solid];
+"2502 ne_10" -> "2503 masked_fill_20"  [label="(4, 64, 64)", style=solid];
+"2503 masked_fill_20" -> "2505 masked_fill_21"  [label="(4, 64, 64)", style=solid];
+"2504 eq_10" -> "2505 masked_fill_21"  [label="(4, 64, 64)", style=solid];
+"2505 masked_fill_21" -> "2507 unsqueeze_64"  [label="(4, 64, 64)", style=solid];
+"2506 view_118" -> "2509 add_74"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2507 unsqueeze_64" -> "2508 unsqueeze_65"  [label="(4, 1, 64, 64)", style=solid];
+"2508 unsqueeze_65" -> "2509 add_74"  [label="(1, 4, 1, 64, 64)", style=solid];
+"2509 add_74" -> "2510 view_119"  [label="(1, 4, 12, 64, 64)", style=solid];
+"2510 view_119" -> "2511 softmax_21"  [label="(4, 12, 64, 64)", style=solid];
+"2511 softmax_21" -> "2512 dropout_84"  [label="(4, 12, 64, 64)", style=solid];
+"2512 dropout_84" -> "2513 matmul_43"  [label="(4, 12, 64, 64)", style=solid];
+"2513 matmul_43" -> "2514 transpose_43"  [label="(4, 12, 64, 32)", style=solid];
+"2514 transpose_43" -> "2515 reshape_97"  [label="(4, 64, 12, 32)", style=solid];
+"2515 reshape_97" -> "2516 linear_131"  [label="(4, 64, 384)", style=solid];
+"2516 linear_131" -> "2517 dropout_85"  [label="(4, 64, 384)", style=solid];
+"2517 dropout_85" -> "2518 view_120"  [label="(4, 64, 384)", style=solid];
+"2518 view_120" -> "2519 permute_99"  [label="(1, 2, 2, 8, 8, 384)", style=solid];
+"2519 permute_99" -> "2520 reshape_98"  [label="(1, 2, 8, 2, 8, 384)", style=solid];
+"2520 reshape_98" -> "2521 roll_21"  [label="(1, 16, 16, 384)", style=solid];
+"2521 roll_21" -> "2522 slice_325"  [label="(1, 16, 16, 384)", style=solid];
+"2522 slice_325" -> "2523 slice_326"  [label="(1, 16, 16, 384)", style=solid];
+"2523 slice_326" -> "2524 slice_327"  [label="(1, 14, 16, 384)", style=solid];
+"2524 slice_327" -> "2525 slice_328"  [label="(1, 14, 14, 384)", style=solid];
+"2525 slice_328" -> "2526 contiguous_41"  [label="(1, 14, 14, 384)", style=solid];
+"2526 contiguous_41" -> "2527 layer_norm_45"  [label="(1, 14, 14, 384)", style=solid];
+"2527 layer_norm_45" -> "2528 add_75"  [label="(1, 14, 14, 384)", style=solid];
+"2528 add_75" -> "2529 linear_132"  [label="(1, 14, 14, 384)", style=solid];
+"2528 add_75" -> "2535 add_76"  [label="(1, 14, 14, 384)", style=solid];
+"2529 linear_132" -> "2530 gelu_21"  [label="(1, 14, 14, 1536)", style=solid];
+"2530 gelu_21" -> "2531 dropout_86"  [label="(1, 14, 14, 1536)", style=solid];
+"2531 dropout_86" -> "2532 linear_133"  [label="(1, 14, 14, 1536)", style=solid];
+"2532 linear_133" -> "2533 dropout_87"  [label="(1, 14, 14, 384)", style=solid];
+"2533 dropout_87" -> "2534 layer_norm_46"  [label="(1, 14, 14, 384)", style=solid];
+"2534 layer_norm_46" -> "2535 add_76"  [label="(1, 14, 14, 384)", style=solid];
+"2535 add_76" -> "2536 pad_24"  [label="(1, 14, 14, 384)", style=solid];
+"2536 pad_24" -> "2537 slice_329"  [label="(1, 14, 14, 384)", style=solid];
+"2536 pad_24" -> "2540 slice_332"  [label="(1, 14, 14, 384)", style=solid];
+"2536 pad_24" -> "2543 slice_335"  [label="(1, 14, 14, 384)", style=solid];
+"2536 pad_24" -> "2546 slice_338"  [label="(1, 14, 14, 384)", style=solid];
+"2537 slice_329" -> "2538 slice_330"  [label="(1, 7, 14, 384)", style=solid];
+"2538 slice_330" -> "2539 slice_331"  [label="(1, 7, 7, 384)", style=solid];
+"2539 slice_331" -> "2549 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2540 slice_332" -> "2541 slice_333"  [label="(1, 7, 14, 384)", style=solid];
+"2541 slice_333" -> "2542 slice_334"  [label="(1, 7, 7, 384)", style=solid];
+"2542 slice_334" -> "2549 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2543 slice_335" -> "2544 slice_336"  [label="(1, 7, 14, 384)", style=solid];
+"2544 slice_336" -> "2545 slice_337"  [label="(1, 7, 7, 384)", style=solid];
+"2545 slice_337" -> "2549 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2546 slice_338" -> "2547 slice_339"  [label="(1, 7, 14, 384)", style=solid];
+"2547 slice_339" -> "2548 slice_340"  [label="(1, 7, 7, 384)", style=solid];
+"2548 slice_340" -> "2549 cat_2"  [label="(1, 7, 7, 384)", style=solid];
+"2549 cat_2" -> "2550 linear_134"  [label="(1, 7, 7, 1536)", style=solid];
+"2550 linear_134" -> "2551 layer_norm_47"  [label="(1, 7, 7, 768)", style=solid];
+"2551 layer_norm_47" -> "2563 pad_25"  [label="(1, 7, 7, 768)", style=solid];
+"2551 layer_norm_47" -> "2606 add_78"  [label="(1, 7, 7, 768)", style=solid];
+"2552 linear_135" -> "2553 relu__22"  [label="(1, 15, 15, 512)", style=solid];
+"2553 relu__22" -> "2554 linear_136"  [label="(1, 15, 15, 512)", style=solid];
+"2554 linear_136" -> "2555 view_121"  [label="(1, 15, 15, 24)", style=solid];
+"2555 view_121" -> "2556 index_22"  [label="(225, 24)", style=solid];
+"2556 index_22" -> "2557 view_122"  [label="(4096, 24)", style=solid];
+"2557 view_122" -> "2558 permute_100"  [label="(64, 64, 24)", style=solid];
+"2558 permute_100" -> "2559 contiguous_42"  [label="(24, 64, 64)", style=solid];
+"2559 contiguous_42" -> "2560 unsqueeze_66"  [label="(24, 64, 64)", style=solid];
+"2560 unsqueeze_66" -> "2561 sigmoid_22"  [label="(1, 24, 64, 64)", style=solid];
+"2561 sigmoid_22" -> "2562 mul_44"  [label="(1, 24, 64, 64)", style=solid];
+"2562 mul_44" -> "2589 add_77"  [label="(1, 24, 64, 64)", style=solid];
+"2563 pad_25" -> "2564 view_123"  [label="(1, 8, 8, 768)", style=solid];
+"2564 view_123" -> "2565 permute_101"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2565 permute_101" -> "2566 reshape_99"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2566 reshape_99" -> "2570 linear_137"  [label="(1, 64, 768)", style=solid];
+"2567 clone_22" -> "2568 slice_341"  [label="(2304,)", style=solid];
+"2567 clone_22" -> "2570 linear_137"  [label="(2304,)", style=solid];
+"2568 slice_341" -> "2569 zero__22"  [label="(768,)", style=solid];
+"2570 linear_137" -> "2571 reshape_100"  [label="(1, 64, 2304)", style=solid];
+"2571 reshape_100" -> "2572 permute_102"  [label="(1, 64, 3, 24, 32)", style=solid];
+"2572 permute_102" -> "2573 select_66"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2572 permute_102" -> "2574 select_67"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2572 permute_102" -> "2575 select_68"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2573 select_66" -> "2576 linalg_vector_norm_44"  [label="(1, 24, 64, 32)", style=solid];
+"2573 select_66" -> "2578 expand_as_44"  [label="(1, 24, 64, 32)", style=solid];
+"2573 select_66" -> "2579 div_44"  [label="(1, 24, 64, 32)", style=solid];
+"2574 select_67" -> "2580 linalg_vector_norm_45"  [label="(1, 24, 64, 32)", style=solid];
+"2574 select_67" -> "2582 expand_as_45"  [label="(1, 24, 64, 32)", style=solid];
+"2574 select_67" -> "2583 div_45"  [label="(1, 24, 64, 32)", style=solid];
+"2575 select_68" -> "2592 matmul_45"  [label="(1, 24, 64, 32)", style=solid];
+"2576 linalg_vector_norm_44" -> "2577 clamp_min_44"  [label="(1, 24, 64, 1)", style=solid];
+"2577 clamp_min_44" -> "2578 expand_as_44"  [label="(1, 24, 64, 1)", style=solid];
+"2578 expand_as_44" -> "2579 div_44"  [label="(1, 24, 64, 32)", style=solid];
+"2579 div_44" -> "2585 matmul_44"  [label="(1, 24, 64, 32)", style=solid];
+"2580 linalg_vector_norm_45" -> "2581 clamp_min_45"  [label="(1, 24, 64, 1)", style=solid];
+"2581 clamp_min_45" -> "2582 expand_as_45"  [label="(1, 24, 64, 1)", style=solid];
+"2582 expand_as_45" -> "2583 div_45"  [label="(1, 24, 64, 32)", style=solid];
+"2583 div_45" -> "2584 transpose_44"  [label="(1, 24, 64, 32)", style=solid];
+"2584 transpose_44" -> "2585 matmul_44"  [label="(1, 24, 32, 64)", style=solid];
+"2585 matmul_44" -> "2588 mul_45"  [label="(1, 24, 64, 64)", style=solid];
+"2586 clamp_22" -> "2587 exp_22"  [label="(24, 1, 1)", style=solid];
+"2587 exp_22" -> "2588 mul_45"  [label="(24, 1, 1)", style=solid];
+"2588 mul_45" -> "2589 add_77"  [label="(1, 24, 64, 64)", style=solid];
+"2589 add_77" -> "2590 softmax_22"  [label="(1, 24, 64, 64)", style=solid];
+"2590 softmax_22" -> "2591 dropout_88"  [label="(1, 24, 64, 64)", style=solid];
+"2591 dropout_88" -> "2592 matmul_45"  [label="(1, 24, 64, 64)", style=solid];
+"2592 matmul_45" -> "2593 transpose_45"  [label="(1, 24, 64, 32)", style=solid];
+"2593 transpose_45" -> "2594 reshape_101"  [label="(1, 64, 24, 32)", style=solid];
+"2594 reshape_101" -> "2595 linear_138"  [label="(1, 64, 768)", style=solid];
+"2595 linear_138" -> "2596 dropout_89"  [label="(1, 64, 768)", style=solid];
+"2596 dropout_89" -> "2597 view_124"  [label="(1, 64, 768)", style=solid];
+"2597 view_124" -> "2598 permute_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2598 permute_103" -> "2599 reshape_102"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2599 reshape_102" -> "2600 slice_342"  [label="(1, 8, 8, 768)", style=solid];
+"2600 slice_342" -> "2601 slice_343"  [label="(1, 8, 8, 768)", style=solid];
+"2601 slice_343" -> "2602 slice_344"  [label="(1, 7, 8, 768)", style=solid];
+"2602 slice_344" -> "2603 slice_345"  [label="(1, 7, 7, 768)", style=solid];
+"2603 slice_345" -> "2604 contiguous_43"  [label="(1, 7, 7, 768)", style=solid];
+"2604 contiguous_43" -> "2605 layer_norm_48"  [label="(1, 7, 7, 768)", style=solid];
+"2605 layer_norm_48" -> "2606 add_78"  [label="(1, 7, 7, 768)", style=solid];
+"2606 add_78" -> "2607 linear_139"  [label="(1, 7, 7, 768)", style=solid];
+"2606 add_78" -> "2613 add_79"  [label="(1, 7, 7, 768)", style=solid];
+"2607 linear_139" -> "2608 gelu_22"  [label="(1, 7, 7, 3072)", style=solid];
+"2608 gelu_22" -> "2609 dropout_90"  [label="(1, 7, 7, 3072)", style=solid];
+"2609 dropout_90" -> "2610 linear_140"  [label="(1, 7, 7, 3072)", style=solid];
+"2610 linear_140" -> "2611 dropout_91"  [label="(1, 7, 7, 768)", style=solid];
+"2611 dropout_91" -> "2612 layer_norm_49"  [label="(1, 7, 7, 768)", style=solid];
+"2612 layer_norm_49" -> "2613 add_79"  [label="(1, 7, 7, 768)", style=solid];
+"2613 add_79" -> "2625 pad_26"  [label="(1, 7, 7, 768)", style=solid];
+"2613 add_79" -> "2668 add_81"  [label="(1, 7, 7, 768)", style=solid];
+"2614 linear_141" -> "2615 relu__23"  [label="(1, 15, 15, 512)", style=solid];
+"2615 relu__23" -> "2616 linear_142"  [label="(1, 15, 15, 512)", style=solid];
+"2616 linear_142" -> "2617 view_125"  [label="(1, 15, 15, 24)", style=solid];
+"2617 view_125" -> "2618 index_23"  [label="(225, 24)", style=solid];
+"2618 index_23" -> "2619 view_126"  [label="(4096, 24)", style=solid];
+"2619 view_126" -> "2620 permute_104"  [label="(64, 64, 24)", style=solid];
+"2620 permute_104" -> "2621 contiguous_44"  [label="(24, 64, 64)", style=solid];
+"2621 contiguous_44" -> "2622 unsqueeze_67"  [label="(24, 64, 64)", style=solid];
+"2622 unsqueeze_67" -> "2623 sigmoid_23"  [label="(1, 24, 64, 64)", style=solid];
+"2623 sigmoid_23" -> "2624 mul_46"  [label="(1, 24, 64, 64)", style=solid];
+"2624 mul_46" -> "2651 add_80"  [label="(1, 24, 64, 64)", style=solid];
+"2625 pad_26" -> "2626 view_127"  [label="(1, 8, 8, 768)", style=solid];
+"2626 view_127" -> "2627 permute_105"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2627 permute_105" -> "2628 reshape_103"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2628 reshape_103" -> "2632 linear_143"  [label="(1, 64, 768)", style=solid];
+"2629 clone_23" -> "2630 slice_346"  [label="(2304,)", style=solid];
+"2629 clone_23" -> "2632 linear_143"  [label="(2304,)", style=solid];
+"2630 slice_346" -> "2631 zero__23"  [label="(768,)", style=solid];
+"2632 linear_143" -> "2633 reshape_104"  [label="(1, 64, 2304)", style=solid];
+"2633 reshape_104" -> "2634 permute_106"  [label="(1, 64, 3, 24, 32)", style=solid];
+"2634 permute_106" -> "2635 select_69"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2634 permute_106" -> "2636 select_70"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2634 permute_106" -> "2637 select_71"  [label="(3, 1, 24, 64, 32)", style=solid];
+"2635 select_69" -> "2638 linalg_vector_norm_46"  [label="(1, 24, 64, 32)", style=solid];
+"2635 select_69" -> "2640 expand_as_46"  [label="(1, 24, 64, 32)", style=solid];
+"2635 select_69" -> "2641 div_46"  [label="(1, 24, 64, 32)", style=solid];
+"2636 select_70" -> "2642 linalg_vector_norm_47"  [label="(1, 24, 64, 32)", style=solid];
+"2636 select_70" -> "2644 expand_as_47"  [label="(1, 24, 64, 32)", style=solid];
+"2636 select_70" -> "2645 div_47"  [label="(1, 24, 64, 32)", style=solid];
+"2637 select_71" -> "2654 matmul_47"  [label="(1, 24, 64, 32)", style=solid];
+"2638 linalg_vector_norm_46" -> "2639 clamp_min_46"  [label="(1, 24, 64, 1)", style=solid];
+"2639 clamp_min_46" -> "2640 expand_as_46"  [label="(1, 24, 64, 1)", style=solid];
+"2640 expand_as_46" -> "2641 div_46"  [label="(1, 24, 64, 32)", style=solid];
+"2641 div_46" -> "2647 matmul_46"  [label="(1, 24, 64, 32)", style=solid];
+"2642 linalg_vector_norm_47" -> "2643 clamp_min_47"  [label="(1, 24, 64, 1)", style=solid];
+"2643 clamp_min_47" -> "2644 expand_as_47"  [label="(1, 24, 64, 1)", style=solid];
+"2644 expand_as_47" -> "2645 div_47"  [label="(1, 24, 64, 32)", style=solid];
+"2645 div_47" -> "2646 transpose_46"  [label="(1, 24, 64, 32)", style=solid];
+"2646 transpose_46" -> "2647 matmul_46"  [label="(1, 24, 32, 64)", style=solid];
+"2647 matmul_46" -> "2650 mul_47"  [label="(1, 24, 64, 64)", style=solid];
+"2648 clamp_23" -> "2649 exp_23"  [label="(24, 1, 1)", style=solid];
+"2649 exp_23" -> "2650 mul_47"  [label="(24, 1, 1)", style=solid];
+"2650 mul_47" -> "2651 add_80"  [label="(1, 24, 64, 64)", style=solid];
+"2651 add_80" -> "2652 softmax_23"  [label="(1, 24, 64, 64)", style=solid];
+"2652 softmax_23" -> "2653 dropout_92"  [label="(1, 24, 64, 64)", style=solid];
+"2653 dropout_92" -> "2654 matmul_47"  [label="(1, 24, 64, 64)", style=solid];
+"2654 matmul_47" -> "2655 transpose_47"  [label="(1, 24, 64, 32)", style=solid];
+"2655 transpose_47" -> "2656 reshape_105"  [label="(1, 64, 24, 32)", style=solid];
+"2656 reshape_105" -> "2657 linear_144"  [label="(1, 64, 768)", style=solid];
+"2657 linear_144" -> "2658 dropout_93"  [label="(1, 64, 768)", style=solid];
+"2658 dropout_93" -> "2659 view_128"  [label="(1, 64, 768)", style=solid];
+"2659 view_128" -> "2660 permute_107"  [label="(1, 1, 1, 8, 8, 768)", style=solid];
+"2660 permute_107" -> "2661 reshape_106"  [label="(1, 1, 8, 1, 8, 768)", style=solid];
+"2661 reshape_106" -> "2662 slice_347"  [label="(1, 8, 8, 768)", style=solid];
+"2662 slice_347" -> "2663 slice_348"  [label="(1, 8, 8, 768)", style=solid];
+"2663 slice_348" -> "2664 slice_349"  [label="(1, 7, 8, 768)", style=solid];
+"2664 slice_349" -> "2665 slice_350"  [label="(1, 7, 7, 768)", style=solid];
+"2665 slice_350" -> "2666 contiguous_45"  [label="(1, 7, 7, 768)", style=solid];
+"2666 contiguous_45" -> "2667 layer_norm_50"  [label="(1, 7, 7, 768)", style=solid];
+"2667 layer_norm_50" -> "2668 add_81"  [label="(1, 7, 7, 768)", style=solid];
+"2668 add_81" -> "2669 linear_145"  [label="(1, 7, 7, 768)", style=solid];
+"2668 add_81" -> "2675 add_82"  [label="(1, 7, 7, 768)", style=solid];
+"2669 linear_145" -> "2670 gelu_23"  [label="(1, 7, 7, 3072)", style=solid];
+"2670 gelu_23" -> "2671 dropout_94"  [label="(1, 7, 7, 3072)", style=solid];
+"2671 dropout_94" -> "2672 linear_146"  [label="(1, 7, 7, 3072)", style=solid];
+"2672 linear_146" -> "2673 dropout_95"  [label="(1, 7, 7, 768)", style=solid];
+"2673 dropout_95" -> "2674 layer_norm_51"  [label="(1, 7, 7, 768)", style=solid];
+"2674 layer_norm_51" -> "2675 add_82"  [label="(1, 7, 7, 768)", style=solid];
+"2675 add_82" -> "2676 layer_norm_52"  [label="(1, 7, 7, 768)", style=solid];
+"2676 layer_norm_52" -> "2677 permute_108"  [label="(1, 7, 7, 768)", style=solid];
+"2677 permute_108" -> "2678 adaptive_avg_pool2d"  [label="(1, 768, 7, 7)", style=solid];
+"2678 adaptive_avg_pool2d" -> "2679 flatten"  [label="(1, 768, 1, 1)", style=solid];
+"2679 flatten" -> "2680 linear_147"  [label="(1, 768)", style=solid];
+"2680 linear_147" -> "2681 output_1"  [label="(1, 1000)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/synthetic_transformer.dot b/tests/torch/data/reference_graphs/fx/synthetic_transformer.dot
index 34212128ec4..41c895e3cc0 100644
--- a/tests/torch/data/reference_graphs/fx/synthetic_transformer.dot
+++ b/tests/torch/data/reference_graphs/fx/synthetic_transformer.dot
@@ -1,21 +1,21 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 embedding" [id=2, type=embedding];
-"3 _param_constant1" [id=3, type=get_attr];
-"4 _param_constant2" [id=4, type=get_attr];
-"5 linear" [id=5, type=linear];
-"6 _param_constant3" [id=6, type=get_attr];
-"7 _param_constant4" [id=7, type=get_attr];
+"0 wte_weight" [id=0, type=get_attr];
+"1 linear_weight" [id=1, type=get_attr];
+"2 linear_bias" [id=2, type=get_attr];
+"3 lm_head_weight" [id=3, type=get_attr];
+"4 lm_head_bias" [id=4, type=get_attr];
+"5 input_ids" [id=5, type=input];
+"6 embedding" [id=6, type=embedding];
+"7 linear" [id=7, type=linear];
 "8 linear_1" [id=8, type=linear];
-"9 output" [id=9, type=output];
-"0 arg0_1" -> "2 embedding"  [label="(5,)", style=solid];
-"1 _param_constant0" -> "2 embedding"  [label="(10, 5)", style=solid];
-"2 embedding" -> "5 linear"  [label="(5, 5)", style=solid];
-"3 _param_constant1" -> "5 linear"  [label="(5, 5)", style=solid];
-"4 _param_constant2" -> "5 linear"  [label="(5,)", style=solid];
-"5 linear" -> "8 linear_1"  [label="(5, 5)", style=solid];
-"6 _param_constant3" -> "8 linear_1"  [label="(10, 5)", style=solid];
-"7 _param_constant4" -> "8 linear_1"  [label="(10,)", style=solid];
-"8 linear_1" -> "9 output"  [label="(5, 10)", style=solid];
+"9 output_1" [id=9, type=output];
+"0 wte_weight" -> "6 embedding"  [label="(10, 5)", style=solid];
+"1 linear_weight" -> "7 linear"  [label="(5, 5)", style=solid];
+"2 linear_bias" -> "7 linear"  [label="(5,)", style=solid];
+"3 lm_head_weight" -> "8 linear_1"  [label="(10, 5)", style=solid];
+"4 lm_head_bias" -> "8 linear_1"  [label="(10,)", style=solid];
+"5 input_ids" -> "6 embedding"  [label="(5,)", style=solid];
+"6 embedding" -> "7 linear"  [label="(5, 5)", style=solid];
+"7 linear" -> "8 linear_1"  [label="(5, 5)", style=solid];
+"8 linear_1" -> "9 output_1"  [label="(5, 10)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_channel_invalid.dot b/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_channel_invalid.dot
index 5db5924962f..e311fcb3773 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_channel_invalid.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_channel_invalid.dot
@@ -1,50 +1,48 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 scale_node0" [id=3, type=get_attr];
-"4 weight_node0" [id=4, type=get_attr];
-"5 quantize_per_channel_default" [id=5, type=quantize_per_channel];
-"6 add_tensor_2" [id=6, type=add];
-"7 dequantize_per_channel_default" [id=7, type=dequantize_per_channel];
-"8 conv2d" [id=8, type=conv2d];
-"9 _param_constant2" [id=9, type=get_attr];
-"10 _param_constant3" [id=10, type=get_attr];
-"11 conv2d_1" [id=11, type=conv2d];
-"12 _tensor_constant0" [id=12, type=get_attr];
-"13 add_" [id=13, type=add_];
-"14 _tensor_constant0_1" [id=14, type=get_attr];
-"15 add__1" [id=15, type=add_];
-"16 add" [id=16, type=add];
-"17 _param_constant4" [id=17, type=get_attr];
-"18 _param_constant5" [id=18, type=get_attr];
-"19 conv2d_2" [id=19, type=conv2d];
-"20 _tensor_constant0_2" [id=20, type=get_attr];
-"21 add_1" [id=21, type=add];
-"22 output" [id=22, type=output];
-"0 arg0_1" -> "8 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"1 _param_constant0" -> "5 quantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "8 conv2d"  [label="(3,)", style=solid];
-"3 scale_node0" -> "5 quantize_per_channel_default"  [label="(3,)", style=solid];
-"3 scale_node0" -> "7 dequantize_per_channel_default"  [label="(3,)", style=solid];
-"4 weight_node0" -> "5 quantize_per_channel_default"  [label="(3,)", style=solid];
-"4 weight_node0" -> "7 dequantize_per_channel_default"  [label="(3,)", style=solid];
-"5 quantize_per_channel_default" -> "6 add_tensor_2"  [label="(3, 3, 1, 1)", style=solid];
-"6 add_tensor_2" -> "7 dequantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
-"7 dequantize_per_channel_default" -> "8 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"8 conv2d" -> "11 conv2d_1"  [label="(1, 3, 224, 224)", style=solid];
-"8 conv2d" -> "13 add_"  [label="(1, 3, 224, 224)", style=solid];
-"9 _param_constant2" -> "11 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"10 _param_constant3" -> "11 conv2d_1"  [label="(3,)", style=solid];
-"11 conv2d_1" -> "15 add__1"  [label="(1, 3, 224, 224)", style=solid];
-"12 _tensor_constant0" -> "13 add_"  [label="(1,)", style=solid];
-"13 add_" -> "16 add"  [label="(1, 3, 224, 224)", style=solid];
-"14 _tensor_constant0_1" -> "15 add__1"  [label="(1,)", style=solid];
-"15 add__1" -> "16 add"  [label="(1, 3, 224, 224)", style=solid];
-"16 add" -> "19 conv2d_2"  [label="(1, 3, 224, 224)", style=solid];
-"17 _param_constant4" -> "19 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"18 _param_constant5" -> "19 conv2d_2"  [label="(3,)", style=solid];
-"19 conv2d_2" -> "21 add_1"  [label="(1, 3, 224, 224)", style=solid];
-"20 _tensor_constant0_2" -> "21 add_1"  [label="(1,)", style=solid];
-"21 add_1" -> "22 output"  [label="(1, 3, 224, 224)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 scale_node0" [id=8, type=get_attr];
+"9 weight_node0" [id=9, type=get_attr];
+"10 quantize_per_channel_default" [id=10, type=quantize_per_channel];
+"11 add_tensor" [id=11, type=add];
+"12 dequantize_per_channel_default" [id=12, type=dequantize_per_channel];
+"13 conv2d" [id=13, type=conv2d];
+"14 conv2d_1" [id=14, type=conv2d];
+"15 add_" [id=15, type=add_];
+"16 add__1" [id=16, type=add_];
+"17 add" [id=17, type=add];
+"18 conv2d_2" [id=18, type=conv2d];
+"19 add_1" [id=19, type=add];
+"20 output_1" [id=20, type=output];
+"0 conv_a_weight" -> "10 quantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "13 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "14 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "14 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "18 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "18 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "15 add_"  [label="(1,)", style=solid];
+"6 bias" -> "16 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "19 add_1"  [label="(1,)", style=solid];
+"7 x" -> "13 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"8 scale_node0" -> "10 quantize_per_channel_default"  [label="(3,)", style=solid];
+"8 scale_node0" -> "12 dequantize_per_channel_default"  [label="(3,)", style=solid];
+"9 weight_node0" -> "10 quantize_per_channel_default"  [label="(3,)", style=solid];
+"9 weight_node0" -> "12 dequantize_per_channel_default"  [label="(3,)", style=solid];
+"10 quantize_per_channel_default" -> "11 add_tensor"  [label="(3, 3, 1, 1)", style=solid];
+"11 add_tensor" -> "12 dequantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
+"12 dequantize_per_channel_default" -> "13 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"13 conv2d" -> "14 conv2d_1"  [label="(1, 3, 224, 224)", style=solid];
+"13 conv2d" -> "15 add_"  [label="(1, 3, 224, 224)", style=solid];
+"14 conv2d_1" -> "16 add__1"  [label="(1, 3, 224, 224)", style=solid];
+"15 add_" -> "17 add"  [label="(1, 3, 224, 224)", style=solid];
+"16 add__1" -> "17 add"  [label="(1, 3, 224, 224)", style=solid];
+"17 add" -> "18 conv2d_2"  [label="(1, 3, 224, 224)", style=solid];
+"18 conv2d_2" -> "19 add_1"  [label="(1, 3, 224, 224)", style=solid];
+"19 add_1" -> "20 output_1"  [label="(1, 3, 224, 224)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_channel_valid.dot b/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_channel_valid.dot
index 232b2ba544a..8ed024a8e1f 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_channel_valid.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_channel_valid.dot
@@ -1,46 +1,44 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant1" [id=1, type=get_attr];
-"2 scale_updated_constant0" [id=2, type=get_attr];
-"3 compressed_weight_updated_constant0" [id=3, type=get_attr];
-"4 mul_tensor" [id=4, type=mul];
-"5 zero_point_updated_constant0" [id=5, type=get_attr];
-"6 sub_tensor" [id=6, type=sub];
-"7 conv2d" [id=7, type=conv2d];
-"8 _param_constant2" [id=8, type=get_attr];
-"9 _param_constant3" [id=9, type=get_attr];
-"10 conv2d_1" [id=10, type=conv2d];
-"11 _tensor_constant0" [id=11, type=get_attr];
-"12 add_" [id=12, type=add_];
-"13 _tensor_constant0_1" [id=13, type=get_attr];
-"14 add__1" [id=14, type=add_];
-"15 add" [id=15, type=add];
-"16 _param_constant4" [id=16, type=get_attr];
-"17 _param_constant5" [id=17, type=get_attr];
-"18 conv2d_2" [id=18, type=conv2d];
-"19 _tensor_constant0_2" [id=19, type=get_attr];
-"20 add_1" [id=20, type=add];
-"21 output" [id=21, type=output];
-"0 arg0_1" -> "7 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"1 _param_constant1" -> "7 conv2d"  [label="(3,)", style=solid];
-"2 scale_updated_constant0" -> "4 mul_tensor"  [label="(3, 1, 1, 1)", style=solid];
-"3 compressed_weight_updated_constant0" -> "4 mul_tensor"  [label="(3, 3, 1, 1)", style=solid];
-"4 mul_tensor" -> "6 sub_tensor"  [label="(3, 3, 1, 1)", style=solid];
-"5 zero_point_updated_constant0" -> "6 sub_tensor"  [label="(3, 1, 1, 1)", style=solid];
-"6 sub_tensor" -> "7 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"7 conv2d" -> "10 conv2d_1"  [label="(1, 3, 224, 224)", style=solid];
-"7 conv2d" -> "12 add_"  [label="(1, 3, 224, 224)", style=solid];
-"8 _param_constant2" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"9 _param_constant3" -> "10 conv2d_1"  [label="(3,)", style=solid];
-"10 conv2d_1" -> "14 add__1"  [label="(1, 3, 224, 224)", style=solid];
-"11 _tensor_constant0" -> "12 add_"  [label="(1,)", style=solid];
-"12 add_" -> "15 add"  [label="(1, 3, 224, 224)", style=solid];
-"13 _tensor_constant0_1" -> "14 add__1"  [label="(1,)", style=solid];
-"14 add__1" -> "15 add"  [label="(1, 3, 224, 224)", style=solid];
-"15 add" -> "18 conv2d_2"  [label="(1, 3, 224, 224)", style=solid];
-"16 _param_constant4" -> "18 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"17 _param_constant5" -> "18 conv2d_2"  [label="(3,)", style=solid];
-"18 conv2d_2" -> "20 add_1"  [label="(1, 3, 224, 224)", style=solid];
-"19 _tensor_constant0_2" -> "20 add_1"  [label="(1,)", style=solid];
-"20 add_1" -> "21 output"  [label="(1, 3, 224, 224)", style=solid];
+"0 conv_a_bias" [id=0, type=get_attr];
+"1 conv_b_weight" [id=1, type=get_attr];
+"2 conv_b_bias" [id=2, type=get_attr];
+"3 conv_c_weight" [id=3, type=get_attr];
+"4 conv_c_bias" [id=4, type=get_attr];
+"5 bias" [id=5, type=get_attr];
+"6 x" [id=6, type=input];
+"7 scale_updated_constant0" [id=7, type=get_attr];
+"8 compressed_weight_updated_constant0" [id=8, type=get_attr];
+"9 mul_tensor" [id=9, type=mul];
+"10 zero_point_updated_constant0" [id=10, type=get_attr];
+"11 sub_tensor" [id=11, type=sub];
+"12 conv2d" [id=12, type=conv2d];
+"13 conv2d_1" [id=13, type=conv2d];
+"14 add_" [id=14, type=add_];
+"15 add__1" [id=15, type=add_];
+"16 add" [id=16, type=add];
+"17 conv2d_2" [id=17, type=conv2d];
+"18 add_1" [id=18, type=add];
+"19 output_1" [id=19, type=output];
+"0 conv_a_bias" -> "12 conv2d"  [label="(3,)", style=solid];
+"1 conv_b_weight" -> "13 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"2 conv_b_bias" -> "13 conv2d_1"  [label="(3,)", style=solid];
+"3 conv_c_weight" -> "17 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"4 conv_c_bias" -> "17 conv2d_2"  [label="(3,)", style=solid];
+"5 bias" -> "14 add_"  [label="(1,)", style=solid];
+"5 bias" -> "15 add__1"  [label="(1,)", style=solid];
+"5 bias" -> "18 add_1"  [label="(1,)", style=solid];
+"6 x" -> "12 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"7 scale_updated_constant0" -> "9 mul_tensor"  [label="(3, 1, 1, 1)", style=solid];
+"8 compressed_weight_updated_constant0" -> "9 mul_tensor"  [label="(3, 3, 1, 1)", style=solid];
+"9 mul_tensor" -> "11 sub_tensor"  [label="(3, 3, 1, 1)", style=solid];
+"10 zero_point_updated_constant0" -> "11 sub_tensor"  [label="(3, 1, 1, 1)", style=solid];
+"11 sub_tensor" -> "12 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"12 conv2d" -> "13 conv2d_1"  [label="(1, 3, 224, 224)", style=solid];
+"12 conv2d" -> "14 add_"  [label="(1, 3, 224, 224)", style=solid];
+"13 conv2d_1" -> "15 add__1"  [label="(1, 3, 224, 224)", style=solid];
+"14 add_" -> "16 add"  [label="(1, 3, 224, 224)", style=solid];
+"15 add__1" -> "16 add"  [label="(1, 3, 224, 224)", style=solid];
+"16 add" -> "17 conv2d_2"  [label="(1, 3, 224, 224)", style=solid];
+"17 conv2d_2" -> "18 add_1"  [label="(1, 3, 224, 224)", style=solid];
+"18 add_1" -> "19 output_1"  [label="(1, 3, 224, 224)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_tensor_invalid.dot b/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_tensor_invalid.dot
index c603317a2d1..0e2b0d16eea 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_tensor_invalid.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_tensor_invalid.dot
@@ -1,44 +1,42 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 quantize_per_tensor_default" [id=3, type=quantize_per_tensor];
-"4 add_tensor_2" [id=4, type=add];
-"5 dequantize_per_tensor_default" [id=5, type=dequantize_per_tensor];
-"6 conv2d" [id=6, type=conv2d];
-"7 _param_constant2" [id=7, type=get_attr];
-"8 _param_constant3" [id=8, type=get_attr];
-"9 conv2d_1" [id=9, type=conv2d];
-"10 _tensor_constant0" [id=10, type=get_attr];
-"11 add_" [id=11, type=add_];
-"12 _tensor_constant0_1" [id=12, type=get_attr];
-"13 add__1" [id=13, type=add_];
-"14 add" [id=14, type=add];
-"15 _param_constant4" [id=15, type=get_attr];
-"16 _param_constant5" [id=16, type=get_attr];
-"17 conv2d_2" [id=17, type=conv2d];
-"18 _tensor_constant0_2" [id=18, type=get_attr];
-"19 add_1" [id=19, type=add];
-"20 output" [id=20, type=output];
-"0 arg0_1" -> "6 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"1 _param_constant0" -> "3 quantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "6 conv2d"  [label="(3,)", style=solid];
-"3 quantize_per_tensor_default" -> "4 add_tensor_2"  [label="(3, 3, 1, 1)", style=solid];
-"4 add_tensor_2" -> "5 dequantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
-"5 dequantize_per_tensor_default" -> "6 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"6 conv2d" -> "9 conv2d_1"  [label="(1, 3, 224, 224)", style=solid];
-"6 conv2d" -> "11 add_"  [label="(1, 3, 224, 224)", style=solid];
-"7 _param_constant2" -> "9 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"8 _param_constant3" -> "9 conv2d_1"  [label="(3,)", style=solid];
-"9 conv2d_1" -> "13 add__1"  [label="(1, 3, 224, 224)", style=solid];
-"10 _tensor_constant0" -> "11 add_"  [label="(1,)", style=solid];
-"11 add_" -> "14 add"  [label="(1, 3, 224, 224)", style=solid];
-"12 _tensor_constant0_1" -> "13 add__1"  [label="(1,)", style=solid];
-"13 add__1" -> "14 add"  [label="(1, 3, 224, 224)", style=solid];
-"14 add" -> "17 conv2d_2"  [label="(1, 3, 224, 224)", style=solid];
-"15 _param_constant4" -> "17 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"16 _param_constant5" -> "17 conv2d_2"  [label="(3,)", style=solid];
-"17 conv2d_2" -> "19 add_1"  [label="(1, 3, 224, 224)", style=solid];
-"18 _tensor_constant0_2" -> "19 add_1"  [label="(1,)", style=solid];
-"19 add_1" -> "20 output"  [label="(1, 3, 224, 224)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 quantize_per_tensor_default" [id=8, type=quantize_per_tensor];
+"9 add_tensor" [id=9, type=add];
+"10 dequantize_per_tensor_default" [id=10, type=dequantize_per_tensor];
+"11 conv2d" [id=11, type=conv2d];
+"12 conv2d_1" [id=12, type=conv2d];
+"13 add_" [id=13, type=add_];
+"14 add__1" [id=14, type=add_];
+"15 add" [id=15, type=add];
+"16 conv2d_2" [id=16, type=conv2d];
+"17 add_1" [id=17, type=add];
+"18 output_1" [id=18, type=output];
+"0 conv_a_weight" -> "8 quantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "11 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "12 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "12 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "16 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "16 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "13 add_"  [label="(1,)", style=solid];
+"6 bias" -> "14 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "17 add_1"  [label="(1,)", style=solid];
+"7 x" -> "11 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"8 quantize_per_tensor_default" -> "9 add_tensor"  [label="(3, 3, 1, 1)", style=solid];
+"9 add_tensor" -> "10 dequantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
+"10 dequantize_per_tensor_default" -> "11 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"11 conv2d" -> "12 conv2d_1"  [label="(1, 3, 224, 224)", style=solid];
+"11 conv2d" -> "13 add_"  [label="(1, 3, 224, 224)", style=solid];
+"12 conv2d_1" -> "14 add__1"  [label="(1, 3, 224, 224)", style=solid];
+"13 add_" -> "15 add"  [label="(1, 3, 224, 224)", style=solid];
+"14 add__1" -> "15 add"  [label="(1, 3, 224, 224)", style=solid];
+"15 add" -> "16 conv2d_2"  [label="(1, 3, 224, 224)", style=solid];
+"16 conv2d_2" -> "17 add_1"  [label="(1, 3, 224, 224)", style=solid];
+"17 add_1" -> "18 output_1"  [label="(1, 3, 224, 224)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_tensor_valid.dot b/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_tensor_valid.dot
index d4b52ec0840..9d0e7e08f2c 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_tensor_valid.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/compress_post_quantize_per_tensor_valid.dot
@@ -1,42 +1,40 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant1" [id=1, type=get_attr];
-"2 compressed_weight_updated_constant0" [id=2, type=get_attr];
-"3 mul_tensor" [id=3, type=mul];
-"4 sub_tensor" [id=4, type=sub];
-"5 conv2d" [id=5, type=conv2d];
-"6 _param_constant2" [id=6, type=get_attr];
-"7 _param_constant3" [id=7, type=get_attr];
-"8 conv2d_1" [id=8, type=conv2d];
-"9 _tensor_constant0" [id=9, type=get_attr];
-"10 add_" [id=10, type=add_];
-"11 _tensor_constant0_1" [id=11, type=get_attr];
-"12 add__1" [id=12, type=add_];
-"13 add" [id=13, type=add];
-"14 _param_constant4" [id=14, type=get_attr];
-"15 _param_constant5" [id=15, type=get_attr];
-"16 conv2d_2" [id=16, type=conv2d];
-"17 _tensor_constant0_2" [id=17, type=get_attr];
-"18 add_1" [id=18, type=add];
-"19 output" [id=19, type=output];
-"0 arg0_1" -> "5 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"1 _param_constant1" -> "5 conv2d"  [label="(3,)", style=solid];
-"2 compressed_weight_updated_constant0" -> "3 mul_tensor"  [label="(3, 3, 1, 1)", style=solid];
-"3 mul_tensor" -> "4 sub_tensor"  [label="(3, 3, 1, 1)", style=solid];
-"4 sub_tensor" -> "5 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"5 conv2d" -> "8 conv2d_1"  [label="(1, 3, 224, 224)", style=solid];
-"5 conv2d" -> "10 add_"  [label="(1, 3, 224, 224)", style=solid];
-"6 _param_constant2" -> "8 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"7 _param_constant3" -> "8 conv2d_1"  [label="(3,)", style=solid];
-"8 conv2d_1" -> "12 add__1"  [label="(1, 3, 224, 224)", style=solid];
-"9 _tensor_constant0" -> "10 add_"  [label="(1,)", style=solid];
-"10 add_" -> "13 add"  [label="(1, 3, 224, 224)", style=solid];
-"11 _tensor_constant0_1" -> "12 add__1"  [label="(1,)", style=solid];
-"12 add__1" -> "13 add"  [label="(1, 3, 224, 224)", style=solid];
-"13 add" -> "16 conv2d_2"  [label="(1, 3, 224, 224)", style=solid];
-"14 _param_constant4" -> "16 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"15 _param_constant5" -> "16 conv2d_2"  [label="(3,)", style=solid];
-"16 conv2d_2" -> "18 add_1"  [label="(1, 3, 224, 224)", style=solid];
-"17 _tensor_constant0_2" -> "18 add_1"  [label="(1,)", style=solid];
-"18 add_1" -> "19 output"  [label="(1, 3, 224, 224)", style=solid];
+"0 conv_a_bias" [id=0, type=get_attr];
+"1 conv_b_weight" [id=1, type=get_attr];
+"2 conv_b_bias" [id=2, type=get_attr];
+"3 conv_c_weight" [id=3, type=get_attr];
+"4 conv_c_bias" [id=4, type=get_attr];
+"5 bias" [id=5, type=get_attr];
+"6 x" [id=6, type=input];
+"7 compressed_weight_updated_constant0" [id=7, type=get_attr];
+"8 mul_tensor" [id=8, type=mul];
+"9 sub_tensor" [id=9, type=sub];
+"10 conv2d" [id=10, type=conv2d];
+"11 conv2d_1" [id=11, type=conv2d];
+"12 add_" [id=12, type=add_];
+"13 add__1" [id=13, type=add_];
+"14 add" [id=14, type=add];
+"15 conv2d_2" [id=15, type=conv2d];
+"16 add_1" [id=16, type=add];
+"17 output_1" [id=17, type=output];
+"0 conv_a_bias" -> "10 conv2d"  [label="(3,)", style=solid];
+"1 conv_b_weight" -> "11 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"2 conv_b_bias" -> "11 conv2d_1"  [label="(3,)", style=solid];
+"3 conv_c_weight" -> "15 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"4 conv_c_bias" -> "15 conv2d_2"  [label="(3,)", style=solid];
+"5 bias" -> "12 add_"  [label="(1,)", style=solid];
+"5 bias" -> "13 add__1"  [label="(1,)", style=solid];
+"5 bias" -> "16 add_1"  [label="(1,)", style=solid];
+"6 x" -> "10 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"7 compressed_weight_updated_constant0" -> "8 mul_tensor"  [label="(3, 3, 1, 1)", style=solid];
+"8 mul_tensor" -> "9 sub_tensor"  [label="(3, 3, 1, 1)", style=solid];
+"9 sub_tensor" -> "10 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"10 conv2d" -> "11 conv2d_1"  [label="(1, 3, 224, 224)", style=solid];
+"10 conv2d" -> "12 add_"  [label="(1, 3, 224, 224)", style=solid];
+"11 conv2d_1" -> "13 add__1"  [label="(1, 3, 224, 224)", style=solid];
+"12 add_" -> "14 add"  [label="(1, 3, 224, 224)", style=solid];
+"13 add__1" -> "14 add"  [label="(1, 3, 224, 224)", style=solid];
+"14 add" -> "15 conv2d_2"  [label="(1, 3, 224, 224)", style=solid];
+"15 conv2d_2" -> "16 add_1"  [label="(1, 3, 224, 224)", style=solid];
+"16 add_1" -> "17 output_1"  [label="(1, 3, 224, 224)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/constant_update.dot b/tests/torch/data/reference_graphs/fx/transformed/constant_update.dot
index 36b9e1bfff9..dde22a7ecba 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/constant_update.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/constant_update.dot
@@ -1,38 +1,36 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d" [id=3, type=conv2d];
-"4 _param_constant2" [id=4, type=get_attr];
-"5 _param_constant3" [id=5, type=get_attr];
-"6 conv2d_1" [id=6, type=conv2d];
-"7 add__updated_constant0" [id=7, type=get_attr];
-"8 add_" [id=8, type=add_];
-"9 _tensor_constant0_1" [id=9, type=get_attr];
-"10 add__1" [id=10, type=add_];
-"11 add" [id=11, type=add];
-"12 _param_constant4" [id=12, type=get_attr];
-"13 _param_constant5" [id=13, type=get_attr];
-"14 conv2d_2" [id=14, type=conv2d];
-"15 _tensor_constant0_2" [id=15, type=get_attr];
-"16 add_1" [id=16, type=add];
-"17 output" [id=17, type=output];
-"0 arg0_1" -> "3 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"3 conv2d" -> "6 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"3 conv2d" -> "8 add_"  [label="(1, 3, 3, 3)", style=solid];
-"4 _param_constant2" -> "6 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"5 _param_constant3" -> "6 conv2d_1"  [label="(3,)", style=solid];
-"6 conv2d_1" -> "10 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"7 add__updated_constant0" -> "8 add_"  [label="(1,)", style=solid];
-"8 add_" -> "11 add"  [label="(1, 3, 3, 3)", style=solid];
-"9 _tensor_constant0_1" -> "10 add__1"  [label="(1,)", style=solid];
-"10 add__1" -> "11 add"  [label="(1, 3, 3, 3)", style=solid];
-"11 add" -> "14 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"12 _param_constant4" -> "14 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"13 _param_constant5" -> "14 conv2d_2"  [label="(3,)", style=solid];
-"14 conv2d_2" -> "16 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"15 _tensor_constant0_2" -> "16 add_1"  [label="(1,)", style=solid];
-"16 add_1" -> "17 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 x" [id=6, type=input];
+"7 conv2d" [id=7, type=conv2d];
+"8 conv2d_1" [id=8, type=conv2d];
+"9 add__updated_constant0" [id=9, type=get_attr];
+"10 add_" [id=10, type=add_];
+"11 add__1" [id=11, type=add_];
+"12 add" [id=12, type=add];
+"13 conv2d_2" [id=13, type=conv2d];
+"14 add_1" [id=14, type=add];
+"15 output_1" [id=15, type=output];
+"0 conv_a_weight" -> "7 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "7 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "8 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "8 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "13 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "13 conv2d_2"  [label="(3,)", style=solid];
+"6 x" -> "7 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"7 conv2d" -> "8 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"7 conv2d" -> "10 add_"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d_1" -> "11 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"9 add__updated_constant0" -> "10 add_"  [label="(1,)", style=solid];
+"9 add__updated_constant0" -> "11 add__1"  [label="(1,)", style=solid];
+"9 add__updated_constant0" -> "14 add_1"  [label="(1,)", style=solid];
+"10 add_" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
+"11 add__1" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
+"12 add" -> "13 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"13 conv2d_2" -> "14 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"14 add_1" -> "15 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/folded_model.dot b/tests/torch/data/reference_graphs/fx/transformed/folded_model.dot
index 5c9bc911d09..2dea2e7bd93 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/folded_model.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/folded_model.dot
@@ -1,15 +1,15 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _frozen_param0" [id=1, type=get_attr];
-"2 _param_constant3" [id=2, type=get_attr];
-"3 _param_constant4" [id=3, type=get_attr];
+"0 linear_act_weight" [id=0, type=get_attr];
+"1 linear_act_bias" [id=1, type=get_attr];
+"2 x" [id=2, type=input];
+"3 _frozen_param0" [id=3, type=get_attr];
 "4 linear_1" [id=4, type=linear];
 "5 add" [id=5, type=add];
 "6 output" [id=6, type=output];
-"0 arg0_1" -> "4 linear_1"  [label="(1, 3, 3, 3)", style=solid];
-"1 _frozen_param0" -> "5 add"  [label="(3, 3)", style=solid];
-"2 _param_constant3" -> "4 linear_1"  [label="(3, 3)", style=solid];
-"3 _param_constant4" -> "4 linear_1"  [label="(3,)", style=solid];
+"0 linear_act_weight" -> "4 linear_1"  [label="(3, 3)", style=solid];
+"1 linear_act_bias" -> "4 linear_1"  [label="(3,)", style=solid];
+"2 x" -> "4 linear_1"  [label="(1, 3, 3, 3)", style=solid];
+"3 _frozen_param0" -> "5 add"  [label="(3, 3)", style=solid];
 "4 linear_1" -> "5 add"  [label="(1, 3, 3, 3)", style=solid];
 "5 add" -> "6 output"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/folded_model_with_constraints_per_channel.dot b/tests/torch/data/reference_graphs/fx/transformed/folded_model_with_constraints_per_channel.dot
index ec743cabec6..43efd89ba6f 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/folded_model_with_constraints_per_channel.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/folded_model_with_constraints_per_channel.dot
@@ -1,19 +1,19 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _frozen_param0" [id=1, type=get_attr];
-"2 _param_constant3" [id=2, type=get_attr];
-"3 _param_constant4" [id=3, type=get_attr];
+"0 linear_act_weight" [id=0, type=get_attr];
+"1 linear_act_bias" [id=1, type=get_attr];
+"2 x" [id=2, type=input];
+"3 _frozen_param0" [id=3, type=get_attr];
 "4 scale_node0" [id=4, type=get_attr];
 "5 weight_node0" [id=5, type=get_attr];
 "6 quantize_per_channel_default" [id=6, type=quantize_per_channel];
 "7 dequantize_per_channel_default" [id=7, type=dequantize_per_channel];
 "8 linear_1" [id=8, type=linear];
 "9 add" [id=9, type=add];
-"10 output" [id=10, type=output];
-"0 arg0_1" -> "8 linear_1"  [label="(1, 3, 3, 3)", style=solid];
-"1 _frozen_param0" -> "9 add"  [label="(3, 3)", style=solid];
-"2 _param_constant3" -> "6 quantize_per_channel_default"  [label="(3, 3)", style=solid];
-"3 _param_constant4" -> "8 linear_1"  [label="(3,)", style=solid];
+"10 output_1" [id=10, type=output];
+"0 linear_act_weight" -> "6 quantize_per_channel_default"  [label="(3, 3)", style=solid];
+"1 linear_act_bias" -> "8 linear_1"  [label="(3,)", style=solid];
+"2 x" -> "8 linear_1"  [label="(1, 3, 3, 3)", style=solid];
+"3 _frozen_param0" -> "9 add"  [label="(3, 3)", style=solid];
 "4 scale_node0" -> "6 quantize_per_channel_default"  [label="(3,)", style=solid];
 "4 scale_node0" -> "7 dequantize_per_channel_default"  [label="(3,)", style=solid];
 "5 weight_node0" -> "6 quantize_per_channel_default"  [label="(3,)", style=solid];
@@ -21,5 +21,5 @@ strict digraph  {
 "6 quantize_per_channel_default" -> "7 dequantize_per_channel_default"  [label="(3, 3)", style=solid];
 "7 dequantize_per_channel_default" -> "8 linear_1"  [label="(3, 3)", style=solid];
 "8 linear_1" -> "9 add"  [label="(1, 3, 3, 3)", style=solid];
-"9 add" -> "10 output"  [label="(1, 3, 3, 3)", style=solid];
+"9 add" -> "10 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/folded_model_with_constraints_per_tensor.dot b/tests/torch/data/reference_graphs/fx/transformed/folded_model_with_constraints_per_tensor.dot
index 91c418afe62..16e7d368683 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/folded_model_with_constraints_per_tensor.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/folded_model_with_constraints_per_tensor.dot
@@ -1,19 +1,19 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _frozen_param0" [id=1, type=get_attr];
-"2 _param_constant3" [id=2, type=get_attr];
-"3 _param_constant4" [id=3, type=get_attr];
+"0 linear_act_weight" [id=0, type=get_attr];
+"1 linear_act_bias" [id=1, type=get_attr];
+"2 x" [id=2, type=input];
+"3 _frozen_param0" [id=3, type=get_attr];
 "4 quantize_per_tensor_default" [id=4, type=quantize_per_tensor];
 "5 dequantize_per_tensor_default" [id=5, type=dequantize_per_tensor];
 "6 linear_1" [id=6, type=linear];
 "7 add" [id=7, type=add];
-"8 output" [id=8, type=output];
-"0 arg0_1" -> "6 linear_1"  [label="(1, 3, 3, 3)", style=solid];
-"1 _frozen_param0" -> "7 add"  [label="(3, 3)", style=solid];
-"2 _param_constant3" -> "4 quantize_per_tensor_default"  [label="(3, 3)", style=solid];
-"3 _param_constant4" -> "6 linear_1"  [label="(3,)", style=solid];
+"8 output_1" [id=8, type=output];
+"0 linear_act_weight" -> "4 quantize_per_tensor_default"  [label="(3, 3)", style=solid];
+"1 linear_act_bias" -> "6 linear_1"  [label="(3,)", style=solid];
+"2 x" -> "6 linear_1"  [label="(1, 3, 3, 3)", style=solid];
+"3 _frozen_param0" -> "7 add"  [label="(3, 3)", style=solid];
 "4 quantize_per_tensor_default" -> "5 dequantize_per_tensor_default"  [label="(3, 3)", style=solid];
 "5 dequantize_per_tensor_default" -> "6 linear_1"  [label="(3, 3)", style=solid];
 "6 linear_1" -> "7 add"  [label="(1, 3, 3, 3)", style=solid];
-"7 add" -> "8 output"  [label="(1, 3, 3, 3)", style=solid];
+"7 add" -> "8 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/model_insertion.dot b/tests/torch/data/reference_graphs/fx/transformed/model_insertion.dot
index a03dce9b464..825f345d634 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/model_insertion.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/model_insertion.dot
@@ -1,46 +1,44 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 TEST_MODULE_0" [id=3, type=call_module];
-"4 TEST_MODULE_1" [id=4, type=call_module];
-"5 conv2d" [id=5, type=conv2d];
-"6 TEST_MODULE_3" [id=6, type=call_module];
-"7 _param_constant2" [id=7, type=get_attr];
-"8 _param_constant3" [id=8, type=get_attr];
-"9 TEST_MODULE_2" [id=9, type=call_module];
-"10 conv2d_1" [id=10, type=conv2d];
-"11 _tensor_constant0" [id=11, type=get_attr];
-"12 add_" [id=12, type=add_];
-"13 _tensor_constant0_1" [id=13, type=get_attr];
-"14 add__1" [id=14, type=add_];
-"15 add" [id=15, type=add];
-"16 _param_constant4" [id=16, type=get_attr];
-"17 _param_constant5" [id=17, type=get_attr];
-"18 conv2d_2" [id=18, type=conv2d];
-"19 _tensor_constant0_2" [id=19, type=get_attr];
-"20 add_1" [id=20, type=add];
-"21 output" [id=21, type=output];
-"0 arg0_1" -> "3 TEST_MODULE_0"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "4 TEST_MODULE_1"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "5 conv2d"  [label="(3,)", style=solid];
-"3 TEST_MODULE_0" -> "5 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"4 TEST_MODULE_1" -> "5 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"5 conv2d" -> "6 TEST_MODULE_3"  [label="(1, 3, 3, 3)", style=solid];
-"6 TEST_MODULE_3" -> "10 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"6 TEST_MODULE_3" -> "12 add_"  [label="(1, 3, 3, 3)", style=solid];
-"7 _param_constant2" -> "9 TEST_MODULE_2"  [label="(3, 3, 1, 1)", style=solid];
-"8 _param_constant3" -> "10 conv2d_1"  [label="(3,)", style=solid];
-"9 TEST_MODULE_2" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"10 conv2d_1" -> "14 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"11 _tensor_constant0" -> "12 add_"  [label="(1,)", style=solid];
-"12 add_" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"13 _tensor_constant0_1" -> "14 add__1"  [label="(1,)", style=solid];
-"14 add__1" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"15 add" -> "18 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"16 _param_constant4" -> "18 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"17 _param_constant5" -> "18 conv2d_2"  [label="(3,)", style=solid];
-"18 conv2d_2" -> "20 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"19 _tensor_constant0_2" -> "20 add_1"  [label="(1,)", style=solid];
-"20 add_1" -> "21 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 TEST_MODULE_0" [id=8, type=call_module];
+"9 TEST_MODULE_1" [id=9, type=call_module];
+"10 conv2d" [id=10, type=conv2d];
+"11 TEST_MODULE_3" [id=11, type=call_module];
+"12 TEST_MODULE_2" [id=12, type=call_module];
+"13 conv2d_1" [id=13, type=conv2d];
+"14 add_" [id=14, type=add_];
+"15 add__1" [id=15, type=add_];
+"16 add" [id=16, type=add];
+"17 conv2d_2" [id=17, type=conv2d];
+"18 add_1" [id=18, type=add];
+"19 output_1" [id=19, type=output];
+"0 conv_a_weight" -> "9 TEST_MODULE_1"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "10 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "12 TEST_MODULE_2"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "13 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "17 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "17 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "14 add_"  [label="(1,)", style=solid];
+"6 bias" -> "15 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "18 add_1"  [label="(1,)", style=solid];
+"7 x" -> "8 TEST_MODULE_0"  [label="(1, 3, 3, 3)", style=solid];
+"8 TEST_MODULE_0" -> "10 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"9 TEST_MODULE_1" -> "10 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"10 conv2d" -> "11 TEST_MODULE_3"  [label="(1, 3, 3, 3)", style=solid];
+"11 TEST_MODULE_3" -> "13 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"11 TEST_MODULE_3" -> "14 add_"  [label="(1, 3, 3, 3)", style=solid];
+"12 TEST_MODULE_2" -> "13 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"13 conv2d_1" -> "15 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"14 add_" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"15 add__1" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"16 add" -> "17 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"17 conv2d_2" -> "18 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"18 add_1" -> "19 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/model_insertion_leaf.dot b/tests/torch/data/reference_graphs/fx/transformed/model_insertion_leaf.dot
index 236ed5d745c..d4eccf226a0 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/model_insertion_leaf.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/model_insertion_leaf.dot
@@ -1,46 +1,44 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 TEST_MODULE_0" [id=3, type=call_module];
-"4 TEST_MODULE_1" [id=4, type=call_module];
-"5 conv2d" [id=5, type=conv2d];
-"6 TEST_MODULE_3" [id=6, type=call_module];
-"7 _param_constant2" [id=7, type=get_attr];
-"8 _param_constant3" [id=8, type=get_attr];
-"9 TEST_MODULE_2" [id=9, type=call_module];
-"10 conv2d_1" [id=10, type=conv2d];
-"11 _tensor_constant0" [id=11, type=get_attr];
-"12 add_" [id=12, type=add_];
-"13 _tensor_constant0_1" [id=13, type=get_attr];
-"14 add__1" [id=14, type=add_];
-"15 add" [id=15, type=add];
-"16 _param_constant4" [id=16, type=get_attr];
-"17 _param_constant5" [id=17, type=get_attr];
-"18 conv2d_2" [id=18, type=conv2d];
-"19 _tensor_constant0_2" [id=19, type=get_attr];
-"20 add_1" [id=20, type=add];
-"21 output" [id=21, type=output];
-"0 arg0_1" -> "3 TEST_MODULE_0"  [label="(1, 3, 3, 3)", style=solid];
-"0 arg0_1" -> "5 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "4 TEST_MODULE_1"  [label="(3, 3, 1, 1)", style=solid];
-"1 _param_constant0" -> "5 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "5 conv2d"  [label="(3,)", style=solid];
-"5 conv2d" -> "6 TEST_MODULE_3"  [label="(1, 3, 3, 3)", style=solid];
-"5 conv2d" -> "10 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"5 conv2d" -> "12 add_"  [label="(1, 3, 3, 3)", style=solid];
-"7 _param_constant2" -> "9 TEST_MODULE_2"  [label="(3, 3, 1, 1)", style=solid];
-"7 _param_constant2" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"8 _param_constant3" -> "10 conv2d_1"  [label="(3,)", style=solid];
-"10 conv2d_1" -> "14 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"11 _tensor_constant0" -> "12 add_"  [label="(1,)", style=solid];
-"12 add_" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"13 _tensor_constant0_1" -> "14 add__1"  [label="(1,)", style=solid];
-"14 add__1" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"15 add" -> "18 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"16 _param_constant4" -> "18 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"17 _param_constant5" -> "18 conv2d_2"  [label="(3,)", style=solid];
-"18 conv2d_2" -> "20 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"19 _tensor_constant0_2" -> "20 add_1"  [label="(1,)", style=solid];
-"20 add_1" -> "21 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 TEST_MODULE_0" [id=8, type=call_module];
+"9 TEST_MODULE_1" [id=9, type=call_module];
+"10 conv2d" [id=10, type=conv2d];
+"11 TEST_MODULE_3" [id=11, type=call_module];
+"12 TEST_MODULE_2" [id=12, type=call_module];
+"13 conv2d_1" [id=13, type=conv2d];
+"14 add_" [id=14, type=add_];
+"15 add__1" [id=15, type=add_];
+"16 add" [id=16, type=add];
+"17 conv2d_2" [id=17, type=conv2d];
+"18 add_1" [id=18, type=add];
+"19 output_1" [id=19, type=output];
+"0 conv_a_weight" -> "9 TEST_MODULE_1"  [label="(3, 3, 1, 1)", style=solid];
+"0 conv_a_weight" -> "10 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "10 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "12 TEST_MODULE_2"  [label="(3, 3, 1, 1)", style=solid];
+"2 conv_b_weight" -> "13 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "13 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "17 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "17 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "14 add_"  [label="(1,)", style=solid];
+"6 bias" -> "15 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "18 add_1"  [label="(1,)", style=solid];
+"7 x" -> "8 TEST_MODULE_0"  [label="(1, 3, 3, 3)", style=solid];
+"7 x" -> "10 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d" -> "11 TEST_MODULE_3"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d" -> "13 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d" -> "14 add_"  [label="(1, 3, 3, 3)", style=solid];
+"13 conv2d_1" -> "15 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"14 add_" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"15 add__1" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"16 add" -> "17 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"17 conv2d_2" -> "18 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"18 add_1" -> "19 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/node_removal_ref.dot b/tests/torch/data/reference_graphs/fx/transformed/node_removal_ref.dot
index ff4dfc03eb1..db8acf6419c 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/node_removal_ref.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/node_removal_ref.dot
@@ -1,32 +1,30 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant2" [id=1, type=get_attr];
-"2 _param_constant3" [id=2, type=get_attr];
-"3 conv2d_1" [id=3, type=conv2d];
-"4 _tensor_constant0" [id=4, type=get_attr];
-"5 add_" [id=5, type=add_];
-"6 _tensor_constant0_1" [id=6, type=get_attr];
-"7 add__1" [id=7, type=add_];
-"8 add" [id=8, type=add];
-"9 _param_constant4" [id=9, type=get_attr];
-"10 _param_constant5" [id=10, type=get_attr];
-"11 conv2d_2" [id=11, type=conv2d];
-"12 _tensor_constant0_2" [id=12, type=get_attr];
-"13 add_1" [id=13, type=add];
-"14 output" [id=14, type=output];
-"0 arg0_1" -> "3 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"0 arg0_1" -> "5 add_"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant2" -> "3 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant3" -> "3 conv2d_1"  [label="(3,)", style=solid];
-"3 conv2d_1" -> "7 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"4 _tensor_constant0" -> "5 add_"  [label="(1,)", style=solid];
-"5 add_" -> "8 add"  [label="(1, 3, 3, 3)", style=solid];
-"6 _tensor_constant0_1" -> "7 add__1"  [label="(1,)", style=solid];
-"7 add__1" -> "8 add"  [label="(1, 3, 3, 3)", style=solid];
-"8 add" -> "11 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"9 _param_constant4" -> "11 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"10 _param_constant5" -> "11 conv2d_2"  [label="(3,)", style=solid];
-"11 conv2d_2" -> "13 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"12 _tensor_constant0_2" -> "13 add_1"  [label="(1,)", style=solid];
-"13 add_1" -> "14 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_b_weight" [id=0, type=get_attr];
+"1 conv_b_bias" [id=1, type=get_attr];
+"2 conv_c_weight" [id=2, type=get_attr];
+"3 conv_c_bias" [id=3, type=get_attr];
+"4 bias" [id=4, type=get_attr];
+"5 x" [id=5, type=input];
+"6 conv2d_1" [id=6, type=conv2d];
+"7 add_" [id=7, type=add_];
+"8 add__1" [id=8, type=add_];
+"9 add" [id=9, type=add];
+"10 conv2d_2" [id=10, type=conv2d];
+"11 add_1" [id=11, type=add];
+"12 output_1" [id=12, type=output];
+"0 conv_b_weight" -> "6 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_b_bias" -> "6 conv2d_1"  [label="(3,)", style=solid];
+"2 conv_c_weight" -> "10 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_c_bias" -> "10 conv2d_2"  [label="(3,)", style=solid];
+"4 bias" -> "7 add_"  [label="(1,)", style=solid];
+"4 bias" -> "8 add__1"  [label="(1,)", style=solid];
+"4 bias" -> "11 add_1"  [label="(1,)", style=solid];
+"5 x" -> "6 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"5 x" -> "7 add_"  [label="(1, 3, 3, 3)", style=solid];
+"6 conv2d_1" -> "8 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"7 add_" -> "9 add"  [label="(1, 3, 3, 3)", style=solid];
+"8 add__1" -> "9 add"  [label="(1, 3, 3, 3)", style=solid];
+"9 add" -> "10 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d_2" -> "11 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"11 add_1" -> "12 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_1_5_1.dot b/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_1_5_1.dot
index 737f0ba99af..8d641d46d41 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_1_5_1.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_1_5_1.dot
@@ -1,41 +1,39 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d" [id=3, type=conv2d];
-"4 _param_constant2" [id=4, type=get_attr];
-"5 _param_constant2_cloned" [id=5, type=clone];
-"6 _param_constant3" [id=6, type=get_attr];
-"7 conv2d_1" [id=7, type=conv2d];
-"8 _tensor_constant0" [id=8, type=get_attr];
-"9 add_" [id=9, type=add_];
-"10 _tensor_constant0_1" [id=10, type=get_attr];
-"11 add__1" [id=11, type=add_];
-"12 add" [id=12, type=add];
-"13 _param_constant4" [id=13, type=get_attr];
-"14 _param_constant5" [id=14, type=get_attr];
-"15 conv2d_2" [id=15, type=conv2d];
-"16 _tensor_constant0_2" [id=16, type=get_attr];
-"17 add_1" [id=17, type=add];
-"18 output" [id=18, type=output];
-"0 arg0_1" -> "3 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"3 conv2d" -> "7 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"3 conv2d" -> "9 add_"  [label="(1, 3, 3, 3)", style=solid];
-"4 _param_constant2" -> "5 _param_constant2_cloned"  [label="(3, 3, 1, 1)", style=solid];
-"4 _param_constant2" -> "7 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"5 _param_constant2_cloned" -> "18 output"  [label="(3, 3, 1, 1)", style=solid];
-"6 _param_constant3" -> "7 conv2d_1"  [label="(3,)", style=solid];
-"7 conv2d_1" -> "11 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"8 _tensor_constant0" -> "9 add_"  [label="(1,)", style=solid];
-"9 add_" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
-"10 _tensor_constant0_1" -> "11 add__1"  [label="(1,)", style=solid];
-"11 add__1" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
-"12 add" -> "15 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"13 _param_constant4" -> "15 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"14 _param_constant5" -> "15 conv2d_2"  [label="(3,)", style=solid];
-"15 conv2d_2" -> "17 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"16 _tensor_constant0_2" -> "17 add_1"  [label="(1,)", style=solid];
-"17 add_1" -> "18 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_weight_cloned" [id=3, type=clone];
+"4 conv_b_bias" [id=4, type=get_attr];
+"5 conv_c_weight" [id=5, type=get_attr];
+"6 conv_c_bias" [id=6, type=get_attr];
+"7 bias" [id=7, type=get_attr];
+"8 x" [id=8, type=input];
+"9 conv2d" [id=9, type=conv2d];
+"10 conv2d_1" [id=10, type=conv2d];
+"11 add_" [id=11, type=add_];
+"12 add__1" [id=12, type=add_];
+"13 add" [id=13, type=add];
+"14 conv2d_2" [id=14, type=conv2d];
+"15 add_1" [id=15, type=add];
+"16 output_1" [id=16, type=output];
+"0 conv_a_weight" -> "9 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "9 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "3 conv_b_weight_cloned"  [label="(3, 3, 1, 1)", style=solid];
+"2 conv_b_weight" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_weight_cloned" -> "16 output_1"  [label="(3, 3, 1, 1)", style=solid];
+"4 conv_b_bias" -> "10 conv2d_1"  [label="(3,)", style=solid];
+"5 conv_c_weight" -> "14 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"6 conv_c_bias" -> "14 conv2d_2"  [label="(3,)", style=solid];
+"7 bias" -> "11 add_"  [label="(1,)", style=solid];
+"7 bias" -> "12 add__1"  [label="(1,)", style=solid];
+"7 bias" -> "15 add_1"  [label="(1,)", style=solid];
+"8 x" -> "9 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"9 conv2d" -> "10 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"9 conv2d" -> "11 add_"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d_1" -> "12 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"11 add_" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
+"12 add__1" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
+"13 add" -> "14 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"14 conv2d_2" -> "15 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"15 add_1" -> "16 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_6_0.dot b/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_6_0.dot
index 79221ec29b3..3a35ddf39cf 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_6_0.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_6_0.dot
@@ -1,41 +1,39 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 arg0_1_cloned" [id=1, type=clone];
-"2 _param_constant0" [id=2, type=get_attr];
-"3 _param_constant1" [id=3, type=get_attr];
-"4 conv2d" [id=4, type=conv2d];
-"5 _param_constant2" [id=5, type=get_attr];
-"6 _param_constant3" [id=6, type=get_attr];
-"7 conv2d_1" [id=7, type=conv2d];
-"8 _tensor_constant0" [id=8, type=get_attr];
-"9 add_" [id=9, type=add_];
-"10 _tensor_constant0_1" [id=10, type=get_attr];
-"11 add__1" [id=11, type=add_];
-"12 add" [id=12, type=add];
-"13 _param_constant4" [id=13, type=get_attr];
-"14 _param_constant5" [id=14, type=get_attr];
-"15 conv2d_2" [id=15, type=conv2d];
-"16 _tensor_constant0_2" [id=16, type=get_attr];
-"17 add_1" [id=17, type=add];
-"18 output" [id=18, type=output];
-"0 arg0_1" -> "1 arg0_1_cloned"  [label="(1, 3, 3, 3)", style=solid];
-"0 arg0_1" -> "4 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 arg0_1_cloned" -> "18 output"  [label="(1, 3, 3, 3)", style=solid];
-"2 _param_constant0" -> "4 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"3 _param_constant1" -> "4 conv2d"  [label="(3,)", style=solid];
-"4 conv2d" -> "7 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"4 conv2d" -> "9 add_"  [label="(1, 3, 3, 3)", style=solid];
-"5 _param_constant2" -> "7 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"6 _param_constant3" -> "7 conv2d_1"  [label="(3,)", style=solid];
-"7 conv2d_1" -> "11 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"8 _tensor_constant0" -> "9 add_"  [label="(1,)", style=solid];
-"9 add_" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
-"10 _tensor_constant0_1" -> "11 add__1"  [label="(1,)", style=solid];
-"11 add__1" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
-"12 add" -> "15 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"13 _param_constant4" -> "15 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"14 _param_constant5" -> "15 conv2d_2"  [label="(3,)", style=solid];
-"15 conv2d_2" -> "17 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"16 _tensor_constant0_2" -> "17 add_1"  [label="(1,)", style=solid];
-"17 add_1" -> "18 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 x_cloned" [id=8, type=clone];
+"9 conv2d" [id=9, type=conv2d];
+"10 conv2d_1" [id=10, type=conv2d];
+"11 add_" [id=11, type=add_];
+"12 add__1" [id=12, type=add_];
+"13 add" [id=13, type=add];
+"14 conv2d_2" [id=14, type=conv2d];
+"15 add_1" [id=15, type=add];
+"16 output_1" [id=16, type=output];
+"0 conv_a_weight" -> "9 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "9 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "10 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "14 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "14 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "11 add_"  [label="(1,)", style=solid];
+"6 bias" -> "12 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "15 add_1"  [label="(1,)", style=solid];
+"7 x" -> "8 x_cloned"  [label="(1, 3, 3, 3)", style=solid];
+"7 x" -> "9 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"8 x_cloned" -> "16 output_1"  [label="(1, 3, 3, 3)", style=solid];
+"9 conv2d" -> "10 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"9 conv2d" -> "11 add_"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d_1" -> "12 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"11 add_" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
+"12 add__1" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
+"13 add" -> "14 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"14 conv2d_2" -> "15 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"15 add_1" -> "16 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_6_1.dot b/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_6_1.dot
index 2860aab488c..290db19728b 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_6_1.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_6_1.dot
@@ -1,41 +1,39 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant0_cloned" [id=2, type=clone];
-"3 _param_constant1" [id=3, type=get_attr];
-"4 conv2d" [id=4, type=conv2d];
-"5 _param_constant2" [id=5, type=get_attr];
-"6 _param_constant3" [id=6, type=get_attr];
-"7 conv2d_1" [id=7, type=conv2d];
-"8 _tensor_constant0" [id=8, type=get_attr];
-"9 add_" [id=9, type=add_];
-"10 _tensor_constant0_1" [id=10, type=get_attr];
-"11 add__1" [id=11, type=add_];
-"12 add" [id=12, type=add];
-"13 _param_constant4" [id=13, type=get_attr];
-"14 _param_constant5" [id=14, type=get_attr];
-"15 conv2d_2" [id=15, type=conv2d];
-"16 _tensor_constant0_2" [id=16, type=get_attr];
-"17 add_1" [id=17, type=add];
-"18 output" [id=18, type=output];
-"0 arg0_1" -> "4 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "2 _param_constant0_cloned"  [label="(3, 3, 1, 1)", style=solid];
-"1 _param_constant0" -> "4 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant0_cloned" -> "18 output"  [label="(3, 3, 1, 1)", style=solid];
-"3 _param_constant1" -> "4 conv2d"  [label="(3,)", style=solid];
-"4 conv2d" -> "7 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"4 conv2d" -> "9 add_"  [label="(1, 3, 3, 3)", style=solid];
-"5 _param_constant2" -> "7 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"6 _param_constant3" -> "7 conv2d_1"  [label="(3,)", style=solid];
-"7 conv2d_1" -> "11 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"8 _tensor_constant0" -> "9 add_"  [label="(1,)", style=solid];
-"9 add_" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
-"10 _tensor_constant0_1" -> "11 add__1"  [label="(1,)", style=solid];
-"11 add__1" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
-"12 add" -> "15 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"13 _param_constant4" -> "15 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"14 _param_constant5" -> "15 conv2d_2"  [label="(3,)", style=solid];
-"15 conv2d_2" -> "17 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"16 _tensor_constant0_2" -> "17 add_1"  [label="(1,)", style=solid];
-"17 add_1" -> "18 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_weight_cloned" [id=1, type=clone];
+"2 conv_a_bias" [id=2, type=get_attr];
+"3 conv_b_weight" [id=3, type=get_attr];
+"4 conv_b_bias" [id=4, type=get_attr];
+"5 conv_c_weight" [id=5, type=get_attr];
+"6 conv_c_bias" [id=6, type=get_attr];
+"7 bias" [id=7, type=get_attr];
+"8 x" [id=8, type=input];
+"9 conv2d" [id=9, type=conv2d];
+"10 conv2d_1" [id=10, type=conv2d];
+"11 add_" [id=11, type=add_];
+"12 add__1" [id=12, type=add_];
+"13 add" [id=13, type=add];
+"14 conv2d_2" [id=14, type=conv2d];
+"15 add_1" [id=15, type=add];
+"16 output_1" [id=16, type=output];
+"0 conv_a_weight" -> "1 conv_a_weight_cloned"  [label="(3, 3, 1, 1)", style=solid];
+"0 conv_a_weight" -> "9 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_weight_cloned" -> "16 output_1"  [label="(3, 3, 1, 1)", style=solid];
+"2 conv_a_bias" -> "9 conv2d"  [label="(3,)", style=solid];
+"3 conv_b_weight" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"4 conv_b_bias" -> "10 conv2d_1"  [label="(3,)", style=solid];
+"5 conv_c_weight" -> "14 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"6 conv_c_bias" -> "14 conv2d_2"  [label="(3,)", style=solid];
+"7 bias" -> "11 add_"  [label="(1,)", style=solid];
+"7 bias" -> "12 add__1"  [label="(1,)", style=solid];
+"7 bias" -> "15 add_1"  [label="(1,)", style=solid];
+"8 x" -> "9 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"9 conv2d" -> "10 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"9 conv2d" -> "11 add_"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d_1" -> "12 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"11 add_" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
+"12 add__1" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
+"13 add" -> "14 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"14 conv2d_2" -> "15 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"15 add_1" -> "16 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_7_None.dot b/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_7_None.dot
index f42688b109a..ff9313f1acc 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_7_None.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/output_insertion_conv2d_7_None.dot
@@ -1,41 +1,39 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d" [id=3, type=conv2d];
-"4 conv2d_cloned" [id=4, type=clone];
-"5 _param_constant2" [id=5, type=get_attr];
-"6 _param_constant3" [id=6, type=get_attr];
-"7 conv2d_1" [id=7, type=conv2d];
-"8 _tensor_constant0" [id=8, type=get_attr];
-"9 add_" [id=9, type=add_];
-"10 _tensor_constant0_1" [id=10, type=get_attr];
-"11 add__1" [id=11, type=add_];
-"12 add" [id=12, type=add];
-"13 _param_constant4" [id=13, type=get_attr];
-"14 _param_constant5" [id=14, type=get_attr];
-"15 conv2d_2" [id=15, type=conv2d];
-"16 _tensor_constant0_2" [id=16, type=get_attr];
-"17 add_1" [id=17, type=add];
-"18 output" [id=18, type=output];
-"0 arg0_1" -> "3 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"3 conv2d" -> "4 conv2d_cloned"  [label="(1, 3, 3, 3)", style=solid];
-"3 conv2d" -> "7 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"3 conv2d" -> "9 add_"  [label="(1, 3, 3, 3)", style=solid];
-"4 conv2d_cloned" -> "18 output"  [label="(1, 3, 3, 3)", style=solid];
-"5 _param_constant2" -> "7 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"6 _param_constant3" -> "7 conv2d_1"  [label="(3,)", style=solid];
-"7 conv2d_1" -> "11 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"8 _tensor_constant0" -> "9 add_"  [label="(1,)", style=solid];
-"9 add_" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
-"10 _tensor_constant0_1" -> "11 add__1"  [label="(1,)", style=solid];
-"11 add__1" -> "12 add"  [label="(1, 3, 3, 3)", style=solid];
-"12 add" -> "15 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"13 _param_constant4" -> "15 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"14 _param_constant5" -> "15 conv2d_2"  [label="(3,)", style=solid];
-"15 conv2d_2" -> "17 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"16 _tensor_constant0_2" -> "17 add_1"  [label="(1,)", style=solid];
-"17 add_1" -> "18 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 conv2d" [id=8, type=conv2d];
+"9 conv2d_cloned" [id=9, type=clone];
+"10 conv2d_1" [id=10, type=conv2d];
+"11 add_" [id=11, type=add_];
+"12 add__1" [id=12, type=add_];
+"13 add" [id=13, type=add];
+"14 conv2d_2" [id=14, type=conv2d];
+"15 add_1" [id=15, type=add];
+"16 output_1" [id=16, type=output];
+"0 conv_a_weight" -> "8 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "8 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "10 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "14 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "14 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "11 add_"  [label="(1,)", style=solid];
+"6 bias" -> "12 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "15 add_1"  [label="(1,)", style=solid];
+"7 x" -> "8 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d" -> "9 conv2d_cloned"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d" -> "10 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d" -> "11 add_"  [label="(1, 3, 3, 3)", style=solid];
+"9 conv2d_cloned" -> "16 output_1"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d_1" -> "12 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"11 add_" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
+"12 add__1" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
+"13 add" -> "14 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"14 conv2d_2" -> "15 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"15 add_1" -> "16 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_1_5_1_per_channel.dot b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_1_5_1_per_channel.dot
index b295abfc940..1843f556984 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_1_5_1_per_channel.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_1_5_1_per_channel.dot
@@ -1,48 +1,46 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d" [id=3, type=conv2d];
-"4 _param_constant2" [id=4, type=get_attr];
-"5 _param_constant3" [id=5, type=get_attr];
-"6 conv2d_1_scale_0" [id=6, type=get_attr];
-"7 conv2d_1_zero_point_0" [id=7, type=get_attr];
-"8 quantize_per_channel_default" [id=8, type=quantize_per_channel];
-"9 dequantize_per_channel_default" [id=9, type=dequantize_per_channel];
-"10 conv2d_1" [id=10, type=conv2d];
-"11 _tensor_constant0" [id=11, type=get_attr];
-"12 add_" [id=12, type=add_];
-"13 _tensor_constant0_1" [id=13, type=get_attr];
-"14 add__1" [id=14, type=add_];
-"15 add" [id=15, type=add];
-"16 _param_constant4" [id=16, type=get_attr];
-"17 _param_constant5" [id=17, type=get_attr];
-"18 conv2d_2" [id=18, type=conv2d];
-"19 _tensor_constant0_2" [id=19, type=get_attr];
-"20 add_1" [id=20, type=add];
-"21 output" [id=21, type=output];
-"0 arg0_1" -> "3 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"3 conv2d" -> "10 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"3 conv2d" -> "12 add_"  [label="(1, 3, 3, 3)", style=solid];
-"4 _param_constant2" -> "8 quantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
-"5 _param_constant3" -> "10 conv2d_1"  [label="(3,)", style=solid];
-"6 conv2d_1_scale_0" -> "8 quantize_per_channel_default"  [label="(1,)", style=solid];
-"6 conv2d_1_scale_0" -> "9 dequantize_per_channel_default"  [label="(1,)", style=solid];
-"7 conv2d_1_zero_point_0" -> "8 quantize_per_channel_default"  [label="(1,)", style=solid];
-"7 conv2d_1_zero_point_0" -> "9 dequantize_per_channel_default"  [label="(1,)", style=solid];
-"8 quantize_per_channel_default" -> "9 dequantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
-"9 dequantize_per_channel_default" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"10 conv2d_1" -> "14 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"11 _tensor_constant0" -> "12 add_"  [label="(1,)", style=solid];
-"12 add_" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"13 _tensor_constant0_1" -> "14 add__1"  [label="(1,)", style=solid];
-"14 add__1" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"15 add" -> "18 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"16 _param_constant4" -> "18 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"17 _param_constant5" -> "18 conv2d_2"  [label="(3,)", style=solid];
-"18 conv2d_2" -> "20 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"19 _tensor_constant0_2" -> "20 add_1"  [label="(1,)", style=solid];
-"20 add_1" -> "21 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 conv2d" [id=8, type=conv2d];
+"9 conv2d_1_scale_0" [id=9, type=get_attr];
+"10 conv2d_1_zero_point_0" [id=10, type=get_attr];
+"11 quantize_per_channel_default" [id=11, type=quantize_per_channel];
+"12 dequantize_per_channel_default" [id=12, type=dequantize_per_channel];
+"13 conv2d_1" [id=13, type=conv2d];
+"14 add_" [id=14, type=add_];
+"15 add__1" [id=15, type=add_];
+"16 add" [id=16, type=add];
+"17 conv2d_2" [id=17, type=conv2d];
+"18 add_1" [id=18, type=add];
+"19 output_1" [id=19, type=output];
+"0 conv_a_weight" -> "8 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "8 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "11 quantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "13 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "17 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "17 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "14 add_"  [label="(1,)", style=solid];
+"6 bias" -> "15 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "18 add_1"  [label="(1,)", style=solid];
+"7 x" -> "8 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d" -> "13 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d" -> "14 add_"  [label="(1, 3, 3, 3)", style=solid];
+"9 conv2d_1_scale_0" -> "11 quantize_per_channel_default"  [label="(1,)", style=solid];
+"9 conv2d_1_scale_0" -> "12 dequantize_per_channel_default"  [label="(1,)", style=solid];
+"10 conv2d_1_zero_point_0" -> "11 quantize_per_channel_default"  [label="(1,)", style=solid];
+"10 conv2d_1_zero_point_0" -> "12 dequantize_per_channel_default"  [label="(1,)", style=solid];
+"11 quantize_per_channel_default" -> "12 dequantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
+"12 dequantize_per_channel_default" -> "13 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"13 conv2d_1" -> "15 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"14 add_" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"15 add__1" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"16 add" -> "17 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"17 conv2d_2" -> "18 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"18 add_1" -> "19 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_1_5_1_per_tensor.dot b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_1_5_1_per_tensor.dot
index 895d90e44b8..1d8a5e629c0 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_1_5_1_per_tensor.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_1_5_1_per_tensor.dot
@@ -1,42 +1,40 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d" [id=3, type=conv2d];
-"4 _param_constant2" [id=4, type=get_attr];
-"5 _param_constant3" [id=5, type=get_attr];
-"6 quantize_per_tensor_default" [id=6, type=quantize_per_tensor];
-"7 dequantize_per_tensor_default" [id=7, type=dequantize_per_tensor];
-"8 conv2d_1" [id=8, type=conv2d];
-"9 _tensor_constant0" [id=9, type=get_attr];
-"10 add_" [id=10, type=add_];
-"11 _tensor_constant0_1" [id=11, type=get_attr];
-"12 add__1" [id=12, type=add_];
-"13 add" [id=13, type=add];
-"14 _param_constant4" [id=14, type=get_attr];
-"15 _param_constant5" [id=15, type=get_attr];
-"16 conv2d_2" [id=16, type=conv2d];
-"17 _tensor_constant0_2" [id=17, type=get_attr];
-"18 add_1" [id=18, type=add];
-"19 output" [id=19, type=output];
-"0 arg0_1" -> "3 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"3 conv2d" -> "8 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"3 conv2d" -> "10 add_"  [label="(1, 3, 3, 3)", style=solid];
-"4 _param_constant2" -> "6 quantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
-"5 _param_constant3" -> "8 conv2d_1"  [label="(3,)", style=solid];
-"6 quantize_per_tensor_default" -> "7 dequantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
-"7 dequantize_per_tensor_default" -> "8 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"8 conv2d_1" -> "12 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"9 _tensor_constant0" -> "10 add_"  [label="(1,)", style=solid];
-"10 add_" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
-"11 _tensor_constant0_1" -> "12 add__1"  [label="(1,)", style=solid];
-"12 add__1" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
-"13 add" -> "16 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"14 _param_constant4" -> "16 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"15 _param_constant5" -> "16 conv2d_2"  [label="(3,)", style=solid];
-"16 conv2d_2" -> "18 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"17 _tensor_constant0_2" -> "18 add_1"  [label="(1,)", style=solid];
-"18 add_1" -> "19 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 conv2d" [id=8, type=conv2d];
+"9 quantize_per_tensor_default" [id=9, type=quantize_per_tensor];
+"10 dequantize_per_tensor_default" [id=10, type=dequantize_per_tensor];
+"11 conv2d_1" [id=11, type=conv2d];
+"12 add_" [id=12, type=add_];
+"13 add__1" [id=13, type=add_];
+"14 add" [id=14, type=add];
+"15 conv2d_2" [id=15, type=conv2d];
+"16 add_1" [id=16, type=add];
+"17 output_1" [id=17, type=output];
+"0 conv_a_weight" -> "8 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "8 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "9 quantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "11 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "15 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "15 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "12 add_"  [label="(1,)", style=solid];
+"6 bias" -> "13 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "16 add_1"  [label="(1,)", style=solid];
+"7 x" -> "8 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d" -> "11 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d" -> "12 add_"  [label="(1, 3, 3, 3)", style=solid];
+"9 quantize_per_tensor_default" -> "10 dequantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
+"10 dequantize_per_tensor_default" -> "11 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"11 conv2d_1" -> "13 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"12 add_" -> "14 add"  [label="(1, 3, 3, 3)", style=solid];
+"13 add__1" -> "14 add"  [label="(1, 3, 3, 3)", style=solid];
+"14 add" -> "15 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"15 conv2d_2" -> "16 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"16 add_1" -> "17 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_0_per_channel.dot b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_0_per_channel.dot
index 91e9f02c918..50e8261f594 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_0_per_channel.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_0_per_channel.dot
@@ -1,48 +1,46 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d_scale_0" [id=3, type=get_attr];
-"4 conv2d_zero_point_0" [id=4, type=get_attr];
-"5 quantize_per_channel_default" [id=5, type=quantize_per_channel];
-"6 dequantize_per_channel_default" [id=6, type=dequantize_per_channel];
-"7 conv2d" [id=7, type=conv2d];
-"8 _param_constant2" [id=8, type=get_attr];
-"9 _param_constant3" [id=9, type=get_attr];
-"10 conv2d_1" [id=10, type=conv2d];
-"11 _tensor_constant0" [id=11, type=get_attr];
-"12 add_" [id=12, type=add_];
-"13 _tensor_constant0_1" [id=13, type=get_attr];
-"14 add__1" [id=14, type=add_];
-"15 add" [id=15, type=add];
-"16 _param_constant4" [id=16, type=get_attr];
-"17 _param_constant5" [id=17, type=get_attr];
-"18 conv2d_2" [id=18, type=conv2d];
-"19 _tensor_constant0_2" [id=19, type=get_attr];
-"20 add_1" [id=20, type=add];
-"21 output" [id=21, type=output];
-"0 arg0_1" -> "5 quantize_per_channel_default"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "7 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "7 conv2d"  [label="(3,)", style=solid];
-"3 conv2d_scale_0" -> "5 quantize_per_channel_default"  [label="(1,)", style=solid];
-"3 conv2d_scale_0" -> "6 dequantize_per_channel_default"  [label="(1,)", style=solid];
-"4 conv2d_zero_point_0" -> "5 quantize_per_channel_default"  [label="(1,)", style=solid];
-"4 conv2d_zero_point_0" -> "6 dequantize_per_channel_default"  [label="(1,)", style=solid];
-"5 quantize_per_channel_default" -> "6 dequantize_per_channel_default"  [label="(1, 3, 3, 3)", style=solid];
-"6 dequantize_per_channel_default" -> "7 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"7 conv2d" -> "10 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"7 conv2d" -> "12 add_"  [label="(1, 3, 3, 3)", style=solid];
-"8 _param_constant2" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"9 _param_constant3" -> "10 conv2d_1"  [label="(3,)", style=solid];
-"10 conv2d_1" -> "14 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"11 _tensor_constant0" -> "12 add_"  [label="(1,)", style=solid];
-"12 add_" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"13 _tensor_constant0_1" -> "14 add__1"  [label="(1,)", style=solid];
-"14 add__1" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"15 add" -> "18 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"16 _param_constant4" -> "18 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"17 _param_constant5" -> "18 conv2d_2"  [label="(3,)", style=solid];
-"18 conv2d_2" -> "20 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"19 _tensor_constant0_2" -> "20 add_1"  [label="(1,)", style=solid];
-"20 add_1" -> "21 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 conv2d_scale_0" [id=8, type=get_attr];
+"9 conv2d_zero_point_0" [id=9, type=get_attr];
+"10 quantize_per_channel_default" [id=10, type=quantize_per_channel];
+"11 dequantize_per_channel_default" [id=11, type=dequantize_per_channel];
+"12 conv2d" [id=12, type=conv2d];
+"13 conv2d_1" [id=13, type=conv2d];
+"14 add_" [id=14, type=add_];
+"15 add__1" [id=15, type=add_];
+"16 add" [id=16, type=add];
+"17 conv2d_2" [id=17, type=conv2d];
+"18 add_1" [id=18, type=add];
+"19 output_1" [id=19, type=output];
+"0 conv_a_weight" -> "12 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "12 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "13 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "13 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "17 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "17 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "14 add_"  [label="(1,)", style=solid];
+"6 bias" -> "15 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "18 add_1"  [label="(1,)", style=solid];
+"7 x" -> "10 quantize_per_channel_default"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d_scale_0" -> "10 quantize_per_channel_default"  [label="(1,)", style=solid];
+"8 conv2d_scale_0" -> "11 dequantize_per_channel_default"  [label="(1,)", style=solid];
+"9 conv2d_zero_point_0" -> "10 quantize_per_channel_default"  [label="(1,)", style=solid];
+"9 conv2d_zero_point_0" -> "11 dequantize_per_channel_default"  [label="(1,)", style=solid];
+"10 quantize_per_channel_default" -> "11 dequantize_per_channel_default"  [label="(1, 3, 3, 3)", style=solid];
+"11 dequantize_per_channel_default" -> "12 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"12 conv2d" -> "13 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"12 conv2d" -> "14 add_"  [label="(1, 3, 3, 3)", style=solid];
+"13 conv2d_1" -> "15 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"14 add_" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"15 add__1" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"16 add" -> "17 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"17 conv2d_2" -> "18 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"18 add_1" -> "19 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_0_per_tensor.dot b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_0_per_tensor.dot
index eec47cacece..01bfb0da32b 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_0_per_tensor.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_0_per_tensor.dot
@@ -1,42 +1,40 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 quantize_per_tensor_default" [id=3, type=quantize_per_tensor];
-"4 dequantize_per_tensor_default" [id=4, type=dequantize_per_tensor];
-"5 conv2d" [id=5, type=conv2d];
-"6 _param_constant2" [id=6, type=get_attr];
-"7 _param_constant3" [id=7, type=get_attr];
-"8 conv2d_1" [id=8, type=conv2d];
-"9 _tensor_constant0" [id=9, type=get_attr];
-"10 add_" [id=10, type=add_];
-"11 _tensor_constant0_1" [id=11, type=get_attr];
-"12 add__1" [id=12, type=add_];
-"13 add" [id=13, type=add];
-"14 _param_constant4" [id=14, type=get_attr];
-"15 _param_constant5" [id=15, type=get_attr];
-"16 conv2d_2" [id=16, type=conv2d];
-"17 _tensor_constant0_2" [id=17, type=get_attr];
-"18 add_1" [id=18, type=add];
-"19 output" [id=19, type=output];
-"0 arg0_1" -> "3 quantize_per_tensor_default"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "5 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "5 conv2d"  [label="(3,)", style=solid];
-"3 quantize_per_tensor_default" -> "4 dequantize_per_tensor_default"  [label="(1, 3, 3, 3)", style=solid];
-"4 dequantize_per_tensor_default" -> "5 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"5 conv2d" -> "8 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"5 conv2d" -> "10 add_"  [label="(1, 3, 3, 3)", style=solid];
-"6 _param_constant2" -> "8 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"7 _param_constant3" -> "8 conv2d_1"  [label="(3,)", style=solid];
-"8 conv2d_1" -> "12 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"9 _tensor_constant0" -> "10 add_"  [label="(1,)", style=solid];
-"10 add_" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
-"11 _tensor_constant0_1" -> "12 add__1"  [label="(1,)", style=solid];
-"12 add__1" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
-"13 add" -> "16 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"14 _param_constant4" -> "16 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"15 _param_constant5" -> "16 conv2d_2"  [label="(3,)", style=solid];
-"16 conv2d_2" -> "18 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"17 _tensor_constant0_2" -> "18 add_1"  [label="(1,)", style=solid];
-"18 add_1" -> "19 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 quantize_per_tensor_default" [id=8, type=quantize_per_tensor];
+"9 dequantize_per_tensor_default" [id=9, type=dequantize_per_tensor];
+"10 conv2d" [id=10, type=conv2d];
+"11 conv2d_1" [id=11, type=conv2d];
+"12 add_" [id=12, type=add_];
+"13 add__1" [id=13, type=add_];
+"14 add" [id=14, type=add];
+"15 conv2d_2" [id=15, type=conv2d];
+"16 add_1" [id=16, type=add];
+"17 output_1" [id=17, type=output];
+"0 conv_a_weight" -> "10 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "10 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "11 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "11 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "15 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "15 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "12 add_"  [label="(1,)", style=solid];
+"6 bias" -> "13 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "16 add_1"  [label="(1,)", style=solid];
+"7 x" -> "8 quantize_per_tensor_default"  [label="(1, 3, 3, 3)", style=solid];
+"8 quantize_per_tensor_default" -> "9 dequantize_per_tensor_default"  [label="(1, 3, 3, 3)", style=solid];
+"9 dequantize_per_tensor_default" -> "10 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d" -> "11 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d" -> "12 add_"  [label="(1, 3, 3, 3)", style=solid];
+"11 conv2d_1" -> "13 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"12 add_" -> "14 add"  [label="(1, 3, 3, 3)", style=solid];
+"13 add__1" -> "14 add"  [label="(1, 3, 3, 3)", style=solid];
+"14 add" -> "15 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"15 conv2d_2" -> "16 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"16 add_1" -> "17 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_1_per_channel.dot b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_1_per_channel.dot
index b6138a0a70b..1c8808cc8c6 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_1_per_channel.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_1_per_channel.dot
@@ -1,48 +1,46 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d_scale_0" [id=3, type=get_attr];
-"4 conv2d_zero_point_0" [id=4, type=get_attr];
-"5 quantize_per_channel_default" [id=5, type=quantize_per_channel];
-"6 dequantize_per_channel_default" [id=6, type=dequantize_per_channel];
-"7 conv2d" [id=7, type=conv2d];
-"8 _param_constant2" [id=8, type=get_attr];
-"9 _param_constant3" [id=9, type=get_attr];
-"10 conv2d_1" [id=10, type=conv2d];
-"11 _tensor_constant0" [id=11, type=get_attr];
-"12 add_" [id=12, type=add_];
-"13 _tensor_constant0_1" [id=13, type=get_attr];
-"14 add__1" [id=14, type=add_];
-"15 add" [id=15, type=add];
-"16 _param_constant4" [id=16, type=get_attr];
-"17 _param_constant5" [id=17, type=get_attr];
-"18 conv2d_2" [id=18, type=conv2d];
-"19 _tensor_constant0_2" [id=19, type=get_attr];
-"20 add_1" [id=20, type=add];
-"21 output" [id=21, type=output];
-"0 arg0_1" -> "7 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "5 quantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "7 conv2d"  [label="(3,)", style=solid];
-"3 conv2d_scale_0" -> "5 quantize_per_channel_default"  [label="(1,)", style=solid];
-"3 conv2d_scale_0" -> "6 dequantize_per_channel_default"  [label="(1,)", style=solid];
-"4 conv2d_zero_point_0" -> "5 quantize_per_channel_default"  [label="(1,)", style=solid];
-"4 conv2d_zero_point_0" -> "6 dequantize_per_channel_default"  [label="(1,)", style=solid];
-"5 quantize_per_channel_default" -> "6 dequantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
-"6 dequantize_per_channel_default" -> "7 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"7 conv2d" -> "10 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"7 conv2d" -> "12 add_"  [label="(1, 3, 3, 3)", style=solid];
-"8 _param_constant2" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"9 _param_constant3" -> "10 conv2d_1"  [label="(3,)", style=solid];
-"10 conv2d_1" -> "14 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"11 _tensor_constant0" -> "12 add_"  [label="(1,)", style=solid];
-"12 add_" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"13 _tensor_constant0_1" -> "14 add__1"  [label="(1,)", style=solid];
-"14 add__1" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"15 add" -> "18 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"16 _param_constant4" -> "18 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"17 _param_constant5" -> "18 conv2d_2"  [label="(3,)", style=solid];
-"18 conv2d_2" -> "20 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"19 _tensor_constant0_2" -> "20 add_1"  [label="(1,)", style=solid];
-"20 add_1" -> "21 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 conv2d_scale_0" [id=8, type=get_attr];
+"9 conv2d_zero_point_0" [id=9, type=get_attr];
+"10 quantize_per_channel_default" [id=10, type=quantize_per_channel];
+"11 dequantize_per_channel_default" [id=11, type=dequantize_per_channel];
+"12 conv2d" [id=12, type=conv2d];
+"13 conv2d_1" [id=13, type=conv2d];
+"14 add_" [id=14, type=add_];
+"15 add__1" [id=15, type=add_];
+"16 add" [id=16, type=add];
+"17 conv2d_2" [id=17, type=conv2d];
+"18 add_1" [id=18, type=add];
+"19 output_1" [id=19, type=output];
+"0 conv_a_weight" -> "10 quantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "12 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "13 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "13 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "17 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "17 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "14 add_"  [label="(1,)", style=solid];
+"6 bias" -> "15 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "18 add_1"  [label="(1,)", style=solid];
+"7 x" -> "12 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d_scale_0" -> "10 quantize_per_channel_default"  [label="(1,)", style=solid];
+"8 conv2d_scale_0" -> "11 dequantize_per_channel_default"  [label="(1,)", style=solid];
+"9 conv2d_zero_point_0" -> "10 quantize_per_channel_default"  [label="(1,)", style=solid];
+"9 conv2d_zero_point_0" -> "11 dequantize_per_channel_default"  [label="(1,)", style=solid];
+"10 quantize_per_channel_default" -> "11 dequantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
+"11 dequantize_per_channel_default" -> "12 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"12 conv2d" -> "13 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"12 conv2d" -> "14 add_"  [label="(1, 3, 3, 3)", style=solid];
+"13 conv2d_1" -> "15 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"14 add_" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"15 add__1" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"16 add" -> "17 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"17 conv2d_2" -> "18 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"18 add_1" -> "19 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_1_per_tensor.dot b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_1_per_tensor.dot
index 62478def9d7..62ab139712c 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_1_per_tensor.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_6_1_per_tensor.dot
@@ -1,42 +1,40 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 quantize_per_tensor_default" [id=3, type=quantize_per_tensor];
-"4 dequantize_per_tensor_default" [id=4, type=dequantize_per_tensor];
-"5 conv2d" [id=5, type=conv2d];
-"6 _param_constant2" [id=6, type=get_attr];
-"7 _param_constant3" [id=7, type=get_attr];
-"8 conv2d_1" [id=8, type=conv2d];
-"9 _tensor_constant0" [id=9, type=get_attr];
-"10 add_" [id=10, type=add_];
-"11 _tensor_constant0_1" [id=11, type=get_attr];
-"12 add__1" [id=12, type=add_];
-"13 add" [id=13, type=add];
-"14 _param_constant4" [id=14, type=get_attr];
-"15 _param_constant5" [id=15, type=get_attr];
-"16 conv2d_2" [id=16, type=conv2d];
-"17 _tensor_constant0_2" [id=17, type=get_attr];
-"18 add_1" [id=18, type=add];
-"19 output" [id=19, type=output];
-"0 arg0_1" -> "5 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "3 quantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "5 conv2d"  [label="(3,)", style=solid];
-"3 quantize_per_tensor_default" -> "4 dequantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
-"4 dequantize_per_tensor_default" -> "5 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"5 conv2d" -> "8 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"5 conv2d" -> "10 add_"  [label="(1, 3, 3, 3)", style=solid];
-"6 _param_constant2" -> "8 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"7 _param_constant3" -> "8 conv2d_1"  [label="(3,)", style=solid];
-"8 conv2d_1" -> "12 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"9 _tensor_constant0" -> "10 add_"  [label="(1,)", style=solid];
-"10 add_" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
-"11 _tensor_constant0_1" -> "12 add__1"  [label="(1,)", style=solid];
-"12 add__1" -> "13 add"  [label="(1, 3, 3, 3)", style=solid];
-"13 add" -> "16 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"14 _param_constant4" -> "16 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"15 _param_constant5" -> "16 conv2d_2"  [label="(3,)", style=solid];
-"16 conv2d_2" -> "18 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"17 _tensor_constant0_2" -> "18 add_1"  [label="(1,)", style=solid];
-"18 add_1" -> "19 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 quantize_per_tensor_default" [id=8, type=quantize_per_tensor];
+"9 dequantize_per_tensor_default" [id=9, type=dequantize_per_tensor];
+"10 conv2d" [id=10, type=conv2d];
+"11 conv2d_1" [id=11, type=conv2d];
+"12 add_" [id=12, type=add_];
+"13 add__1" [id=13, type=add_];
+"14 add" [id=14, type=add];
+"15 conv2d_2" [id=15, type=conv2d];
+"16 add_1" [id=16, type=add];
+"17 output_1" [id=17, type=output];
+"0 conv_a_weight" -> "8 quantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "10 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "11 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "11 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "15 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "15 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "12 add_"  [label="(1,)", style=solid];
+"6 bias" -> "13 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "16 add_1"  [label="(1,)", style=solid];
+"7 x" -> "10 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"8 quantize_per_tensor_default" -> "9 dequantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
+"9 dequantize_per_tensor_default" -> "10 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"10 conv2d" -> "11 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d" -> "12 add_"  [label="(1, 3, 3, 3)", style=solid];
+"11 conv2d_1" -> "13 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"12 add_" -> "14 add"  [label="(1, 3, 3, 3)", style=solid];
+"13 add__1" -> "14 add"  [label="(1, 3, 3, 3)", style=solid];
+"14 add" -> "15 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"15 conv2d_2" -> "16 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"16 add_1" -> "17 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_7_None_per_channel.dot b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_7_None_per_channel.dot
index af758b9085b..5c9e516a456 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_7_None_per_channel.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_7_None_per_channel.dot
@@ -1,52 +1,50 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d_scale_0" [id=3, type=get_attr];
-"4 conv2d_zero_point_0" [id=4, type=get_attr];
-"5 conv2d" [id=5, type=conv2d];
-"6 quantize_per_channel_default" [id=6, type=quantize_per_channel];
-"7 dequantize_per_channel_default_1" [id=7, type=dequantize_per_channel];
-"8 dequantize_per_channel_default" [id=8, type=dequantize_per_channel];
-"9 _param_constant2" [id=9, type=get_attr];
-"10 _param_constant3" [id=10, type=get_attr];
-"11 conv2d_1" [id=11, type=conv2d];
-"12 _tensor_constant0" [id=12, type=get_attr];
-"13 add_" [id=13, type=add_];
-"14 _tensor_constant0_1" [id=14, type=get_attr];
-"15 add__1" [id=15, type=add_];
-"16 add" [id=16, type=add];
-"17 _param_constant4" [id=17, type=get_attr];
-"18 _param_constant5" [id=18, type=get_attr];
-"19 conv2d_2" [id=19, type=conv2d];
-"20 _tensor_constant0_2" [id=20, type=get_attr];
-"21 add_1" [id=21, type=add];
-"22 output" [id=22, type=output];
-"0 arg0_1" -> "5 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "5 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "5 conv2d"  [label="(3,)", style=solid];
-"3 conv2d_scale_0" -> "6 quantize_per_channel_default"  [label="(1,)", style=solid];
-"3 conv2d_scale_0" -> "7 dequantize_per_channel_default_1"  [label="(1,)", style=solid];
-"3 conv2d_scale_0" -> "8 dequantize_per_channel_default"  [label="(1,)", style=solid];
-"4 conv2d_zero_point_0" -> "6 quantize_per_channel_default"  [label="(1,)", style=solid];
-"4 conv2d_zero_point_0" -> "7 dequantize_per_channel_default_1"  [label="(1,)", style=solid];
-"4 conv2d_zero_point_0" -> "8 dequantize_per_channel_default"  [label="(1,)", style=solid];
-"5 conv2d" -> "6 quantize_per_channel_default"  [label="(1, 3, 3, 3)", style=solid];
-"6 quantize_per_channel_default" -> "7 dequantize_per_channel_default_1"  [label="(1, 3, 3, 3)", style=solid];
-"6 quantize_per_channel_default" -> "8 dequantize_per_channel_default"  [label="(1, 3, 3, 3)", style=solid];
-"7 dequantize_per_channel_default_1" -> "13 add_"  [label="(1, 3, 3, 3)", style=solid];
-"8 dequantize_per_channel_default" -> "11 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"9 _param_constant2" -> "11 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"10 _param_constant3" -> "11 conv2d_1"  [label="(3,)", style=solid];
-"11 conv2d_1" -> "15 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"12 _tensor_constant0" -> "13 add_"  [label="(1,)", style=solid];
-"13 add_" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
-"14 _tensor_constant0_1" -> "15 add__1"  [label="(1,)", style=solid];
-"15 add__1" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
-"16 add" -> "19 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"17 _param_constant4" -> "19 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"18 _param_constant5" -> "19 conv2d_2"  [label="(3,)", style=solid];
-"19 conv2d_2" -> "21 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"20 _tensor_constant0_2" -> "21 add_1"  [label="(1,)", style=solid];
-"21 add_1" -> "22 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 conv2d_scale_0" [id=8, type=get_attr];
+"9 conv2d_zero_point_0" [id=9, type=get_attr];
+"10 conv2d" [id=10, type=conv2d];
+"11 quantize_per_channel_default" [id=11, type=quantize_per_channel];
+"12 dequantize_per_channel_default_1" [id=12, type=dequantize_per_channel];
+"13 dequantize_per_channel_default" [id=13, type=dequantize_per_channel];
+"14 conv2d_1" [id=14, type=conv2d];
+"15 add_" [id=15, type=add_];
+"16 add__1" [id=16, type=add_];
+"17 add" [id=17, type=add];
+"18 conv2d_2" [id=18, type=conv2d];
+"19 add_1" [id=19, type=add];
+"20 output_1" [id=20, type=output];
+"0 conv_a_weight" -> "10 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "10 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "14 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "14 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "18 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "18 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "15 add_"  [label="(1,)", style=solid];
+"6 bias" -> "16 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "19 add_1"  [label="(1,)", style=solid];
+"7 x" -> "10 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d_scale_0" -> "11 quantize_per_channel_default"  [label="(1,)", style=solid];
+"8 conv2d_scale_0" -> "12 dequantize_per_channel_default_1"  [label="(1,)", style=solid];
+"8 conv2d_scale_0" -> "13 dequantize_per_channel_default"  [label="(1,)", style=solid];
+"9 conv2d_zero_point_0" -> "11 quantize_per_channel_default"  [label="(1,)", style=solid];
+"9 conv2d_zero_point_0" -> "12 dequantize_per_channel_default_1"  [label="(1,)", style=solid];
+"9 conv2d_zero_point_0" -> "13 dequantize_per_channel_default"  [label="(1,)", style=solid];
+"10 conv2d" -> "11 quantize_per_channel_default"  [label="(1, 3, 3, 3)", style=solid];
+"11 quantize_per_channel_default" -> "12 dequantize_per_channel_default_1"  [label="(1, 3, 3, 3)", style=solid];
+"11 quantize_per_channel_default" -> "13 dequantize_per_channel_default"  [label="(1, 3, 3, 3)", style=solid];
+"12 dequantize_per_channel_default_1" -> "15 add_"  [label="(1, 3, 3, 3)", style=solid];
+"13 dequantize_per_channel_default" -> "14 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"14 conv2d_1" -> "16 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"15 add_" -> "17 add"  [label="(1, 3, 3, 3)", style=solid];
+"16 add__1" -> "17 add"  [label="(1, 3, 3, 3)", style=solid];
+"17 add" -> "18 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"18 conv2d_2" -> "19 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"19 add_1" -> "20 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_7_None_per_tensor.dot b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_7_None_per_tensor.dot
index 7d4f98cd99b..b0d96c62313 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_7_None_per_tensor.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/qdq_insert_conv2d_7_None_per_tensor.dot
@@ -1,44 +1,42 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d" [id=3, type=conv2d];
-"4 quantize_per_tensor_default" [id=4, type=quantize_per_tensor];
-"5 dequantize_per_tensor_default_1" [id=5, type=dequantize_per_tensor];
-"6 dequantize_per_tensor_default" [id=6, type=dequantize_per_tensor];
-"7 _param_constant2" [id=7, type=get_attr];
-"8 _param_constant3" [id=8, type=get_attr];
-"9 conv2d_1" [id=9, type=conv2d];
-"10 _tensor_constant0" [id=10, type=get_attr];
-"11 add_" [id=11, type=add_];
-"12 _tensor_constant0_1" [id=12, type=get_attr];
-"13 add__1" [id=13, type=add_];
-"14 add" [id=14, type=add];
-"15 _param_constant4" [id=15, type=get_attr];
-"16 _param_constant5" [id=16, type=get_attr];
-"17 conv2d_2" [id=17, type=conv2d];
-"18 _tensor_constant0_2" [id=18, type=get_attr];
-"19 add_1" [id=19, type=add];
-"20 output" [id=20, type=output];
-"0 arg0_1" -> "3 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"3 conv2d" -> "4 quantize_per_tensor_default"  [label="(1, 3, 3, 3)", style=solid];
-"4 quantize_per_tensor_default" -> "5 dequantize_per_tensor_default_1"  [label="(1, 3, 3, 3)", style=solid];
-"4 quantize_per_tensor_default" -> "6 dequantize_per_tensor_default"  [label="(1, 3, 3, 3)", style=solid];
-"5 dequantize_per_tensor_default_1" -> "11 add_"  [label="(1, 3, 3, 3)", style=solid];
-"6 dequantize_per_tensor_default" -> "9 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"7 _param_constant2" -> "9 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"8 _param_constant3" -> "9 conv2d_1"  [label="(3,)", style=solid];
-"9 conv2d_1" -> "13 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"10 _tensor_constant0" -> "11 add_"  [label="(1,)", style=solid];
-"11 add_" -> "14 add"  [label="(1, 3, 3, 3)", style=solid];
-"12 _tensor_constant0_1" -> "13 add__1"  [label="(1,)", style=solid];
-"13 add__1" -> "14 add"  [label="(1, 3, 3, 3)", style=solid];
-"14 add" -> "17 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"15 _param_constant4" -> "17 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"16 _param_constant5" -> "17 conv2d_2"  [label="(3,)", style=solid];
-"17 conv2d_2" -> "19 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"18 _tensor_constant0_2" -> "19 add_1"  [label="(1,)", style=solid];
-"19 add_1" -> "20 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 conv2d" [id=8, type=conv2d];
+"9 quantize_per_tensor_default" [id=9, type=quantize_per_tensor];
+"10 dequantize_per_tensor_default_1" [id=10, type=dequantize_per_tensor];
+"11 dequantize_per_tensor_default" [id=11, type=dequantize_per_tensor];
+"12 conv2d_1" [id=12, type=conv2d];
+"13 add_" [id=13, type=add_];
+"14 add__1" [id=14, type=add_];
+"15 add" [id=15, type=add];
+"16 conv2d_2" [id=16, type=conv2d];
+"17 add_1" [id=17, type=add];
+"18 output_1" [id=18, type=output];
+"0 conv_a_weight" -> "8 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "8 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "12 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "12 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "16 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "16 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "13 add_"  [label="(1,)", style=solid];
+"6 bias" -> "14 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "17 add_1"  [label="(1,)", style=solid];
+"7 x" -> "8 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d" -> "9 quantize_per_tensor_default"  [label="(1, 3, 3, 3)", style=solid];
+"9 quantize_per_tensor_default" -> "10 dequantize_per_tensor_default_1"  [label="(1, 3, 3, 3)", style=solid];
+"9 quantize_per_tensor_default" -> "11 dequantize_per_tensor_default"  [label="(1, 3, 3, 3)", style=solid];
+"10 dequantize_per_tensor_default_1" -> "13 add_"  [label="(1, 3, 3, 3)", style=solid];
+"11 dequantize_per_tensor_default" -> "12 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"12 conv2d_1" -> "14 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"13 add_" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
+"14 add__1" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
+"15 add" -> "16 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"16 conv2d_2" -> "17 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"17 add_1" -> "18 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/qdq_shared_insert_weights_per_channel.dot b/tests/torch/data/reference_graphs/fx/transformed/qdq_shared_insert_weights_per_channel.dot
index b2333db9ef0..38f5b8a0c91 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/qdq_shared_insert_weights_per_channel.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/qdq_shared_insert_weights_per_channel.dot
@@ -1,58 +1,56 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d_scale_0" [id=3, type=get_attr];
-"4 conv2d_zero_point_0" [id=4, type=get_attr];
-"5 quantize_per_channel_default" [id=5, type=quantize_per_channel];
-"6 dequantize_per_channel_default" [id=6, type=dequantize_per_channel];
-"7 conv2d" [id=7, type=conv2d];
-"8 _param_constant2" [id=8, type=get_attr];
-"9 _param_constant3" [id=9, type=get_attr];
-"10 conv2d_1_scale_0" [id=10, type=get_attr];
-"11 conv2d_1_zero_point_0" [id=11, type=get_attr];
-"12 quantize_per_channel_default_1" [id=12, type=quantize_per_channel];
-"13 dequantize_per_channel_default_1" [id=13, type=dequantize_per_channel];
-"14 conv2d_1" [id=14, type=conv2d];
-"15 _tensor_constant0" [id=15, type=get_attr];
-"16 add_" [id=16, type=add_];
-"17 _tensor_constant0_1" [id=17, type=get_attr];
-"18 add__1" [id=18, type=add_];
-"19 add" [id=19, type=add];
-"20 _param_constant4" [id=20, type=get_attr];
-"21 _param_constant5" [id=21, type=get_attr];
-"22 conv2d_2" [id=22, type=conv2d];
-"23 _tensor_constant0_2" [id=23, type=get_attr];
-"24 add_1" [id=24, type=add];
-"25 output" [id=25, type=output];
-"0 arg0_1" -> "7 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "5 quantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "7 conv2d"  [label="(3,)", style=solid];
-"3 conv2d_scale_0" -> "5 quantize_per_channel_default"  [label="(1,)", style=solid];
-"3 conv2d_scale_0" -> "6 dequantize_per_channel_default"  [label="(1,)", style=solid];
-"4 conv2d_zero_point_0" -> "5 quantize_per_channel_default"  [label="(1,)", style=solid];
-"4 conv2d_zero_point_0" -> "6 dequantize_per_channel_default"  [label="(1,)", style=solid];
-"5 quantize_per_channel_default" -> "6 dequantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
-"6 dequantize_per_channel_default" -> "7 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"7 conv2d" -> "14 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"7 conv2d" -> "16 add_"  [label="(1, 3, 3, 3)", style=solid];
-"8 _param_constant2" -> "12 quantize_per_channel_default_1"  [label="(3, 3, 1, 1)", style=solid];
-"9 _param_constant3" -> "14 conv2d_1"  [label="(3,)", style=solid];
-"10 conv2d_1_scale_0" -> "12 quantize_per_channel_default_1"  [label="(1,)", style=solid];
-"10 conv2d_1_scale_0" -> "13 dequantize_per_channel_default_1"  [label="(1,)", style=solid];
-"11 conv2d_1_zero_point_0" -> "12 quantize_per_channel_default_1"  [label="(1,)", style=solid];
-"11 conv2d_1_zero_point_0" -> "13 dequantize_per_channel_default_1"  [label="(1,)", style=solid];
-"12 quantize_per_channel_default_1" -> "13 dequantize_per_channel_default_1"  [label="(3, 3, 1, 1)", style=solid];
-"13 dequantize_per_channel_default_1" -> "14 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"14 conv2d_1" -> "18 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"15 _tensor_constant0" -> "16 add_"  [label="(1,)", style=solid];
-"16 add_" -> "19 add"  [label="(1, 3, 3, 3)", style=solid];
-"17 _tensor_constant0_1" -> "18 add__1"  [label="(1,)", style=solid];
-"18 add__1" -> "19 add"  [label="(1, 3, 3, 3)", style=solid];
-"19 add" -> "22 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"20 _param_constant4" -> "22 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"21 _param_constant5" -> "22 conv2d_2"  [label="(3,)", style=solid];
-"22 conv2d_2" -> "24 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"23 _tensor_constant0_2" -> "24 add_1"  [label="(1,)", style=solid];
-"24 add_1" -> "25 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 conv2d_scale_0" [id=8, type=get_attr];
+"9 conv2d_zero_point_0" [id=9, type=get_attr];
+"10 quantize_per_channel_default" [id=10, type=quantize_per_channel];
+"11 dequantize_per_channel_default" [id=11, type=dequantize_per_channel];
+"12 conv2d" [id=12, type=conv2d];
+"13 conv2d_1_scale_0" [id=13, type=get_attr];
+"14 conv2d_1_zero_point_0" [id=14, type=get_attr];
+"15 quantize_per_channel_default_1" [id=15, type=quantize_per_channel];
+"16 dequantize_per_channel_default_1" [id=16, type=dequantize_per_channel];
+"17 conv2d_1" [id=17, type=conv2d];
+"18 add_" [id=18, type=add_];
+"19 add__1" [id=19, type=add_];
+"20 add" [id=20, type=add];
+"21 conv2d_2" [id=21, type=conv2d];
+"22 add_1" [id=22, type=add];
+"23 output_1" [id=23, type=output];
+"0 conv_a_weight" -> "10 quantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "12 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "15 quantize_per_channel_default_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "17 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "21 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "21 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "18 add_"  [label="(1,)", style=solid];
+"6 bias" -> "19 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "22 add_1"  [label="(1,)", style=solid];
+"7 x" -> "12 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"8 conv2d_scale_0" -> "10 quantize_per_channel_default"  [label="(1,)", style=solid];
+"8 conv2d_scale_0" -> "11 dequantize_per_channel_default"  [label="(1,)", style=solid];
+"9 conv2d_zero_point_0" -> "10 quantize_per_channel_default"  [label="(1,)", style=solid];
+"9 conv2d_zero_point_0" -> "11 dequantize_per_channel_default"  [label="(1,)", style=solid];
+"10 quantize_per_channel_default" -> "11 dequantize_per_channel_default"  [label="(3, 3, 1, 1)", style=solid];
+"11 dequantize_per_channel_default" -> "12 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"12 conv2d" -> "17 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"12 conv2d" -> "18 add_"  [label="(1, 3, 3, 3)", style=solid];
+"13 conv2d_1_scale_0" -> "15 quantize_per_channel_default_1"  [label="(1,)", style=solid];
+"13 conv2d_1_scale_0" -> "16 dequantize_per_channel_default_1"  [label="(1,)", style=solid];
+"14 conv2d_1_zero_point_0" -> "15 quantize_per_channel_default_1"  [label="(1,)", style=solid];
+"14 conv2d_1_zero_point_0" -> "16 dequantize_per_channel_default_1"  [label="(1,)", style=solid];
+"15 quantize_per_channel_default_1" -> "16 dequantize_per_channel_default_1"  [label="(3, 3, 1, 1)", style=solid];
+"16 dequantize_per_channel_default_1" -> "17 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"17 conv2d_1" -> "19 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"18 add_" -> "20 add"  [label="(1, 3, 3, 3)", style=solid];
+"19 add__1" -> "20 add"  [label="(1, 3, 3, 3)", style=solid];
+"20 add" -> "21 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"21 conv2d_2" -> "22 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"22 add_1" -> "23 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/qdq_shared_insert_weights_per_tensor.dot b/tests/torch/data/reference_graphs/fx/transformed/qdq_shared_insert_weights_per_tensor.dot
index 6c8925fe8d2..952b7087ef5 100644
--- a/tests/torch/data/reference_graphs/fx/transformed/qdq_shared_insert_weights_per_tensor.dot
+++ b/tests/torch/data/reference_graphs/fx/transformed/qdq_shared_insert_weights_per_tensor.dot
@@ -1,46 +1,44 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 quantize_per_tensor_default" [id=3, type=quantize_per_tensor];
-"4 dequantize_per_tensor_default" [id=4, type=dequantize_per_tensor];
-"5 conv2d" [id=5, type=conv2d];
-"6 _param_constant2" [id=6, type=get_attr];
-"7 _param_constant3" [id=7, type=get_attr];
-"8 quantize_per_tensor_default_1" [id=8, type=quantize_per_tensor];
-"9 dequantize_per_tensor_default_1" [id=9, type=dequantize_per_tensor];
-"10 conv2d_1" [id=10, type=conv2d];
-"11 _tensor_constant0" [id=11, type=get_attr];
-"12 add_" [id=12, type=add_];
-"13 _tensor_constant0_1" [id=13, type=get_attr];
-"14 add__1" [id=14, type=add_];
-"15 add" [id=15, type=add];
-"16 _param_constant4" [id=16, type=get_attr];
-"17 _param_constant5" [id=17, type=get_attr];
-"18 conv2d_2" [id=18, type=conv2d];
-"19 _tensor_constant0_2" [id=19, type=get_attr];
-"20 add_1" [id=20, type=add];
-"21 output" [id=21, type=output];
-"0 arg0_1" -> "5 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "3 quantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "5 conv2d"  [label="(3,)", style=solid];
-"3 quantize_per_tensor_default" -> "4 dequantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
-"4 dequantize_per_tensor_default" -> "5 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"5 conv2d" -> "10 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"5 conv2d" -> "12 add_"  [label="(1, 3, 3, 3)", style=solid];
-"6 _param_constant2" -> "8 quantize_per_tensor_default_1"  [label="(3, 3, 1, 1)", style=solid];
-"7 _param_constant3" -> "10 conv2d_1"  [label="(3,)", style=solid];
-"8 quantize_per_tensor_default_1" -> "9 dequantize_per_tensor_default_1"  [label="(3, 3, 1, 1)", style=solid];
-"9 dequantize_per_tensor_default_1" -> "10 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"10 conv2d_1" -> "14 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"11 _tensor_constant0" -> "12 add_"  [label="(1,)", style=solid];
-"12 add_" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"13 _tensor_constant0_1" -> "14 add__1"  [label="(1,)", style=solid];
-"14 add__1" -> "15 add"  [label="(1, 3, 3, 3)", style=solid];
-"15 add" -> "18 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"16 _param_constant4" -> "18 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"17 _param_constant5" -> "18 conv2d_2"  [label="(3,)", style=solid];
-"18 conv2d_2" -> "20 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"19 _tensor_constant0_2" -> "20 add_1"  [label="(1,)", style=solid];
-"20 add_1" -> "21 output"  [label="(1, 3, 3, 3)", style=solid];
+"0 conv_a_weight" [id=0, type=get_attr];
+"1 conv_a_bias" [id=1, type=get_attr];
+"2 conv_b_weight" [id=2, type=get_attr];
+"3 conv_b_bias" [id=3, type=get_attr];
+"4 conv_c_weight" [id=4, type=get_attr];
+"5 conv_c_bias" [id=5, type=get_attr];
+"6 bias" [id=6, type=get_attr];
+"7 x" [id=7, type=input];
+"8 quantize_per_tensor_default" [id=8, type=quantize_per_tensor];
+"9 dequantize_per_tensor_default" [id=9, type=dequantize_per_tensor];
+"10 conv2d" [id=10, type=conv2d];
+"11 quantize_per_tensor_default_1" [id=11, type=quantize_per_tensor];
+"12 dequantize_per_tensor_default_1" [id=12, type=dequantize_per_tensor];
+"13 conv2d_1" [id=13, type=conv2d];
+"14 add_" [id=14, type=add_];
+"15 add__1" [id=15, type=add_];
+"16 add" [id=16, type=add];
+"17 conv2d_2" [id=17, type=conv2d];
+"18 add_1" [id=18, type=add];
+"19 output_1" [id=19, type=output];
+"0 conv_a_weight" -> "8 quantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
+"1 conv_a_bias" -> "10 conv2d"  [label="(3,)", style=solid];
+"2 conv_b_weight" -> "11 quantize_per_tensor_default_1"  [label="(3, 3, 1, 1)", style=solid];
+"3 conv_b_bias" -> "13 conv2d_1"  [label="(3,)", style=solid];
+"4 conv_c_weight" -> "17 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
+"5 conv_c_bias" -> "17 conv2d_2"  [label="(3,)", style=solid];
+"6 bias" -> "14 add_"  [label="(1,)", style=solid];
+"6 bias" -> "15 add__1"  [label="(1,)", style=solid];
+"6 bias" -> "18 add_1"  [label="(1,)", style=solid];
+"7 x" -> "10 conv2d"  [label="(1, 3, 3, 3)", style=solid];
+"8 quantize_per_tensor_default" -> "9 dequantize_per_tensor_default"  [label="(3, 3, 1, 1)", style=solid];
+"9 dequantize_per_tensor_default" -> "10 conv2d"  [label="(3, 3, 1, 1)", style=solid];
+"10 conv2d" -> "13 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
+"10 conv2d" -> "14 add_"  [label="(1, 3, 3, 3)", style=solid];
+"11 quantize_per_tensor_default_1" -> "12 dequantize_per_tensor_default_1"  [label="(3, 3, 1, 1)", style=solid];
+"12 dequantize_per_tensor_default_1" -> "13 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
+"13 conv2d_1" -> "15 add__1"  [label="(1, 3, 3, 3)", style=solid];
+"14 add_" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"15 add__1" -> "16 add"  [label="(1, 3, 3, 3)", style=solid];
+"16 add" -> "17 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
+"17 conv2d_2" -> "18 add_1"  [label="(1, 3, 3, 3)", style=solid];
+"18 add_1" -> "19 output_1"  [label="(1, 3, 3, 3)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/transformed/shared_constants_unification_transformation_test.dot b/tests/torch/data/reference_graphs/fx/transformed/shared_constants_unification_transformation_test.dot
deleted file mode 100644
index 996b76cf4be..00000000000
--- a/tests/torch/data/reference_graphs/fx/transformed/shared_constants_unification_transformation_test.dot
+++ /dev/null
@@ -1,36 +0,0 @@
-strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d" [id=3, type=conv2d];
-"4 _param_constant2" [id=4, type=get_attr];
-"5 _param_constant3" [id=5, type=get_attr];
-"6 conv2d_1" [id=6, type=conv2d];
-"7 _tensor_constant0" [id=7, type=get_attr];
-"8 add_" [id=8, type=add_];
-"9 add__1" [id=9, type=add_];
-"10 add" [id=10, type=add];
-"11 _param_constant4" [id=11, type=get_attr];
-"12 _param_constant5" [id=12, type=get_attr];
-"13 conv2d_2" [id=13, type=conv2d];
-"14 add_1" [id=14, type=add];
-"15 output" [id=15, type=output];
-"0 arg0_1" -> "3 conv2d"  [label="(1, 3, 3, 3)", style=solid];
-"1 _param_constant0" -> "3 conv2d"  [label="(3, 3, 1, 1)", style=solid];
-"2 _param_constant1" -> "3 conv2d"  [label="(3,)", style=solid];
-"3 conv2d" -> "6 conv2d_1"  [label="(1, 3, 3, 3)", style=solid];
-"3 conv2d" -> "8 add_"  [label="(1, 3, 3, 3)", style=solid];
-"4 _param_constant2" -> "6 conv2d_1"  [label="(3, 3, 1, 1)", style=solid];
-"5 _param_constant3" -> "6 conv2d_1"  [label="(3,)", style=solid];
-"6 conv2d_1" -> "9 add__1"  [label="(1, 3, 3, 3)", style=solid];
-"7 _tensor_constant0" -> "8 add_"  [label="(1,)", style=solid];
-"7 _tensor_constant0" -> "9 add__1"  [label="(1,)", style=solid];
-"7 _tensor_constant0" -> "14 add_1"  [label="(1,)", style=solid];
-"8 add_" -> "10 add"  [label="(1, 3, 3, 3)", style=solid];
-"9 add__1" -> "10 add"  [label="(1, 3, 3, 3)", style=solid];
-"10 add" -> "13 conv2d_2"  [label="(1, 3, 3, 3)", style=solid];
-"11 _param_constant4" -> "13 conv2d_2"  [label="(3, 3, 1, 1)", style=solid];
-"12 _param_constant5" -> "13 conv2d_2"  [label="(3,)", style=solid];
-"13 conv2d_2" -> "14 add_1"  [label="(1, 3, 3, 3)", style=solid];
-"14 add_1" -> "15 output"  [label="(1, 3, 3, 3)", style=solid];
-}
diff --git a/tests/torch/data/reference_graphs/fx/unet.dot b/tests/torch/data/reference_graphs/fx/unet.dot
index 1412ad8f8b1..730441ca545 100644
--- a/tests/torch/data/reference_graphs/fx/unet.dot
+++ b/tests/torch/data/reference_graphs/fx/unet.dot
@@ -1,537 +1,429 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d" [id=3, type=conv2d];
-"4 empty" [id=4, type=empty];
-"5 _param_constant2" [id=5, type=get_attr];
-"6 _param_constant3" [id=6, type=get_attr];
-"7 _tensor_constant0" [id=7, type=get_attr];
-"8 _tensor_constant1" [id=8, type=get_attr];
-"9 _native_batch_norm_legit_no_training" [id=9, type=_native_batch_norm_legit_no_training];
-"10 getitem" [id=10, type=__getitem__];
-"11 getitem_1" [id=11, type=__getitem__];
-"12 getitem_2" [id=12, type=__getitem__];
-"13 relu" [id=13, type=relu];
-"14 _param_constant4" [id=14, type=get_attr];
-"15 _param_constant5" [id=15, type=get_attr];
-"16 conv2d_1" [id=16, type=conv2d];
-"17 empty_1" [id=17, type=empty];
-"18 _param_constant6" [id=18, type=get_attr];
-"19 _param_constant7" [id=19, type=get_attr];
-"20 _tensor_constant2" [id=20, type=get_attr];
-"21 _tensor_constant3" [id=21, type=get_attr];
-"22 _native_batch_norm_legit_no_training_1" [id=22, type=_native_batch_norm_legit_no_training];
-"23 getitem_3" [id=23, type=__getitem__];
-"24 getitem_4" [id=24, type=__getitem__];
-"25 getitem_5" [id=25, type=__getitem__];
-"26 relu_1" [id=26, type=relu];
-"27 max_pool2d" [id=27, type=max_pool2d];
-"28 _param_constant8" [id=28, type=get_attr];
-"29 _param_constant9" [id=29, type=get_attr];
-"30 conv2d_2" [id=30, type=conv2d];
-"31 empty_2" [id=31, type=empty];
-"32 _param_constant10" [id=32, type=get_attr];
-"33 _param_constant11" [id=33, type=get_attr];
-"34 _tensor_constant4" [id=34, type=get_attr];
-"35 _tensor_constant5" [id=35, type=get_attr];
-"36 _native_batch_norm_legit_no_training_2" [id=36, type=_native_batch_norm_legit_no_training];
-"37 getitem_6" [id=37, type=__getitem__];
-"38 getitem_7" [id=38, type=__getitem__];
-"39 getitem_8" [id=39, type=__getitem__];
-"40 relu_2" [id=40, type=relu];
-"41 _param_constant12" [id=41, type=get_attr];
-"42 _param_constant13" [id=42, type=get_attr];
-"43 conv2d_3" [id=43, type=conv2d];
-"44 empty_3" [id=44, type=empty];
-"45 _param_constant14" [id=45, type=get_attr];
-"46 _param_constant15" [id=46, type=get_attr];
-"47 _tensor_constant6" [id=47, type=get_attr];
-"48 _tensor_constant7" [id=48, type=get_attr];
-"49 _native_batch_norm_legit_no_training_3" [id=49, type=_native_batch_norm_legit_no_training];
-"50 getitem_9" [id=50, type=__getitem__];
-"51 getitem_10" [id=51, type=__getitem__];
-"52 getitem_11" [id=52, type=__getitem__];
-"53 relu_3" [id=53, type=relu];
-"54 max_pool2d_1" [id=54, type=max_pool2d];
-"55 _param_constant16" [id=55, type=get_attr];
-"56 _param_constant17" [id=56, type=get_attr];
-"57 conv2d_4" [id=57, type=conv2d];
-"58 empty_4" [id=58, type=empty];
-"59 _param_constant18" [id=59, type=get_attr];
-"60 _param_constant19" [id=60, type=get_attr];
-"61 _tensor_constant8" [id=61, type=get_attr];
-"62 _tensor_constant9" [id=62, type=get_attr];
-"63 _native_batch_norm_legit_no_training_4" [id=63, type=_native_batch_norm_legit_no_training];
-"64 getitem_12" [id=64, type=__getitem__];
-"65 getitem_13" [id=65, type=__getitem__];
-"66 getitem_14" [id=66, type=__getitem__];
-"67 relu_4" [id=67, type=relu];
-"68 _param_constant20" [id=68, type=get_attr];
-"69 _param_constant21" [id=69, type=get_attr];
-"70 conv2d_5" [id=70, type=conv2d];
-"71 empty_5" [id=71, type=empty];
-"72 _param_constant22" [id=72, type=get_attr];
-"73 _param_constant23" [id=73, type=get_attr];
-"74 _tensor_constant10" [id=74, type=get_attr];
-"75 _tensor_constant11" [id=75, type=get_attr];
-"76 _native_batch_norm_legit_no_training_5" [id=76, type=_native_batch_norm_legit_no_training];
-"77 getitem_15" [id=77, type=__getitem__];
-"78 getitem_16" [id=78, type=__getitem__];
-"79 getitem_17" [id=79, type=__getitem__];
-"80 relu_5" [id=80, type=relu];
-"81 max_pool2d_2" [id=81, type=max_pool2d];
-"82 _param_constant24" [id=82, type=get_attr];
-"83 _param_constant25" [id=83, type=get_attr];
-"84 conv2d_6" [id=84, type=conv2d];
-"85 empty_6" [id=85, type=empty];
-"86 _param_constant26" [id=86, type=get_attr];
-"87 _param_constant27" [id=87, type=get_attr];
-"88 _tensor_constant12" [id=88, type=get_attr];
-"89 _tensor_constant13" [id=89, type=get_attr];
-"90 _native_batch_norm_legit_no_training_6" [id=90, type=_native_batch_norm_legit_no_training];
-"91 getitem_18" [id=91, type=__getitem__];
-"92 getitem_19" [id=92, type=__getitem__];
-"93 getitem_20" [id=93, type=__getitem__];
-"94 relu_6" [id=94, type=relu];
-"95 _param_constant28" [id=95, type=get_attr];
-"96 _param_constant29" [id=96, type=get_attr];
-"97 conv2d_7" [id=97, type=conv2d];
-"98 empty_7" [id=98, type=empty];
-"99 _param_constant30" [id=99, type=get_attr];
-"100 _param_constant31" [id=100, type=get_attr];
-"101 _tensor_constant14" [id=101, type=get_attr];
-"102 _tensor_constant15" [id=102, type=get_attr];
-"103 _native_batch_norm_legit_no_training_7" [id=103, type=_native_batch_norm_legit_no_training];
-"104 getitem_21" [id=104, type=__getitem__];
-"105 getitem_22" [id=105, type=__getitem__];
-"106 getitem_23" [id=106, type=__getitem__];
-"107 relu_7" [id=107, type=relu];
-"108 max_pool2d_3" [id=108, type=max_pool2d];
-"109 _param_constant32" [id=109, type=get_attr];
-"110 _param_constant33" [id=110, type=get_attr];
-"111 conv2d_8" [id=111, type=conv2d];
-"112 empty_8" [id=112, type=empty];
-"113 _param_constant34" [id=113, type=get_attr];
-"114 _param_constant35" [id=114, type=get_attr];
-"115 _tensor_constant16" [id=115, type=get_attr];
-"116 _tensor_constant17" [id=116, type=get_attr];
-"117 _native_batch_norm_legit_no_training_8" [id=117, type=_native_batch_norm_legit_no_training];
-"118 getitem_24" [id=118, type=__getitem__];
-"119 getitem_25" [id=119, type=__getitem__];
-"120 getitem_26" [id=120, type=__getitem__];
-"121 relu_8" [id=121, type=relu];
-"122 _param_constant36" [id=122, type=get_attr];
-"123 _param_constant37" [id=123, type=get_attr];
-"124 conv2d_9" [id=124, type=conv2d];
-"125 empty_9" [id=125, type=empty];
-"126 _param_constant38" [id=126, type=get_attr];
-"127 _param_constant39" [id=127, type=get_attr];
-"128 _tensor_constant18" [id=128, type=get_attr];
-"129 _tensor_constant19" [id=129, type=get_attr];
-"130 _native_batch_norm_legit_no_training_9" [id=130, type=_native_batch_norm_legit_no_training];
-"131 getitem_27" [id=131, type=__getitem__];
-"132 getitem_28" [id=132, type=__getitem__];
-"133 getitem_29" [id=133, type=__getitem__];
-"134 relu_9" [id=134, type=relu];
-"135 _param_constant40" [id=135, type=get_attr];
-"136 _param_constant41" [id=136, type=get_attr];
-"137 conv_transpose2d" [id=137, type=conv_transpose2d];
-"138 slice_1" [id=138, type=slice];
-"139 slice_2" [id=139, type=slice];
-"140 slice_3" [id=140, type=slice];
-"141 slice_4" [id=141, type=slice];
-"142 cat" [id=142, type=cat];
-"143 _param_constant42" [id=143, type=get_attr];
-"144 _param_constant43" [id=144, type=get_attr];
-"145 conv2d_10" [id=145, type=conv2d];
-"146 empty_10" [id=146, type=empty];
-"147 _param_constant44" [id=147, type=get_attr];
-"148 _param_constant45" [id=148, type=get_attr];
-"149 _tensor_constant20" [id=149, type=get_attr];
-"150 _tensor_constant21" [id=150, type=get_attr];
-"151 _native_batch_norm_legit_no_training_10" [id=151, type=_native_batch_norm_legit_no_training];
-"152 getitem_30" [id=152, type=__getitem__];
-"153 getitem_31" [id=153, type=__getitem__];
-"154 getitem_32" [id=154, type=__getitem__];
-"155 relu_10" [id=155, type=relu];
-"156 _param_constant46" [id=156, type=get_attr];
-"157 _param_constant47" [id=157, type=get_attr];
-"158 conv2d_11" [id=158, type=conv2d];
-"159 empty_11" [id=159, type=empty];
-"160 _param_constant48" [id=160, type=get_attr];
-"161 _param_constant49" [id=161, type=get_attr];
-"162 _tensor_constant22" [id=162, type=get_attr];
-"163 _tensor_constant23" [id=163, type=get_attr];
-"164 _native_batch_norm_legit_no_training_11" [id=164, type=_native_batch_norm_legit_no_training];
-"165 getitem_33" [id=165, type=__getitem__];
-"166 getitem_34" [id=166, type=__getitem__];
-"167 getitem_35" [id=167, type=__getitem__];
-"168 relu_11" [id=168, type=relu];
-"169 _param_constant50" [id=169, type=get_attr];
-"170 _param_constant51" [id=170, type=get_attr];
-"171 conv_transpose2d_1" [id=171, type=conv_transpose2d];
-"172 slice_5" [id=172, type=slice];
-"173 slice_6" [id=173, type=slice];
-"174 slice_7" [id=174, type=slice];
-"175 slice_8" [id=175, type=slice];
-"176 cat_1" [id=176, type=cat];
-"177 _param_constant52" [id=177, type=get_attr];
-"178 _param_constant53" [id=178, type=get_attr];
-"179 conv2d_12" [id=179, type=conv2d];
-"180 empty_12" [id=180, type=empty];
-"181 _param_constant54" [id=181, type=get_attr];
-"182 _param_constant55" [id=182, type=get_attr];
-"183 _tensor_constant24" [id=183, type=get_attr];
-"184 _tensor_constant25" [id=184, type=get_attr];
-"185 _native_batch_norm_legit_no_training_12" [id=185, type=_native_batch_norm_legit_no_training];
-"186 getitem_36" [id=186, type=__getitem__];
-"187 getitem_37" [id=187, type=__getitem__];
-"188 getitem_38" [id=188, type=__getitem__];
-"189 relu_12" [id=189, type=relu];
-"190 _param_constant56" [id=190, type=get_attr];
-"191 _param_constant57" [id=191, type=get_attr];
+"0 down_path_0_block_0_weight" [id=0, type=get_attr];
+"1 down_path_0_block_0_bias" [id=1, type=get_attr];
+"2 down_path_0_block_1_weight" [id=2, type=get_attr];
+"3 down_path_0_block_1_bias" [id=3, type=get_attr];
+"4 down_path_0_block_3_weight" [id=4, type=get_attr];
+"5 down_path_0_block_3_bias" [id=5, type=get_attr];
+"6 down_path_0_block_4_weight" [id=6, type=get_attr];
+"7 down_path_0_block_4_bias" [id=7, type=get_attr];
+"8 down_path_1_block_0_weight" [id=8, type=get_attr];
+"9 down_path_1_block_0_bias" [id=9, type=get_attr];
+"10 down_path_1_block_1_weight" [id=10, type=get_attr];
+"11 down_path_1_block_1_bias" [id=11, type=get_attr];
+"12 down_path_1_block_3_weight" [id=12, type=get_attr];
+"13 down_path_1_block_3_bias" [id=13, type=get_attr];
+"14 down_path_1_block_4_weight" [id=14, type=get_attr];
+"15 down_path_1_block_4_bias" [id=15, type=get_attr];
+"16 down_path_2_block_0_weight" [id=16, type=get_attr];
+"17 down_path_2_block_0_bias" [id=17, type=get_attr];
+"18 down_path_2_block_1_weight" [id=18, type=get_attr];
+"19 down_path_2_block_1_bias" [id=19, type=get_attr];
+"20 down_path_2_block_3_weight" [id=20, type=get_attr];
+"21 down_path_2_block_3_bias" [id=21, type=get_attr];
+"22 down_path_2_block_4_weight" [id=22, type=get_attr];
+"23 down_path_2_block_4_bias" [id=23, type=get_attr];
+"24 down_path_3_block_0_weight" [id=24, type=get_attr];
+"25 down_path_3_block_0_bias" [id=25, type=get_attr];
+"26 down_path_3_block_1_weight" [id=26, type=get_attr];
+"27 down_path_3_block_1_bias" [id=27, type=get_attr];
+"28 down_path_3_block_3_weight" [id=28, type=get_attr];
+"29 down_path_3_block_3_bias" [id=29, type=get_attr];
+"30 down_path_3_block_4_weight" [id=30, type=get_attr];
+"31 down_path_3_block_4_bias" [id=31, type=get_attr];
+"32 down_path_4_block_0_weight" [id=32, type=get_attr];
+"33 down_path_4_block_0_bias" [id=33, type=get_attr];
+"34 down_path_4_block_1_weight" [id=34, type=get_attr];
+"35 down_path_4_block_1_bias" [id=35, type=get_attr];
+"36 down_path_4_block_3_weight" [id=36, type=get_attr];
+"37 down_path_4_block_3_bias" [id=37, type=get_attr];
+"38 down_path_4_block_4_weight" [id=38, type=get_attr];
+"39 down_path_4_block_4_bias" [id=39, type=get_attr];
+"40 up_path_0_up_weight" [id=40, type=get_attr];
+"41 up_path_0_up_bias" [id=41, type=get_attr];
+"42 up_path_0_conv_block_block_0_weight" [id=42, type=get_attr];
+"43 up_path_0_conv_block_block_0_bias" [id=43, type=get_attr];
+"44 up_path_0_conv_block_block_1_weight" [id=44, type=get_attr];
+"45 up_path_0_conv_block_block_1_bias" [id=45, type=get_attr];
+"46 up_path_0_conv_block_block_3_weight" [id=46, type=get_attr];
+"47 up_path_0_conv_block_block_3_bias" [id=47, type=get_attr];
+"48 up_path_0_conv_block_block_4_weight" [id=48, type=get_attr];
+"49 up_path_0_conv_block_block_4_bias" [id=49, type=get_attr];
+"50 up_path_1_up_weight" [id=50, type=get_attr];
+"51 up_path_1_up_bias" [id=51, type=get_attr];
+"52 up_path_1_conv_block_block_0_weight" [id=52, type=get_attr];
+"53 up_path_1_conv_block_block_0_bias" [id=53, type=get_attr];
+"54 up_path_1_conv_block_block_1_weight" [id=54, type=get_attr];
+"55 up_path_1_conv_block_block_1_bias" [id=55, type=get_attr];
+"56 up_path_1_conv_block_block_3_weight" [id=56, type=get_attr];
+"57 up_path_1_conv_block_block_3_bias" [id=57, type=get_attr];
+"58 up_path_1_conv_block_block_4_weight" [id=58, type=get_attr];
+"59 up_path_1_conv_block_block_4_bias" [id=59, type=get_attr];
+"60 up_path_2_up_weight" [id=60, type=get_attr];
+"61 up_path_2_up_bias" [id=61, type=get_attr];
+"62 up_path_2_conv_block_block_0_weight" [id=62, type=get_attr];
+"63 up_path_2_conv_block_block_0_bias" [id=63, type=get_attr];
+"64 up_path_2_conv_block_block_1_weight" [id=64, type=get_attr];
+"65 up_path_2_conv_block_block_1_bias" [id=65, type=get_attr];
+"66 up_path_2_conv_block_block_3_weight" [id=66, type=get_attr];
+"67 up_path_2_conv_block_block_3_bias" [id=67, type=get_attr];
+"68 up_path_2_conv_block_block_4_weight" [id=68, type=get_attr];
+"69 up_path_2_conv_block_block_4_bias" [id=69, type=get_attr];
+"70 up_path_3_up_weight" [id=70, type=get_attr];
+"71 up_path_3_up_bias" [id=71, type=get_attr];
+"72 up_path_3_conv_block_block_0_weight" [id=72, type=get_attr];
+"73 up_path_3_conv_block_block_0_bias" [id=73, type=get_attr];
+"74 up_path_3_conv_block_block_1_weight" [id=74, type=get_attr];
+"75 up_path_3_conv_block_block_1_bias" [id=75, type=get_attr];
+"76 up_path_3_conv_block_block_3_weight" [id=76, type=get_attr];
+"77 up_path_3_conv_block_block_3_bias" [id=77, type=get_attr];
+"78 up_path_3_conv_block_block_4_weight" [id=78, type=get_attr];
+"79 up_path_3_conv_block_block_4_bias" [id=79, type=get_attr];
+"80 last_weight" [id=80, type=get_attr];
+"81 last_bias" [id=81, type=get_attr];
+"82 down_path_0_block_1_running_mean" [id=82, type=get_attr];
+"83 down_path_0_block_1_running_var" [id=83, type=get_attr];
+"84 down_path_0_block_1_num_batches_tracked" [id=84, type=get_attr];
+"85 down_path_0_block_4_running_mean" [id=85, type=get_attr];
+"86 down_path_0_block_4_running_var" [id=86, type=get_attr];
+"87 down_path_0_block_4_num_batches_tracked" [id=87, type=get_attr];
+"88 down_path_1_block_1_running_mean" [id=88, type=get_attr];
+"89 down_path_1_block_1_running_var" [id=89, type=get_attr];
+"90 down_path_1_block_1_num_batches_tracked" [id=90, type=get_attr];
+"91 down_path_1_block_4_running_mean" [id=91, type=get_attr];
+"92 down_path_1_block_4_running_var" [id=92, type=get_attr];
+"93 down_path_1_block_4_num_batches_tracked" [id=93, type=get_attr];
+"94 down_path_2_block_1_running_mean" [id=94, type=get_attr];
+"95 down_path_2_block_1_running_var" [id=95, type=get_attr];
+"96 down_path_2_block_1_num_batches_tracked" [id=96, type=get_attr];
+"97 down_path_2_block_4_running_mean" [id=97, type=get_attr];
+"98 down_path_2_block_4_running_var" [id=98, type=get_attr];
+"99 down_path_2_block_4_num_batches_tracked" [id=99, type=get_attr];
+"100 down_path_3_block_1_running_mean" [id=100, type=get_attr];
+"101 down_path_3_block_1_running_var" [id=101, type=get_attr];
+"102 down_path_3_block_1_num_batches_tracked" [id=102, type=get_attr];
+"103 down_path_3_block_4_running_mean" [id=103, type=get_attr];
+"104 down_path_3_block_4_running_var" [id=104, type=get_attr];
+"105 down_path_3_block_4_num_batches_tracked" [id=105, type=get_attr];
+"106 down_path_4_block_1_running_mean" [id=106, type=get_attr];
+"107 down_path_4_block_1_running_var" [id=107, type=get_attr];
+"108 down_path_4_block_1_num_batches_tracked" [id=108, type=get_attr];
+"109 down_path_4_block_4_running_mean" [id=109, type=get_attr];
+"110 down_path_4_block_4_running_var" [id=110, type=get_attr];
+"111 down_path_4_block_4_num_batches_tracked" [id=111, type=get_attr];
+"112 up_path_0_conv_block_block_1_running_mean" [id=112, type=get_attr];
+"113 up_path_0_conv_block_block_1_running_var" [id=113, type=get_attr];
+"114 up_path_0_conv_block_block_1_num_batches_tracked" [id=114, type=get_attr];
+"115 up_path_0_conv_block_block_4_running_mean" [id=115, type=get_attr];
+"116 up_path_0_conv_block_block_4_running_var" [id=116, type=get_attr];
+"117 up_path_0_conv_block_block_4_num_batches_tracked" [id=117, type=get_attr];
+"118 up_path_1_conv_block_block_1_running_mean" [id=118, type=get_attr];
+"119 up_path_1_conv_block_block_1_running_var" [id=119, type=get_attr];
+"120 up_path_1_conv_block_block_1_num_batches_tracked" [id=120, type=get_attr];
+"121 up_path_1_conv_block_block_4_running_mean" [id=121, type=get_attr];
+"122 up_path_1_conv_block_block_4_running_var" [id=122, type=get_attr];
+"123 up_path_1_conv_block_block_4_num_batches_tracked" [id=123, type=get_attr];
+"124 up_path_2_conv_block_block_1_running_mean" [id=124, type=get_attr];
+"125 up_path_2_conv_block_block_1_running_var" [id=125, type=get_attr];
+"126 up_path_2_conv_block_block_1_num_batches_tracked" [id=126, type=get_attr];
+"127 up_path_2_conv_block_block_4_running_mean" [id=127, type=get_attr];
+"128 up_path_2_conv_block_block_4_running_var" [id=128, type=get_attr];
+"129 up_path_2_conv_block_block_4_num_batches_tracked" [id=129, type=get_attr];
+"130 up_path_3_conv_block_block_1_running_mean" [id=130, type=get_attr];
+"131 up_path_3_conv_block_block_1_running_var" [id=131, type=get_attr];
+"132 up_path_3_conv_block_block_1_num_batches_tracked" [id=132, type=get_attr];
+"133 up_path_3_conv_block_block_4_running_mean" [id=133, type=get_attr];
+"134 up_path_3_conv_block_block_4_running_var" [id=134, type=get_attr];
+"135 up_path_3_conv_block_block_4_num_batches_tracked" [id=135, type=get_attr];
+"136 x" [id=136, type=input];
+"137 conv2d" [id=137, type=conv2d];
+"138 batch_norm" [id=138, type=batch_norm];
+"139 relu" [id=139, type=relu];
+"140 conv2d_1" [id=140, type=conv2d];
+"141 batch_norm_1" [id=141, type=batch_norm];
+"142 relu_1" [id=142, type=relu];
+"143 max_pool2d" [id=143, type=max_pool2d];
+"144 conv2d_2" [id=144, type=conv2d];
+"145 batch_norm_2" [id=145, type=batch_norm];
+"146 relu_2" [id=146, type=relu];
+"147 conv2d_3" [id=147, type=conv2d];
+"148 batch_norm_3" [id=148, type=batch_norm];
+"149 relu_3" [id=149, type=relu];
+"150 max_pool2d_1" [id=150, type=max_pool2d];
+"151 conv2d_4" [id=151, type=conv2d];
+"152 batch_norm_4" [id=152, type=batch_norm];
+"153 relu_4" [id=153, type=relu];
+"154 conv2d_5" [id=154, type=conv2d];
+"155 batch_norm_5" [id=155, type=batch_norm];
+"156 relu_5" [id=156, type=relu];
+"157 max_pool2d_2" [id=157, type=max_pool2d];
+"158 conv2d_6" [id=158, type=conv2d];
+"159 batch_norm_6" [id=159, type=batch_norm];
+"160 relu_6" [id=160, type=relu];
+"161 conv2d_7" [id=161, type=conv2d];
+"162 batch_norm_7" [id=162, type=batch_norm];
+"163 relu_7" [id=163, type=relu];
+"164 max_pool2d_3" [id=164, type=max_pool2d];
+"165 conv2d_8" [id=165, type=conv2d];
+"166 batch_norm_8" [id=166, type=batch_norm];
+"167 relu_8" [id=167, type=relu];
+"168 conv2d_9" [id=168, type=conv2d];
+"169 batch_norm_9" [id=169, type=batch_norm];
+"170 relu_9" [id=170, type=relu];
+"171 conv_transpose2d" [id=171, type=conv_transpose2d];
+"172 slice_1" [id=172, type=slice];
+"173 slice_2" [id=173, type=slice];
+"174 slice_3" [id=174, type=slice];
+"175 slice_4" [id=175, type=slice];
+"176 cat" [id=176, type=cat];
+"177 conv2d_10" [id=177, type=conv2d];
+"178 batch_norm_10" [id=178, type=batch_norm];
+"179 relu_10" [id=179, type=relu];
+"180 conv2d_11" [id=180, type=conv2d];
+"181 batch_norm_11" [id=181, type=batch_norm];
+"182 relu_11" [id=182, type=relu];
+"183 conv_transpose2d_1" [id=183, type=conv_transpose2d];
+"184 slice_5" [id=184, type=slice];
+"185 slice_6" [id=185, type=slice];
+"186 slice_7" [id=186, type=slice];
+"187 slice_8" [id=187, type=slice];
+"188 cat_1" [id=188, type=cat];
+"189 conv2d_12" [id=189, type=conv2d];
+"190 batch_norm_12" [id=190, type=batch_norm];
+"191 relu_12" [id=191, type=relu];
 "192 conv2d_13" [id=192, type=conv2d];
-"193 empty_13" [id=193, type=empty];
-"194 _param_constant58" [id=194, type=get_attr];
-"195 _param_constant59" [id=195, type=get_attr];
-"196 _tensor_constant26" [id=196, type=get_attr];
-"197 _tensor_constant27" [id=197, type=get_attr];
-"198 _native_batch_norm_legit_no_training_13" [id=198, type=_native_batch_norm_legit_no_training];
-"199 getitem_39" [id=199, type=__getitem__];
-"200 getitem_40" [id=200, type=__getitem__];
-"201 getitem_41" [id=201, type=__getitem__];
-"202 relu_13" [id=202, type=relu];
-"203 _param_constant60" [id=203, type=get_attr];
-"204 _param_constant61" [id=204, type=get_attr];
-"205 conv_transpose2d_2" [id=205, type=conv_transpose2d];
-"206 slice_9" [id=206, type=slice];
-"207 slice_10" [id=207, type=slice];
-"208 slice_11" [id=208, type=slice];
-"209 slice_12" [id=209, type=slice];
-"210 cat_2" [id=210, type=cat];
-"211 _param_constant62" [id=211, type=get_attr];
-"212 _param_constant63" [id=212, type=get_attr];
-"213 conv2d_14" [id=213, type=conv2d];
-"214 empty_14" [id=214, type=empty];
-"215 _param_constant64" [id=215, type=get_attr];
-"216 _param_constant65" [id=216, type=get_attr];
-"217 _tensor_constant28" [id=217, type=get_attr];
-"218 _tensor_constant29" [id=218, type=get_attr];
-"219 _native_batch_norm_legit_no_training_14" [id=219, type=_native_batch_norm_legit_no_training];
-"220 getitem_42" [id=220, type=__getitem__];
-"221 getitem_43" [id=221, type=__getitem__];
-"222 getitem_44" [id=222, type=__getitem__];
-"223 relu_14" [id=223, type=relu];
-"224 _param_constant66" [id=224, type=get_attr];
-"225 _param_constant67" [id=225, type=get_attr];
-"226 conv2d_15" [id=226, type=conv2d];
-"227 empty_15" [id=227, type=empty];
-"228 _param_constant68" [id=228, type=get_attr];
-"229 _param_constant69" [id=229, type=get_attr];
-"230 _tensor_constant30" [id=230, type=get_attr];
-"231 _tensor_constant31" [id=231, type=get_attr];
-"232 _native_batch_norm_legit_no_training_15" [id=232, type=_native_batch_norm_legit_no_training];
-"233 getitem_45" [id=233, type=__getitem__];
-"234 getitem_46" [id=234, type=__getitem__];
-"235 getitem_47" [id=235, type=__getitem__];
-"236 relu_15" [id=236, type=relu];
-"237 _param_constant70" [id=237, type=get_attr];
-"238 _param_constant71" [id=238, type=get_attr];
-"239 conv_transpose2d_3" [id=239, type=conv_transpose2d];
-"240 slice_13" [id=240, type=slice];
-"241 slice_14" [id=241, type=slice];
-"242 slice_15" [id=242, type=slice];
-"243 slice_16" [id=243, type=slice];
-"244 cat_3" [id=244, type=cat];
-"245 _param_constant72" [id=245, type=get_attr];
-"246 _param_constant73" [id=246, type=get_attr];
-"247 conv2d_16" [id=247, type=conv2d];
-"248 empty_16" [id=248, type=empty];
-"249 _param_constant74" [id=249, type=get_attr];
-"250 _param_constant75" [id=250, type=get_attr];
-"251 _tensor_constant32" [id=251, type=get_attr];
-"252 _tensor_constant33" [id=252, type=get_attr];
-"253 _native_batch_norm_legit_no_training_16" [id=253, type=_native_batch_norm_legit_no_training];
-"254 getitem_48" [id=254, type=__getitem__];
-"255 getitem_49" [id=255, type=__getitem__];
-"256 getitem_50" [id=256, type=__getitem__];
-"257 relu_16" [id=257, type=relu];
-"258 _param_constant76" [id=258, type=get_attr];
-"259 _param_constant77" [id=259, type=get_attr];
-"260 conv2d_17" [id=260, type=conv2d];
-"261 empty_17" [id=261, type=empty];
-"262 _param_constant78" [id=262, type=get_attr];
-"263 _param_constant79" [id=263, type=get_attr];
-"264 _tensor_constant34" [id=264, type=get_attr];
-"265 _tensor_constant35" [id=265, type=get_attr];
-"266 _native_batch_norm_legit_no_training_17" [id=266, type=_native_batch_norm_legit_no_training];
-"267 getitem_51" [id=267, type=__getitem__];
-"268 getitem_52" [id=268, type=__getitem__];
-"269 getitem_53" [id=269, type=__getitem__];
-"270 relu_17" [id=270, type=relu];
-"271 _param_constant80" [id=271, type=get_attr];
-"272 _param_constant81" [id=272, type=get_attr];
-"273 conv2d_18" [id=273, type=conv2d];
-"274 output" [id=274, type=output];
-"0 arg0_1" -> "3 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"1 _param_constant0" -> "3 conv2d"  [label="(64, 3, 3, 3)", style=solid];
-"2 _param_constant1" -> "3 conv2d"  [label="(64,)", style=solid];
-"3 conv2d" -> "9 _native_batch_norm_legit_no_training"  [label="(1, 64, 222, 222)", style=solid];
-"5 _param_constant2" -> "9 _native_batch_norm_legit_no_training"  [label="(64,)", style=solid];
-"6 _param_constant3" -> "9 _native_batch_norm_legit_no_training"  [label="(64,)", style=solid];
-"7 _tensor_constant0" -> "9 _native_batch_norm_legit_no_training"  [label="(64,)", style=solid];
-"8 _tensor_constant1" -> "9 _native_batch_norm_legit_no_training"  [label="(64,)", style=solid];
-"9 _native_batch_norm_legit_no_training" -> "10 getitem"  [label="(1, 64, 222, 222)", style=solid];
-"9 _native_batch_norm_legit_no_training" -> "11 getitem_1"  [label="(1, 64, 222, 222)", style=solid];
-"9 _native_batch_norm_legit_no_training" -> "12 getitem_2"  [label="(1, 64, 222, 222)", style=solid];
-"10 getitem" -> "13 relu"  [label="(1, 64, 222, 222)", style=solid];
-"13 relu" -> "16 conv2d_1"  [label="(1, 64, 222, 222)", style=solid];
-"14 _param_constant4" -> "16 conv2d_1"  [label="(64, 64, 3, 3)", style=solid];
-"15 _param_constant5" -> "16 conv2d_1"  [label="(64,)", style=solid];
-"16 conv2d_1" -> "22 _native_batch_norm_legit_no_training_1"  [label="(1, 64, 220, 220)", style=solid];
-"18 _param_constant6" -> "22 _native_batch_norm_legit_no_training_1"  [label="(64,)", style=solid];
-"19 _param_constant7" -> "22 _native_batch_norm_legit_no_training_1"  [label="(64,)", style=solid];
-"20 _tensor_constant2" -> "22 _native_batch_norm_legit_no_training_1"  [label="(64,)", style=solid];
-"21 _tensor_constant3" -> "22 _native_batch_norm_legit_no_training_1"  [label="(64,)", style=solid];
-"22 _native_batch_norm_legit_no_training_1" -> "23 getitem_3"  [label="(1, 64, 220, 220)", style=solid];
-"22 _native_batch_norm_legit_no_training_1" -> "24 getitem_4"  [label="(1, 64, 220, 220)", style=solid];
-"22 _native_batch_norm_legit_no_training_1" -> "25 getitem_5"  [label="(1, 64, 220, 220)", style=solid];
-"23 getitem_3" -> "26 relu_1"  [label="(1, 64, 220, 220)", style=solid];
-"26 relu_1" -> "27 max_pool2d"  [label="(1, 64, 220, 220)", style=solid];
-"26 relu_1" -> "240 slice_13"  [label="(1, 64, 220, 220)", style=solid];
-"27 max_pool2d" -> "30 conv2d_2"  [label="(1, 64, 110, 110)", style=solid];
-"28 _param_constant8" -> "30 conv2d_2"  [label="(128, 64, 3, 3)", style=solid];
-"29 _param_constant9" -> "30 conv2d_2"  [label="(128,)", style=solid];
-"30 conv2d_2" -> "36 _native_batch_norm_legit_no_training_2"  [label="(1, 128, 108, 108)", style=solid];
-"32 _param_constant10" -> "36 _native_batch_norm_legit_no_training_2"  [label="(128,)", style=solid];
-"33 _param_constant11" -> "36 _native_batch_norm_legit_no_training_2"  [label="(128,)", style=solid];
-"34 _tensor_constant4" -> "36 _native_batch_norm_legit_no_training_2"  [label="(128,)", style=solid];
-"35 _tensor_constant5" -> "36 _native_batch_norm_legit_no_training_2"  [label="(128,)", style=solid];
-"36 _native_batch_norm_legit_no_training_2" -> "37 getitem_6"  [label="(1, 128, 108, 108)", style=solid];
-"36 _native_batch_norm_legit_no_training_2" -> "38 getitem_7"  [label="(1, 128, 108, 108)", style=solid];
-"36 _native_batch_norm_legit_no_training_2" -> "39 getitem_8"  [label="(1, 128, 108, 108)", style=solid];
-"37 getitem_6" -> "40 relu_2"  [label="(1, 128, 108, 108)", style=solid];
-"40 relu_2" -> "43 conv2d_3"  [label="(1, 128, 108, 108)", style=solid];
-"41 _param_constant12" -> "43 conv2d_3"  [label="(128, 128, 3, 3)", style=solid];
-"42 _param_constant13" -> "43 conv2d_3"  [label="(128,)", style=solid];
-"43 conv2d_3" -> "49 _native_batch_norm_legit_no_training_3"  [label="(1, 128, 106, 106)", style=solid];
-"45 _param_constant14" -> "49 _native_batch_norm_legit_no_training_3"  [label="(128,)", style=solid];
-"46 _param_constant15" -> "49 _native_batch_norm_legit_no_training_3"  [label="(128,)", style=solid];
-"47 _tensor_constant6" -> "49 _native_batch_norm_legit_no_training_3"  [label="(128,)", style=solid];
-"48 _tensor_constant7" -> "49 _native_batch_norm_legit_no_training_3"  [label="(128,)", style=solid];
-"49 _native_batch_norm_legit_no_training_3" -> "50 getitem_9"  [label="(1, 128, 106, 106)", style=solid];
-"49 _native_batch_norm_legit_no_training_3" -> "51 getitem_10"  [label="(1, 128, 106, 106)", style=solid];
-"49 _native_batch_norm_legit_no_training_3" -> "52 getitem_11"  [label="(1, 128, 106, 106)", style=solid];
-"50 getitem_9" -> "53 relu_3"  [label="(1, 128, 106, 106)", style=solid];
-"53 relu_3" -> "54 max_pool2d_1"  [label="(1, 128, 106, 106)", style=solid];
-"53 relu_3" -> "206 slice_9"  [label="(1, 128, 106, 106)", style=solid];
-"54 max_pool2d_1" -> "57 conv2d_4"  [label="(1, 128, 53, 53)", style=solid];
-"55 _param_constant16" -> "57 conv2d_4"  [label="(256, 128, 3, 3)", style=solid];
-"56 _param_constant17" -> "57 conv2d_4"  [label="(256,)", style=solid];
-"57 conv2d_4" -> "63 _native_batch_norm_legit_no_training_4"  [label="(1, 256, 51, 51)", style=solid];
-"59 _param_constant18" -> "63 _native_batch_norm_legit_no_training_4"  [label="(256,)", style=solid];
-"60 _param_constant19" -> "63 _native_batch_norm_legit_no_training_4"  [label="(256,)", style=solid];
-"61 _tensor_constant8" -> "63 _native_batch_norm_legit_no_training_4"  [label="(256,)", style=solid];
-"62 _tensor_constant9" -> "63 _native_batch_norm_legit_no_training_4"  [label="(256,)", style=solid];
-"63 _native_batch_norm_legit_no_training_4" -> "64 getitem_12"  [label="(1, 256, 51, 51)", style=solid];
-"63 _native_batch_norm_legit_no_training_4" -> "65 getitem_13"  [label="(1, 256, 51, 51)", style=solid];
-"63 _native_batch_norm_legit_no_training_4" -> "66 getitem_14"  [label="(1, 256, 51, 51)", style=solid];
-"64 getitem_12" -> "67 relu_4"  [label="(1, 256, 51, 51)", style=solid];
-"67 relu_4" -> "70 conv2d_5"  [label="(1, 256, 51, 51)", style=solid];
-"68 _param_constant20" -> "70 conv2d_5"  [label="(256, 256, 3, 3)", style=solid];
-"69 _param_constant21" -> "70 conv2d_5"  [label="(256,)", style=solid];
-"70 conv2d_5" -> "76 _native_batch_norm_legit_no_training_5"  [label="(1, 256, 49, 49)", style=solid];
-"72 _param_constant22" -> "76 _native_batch_norm_legit_no_training_5"  [label="(256,)", style=solid];
-"73 _param_constant23" -> "76 _native_batch_norm_legit_no_training_5"  [label="(256,)", style=solid];
-"74 _tensor_constant10" -> "76 _native_batch_norm_legit_no_training_5"  [label="(256,)", style=solid];
-"75 _tensor_constant11" -> "76 _native_batch_norm_legit_no_training_5"  [label="(256,)", style=solid];
-"76 _native_batch_norm_legit_no_training_5" -> "77 getitem_15"  [label="(1, 256, 49, 49)", style=solid];
-"76 _native_batch_norm_legit_no_training_5" -> "78 getitem_16"  [label="(1, 256, 49, 49)", style=solid];
-"76 _native_batch_norm_legit_no_training_5" -> "79 getitem_17"  [label="(1, 256, 49, 49)", style=solid];
-"77 getitem_15" -> "80 relu_5"  [label="(1, 256, 49, 49)", style=solid];
-"80 relu_5" -> "81 max_pool2d_2"  [label="(1, 256, 49, 49)", style=solid];
-"80 relu_5" -> "172 slice_5"  [label="(1, 256, 49, 49)", style=solid];
-"81 max_pool2d_2" -> "84 conv2d_6"  [label="(1, 256, 24, 24)", style=solid];
-"82 _param_constant24" -> "84 conv2d_6"  [label="(512, 256, 3, 3)", style=solid];
-"83 _param_constant25" -> "84 conv2d_6"  [label="(512,)", style=solid];
-"84 conv2d_6" -> "90 _native_batch_norm_legit_no_training_6"  [label="(1, 512, 22, 22)", style=solid];
-"86 _param_constant26" -> "90 _native_batch_norm_legit_no_training_6"  [label="(512,)", style=solid];
-"87 _param_constant27" -> "90 _native_batch_norm_legit_no_training_6"  [label="(512,)", style=solid];
-"88 _tensor_constant12" -> "90 _native_batch_norm_legit_no_training_6"  [label="(512,)", style=solid];
-"89 _tensor_constant13" -> "90 _native_batch_norm_legit_no_training_6"  [label="(512,)", style=solid];
-"90 _native_batch_norm_legit_no_training_6" -> "91 getitem_18"  [label="(1, 512, 22, 22)", style=solid];
-"90 _native_batch_norm_legit_no_training_6" -> "92 getitem_19"  [label="(1, 512, 22, 22)", style=solid];
-"90 _native_batch_norm_legit_no_training_6" -> "93 getitem_20"  [label="(1, 512, 22, 22)", style=solid];
-"91 getitem_18" -> "94 relu_6"  [label="(1, 512, 22, 22)", style=solid];
-"94 relu_6" -> "97 conv2d_7"  [label="(1, 512, 22, 22)", style=solid];
-"95 _param_constant28" -> "97 conv2d_7"  [label="(512, 512, 3, 3)", style=solid];
-"96 _param_constant29" -> "97 conv2d_7"  [label="(512,)", style=solid];
-"97 conv2d_7" -> "103 _native_batch_norm_legit_no_training_7"  [label="(1, 512, 20, 20)", style=solid];
-"99 _param_constant30" -> "103 _native_batch_norm_legit_no_training_7"  [label="(512,)", style=solid];
-"100 _param_constant31" -> "103 _native_batch_norm_legit_no_training_7"  [label="(512,)", style=solid];
-"101 _tensor_constant14" -> "103 _native_batch_norm_legit_no_training_7"  [label="(512,)", style=solid];
-"102 _tensor_constant15" -> "103 _native_batch_norm_legit_no_training_7"  [label="(512,)", style=solid];
-"103 _native_batch_norm_legit_no_training_7" -> "104 getitem_21"  [label="(1, 512, 20, 20)", style=solid];
-"103 _native_batch_norm_legit_no_training_7" -> "105 getitem_22"  [label="(1, 512, 20, 20)", style=solid];
-"103 _native_batch_norm_legit_no_training_7" -> "106 getitem_23"  [label="(1, 512, 20, 20)", style=solid];
-"104 getitem_21" -> "107 relu_7"  [label="(1, 512, 20, 20)", style=solid];
-"107 relu_7" -> "108 max_pool2d_3"  [label="(1, 512, 20, 20)", style=solid];
-"107 relu_7" -> "138 slice_1"  [label="(1, 512, 20, 20)", style=solid];
-"108 max_pool2d_3" -> "111 conv2d_8"  [label="(1, 512, 10, 10)", style=solid];
-"109 _param_constant32" -> "111 conv2d_8"  [label="(1024, 512, 3, 3)", style=solid];
-"110 _param_constant33" -> "111 conv2d_8"  [label="(1024,)", style=solid];
-"111 conv2d_8" -> "117 _native_batch_norm_legit_no_training_8"  [label="(1, 1024, 8, 8)", style=solid];
-"113 _param_constant34" -> "117 _native_batch_norm_legit_no_training_8"  [label="(1024,)", style=solid];
-"114 _param_constant35" -> "117 _native_batch_norm_legit_no_training_8"  [label="(1024,)", style=solid];
-"115 _tensor_constant16" -> "117 _native_batch_norm_legit_no_training_8"  [label="(1024,)", style=solid];
-"116 _tensor_constant17" -> "117 _native_batch_norm_legit_no_training_8"  [label="(1024,)", style=solid];
-"117 _native_batch_norm_legit_no_training_8" -> "118 getitem_24"  [label="(1, 1024, 8, 8)", style=solid];
-"117 _native_batch_norm_legit_no_training_8" -> "119 getitem_25"  [label="(1, 1024, 8, 8)", style=solid];
-"117 _native_batch_norm_legit_no_training_8" -> "120 getitem_26"  [label="(1, 1024, 8, 8)", style=solid];
-"118 getitem_24" -> "121 relu_8"  [label="(1, 1024, 8, 8)", style=solid];
-"121 relu_8" -> "124 conv2d_9"  [label="(1, 1024, 8, 8)", style=solid];
-"122 _param_constant36" -> "124 conv2d_9"  [label="(1024, 1024, 3, 3)", style=solid];
-"123 _param_constant37" -> "124 conv2d_9"  [label="(1024,)", style=solid];
-"124 conv2d_9" -> "130 _native_batch_norm_legit_no_training_9"  [label="(1, 1024, 6, 6)", style=solid];
-"126 _param_constant38" -> "130 _native_batch_norm_legit_no_training_9"  [label="(1024,)", style=solid];
-"127 _param_constant39" -> "130 _native_batch_norm_legit_no_training_9"  [label="(1024,)", style=solid];
-"128 _tensor_constant18" -> "130 _native_batch_norm_legit_no_training_9"  [label="(1024,)", style=solid];
-"129 _tensor_constant19" -> "130 _native_batch_norm_legit_no_training_9"  [label="(1024,)", style=solid];
-"130 _native_batch_norm_legit_no_training_9" -> "131 getitem_27"  [label="(1, 1024, 6, 6)", style=solid];
-"130 _native_batch_norm_legit_no_training_9" -> "132 getitem_28"  [label="(1, 1024, 6, 6)", style=solid];
-"130 _native_batch_norm_legit_no_training_9" -> "133 getitem_29"  [label="(1, 1024, 6, 6)", style=solid];
-"131 getitem_27" -> "134 relu_9"  [label="(1, 1024, 6, 6)", style=solid];
-"134 relu_9" -> "137 conv_transpose2d"  [label="(1, 1024, 6, 6)", style=solid];
-"135 _param_constant40" -> "137 conv_transpose2d"  [label="(1024, 512, 2, 2)", style=solid];
-"136 _param_constant41" -> "137 conv_transpose2d"  [label="(512,)", style=solid];
-"137 conv_transpose2d" -> "142 cat"  [label="(1, 512, 12, 12)", style=solid];
-"138 slice_1" -> "139 slice_2"  [label="(1, 512, 20, 20)", style=solid];
-"139 slice_2" -> "140 slice_3"  [label="(1, 512, 20, 20)", style=solid];
-"140 slice_3" -> "141 slice_4"  [label="(1, 512, 12, 20)", style=solid];
-"141 slice_4" -> "142 cat"  [label="(1, 512, 12, 12)", style=solid];
-"142 cat" -> "145 conv2d_10"  [label="(1, 1024, 12, 12)", style=solid];
-"143 _param_constant42" -> "145 conv2d_10"  [label="(512, 1024, 3, 3)", style=solid];
-"144 _param_constant43" -> "145 conv2d_10"  [label="(512,)", style=solid];
-"145 conv2d_10" -> "151 _native_batch_norm_legit_no_training_10"  [label="(1, 512, 10, 10)", style=solid];
-"147 _param_constant44" -> "151 _native_batch_norm_legit_no_training_10"  [label="(512,)", style=solid];
-"148 _param_constant45" -> "151 _native_batch_norm_legit_no_training_10"  [label="(512,)", style=solid];
-"149 _tensor_constant20" -> "151 _native_batch_norm_legit_no_training_10"  [label="(512,)", style=solid];
-"150 _tensor_constant21" -> "151 _native_batch_norm_legit_no_training_10"  [label="(512,)", style=solid];
-"151 _native_batch_norm_legit_no_training_10" -> "152 getitem_30"  [label="(1, 512, 10, 10)", style=solid];
-"151 _native_batch_norm_legit_no_training_10" -> "153 getitem_31"  [label="(1, 512, 10, 10)", style=solid];
-"151 _native_batch_norm_legit_no_training_10" -> "154 getitem_32"  [label="(1, 512, 10, 10)", style=solid];
-"152 getitem_30" -> "155 relu_10"  [label="(1, 512, 10, 10)", style=solid];
-"155 relu_10" -> "158 conv2d_11"  [label="(1, 512, 10, 10)", style=solid];
-"156 _param_constant46" -> "158 conv2d_11"  [label="(512, 512, 3, 3)", style=solid];
-"157 _param_constant47" -> "158 conv2d_11"  [label="(512,)", style=solid];
-"158 conv2d_11" -> "164 _native_batch_norm_legit_no_training_11"  [label="(1, 512, 8, 8)", style=solid];
-"160 _param_constant48" -> "164 _native_batch_norm_legit_no_training_11"  [label="(512,)", style=solid];
-"161 _param_constant49" -> "164 _native_batch_norm_legit_no_training_11"  [label="(512,)", style=solid];
-"162 _tensor_constant22" -> "164 _native_batch_norm_legit_no_training_11"  [label="(512,)", style=solid];
-"163 _tensor_constant23" -> "164 _native_batch_norm_legit_no_training_11"  [label="(512,)", style=solid];
-"164 _native_batch_norm_legit_no_training_11" -> "165 getitem_33"  [label="(1, 512, 8, 8)", style=solid];
-"164 _native_batch_norm_legit_no_training_11" -> "166 getitem_34"  [label="(1, 512, 8, 8)", style=solid];
-"164 _native_batch_norm_legit_no_training_11" -> "167 getitem_35"  [label="(1, 512, 8, 8)", style=solid];
-"165 getitem_33" -> "168 relu_11"  [label="(1, 512, 8, 8)", style=solid];
-"168 relu_11" -> "171 conv_transpose2d_1"  [label="(1, 512, 8, 8)", style=solid];
-"169 _param_constant50" -> "171 conv_transpose2d_1"  [label="(512, 256, 2, 2)", style=solid];
-"170 _param_constant51" -> "171 conv_transpose2d_1"  [label="(256,)", style=solid];
-"171 conv_transpose2d_1" -> "176 cat_1"  [label="(1, 256, 16, 16)", style=solid];
-"172 slice_5" -> "173 slice_6"  [label="(1, 256, 49, 49)", style=solid];
-"173 slice_6" -> "174 slice_7"  [label="(1, 256, 49, 49)", style=solid];
-"174 slice_7" -> "175 slice_8"  [label="(1, 256, 16, 49)", style=solid];
-"175 slice_8" -> "176 cat_1"  [label="(1, 256, 16, 16)", style=solid];
-"176 cat_1" -> "179 conv2d_12"  [label="(1, 512, 16, 16)", style=solid];
-"177 _param_constant52" -> "179 conv2d_12"  [label="(256, 512, 3, 3)", style=solid];
-"178 _param_constant53" -> "179 conv2d_12"  [label="(256,)", style=solid];
-"179 conv2d_12" -> "185 _native_batch_norm_legit_no_training_12"  [label="(1, 256, 14, 14)", style=solid];
-"181 _param_constant54" -> "185 _native_batch_norm_legit_no_training_12"  [label="(256,)", style=solid];
-"182 _param_constant55" -> "185 _native_batch_norm_legit_no_training_12"  [label="(256,)", style=solid];
-"183 _tensor_constant24" -> "185 _native_batch_norm_legit_no_training_12"  [label="(256,)", style=solid];
-"184 _tensor_constant25" -> "185 _native_batch_norm_legit_no_training_12"  [label="(256,)", style=solid];
-"185 _native_batch_norm_legit_no_training_12" -> "186 getitem_36"  [label="(1, 256, 14, 14)", style=solid];
-"185 _native_batch_norm_legit_no_training_12" -> "187 getitem_37"  [label="(1, 256, 14, 14)", style=solid];
-"185 _native_batch_norm_legit_no_training_12" -> "188 getitem_38"  [label="(1, 256, 14, 14)", style=solid];
-"186 getitem_36" -> "189 relu_12"  [label="(1, 256, 14, 14)", style=solid];
-"189 relu_12" -> "192 conv2d_13"  [label="(1, 256, 14, 14)", style=solid];
-"190 _param_constant56" -> "192 conv2d_13"  [label="(256, 256, 3, 3)", style=solid];
-"191 _param_constant57" -> "192 conv2d_13"  [label="(256,)", style=solid];
-"192 conv2d_13" -> "198 _native_batch_norm_legit_no_training_13"  [label="(1, 256, 12, 12)", style=solid];
-"194 _param_constant58" -> "198 _native_batch_norm_legit_no_training_13"  [label="(256,)", style=solid];
-"195 _param_constant59" -> "198 _native_batch_norm_legit_no_training_13"  [label="(256,)", style=solid];
-"196 _tensor_constant26" -> "198 _native_batch_norm_legit_no_training_13"  [label="(256,)", style=solid];
-"197 _tensor_constant27" -> "198 _native_batch_norm_legit_no_training_13"  [label="(256,)", style=solid];
-"198 _native_batch_norm_legit_no_training_13" -> "199 getitem_39"  [label="(1, 256, 12, 12)", style=solid];
-"198 _native_batch_norm_legit_no_training_13" -> "200 getitem_40"  [label="(1, 256, 12, 12)", style=solid];
-"198 _native_batch_norm_legit_no_training_13" -> "201 getitem_41"  [label="(1, 256, 12, 12)", style=solid];
-"199 getitem_39" -> "202 relu_13"  [label="(1, 256, 12, 12)", style=solid];
-"202 relu_13" -> "205 conv_transpose2d_2"  [label="(1, 256, 12, 12)", style=solid];
-"203 _param_constant60" -> "205 conv_transpose2d_2"  [label="(256, 128, 2, 2)", style=solid];
-"204 _param_constant61" -> "205 conv_transpose2d_2"  [label="(128,)", style=solid];
-"205 conv_transpose2d_2" -> "210 cat_2"  [label="(1, 128, 24, 24)", style=solid];
-"206 slice_9" -> "207 slice_10"  [label="(1, 128, 106, 106)", style=solid];
-"207 slice_10" -> "208 slice_11"  [label="(1, 128, 106, 106)", style=solid];
-"208 slice_11" -> "209 slice_12"  [label="(1, 128, 24, 106)", style=solid];
-"209 slice_12" -> "210 cat_2"  [label="(1, 128, 24, 24)", style=solid];
-"210 cat_2" -> "213 conv2d_14"  [label="(1, 256, 24, 24)", style=solid];
-"211 _param_constant62" -> "213 conv2d_14"  [label="(128, 256, 3, 3)", style=solid];
-"212 _param_constant63" -> "213 conv2d_14"  [label="(128,)", style=solid];
-"213 conv2d_14" -> "219 _native_batch_norm_legit_no_training_14"  [label="(1, 128, 22, 22)", style=solid];
-"215 _param_constant64" -> "219 _native_batch_norm_legit_no_training_14"  [label="(128,)", style=solid];
-"216 _param_constant65" -> "219 _native_batch_norm_legit_no_training_14"  [label="(128,)", style=solid];
-"217 _tensor_constant28" -> "219 _native_batch_norm_legit_no_training_14"  [label="(128,)", style=solid];
-"218 _tensor_constant29" -> "219 _native_batch_norm_legit_no_training_14"  [label="(128,)", style=solid];
-"219 _native_batch_norm_legit_no_training_14" -> "220 getitem_42"  [label="(1, 128, 22, 22)", style=solid];
-"219 _native_batch_norm_legit_no_training_14" -> "221 getitem_43"  [label="(1, 128, 22, 22)", style=solid];
-"219 _native_batch_norm_legit_no_training_14" -> "222 getitem_44"  [label="(1, 128, 22, 22)", style=solid];
-"220 getitem_42" -> "223 relu_14"  [label="(1, 128, 22, 22)", style=solid];
-"223 relu_14" -> "226 conv2d_15"  [label="(1, 128, 22, 22)", style=solid];
-"224 _param_constant66" -> "226 conv2d_15"  [label="(128, 128, 3, 3)", style=solid];
-"225 _param_constant67" -> "226 conv2d_15"  [label="(128,)", style=solid];
-"226 conv2d_15" -> "232 _native_batch_norm_legit_no_training_15"  [label="(1, 128, 20, 20)", style=solid];
-"228 _param_constant68" -> "232 _native_batch_norm_legit_no_training_15"  [label="(128,)", style=solid];
-"229 _param_constant69" -> "232 _native_batch_norm_legit_no_training_15"  [label="(128,)", style=solid];
-"230 _tensor_constant30" -> "232 _native_batch_norm_legit_no_training_15"  [label="(128,)", style=solid];
-"231 _tensor_constant31" -> "232 _native_batch_norm_legit_no_training_15"  [label="(128,)", style=solid];
-"232 _native_batch_norm_legit_no_training_15" -> "233 getitem_45"  [label="(1, 128, 20, 20)", style=solid];
-"232 _native_batch_norm_legit_no_training_15" -> "234 getitem_46"  [label="(1, 128, 20, 20)", style=solid];
-"232 _native_batch_norm_legit_no_training_15" -> "235 getitem_47"  [label="(1, 128, 20, 20)", style=solid];
-"233 getitem_45" -> "236 relu_15"  [label="(1, 128, 20, 20)", style=solid];
-"236 relu_15" -> "239 conv_transpose2d_3"  [label="(1, 128, 20, 20)", style=solid];
-"237 _param_constant70" -> "239 conv_transpose2d_3"  [label="(128, 64, 2, 2)", style=solid];
-"238 _param_constant71" -> "239 conv_transpose2d_3"  [label="(64,)", style=solid];
-"239 conv_transpose2d_3" -> "244 cat_3"  [label="(1, 64, 40, 40)", style=solid];
-"240 slice_13" -> "241 slice_14"  [label="(1, 64, 220, 220)", style=solid];
-"241 slice_14" -> "242 slice_15"  [label="(1, 64, 220, 220)", style=solid];
-"242 slice_15" -> "243 slice_16"  [label="(1, 64, 40, 220)", style=solid];
-"243 slice_16" -> "244 cat_3"  [label="(1, 64, 40, 40)", style=solid];
-"244 cat_3" -> "247 conv2d_16"  [label="(1, 128, 40, 40)", style=solid];
-"245 _param_constant72" -> "247 conv2d_16"  [label="(64, 128, 3, 3)", style=solid];
-"246 _param_constant73" -> "247 conv2d_16"  [label="(64,)", style=solid];
-"247 conv2d_16" -> "253 _native_batch_norm_legit_no_training_16"  [label="(1, 64, 38, 38)", style=solid];
-"249 _param_constant74" -> "253 _native_batch_norm_legit_no_training_16"  [label="(64,)", style=solid];
-"250 _param_constant75" -> "253 _native_batch_norm_legit_no_training_16"  [label="(64,)", style=solid];
-"251 _tensor_constant32" -> "253 _native_batch_norm_legit_no_training_16"  [label="(64,)", style=solid];
-"252 _tensor_constant33" -> "253 _native_batch_norm_legit_no_training_16"  [label="(64,)", style=solid];
-"253 _native_batch_norm_legit_no_training_16" -> "254 getitem_48"  [label="(1, 64, 38, 38)", style=solid];
-"253 _native_batch_norm_legit_no_training_16" -> "255 getitem_49"  [label="(1, 64, 38, 38)", style=solid];
-"253 _native_batch_norm_legit_no_training_16" -> "256 getitem_50"  [label="(1, 64, 38, 38)", style=solid];
-"254 getitem_48" -> "257 relu_16"  [label="(1, 64, 38, 38)", style=solid];
-"257 relu_16" -> "260 conv2d_17"  [label="(1, 64, 38, 38)", style=solid];
-"258 _param_constant76" -> "260 conv2d_17"  [label="(64, 64, 3, 3)", style=solid];
-"259 _param_constant77" -> "260 conv2d_17"  [label="(64,)", style=solid];
-"260 conv2d_17" -> "266 _native_batch_norm_legit_no_training_17"  [label="(1, 64, 36, 36)", style=solid];
-"262 _param_constant78" -> "266 _native_batch_norm_legit_no_training_17"  [label="(64,)", style=solid];
-"263 _param_constant79" -> "266 _native_batch_norm_legit_no_training_17"  [label="(64,)", style=solid];
-"264 _tensor_constant34" -> "266 _native_batch_norm_legit_no_training_17"  [label="(64,)", style=solid];
-"265 _tensor_constant35" -> "266 _native_batch_norm_legit_no_training_17"  [label="(64,)", style=solid];
-"266 _native_batch_norm_legit_no_training_17" -> "267 getitem_51"  [label="(1, 64, 36, 36)", style=solid];
-"266 _native_batch_norm_legit_no_training_17" -> "268 getitem_52"  [label="(1, 64, 36, 36)", style=solid];
-"266 _native_batch_norm_legit_no_training_17" -> "269 getitem_53"  [label="(1, 64, 36, 36)", style=solid];
-"267 getitem_51" -> "270 relu_17"  [label="(1, 64, 36, 36)", style=solid];
-"270 relu_17" -> "273 conv2d_18"  [label="(1, 64, 36, 36)", style=solid];
-"271 _param_constant80" -> "273 conv2d_18"  [label="(12, 64, 1, 1)", style=solid];
-"272 _param_constant81" -> "273 conv2d_18"  [label="(12,)", style=solid];
-"273 conv2d_18" -> "274 output"  [label="(1, 12, 36, 36)", style=solid];
+"193 batch_norm_13" [id=193, type=batch_norm];
+"194 relu_13" [id=194, type=relu];
+"195 conv_transpose2d_2" [id=195, type=conv_transpose2d];
+"196 slice_9" [id=196, type=slice];
+"197 slice_10" [id=197, type=slice];
+"198 slice_11" [id=198, type=slice];
+"199 slice_12" [id=199, type=slice];
+"200 cat_2" [id=200, type=cat];
+"201 conv2d_14" [id=201, type=conv2d];
+"202 batch_norm_14" [id=202, type=batch_norm];
+"203 relu_14" [id=203, type=relu];
+"204 conv2d_15" [id=204, type=conv2d];
+"205 batch_norm_15" [id=205, type=batch_norm];
+"206 relu_15" [id=206, type=relu];
+"207 conv_transpose2d_3" [id=207, type=conv_transpose2d];
+"208 slice_13" [id=208, type=slice];
+"209 slice_14" [id=209, type=slice];
+"210 slice_15" [id=210, type=slice];
+"211 slice_16" [id=211, type=slice];
+"212 cat_3" [id=212, type=cat];
+"213 conv2d_16" [id=213, type=conv2d];
+"214 batch_norm_16" [id=214, type=batch_norm];
+"215 relu_16" [id=215, type=relu];
+"216 conv2d_17" [id=216, type=conv2d];
+"217 batch_norm_17" [id=217, type=batch_norm];
+"218 relu_17" [id=218, type=relu];
+"219 conv2d_18" [id=219, type=conv2d];
+"220 output_1" [id=220, type=output];
+"0 down_path_0_block_0_weight" -> "137 conv2d"  [label="(64, 3, 3, 3)", style=solid];
+"1 down_path_0_block_0_bias" -> "137 conv2d"  [label="(64,)", style=solid];
+"2 down_path_0_block_1_weight" -> "138 batch_norm"  [label="(64,)", style=solid];
+"3 down_path_0_block_1_bias" -> "138 batch_norm"  [label="(64,)", style=solid];
+"4 down_path_0_block_3_weight" -> "140 conv2d_1"  [label="(64, 64, 3, 3)", style=solid];
+"5 down_path_0_block_3_bias" -> "140 conv2d_1"  [label="(64,)", style=solid];
+"6 down_path_0_block_4_weight" -> "141 batch_norm_1"  [label="(64,)", style=solid];
+"7 down_path_0_block_4_bias" -> "141 batch_norm_1"  [label="(64,)", style=solid];
+"8 down_path_1_block_0_weight" -> "144 conv2d_2"  [label="(128, 64, 3, 3)", style=solid];
+"9 down_path_1_block_0_bias" -> "144 conv2d_2"  [label="(128,)", style=solid];
+"10 down_path_1_block_1_weight" -> "145 batch_norm_2"  [label="(128,)", style=solid];
+"11 down_path_1_block_1_bias" -> "145 batch_norm_2"  [label="(128,)", style=solid];
+"12 down_path_1_block_3_weight" -> "147 conv2d_3"  [label="(128, 128, 3, 3)", style=solid];
+"13 down_path_1_block_3_bias" -> "147 conv2d_3"  [label="(128,)", style=solid];
+"14 down_path_1_block_4_weight" -> "148 batch_norm_3"  [label="(128,)", style=solid];
+"15 down_path_1_block_4_bias" -> "148 batch_norm_3"  [label="(128,)", style=solid];
+"16 down_path_2_block_0_weight" -> "151 conv2d_4"  [label="(256, 128, 3, 3)", style=solid];
+"17 down_path_2_block_0_bias" -> "151 conv2d_4"  [label="(256,)", style=solid];
+"18 down_path_2_block_1_weight" -> "152 batch_norm_4"  [label="(256,)", style=solid];
+"19 down_path_2_block_1_bias" -> "152 batch_norm_4"  [label="(256,)", style=solid];
+"20 down_path_2_block_3_weight" -> "154 conv2d_5"  [label="(256, 256, 3, 3)", style=solid];
+"21 down_path_2_block_3_bias" -> "154 conv2d_5"  [label="(256,)", style=solid];
+"22 down_path_2_block_4_weight" -> "155 batch_norm_5"  [label="(256,)", style=solid];
+"23 down_path_2_block_4_bias" -> "155 batch_norm_5"  [label="(256,)", style=solid];
+"24 down_path_3_block_0_weight" -> "158 conv2d_6"  [label="(512, 256, 3, 3)", style=solid];
+"25 down_path_3_block_0_bias" -> "158 conv2d_6"  [label="(512,)", style=solid];
+"26 down_path_3_block_1_weight" -> "159 batch_norm_6"  [label="(512,)", style=solid];
+"27 down_path_3_block_1_bias" -> "159 batch_norm_6"  [label="(512,)", style=solid];
+"28 down_path_3_block_3_weight" -> "161 conv2d_7"  [label="(512, 512, 3, 3)", style=solid];
+"29 down_path_3_block_3_bias" -> "161 conv2d_7"  [label="(512,)", style=solid];
+"30 down_path_3_block_4_weight" -> "162 batch_norm_7"  [label="(512,)", style=solid];
+"31 down_path_3_block_4_bias" -> "162 batch_norm_7"  [label="(512,)", style=solid];
+"32 down_path_4_block_0_weight" -> "165 conv2d_8"  [label="(1024, 512, 3, 3)", style=solid];
+"33 down_path_4_block_0_bias" -> "165 conv2d_8"  [label="(1024,)", style=solid];
+"34 down_path_4_block_1_weight" -> "166 batch_norm_8"  [label="(1024,)", style=solid];
+"35 down_path_4_block_1_bias" -> "166 batch_norm_8"  [label="(1024,)", style=solid];
+"36 down_path_4_block_3_weight" -> "168 conv2d_9"  [label="(1024, 1024, 3, 3)", style=solid];
+"37 down_path_4_block_3_bias" -> "168 conv2d_9"  [label="(1024,)", style=solid];
+"38 down_path_4_block_4_weight" -> "169 batch_norm_9"  [label="(1024,)", style=solid];
+"39 down_path_4_block_4_bias" -> "169 batch_norm_9"  [label="(1024,)", style=solid];
+"40 up_path_0_up_weight" -> "171 conv_transpose2d"  [label="(1024, 512, 2, 2)", style=solid];
+"41 up_path_0_up_bias" -> "171 conv_transpose2d"  [label="(512,)", style=solid];
+"42 up_path_0_conv_block_block_0_weight" -> "177 conv2d_10"  [label="(512, 1024, 3, 3)", style=solid];
+"43 up_path_0_conv_block_block_0_bias" -> "177 conv2d_10"  [label="(512,)", style=solid];
+"44 up_path_0_conv_block_block_1_weight" -> "178 batch_norm_10"  [label="(512,)", style=solid];
+"45 up_path_0_conv_block_block_1_bias" -> "178 batch_norm_10"  [label="(512,)", style=solid];
+"46 up_path_0_conv_block_block_3_weight" -> "180 conv2d_11"  [label="(512, 512, 3, 3)", style=solid];
+"47 up_path_0_conv_block_block_3_bias" -> "180 conv2d_11"  [label="(512,)", style=solid];
+"48 up_path_0_conv_block_block_4_weight" -> "181 batch_norm_11"  [label="(512,)", style=solid];
+"49 up_path_0_conv_block_block_4_bias" -> "181 batch_norm_11"  [label="(512,)", style=solid];
+"50 up_path_1_up_weight" -> "183 conv_transpose2d_1"  [label="(512, 256, 2, 2)", style=solid];
+"51 up_path_1_up_bias" -> "183 conv_transpose2d_1"  [label="(256,)", style=solid];
+"52 up_path_1_conv_block_block_0_weight" -> "189 conv2d_12"  [label="(256, 512, 3, 3)", style=solid];
+"53 up_path_1_conv_block_block_0_bias" -> "189 conv2d_12"  [label="(256,)", style=solid];
+"54 up_path_1_conv_block_block_1_weight" -> "190 batch_norm_12"  [label="(256,)", style=solid];
+"55 up_path_1_conv_block_block_1_bias" -> "190 batch_norm_12"  [label="(256,)", style=solid];
+"56 up_path_1_conv_block_block_3_weight" -> "192 conv2d_13"  [label="(256, 256, 3, 3)", style=solid];
+"57 up_path_1_conv_block_block_3_bias" -> "192 conv2d_13"  [label="(256,)", style=solid];
+"58 up_path_1_conv_block_block_4_weight" -> "193 batch_norm_13"  [label="(256,)", style=solid];
+"59 up_path_1_conv_block_block_4_bias" -> "193 batch_norm_13"  [label="(256,)", style=solid];
+"60 up_path_2_up_weight" -> "195 conv_transpose2d_2"  [label="(256, 128, 2, 2)", style=solid];
+"61 up_path_2_up_bias" -> "195 conv_transpose2d_2"  [label="(128,)", style=solid];
+"62 up_path_2_conv_block_block_0_weight" -> "201 conv2d_14"  [label="(128, 256, 3, 3)", style=solid];
+"63 up_path_2_conv_block_block_0_bias" -> "201 conv2d_14"  [label="(128,)", style=solid];
+"64 up_path_2_conv_block_block_1_weight" -> "202 batch_norm_14"  [label="(128,)", style=solid];
+"65 up_path_2_conv_block_block_1_bias" -> "202 batch_norm_14"  [label="(128,)", style=solid];
+"66 up_path_2_conv_block_block_3_weight" -> "204 conv2d_15"  [label="(128, 128, 3, 3)", style=solid];
+"67 up_path_2_conv_block_block_3_bias" -> "204 conv2d_15"  [label="(128,)", style=solid];
+"68 up_path_2_conv_block_block_4_weight" -> "205 batch_norm_15"  [label="(128,)", style=solid];
+"69 up_path_2_conv_block_block_4_bias" -> "205 batch_norm_15"  [label="(128,)", style=solid];
+"70 up_path_3_up_weight" -> "207 conv_transpose2d_3"  [label="(128, 64, 2, 2)", style=solid];
+"71 up_path_3_up_bias" -> "207 conv_transpose2d_3"  [label="(64,)", style=solid];
+"72 up_path_3_conv_block_block_0_weight" -> "213 conv2d_16"  [label="(64, 128, 3, 3)", style=solid];
+"73 up_path_3_conv_block_block_0_bias" -> "213 conv2d_16"  [label="(64,)", style=solid];
+"74 up_path_3_conv_block_block_1_weight" -> "214 batch_norm_16"  [label="(64,)", style=solid];
+"75 up_path_3_conv_block_block_1_bias" -> "214 batch_norm_16"  [label="(64,)", style=solid];
+"76 up_path_3_conv_block_block_3_weight" -> "216 conv2d_17"  [label="(64, 64, 3, 3)", style=solid];
+"77 up_path_3_conv_block_block_3_bias" -> "216 conv2d_17"  [label="(64,)", style=solid];
+"78 up_path_3_conv_block_block_4_weight" -> "217 batch_norm_17"  [label="(64,)", style=solid];
+"79 up_path_3_conv_block_block_4_bias" -> "217 batch_norm_17"  [label="(64,)", style=solid];
+"80 last_weight" -> "219 conv2d_18"  [label="(12, 64, 1, 1)", style=solid];
+"81 last_bias" -> "219 conv2d_18"  [label="(12,)", style=solid];
+"82 down_path_0_block_1_running_mean" -> "138 batch_norm"  [label="(64,)", style=solid];
+"83 down_path_0_block_1_running_var" -> "138 batch_norm"  [label="(64,)", style=solid];
+"85 down_path_0_block_4_running_mean" -> "141 batch_norm_1"  [label="(64,)", style=solid];
+"86 down_path_0_block_4_running_var" -> "141 batch_norm_1"  [label="(64,)", style=solid];
+"88 down_path_1_block_1_running_mean" -> "145 batch_norm_2"  [label="(128,)", style=solid];
+"89 down_path_1_block_1_running_var" -> "145 batch_norm_2"  [label="(128,)", style=solid];
+"91 down_path_1_block_4_running_mean" -> "148 batch_norm_3"  [label="(128,)", style=solid];
+"92 down_path_1_block_4_running_var" -> "148 batch_norm_3"  [label="(128,)", style=solid];
+"94 down_path_2_block_1_running_mean" -> "152 batch_norm_4"  [label="(256,)", style=solid];
+"95 down_path_2_block_1_running_var" -> "152 batch_norm_4"  [label="(256,)", style=solid];
+"97 down_path_2_block_4_running_mean" -> "155 batch_norm_5"  [label="(256,)", style=solid];
+"98 down_path_2_block_4_running_var" -> "155 batch_norm_5"  [label="(256,)", style=solid];
+"100 down_path_3_block_1_running_mean" -> "159 batch_norm_6"  [label="(512,)", style=solid];
+"101 down_path_3_block_1_running_var" -> "159 batch_norm_6"  [label="(512,)", style=solid];
+"103 down_path_3_block_4_running_mean" -> "162 batch_norm_7"  [label="(512,)", style=solid];
+"104 down_path_3_block_4_running_var" -> "162 batch_norm_7"  [label="(512,)", style=solid];
+"106 down_path_4_block_1_running_mean" -> "166 batch_norm_8"  [label="(1024,)", style=solid];
+"107 down_path_4_block_1_running_var" -> "166 batch_norm_8"  [label="(1024,)", style=solid];
+"109 down_path_4_block_4_running_mean" -> "169 batch_norm_9"  [label="(1024,)", style=solid];
+"110 down_path_4_block_4_running_var" -> "169 batch_norm_9"  [label="(1024,)", style=solid];
+"112 up_path_0_conv_block_block_1_running_mean" -> "178 batch_norm_10"  [label="(512,)", style=solid];
+"113 up_path_0_conv_block_block_1_running_var" -> "178 batch_norm_10"  [label="(512,)", style=solid];
+"115 up_path_0_conv_block_block_4_running_mean" -> "181 batch_norm_11"  [label="(512,)", style=solid];
+"116 up_path_0_conv_block_block_4_running_var" -> "181 batch_norm_11"  [label="(512,)", style=solid];
+"118 up_path_1_conv_block_block_1_running_mean" -> "190 batch_norm_12"  [label="(256,)", style=solid];
+"119 up_path_1_conv_block_block_1_running_var" -> "190 batch_norm_12"  [label="(256,)", style=solid];
+"121 up_path_1_conv_block_block_4_running_mean" -> "193 batch_norm_13"  [label="(256,)", style=solid];
+"122 up_path_1_conv_block_block_4_running_var" -> "193 batch_norm_13"  [label="(256,)", style=solid];
+"124 up_path_2_conv_block_block_1_running_mean" -> "202 batch_norm_14"  [label="(128,)", style=solid];
+"125 up_path_2_conv_block_block_1_running_var" -> "202 batch_norm_14"  [label="(128,)", style=solid];
+"127 up_path_2_conv_block_block_4_running_mean" -> "205 batch_norm_15"  [label="(128,)", style=solid];
+"128 up_path_2_conv_block_block_4_running_var" -> "205 batch_norm_15"  [label="(128,)", style=solid];
+"130 up_path_3_conv_block_block_1_running_mean" -> "214 batch_norm_16"  [label="(64,)", style=solid];
+"131 up_path_3_conv_block_block_1_running_var" -> "214 batch_norm_16"  [label="(64,)", style=solid];
+"133 up_path_3_conv_block_block_4_running_mean" -> "217 batch_norm_17"  [label="(64,)", style=solid];
+"134 up_path_3_conv_block_block_4_running_var" -> "217 batch_norm_17"  [label="(64,)", style=solid];
+"136 x" -> "137 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"137 conv2d" -> "138 batch_norm"  [label="(1, 64, 222, 222)", style=solid];
+"138 batch_norm" -> "139 relu"  [label="(1, 64, 222, 222)", style=solid];
+"139 relu" -> "140 conv2d_1"  [label="(1, 64, 222, 222)", style=solid];
+"140 conv2d_1" -> "141 batch_norm_1"  [label="(1, 64, 220, 220)", style=solid];
+"141 batch_norm_1" -> "142 relu_1"  [label="(1, 64, 220, 220)", style=solid];
+"142 relu_1" -> "143 max_pool2d"  [label="(1, 64, 220, 220)", style=solid];
+"142 relu_1" -> "208 slice_13"  [label="(1, 64, 220, 220)", style=solid];
+"143 max_pool2d" -> "144 conv2d_2"  [label="(1, 64, 110, 110)", style=solid];
+"144 conv2d_2" -> "145 batch_norm_2"  [label="(1, 128, 108, 108)", style=solid];
+"145 batch_norm_2" -> "146 relu_2"  [label="(1, 128, 108, 108)", style=solid];
+"146 relu_2" -> "147 conv2d_3"  [label="(1, 128, 108, 108)", style=solid];
+"147 conv2d_3" -> "148 batch_norm_3"  [label="(1, 128, 106, 106)", style=solid];
+"148 batch_norm_3" -> "149 relu_3"  [label="(1, 128, 106, 106)", style=solid];
+"149 relu_3" -> "150 max_pool2d_1"  [label="(1, 128, 106, 106)", style=solid];
+"149 relu_3" -> "196 slice_9"  [label="(1, 128, 106, 106)", style=solid];
+"150 max_pool2d_1" -> "151 conv2d_4"  [label="(1, 128, 53, 53)", style=solid];
+"151 conv2d_4" -> "152 batch_norm_4"  [label="(1, 256, 51, 51)", style=solid];
+"152 batch_norm_4" -> "153 relu_4"  [label="(1, 256, 51, 51)", style=solid];
+"153 relu_4" -> "154 conv2d_5"  [label="(1, 256, 51, 51)", style=solid];
+"154 conv2d_5" -> "155 batch_norm_5"  [label="(1, 256, 49, 49)", style=solid];
+"155 batch_norm_5" -> "156 relu_5"  [label="(1, 256, 49, 49)", style=solid];
+"156 relu_5" -> "157 max_pool2d_2"  [label="(1, 256, 49, 49)", style=solid];
+"156 relu_5" -> "184 slice_5"  [label="(1, 256, 49, 49)", style=solid];
+"157 max_pool2d_2" -> "158 conv2d_6"  [label="(1, 256, 24, 24)", style=solid];
+"158 conv2d_6" -> "159 batch_norm_6"  [label="(1, 512, 22, 22)", style=solid];
+"159 batch_norm_6" -> "160 relu_6"  [label="(1, 512, 22, 22)", style=solid];
+"160 relu_6" -> "161 conv2d_7"  [label="(1, 512, 22, 22)", style=solid];
+"161 conv2d_7" -> "162 batch_norm_7"  [label="(1, 512, 20, 20)", style=solid];
+"162 batch_norm_7" -> "163 relu_7"  [label="(1, 512, 20, 20)", style=solid];
+"163 relu_7" -> "164 max_pool2d_3"  [label="(1, 512, 20, 20)", style=solid];
+"163 relu_7" -> "172 slice_1"  [label="(1, 512, 20, 20)", style=solid];
+"164 max_pool2d_3" -> "165 conv2d_8"  [label="(1, 512, 10, 10)", style=solid];
+"165 conv2d_8" -> "166 batch_norm_8"  [label="(1, 1024, 8, 8)", style=solid];
+"166 batch_norm_8" -> "167 relu_8"  [label="(1, 1024, 8, 8)", style=solid];
+"167 relu_8" -> "168 conv2d_9"  [label="(1, 1024, 8, 8)", style=solid];
+"168 conv2d_9" -> "169 batch_norm_9"  [label="(1, 1024, 6, 6)", style=solid];
+"169 batch_norm_9" -> "170 relu_9"  [label="(1, 1024, 6, 6)", style=solid];
+"170 relu_9" -> "171 conv_transpose2d"  [label="(1, 1024, 6, 6)", style=solid];
+"171 conv_transpose2d" -> "176 cat"  [label="(1, 512, 12, 12)", style=solid];
+"172 slice_1" -> "173 slice_2"  [label="(1, 512, 20, 20)", style=solid];
+"173 slice_2" -> "174 slice_3"  [label="(1, 512, 20, 20)", style=solid];
+"174 slice_3" -> "175 slice_4"  [label="(1, 512, 12, 20)", style=solid];
+"175 slice_4" -> "176 cat"  [label="(1, 512, 12, 12)", style=solid];
+"176 cat" -> "177 conv2d_10"  [label="(1, 1024, 12, 12)", style=solid];
+"177 conv2d_10" -> "178 batch_norm_10"  [label="(1, 512, 10, 10)", style=solid];
+"178 batch_norm_10" -> "179 relu_10"  [label="(1, 512, 10, 10)", style=solid];
+"179 relu_10" -> "180 conv2d_11"  [label="(1, 512, 10, 10)", style=solid];
+"180 conv2d_11" -> "181 batch_norm_11"  [label="(1, 512, 8, 8)", style=solid];
+"181 batch_norm_11" -> "182 relu_11"  [label="(1, 512, 8, 8)", style=solid];
+"182 relu_11" -> "183 conv_transpose2d_1"  [label="(1, 512, 8, 8)", style=solid];
+"183 conv_transpose2d_1" -> "188 cat_1"  [label="(1, 256, 16, 16)", style=solid];
+"184 slice_5" -> "185 slice_6"  [label="(1, 256, 49, 49)", style=solid];
+"185 slice_6" -> "186 slice_7"  [label="(1, 256, 49, 49)", style=solid];
+"186 slice_7" -> "187 slice_8"  [label="(1, 256, 16, 49)", style=solid];
+"187 slice_8" -> "188 cat_1"  [label="(1, 256, 16, 16)", style=solid];
+"188 cat_1" -> "189 conv2d_12"  [label="(1, 512, 16, 16)", style=solid];
+"189 conv2d_12" -> "190 batch_norm_12"  [label="(1, 256, 14, 14)", style=solid];
+"190 batch_norm_12" -> "191 relu_12"  [label="(1, 256, 14, 14)", style=solid];
+"191 relu_12" -> "192 conv2d_13"  [label="(1, 256, 14, 14)", style=solid];
+"192 conv2d_13" -> "193 batch_norm_13"  [label="(1, 256, 12, 12)", style=solid];
+"193 batch_norm_13" -> "194 relu_13"  [label="(1, 256, 12, 12)", style=solid];
+"194 relu_13" -> "195 conv_transpose2d_2"  [label="(1, 256, 12, 12)", style=solid];
+"195 conv_transpose2d_2" -> "200 cat_2"  [label="(1, 128, 24, 24)", style=solid];
+"196 slice_9" -> "197 slice_10"  [label="(1, 128, 106, 106)", style=solid];
+"197 slice_10" -> "198 slice_11"  [label="(1, 128, 106, 106)", style=solid];
+"198 slice_11" -> "199 slice_12"  [label="(1, 128, 24, 106)", style=solid];
+"199 slice_12" -> "200 cat_2"  [label="(1, 128, 24, 24)", style=solid];
+"200 cat_2" -> "201 conv2d_14"  [label="(1, 256, 24, 24)", style=solid];
+"201 conv2d_14" -> "202 batch_norm_14"  [label="(1, 128, 22, 22)", style=solid];
+"202 batch_norm_14" -> "203 relu_14"  [label="(1, 128, 22, 22)", style=solid];
+"203 relu_14" -> "204 conv2d_15"  [label="(1, 128, 22, 22)", style=solid];
+"204 conv2d_15" -> "205 batch_norm_15"  [label="(1, 128, 20, 20)", style=solid];
+"205 batch_norm_15" -> "206 relu_15"  [label="(1, 128, 20, 20)", style=solid];
+"206 relu_15" -> "207 conv_transpose2d_3"  [label="(1, 128, 20, 20)", style=solid];
+"207 conv_transpose2d_3" -> "212 cat_3"  [label="(1, 64, 40, 40)", style=solid];
+"208 slice_13" -> "209 slice_14"  [label="(1, 64, 220, 220)", style=solid];
+"209 slice_14" -> "210 slice_15"  [label="(1, 64, 220, 220)", style=solid];
+"210 slice_15" -> "211 slice_16"  [label="(1, 64, 40, 220)", style=solid];
+"211 slice_16" -> "212 cat_3"  [label="(1, 64, 40, 40)", style=solid];
+"212 cat_3" -> "213 conv2d_16"  [label="(1, 128, 40, 40)", style=solid];
+"213 conv2d_16" -> "214 batch_norm_16"  [label="(1, 64, 38, 38)", style=solid];
+"214 batch_norm_16" -> "215 relu_16"  [label="(1, 64, 38, 38)", style=solid];
+"215 relu_16" -> "216 conv2d_17"  [label="(1, 64, 38, 38)", style=solid];
+"216 conv2d_17" -> "217 batch_norm_17"  [label="(1, 64, 36, 36)", style=solid];
+"217 batch_norm_17" -> "218 relu_17"  [label="(1, 64, 36, 36)", style=solid];
+"218 relu_17" -> "219 conv2d_18"  [label="(1, 64, 36, 36)", style=solid];
+"219 conv2d_18" -> "220 output_1"  [label="(1, 12, 36, 36)", style=solid];
 }
diff --git a/tests/torch/data/reference_graphs/fx/vit_b_16.dot b/tests/torch/data/reference_graphs/fx/vit_b_16.dot
index 38920ce2ff3..b29787ce0c4 100644
--- a/tests/torch/data/reference_graphs/fx/vit_b_16.dot
+++ b/tests/torch/data/reference_graphs/fx/vit_b_16.dot
@@ -1,1219 +1,1219 @@
 strict digraph  {
-"0 arg0_1" [id=0, type=input];
-"1 _param_constant0" [id=1, type=get_attr];
-"2 _param_constant1" [id=2, type=get_attr];
-"3 conv2d" [id=3, type=conv2d];
-"4 reshape" [id=4, type=reshape];
-"5 permute" [id=5, type=permute];
-"6 _param_constant2" [id=6, type=get_attr];
-"7 expand" [id=7, type=expand];
-"8 cat" [id=8, type=cat];
-"9 _param_constant3" [id=9, type=get_attr];
-"10 add" [id=10, type=add];
-"11 dropout" [id=11, type=dropout];
-"12 _param_constant4" [id=12, type=get_attr];
-"13 _param_constant5" [id=13, type=get_attr];
-"14 layer_norm" [id=14, type=layer_norm];
-"15 transpose" [id=15, type=transpose];
-"16 _param_constant6" [id=16, type=get_attr];
-"17 _param_constant7" [id=17, type=get_attr];
-"18 linear" [id=18, type=linear];
-"19 unflatten" [id=19, type=unflatten];
-"20 unsqueeze" [id=20, type=unsqueeze];
-"21 transpose_1" [id=21, type=transpose];
-"22 squeeze" [id=22, type=squeeze];
-"23 contiguous" [id=23, type=contiguous];
-"24 select" [id=24, type=select];
-"25 select_1" [id=25, type=select];
-"26 select_2" [id=26, type=select];
-"27 view" [id=27, type=view];
-"28 transpose_2" [id=28, type=transpose];
-"29 view_1" [id=29, type=view];
-"30 transpose_3" [id=30, type=transpose];
-"31 view_2" [id=31, type=view];
-"32 transpose_4" [id=32, type=transpose];
-"33 view_3" [id=33, type=view];
-"34 view_4" [id=34, type=view];
-"35 view_5" [id=35, type=view];
-"36 scaled_dot_product_attention" [id=36, type=scaled_dot_product_attention];
-"37 permute_1" [id=37, type=permute];
-"38 view_6" [id=38, type=view];
-"39 _param_constant8" [id=39, type=get_attr];
-"40 _param_constant9" [id=40, type=get_attr];
-"41 linear_1" [id=41, type=linear];
-"42 view_7" [id=42, type=view];
-"43 transpose_5" [id=43, type=transpose];
-"44 dropout_1" [id=44, type=dropout];
-"45 add_1" [id=45, type=add];
-"46 _param_constant10" [id=46, type=get_attr];
-"47 _param_constant11" [id=47, type=get_attr];
-"48 layer_norm_1" [id=48, type=layer_norm];
-"49 _param_constant12" [id=49, type=get_attr];
-"50 _param_constant13" [id=50, type=get_attr];
-"51 linear_2" [id=51, type=linear];
-"52 gelu" [id=52, type=gelu];
-"53 dropout_2" [id=53, type=dropout];
-"54 _param_constant14" [id=54, type=get_attr];
-"55 _param_constant15" [id=55, type=get_attr];
-"56 linear_3" [id=56, type=linear];
-"57 dropout_3" [id=57, type=dropout];
-"58 add_2" [id=58, type=add];
-"59 _param_constant16" [id=59, type=get_attr];
-"60 _param_constant17" [id=60, type=get_attr];
-"61 layer_norm_2" [id=61, type=layer_norm];
-"62 transpose_6" [id=62, type=transpose];
-"63 _param_constant18" [id=63, type=get_attr];
-"64 _param_constant19" [id=64, type=get_attr];
-"65 linear_4" [id=65, type=linear];
-"66 unflatten_1" [id=66, type=unflatten];
-"67 unsqueeze_1" [id=67, type=unsqueeze];
-"68 transpose_7" [id=68, type=transpose];
-"69 squeeze_1" [id=69, type=squeeze];
-"70 contiguous_1" [id=70, type=contiguous];
-"71 select_3" [id=71, type=select];
-"72 select_4" [id=72, type=select];
-"73 select_5" [id=73, type=select];
-"74 view_8" [id=74, type=view];
-"75 transpose_8" [id=75, type=transpose];
-"76 view_9" [id=76, type=view];
-"77 transpose_9" [id=77, type=transpose];
-"78 view_10" [id=78, type=view];
-"79 transpose_10" [id=79, type=transpose];
-"80 view_11" [id=80, type=view];
-"81 view_12" [id=81, type=view];
-"82 view_13" [id=82, type=view];
-"83 scaled_dot_product_attention_1" [id=83, type=scaled_dot_product_attention];
-"84 permute_2" [id=84, type=permute];
-"85 view_14" [id=85, type=view];
-"86 _param_constant20" [id=86, type=get_attr];
-"87 _param_constant21" [id=87, type=get_attr];
-"88 linear_5" [id=88, type=linear];
-"89 view_15" [id=89, type=view];
-"90 transpose_11" [id=90, type=transpose];
-"91 dropout_4" [id=91, type=dropout];
-"92 add_3" [id=92, type=add];
-"93 _param_constant22" [id=93, type=get_attr];
-"94 _param_constant23" [id=94, type=get_attr];
-"95 layer_norm_3" [id=95, type=layer_norm];
-"96 _param_constant24" [id=96, type=get_attr];
-"97 _param_constant25" [id=97, type=get_attr];
-"98 linear_6" [id=98, type=linear];
-"99 gelu_1" [id=99, type=gelu];
-"100 dropout_5" [id=100, type=dropout];
-"101 _param_constant26" [id=101, type=get_attr];
-"102 _param_constant27" [id=102, type=get_attr];
-"103 linear_7" [id=103, type=linear];
-"104 dropout_6" [id=104, type=dropout];
-"105 add_4" [id=105, type=add];
-"106 _param_constant28" [id=106, type=get_attr];
-"107 _param_constant29" [id=107, type=get_attr];
-"108 layer_norm_4" [id=108, type=layer_norm];
-"109 transpose_12" [id=109, type=transpose];
-"110 _param_constant30" [id=110, type=get_attr];
-"111 _param_constant31" [id=111, type=get_attr];
-"112 linear_8" [id=112, type=linear];
-"113 unflatten_2" [id=113, type=unflatten];
-"114 unsqueeze_2" [id=114, type=unsqueeze];
-"115 transpose_13" [id=115, type=transpose];
-"116 squeeze_2" [id=116, type=squeeze];
-"117 contiguous_2" [id=117, type=contiguous];
-"118 select_6" [id=118, type=select];
-"119 select_7" [id=119, type=select];
-"120 select_8" [id=120, type=select];
-"121 view_16" [id=121, type=view];
-"122 transpose_14" [id=122, type=transpose];
-"123 view_17" [id=123, type=view];
-"124 transpose_15" [id=124, type=transpose];
-"125 view_18" [id=125, type=view];
-"126 transpose_16" [id=126, type=transpose];
-"127 view_19" [id=127, type=view];
-"128 view_20" [id=128, type=view];
-"129 view_21" [id=129, type=view];
-"130 scaled_dot_product_attention_2" [id=130, type=scaled_dot_product_attention];
-"131 permute_3" [id=131, type=permute];
-"132 view_22" [id=132, type=view];
-"133 _param_constant32" [id=133, type=get_attr];
-"134 _param_constant33" [id=134, type=get_attr];
-"135 linear_9" [id=135, type=linear];
-"136 view_23" [id=136, type=view];
-"137 transpose_17" [id=137, type=transpose];
-"138 dropout_7" [id=138, type=dropout];
-"139 add_5" [id=139, type=add];
-"140 _param_constant34" [id=140, type=get_attr];
-"141 _param_constant35" [id=141, type=get_attr];
-"142 layer_norm_5" [id=142, type=layer_norm];
-"143 _param_constant36" [id=143, type=get_attr];
-"144 _param_constant37" [id=144, type=get_attr];
-"145 linear_10" [id=145, type=linear];
-"146 gelu_2" [id=146, type=gelu];
-"147 dropout_8" [id=147, type=dropout];
-"148 _param_constant38" [id=148, type=get_attr];
-"149 _param_constant39" [id=149, type=get_attr];
-"150 linear_11" [id=150, type=linear];
-"151 dropout_9" [id=151, type=dropout];
-"152 add_6" [id=152, type=add];
-"153 _param_constant40" [id=153, type=get_attr];
-"154 _param_constant41" [id=154, type=get_attr];
-"155 layer_norm_6" [id=155, type=layer_norm];
-"156 transpose_18" [id=156, type=transpose];
-"157 _param_constant42" [id=157, type=get_attr];
-"158 _param_constant43" [id=158, type=get_attr];
-"159 linear_12" [id=159, type=linear];
-"160 unflatten_3" [id=160, type=unflatten];
-"161 unsqueeze_3" [id=161, type=unsqueeze];
-"162 transpose_19" [id=162, type=transpose];
-"163 squeeze_3" [id=163, type=squeeze];
-"164 contiguous_3" [id=164, type=contiguous];
-"165 select_9" [id=165, type=select];
-"166 select_10" [id=166, type=select];
-"167 select_11" [id=167, type=select];
-"168 view_24" [id=168, type=view];
-"169 transpose_20" [id=169, type=transpose];
-"170 view_25" [id=170, type=view];
-"171 transpose_21" [id=171, type=transpose];
-"172 view_26" [id=172, type=view];
-"173 transpose_22" [id=173, type=transpose];
-"174 view_27" [id=174, type=view];
-"175 view_28" [id=175, type=view];
-"176 view_29" [id=176, type=view];
-"177 scaled_dot_product_attention_3" [id=177, type=scaled_dot_product_attention];
-"178 permute_4" [id=178, type=permute];
-"179 view_30" [id=179, type=view];
-"180 _param_constant44" [id=180, type=get_attr];
-"181 _param_constant45" [id=181, type=get_attr];
-"182 linear_13" [id=182, type=linear];
-"183 view_31" [id=183, type=view];
-"184 transpose_23" [id=184, type=transpose];
-"185 dropout_10" [id=185, type=dropout];
-"186 add_7" [id=186, type=add];
-"187 _param_constant46" [id=187, type=get_attr];
-"188 _param_constant47" [id=188, type=get_attr];
-"189 layer_norm_7" [id=189, type=layer_norm];
-"190 _param_constant48" [id=190, type=get_attr];
-"191 _param_constant49" [id=191, type=get_attr];
-"192 linear_14" [id=192, type=linear];
-"193 gelu_3" [id=193, type=gelu];
-"194 dropout_11" [id=194, type=dropout];
-"195 _param_constant50" [id=195, type=get_attr];
-"196 _param_constant51" [id=196, type=get_attr];
-"197 linear_15" [id=197, type=linear];
-"198 dropout_12" [id=198, type=dropout];
-"199 add_8" [id=199, type=add];
-"200 _param_constant52" [id=200, type=get_attr];
-"201 _param_constant53" [id=201, type=get_attr];
-"202 layer_norm_8" [id=202, type=layer_norm];
-"203 transpose_24" [id=203, type=transpose];
-"204 _param_constant54" [id=204, type=get_attr];
-"205 _param_constant55" [id=205, type=get_attr];
-"206 linear_16" [id=206, type=linear];
-"207 unflatten_4" [id=207, type=unflatten];
-"208 unsqueeze_4" [id=208, type=unsqueeze];
-"209 transpose_25" [id=209, type=transpose];
-"210 squeeze_4" [id=210, type=squeeze];
-"211 contiguous_4" [id=211, type=contiguous];
-"212 select_12" [id=212, type=select];
-"213 select_13" [id=213, type=select];
-"214 select_14" [id=214, type=select];
-"215 view_32" [id=215, type=view];
-"216 transpose_26" [id=216, type=transpose];
-"217 view_33" [id=217, type=view];
-"218 transpose_27" [id=218, type=transpose];
-"219 view_34" [id=219, type=view];
-"220 transpose_28" [id=220, type=transpose];
-"221 view_35" [id=221, type=view];
-"222 view_36" [id=222, type=view];
-"223 view_37" [id=223, type=view];
-"224 scaled_dot_product_attention_4" [id=224, type=scaled_dot_product_attention];
-"225 permute_5" [id=225, type=permute];
-"226 view_38" [id=226, type=view];
-"227 _param_constant56" [id=227, type=get_attr];
-"228 _param_constant57" [id=228, type=get_attr];
-"229 linear_17" [id=229, type=linear];
-"230 view_39" [id=230, type=view];
-"231 transpose_29" [id=231, type=transpose];
-"232 dropout_13" [id=232, type=dropout];
-"233 add_9" [id=233, type=add];
-"234 _param_constant58" [id=234, type=get_attr];
-"235 _param_constant59" [id=235, type=get_attr];
-"236 layer_norm_9" [id=236, type=layer_norm];
-"237 _param_constant60" [id=237, type=get_attr];
-"238 _param_constant61" [id=238, type=get_attr];
-"239 linear_18" [id=239, type=linear];
-"240 gelu_4" [id=240, type=gelu];
-"241 dropout_14" [id=241, type=dropout];
-"242 _param_constant62" [id=242, type=get_attr];
-"243 _param_constant63" [id=243, type=get_attr];
-"244 linear_19" [id=244, type=linear];
-"245 dropout_15" [id=245, type=dropout];
-"246 add_10" [id=246, type=add];
-"247 _param_constant64" [id=247, type=get_attr];
-"248 _param_constant65" [id=248, type=get_attr];
-"249 layer_norm_10" [id=249, type=layer_norm];
-"250 transpose_30" [id=250, type=transpose];
-"251 _param_constant66" [id=251, type=get_attr];
-"252 _param_constant67" [id=252, type=get_attr];
-"253 linear_20" [id=253, type=linear];
-"254 unflatten_5" [id=254, type=unflatten];
-"255 unsqueeze_5" [id=255, type=unsqueeze];
-"256 transpose_31" [id=256, type=transpose];
-"257 squeeze_5" [id=257, type=squeeze];
-"258 contiguous_5" [id=258, type=contiguous];
-"259 select_15" [id=259, type=select];
-"260 select_16" [id=260, type=select];
-"261 select_17" [id=261, type=select];
-"262 view_40" [id=262, type=view];
-"263 transpose_32" [id=263, type=transpose];
-"264 view_41" [id=264, type=view];
-"265 transpose_33" [id=265, type=transpose];
-"266 view_42" [id=266, type=view];
-"267 transpose_34" [id=267, type=transpose];
-"268 view_43" [id=268, type=view];
-"269 view_44" [id=269, type=view];
-"270 view_45" [id=270, type=view];
-"271 scaled_dot_product_attention_5" [id=271, type=scaled_dot_product_attention];
-"272 permute_6" [id=272, type=permute];
-"273 view_46" [id=273, type=view];
-"274 _param_constant68" [id=274, type=get_attr];
-"275 _param_constant69" [id=275, type=get_attr];
-"276 linear_21" [id=276, type=linear];
-"277 view_47" [id=277, type=view];
-"278 transpose_35" [id=278, type=transpose];
-"279 dropout_16" [id=279, type=dropout];
-"280 add_11" [id=280, type=add];
-"281 _param_constant70" [id=281, type=get_attr];
-"282 _param_constant71" [id=282, type=get_attr];
-"283 layer_norm_11" [id=283, type=layer_norm];
-"284 _param_constant72" [id=284, type=get_attr];
-"285 _param_constant73" [id=285, type=get_attr];
-"286 linear_22" [id=286, type=linear];
-"287 gelu_5" [id=287, type=gelu];
-"288 dropout_17" [id=288, type=dropout];
-"289 _param_constant74" [id=289, type=get_attr];
-"290 _param_constant75" [id=290, type=get_attr];
-"291 linear_23" [id=291, type=linear];
-"292 dropout_18" [id=292, type=dropout];
-"293 add_12" [id=293, type=add];
-"294 _param_constant76" [id=294, type=get_attr];
-"295 _param_constant77" [id=295, type=get_attr];
-"296 layer_norm_12" [id=296, type=layer_norm];
-"297 transpose_36" [id=297, type=transpose];
-"298 _param_constant78" [id=298, type=get_attr];
-"299 _param_constant79" [id=299, type=get_attr];
-"300 linear_24" [id=300, type=linear];
-"301 unflatten_6" [id=301, type=unflatten];
-"302 unsqueeze_6" [id=302, type=unsqueeze];
-"303 transpose_37" [id=303, type=transpose];
-"304 squeeze_6" [id=304, type=squeeze];
-"305 contiguous_6" [id=305, type=contiguous];
-"306 select_18" [id=306, type=select];
-"307 select_19" [id=307, type=select];
-"308 select_20" [id=308, type=select];
-"309 view_48" [id=309, type=view];
-"310 transpose_38" [id=310, type=transpose];
-"311 view_49" [id=311, type=view];
-"312 transpose_39" [id=312, type=transpose];
-"313 view_50" [id=313, type=view];
-"314 transpose_40" [id=314, type=transpose];
-"315 view_51" [id=315, type=view];
-"316 view_52" [id=316, type=view];
-"317 view_53" [id=317, type=view];
-"318 scaled_dot_product_attention_6" [id=318, type=scaled_dot_product_attention];
-"319 permute_7" [id=319, type=permute];
-"320 view_54" [id=320, type=view];
-"321 _param_constant80" [id=321, type=get_attr];
-"322 _param_constant81" [id=322, type=get_attr];
-"323 linear_25" [id=323, type=linear];
-"324 view_55" [id=324, type=view];
-"325 transpose_41" [id=325, type=transpose];
-"326 dropout_19" [id=326, type=dropout];
-"327 add_13" [id=327, type=add];
-"328 _param_constant82" [id=328, type=get_attr];
-"329 _param_constant83" [id=329, type=get_attr];
-"330 layer_norm_13" [id=330, type=layer_norm];
-"331 _param_constant84" [id=331, type=get_attr];
-"332 _param_constant85" [id=332, type=get_attr];
-"333 linear_26" [id=333, type=linear];
-"334 gelu_6" [id=334, type=gelu];
-"335 dropout_20" [id=335, type=dropout];
-"336 _param_constant86" [id=336, type=get_attr];
-"337 _param_constant87" [id=337, type=get_attr];
-"338 linear_27" [id=338, type=linear];
-"339 dropout_21" [id=339, type=dropout];
-"340 add_14" [id=340, type=add];
-"341 _param_constant88" [id=341, type=get_attr];
-"342 _param_constant89" [id=342, type=get_attr];
-"343 layer_norm_14" [id=343, type=layer_norm];
-"344 transpose_42" [id=344, type=transpose];
-"345 _param_constant90" [id=345, type=get_attr];
-"346 _param_constant91" [id=346, type=get_attr];
-"347 linear_28" [id=347, type=linear];
-"348 unflatten_7" [id=348, type=unflatten];
-"349 unsqueeze_7" [id=349, type=unsqueeze];
-"350 transpose_43" [id=350, type=transpose];
-"351 squeeze_7" [id=351, type=squeeze];
-"352 contiguous_7" [id=352, type=contiguous];
-"353 select_21" [id=353, type=select];
-"354 select_22" [id=354, type=select];
-"355 select_23" [id=355, type=select];
-"356 view_56" [id=356, type=view];
-"357 transpose_44" [id=357, type=transpose];
-"358 view_57" [id=358, type=view];
-"359 transpose_45" [id=359, type=transpose];
-"360 view_58" [id=360, type=view];
-"361 transpose_46" [id=361, type=transpose];
-"362 view_59" [id=362, type=view];
-"363 view_60" [id=363, type=view];
-"364 view_61" [id=364, type=view];
-"365 scaled_dot_product_attention_7" [id=365, type=scaled_dot_product_attention];
-"366 permute_8" [id=366, type=permute];
-"367 view_62" [id=367, type=view];
-"368 _param_constant92" [id=368, type=get_attr];
-"369 _param_constant93" [id=369, type=get_attr];
-"370 linear_29" [id=370, type=linear];
-"371 view_63" [id=371, type=view];
-"372 transpose_47" [id=372, type=transpose];
-"373 dropout_22" [id=373, type=dropout];
-"374 add_15" [id=374, type=add];
-"375 _param_constant94" [id=375, type=get_attr];
-"376 _param_constant95" [id=376, type=get_attr];
-"377 layer_norm_15" [id=377, type=layer_norm];
-"378 _param_constant96" [id=378, type=get_attr];
-"379 _param_constant97" [id=379, type=get_attr];
-"380 linear_30" [id=380, type=linear];
-"381 gelu_7" [id=381, type=gelu];
-"382 dropout_23" [id=382, type=dropout];
-"383 _param_constant98" [id=383, type=get_attr];
-"384 _param_constant99" [id=384, type=get_attr];
-"385 linear_31" [id=385, type=linear];
-"386 dropout_24" [id=386, type=dropout];
-"387 add_16" [id=387, type=add];
-"388 _param_constant100" [id=388, type=get_attr];
-"389 _param_constant101" [id=389, type=get_attr];
-"390 layer_norm_16" [id=390, type=layer_norm];
-"391 transpose_48" [id=391, type=transpose];
-"392 _param_constant102" [id=392, type=get_attr];
-"393 _param_constant103" [id=393, type=get_attr];
-"394 linear_32" [id=394, type=linear];
-"395 unflatten_8" [id=395, type=unflatten];
-"396 unsqueeze_8" [id=396, type=unsqueeze];
-"397 transpose_49" [id=397, type=transpose];
-"398 squeeze_8" [id=398, type=squeeze];
-"399 contiguous_8" [id=399, type=contiguous];
-"400 select_24" [id=400, type=select];
-"401 select_25" [id=401, type=select];
-"402 select_26" [id=402, type=select];
-"403 view_64" [id=403, type=view];
-"404 transpose_50" [id=404, type=transpose];
-"405 view_65" [id=405, type=view];
-"406 transpose_51" [id=406, type=transpose];
-"407 view_66" [id=407, type=view];
-"408 transpose_52" [id=408, type=transpose];
-"409 view_67" [id=409, type=view];
-"410 view_68" [id=410, type=view];
-"411 view_69" [id=411, type=view];
-"412 scaled_dot_product_attention_8" [id=412, type=scaled_dot_product_attention];
-"413 permute_9" [id=413, type=permute];
-"414 view_70" [id=414, type=view];
-"415 _param_constant104" [id=415, type=get_attr];
-"416 _param_constant105" [id=416, type=get_attr];
-"417 linear_33" [id=417, type=linear];
-"418 view_71" [id=418, type=view];
-"419 transpose_53" [id=419, type=transpose];
-"420 dropout_25" [id=420, type=dropout];
-"421 add_17" [id=421, type=add];
-"422 _param_constant106" [id=422, type=get_attr];
-"423 _param_constant107" [id=423, type=get_attr];
-"424 layer_norm_17" [id=424, type=layer_norm];
-"425 _param_constant108" [id=425, type=get_attr];
-"426 _param_constant109" [id=426, type=get_attr];
-"427 linear_34" [id=427, type=linear];
-"428 gelu_8" [id=428, type=gelu];
-"429 dropout_26" [id=429, type=dropout];
-"430 _param_constant110" [id=430, type=get_attr];
-"431 _param_constant111" [id=431, type=get_attr];
-"432 linear_35" [id=432, type=linear];
-"433 dropout_27" [id=433, type=dropout];
-"434 add_18" [id=434, type=add];
-"435 _param_constant112" [id=435, type=get_attr];
-"436 _param_constant113" [id=436, type=get_attr];
-"437 layer_norm_18" [id=437, type=layer_norm];
-"438 transpose_54" [id=438, type=transpose];
-"439 _param_constant114" [id=439, type=get_attr];
-"440 _param_constant115" [id=440, type=get_attr];
-"441 linear_36" [id=441, type=linear];
-"442 unflatten_9" [id=442, type=unflatten];
-"443 unsqueeze_9" [id=443, type=unsqueeze];
-"444 transpose_55" [id=444, type=transpose];
-"445 squeeze_9" [id=445, type=squeeze];
-"446 contiguous_9" [id=446, type=contiguous];
-"447 select_27" [id=447, type=select];
-"448 select_28" [id=448, type=select];
-"449 select_29" [id=449, type=select];
-"450 view_72" [id=450, type=view];
-"451 transpose_56" [id=451, type=transpose];
-"452 view_73" [id=452, type=view];
-"453 transpose_57" [id=453, type=transpose];
-"454 view_74" [id=454, type=view];
-"455 transpose_58" [id=455, type=transpose];
-"456 view_75" [id=456, type=view];
-"457 view_76" [id=457, type=view];
-"458 view_77" [id=458, type=view];
-"459 scaled_dot_product_attention_9" [id=459, type=scaled_dot_product_attention];
-"460 permute_10" [id=460, type=permute];
-"461 view_78" [id=461, type=view];
-"462 _param_constant116" [id=462, type=get_attr];
-"463 _param_constant117" [id=463, type=get_attr];
-"464 linear_37" [id=464, type=linear];
-"465 view_79" [id=465, type=view];
-"466 transpose_59" [id=466, type=transpose];
-"467 dropout_28" [id=467, type=dropout];
-"468 add_19" [id=468, type=add];
-"469 _param_constant118" [id=469, type=get_attr];
-"470 _param_constant119" [id=470, type=get_attr];
-"471 layer_norm_19" [id=471, type=layer_norm];
-"472 _param_constant120" [id=472, type=get_attr];
-"473 _param_constant121" [id=473, type=get_attr];
-"474 linear_38" [id=474, type=linear];
-"475 gelu_9" [id=475, type=gelu];
-"476 dropout_29" [id=476, type=dropout];
-"477 _param_constant122" [id=477, type=get_attr];
-"478 _param_constant123" [id=478, type=get_attr];
-"479 linear_39" [id=479, type=linear];
-"480 dropout_30" [id=480, type=dropout];
-"481 add_20" [id=481, type=add];
-"482 _param_constant124" [id=482, type=get_attr];
-"483 _param_constant125" [id=483, type=get_attr];
-"484 layer_norm_20" [id=484, type=layer_norm];
-"485 transpose_60" [id=485, type=transpose];
-"486 _param_constant126" [id=486, type=get_attr];
-"487 _param_constant127" [id=487, type=get_attr];
-"488 linear_40" [id=488, type=linear];
-"489 unflatten_10" [id=489, type=unflatten];
-"490 unsqueeze_10" [id=490, type=unsqueeze];
-"491 transpose_61" [id=491, type=transpose];
-"492 squeeze_10" [id=492, type=squeeze];
-"493 contiguous_10" [id=493, type=contiguous];
-"494 select_30" [id=494, type=select];
-"495 select_31" [id=495, type=select];
-"496 select_32" [id=496, type=select];
-"497 view_80" [id=497, type=view];
-"498 transpose_62" [id=498, type=transpose];
-"499 view_81" [id=499, type=view];
-"500 transpose_63" [id=500, type=transpose];
-"501 view_82" [id=501, type=view];
-"502 transpose_64" [id=502, type=transpose];
-"503 view_83" [id=503, type=view];
-"504 view_84" [id=504, type=view];
-"505 view_85" [id=505, type=view];
-"506 scaled_dot_product_attention_10" [id=506, type=scaled_dot_product_attention];
-"507 permute_11" [id=507, type=permute];
-"508 view_86" [id=508, type=view];
-"509 _param_constant128" [id=509, type=get_attr];
-"510 _param_constant129" [id=510, type=get_attr];
-"511 linear_41" [id=511, type=linear];
-"512 view_87" [id=512, type=view];
-"513 transpose_65" [id=513, type=transpose];
-"514 dropout_31" [id=514, type=dropout];
-"515 add_21" [id=515, type=add];
-"516 _param_constant130" [id=516, type=get_attr];
-"517 _param_constant131" [id=517, type=get_attr];
-"518 layer_norm_21" [id=518, type=layer_norm];
-"519 _param_constant132" [id=519, type=get_attr];
-"520 _param_constant133" [id=520, type=get_attr];
-"521 linear_42" [id=521, type=linear];
-"522 gelu_10" [id=522, type=gelu];
-"523 dropout_32" [id=523, type=dropout];
-"524 _param_constant134" [id=524, type=get_attr];
-"525 _param_constant135" [id=525, type=get_attr];
-"526 linear_43" [id=526, type=linear];
-"527 dropout_33" [id=527, type=dropout];
-"528 add_22" [id=528, type=add];
-"529 _param_constant136" [id=529, type=get_attr];
-"530 _param_constant137" [id=530, type=get_attr];
-"531 layer_norm_22" [id=531, type=layer_norm];
-"532 transpose_66" [id=532, type=transpose];
-"533 _param_constant138" [id=533, type=get_attr];
-"534 _param_constant139" [id=534, type=get_attr];
-"535 linear_44" [id=535, type=linear];
-"536 unflatten_11" [id=536, type=unflatten];
-"537 unsqueeze_11" [id=537, type=unsqueeze];
-"538 transpose_67" [id=538, type=transpose];
-"539 squeeze_11" [id=539, type=squeeze];
-"540 contiguous_11" [id=540, type=contiguous];
-"541 select_33" [id=541, type=select];
-"542 select_34" [id=542, type=select];
-"543 select_35" [id=543, type=select];
-"544 view_88" [id=544, type=view];
-"545 transpose_68" [id=545, type=transpose];
-"546 view_89" [id=546, type=view];
-"547 transpose_69" [id=547, type=transpose];
-"548 view_90" [id=548, type=view];
-"549 transpose_70" [id=549, type=transpose];
-"550 view_91" [id=550, type=view];
-"551 view_92" [id=551, type=view];
-"552 view_93" [id=552, type=view];
-"553 scaled_dot_product_attention_11" [id=553, type=scaled_dot_product_attention];
-"554 permute_12" [id=554, type=permute];
-"555 view_94" [id=555, type=view];
-"556 _param_constant140" [id=556, type=get_attr];
-"557 _param_constant141" [id=557, type=get_attr];
-"558 linear_45" [id=558, type=linear];
-"559 view_95" [id=559, type=view];
-"560 transpose_71" [id=560, type=transpose];
-"561 dropout_34" [id=561, type=dropout];
-"562 add_23" [id=562, type=add];
-"563 _param_constant142" [id=563, type=get_attr];
-"564 _param_constant143" [id=564, type=get_attr];
-"565 layer_norm_23" [id=565, type=layer_norm];
-"566 _param_constant144" [id=566, type=get_attr];
-"567 _param_constant145" [id=567, type=get_attr];
-"568 linear_46" [id=568, type=linear];
-"569 gelu_11" [id=569, type=gelu];
-"570 dropout_35" [id=570, type=dropout];
-"571 _param_constant146" [id=571, type=get_attr];
-"572 _param_constant147" [id=572, type=get_attr];
-"573 linear_47" [id=573, type=linear];
-"574 dropout_36" [id=574, type=dropout];
-"575 add_24" [id=575, type=add];
-"576 _param_constant148" [id=576, type=get_attr];
-"577 _param_constant149" [id=577, type=get_attr];
-"578 layer_norm_24" [id=578, type=layer_norm];
-"579 slice_1" [id=579, type=slice];
-"580 select_36" [id=580, type=select];
-"581 _param_constant150" [id=581, type=get_attr];
-"582 _param_constant151" [id=582, type=get_attr];
+"0 class_token" [id=0, type=get_attr];
+"1 encoder_pos_embedding" [id=1, type=get_attr];
+"2 conv_proj_weight" [id=2, type=get_attr];
+"3 conv_proj_bias" [id=3, type=get_attr];
+"4 encoder_layers_encoder_layer_0_ln_1_weight" [id=4, type=get_attr];
+"5 encoder_layers_encoder_layer_0_ln_1_bias" [id=5, type=get_attr];
+"6 encoder_layers_encoder_layer_0_self_attention_in_proj_weight" [id=6, type=get_attr];
+"7 encoder_layers_encoder_layer_0_self_attention_in_proj_bias" [id=7, type=get_attr];
+"8 encoder_layers_encoder_layer_0_self_attention_out_proj_weight" [id=8, type=get_attr];
+"9 encoder_layers_encoder_layer_0_self_attention_out_proj_bias" [id=9, type=get_attr];
+"10 encoder_layers_encoder_layer_0_ln_2_weight" [id=10, type=get_attr];
+"11 encoder_layers_encoder_layer_0_ln_2_bias" [id=11, type=get_attr];
+"12 encoder_layers_encoder_layer_0_mlp_0_weight" [id=12, type=get_attr];
+"13 encoder_layers_encoder_layer_0_mlp_0_bias" [id=13, type=get_attr];
+"14 encoder_layers_encoder_layer_0_mlp_3_weight" [id=14, type=get_attr];
+"15 encoder_layers_encoder_layer_0_mlp_3_bias" [id=15, type=get_attr];
+"16 encoder_layers_encoder_layer_1_ln_1_weight" [id=16, type=get_attr];
+"17 encoder_layers_encoder_layer_1_ln_1_bias" [id=17, type=get_attr];
+"18 encoder_layers_encoder_layer_1_self_attention_in_proj_weight" [id=18, type=get_attr];
+"19 encoder_layers_encoder_layer_1_self_attention_in_proj_bias" [id=19, type=get_attr];
+"20 encoder_layers_encoder_layer_1_self_attention_out_proj_weight" [id=20, type=get_attr];
+"21 encoder_layers_encoder_layer_1_self_attention_out_proj_bias" [id=21, type=get_attr];
+"22 encoder_layers_encoder_layer_1_ln_2_weight" [id=22, type=get_attr];
+"23 encoder_layers_encoder_layer_1_ln_2_bias" [id=23, type=get_attr];
+"24 encoder_layers_encoder_layer_1_mlp_0_weight" [id=24, type=get_attr];
+"25 encoder_layers_encoder_layer_1_mlp_0_bias" [id=25, type=get_attr];
+"26 encoder_layers_encoder_layer_1_mlp_3_weight" [id=26, type=get_attr];
+"27 encoder_layers_encoder_layer_1_mlp_3_bias" [id=27, type=get_attr];
+"28 encoder_layers_encoder_layer_2_ln_1_weight" [id=28, type=get_attr];
+"29 encoder_layers_encoder_layer_2_ln_1_bias" [id=29, type=get_attr];
+"30 encoder_layers_encoder_layer_2_self_attention_in_proj_weight" [id=30, type=get_attr];
+"31 encoder_layers_encoder_layer_2_self_attention_in_proj_bias" [id=31, type=get_attr];
+"32 encoder_layers_encoder_layer_2_self_attention_out_proj_weight" [id=32, type=get_attr];
+"33 encoder_layers_encoder_layer_2_self_attention_out_proj_bias" [id=33, type=get_attr];
+"34 encoder_layers_encoder_layer_2_ln_2_weight" [id=34, type=get_attr];
+"35 encoder_layers_encoder_layer_2_ln_2_bias" [id=35, type=get_attr];
+"36 encoder_layers_encoder_layer_2_mlp_0_weight" [id=36, type=get_attr];
+"37 encoder_layers_encoder_layer_2_mlp_0_bias" [id=37, type=get_attr];
+"38 encoder_layers_encoder_layer_2_mlp_3_weight" [id=38, type=get_attr];
+"39 encoder_layers_encoder_layer_2_mlp_3_bias" [id=39, type=get_attr];
+"40 encoder_layers_encoder_layer_3_ln_1_weight" [id=40, type=get_attr];
+"41 encoder_layers_encoder_layer_3_ln_1_bias" [id=41, type=get_attr];
+"42 encoder_layers_encoder_layer_3_self_attention_in_proj_weight" [id=42, type=get_attr];
+"43 encoder_layers_encoder_layer_3_self_attention_in_proj_bias" [id=43, type=get_attr];
+"44 encoder_layers_encoder_layer_3_self_attention_out_proj_weight" [id=44, type=get_attr];
+"45 encoder_layers_encoder_layer_3_self_attention_out_proj_bias" [id=45, type=get_attr];
+"46 encoder_layers_encoder_layer_3_ln_2_weight" [id=46, type=get_attr];
+"47 encoder_layers_encoder_layer_3_ln_2_bias" [id=47, type=get_attr];
+"48 encoder_layers_encoder_layer_3_mlp_0_weight" [id=48, type=get_attr];
+"49 encoder_layers_encoder_layer_3_mlp_0_bias" [id=49, type=get_attr];
+"50 encoder_layers_encoder_layer_3_mlp_3_weight" [id=50, type=get_attr];
+"51 encoder_layers_encoder_layer_3_mlp_3_bias" [id=51, type=get_attr];
+"52 encoder_layers_encoder_layer_4_ln_1_weight" [id=52, type=get_attr];
+"53 encoder_layers_encoder_layer_4_ln_1_bias" [id=53, type=get_attr];
+"54 encoder_layers_encoder_layer_4_self_attention_in_proj_weight" [id=54, type=get_attr];
+"55 encoder_layers_encoder_layer_4_self_attention_in_proj_bias" [id=55, type=get_attr];
+"56 encoder_layers_encoder_layer_4_self_attention_out_proj_weight" [id=56, type=get_attr];
+"57 encoder_layers_encoder_layer_4_self_attention_out_proj_bias" [id=57, type=get_attr];
+"58 encoder_layers_encoder_layer_4_ln_2_weight" [id=58, type=get_attr];
+"59 encoder_layers_encoder_layer_4_ln_2_bias" [id=59, type=get_attr];
+"60 encoder_layers_encoder_layer_4_mlp_0_weight" [id=60, type=get_attr];
+"61 encoder_layers_encoder_layer_4_mlp_0_bias" [id=61, type=get_attr];
+"62 encoder_layers_encoder_layer_4_mlp_3_weight" [id=62, type=get_attr];
+"63 encoder_layers_encoder_layer_4_mlp_3_bias" [id=63, type=get_attr];
+"64 encoder_layers_encoder_layer_5_ln_1_weight" [id=64, type=get_attr];
+"65 encoder_layers_encoder_layer_5_ln_1_bias" [id=65, type=get_attr];
+"66 encoder_layers_encoder_layer_5_self_attention_in_proj_weight" [id=66, type=get_attr];
+"67 encoder_layers_encoder_layer_5_self_attention_in_proj_bias" [id=67, type=get_attr];
+"68 encoder_layers_encoder_layer_5_self_attention_out_proj_weight" [id=68, type=get_attr];
+"69 encoder_layers_encoder_layer_5_self_attention_out_proj_bias" [id=69, type=get_attr];
+"70 encoder_layers_encoder_layer_5_ln_2_weight" [id=70, type=get_attr];
+"71 encoder_layers_encoder_layer_5_ln_2_bias" [id=71, type=get_attr];
+"72 encoder_layers_encoder_layer_5_mlp_0_weight" [id=72, type=get_attr];
+"73 encoder_layers_encoder_layer_5_mlp_0_bias" [id=73, type=get_attr];
+"74 encoder_layers_encoder_layer_5_mlp_3_weight" [id=74, type=get_attr];
+"75 encoder_layers_encoder_layer_5_mlp_3_bias" [id=75, type=get_attr];
+"76 encoder_layers_encoder_layer_6_ln_1_weight" [id=76, type=get_attr];
+"77 encoder_layers_encoder_layer_6_ln_1_bias" [id=77, type=get_attr];
+"78 encoder_layers_encoder_layer_6_self_attention_in_proj_weight" [id=78, type=get_attr];
+"79 encoder_layers_encoder_layer_6_self_attention_in_proj_bias" [id=79, type=get_attr];
+"80 encoder_layers_encoder_layer_6_self_attention_out_proj_weight" [id=80, type=get_attr];
+"81 encoder_layers_encoder_layer_6_self_attention_out_proj_bias" [id=81, type=get_attr];
+"82 encoder_layers_encoder_layer_6_ln_2_weight" [id=82, type=get_attr];
+"83 encoder_layers_encoder_layer_6_ln_2_bias" [id=83, type=get_attr];
+"84 encoder_layers_encoder_layer_6_mlp_0_weight" [id=84, type=get_attr];
+"85 encoder_layers_encoder_layer_6_mlp_0_bias" [id=85, type=get_attr];
+"86 encoder_layers_encoder_layer_6_mlp_3_weight" [id=86, type=get_attr];
+"87 encoder_layers_encoder_layer_6_mlp_3_bias" [id=87, type=get_attr];
+"88 encoder_layers_encoder_layer_7_ln_1_weight" [id=88, type=get_attr];
+"89 encoder_layers_encoder_layer_7_ln_1_bias" [id=89, type=get_attr];
+"90 encoder_layers_encoder_layer_7_self_attention_in_proj_weight" [id=90, type=get_attr];
+"91 encoder_layers_encoder_layer_7_self_attention_in_proj_bias" [id=91, type=get_attr];
+"92 encoder_layers_encoder_layer_7_self_attention_out_proj_weight" [id=92, type=get_attr];
+"93 encoder_layers_encoder_layer_7_self_attention_out_proj_bias" [id=93, type=get_attr];
+"94 encoder_layers_encoder_layer_7_ln_2_weight" [id=94, type=get_attr];
+"95 encoder_layers_encoder_layer_7_ln_2_bias" [id=95, type=get_attr];
+"96 encoder_layers_encoder_layer_7_mlp_0_weight" [id=96, type=get_attr];
+"97 encoder_layers_encoder_layer_7_mlp_0_bias" [id=97, type=get_attr];
+"98 encoder_layers_encoder_layer_7_mlp_3_weight" [id=98, type=get_attr];
+"99 encoder_layers_encoder_layer_7_mlp_3_bias" [id=99, type=get_attr];
+"100 encoder_layers_encoder_layer_8_ln_1_weight" [id=100, type=get_attr];
+"101 encoder_layers_encoder_layer_8_ln_1_bias" [id=101, type=get_attr];
+"102 encoder_layers_encoder_layer_8_self_attention_in_proj_weight" [id=102, type=get_attr];
+"103 encoder_layers_encoder_layer_8_self_attention_in_proj_bias" [id=103, type=get_attr];
+"104 encoder_layers_encoder_layer_8_self_attention_out_proj_weight" [id=104, type=get_attr];
+"105 encoder_layers_encoder_layer_8_self_attention_out_proj_bias" [id=105, type=get_attr];
+"106 encoder_layers_encoder_layer_8_ln_2_weight" [id=106, type=get_attr];
+"107 encoder_layers_encoder_layer_8_ln_2_bias" [id=107, type=get_attr];
+"108 encoder_layers_encoder_layer_8_mlp_0_weight" [id=108, type=get_attr];
+"109 encoder_layers_encoder_layer_8_mlp_0_bias" [id=109, type=get_attr];
+"110 encoder_layers_encoder_layer_8_mlp_3_weight" [id=110, type=get_attr];
+"111 encoder_layers_encoder_layer_8_mlp_3_bias" [id=111, type=get_attr];
+"112 encoder_layers_encoder_layer_9_ln_1_weight" [id=112, type=get_attr];
+"113 encoder_layers_encoder_layer_9_ln_1_bias" [id=113, type=get_attr];
+"114 encoder_layers_encoder_layer_9_self_attention_in_proj_weight" [id=114, type=get_attr];
+"115 encoder_layers_encoder_layer_9_self_attention_in_proj_bias" [id=115, type=get_attr];
+"116 encoder_layers_encoder_layer_9_self_attention_out_proj_weight" [id=116, type=get_attr];
+"117 encoder_layers_encoder_layer_9_self_attention_out_proj_bias" [id=117, type=get_attr];
+"118 encoder_layers_encoder_layer_9_ln_2_weight" [id=118, type=get_attr];
+"119 encoder_layers_encoder_layer_9_ln_2_bias" [id=119, type=get_attr];
+"120 encoder_layers_encoder_layer_9_mlp_0_weight" [id=120, type=get_attr];
+"121 encoder_layers_encoder_layer_9_mlp_0_bias" [id=121, type=get_attr];
+"122 encoder_layers_encoder_layer_9_mlp_3_weight" [id=122, type=get_attr];
+"123 encoder_layers_encoder_layer_9_mlp_3_bias" [id=123, type=get_attr];
+"124 encoder_layers_encoder_layer_10_ln_1_weight" [id=124, type=get_attr];
+"125 encoder_layers_encoder_layer_10_ln_1_bias" [id=125, type=get_attr];
+"126 encoder_layers_encoder_layer_10_self_attention_in_proj_weight" [id=126, type=get_attr];
+"127 encoder_layers_encoder_layer_10_self_attention_in_proj_bias" [id=127, type=get_attr];
+"128 encoder_layers_encoder_layer_10_self_attention_out_proj_weight" [id=128, type=get_attr];
+"129 encoder_layers_encoder_layer_10_self_attention_out_proj_bias" [id=129, type=get_attr];
+"130 encoder_layers_encoder_layer_10_ln_2_weight" [id=130, type=get_attr];
+"131 encoder_layers_encoder_layer_10_ln_2_bias" [id=131, type=get_attr];
+"132 encoder_layers_encoder_layer_10_mlp_0_weight" [id=132, type=get_attr];
+"133 encoder_layers_encoder_layer_10_mlp_0_bias" [id=133, type=get_attr];
+"134 encoder_layers_encoder_layer_10_mlp_3_weight" [id=134, type=get_attr];
+"135 encoder_layers_encoder_layer_10_mlp_3_bias" [id=135, type=get_attr];
+"136 encoder_layers_encoder_layer_11_ln_1_weight" [id=136, type=get_attr];
+"137 encoder_layers_encoder_layer_11_ln_1_bias" [id=137, type=get_attr];
+"138 encoder_layers_encoder_layer_11_self_attention_in_proj_weight" [id=138, type=get_attr];
+"139 encoder_layers_encoder_layer_11_self_attention_in_proj_bias" [id=139, type=get_attr];
+"140 encoder_layers_encoder_layer_11_self_attention_out_proj_weight" [id=140, type=get_attr];
+"141 encoder_layers_encoder_layer_11_self_attention_out_proj_bias" [id=141, type=get_attr];
+"142 encoder_layers_encoder_layer_11_ln_2_weight" [id=142, type=get_attr];
+"143 encoder_layers_encoder_layer_11_ln_2_bias" [id=143, type=get_attr];
+"144 encoder_layers_encoder_layer_11_mlp_0_weight" [id=144, type=get_attr];
+"145 encoder_layers_encoder_layer_11_mlp_0_bias" [id=145, type=get_attr];
+"146 encoder_layers_encoder_layer_11_mlp_3_weight" [id=146, type=get_attr];
+"147 encoder_layers_encoder_layer_11_mlp_3_bias" [id=147, type=get_attr];
+"148 encoder_ln_weight" [id=148, type=get_attr];
+"149 encoder_ln_bias" [id=149, type=get_attr];
+"150 heads_head_weight" [id=150, type=get_attr];
+"151 heads_head_bias" [id=151, type=get_attr];
+"152 x" [id=152, type=input];
+"153 conv2d" [id=153, type=conv2d];
+"154 reshape" [id=154, type=reshape];
+"155 permute" [id=155, type=permute];
+"156 expand" [id=156, type=expand];
+"157 cat" [id=157, type=cat];
+"158 add" [id=158, type=add];
+"159 dropout" [id=159, type=dropout];
+"160 layer_norm" [id=160, type=layer_norm];
+"161 transpose" [id=161, type=transpose];
+"162 linear" [id=162, type=linear];
+"163 unflatten" [id=163, type=unflatten];
+"164 unsqueeze" [id=164, type=unsqueeze];
+"165 transpose_1" [id=165, type=transpose];
+"166 squeeze" [id=166, type=squeeze];
+"167 contiguous" [id=167, type=contiguous];
+"168 select" [id=168, type=select];
+"169 select_1" [id=169, type=select];
+"170 select_2" [id=170, type=select];
+"171 view" [id=171, type=view];
+"172 transpose_2" [id=172, type=transpose];
+"173 view_1" [id=173, type=view];
+"174 transpose_3" [id=174, type=transpose];
+"175 view_2" [id=175, type=view];
+"176 transpose_4" [id=176, type=transpose];
+"177 view_3" [id=177, type=view];
+"178 view_4" [id=178, type=view];
+"179 view_5" [id=179, type=view];
+"180 scaled_dot_product_attention" [id=180, type=scaled_dot_product_attention];
+"181 permute_1" [id=181, type=permute];
+"182 view_6" [id=182, type=view];
+"183 linear_1" [id=183, type=linear];
+"184 view_7" [id=184, type=view];
+"185 transpose_5" [id=185, type=transpose];
+"186 dropout_1" [id=186, type=dropout];
+"187 add_1" [id=187, type=add];
+"188 layer_norm_1" [id=188, type=layer_norm];
+"189 linear_2" [id=189, type=linear];
+"190 gelu" [id=190, type=gelu];
+"191 dropout_2" [id=191, type=dropout];
+"192 linear_3" [id=192, type=linear];
+"193 dropout_3" [id=193, type=dropout];
+"194 add_2" [id=194, type=add];
+"195 layer_norm_2" [id=195, type=layer_norm];
+"196 transpose_6" [id=196, type=transpose];
+"197 linear_4" [id=197, type=linear];
+"198 unflatten_1" [id=198, type=unflatten];
+"199 unsqueeze_1" [id=199, type=unsqueeze];
+"200 transpose_7" [id=200, type=transpose];
+"201 squeeze_1" [id=201, type=squeeze];
+"202 contiguous_1" [id=202, type=contiguous];
+"203 select_3" [id=203, type=select];
+"204 select_4" [id=204, type=select];
+"205 select_5" [id=205, type=select];
+"206 view_8" [id=206, type=view];
+"207 transpose_8" [id=207, type=transpose];
+"208 view_9" [id=208, type=view];
+"209 transpose_9" [id=209, type=transpose];
+"210 view_10" [id=210, type=view];
+"211 transpose_10" [id=211, type=transpose];
+"212 view_11" [id=212, type=view];
+"213 view_12" [id=213, type=view];
+"214 view_13" [id=214, type=view];
+"215 scaled_dot_product_attention_1" [id=215, type=scaled_dot_product_attention];
+"216 permute_2" [id=216, type=permute];
+"217 view_14" [id=217, type=view];
+"218 linear_5" [id=218, type=linear];
+"219 view_15" [id=219, type=view];
+"220 transpose_11" [id=220, type=transpose];
+"221 dropout_4" [id=221, type=dropout];
+"222 add_3" [id=222, type=add];
+"223 layer_norm_3" [id=223, type=layer_norm];
+"224 linear_6" [id=224, type=linear];
+"225 gelu_1" [id=225, type=gelu];
+"226 dropout_5" [id=226, type=dropout];
+"227 linear_7" [id=227, type=linear];
+"228 dropout_6" [id=228, type=dropout];
+"229 add_4" [id=229, type=add];
+"230 layer_norm_4" [id=230, type=layer_norm];
+"231 transpose_12" [id=231, type=transpose];
+"232 linear_8" [id=232, type=linear];
+"233 unflatten_2" [id=233, type=unflatten];
+"234 unsqueeze_2" [id=234, type=unsqueeze];
+"235 transpose_13" [id=235, type=transpose];
+"236 squeeze_2" [id=236, type=squeeze];
+"237 contiguous_2" [id=237, type=contiguous];
+"238 select_6" [id=238, type=select];
+"239 select_7" [id=239, type=select];
+"240 select_8" [id=240, type=select];
+"241 view_16" [id=241, type=view];
+"242 transpose_14" [id=242, type=transpose];
+"243 view_17" [id=243, type=view];
+"244 transpose_15" [id=244, type=transpose];
+"245 view_18" [id=245, type=view];
+"246 transpose_16" [id=246, type=transpose];
+"247 view_19" [id=247, type=view];
+"248 view_20" [id=248, type=view];
+"249 view_21" [id=249, type=view];
+"250 scaled_dot_product_attention_2" [id=250, type=scaled_dot_product_attention];
+"251 permute_3" [id=251, type=permute];
+"252 view_22" [id=252, type=view];
+"253 linear_9" [id=253, type=linear];
+"254 view_23" [id=254, type=view];
+"255 transpose_17" [id=255, type=transpose];
+"256 dropout_7" [id=256, type=dropout];
+"257 add_5" [id=257, type=add];
+"258 layer_norm_5" [id=258, type=layer_norm];
+"259 linear_10" [id=259, type=linear];
+"260 gelu_2" [id=260, type=gelu];
+"261 dropout_8" [id=261, type=dropout];
+"262 linear_11" [id=262, type=linear];
+"263 dropout_9" [id=263, type=dropout];
+"264 add_6" [id=264, type=add];
+"265 layer_norm_6" [id=265, type=layer_norm];
+"266 transpose_18" [id=266, type=transpose];
+"267 linear_12" [id=267, type=linear];
+"268 unflatten_3" [id=268, type=unflatten];
+"269 unsqueeze_3" [id=269, type=unsqueeze];
+"270 transpose_19" [id=270, type=transpose];
+"271 squeeze_3" [id=271, type=squeeze];
+"272 contiguous_3" [id=272, type=contiguous];
+"273 select_9" [id=273, type=select];
+"274 select_10" [id=274, type=select];
+"275 select_11" [id=275, type=select];
+"276 view_24" [id=276, type=view];
+"277 transpose_20" [id=277, type=transpose];
+"278 view_25" [id=278, type=view];
+"279 transpose_21" [id=279, type=transpose];
+"280 view_26" [id=280, type=view];
+"281 transpose_22" [id=281, type=transpose];
+"282 view_27" [id=282, type=view];
+"283 view_28" [id=283, type=view];
+"284 view_29" [id=284, type=view];
+"285 scaled_dot_product_attention_3" [id=285, type=scaled_dot_product_attention];
+"286 permute_4" [id=286, type=permute];
+"287 view_30" [id=287, type=view];
+"288 linear_13" [id=288, type=linear];
+"289 view_31" [id=289, type=view];
+"290 transpose_23" [id=290, type=transpose];
+"291 dropout_10" [id=291, type=dropout];
+"292 add_7" [id=292, type=add];
+"293 layer_norm_7" [id=293, type=layer_norm];
+"294 linear_14" [id=294, type=linear];
+"295 gelu_3" [id=295, type=gelu];
+"296 dropout_11" [id=296, type=dropout];
+"297 linear_15" [id=297, type=linear];
+"298 dropout_12" [id=298, type=dropout];
+"299 add_8" [id=299, type=add];
+"300 layer_norm_8" [id=300, type=layer_norm];
+"301 transpose_24" [id=301, type=transpose];
+"302 linear_16" [id=302, type=linear];
+"303 unflatten_4" [id=303, type=unflatten];
+"304 unsqueeze_4" [id=304, type=unsqueeze];
+"305 transpose_25" [id=305, type=transpose];
+"306 squeeze_4" [id=306, type=squeeze];
+"307 contiguous_4" [id=307, type=contiguous];
+"308 select_12" [id=308, type=select];
+"309 select_13" [id=309, type=select];
+"310 select_14" [id=310, type=select];
+"311 view_32" [id=311, type=view];
+"312 transpose_26" [id=312, type=transpose];
+"313 view_33" [id=313, type=view];
+"314 transpose_27" [id=314, type=transpose];
+"315 view_34" [id=315, type=view];
+"316 transpose_28" [id=316, type=transpose];
+"317 view_35" [id=317, type=view];
+"318 view_36" [id=318, type=view];
+"319 view_37" [id=319, type=view];
+"320 scaled_dot_product_attention_4" [id=320, type=scaled_dot_product_attention];
+"321 permute_5" [id=321, type=permute];
+"322 view_38" [id=322, type=view];
+"323 linear_17" [id=323, type=linear];
+"324 view_39" [id=324, type=view];
+"325 transpose_29" [id=325, type=transpose];
+"326 dropout_13" [id=326, type=dropout];
+"327 add_9" [id=327, type=add];
+"328 layer_norm_9" [id=328, type=layer_norm];
+"329 linear_18" [id=329, type=linear];
+"330 gelu_4" [id=330, type=gelu];
+"331 dropout_14" [id=331, type=dropout];
+"332 linear_19" [id=332, type=linear];
+"333 dropout_15" [id=333, type=dropout];
+"334 add_10" [id=334, type=add];
+"335 layer_norm_10" [id=335, type=layer_norm];
+"336 transpose_30" [id=336, type=transpose];
+"337 linear_20" [id=337, type=linear];
+"338 unflatten_5" [id=338, type=unflatten];
+"339 unsqueeze_5" [id=339, type=unsqueeze];
+"340 transpose_31" [id=340, type=transpose];
+"341 squeeze_5" [id=341, type=squeeze];
+"342 contiguous_5" [id=342, type=contiguous];
+"343 select_15" [id=343, type=select];
+"344 select_16" [id=344, type=select];
+"345 select_17" [id=345, type=select];
+"346 view_40" [id=346, type=view];
+"347 transpose_32" [id=347, type=transpose];
+"348 view_41" [id=348, type=view];
+"349 transpose_33" [id=349, type=transpose];
+"350 view_42" [id=350, type=view];
+"351 transpose_34" [id=351, type=transpose];
+"352 view_43" [id=352, type=view];
+"353 view_44" [id=353, type=view];
+"354 view_45" [id=354, type=view];
+"355 scaled_dot_product_attention_5" [id=355, type=scaled_dot_product_attention];
+"356 permute_6" [id=356, type=permute];
+"357 view_46" [id=357, type=view];
+"358 linear_21" [id=358, type=linear];
+"359 view_47" [id=359, type=view];
+"360 transpose_35" [id=360, type=transpose];
+"361 dropout_16" [id=361, type=dropout];
+"362 add_11" [id=362, type=add];
+"363 layer_norm_11" [id=363, type=layer_norm];
+"364 linear_22" [id=364, type=linear];
+"365 gelu_5" [id=365, type=gelu];
+"366 dropout_17" [id=366, type=dropout];
+"367 linear_23" [id=367, type=linear];
+"368 dropout_18" [id=368, type=dropout];
+"369 add_12" [id=369, type=add];
+"370 layer_norm_12" [id=370, type=layer_norm];
+"371 transpose_36" [id=371, type=transpose];
+"372 linear_24" [id=372, type=linear];
+"373 unflatten_6" [id=373, type=unflatten];
+"374 unsqueeze_6" [id=374, type=unsqueeze];
+"375 transpose_37" [id=375, type=transpose];
+"376 squeeze_6" [id=376, type=squeeze];
+"377 contiguous_6" [id=377, type=contiguous];
+"378 select_18" [id=378, type=select];
+"379 select_19" [id=379, type=select];
+"380 select_20" [id=380, type=select];
+"381 view_48" [id=381, type=view];
+"382 transpose_38" [id=382, type=transpose];
+"383 view_49" [id=383, type=view];
+"384 transpose_39" [id=384, type=transpose];
+"385 view_50" [id=385, type=view];
+"386 transpose_40" [id=386, type=transpose];
+"387 view_51" [id=387, type=view];
+"388 view_52" [id=388, type=view];
+"389 view_53" [id=389, type=view];
+"390 scaled_dot_product_attention_6" [id=390, type=scaled_dot_product_attention];
+"391 permute_7" [id=391, type=permute];
+"392 view_54" [id=392, type=view];
+"393 linear_25" [id=393, type=linear];
+"394 view_55" [id=394, type=view];
+"395 transpose_41" [id=395, type=transpose];
+"396 dropout_19" [id=396, type=dropout];
+"397 add_13" [id=397, type=add];
+"398 layer_norm_13" [id=398, type=layer_norm];
+"399 linear_26" [id=399, type=linear];
+"400 gelu_6" [id=400, type=gelu];
+"401 dropout_20" [id=401, type=dropout];
+"402 linear_27" [id=402, type=linear];
+"403 dropout_21" [id=403, type=dropout];
+"404 add_14" [id=404, type=add];
+"405 layer_norm_14" [id=405, type=layer_norm];
+"406 transpose_42" [id=406, type=transpose];
+"407 linear_28" [id=407, type=linear];
+"408 unflatten_7" [id=408, type=unflatten];
+"409 unsqueeze_7" [id=409, type=unsqueeze];
+"410 transpose_43" [id=410, type=transpose];
+"411 squeeze_7" [id=411, type=squeeze];
+"412 contiguous_7" [id=412, type=contiguous];
+"413 select_21" [id=413, type=select];
+"414 select_22" [id=414, type=select];
+"415 select_23" [id=415, type=select];
+"416 view_56" [id=416, type=view];
+"417 transpose_44" [id=417, type=transpose];
+"418 view_57" [id=418, type=view];
+"419 transpose_45" [id=419, type=transpose];
+"420 view_58" [id=420, type=view];
+"421 transpose_46" [id=421, type=transpose];
+"422 view_59" [id=422, type=view];
+"423 view_60" [id=423, type=view];
+"424 view_61" [id=424, type=view];
+"425 scaled_dot_product_attention_7" [id=425, type=scaled_dot_product_attention];
+"426 permute_8" [id=426, type=permute];
+"427 view_62" [id=427, type=view];
+"428 linear_29" [id=428, type=linear];
+"429 view_63" [id=429, type=view];
+"430 transpose_47" [id=430, type=transpose];
+"431 dropout_22" [id=431, type=dropout];
+"432 add_15" [id=432, type=add];
+"433 layer_norm_15" [id=433, type=layer_norm];
+"434 linear_30" [id=434, type=linear];
+"435 gelu_7" [id=435, type=gelu];
+"436 dropout_23" [id=436, type=dropout];
+"437 linear_31" [id=437, type=linear];
+"438 dropout_24" [id=438, type=dropout];
+"439 add_16" [id=439, type=add];
+"440 layer_norm_16" [id=440, type=layer_norm];
+"441 transpose_48" [id=441, type=transpose];
+"442 linear_32" [id=442, type=linear];
+"443 unflatten_8" [id=443, type=unflatten];
+"444 unsqueeze_8" [id=444, type=unsqueeze];
+"445 transpose_49" [id=445, type=transpose];
+"446 squeeze_8" [id=446, type=squeeze];
+"447 contiguous_8" [id=447, type=contiguous];
+"448 select_24" [id=448, type=select];
+"449 select_25" [id=449, type=select];
+"450 select_26" [id=450, type=select];
+"451 view_64" [id=451, type=view];
+"452 transpose_50" [id=452, type=transpose];
+"453 view_65" [id=453, type=view];
+"454 transpose_51" [id=454, type=transpose];
+"455 view_66" [id=455, type=view];
+"456 transpose_52" [id=456, type=transpose];
+"457 view_67" [id=457, type=view];
+"458 view_68" [id=458, type=view];
+"459 view_69" [id=459, type=view];
+"460 scaled_dot_product_attention_8" [id=460, type=scaled_dot_product_attention];
+"461 permute_9" [id=461, type=permute];
+"462 view_70" [id=462, type=view];
+"463 linear_33" [id=463, type=linear];
+"464 view_71" [id=464, type=view];
+"465 transpose_53" [id=465, type=transpose];
+"466 dropout_25" [id=466, type=dropout];
+"467 add_17" [id=467, type=add];
+"468 layer_norm_17" [id=468, type=layer_norm];
+"469 linear_34" [id=469, type=linear];
+"470 gelu_8" [id=470, type=gelu];
+"471 dropout_26" [id=471, type=dropout];
+"472 linear_35" [id=472, type=linear];
+"473 dropout_27" [id=473, type=dropout];
+"474 add_18" [id=474, type=add];
+"475 layer_norm_18" [id=475, type=layer_norm];
+"476 transpose_54" [id=476, type=transpose];
+"477 linear_36" [id=477, type=linear];
+"478 unflatten_9" [id=478, type=unflatten];
+"479 unsqueeze_9" [id=479, type=unsqueeze];
+"480 transpose_55" [id=480, type=transpose];
+"481 squeeze_9" [id=481, type=squeeze];
+"482 contiguous_9" [id=482, type=contiguous];
+"483 select_27" [id=483, type=select];
+"484 select_28" [id=484, type=select];
+"485 select_29" [id=485, type=select];
+"486 view_72" [id=486, type=view];
+"487 transpose_56" [id=487, type=transpose];
+"488 view_73" [id=488, type=view];
+"489 transpose_57" [id=489, type=transpose];
+"490 view_74" [id=490, type=view];
+"491 transpose_58" [id=491, type=transpose];
+"492 view_75" [id=492, type=view];
+"493 view_76" [id=493, type=view];
+"494 view_77" [id=494, type=view];
+"495 scaled_dot_product_attention_9" [id=495, type=scaled_dot_product_attention];
+"496 permute_10" [id=496, type=permute];
+"497 view_78" [id=497, type=view];
+"498 linear_37" [id=498, type=linear];
+"499 view_79" [id=499, type=view];
+"500 transpose_59" [id=500, type=transpose];
+"501 dropout_28" [id=501, type=dropout];
+"502 add_19" [id=502, type=add];
+"503 layer_norm_19" [id=503, type=layer_norm];
+"504 linear_38" [id=504, type=linear];
+"505 gelu_9" [id=505, type=gelu];
+"506 dropout_29" [id=506, type=dropout];
+"507 linear_39" [id=507, type=linear];
+"508 dropout_30" [id=508, type=dropout];
+"509 add_20" [id=509, type=add];
+"510 layer_norm_20" [id=510, type=layer_norm];
+"511 transpose_60" [id=511, type=transpose];
+"512 linear_40" [id=512, type=linear];
+"513 unflatten_10" [id=513, type=unflatten];
+"514 unsqueeze_10" [id=514, type=unsqueeze];
+"515 transpose_61" [id=515, type=transpose];
+"516 squeeze_10" [id=516, type=squeeze];
+"517 contiguous_10" [id=517, type=contiguous];
+"518 select_30" [id=518, type=select];
+"519 select_31" [id=519, type=select];
+"520 select_32" [id=520, type=select];
+"521 view_80" [id=521, type=view];
+"522 transpose_62" [id=522, type=transpose];
+"523 view_81" [id=523, type=view];
+"524 transpose_63" [id=524, type=transpose];
+"525 view_82" [id=525, type=view];
+"526 transpose_64" [id=526, type=transpose];
+"527 view_83" [id=527, type=view];
+"528 view_84" [id=528, type=view];
+"529 view_85" [id=529, type=view];
+"530 scaled_dot_product_attention_10" [id=530, type=scaled_dot_product_attention];
+"531 permute_11" [id=531, type=permute];
+"532 view_86" [id=532, type=view];
+"533 linear_41" [id=533, type=linear];
+"534 view_87" [id=534, type=view];
+"535 transpose_65" [id=535, type=transpose];
+"536 dropout_31" [id=536, type=dropout];
+"537 add_21" [id=537, type=add];
+"538 layer_norm_21" [id=538, type=layer_norm];
+"539 linear_42" [id=539, type=linear];
+"540 gelu_10" [id=540, type=gelu];
+"541 dropout_32" [id=541, type=dropout];
+"542 linear_43" [id=542, type=linear];
+"543 dropout_33" [id=543, type=dropout];
+"544 add_22" [id=544, type=add];
+"545 layer_norm_22" [id=545, type=layer_norm];
+"546 transpose_66" [id=546, type=transpose];
+"547 linear_44" [id=547, type=linear];
+"548 unflatten_11" [id=548, type=unflatten];
+"549 unsqueeze_11" [id=549, type=unsqueeze];
+"550 transpose_67" [id=550, type=transpose];
+"551 squeeze_11" [id=551, type=squeeze];
+"552 contiguous_11" [id=552, type=contiguous];
+"553 select_33" [id=553, type=select];
+"554 select_34" [id=554, type=select];
+"555 select_35" [id=555, type=select];
+"556 view_88" [id=556, type=view];
+"557 transpose_68" [id=557, type=transpose];
+"558 view_89" [id=558, type=view];
+"559 transpose_69" [id=559, type=transpose];
+"560 view_90" [id=560, type=view];
+"561 transpose_70" [id=561, type=transpose];
+"562 view_91" [id=562, type=view];
+"563 view_92" [id=563, type=view];
+"564 view_93" [id=564, type=view];
+"565 scaled_dot_product_attention_11" [id=565, type=scaled_dot_product_attention];
+"566 permute_12" [id=566, type=permute];
+"567 view_94" [id=567, type=view];
+"568 linear_45" [id=568, type=linear];
+"569 view_95" [id=569, type=view];
+"570 transpose_71" [id=570, type=transpose];
+"571 dropout_34" [id=571, type=dropout];
+"572 add_23" [id=572, type=add];
+"573 layer_norm_23" [id=573, type=layer_norm];
+"574 linear_46" [id=574, type=linear];
+"575 gelu_11" [id=575, type=gelu];
+"576 dropout_35" [id=576, type=dropout];
+"577 linear_47" [id=577, type=linear];
+"578 dropout_36" [id=578, type=dropout];
+"579 add_24" [id=579, type=add];
+"580 layer_norm_24" [id=580, type=layer_norm];
+"581 slice_1" [id=581, type=slice];
+"582 select_36" [id=582, type=select];
 "583 linear_48" [id=583, type=linear];
-"584 output" [id=584, type=output];
-"0 arg0_1" -> "3 conv2d"  [label="(1, 3, 224, 224)", style=solid];
-"1 _param_constant0" -> "3 conv2d"  [label="(768, 3, 16, 16)", style=solid];
-"2 _param_constant1" -> "3 conv2d"  [label="(768,)", style=solid];
-"3 conv2d" -> "4 reshape"  [label="(1, 768, 14, 14)", style=solid];
-"4 reshape" -> "5 permute"  [label="(1, 768, 196)", style=solid];
-"5 permute" -> "8 cat"  [label="(1, 196, 768)", style=solid];
-"6 _param_constant2" -> "7 expand"  [label="(1, 1, 768)", style=solid];
-"7 expand" -> "8 cat"  [label="(1, 1, 768)", style=solid];
-"8 cat" -> "10 add"  [label="(1, 197, 768)", style=solid];
-"9 _param_constant3" -> "10 add"  [label="(1, 197, 768)", style=solid];
-"10 add" -> "11 dropout"  [label="(1, 197, 768)", style=solid];
-"11 dropout" -> "14 layer_norm"  [label="(1, 197, 768)", style=solid];
-"11 dropout" -> "45 add_1"  [label="(1, 197, 768)", style=solid];
-"12 _param_constant4" -> "14 layer_norm"  [label="(768,)", style=solid];
-"13 _param_constant5" -> "14 layer_norm"  [label="(768,)", style=solid];
-"14 layer_norm" -> "15 transpose"  [label="(1, 197, 768)", style=solid];
-"15 transpose" -> "18 linear"  [label="(197, 1, 768)", style=solid];
-"16 _param_constant6" -> "18 linear"  [label="(2304, 768)", style=solid];
-"17 _param_constant7" -> "18 linear"  [label="(2304,)", style=solid];
-"18 linear" -> "19 unflatten"  [label="(197, 1, 2304)", style=solid];
-"19 unflatten" -> "20 unsqueeze"  [label="(197, 1, 3, 768)", style=solid];
-"20 unsqueeze" -> "21 transpose_1"  [label="(1, 197, 1, 3, 768)", style=solid];
-"21 transpose_1" -> "22 squeeze"  [label="(3, 197, 1, 1, 768)", style=solid];
-"22 squeeze" -> "23 contiguous"  [label="(3, 197, 1, 768)", style=solid];
-"23 contiguous" -> "24 select"  [label="(3, 197, 1, 768)", style=solid];
-"23 contiguous" -> "25 select_1"  [label="(3, 197, 1, 768)", style=solid];
-"23 contiguous" -> "26 select_2"  [label="(3, 197, 1, 768)", style=solid];
-"24 select" -> "27 view"  [label="(197, 1, 768)", style=solid];
-"25 select_1" -> "29 view_1"  [label="(197, 1, 768)", style=solid];
-"26 select_2" -> "31 view_2"  [label="(197, 1, 768)", style=solid];
-"27 view" -> "28 transpose_2"  [label="(197, 12, 64)", style=solid];
-"28 transpose_2" -> "33 view_3"  [label="(12, 197, 64)", style=solid];
-"29 view_1" -> "30 transpose_3"  [label="(197, 12, 64)", style=solid];
-"30 transpose_3" -> "34 view_4"  [label="(12, 197, 64)", style=solid];
-"31 view_2" -> "32 transpose_4"  [label="(197, 12, 64)", style=solid];
-"32 transpose_4" -> "35 view_5"  [label="(12, 197, 64)", style=solid];
-"33 view_3" -> "36 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
-"34 view_4" -> "36 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
-"35 view_5" -> "36 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
-"36 scaled_dot_product_attention" -> "37 permute_1"  [label="(1, 12, 197, 64)", style=solid];
-"37 permute_1" -> "38 view_6"  [label="(197, 1, 12, 64)", style=solid];
-"38 view_6" -> "41 linear_1"  [label="(197, 768)", style=solid];
-"39 _param_constant8" -> "41 linear_1"  [label="(768, 768)", style=solid];
-"40 _param_constant9" -> "41 linear_1"  [label="(768,)", style=solid];
-"41 linear_1" -> "42 view_7"  [label="(197, 768)", style=solid];
-"42 view_7" -> "43 transpose_5"  [label="(197, 1, 768)", style=solid];
-"43 transpose_5" -> "44 dropout_1"  [label="(1, 197, 768)", style=solid];
-"44 dropout_1" -> "45 add_1"  [label="(1, 197, 768)", style=solid];
-"45 add_1" -> "48 layer_norm_1"  [label="(1, 197, 768)", style=solid];
-"45 add_1" -> "58 add_2"  [label="(1, 197, 768)", style=solid];
-"46 _param_constant10" -> "48 layer_norm_1"  [label="(768,)", style=solid];
-"47 _param_constant11" -> "48 layer_norm_1"  [label="(768,)", style=solid];
-"48 layer_norm_1" -> "51 linear_2"  [label="(1, 197, 768)", style=solid];
-"49 _param_constant12" -> "51 linear_2"  [label="(3072, 768)", style=solid];
-"50 _param_constant13" -> "51 linear_2"  [label="(3072,)", style=solid];
-"51 linear_2" -> "52 gelu"  [label="(1, 197, 3072)", style=solid];
-"52 gelu" -> "53 dropout_2"  [label="(1, 197, 3072)", style=solid];
-"53 dropout_2" -> "56 linear_3"  [label="(1, 197, 3072)", style=solid];
-"54 _param_constant14" -> "56 linear_3"  [label="(768, 3072)", style=solid];
-"55 _param_constant15" -> "56 linear_3"  [label="(768,)", style=solid];
-"56 linear_3" -> "57 dropout_3"  [label="(1, 197, 768)", style=solid];
-"57 dropout_3" -> "58 add_2"  [label="(1, 197, 768)", style=solid];
-"58 add_2" -> "61 layer_norm_2"  [label="(1, 197, 768)", style=solid];
-"58 add_2" -> "92 add_3"  [label="(1, 197, 768)", style=solid];
-"59 _param_constant16" -> "61 layer_norm_2"  [label="(768,)", style=solid];
-"60 _param_constant17" -> "61 layer_norm_2"  [label="(768,)", style=solid];
-"61 layer_norm_2" -> "62 transpose_6"  [label="(1, 197, 768)", style=solid];
-"62 transpose_6" -> "65 linear_4"  [label="(197, 1, 768)", style=solid];
-"63 _param_constant18" -> "65 linear_4"  [label="(2304, 768)", style=solid];
-"64 _param_constant19" -> "65 linear_4"  [label="(2304,)", style=solid];
-"65 linear_4" -> "66 unflatten_1"  [label="(197, 1, 2304)", style=solid];
-"66 unflatten_1" -> "67 unsqueeze_1"  [label="(197, 1, 3, 768)", style=solid];
-"67 unsqueeze_1" -> "68 transpose_7"  [label="(1, 197, 1, 3, 768)", style=solid];
-"68 transpose_7" -> "69 squeeze_1"  [label="(3, 197, 1, 1, 768)", style=solid];
-"69 squeeze_1" -> "70 contiguous_1"  [label="(3, 197, 1, 768)", style=solid];
-"70 contiguous_1" -> "71 select_3"  [label="(3, 197, 1, 768)", style=solid];
-"70 contiguous_1" -> "72 select_4"  [label="(3, 197, 1, 768)", style=solid];
-"70 contiguous_1" -> "73 select_5"  [label="(3, 197, 1, 768)", style=solid];
-"71 select_3" -> "74 view_8"  [label="(197, 1, 768)", style=solid];
-"72 select_4" -> "76 view_9"  [label="(197, 1, 768)", style=solid];
-"73 select_5" -> "78 view_10"  [label="(197, 1, 768)", style=solid];
-"74 view_8" -> "75 transpose_8"  [label="(197, 12, 64)", style=solid];
-"75 transpose_8" -> "80 view_11"  [label="(12, 197, 64)", style=solid];
-"76 view_9" -> "77 transpose_9"  [label="(197, 12, 64)", style=solid];
-"77 transpose_9" -> "81 view_12"  [label="(12, 197, 64)", style=solid];
-"78 view_10" -> "79 transpose_10"  [label="(197, 12, 64)", style=solid];
-"79 transpose_10" -> "82 view_13"  [label="(12, 197, 64)", style=solid];
-"80 view_11" -> "83 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
-"81 view_12" -> "83 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
-"82 view_13" -> "83 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
-"83 scaled_dot_product_attention_1" -> "84 permute_2"  [label="(1, 12, 197, 64)", style=solid];
-"84 permute_2" -> "85 view_14"  [label="(197, 1, 12, 64)", style=solid];
-"85 view_14" -> "88 linear_5"  [label="(197, 768)", style=solid];
-"86 _param_constant20" -> "88 linear_5"  [label="(768, 768)", style=solid];
-"87 _param_constant21" -> "88 linear_5"  [label="(768,)", style=solid];
-"88 linear_5" -> "89 view_15"  [label="(197, 768)", style=solid];
-"89 view_15" -> "90 transpose_11"  [label="(197, 1, 768)", style=solid];
-"90 transpose_11" -> "91 dropout_4"  [label="(1, 197, 768)", style=solid];
-"91 dropout_4" -> "92 add_3"  [label="(1, 197, 768)", style=solid];
-"92 add_3" -> "95 layer_norm_3"  [label="(1, 197, 768)", style=solid];
-"92 add_3" -> "105 add_4"  [label="(1, 197, 768)", style=solid];
-"93 _param_constant22" -> "95 layer_norm_3"  [label="(768,)", style=solid];
-"94 _param_constant23" -> "95 layer_norm_3"  [label="(768,)", style=solid];
-"95 layer_norm_3" -> "98 linear_6"  [label="(1, 197, 768)", style=solid];
-"96 _param_constant24" -> "98 linear_6"  [label="(3072, 768)", style=solid];
-"97 _param_constant25" -> "98 linear_6"  [label="(3072,)", style=solid];
-"98 linear_6" -> "99 gelu_1"  [label="(1, 197, 3072)", style=solid];
-"99 gelu_1" -> "100 dropout_5"  [label="(1, 197, 3072)", style=solid];
-"100 dropout_5" -> "103 linear_7"  [label="(1, 197, 3072)", style=solid];
-"101 _param_constant26" -> "103 linear_7"  [label="(768, 3072)", style=solid];
-"102 _param_constant27" -> "103 linear_7"  [label="(768,)", style=solid];
-"103 linear_7" -> "104 dropout_6"  [label="(1, 197, 768)", style=solid];
-"104 dropout_6" -> "105 add_4"  [label="(1, 197, 768)", style=solid];
-"105 add_4" -> "108 layer_norm_4"  [label="(1, 197, 768)", style=solid];
-"105 add_4" -> "139 add_5"  [label="(1, 197, 768)", style=solid];
-"106 _param_constant28" -> "108 layer_norm_4"  [label="(768,)", style=solid];
-"107 _param_constant29" -> "108 layer_norm_4"  [label="(768,)", style=solid];
-"108 layer_norm_4" -> "109 transpose_12"  [label="(1, 197, 768)", style=solid];
-"109 transpose_12" -> "112 linear_8"  [label="(197, 1, 768)", style=solid];
-"110 _param_constant30" -> "112 linear_8"  [label="(2304, 768)", style=solid];
-"111 _param_constant31" -> "112 linear_8"  [label="(2304,)", style=solid];
-"112 linear_8" -> "113 unflatten_2"  [label="(197, 1, 2304)", style=solid];
-"113 unflatten_2" -> "114 unsqueeze_2"  [label="(197, 1, 3, 768)", style=solid];
-"114 unsqueeze_2" -> "115 transpose_13"  [label="(1, 197, 1, 3, 768)", style=solid];
-"115 transpose_13" -> "116 squeeze_2"  [label="(3, 197, 1, 1, 768)", style=solid];
-"116 squeeze_2" -> "117 contiguous_2"  [label="(3, 197, 1, 768)", style=solid];
-"117 contiguous_2" -> "118 select_6"  [label="(3, 197, 1, 768)", style=solid];
-"117 contiguous_2" -> "119 select_7"  [label="(3, 197, 1, 768)", style=solid];
-"117 contiguous_2" -> "120 select_8"  [label="(3, 197, 1, 768)", style=solid];
-"118 select_6" -> "121 view_16"  [label="(197, 1, 768)", style=solid];
-"119 select_7" -> "123 view_17"  [label="(197, 1, 768)", style=solid];
-"120 select_8" -> "125 view_18"  [label="(197, 1, 768)", style=solid];
-"121 view_16" -> "122 transpose_14"  [label="(197, 12, 64)", style=solid];
-"122 transpose_14" -> "127 view_19"  [label="(12, 197, 64)", style=solid];
-"123 view_17" -> "124 transpose_15"  [label="(197, 12, 64)", style=solid];
-"124 transpose_15" -> "128 view_20"  [label="(12, 197, 64)", style=solid];
-"125 view_18" -> "126 transpose_16"  [label="(197, 12, 64)", style=solid];
-"126 transpose_16" -> "129 view_21"  [label="(12, 197, 64)", style=solid];
-"127 view_19" -> "130 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
-"128 view_20" -> "130 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
-"129 view_21" -> "130 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
-"130 scaled_dot_product_attention_2" -> "131 permute_3"  [label="(1, 12, 197, 64)", style=solid];
-"131 permute_3" -> "132 view_22"  [label="(197, 1, 12, 64)", style=solid];
-"132 view_22" -> "135 linear_9"  [label="(197, 768)", style=solid];
-"133 _param_constant32" -> "135 linear_9"  [label="(768, 768)", style=solid];
-"134 _param_constant33" -> "135 linear_9"  [label="(768,)", style=solid];
-"135 linear_9" -> "136 view_23"  [label="(197, 768)", style=solid];
-"136 view_23" -> "137 transpose_17"  [label="(197, 1, 768)", style=solid];
-"137 transpose_17" -> "138 dropout_7"  [label="(1, 197, 768)", style=solid];
-"138 dropout_7" -> "139 add_5"  [label="(1, 197, 768)", style=solid];
-"139 add_5" -> "142 layer_norm_5"  [label="(1, 197, 768)", style=solid];
-"139 add_5" -> "152 add_6"  [label="(1, 197, 768)", style=solid];
-"140 _param_constant34" -> "142 layer_norm_5"  [label="(768,)", style=solid];
-"141 _param_constant35" -> "142 layer_norm_5"  [label="(768,)", style=solid];
-"142 layer_norm_5" -> "145 linear_10"  [label="(1, 197, 768)", style=solid];
-"143 _param_constant36" -> "145 linear_10"  [label="(3072, 768)", style=solid];
-"144 _param_constant37" -> "145 linear_10"  [label="(3072,)", style=solid];
-"145 linear_10" -> "146 gelu_2"  [label="(1, 197, 3072)", style=solid];
-"146 gelu_2" -> "147 dropout_8"  [label="(1, 197, 3072)", style=solid];
-"147 dropout_8" -> "150 linear_11"  [label="(1, 197, 3072)", style=solid];
-"148 _param_constant38" -> "150 linear_11"  [label="(768, 3072)", style=solid];
-"149 _param_constant39" -> "150 linear_11"  [label="(768,)", style=solid];
-"150 linear_11" -> "151 dropout_9"  [label="(1, 197, 768)", style=solid];
-"151 dropout_9" -> "152 add_6"  [label="(1, 197, 768)", style=solid];
-"152 add_6" -> "155 layer_norm_6"  [label="(1, 197, 768)", style=solid];
-"152 add_6" -> "186 add_7"  [label="(1, 197, 768)", style=solid];
-"153 _param_constant40" -> "155 layer_norm_6"  [label="(768,)", style=solid];
-"154 _param_constant41" -> "155 layer_norm_6"  [label="(768,)", style=solid];
-"155 layer_norm_6" -> "156 transpose_18"  [label="(1, 197, 768)", style=solid];
-"156 transpose_18" -> "159 linear_12"  [label="(197, 1, 768)", style=solid];
-"157 _param_constant42" -> "159 linear_12"  [label="(2304, 768)", style=solid];
-"158 _param_constant43" -> "159 linear_12"  [label="(2304,)", style=solid];
-"159 linear_12" -> "160 unflatten_3"  [label="(197, 1, 2304)", style=solid];
-"160 unflatten_3" -> "161 unsqueeze_3"  [label="(197, 1, 3, 768)", style=solid];
-"161 unsqueeze_3" -> "162 transpose_19"  [label="(1, 197, 1, 3, 768)", style=solid];
-"162 transpose_19" -> "163 squeeze_3"  [label="(3, 197, 1, 1, 768)", style=solid];
-"163 squeeze_3" -> "164 contiguous_3"  [label="(3, 197, 1, 768)", style=solid];
-"164 contiguous_3" -> "165 select_9"  [label="(3, 197, 1, 768)", style=solid];
-"164 contiguous_3" -> "166 select_10"  [label="(3, 197, 1, 768)", style=solid];
-"164 contiguous_3" -> "167 select_11"  [label="(3, 197, 1, 768)", style=solid];
-"165 select_9" -> "168 view_24"  [label="(197, 1, 768)", style=solid];
-"166 select_10" -> "170 view_25"  [label="(197, 1, 768)", style=solid];
-"167 select_11" -> "172 view_26"  [label="(197, 1, 768)", style=solid];
-"168 view_24" -> "169 transpose_20"  [label="(197, 12, 64)", style=solid];
-"169 transpose_20" -> "174 view_27"  [label="(12, 197, 64)", style=solid];
-"170 view_25" -> "171 transpose_21"  [label="(197, 12, 64)", style=solid];
-"171 transpose_21" -> "175 view_28"  [label="(12, 197, 64)", style=solid];
-"172 view_26" -> "173 transpose_22"  [label="(197, 12, 64)", style=solid];
-"173 transpose_22" -> "176 view_29"  [label="(12, 197, 64)", style=solid];
-"174 view_27" -> "177 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
-"175 view_28" -> "177 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
-"176 view_29" -> "177 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
-"177 scaled_dot_product_attention_3" -> "178 permute_4"  [label="(1, 12, 197, 64)", style=solid];
-"178 permute_4" -> "179 view_30"  [label="(197, 1, 12, 64)", style=solid];
-"179 view_30" -> "182 linear_13"  [label="(197, 768)", style=solid];
-"180 _param_constant44" -> "182 linear_13"  [label="(768, 768)", style=solid];
-"181 _param_constant45" -> "182 linear_13"  [label="(768,)", style=solid];
-"182 linear_13" -> "183 view_31"  [label="(197, 768)", style=solid];
-"183 view_31" -> "184 transpose_23"  [label="(197, 1, 768)", style=solid];
-"184 transpose_23" -> "185 dropout_10"  [label="(1, 197, 768)", style=solid];
-"185 dropout_10" -> "186 add_7"  [label="(1, 197, 768)", style=solid];
-"186 add_7" -> "189 layer_norm_7"  [label="(1, 197, 768)", style=solid];
-"186 add_7" -> "199 add_8"  [label="(1, 197, 768)", style=solid];
-"187 _param_constant46" -> "189 layer_norm_7"  [label="(768,)", style=solid];
-"188 _param_constant47" -> "189 layer_norm_7"  [label="(768,)", style=solid];
-"189 layer_norm_7" -> "192 linear_14"  [label="(1, 197, 768)", style=solid];
-"190 _param_constant48" -> "192 linear_14"  [label="(3072, 768)", style=solid];
-"191 _param_constant49" -> "192 linear_14"  [label="(3072,)", style=solid];
-"192 linear_14" -> "193 gelu_3"  [label="(1, 197, 3072)", style=solid];
-"193 gelu_3" -> "194 dropout_11"  [label="(1, 197, 3072)", style=solid];
-"194 dropout_11" -> "197 linear_15"  [label="(1, 197, 3072)", style=solid];
-"195 _param_constant50" -> "197 linear_15"  [label="(768, 3072)", style=solid];
-"196 _param_constant51" -> "197 linear_15"  [label="(768,)", style=solid];
-"197 linear_15" -> "198 dropout_12"  [label="(1, 197, 768)", style=solid];
-"198 dropout_12" -> "199 add_8"  [label="(1, 197, 768)", style=solid];
-"199 add_8" -> "202 layer_norm_8"  [label="(1, 197, 768)", style=solid];
-"199 add_8" -> "233 add_9"  [label="(1, 197, 768)", style=solid];
-"200 _param_constant52" -> "202 layer_norm_8"  [label="(768,)", style=solid];
-"201 _param_constant53" -> "202 layer_norm_8"  [label="(768,)", style=solid];
-"202 layer_norm_8" -> "203 transpose_24"  [label="(1, 197, 768)", style=solid];
-"203 transpose_24" -> "206 linear_16"  [label="(197, 1, 768)", style=solid];
-"204 _param_constant54" -> "206 linear_16"  [label="(2304, 768)", style=solid];
-"205 _param_constant55" -> "206 linear_16"  [label="(2304,)", style=solid];
-"206 linear_16" -> "207 unflatten_4"  [label="(197, 1, 2304)", style=solid];
-"207 unflatten_4" -> "208 unsqueeze_4"  [label="(197, 1, 3, 768)", style=solid];
-"208 unsqueeze_4" -> "209 transpose_25"  [label="(1, 197, 1, 3, 768)", style=solid];
-"209 transpose_25" -> "210 squeeze_4"  [label="(3, 197, 1, 1, 768)", style=solid];
-"210 squeeze_4" -> "211 contiguous_4"  [label="(3, 197, 1, 768)", style=solid];
-"211 contiguous_4" -> "212 select_12"  [label="(3, 197, 1, 768)", style=solid];
-"211 contiguous_4" -> "213 select_13"  [label="(3, 197, 1, 768)", style=solid];
-"211 contiguous_4" -> "214 select_14"  [label="(3, 197, 1, 768)", style=solid];
-"212 select_12" -> "215 view_32"  [label="(197, 1, 768)", style=solid];
-"213 select_13" -> "217 view_33"  [label="(197, 1, 768)", style=solid];
-"214 select_14" -> "219 view_34"  [label="(197, 1, 768)", style=solid];
-"215 view_32" -> "216 transpose_26"  [label="(197, 12, 64)", style=solid];
-"216 transpose_26" -> "221 view_35"  [label="(12, 197, 64)", style=solid];
-"217 view_33" -> "218 transpose_27"  [label="(197, 12, 64)", style=solid];
-"218 transpose_27" -> "222 view_36"  [label="(12, 197, 64)", style=solid];
-"219 view_34" -> "220 transpose_28"  [label="(197, 12, 64)", style=solid];
-"220 transpose_28" -> "223 view_37"  [label="(12, 197, 64)", style=solid];
-"221 view_35" -> "224 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
-"222 view_36" -> "224 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
-"223 view_37" -> "224 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
-"224 scaled_dot_product_attention_4" -> "225 permute_5"  [label="(1, 12, 197, 64)", style=solid];
-"225 permute_5" -> "226 view_38"  [label="(197, 1, 12, 64)", style=solid];
-"226 view_38" -> "229 linear_17"  [label="(197, 768)", style=solid];
-"227 _param_constant56" -> "229 linear_17"  [label="(768, 768)", style=solid];
-"228 _param_constant57" -> "229 linear_17"  [label="(768,)", style=solid];
-"229 linear_17" -> "230 view_39"  [label="(197, 768)", style=solid];
-"230 view_39" -> "231 transpose_29"  [label="(197, 1, 768)", style=solid];
-"231 transpose_29" -> "232 dropout_13"  [label="(1, 197, 768)", style=solid];
-"232 dropout_13" -> "233 add_9"  [label="(1, 197, 768)", style=solid];
-"233 add_9" -> "236 layer_norm_9"  [label="(1, 197, 768)", style=solid];
-"233 add_9" -> "246 add_10"  [label="(1, 197, 768)", style=solid];
-"234 _param_constant58" -> "236 layer_norm_9"  [label="(768,)", style=solid];
-"235 _param_constant59" -> "236 layer_norm_9"  [label="(768,)", style=solid];
-"236 layer_norm_9" -> "239 linear_18"  [label="(1, 197, 768)", style=solid];
-"237 _param_constant60" -> "239 linear_18"  [label="(3072, 768)", style=solid];
-"238 _param_constant61" -> "239 linear_18"  [label="(3072,)", style=solid];
-"239 linear_18" -> "240 gelu_4"  [label="(1, 197, 3072)", style=solid];
-"240 gelu_4" -> "241 dropout_14"  [label="(1, 197, 3072)", style=solid];
-"241 dropout_14" -> "244 linear_19"  [label="(1, 197, 3072)", style=solid];
-"242 _param_constant62" -> "244 linear_19"  [label="(768, 3072)", style=solid];
-"243 _param_constant63" -> "244 linear_19"  [label="(768,)", style=solid];
-"244 linear_19" -> "245 dropout_15"  [label="(1, 197, 768)", style=solid];
-"245 dropout_15" -> "246 add_10"  [label="(1, 197, 768)", style=solid];
-"246 add_10" -> "249 layer_norm_10"  [label="(1, 197, 768)", style=solid];
-"246 add_10" -> "280 add_11"  [label="(1, 197, 768)", style=solid];
-"247 _param_constant64" -> "249 layer_norm_10"  [label="(768,)", style=solid];
-"248 _param_constant65" -> "249 layer_norm_10"  [label="(768,)", style=solid];
-"249 layer_norm_10" -> "250 transpose_30"  [label="(1, 197, 768)", style=solid];
-"250 transpose_30" -> "253 linear_20"  [label="(197, 1, 768)", style=solid];
-"251 _param_constant66" -> "253 linear_20"  [label="(2304, 768)", style=solid];
-"252 _param_constant67" -> "253 linear_20"  [label="(2304,)", style=solid];
-"253 linear_20" -> "254 unflatten_5"  [label="(197, 1, 2304)", style=solid];
-"254 unflatten_5" -> "255 unsqueeze_5"  [label="(197, 1, 3, 768)", style=solid];
-"255 unsqueeze_5" -> "256 transpose_31"  [label="(1, 197, 1, 3, 768)", style=solid];
-"256 transpose_31" -> "257 squeeze_5"  [label="(3, 197, 1, 1, 768)", style=solid];
-"257 squeeze_5" -> "258 contiguous_5"  [label="(3, 197, 1, 768)", style=solid];
-"258 contiguous_5" -> "259 select_15"  [label="(3, 197, 1, 768)", style=solid];
-"258 contiguous_5" -> "260 select_16"  [label="(3, 197, 1, 768)", style=solid];
-"258 contiguous_5" -> "261 select_17"  [label="(3, 197, 1, 768)", style=solid];
-"259 select_15" -> "262 view_40"  [label="(197, 1, 768)", style=solid];
-"260 select_16" -> "264 view_41"  [label="(197, 1, 768)", style=solid];
-"261 select_17" -> "266 view_42"  [label="(197, 1, 768)", style=solid];
-"262 view_40" -> "263 transpose_32"  [label="(197, 12, 64)", style=solid];
-"263 transpose_32" -> "268 view_43"  [label="(12, 197, 64)", style=solid];
-"264 view_41" -> "265 transpose_33"  [label="(197, 12, 64)", style=solid];
-"265 transpose_33" -> "269 view_44"  [label="(12, 197, 64)", style=solid];
-"266 view_42" -> "267 transpose_34"  [label="(197, 12, 64)", style=solid];
-"267 transpose_34" -> "270 view_45"  [label="(12, 197, 64)", style=solid];
-"268 view_43" -> "271 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
-"269 view_44" -> "271 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
-"270 view_45" -> "271 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
-"271 scaled_dot_product_attention_5" -> "272 permute_6"  [label="(1, 12, 197, 64)", style=solid];
-"272 permute_6" -> "273 view_46"  [label="(197, 1, 12, 64)", style=solid];
-"273 view_46" -> "276 linear_21"  [label="(197, 768)", style=solid];
-"274 _param_constant68" -> "276 linear_21"  [label="(768, 768)", style=solid];
-"275 _param_constant69" -> "276 linear_21"  [label="(768,)", style=solid];
-"276 linear_21" -> "277 view_47"  [label="(197, 768)", style=solid];
-"277 view_47" -> "278 transpose_35"  [label="(197, 1, 768)", style=solid];
-"278 transpose_35" -> "279 dropout_16"  [label="(1, 197, 768)", style=solid];
-"279 dropout_16" -> "280 add_11"  [label="(1, 197, 768)", style=solid];
-"280 add_11" -> "283 layer_norm_11"  [label="(1, 197, 768)", style=solid];
-"280 add_11" -> "293 add_12"  [label="(1, 197, 768)", style=solid];
-"281 _param_constant70" -> "283 layer_norm_11"  [label="(768,)", style=solid];
-"282 _param_constant71" -> "283 layer_norm_11"  [label="(768,)", style=solid];
-"283 layer_norm_11" -> "286 linear_22"  [label="(1, 197, 768)", style=solid];
-"284 _param_constant72" -> "286 linear_22"  [label="(3072, 768)", style=solid];
-"285 _param_constant73" -> "286 linear_22"  [label="(3072,)", style=solid];
-"286 linear_22" -> "287 gelu_5"  [label="(1, 197, 3072)", style=solid];
-"287 gelu_5" -> "288 dropout_17"  [label="(1, 197, 3072)", style=solid];
-"288 dropout_17" -> "291 linear_23"  [label="(1, 197, 3072)", style=solid];
-"289 _param_constant74" -> "291 linear_23"  [label="(768, 3072)", style=solid];
-"290 _param_constant75" -> "291 linear_23"  [label="(768,)", style=solid];
-"291 linear_23" -> "292 dropout_18"  [label="(1, 197, 768)", style=solid];
-"292 dropout_18" -> "293 add_12"  [label="(1, 197, 768)", style=solid];
-"293 add_12" -> "296 layer_norm_12"  [label="(1, 197, 768)", style=solid];
-"293 add_12" -> "327 add_13"  [label="(1, 197, 768)", style=solid];
-"294 _param_constant76" -> "296 layer_norm_12"  [label="(768,)", style=solid];
-"295 _param_constant77" -> "296 layer_norm_12"  [label="(768,)", style=solid];
-"296 layer_norm_12" -> "297 transpose_36"  [label="(1, 197, 768)", style=solid];
-"297 transpose_36" -> "300 linear_24"  [label="(197, 1, 768)", style=solid];
-"298 _param_constant78" -> "300 linear_24"  [label="(2304, 768)", style=solid];
-"299 _param_constant79" -> "300 linear_24"  [label="(2304,)", style=solid];
-"300 linear_24" -> "301 unflatten_6"  [label="(197, 1, 2304)", style=solid];
-"301 unflatten_6" -> "302 unsqueeze_6"  [label="(197, 1, 3, 768)", style=solid];
-"302 unsqueeze_6" -> "303 transpose_37"  [label="(1, 197, 1, 3, 768)", style=solid];
-"303 transpose_37" -> "304 squeeze_6"  [label="(3, 197, 1, 1, 768)", style=solid];
-"304 squeeze_6" -> "305 contiguous_6"  [label="(3, 197, 1, 768)", style=solid];
-"305 contiguous_6" -> "306 select_18"  [label="(3, 197, 1, 768)", style=solid];
-"305 contiguous_6" -> "307 select_19"  [label="(3, 197, 1, 768)", style=solid];
-"305 contiguous_6" -> "308 select_20"  [label="(3, 197, 1, 768)", style=solid];
-"306 select_18" -> "309 view_48"  [label="(197, 1, 768)", style=solid];
-"307 select_19" -> "311 view_49"  [label="(197, 1, 768)", style=solid];
-"308 select_20" -> "313 view_50"  [label="(197, 1, 768)", style=solid];
-"309 view_48" -> "310 transpose_38"  [label="(197, 12, 64)", style=solid];
-"310 transpose_38" -> "315 view_51"  [label="(12, 197, 64)", style=solid];
-"311 view_49" -> "312 transpose_39"  [label="(197, 12, 64)", style=solid];
-"312 transpose_39" -> "316 view_52"  [label="(12, 197, 64)", style=solid];
-"313 view_50" -> "314 transpose_40"  [label="(197, 12, 64)", style=solid];
-"314 transpose_40" -> "317 view_53"  [label="(12, 197, 64)", style=solid];
-"315 view_51" -> "318 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
-"316 view_52" -> "318 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
-"317 view_53" -> "318 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
-"318 scaled_dot_product_attention_6" -> "319 permute_7"  [label="(1, 12, 197, 64)", style=solid];
-"319 permute_7" -> "320 view_54"  [label="(197, 1, 12, 64)", style=solid];
-"320 view_54" -> "323 linear_25"  [label="(197, 768)", style=solid];
-"321 _param_constant80" -> "323 linear_25"  [label="(768, 768)", style=solid];
-"322 _param_constant81" -> "323 linear_25"  [label="(768,)", style=solid];
-"323 linear_25" -> "324 view_55"  [label="(197, 768)", style=solid];
-"324 view_55" -> "325 transpose_41"  [label="(197, 1, 768)", style=solid];
-"325 transpose_41" -> "326 dropout_19"  [label="(1, 197, 768)", style=solid];
-"326 dropout_19" -> "327 add_13"  [label="(1, 197, 768)", style=solid];
-"327 add_13" -> "330 layer_norm_13"  [label="(1, 197, 768)", style=solid];
-"327 add_13" -> "340 add_14"  [label="(1, 197, 768)", style=solid];
-"328 _param_constant82" -> "330 layer_norm_13"  [label="(768,)", style=solid];
-"329 _param_constant83" -> "330 layer_norm_13"  [label="(768,)", style=solid];
-"330 layer_norm_13" -> "333 linear_26"  [label="(1, 197, 768)", style=solid];
-"331 _param_constant84" -> "333 linear_26"  [label="(3072, 768)", style=solid];
-"332 _param_constant85" -> "333 linear_26"  [label="(3072,)", style=solid];
-"333 linear_26" -> "334 gelu_6"  [label="(1, 197, 3072)", style=solid];
-"334 gelu_6" -> "335 dropout_20"  [label="(1, 197, 3072)", style=solid];
-"335 dropout_20" -> "338 linear_27"  [label="(1, 197, 3072)", style=solid];
-"336 _param_constant86" -> "338 linear_27"  [label="(768, 3072)", style=solid];
-"337 _param_constant87" -> "338 linear_27"  [label="(768,)", style=solid];
-"338 linear_27" -> "339 dropout_21"  [label="(1, 197, 768)", style=solid];
-"339 dropout_21" -> "340 add_14"  [label="(1, 197, 768)", style=solid];
-"340 add_14" -> "343 layer_norm_14"  [label="(1, 197, 768)", style=solid];
-"340 add_14" -> "374 add_15"  [label="(1, 197, 768)", style=solid];
-"341 _param_constant88" -> "343 layer_norm_14"  [label="(768,)", style=solid];
-"342 _param_constant89" -> "343 layer_norm_14"  [label="(768,)", style=solid];
-"343 layer_norm_14" -> "344 transpose_42"  [label="(1, 197, 768)", style=solid];
-"344 transpose_42" -> "347 linear_28"  [label="(197, 1, 768)", style=solid];
-"345 _param_constant90" -> "347 linear_28"  [label="(2304, 768)", style=solid];
-"346 _param_constant91" -> "347 linear_28"  [label="(2304,)", style=solid];
-"347 linear_28" -> "348 unflatten_7"  [label="(197, 1, 2304)", style=solid];
-"348 unflatten_7" -> "349 unsqueeze_7"  [label="(197, 1, 3, 768)", style=solid];
-"349 unsqueeze_7" -> "350 transpose_43"  [label="(1, 197, 1, 3, 768)", style=solid];
-"350 transpose_43" -> "351 squeeze_7"  [label="(3, 197, 1, 1, 768)", style=solid];
-"351 squeeze_7" -> "352 contiguous_7"  [label="(3, 197, 1, 768)", style=solid];
-"352 contiguous_7" -> "353 select_21"  [label="(3, 197, 1, 768)", style=solid];
-"352 contiguous_7" -> "354 select_22"  [label="(3, 197, 1, 768)", style=solid];
-"352 contiguous_7" -> "355 select_23"  [label="(3, 197, 1, 768)", style=solid];
-"353 select_21" -> "356 view_56"  [label="(197, 1, 768)", style=solid];
-"354 select_22" -> "358 view_57"  [label="(197, 1, 768)", style=solid];
-"355 select_23" -> "360 view_58"  [label="(197, 1, 768)", style=solid];
-"356 view_56" -> "357 transpose_44"  [label="(197, 12, 64)", style=solid];
-"357 transpose_44" -> "362 view_59"  [label="(12, 197, 64)", style=solid];
-"358 view_57" -> "359 transpose_45"  [label="(197, 12, 64)", style=solid];
-"359 transpose_45" -> "363 view_60"  [label="(12, 197, 64)", style=solid];
-"360 view_58" -> "361 transpose_46"  [label="(197, 12, 64)", style=solid];
-"361 transpose_46" -> "364 view_61"  [label="(12, 197, 64)", style=solid];
-"362 view_59" -> "365 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
-"363 view_60" -> "365 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
-"364 view_61" -> "365 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
-"365 scaled_dot_product_attention_7" -> "366 permute_8"  [label="(1, 12, 197, 64)", style=solid];
-"366 permute_8" -> "367 view_62"  [label="(197, 1, 12, 64)", style=solid];
-"367 view_62" -> "370 linear_29"  [label="(197, 768)", style=solid];
-"368 _param_constant92" -> "370 linear_29"  [label="(768, 768)", style=solid];
-"369 _param_constant93" -> "370 linear_29"  [label="(768,)", style=solid];
-"370 linear_29" -> "371 view_63"  [label="(197, 768)", style=solid];
-"371 view_63" -> "372 transpose_47"  [label="(197, 1, 768)", style=solid];
-"372 transpose_47" -> "373 dropout_22"  [label="(1, 197, 768)", style=solid];
-"373 dropout_22" -> "374 add_15"  [label="(1, 197, 768)", style=solid];
-"374 add_15" -> "377 layer_norm_15"  [label="(1, 197, 768)", style=solid];
-"374 add_15" -> "387 add_16"  [label="(1, 197, 768)", style=solid];
-"375 _param_constant94" -> "377 layer_norm_15"  [label="(768,)", style=solid];
-"376 _param_constant95" -> "377 layer_norm_15"  [label="(768,)", style=solid];
-"377 layer_norm_15" -> "380 linear_30"  [label="(1, 197, 768)", style=solid];
-"378 _param_constant96" -> "380 linear_30"  [label="(3072, 768)", style=solid];
-"379 _param_constant97" -> "380 linear_30"  [label="(3072,)", style=solid];
-"380 linear_30" -> "381 gelu_7"  [label="(1, 197, 3072)", style=solid];
-"381 gelu_7" -> "382 dropout_23"  [label="(1, 197, 3072)", style=solid];
-"382 dropout_23" -> "385 linear_31"  [label="(1, 197, 3072)", style=solid];
-"383 _param_constant98" -> "385 linear_31"  [label="(768, 3072)", style=solid];
-"384 _param_constant99" -> "385 linear_31"  [label="(768,)", style=solid];
-"385 linear_31" -> "386 dropout_24"  [label="(1, 197, 768)", style=solid];
-"386 dropout_24" -> "387 add_16"  [label="(1, 197, 768)", style=solid];
-"387 add_16" -> "390 layer_norm_16"  [label="(1, 197, 768)", style=solid];
-"387 add_16" -> "421 add_17"  [label="(1, 197, 768)", style=solid];
-"388 _param_constant100" -> "390 layer_norm_16"  [label="(768,)", style=solid];
-"389 _param_constant101" -> "390 layer_norm_16"  [label="(768,)", style=solid];
-"390 layer_norm_16" -> "391 transpose_48"  [label="(1, 197, 768)", style=solid];
-"391 transpose_48" -> "394 linear_32"  [label="(197, 1, 768)", style=solid];
-"392 _param_constant102" -> "394 linear_32"  [label="(2304, 768)", style=solid];
-"393 _param_constant103" -> "394 linear_32"  [label="(2304,)", style=solid];
-"394 linear_32" -> "395 unflatten_8"  [label="(197, 1, 2304)", style=solid];
-"395 unflatten_8" -> "396 unsqueeze_8"  [label="(197, 1, 3, 768)", style=solid];
-"396 unsqueeze_8" -> "397 transpose_49"  [label="(1, 197, 1, 3, 768)", style=solid];
-"397 transpose_49" -> "398 squeeze_8"  [label="(3, 197, 1, 1, 768)", style=solid];
-"398 squeeze_8" -> "399 contiguous_8"  [label="(3, 197, 1, 768)", style=solid];
-"399 contiguous_8" -> "400 select_24"  [label="(3, 197, 1, 768)", style=solid];
-"399 contiguous_8" -> "401 select_25"  [label="(3, 197, 1, 768)", style=solid];
-"399 contiguous_8" -> "402 select_26"  [label="(3, 197, 1, 768)", style=solid];
-"400 select_24" -> "403 view_64"  [label="(197, 1, 768)", style=solid];
-"401 select_25" -> "405 view_65"  [label="(197, 1, 768)", style=solid];
-"402 select_26" -> "407 view_66"  [label="(197, 1, 768)", style=solid];
-"403 view_64" -> "404 transpose_50"  [label="(197, 12, 64)", style=solid];
-"404 transpose_50" -> "409 view_67"  [label="(12, 197, 64)", style=solid];
-"405 view_65" -> "406 transpose_51"  [label="(197, 12, 64)", style=solid];
-"406 transpose_51" -> "410 view_68"  [label="(12, 197, 64)", style=solid];
-"407 view_66" -> "408 transpose_52"  [label="(197, 12, 64)", style=solid];
-"408 transpose_52" -> "411 view_69"  [label="(12, 197, 64)", style=solid];
-"409 view_67" -> "412 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
-"410 view_68" -> "412 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
-"411 view_69" -> "412 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
-"412 scaled_dot_product_attention_8" -> "413 permute_9"  [label="(1, 12, 197, 64)", style=solid];
-"413 permute_9" -> "414 view_70"  [label="(197, 1, 12, 64)", style=solid];
-"414 view_70" -> "417 linear_33"  [label="(197, 768)", style=solid];
-"415 _param_constant104" -> "417 linear_33"  [label="(768, 768)", style=solid];
-"416 _param_constant105" -> "417 linear_33"  [label="(768,)", style=solid];
-"417 linear_33" -> "418 view_71"  [label="(197, 768)", style=solid];
-"418 view_71" -> "419 transpose_53"  [label="(197, 1, 768)", style=solid];
-"419 transpose_53" -> "420 dropout_25"  [label="(1, 197, 768)", style=solid];
-"420 dropout_25" -> "421 add_17"  [label="(1, 197, 768)", style=solid];
-"421 add_17" -> "424 layer_norm_17"  [label="(1, 197, 768)", style=solid];
-"421 add_17" -> "434 add_18"  [label="(1, 197, 768)", style=solid];
-"422 _param_constant106" -> "424 layer_norm_17"  [label="(768,)", style=solid];
-"423 _param_constant107" -> "424 layer_norm_17"  [label="(768,)", style=solid];
-"424 layer_norm_17" -> "427 linear_34"  [label="(1, 197, 768)", style=solid];
-"425 _param_constant108" -> "427 linear_34"  [label="(3072, 768)", style=solid];
-"426 _param_constant109" -> "427 linear_34"  [label="(3072,)", style=solid];
-"427 linear_34" -> "428 gelu_8"  [label="(1, 197, 3072)", style=solid];
-"428 gelu_8" -> "429 dropout_26"  [label="(1, 197, 3072)", style=solid];
-"429 dropout_26" -> "432 linear_35"  [label="(1, 197, 3072)", style=solid];
-"430 _param_constant110" -> "432 linear_35"  [label="(768, 3072)", style=solid];
-"431 _param_constant111" -> "432 linear_35"  [label="(768,)", style=solid];
-"432 linear_35" -> "433 dropout_27"  [label="(1, 197, 768)", style=solid];
-"433 dropout_27" -> "434 add_18"  [label="(1, 197, 768)", style=solid];
-"434 add_18" -> "437 layer_norm_18"  [label="(1, 197, 768)", style=solid];
-"434 add_18" -> "468 add_19"  [label="(1, 197, 768)", style=solid];
-"435 _param_constant112" -> "437 layer_norm_18"  [label="(768,)", style=solid];
-"436 _param_constant113" -> "437 layer_norm_18"  [label="(768,)", style=solid];
-"437 layer_norm_18" -> "438 transpose_54"  [label="(1, 197, 768)", style=solid];
-"438 transpose_54" -> "441 linear_36"  [label="(197, 1, 768)", style=solid];
-"439 _param_constant114" -> "441 linear_36"  [label="(2304, 768)", style=solid];
-"440 _param_constant115" -> "441 linear_36"  [label="(2304,)", style=solid];
-"441 linear_36" -> "442 unflatten_9"  [label="(197, 1, 2304)", style=solid];
-"442 unflatten_9" -> "443 unsqueeze_9"  [label="(197, 1, 3, 768)", style=solid];
-"443 unsqueeze_9" -> "444 transpose_55"  [label="(1, 197, 1, 3, 768)", style=solid];
-"444 transpose_55" -> "445 squeeze_9"  [label="(3, 197, 1, 1, 768)", style=solid];
-"445 squeeze_9" -> "446 contiguous_9"  [label="(3, 197, 1, 768)", style=solid];
-"446 contiguous_9" -> "447 select_27"  [label="(3, 197, 1, 768)", style=solid];
-"446 contiguous_9" -> "448 select_28"  [label="(3, 197, 1, 768)", style=solid];
-"446 contiguous_9" -> "449 select_29"  [label="(3, 197, 1, 768)", style=solid];
-"447 select_27" -> "450 view_72"  [label="(197, 1, 768)", style=solid];
-"448 select_28" -> "452 view_73"  [label="(197, 1, 768)", style=solid];
-"449 select_29" -> "454 view_74"  [label="(197, 1, 768)", style=solid];
-"450 view_72" -> "451 transpose_56"  [label="(197, 12, 64)", style=solid];
-"451 transpose_56" -> "456 view_75"  [label="(12, 197, 64)", style=solid];
-"452 view_73" -> "453 transpose_57"  [label="(197, 12, 64)", style=solid];
-"453 transpose_57" -> "457 view_76"  [label="(12, 197, 64)", style=solid];
-"454 view_74" -> "455 transpose_58"  [label="(197, 12, 64)", style=solid];
-"455 transpose_58" -> "458 view_77"  [label="(12, 197, 64)", style=solid];
-"456 view_75" -> "459 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
-"457 view_76" -> "459 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
-"458 view_77" -> "459 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
-"459 scaled_dot_product_attention_9" -> "460 permute_10"  [label="(1, 12, 197, 64)", style=solid];
-"460 permute_10" -> "461 view_78"  [label="(197, 1, 12, 64)", style=solid];
-"461 view_78" -> "464 linear_37"  [label="(197, 768)", style=solid];
-"462 _param_constant116" -> "464 linear_37"  [label="(768, 768)", style=solid];
-"463 _param_constant117" -> "464 linear_37"  [label="(768,)", style=solid];
-"464 linear_37" -> "465 view_79"  [label="(197, 768)", style=solid];
-"465 view_79" -> "466 transpose_59"  [label="(197, 1, 768)", style=solid];
-"466 transpose_59" -> "467 dropout_28"  [label="(1, 197, 768)", style=solid];
-"467 dropout_28" -> "468 add_19"  [label="(1, 197, 768)", style=solid];
-"468 add_19" -> "471 layer_norm_19"  [label="(1, 197, 768)", style=solid];
-"468 add_19" -> "481 add_20"  [label="(1, 197, 768)", style=solid];
-"469 _param_constant118" -> "471 layer_norm_19"  [label="(768,)", style=solid];
-"470 _param_constant119" -> "471 layer_norm_19"  [label="(768,)", style=solid];
-"471 layer_norm_19" -> "474 linear_38"  [label="(1, 197, 768)", style=solid];
-"472 _param_constant120" -> "474 linear_38"  [label="(3072, 768)", style=solid];
-"473 _param_constant121" -> "474 linear_38"  [label="(3072,)", style=solid];
-"474 linear_38" -> "475 gelu_9"  [label="(1, 197, 3072)", style=solid];
-"475 gelu_9" -> "476 dropout_29"  [label="(1, 197, 3072)", style=solid];
-"476 dropout_29" -> "479 linear_39"  [label="(1, 197, 3072)", style=solid];
-"477 _param_constant122" -> "479 linear_39"  [label="(768, 3072)", style=solid];
-"478 _param_constant123" -> "479 linear_39"  [label="(768,)", style=solid];
-"479 linear_39" -> "480 dropout_30"  [label="(1, 197, 768)", style=solid];
-"480 dropout_30" -> "481 add_20"  [label="(1, 197, 768)", style=solid];
-"481 add_20" -> "484 layer_norm_20"  [label="(1, 197, 768)", style=solid];
-"481 add_20" -> "515 add_21"  [label="(1, 197, 768)", style=solid];
-"482 _param_constant124" -> "484 layer_norm_20"  [label="(768,)", style=solid];
-"483 _param_constant125" -> "484 layer_norm_20"  [label="(768,)", style=solid];
-"484 layer_norm_20" -> "485 transpose_60"  [label="(1, 197, 768)", style=solid];
-"485 transpose_60" -> "488 linear_40"  [label="(197, 1, 768)", style=solid];
-"486 _param_constant126" -> "488 linear_40"  [label="(2304, 768)", style=solid];
-"487 _param_constant127" -> "488 linear_40"  [label="(2304,)", style=solid];
-"488 linear_40" -> "489 unflatten_10"  [label="(197, 1, 2304)", style=solid];
-"489 unflatten_10" -> "490 unsqueeze_10"  [label="(197, 1, 3, 768)", style=solid];
-"490 unsqueeze_10" -> "491 transpose_61"  [label="(1, 197, 1, 3, 768)", style=solid];
-"491 transpose_61" -> "492 squeeze_10"  [label="(3, 197, 1, 1, 768)", style=solid];
-"492 squeeze_10" -> "493 contiguous_10"  [label="(3, 197, 1, 768)", style=solid];
-"493 contiguous_10" -> "494 select_30"  [label="(3, 197, 1, 768)", style=solid];
-"493 contiguous_10" -> "495 select_31"  [label="(3, 197, 1, 768)", style=solid];
-"493 contiguous_10" -> "496 select_32"  [label="(3, 197, 1, 768)", style=solid];
-"494 select_30" -> "497 view_80"  [label="(197, 1, 768)", style=solid];
-"495 select_31" -> "499 view_81"  [label="(197, 1, 768)", style=solid];
-"496 select_32" -> "501 view_82"  [label="(197, 1, 768)", style=solid];
-"497 view_80" -> "498 transpose_62"  [label="(197, 12, 64)", style=solid];
-"498 transpose_62" -> "503 view_83"  [label="(12, 197, 64)", style=solid];
-"499 view_81" -> "500 transpose_63"  [label="(197, 12, 64)", style=solid];
-"500 transpose_63" -> "504 view_84"  [label="(12, 197, 64)", style=solid];
-"501 view_82" -> "502 transpose_64"  [label="(197, 12, 64)", style=solid];
-"502 transpose_64" -> "505 view_85"  [label="(12, 197, 64)", style=solid];
-"503 view_83" -> "506 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
-"504 view_84" -> "506 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
-"505 view_85" -> "506 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
-"506 scaled_dot_product_attention_10" -> "507 permute_11"  [label="(1, 12, 197, 64)", style=solid];
-"507 permute_11" -> "508 view_86"  [label="(197, 1, 12, 64)", style=solid];
-"508 view_86" -> "511 linear_41"  [label="(197, 768)", style=solid];
-"509 _param_constant128" -> "511 linear_41"  [label="(768, 768)", style=solid];
-"510 _param_constant129" -> "511 linear_41"  [label="(768,)", style=solid];
-"511 linear_41" -> "512 view_87"  [label="(197, 768)", style=solid];
-"512 view_87" -> "513 transpose_65"  [label="(197, 1, 768)", style=solid];
-"513 transpose_65" -> "514 dropout_31"  [label="(1, 197, 768)", style=solid];
-"514 dropout_31" -> "515 add_21"  [label="(1, 197, 768)", style=solid];
-"515 add_21" -> "518 layer_norm_21"  [label="(1, 197, 768)", style=solid];
-"515 add_21" -> "528 add_22"  [label="(1, 197, 768)", style=solid];
-"516 _param_constant130" -> "518 layer_norm_21"  [label="(768,)", style=solid];
-"517 _param_constant131" -> "518 layer_norm_21"  [label="(768,)", style=solid];
-"518 layer_norm_21" -> "521 linear_42"  [label="(1, 197, 768)", style=solid];
-"519 _param_constant132" -> "521 linear_42"  [label="(3072, 768)", style=solid];
-"520 _param_constant133" -> "521 linear_42"  [label="(3072,)", style=solid];
-"521 linear_42" -> "522 gelu_10"  [label="(1, 197, 3072)", style=solid];
-"522 gelu_10" -> "523 dropout_32"  [label="(1, 197, 3072)", style=solid];
-"523 dropout_32" -> "526 linear_43"  [label="(1, 197, 3072)", style=solid];
-"524 _param_constant134" -> "526 linear_43"  [label="(768, 3072)", style=solid];
-"525 _param_constant135" -> "526 linear_43"  [label="(768,)", style=solid];
-"526 linear_43" -> "527 dropout_33"  [label="(1, 197, 768)", style=solid];
-"527 dropout_33" -> "528 add_22"  [label="(1, 197, 768)", style=solid];
-"528 add_22" -> "531 layer_norm_22"  [label="(1, 197, 768)", style=solid];
-"528 add_22" -> "562 add_23"  [label="(1, 197, 768)", style=solid];
-"529 _param_constant136" -> "531 layer_norm_22"  [label="(768,)", style=solid];
-"530 _param_constant137" -> "531 layer_norm_22"  [label="(768,)", style=solid];
-"531 layer_norm_22" -> "532 transpose_66"  [label="(1, 197, 768)", style=solid];
-"532 transpose_66" -> "535 linear_44"  [label="(197, 1, 768)", style=solid];
-"533 _param_constant138" -> "535 linear_44"  [label="(2304, 768)", style=solid];
-"534 _param_constant139" -> "535 linear_44"  [label="(2304,)", style=solid];
-"535 linear_44" -> "536 unflatten_11"  [label="(197, 1, 2304)", style=solid];
-"536 unflatten_11" -> "537 unsqueeze_11"  [label="(197, 1, 3, 768)", style=solid];
-"537 unsqueeze_11" -> "538 transpose_67"  [label="(1, 197, 1, 3, 768)", style=solid];
-"538 transpose_67" -> "539 squeeze_11"  [label="(3, 197, 1, 1, 768)", style=solid];
-"539 squeeze_11" -> "540 contiguous_11"  [label="(3, 197, 1, 768)", style=solid];
-"540 contiguous_11" -> "541 select_33"  [label="(3, 197, 1, 768)", style=solid];
-"540 contiguous_11" -> "542 select_34"  [label="(3, 197, 1, 768)", style=solid];
-"540 contiguous_11" -> "543 select_35"  [label="(3, 197, 1, 768)", style=solid];
-"541 select_33" -> "544 view_88"  [label="(197, 1, 768)", style=solid];
-"542 select_34" -> "546 view_89"  [label="(197, 1, 768)", style=solid];
-"543 select_35" -> "548 view_90"  [label="(197, 1, 768)", style=solid];
-"544 view_88" -> "545 transpose_68"  [label="(197, 12, 64)", style=solid];
-"545 transpose_68" -> "550 view_91"  [label="(12, 197, 64)", style=solid];
-"546 view_89" -> "547 transpose_69"  [label="(197, 12, 64)", style=solid];
-"547 transpose_69" -> "551 view_92"  [label="(12, 197, 64)", style=solid];
-"548 view_90" -> "549 transpose_70"  [label="(197, 12, 64)", style=solid];
-"549 transpose_70" -> "552 view_93"  [label="(12, 197, 64)", style=solid];
-"550 view_91" -> "553 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
-"551 view_92" -> "553 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
-"552 view_93" -> "553 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
-"553 scaled_dot_product_attention_11" -> "554 permute_12"  [label="(1, 12, 197, 64)", style=solid];
-"554 permute_12" -> "555 view_94"  [label="(197, 1, 12, 64)", style=solid];
-"555 view_94" -> "558 linear_45"  [label="(197, 768)", style=solid];
-"556 _param_constant140" -> "558 linear_45"  [label="(768, 768)", style=solid];
-"557 _param_constant141" -> "558 linear_45"  [label="(768,)", style=solid];
-"558 linear_45" -> "559 view_95"  [label="(197, 768)", style=solid];
-"559 view_95" -> "560 transpose_71"  [label="(197, 1, 768)", style=solid];
-"560 transpose_71" -> "561 dropout_34"  [label="(1, 197, 768)", style=solid];
-"561 dropout_34" -> "562 add_23"  [label="(1, 197, 768)", style=solid];
-"562 add_23" -> "565 layer_norm_23"  [label="(1, 197, 768)", style=solid];
-"562 add_23" -> "575 add_24"  [label="(1, 197, 768)", style=solid];
-"563 _param_constant142" -> "565 layer_norm_23"  [label="(768,)", style=solid];
-"564 _param_constant143" -> "565 layer_norm_23"  [label="(768,)", style=solid];
-"565 layer_norm_23" -> "568 linear_46"  [label="(1, 197, 768)", style=solid];
-"566 _param_constant144" -> "568 linear_46"  [label="(3072, 768)", style=solid];
-"567 _param_constant145" -> "568 linear_46"  [label="(3072,)", style=solid];
-"568 linear_46" -> "569 gelu_11"  [label="(1, 197, 3072)", style=solid];
-"569 gelu_11" -> "570 dropout_35"  [label="(1, 197, 3072)", style=solid];
-"570 dropout_35" -> "573 linear_47"  [label="(1, 197, 3072)", style=solid];
-"571 _param_constant146" -> "573 linear_47"  [label="(768, 3072)", style=solid];
-"572 _param_constant147" -> "573 linear_47"  [label="(768,)", style=solid];
-"573 linear_47" -> "574 dropout_36"  [label="(1, 197, 768)", style=solid];
-"574 dropout_36" -> "575 add_24"  [label="(1, 197, 768)", style=solid];
-"575 add_24" -> "578 layer_norm_24"  [label="(1, 197, 768)", style=solid];
-"576 _param_constant148" -> "578 layer_norm_24"  [label="(768,)", style=solid];
-"577 _param_constant149" -> "578 layer_norm_24"  [label="(768,)", style=solid];
-"578 layer_norm_24" -> "579 slice_1"  [label="(1, 197, 768)", style=solid];
-"579 slice_1" -> "580 select_36"  [label="(1, 197, 768)", style=solid];
-"580 select_36" -> "583 linear_48"  [label="(1, 768)", style=solid];
-"581 _param_constant150" -> "583 linear_48"  [label="(1000, 768)", style=solid];
-"582 _param_constant151" -> "583 linear_48"  [label="(1000,)", style=solid];
-"583 linear_48" -> "584 output"  [label="(1, 1000)", style=solid];
+"584 output_1" [id=584, type=output];
+"0 class_token" -> "156 expand"  [label="(1, 1, 768)", style=solid];
+"1 encoder_pos_embedding" -> "158 add"  [label="(1, 197, 768)", style=solid];
+"2 conv_proj_weight" -> "153 conv2d"  [label="(768, 3, 16, 16)", style=solid];
+"3 conv_proj_bias" -> "153 conv2d"  [label="(768,)", style=solid];
+"4 encoder_layers_encoder_layer_0_ln_1_weight" -> "160 layer_norm"  [label="(768,)", style=solid];
+"5 encoder_layers_encoder_layer_0_ln_1_bias" -> "160 layer_norm"  [label="(768,)", style=solid];
+"6 encoder_layers_encoder_layer_0_self_attention_in_proj_weight" -> "162 linear"  [label="(2304, 768)", style=solid];
+"7 encoder_layers_encoder_layer_0_self_attention_in_proj_bias" -> "162 linear"  [label="(2304,)", style=solid];
+"8 encoder_layers_encoder_layer_0_self_attention_out_proj_weight" -> "183 linear_1"  [label="(768, 768)", style=solid];
+"9 encoder_layers_encoder_layer_0_self_attention_out_proj_bias" -> "183 linear_1"  [label="(768,)", style=solid];
+"10 encoder_layers_encoder_layer_0_ln_2_weight" -> "188 layer_norm_1"  [label="(768,)", style=solid];
+"11 encoder_layers_encoder_layer_0_ln_2_bias" -> "188 layer_norm_1"  [label="(768,)", style=solid];
+"12 encoder_layers_encoder_layer_0_mlp_0_weight" -> "189 linear_2"  [label="(3072, 768)", style=solid];
+"13 encoder_layers_encoder_layer_0_mlp_0_bias" -> "189 linear_2"  [label="(3072,)", style=solid];
+"14 encoder_layers_encoder_layer_0_mlp_3_weight" -> "192 linear_3"  [label="(768, 3072)", style=solid];
+"15 encoder_layers_encoder_layer_0_mlp_3_bias" -> "192 linear_3"  [label="(768,)", style=solid];
+"16 encoder_layers_encoder_layer_1_ln_1_weight" -> "195 layer_norm_2"  [label="(768,)", style=solid];
+"17 encoder_layers_encoder_layer_1_ln_1_bias" -> "195 layer_norm_2"  [label="(768,)", style=solid];
+"18 encoder_layers_encoder_layer_1_self_attention_in_proj_weight" -> "197 linear_4"  [label="(2304, 768)", style=solid];
+"19 encoder_layers_encoder_layer_1_self_attention_in_proj_bias" -> "197 linear_4"  [label="(2304,)", style=solid];
+"20 encoder_layers_encoder_layer_1_self_attention_out_proj_weight" -> "218 linear_5"  [label="(768, 768)", style=solid];
+"21 encoder_layers_encoder_layer_1_self_attention_out_proj_bias" -> "218 linear_5"  [label="(768,)", style=solid];
+"22 encoder_layers_encoder_layer_1_ln_2_weight" -> "223 layer_norm_3"  [label="(768,)", style=solid];
+"23 encoder_layers_encoder_layer_1_ln_2_bias" -> "223 layer_norm_3"  [label="(768,)", style=solid];
+"24 encoder_layers_encoder_layer_1_mlp_0_weight" -> "224 linear_6"  [label="(3072, 768)", style=solid];
+"25 encoder_layers_encoder_layer_1_mlp_0_bias" -> "224 linear_6"  [label="(3072,)", style=solid];
+"26 encoder_layers_encoder_layer_1_mlp_3_weight" -> "227 linear_7"  [label="(768, 3072)", style=solid];
+"27 encoder_layers_encoder_layer_1_mlp_3_bias" -> "227 linear_7"  [label="(768,)", style=solid];
+"28 encoder_layers_encoder_layer_2_ln_1_weight" -> "230 layer_norm_4"  [label="(768,)", style=solid];
+"29 encoder_layers_encoder_layer_2_ln_1_bias" -> "230 layer_norm_4"  [label="(768,)", style=solid];
+"30 encoder_layers_encoder_layer_2_self_attention_in_proj_weight" -> "232 linear_8"  [label="(2304, 768)", style=solid];
+"31 encoder_layers_encoder_layer_2_self_attention_in_proj_bias" -> "232 linear_8"  [label="(2304,)", style=solid];
+"32 encoder_layers_encoder_layer_2_self_attention_out_proj_weight" -> "253 linear_9"  [label="(768, 768)", style=solid];
+"33 encoder_layers_encoder_layer_2_self_attention_out_proj_bias" -> "253 linear_9"  [label="(768,)", style=solid];
+"34 encoder_layers_encoder_layer_2_ln_2_weight" -> "258 layer_norm_5"  [label="(768,)", style=solid];
+"35 encoder_layers_encoder_layer_2_ln_2_bias" -> "258 layer_norm_5"  [label="(768,)", style=solid];
+"36 encoder_layers_encoder_layer_2_mlp_0_weight" -> "259 linear_10"  [label="(3072, 768)", style=solid];
+"37 encoder_layers_encoder_layer_2_mlp_0_bias" -> "259 linear_10"  [label="(3072,)", style=solid];
+"38 encoder_layers_encoder_layer_2_mlp_3_weight" -> "262 linear_11"  [label="(768, 3072)", style=solid];
+"39 encoder_layers_encoder_layer_2_mlp_3_bias" -> "262 linear_11"  [label="(768,)", style=solid];
+"40 encoder_layers_encoder_layer_3_ln_1_weight" -> "265 layer_norm_6"  [label="(768,)", style=solid];
+"41 encoder_layers_encoder_layer_3_ln_1_bias" -> "265 layer_norm_6"  [label="(768,)", style=solid];
+"42 encoder_layers_encoder_layer_3_self_attention_in_proj_weight" -> "267 linear_12"  [label="(2304, 768)", style=solid];
+"43 encoder_layers_encoder_layer_3_self_attention_in_proj_bias" -> "267 linear_12"  [label="(2304,)", style=solid];
+"44 encoder_layers_encoder_layer_3_self_attention_out_proj_weight" -> "288 linear_13"  [label="(768, 768)", style=solid];
+"45 encoder_layers_encoder_layer_3_self_attention_out_proj_bias" -> "288 linear_13"  [label="(768,)", style=solid];
+"46 encoder_layers_encoder_layer_3_ln_2_weight" -> "293 layer_norm_7"  [label="(768,)", style=solid];
+"47 encoder_layers_encoder_layer_3_ln_2_bias" -> "293 layer_norm_7"  [label="(768,)", style=solid];
+"48 encoder_layers_encoder_layer_3_mlp_0_weight" -> "294 linear_14"  [label="(3072, 768)", style=solid];
+"49 encoder_layers_encoder_layer_3_mlp_0_bias" -> "294 linear_14"  [label="(3072,)", style=solid];
+"50 encoder_layers_encoder_layer_3_mlp_3_weight" -> "297 linear_15"  [label="(768, 3072)", style=solid];
+"51 encoder_layers_encoder_layer_3_mlp_3_bias" -> "297 linear_15"  [label="(768,)", style=solid];
+"52 encoder_layers_encoder_layer_4_ln_1_weight" -> "300 layer_norm_8"  [label="(768,)", style=solid];
+"53 encoder_layers_encoder_layer_4_ln_1_bias" -> "300 layer_norm_8"  [label="(768,)", style=solid];
+"54 encoder_layers_encoder_layer_4_self_attention_in_proj_weight" -> "302 linear_16"  [label="(2304, 768)", style=solid];
+"55 encoder_layers_encoder_layer_4_self_attention_in_proj_bias" -> "302 linear_16"  [label="(2304,)", style=solid];
+"56 encoder_layers_encoder_layer_4_self_attention_out_proj_weight" -> "323 linear_17"  [label="(768, 768)", style=solid];
+"57 encoder_layers_encoder_layer_4_self_attention_out_proj_bias" -> "323 linear_17"  [label="(768,)", style=solid];
+"58 encoder_layers_encoder_layer_4_ln_2_weight" -> "328 layer_norm_9"  [label="(768,)", style=solid];
+"59 encoder_layers_encoder_layer_4_ln_2_bias" -> "328 layer_norm_9"  [label="(768,)", style=solid];
+"60 encoder_layers_encoder_layer_4_mlp_0_weight" -> "329 linear_18"  [label="(3072, 768)", style=solid];
+"61 encoder_layers_encoder_layer_4_mlp_0_bias" -> "329 linear_18"  [label="(3072,)", style=solid];
+"62 encoder_layers_encoder_layer_4_mlp_3_weight" -> "332 linear_19"  [label="(768, 3072)", style=solid];
+"63 encoder_layers_encoder_layer_4_mlp_3_bias" -> "332 linear_19"  [label="(768,)", style=solid];
+"64 encoder_layers_encoder_layer_5_ln_1_weight" -> "335 layer_norm_10"  [label="(768,)", style=solid];
+"65 encoder_layers_encoder_layer_5_ln_1_bias" -> "335 layer_norm_10"  [label="(768,)", style=solid];
+"66 encoder_layers_encoder_layer_5_self_attention_in_proj_weight" -> "337 linear_20"  [label="(2304, 768)", style=solid];
+"67 encoder_layers_encoder_layer_5_self_attention_in_proj_bias" -> "337 linear_20"  [label="(2304,)", style=solid];
+"68 encoder_layers_encoder_layer_5_self_attention_out_proj_weight" -> "358 linear_21"  [label="(768, 768)", style=solid];
+"69 encoder_layers_encoder_layer_5_self_attention_out_proj_bias" -> "358 linear_21"  [label="(768,)", style=solid];
+"70 encoder_layers_encoder_layer_5_ln_2_weight" -> "363 layer_norm_11"  [label="(768,)", style=solid];
+"71 encoder_layers_encoder_layer_5_ln_2_bias" -> "363 layer_norm_11"  [label="(768,)", style=solid];
+"72 encoder_layers_encoder_layer_5_mlp_0_weight" -> "364 linear_22"  [label="(3072, 768)", style=solid];
+"73 encoder_layers_encoder_layer_5_mlp_0_bias" -> "364 linear_22"  [label="(3072,)", style=solid];
+"74 encoder_layers_encoder_layer_5_mlp_3_weight" -> "367 linear_23"  [label="(768, 3072)", style=solid];
+"75 encoder_layers_encoder_layer_5_mlp_3_bias" -> "367 linear_23"  [label="(768,)", style=solid];
+"76 encoder_layers_encoder_layer_6_ln_1_weight" -> "370 layer_norm_12"  [label="(768,)", style=solid];
+"77 encoder_layers_encoder_layer_6_ln_1_bias" -> "370 layer_norm_12"  [label="(768,)", style=solid];
+"78 encoder_layers_encoder_layer_6_self_attention_in_proj_weight" -> "372 linear_24"  [label="(2304, 768)", style=solid];
+"79 encoder_layers_encoder_layer_6_self_attention_in_proj_bias" -> "372 linear_24"  [label="(2304,)", style=solid];
+"80 encoder_layers_encoder_layer_6_self_attention_out_proj_weight" -> "393 linear_25"  [label="(768, 768)", style=solid];
+"81 encoder_layers_encoder_layer_6_self_attention_out_proj_bias" -> "393 linear_25"  [label="(768,)", style=solid];
+"82 encoder_layers_encoder_layer_6_ln_2_weight" -> "398 layer_norm_13"  [label="(768,)", style=solid];
+"83 encoder_layers_encoder_layer_6_ln_2_bias" -> "398 layer_norm_13"  [label="(768,)", style=solid];
+"84 encoder_layers_encoder_layer_6_mlp_0_weight" -> "399 linear_26"  [label="(3072, 768)", style=solid];
+"85 encoder_layers_encoder_layer_6_mlp_0_bias" -> "399 linear_26"  [label="(3072,)", style=solid];
+"86 encoder_layers_encoder_layer_6_mlp_3_weight" -> "402 linear_27"  [label="(768, 3072)", style=solid];
+"87 encoder_layers_encoder_layer_6_mlp_3_bias" -> "402 linear_27"  [label="(768,)", style=solid];
+"88 encoder_layers_encoder_layer_7_ln_1_weight" -> "405 layer_norm_14"  [label="(768,)", style=solid];
+"89 encoder_layers_encoder_layer_7_ln_1_bias" -> "405 layer_norm_14"  [label="(768,)", style=solid];
+"90 encoder_layers_encoder_layer_7_self_attention_in_proj_weight" -> "407 linear_28"  [label="(2304, 768)", style=solid];
+"91 encoder_layers_encoder_layer_7_self_attention_in_proj_bias" -> "407 linear_28"  [label="(2304,)", style=solid];
+"92 encoder_layers_encoder_layer_7_self_attention_out_proj_weight" -> "428 linear_29"  [label="(768, 768)", style=solid];
+"93 encoder_layers_encoder_layer_7_self_attention_out_proj_bias" -> "428 linear_29"  [label="(768,)", style=solid];
+"94 encoder_layers_encoder_layer_7_ln_2_weight" -> "433 layer_norm_15"  [label="(768,)", style=solid];
+"95 encoder_layers_encoder_layer_7_ln_2_bias" -> "433 layer_norm_15"  [label="(768,)", style=solid];
+"96 encoder_layers_encoder_layer_7_mlp_0_weight" -> "434 linear_30"  [label="(3072, 768)", style=solid];
+"97 encoder_layers_encoder_layer_7_mlp_0_bias" -> "434 linear_30"  [label="(3072,)", style=solid];
+"98 encoder_layers_encoder_layer_7_mlp_3_weight" -> "437 linear_31"  [label="(768, 3072)", style=solid];
+"99 encoder_layers_encoder_layer_7_mlp_3_bias" -> "437 linear_31"  [label="(768,)", style=solid];
+"100 encoder_layers_encoder_layer_8_ln_1_weight" -> "440 layer_norm_16"  [label="(768,)", style=solid];
+"101 encoder_layers_encoder_layer_8_ln_1_bias" -> "440 layer_norm_16"  [label="(768,)", style=solid];
+"102 encoder_layers_encoder_layer_8_self_attention_in_proj_weight" -> "442 linear_32"  [label="(2304, 768)", style=solid];
+"103 encoder_layers_encoder_layer_8_self_attention_in_proj_bias" -> "442 linear_32"  [label="(2304,)", style=solid];
+"104 encoder_layers_encoder_layer_8_self_attention_out_proj_weight" -> "463 linear_33"  [label="(768, 768)", style=solid];
+"105 encoder_layers_encoder_layer_8_self_attention_out_proj_bias" -> "463 linear_33"  [label="(768,)", style=solid];
+"106 encoder_layers_encoder_layer_8_ln_2_weight" -> "468 layer_norm_17"  [label="(768,)", style=solid];
+"107 encoder_layers_encoder_layer_8_ln_2_bias" -> "468 layer_norm_17"  [label="(768,)", style=solid];
+"108 encoder_layers_encoder_layer_8_mlp_0_weight" -> "469 linear_34"  [label="(3072, 768)", style=solid];
+"109 encoder_layers_encoder_layer_8_mlp_0_bias" -> "469 linear_34"  [label="(3072,)", style=solid];
+"110 encoder_layers_encoder_layer_8_mlp_3_weight" -> "472 linear_35"  [label="(768, 3072)", style=solid];
+"111 encoder_layers_encoder_layer_8_mlp_3_bias" -> "472 linear_35"  [label="(768,)", style=solid];
+"112 encoder_layers_encoder_layer_9_ln_1_weight" -> "475 layer_norm_18"  [label="(768,)", style=solid];
+"113 encoder_layers_encoder_layer_9_ln_1_bias" -> "475 layer_norm_18"  [label="(768,)", style=solid];
+"114 encoder_layers_encoder_layer_9_self_attention_in_proj_weight" -> "477 linear_36"  [label="(2304, 768)", style=solid];
+"115 encoder_layers_encoder_layer_9_self_attention_in_proj_bias" -> "477 linear_36"  [label="(2304,)", style=solid];
+"116 encoder_layers_encoder_layer_9_self_attention_out_proj_weight" -> "498 linear_37"  [label="(768, 768)", style=solid];
+"117 encoder_layers_encoder_layer_9_self_attention_out_proj_bias" -> "498 linear_37"  [label="(768,)", style=solid];
+"118 encoder_layers_encoder_layer_9_ln_2_weight" -> "503 layer_norm_19"  [label="(768,)", style=solid];
+"119 encoder_layers_encoder_layer_9_ln_2_bias" -> "503 layer_norm_19"  [label="(768,)", style=solid];
+"120 encoder_layers_encoder_layer_9_mlp_0_weight" -> "504 linear_38"  [label="(3072, 768)", style=solid];
+"121 encoder_layers_encoder_layer_9_mlp_0_bias" -> "504 linear_38"  [label="(3072,)", style=solid];
+"122 encoder_layers_encoder_layer_9_mlp_3_weight" -> "507 linear_39"  [label="(768, 3072)", style=solid];
+"123 encoder_layers_encoder_layer_9_mlp_3_bias" -> "507 linear_39"  [label="(768,)", style=solid];
+"124 encoder_layers_encoder_layer_10_ln_1_weight" -> "510 layer_norm_20"  [label="(768,)", style=solid];
+"125 encoder_layers_encoder_layer_10_ln_1_bias" -> "510 layer_norm_20"  [label="(768,)", style=solid];
+"126 encoder_layers_encoder_layer_10_self_attention_in_proj_weight" -> "512 linear_40"  [label="(2304, 768)", style=solid];
+"127 encoder_layers_encoder_layer_10_self_attention_in_proj_bias" -> "512 linear_40"  [label="(2304,)", style=solid];
+"128 encoder_layers_encoder_layer_10_self_attention_out_proj_weight" -> "533 linear_41"  [label="(768, 768)", style=solid];
+"129 encoder_layers_encoder_layer_10_self_attention_out_proj_bias" -> "533 linear_41"  [label="(768,)", style=solid];
+"130 encoder_layers_encoder_layer_10_ln_2_weight" -> "538 layer_norm_21"  [label="(768,)", style=solid];
+"131 encoder_layers_encoder_layer_10_ln_2_bias" -> "538 layer_norm_21"  [label="(768,)", style=solid];
+"132 encoder_layers_encoder_layer_10_mlp_0_weight" -> "539 linear_42"  [label="(3072, 768)", style=solid];
+"133 encoder_layers_encoder_layer_10_mlp_0_bias" -> "539 linear_42"  [label="(3072,)", style=solid];
+"134 encoder_layers_encoder_layer_10_mlp_3_weight" -> "542 linear_43"  [label="(768, 3072)", style=solid];
+"135 encoder_layers_encoder_layer_10_mlp_3_bias" -> "542 linear_43"  [label="(768,)", style=solid];
+"136 encoder_layers_encoder_layer_11_ln_1_weight" -> "545 layer_norm_22"  [label="(768,)", style=solid];
+"137 encoder_layers_encoder_layer_11_ln_1_bias" -> "545 layer_norm_22"  [label="(768,)", style=solid];
+"138 encoder_layers_encoder_layer_11_self_attention_in_proj_weight" -> "547 linear_44"  [label="(2304, 768)", style=solid];
+"139 encoder_layers_encoder_layer_11_self_attention_in_proj_bias" -> "547 linear_44"  [label="(2304,)", style=solid];
+"140 encoder_layers_encoder_layer_11_self_attention_out_proj_weight" -> "568 linear_45"  [label="(768, 768)", style=solid];
+"141 encoder_layers_encoder_layer_11_self_attention_out_proj_bias" -> "568 linear_45"  [label="(768,)", style=solid];
+"142 encoder_layers_encoder_layer_11_ln_2_weight" -> "573 layer_norm_23"  [label="(768,)", style=solid];
+"143 encoder_layers_encoder_layer_11_ln_2_bias" -> "573 layer_norm_23"  [label="(768,)", style=solid];
+"144 encoder_layers_encoder_layer_11_mlp_0_weight" -> "574 linear_46"  [label="(3072, 768)", style=solid];
+"145 encoder_layers_encoder_layer_11_mlp_0_bias" -> "574 linear_46"  [label="(3072,)", style=solid];
+"146 encoder_layers_encoder_layer_11_mlp_3_weight" -> "577 linear_47"  [label="(768, 3072)", style=solid];
+"147 encoder_layers_encoder_layer_11_mlp_3_bias" -> "577 linear_47"  [label="(768,)", style=solid];
+"148 encoder_ln_weight" -> "580 layer_norm_24"  [label="(768,)", style=solid];
+"149 encoder_ln_bias" -> "580 layer_norm_24"  [label="(768,)", style=solid];
+"150 heads_head_weight" -> "583 linear_48"  [label="(1000, 768)", style=solid];
+"151 heads_head_bias" -> "583 linear_48"  [label="(1000,)", style=solid];
+"152 x" -> "153 conv2d"  [label="(1, 3, 224, 224)", style=solid];
+"153 conv2d" -> "154 reshape"  [label="(1, 768, 14, 14)", style=solid];
+"154 reshape" -> "155 permute"  [label="(1, 768, 196)", style=solid];
+"155 permute" -> "157 cat"  [label="(1, 196, 768)", style=solid];
+"156 expand" -> "157 cat"  [label="(1, 1, 768)", style=solid];
+"157 cat" -> "158 add"  [label="(1, 197, 768)", style=solid];
+"158 add" -> "159 dropout"  [label="(1, 197, 768)", style=solid];
+"159 dropout" -> "160 layer_norm"  [label="(1, 197, 768)", style=solid];
+"159 dropout" -> "187 add_1"  [label="(1, 197, 768)", style=solid];
+"160 layer_norm" -> "161 transpose"  [label="(1, 197, 768)", style=solid];
+"161 transpose" -> "162 linear"  [label="(197, 1, 768)", style=solid];
+"162 linear" -> "163 unflatten"  [label="(197, 1, 2304)", style=solid];
+"163 unflatten" -> "164 unsqueeze"  [label="(197, 1, 3, 768)", style=solid];
+"164 unsqueeze" -> "165 transpose_1"  [label="(1, 197, 1, 3, 768)", style=solid];
+"165 transpose_1" -> "166 squeeze"  [label="(3, 197, 1, 1, 768)", style=solid];
+"166 squeeze" -> "167 contiguous"  [label="(3, 197, 1, 768)", style=solid];
+"167 contiguous" -> "168 select"  [label="(3, 197, 1, 768)", style=solid];
+"167 contiguous" -> "169 select_1"  [label="(3, 197, 1, 768)", style=solid];
+"167 contiguous" -> "170 select_2"  [label="(3, 197, 1, 768)", style=solid];
+"168 select" -> "171 view"  [label="(197, 1, 768)", style=solid];
+"169 select_1" -> "173 view_1"  [label="(197, 1, 768)", style=solid];
+"170 select_2" -> "175 view_2"  [label="(197, 1, 768)", style=solid];
+"171 view" -> "172 transpose_2"  [label="(197, 12, 64)", style=solid];
+"172 transpose_2" -> "177 view_3"  [label="(12, 197, 64)", style=solid];
+"173 view_1" -> "174 transpose_3"  [label="(197, 12, 64)", style=solid];
+"174 transpose_3" -> "178 view_4"  [label="(12, 197, 64)", style=solid];
+"175 view_2" -> "176 transpose_4"  [label="(197, 12, 64)", style=solid];
+"176 transpose_4" -> "179 view_5"  [label="(12, 197, 64)", style=solid];
+"177 view_3" -> "180 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
+"178 view_4" -> "180 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
+"179 view_5" -> "180 scaled_dot_product_attention"  [label="(1, 12, 197, 64)", style=solid];
+"180 scaled_dot_product_attention" -> "181 permute_1"  [label="(1, 12, 197, 64)", style=solid];
+"181 permute_1" -> "182 view_6"  [label="(197, 1, 12, 64)", style=solid];
+"182 view_6" -> "183 linear_1"  [label="(197, 768)", style=solid];
+"183 linear_1" -> "184 view_7"  [label="(197, 768)", style=solid];
+"184 view_7" -> "185 transpose_5"  [label="(197, 1, 768)", style=solid];
+"185 transpose_5" -> "186 dropout_1"  [label="(1, 197, 768)", style=solid];
+"186 dropout_1" -> "187 add_1"  [label="(1, 197, 768)", style=solid];
+"187 add_1" -> "188 layer_norm_1"  [label="(1, 197, 768)", style=solid];
+"187 add_1" -> "194 add_2"  [label="(1, 197, 768)", style=solid];
+"188 layer_norm_1" -> "189 linear_2"  [label="(1, 197, 768)", style=solid];
+"189 linear_2" -> "190 gelu"  [label="(1, 197, 3072)", style=solid];
+"190 gelu" -> "191 dropout_2"  [label="(1, 197, 3072)", style=solid];
+"191 dropout_2" -> "192 linear_3"  [label="(1, 197, 3072)", style=solid];
+"192 linear_3" -> "193 dropout_3"  [label="(1, 197, 768)", style=solid];
+"193 dropout_3" -> "194 add_2"  [label="(1, 197, 768)", style=solid];
+"194 add_2" -> "195 layer_norm_2"  [label="(1, 197, 768)", style=solid];
+"194 add_2" -> "222 add_3"  [label="(1, 197, 768)", style=solid];
+"195 layer_norm_2" -> "196 transpose_6"  [label="(1, 197, 768)", style=solid];
+"196 transpose_6" -> "197 linear_4"  [label="(197, 1, 768)", style=solid];
+"197 linear_4" -> "198 unflatten_1"  [label="(197, 1, 2304)", style=solid];
+"198 unflatten_1" -> "199 unsqueeze_1"  [label="(197, 1, 3, 768)", style=solid];
+"199 unsqueeze_1" -> "200 transpose_7"  [label="(1, 197, 1, 3, 768)", style=solid];
+"200 transpose_7" -> "201 squeeze_1"  [label="(3, 197, 1, 1, 768)", style=solid];
+"201 squeeze_1" -> "202 contiguous_1"  [label="(3, 197, 1, 768)", style=solid];
+"202 contiguous_1" -> "203 select_3"  [label="(3, 197, 1, 768)", style=solid];
+"202 contiguous_1" -> "204 select_4"  [label="(3, 197, 1, 768)", style=solid];
+"202 contiguous_1" -> "205 select_5"  [label="(3, 197, 1, 768)", style=solid];
+"203 select_3" -> "206 view_8"  [label="(197, 1, 768)", style=solid];
+"204 select_4" -> "208 view_9"  [label="(197, 1, 768)", style=solid];
+"205 select_5" -> "210 view_10"  [label="(197, 1, 768)", style=solid];
+"206 view_8" -> "207 transpose_8"  [label="(197, 12, 64)", style=solid];
+"207 transpose_8" -> "212 view_11"  [label="(12, 197, 64)", style=solid];
+"208 view_9" -> "209 transpose_9"  [label="(197, 12, 64)", style=solid];
+"209 transpose_9" -> "213 view_12"  [label="(12, 197, 64)", style=solid];
+"210 view_10" -> "211 transpose_10"  [label="(197, 12, 64)", style=solid];
+"211 transpose_10" -> "214 view_13"  [label="(12, 197, 64)", style=solid];
+"212 view_11" -> "215 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
+"213 view_12" -> "215 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
+"214 view_13" -> "215 scaled_dot_product_attention_1"  [label="(1, 12, 197, 64)", style=solid];
+"215 scaled_dot_product_attention_1" -> "216 permute_2"  [label="(1, 12, 197, 64)", style=solid];
+"216 permute_2" -> "217 view_14"  [label="(197, 1, 12, 64)", style=solid];
+"217 view_14" -> "218 linear_5"  [label="(197, 768)", style=solid];
+"218 linear_5" -> "219 view_15"  [label="(197, 768)", style=solid];
+"219 view_15" -> "220 transpose_11"  [label="(197, 1, 768)", style=solid];
+"220 transpose_11" -> "221 dropout_4"  [label="(1, 197, 768)", style=solid];
+"221 dropout_4" -> "222 add_3"  [label="(1, 197, 768)", style=solid];
+"222 add_3" -> "223 layer_norm_3"  [label="(1, 197, 768)", style=solid];
+"222 add_3" -> "229 add_4"  [label="(1, 197, 768)", style=solid];
+"223 layer_norm_3" -> "224 linear_6"  [label="(1, 197, 768)", style=solid];
+"224 linear_6" -> "225 gelu_1"  [label="(1, 197, 3072)", style=solid];
+"225 gelu_1" -> "226 dropout_5"  [label="(1, 197, 3072)", style=solid];
+"226 dropout_5" -> "227 linear_7"  [label="(1, 197, 3072)", style=solid];
+"227 linear_7" -> "228 dropout_6"  [label="(1, 197, 768)", style=solid];
+"228 dropout_6" -> "229 add_4"  [label="(1, 197, 768)", style=solid];
+"229 add_4" -> "230 layer_norm_4"  [label="(1, 197, 768)", style=solid];
+"229 add_4" -> "257 add_5"  [label="(1, 197, 768)", style=solid];
+"230 layer_norm_4" -> "231 transpose_12"  [label="(1, 197, 768)", style=solid];
+"231 transpose_12" -> "232 linear_8"  [label="(197, 1, 768)", style=solid];
+"232 linear_8" -> "233 unflatten_2"  [label="(197, 1, 2304)", style=solid];
+"233 unflatten_2" -> "234 unsqueeze_2"  [label="(197, 1, 3, 768)", style=solid];
+"234 unsqueeze_2" -> "235 transpose_13"  [label="(1, 197, 1, 3, 768)", style=solid];
+"235 transpose_13" -> "236 squeeze_2"  [label="(3, 197, 1, 1, 768)", style=solid];
+"236 squeeze_2" -> "237 contiguous_2"  [label="(3, 197, 1, 768)", style=solid];
+"237 contiguous_2" -> "238 select_6"  [label="(3, 197, 1, 768)", style=solid];
+"237 contiguous_2" -> "239 select_7"  [label="(3, 197, 1, 768)", style=solid];
+"237 contiguous_2" -> "240 select_8"  [label="(3, 197, 1, 768)", style=solid];
+"238 select_6" -> "241 view_16"  [label="(197, 1, 768)", style=solid];
+"239 select_7" -> "243 view_17"  [label="(197, 1, 768)", style=solid];
+"240 select_8" -> "245 view_18"  [label="(197, 1, 768)", style=solid];
+"241 view_16" -> "242 transpose_14"  [label="(197, 12, 64)", style=solid];
+"242 transpose_14" -> "247 view_19"  [label="(12, 197, 64)", style=solid];
+"243 view_17" -> "244 transpose_15"  [label="(197, 12, 64)", style=solid];
+"244 transpose_15" -> "248 view_20"  [label="(12, 197, 64)", style=solid];
+"245 view_18" -> "246 transpose_16"  [label="(197, 12, 64)", style=solid];
+"246 transpose_16" -> "249 view_21"  [label="(12, 197, 64)", style=solid];
+"247 view_19" -> "250 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
+"248 view_20" -> "250 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
+"249 view_21" -> "250 scaled_dot_product_attention_2"  [label="(1, 12, 197, 64)", style=solid];
+"250 scaled_dot_product_attention_2" -> "251 permute_3"  [label="(1, 12, 197, 64)", style=solid];
+"251 permute_3" -> "252 view_22"  [label="(197, 1, 12, 64)", style=solid];
+"252 view_22" -> "253 linear_9"  [label="(197, 768)", style=solid];
+"253 linear_9" -> "254 view_23"  [label="(197, 768)", style=solid];
+"254 view_23" -> "255 transpose_17"  [label="(197, 1, 768)", style=solid];
+"255 transpose_17" -> "256 dropout_7"  [label="(1, 197, 768)", style=solid];
+"256 dropout_7" -> "257 add_5"  [label="(1, 197, 768)", style=solid];
+"257 add_5" -> "258 layer_norm_5"  [label="(1, 197, 768)", style=solid];
+"257 add_5" -> "264 add_6"  [label="(1, 197, 768)", style=solid];
+"258 layer_norm_5" -> "259 linear_10"  [label="(1, 197, 768)", style=solid];
+"259 linear_10" -> "260 gelu_2"  [label="(1, 197, 3072)", style=solid];
+"260 gelu_2" -> "261 dropout_8"  [label="(1, 197, 3072)", style=solid];
+"261 dropout_8" -> "262 linear_11"  [label="(1, 197, 3072)", style=solid];
+"262 linear_11" -> "263 dropout_9"  [label="(1, 197, 768)", style=solid];
+"263 dropout_9" -> "264 add_6"  [label="(1, 197, 768)", style=solid];
+"264 add_6" -> "265 layer_norm_6"  [label="(1, 197, 768)", style=solid];
+"264 add_6" -> "292 add_7"  [label="(1, 197, 768)", style=solid];
+"265 layer_norm_6" -> "266 transpose_18"  [label="(1, 197, 768)", style=solid];
+"266 transpose_18" -> "267 linear_12"  [label="(197, 1, 768)", style=solid];
+"267 linear_12" -> "268 unflatten_3"  [label="(197, 1, 2304)", style=solid];
+"268 unflatten_3" -> "269 unsqueeze_3"  [label="(197, 1, 3, 768)", style=solid];
+"269 unsqueeze_3" -> "270 transpose_19"  [label="(1, 197, 1, 3, 768)", style=solid];
+"270 transpose_19" -> "271 squeeze_3"  [label="(3, 197, 1, 1, 768)", style=solid];
+"271 squeeze_3" -> "272 contiguous_3"  [label="(3, 197, 1, 768)", style=solid];
+"272 contiguous_3" -> "273 select_9"  [label="(3, 197, 1, 768)", style=solid];
+"272 contiguous_3" -> "274 select_10"  [label="(3, 197, 1, 768)", style=solid];
+"272 contiguous_3" -> "275 select_11"  [label="(3, 197, 1, 768)", style=solid];
+"273 select_9" -> "276 view_24"  [label="(197, 1, 768)", style=solid];
+"274 select_10" -> "278 view_25"  [label="(197, 1, 768)", style=solid];
+"275 select_11" -> "280 view_26"  [label="(197, 1, 768)", style=solid];
+"276 view_24" -> "277 transpose_20"  [label="(197, 12, 64)", style=solid];
+"277 transpose_20" -> "282 view_27"  [label="(12, 197, 64)", style=solid];
+"278 view_25" -> "279 transpose_21"  [label="(197, 12, 64)", style=solid];
+"279 transpose_21" -> "283 view_28"  [label="(12, 197, 64)", style=solid];
+"280 view_26" -> "281 transpose_22"  [label="(197, 12, 64)", style=solid];
+"281 transpose_22" -> "284 view_29"  [label="(12, 197, 64)", style=solid];
+"282 view_27" -> "285 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
+"283 view_28" -> "285 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
+"284 view_29" -> "285 scaled_dot_product_attention_3"  [label="(1, 12, 197, 64)", style=solid];
+"285 scaled_dot_product_attention_3" -> "286 permute_4"  [label="(1, 12, 197, 64)", style=solid];
+"286 permute_4" -> "287 view_30"  [label="(197, 1, 12, 64)", style=solid];
+"287 view_30" -> "288 linear_13"  [label="(197, 768)", style=solid];
+"288 linear_13" -> "289 view_31"  [label="(197, 768)", style=solid];
+"289 view_31" -> "290 transpose_23"  [label="(197, 1, 768)", style=solid];
+"290 transpose_23" -> "291 dropout_10"  [label="(1, 197, 768)", style=solid];
+"291 dropout_10" -> "292 add_7"  [label="(1, 197, 768)", style=solid];
+"292 add_7" -> "293 layer_norm_7"  [label="(1, 197, 768)", style=solid];
+"292 add_7" -> "299 add_8"  [label="(1, 197, 768)", style=solid];
+"293 layer_norm_7" -> "294 linear_14"  [label="(1, 197, 768)", style=solid];
+"294 linear_14" -> "295 gelu_3"  [label="(1, 197, 3072)", style=solid];
+"295 gelu_3" -> "296 dropout_11"  [label="(1, 197, 3072)", style=solid];
+"296 dropout_11" -> "297 linear_15"  [label="(1, 197, 3072)", style=solid];
+"297 linear_15" -> "298 dropout_12"  [label="(1, 197, 768)", style=solid];
+"298 dropout_12" -> "299 add_8"  [label="(1, 197, 768)", style=solid];
+"299 add_8" -> "300 layer_norm_8"  [label="(1, 197, 768)", style=solid];
+"299 add_8" -> "327 add_9"  [label="(1, 197, 768)", style=solid];
+"300 layer_norm_8" -> "301 transpose_24"  [label="(1, 197, 768)", style=solid];
+"301 transpose_24" -> "302 linear_16"  [label="(197, 1, 768)", style=solid];
+"302 linear_16" -> "303 unflatten_4"  [label="(197, 1, 2304)", style=solid];
+"303 unflatten_4" -> "304 unsqueeze_4"  [label="(197, 1, 3, 768)", style=solid];
+"304 unsqueeze_4" -> "305 transpose_25"  [label="(1, 197, 1, 3, 768)", style=solid];
+"305 transpose_25" -> "306 squeeze_4"  [label="(3, 197, 1, 1, 768)", style=solid];
+"306 squeeze_4" -> "307 contiguous_4"  [label="(3, 197, 1, 768)", style=solid];
+"307 contiguous_4" -> "308 select_12"  [label="(3, 197, 1, 768)", style=solid];
+"307 contiguous_4" -> "309 select_13"  [label="(3, 197, 1, 768)", style=solid];
+"307 contiguous_4" -> "310 select_14"  [label="(3, 197, 1, 768)", style=solid];
+"308 select_12" -> "311 view_32"  [label="(197, 1, 768)", style=solid];
+"309 select_13" -> "313 view_33"  [label="(197, 1, 768)", style=solid];
+"310 select_14" -> "315 view_34"  [label="(197, 1, 768)", style=solid];
+"311 view_32" -> "312 transpose_26"  [label="(197, 12, 64)", style=solid];
+"312 transpose_26" -> "317 view_35"  [label="(12, 197, 64)", style=solid];
+"313 view_33" -> "314 transpose_27"  [label="(197, 12, 64)", style=solid];
+"314 transpose_27" -> "318 view_36"  [label="(12, 197, 64)", style=solid];
+"315 view_34" -> "316 transpose_28"  [label="(197, 12, 64)", style=solid];
+"316 transpose_28" -> "319 view_37"  [label="(12, 197, 64)", style=solid];
+"317 view_35" -> "320 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
+"318 view_36" -> "320 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
+"319 view_37" -> "320 scaled_dot_product_attention_4"  [label="(1, 12, 197, 64)", style=solid];
+"320 scaled_dot_product_attention_4" -> "321 permute_5"  [label="(1, 12, 197, 64)", style=solid];
+"321 permute_5" -> "322 view_38"  [label="(197, 1, 12, 64)", style=solid];
+"322 view_38" -> "323 linear_17"  [label="(197, 768)", style=solid];
+"323 linear_17" -> "324 view_39"  [label="(197, 768)", style=solid];
+"324 view_39" -> "325 transpose_29"  [label="(197, 1, 768)", style=solid];
+"325 transpose_29" -> "326 dropout_13"  [label="(1, 197, 768)", style=solid];
+"326 dropout_13" -> "327 add_9"  [label="(1, 197, 768)", style=solid];
+"327 add_9" -> "328 layer_norm_9"  [label="(1, 197, 768)", style=solid];
+"327 add_9" -> "334 add_10"  [label="(1, 197, 768)", style=solid];
+"328 layer_norm_9" -> "329 linear_18"  [label="(1, 197, 768)", style=solid];
+"329 linear_18" -> "330 gelu_4"  [label="(1, 197, 3072)", style=solid];
+"330 gelu_4" -> "331 dropout_14"  [label="(1, 197, 3072)", style=solid];
+"331 dropout_14" -> "332 linear_19"  [label="(1, 197, 3072)", style=solid];
+"332 linear_19" -> "333 dropout_15"  [label="(1, 197, 768)", style=solid];
+"333 dropout_15" -> "334 add_10"  [label="(1, 197, 768)", style=solid];
+"334 add_10" -> "335 layer_norm_10"  [label="(1, 197, 768)", style=solid];
+"334 add_10" -> "362 add_11"  [label="(1, 197, 768)", style=solid];
+"335 layer_norm_10" -> "336 transpose_30"  [label="(1, 197, 768)", style=solid];
+"336 transpose_30" -> "337 linear_20"  [label="(197, 1, 768)", style=solid];
+"337 linear_20" -> "338 unflatten_5"  [label="(197, 1, 2304)", style=solid];
+"338 unflatten_5" -> "339 unsqueeze_5"  [label="(197, 1, 3, 768)", style=solid];
+"339 unsqueeze_5" -> "340 transpose_31"  [label="(1, 197, 1, 3, 768)", style=solid];
+"340 transpose_31" -> "341 squeeze_5"  [label="(3, 197, 1, 1, 768)", style=solid];
+"341 squeeze_5" -> "342 contiguous_5"  [label="(3, 197, 1, 768)", style=solid];
+"342 contiguous_5" -> "343 select_15"  [label="(3, 197, 1, 768)", style=solid];
+"342 contiguous_5" -> "344 select_16"  [label="(3, 197, 1, 768)", style=solid];
+"342 contiguous_5" -> "345 select_17"  [label="(3, 197, 1, 768)", style=solid];
+"343 select_15" -> "346 view_40"  [label="(197, 1, 768)", style=solid];
+"344 select_16" -> "348 view_41"  [label="(197, 1, 768)", style=solid];
+"345 select_17" -> "350 view_42"  [label="(197, 1, 768)", style=solid];
+"346 view_40" -> "347 transpose_32"  [label="(197, 12, 64)", style=solid];
+"347 transpose_32" -> "352 view_43"  [label="(12, 197, 64)", style=solid];
+"348 view_41" -> "349 transpose_33"  [label="(197, 12, 64)", style=solid];
+"349 transpose_33" -> "353 view_44"  [label="(12, 197, 64)", style=solid];
+"350 view_42" -> "351 transpose_34"  [label="(197, 12, 64)", style=solid];
+"351 transpose_34" -> "354 view_45"  [label="(12, 197, 64)", style=solid];
+"352 view_43" -> "355 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
+"353 view_44" -> "355 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
+"354 view_45" -> "355 scaled_dot_product_attention_5"  [label="(1, 12, 197, 64)", style=solid];
+"355 scaled_dot_product_attention_5" -> "356 permute_6"  [label="(1, 12, 197, 64)", style=solid];
+"356 permute_6" -> "357 view_46"  [label="(197, 1, 12, 64)", style=solid];
+"357 view_46" -> "358 linear_21"  [label="(197, 768)", style=solid];
+"358 linear_21" -> "359 view_47"  [label="(197, 768)", style=solid];
+"359 view_47" -> "360 transpose_35"  [label="(197, 1, 768)", style=solid];
+"360 transpose_35" -> "361 dropout_16"  [label="(1, 197, 768)", style=solid];
+"361 dropout_16" -> "362 add_11"  [label="(1, 197, 768)", style=solid];
+"362 add_11" -> "363 layer_norm_11"  [label="(1, 197, 768)", style=solid];
+"362 add_11" -> "369 add_12"  [label="(1, 197, 768)", style=solid];
+"363 layer_norm_11" -> "364 linear_22"  [label="(1, 197, 768)", style=solid];
+"364 linear_22" -> "365 gelu_5"  [label="(1, 197, 3072)", style=solid];
+"365 gelu_5" -> "366 dropout_17"  [label="(1, 197, 3072)", style=solid];
+"366 dropout_17" -> "367 linear_23"  [label="(1, 197, 3072)", style=solid];
+"367 linear_23" -> "368 dropout_18"  [label="(1, 197, 768)", style=solid];
+"368 dropout_18" -> "369 add_12"  [label="(1, 197, 768)", style=solid];
+"369 add_12" -> "370 layer_norm_12"  [label="(1, 197, 768)", style=solid];
+"369 add_12" -> "397 add_13"  [label="(1, 197, 768)", style=solid];
+"370 layer_norm_12" -> "371 transpose_36"  [label="(1, 197, 768)", style=solid];
+"371 transpose_36" -> "372 linear_24"  [label="(197, 1, 768)", style=solid];
+"372 linear_24" -> "373 unflatten_6"  [label="(197, 1, 2304)", style=solid];
+"373 unflatten_6" -> "374 unsqueeze_6"  [label="(197, 1, 3, 768)", style=solid];
+"374 unsqueeze_6" -> "375 transpose_37"  [label="(1, 197, 1, 3, 768)", style=solid];
+"375 transpose_37" -> "376 squeeze_6"  [label="(3, 197, 1, 1, 768)", style=solid];
+"376 squeeze_6" -> "377 contiguous_6"  [label="(3, 197, 1, 768)", style=solid];
+"377 contiguous_6" -> "378 select_18"  [label="(3, 197, 1, 768)", style=solid];
+"377 contiguous_6" -> "379 select_19"  [label="(3, 197, 1, 768)", style=solid];
+"377 contiguous_6" -> "380 select_20"  [label="(3, 197, 1, 768)", style=solid];
+"378 select_18" -> "381 view_48"  [label="(197, 1, 768)", style=solid];
+"379 select_19" -> "383 view_49"  [label="(197, 1, 768)", style=solid];
+"380 select_20" -> "385 view_50"  [label="(197, 1, 768)", style=solid];
+"381 view_48" -> "382 transpose_38"  [label="(197, 12, 64)", style=solid];
+"382 transpose_38" -> "387 view_51"  [label="(12, 197, 64)", style=solid];
+"383 view_49" -> "384 transpose_39"  [label="(197, 12, 64)", style=solid];
+"384 transpose_39" -> "388 view_52"  [label="(12, 197, 64)", style=solid];
+"385 view_50" -> "386 transpose_40"  [label="(197, 12, 64)", style=solid];
+"386 transpose_40" -> "389 view_53"  [label="(12, 197, 64)", style=solid];
+"387 view_51" -> "390 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
+"388 view_52" -> "390 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
+"389 view_53" -> "390 scaled_dot_product_attention_6"  [label="(1, 12, 197, 64)", style=solid];
+"390 scaled_dot_product_attention_6" -> "391 permute_7"  [label="(1, 12, 197, 64)", style=solid];
+"391 permute_7" -> "392 view_54"  [label="(197, 1, 12, 64)", style=solid];
+"392 view_54" -> "393 linear_25"  [label="(197, 768)", style=solid];
+"393 linear_25" -> "394 view_55"  [label="(197, 768)", style=solid];
+"394 view_55" -> "395 transpose_41"  [label="(197, 1, 768)", style=solid];
+"395 transpose_41" -> "396 dropout_19"  [label="(1, 197, 768)", style=solid];
+"396 dropout_19" -> "397 add_13"  [label="(1, 197, 768)", style=solid];
+"397 add_13" -> "398 layer_norm_13"  [label="(1, 197, 768)", style=solid];
+"397 add_13" -> "404 add_14"  [label="(1, 197, 768)", style=solid];
+"398 layer_norm_13" -> "399 linear_26"  [label="(1, 197, 768)", style=solid];
+"399 linear_26" -> "400 gelu_6"  [label="(1, 197, 3072)", style=solid];
+"400 gelu_6" -> "401 dropout_20"  [label="(1, 197, 3072)", style=solid];
+"401 dropout_20" -> "402 linear_27"  [label="(1, 197, 3072)", style=solid];
+"402 linear_27" -> "403 dropout_21"  [label="(1, 197, 768)", style=solid];
+"403 dropout_21" -> "404 add_14"  [label="(1, 197, 768)", style=solid];
+"404 add_14" -> "405 layer_norm_14"  [label="(1, 197, 768)", style=solid];
+"404 add_14" -> "432 add_15"  [label="(1, 197, 768)", style=solid];
+"405 layer_norm_14" -> "406 transpose_42"  [label="(1, 197, 768)", style=solid];
+"406 transpose_42" -> "407 linear_28"  [label="(197, 1, 768)", style=solid];
+"407 linear_28" -> "408 unflatten_7"  [label="(197, 1, 2304)", style=solid];
+"408 unflatten_7" -> "409 unsqueeze_7"  [label="(197, 1, 3, 768)", style=solid];
+"409 unsqueeze_7" -> "410 transpose_43"  [label="(1, 197, 1, 3, 768)", style=solid];
+"410 transpose_43" -> "411 squeeze_7"  [label="(3, 197, 1, 1, 768)", style=solid];
+"411 squeeze_7" -> "412 contiguous_7"  [label="(3, 197, 1, 768)", style=solid];
+"412 contiguous_7" -> "413 select_21"  [label="(3, 197, 1, 768)", style=solid];
+"412 contiguous_7" -> "414 select_22"  [label="(3, 197, 1, 768)", style=solid];
+"412 contiguous_7" -> "415 select_23"  [label="(3, 197, 1, 768)", style=solid];
+"413 select_21" -> "416 view_56"  [label="(197, 1, 768)", style=solid];
+"414 select_22" -> "418 view_57"  [label="(197, 1, 768)", style=solid];
+"415 select_23" -> "420 view_58"  [label="(197, 1, 768)", style=solid];
+"416 view_56" -> "417 transpose_44"  [label="(197, 12, 64)", style=solid];
+"417 transpose_44" -> "422 view_59"  [label="(12, 197, 64)", style=solid];
+"418 view_57" -> "419 transpose_45"  [label="(197, 12, 64)", style=solid];
+"419 transpose_45" -> "423 view_60"  [label="(12, 197, 64)", style=solid];
+"420 view_58" -> "421 transpose_46"  [label="(197, 12, 64)", style=solid];
+"421 transpose_46" -> "424 view_61"  [label="(12, 197, 64)", style=solid];
+"422 view_59" -> "425 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
+"423 view_60" -> "425 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
+"424 view_61" -> "425 scaled_dot_product_attention_7"  [label="(1, 12, 197, 64)", style=solid];
+"425 scaled_dot_product_attention_7" -> "426 permute_8"  [label="(1, 12, 197, 64)", style=solid];
+"426 permute_8" -> "427 view_62"  [label="(197, 1, 12, 64)", style=solid];
+"427 view_62" -> "428 linear_29"  [label="(197, 768)", style=solid];
+"428 linear_29" -> "429 view_63"  [label="(197, 768)", style=solid];
+"429 view_63" -> "430 transpose_47"  [label="(197, 1, 768)", style=solid];
+"430 transpose_47" -> "431 dropout_22"  [label="(1, 197, 768)", style=solid];
+"431 dropout_22" -> "432 add_15"  [label="(1, 197, 768)", style=solid];
+"432 add_15" -> "433 layer_norm_15"  [label="(1, 197, 768)", style=solid];
+"432 add_15" -> "439 add_16"  [label="(1, 197, 768)", style=solid];
+"433 layer_norm_15" -> "434 linear_30"  [label="(1, 197, 768)", style=solid];
+"434 linear_30" -> "435 gelu_7"  [label="(1, 197, 3072)", style=solid];
+"435 gelu_7" -> "436 dropout_23"  [label="(1, 197, 3072)", style=solid];
+"436 dropout_23" -> "437 linear_31"  [label="(1, 197, 3072)", style=solid];
+"437 linear_31" -> "438 dropout_24"  [label="(1, 197, 768)", style=solid];
+"438 dropout_24" -> "439 add_16"  [label="(1, 197, 768)", style=solid];
+"439 add_16" -> "440 layer_norm_16"  [label="(1, 197, 768)", style=solid];
+"439 add_16" -> "467 add_17"  [label="(1, 197, 768)", style=solid];
+"440 layer_norm_16" -> "441 transpose_48"  [label="(1, 197, 768)", style=solid];
+"441 transpose_48" -> "442 linear_32"  [label="(197, 1, 768)", style=solid];
+"442 linear_32" -> "443 unflatten_8"  [label="(197, 1, 2304)", style=solid];
+"443 unflatten_8" -> "444 unsqueeze_8"  [label="(197, 1, 3, 768)", style=solid];
+"444 unsqueeze_8" -> "445 transpose_49"  [label="(1, 197, 1, 3, 768)", style=solid];
+"445 transpose_49" -> "446 squeeze_8"  [label="(3, 197, 1, 1, 768)", style=solid];
+"446 squeeze_8" -> "447 contiguous_8"  [label="(3, 197, 1, 768)", style=solid];
+"447 contiguous_8" -> "448 select_24"  [label="(3, 197, 1, 768)", style=solid];
+"447 contiguous_8" -> "449 select_25"  [label="(3, 197, 1, 768)", style=solid];
+"447 contiguous_8" -> "450 select_26"  [label="(3, 197, 1, 768)", style=solid];
+"448 select_24" -> "451 view_64"  [label="(197, 1, 768)", style=solid];
+"449 select_25" -> "453 view_65"  [label="(197, 1, 768)", style=solid];
+"450 select_26" -> "455 view_66"  [label="(197, 1, 768)", style=solid];
+"451 view_64" -> "452 transpose_50"  [label="(197, 12, 64)", style=solid];
+"452 transpose_50" -> "457 view_67"  [label="(12, 197, 64)", style=solid];
+"453 view_65" -> "454 transpose_51"  [label="(197, 12, 64)", style=solid];
+"454 transpose_51" -> "458 view_68"  [label="(12, 197, 64)", style=solid];
+"455 view_66" -> "456 transpose_52"  [label="(197, 12, 64)", style=solid];
+"456 transpose_52" -> "459 view_69"  [label="(12, 197, 64)", style=solid];
+"457 view_67" -> "460 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
+"458 view_68" -> "460 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
+"459 view_69" -> "460 scaled_dot_product_attention_8"  [label="(1, 12, 197, 64)", style=solid];
+"460 scaled_dot_product_attention_8" -> "461 permute_9"  [label="(1, 12, 197, 64)", style=solid];
+"461 permute_9" -> "462 view_70"  [label="(197, 1, 12, 64)", style=solid];
+"462 view_70" -> "463 linear_33"  [label="(197, 768)", style=solid];
+"463 linear_33" -> "464 view_71"  [label="(197, 768)", style=solid];
+"464 view_71" -> "465 transpose_53"  [label="(197, 1, 768)", style=solid];
+"465 transpose_53" -> "466 dropout_25"  [label="(1, 197, 768)", style=solid];
+"466 dropout_25" -> "467 add_17"  [label="(1, 197, 768)", style=solid];
+"467 add_17" -> "468 layer_norm_17"  [label="(1, 197, 768)", style=solid];
+"467 add_17" -> "474 add_18"  [label="(1, 197, 768)", style=solid];
+"468 layer_norm_17" -> "469 linear_34"  [label="(1, 197, 768)", style=solid];
+"469 linear_34" -> "470 gelu_8"  [label="(1, 197, 3072)", style=solid];
+"470 gelu_8" -> "471 dropout_26"  [label="(1, 197, 3072)", style=solid];
+"471 dropout_26" -> "472 linear_35"  [label="(1, 197, 3072)", style=solid];
+"472 linear_35" -> "473 dropout_27"  [label="(1, 197, 768)", style=solid];
+"473 dropout_27" -> "474 add_18"  [label="(1, 197, 768)", style=solid];
+"474 add_18" -> "475 layer_norm_18"  [label="(1, 197, 768)", style=solid];
+"474 add_18" -> "502 add_19"  [label="(1, 197, 768)", style=solid];
+"475 layer_norm_18" -> "476 transpose_54"  [label="(1, 197, 768)", style=solid];
+"476 transpose_54" -> "477 linear_36"  [label="(197, 1, 768)", style=solid];
+"477 linear_36" -> "478 unflatten_9"  [label="(197, 1, 2304)", style=solid];
+"478 unflatten_9" -> "479 unsqueeze_9"  [label="(197, 1, 3, 768)", style=solid];
+"479 unsqueeze_9" -> "480 transpose_55"  [label="(1, 197, 1, 3, 768)", style=solid];
+"480 transpose_55" -> "481 squeeze_9"  [label="(3, 197, 1, 1, 768)", style=solid];
+"481 squeeze_9" -> "482 contiguous_9"  [label="(3, 197, 1, 768)", style=solid];
+"482 contiguous_9" -> "483 select_27"  [label="(3, 197, 1, 768)", style=solid];
+"482 contiguous_9" -> "484 select_28"  [label="(3, 197, 1, 768)", style=solid];
+"482 contiguous_9" -> "485 select_29"  [label="(3, 197, 1, 768)", style=solid];
+"483 select_27" -> "486 view_72"  [label="(197, 1, 768)", style=solid];
+"484 select_28" -> "488 view_73"  [label="(197, 1, 768)", style=solid];
+"485 select_29" -> "490 view_74"  [label="(197, 1, 768)", style=solid];
+"486 view_72" -> "487 transpose_56"  [label="(197, 12, 64)", style=solid];
+"487 transpose_56" -> "492 view_75"  [label="(12, 197, 64)", style=solid];
+"488 view_73" -> "489 transpose_57"  [label="(197, 12, 64)", style=solid];
+"489 transpose_57" -> "493 view_76"  [label="(12, 197, 64)", style=solid];
+"490 view_74" -> "491 transpose_58"  [label="(197, 12, 64)", style=solid];
+"491 transpose_58" -> "494 view_77"  [label="(12, 197, 64)", style=solid];
+"492 view_75" -> "495 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
+"493 view_76" -> "495 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
+"494 view_77" -> "495 scaled_dot_product_attention_9"  [label="(1, 12, 197, 64)", style=solid];
+"495 scaled_dot_product_attention_9" -> "496 permute_10"  [label="(1, 12, 197, 64)", style=solid];
+"496 permute_10" -> "497 view_78"  [label="(197, 1, 12, 64)", style=solid];
+"497 view_78" -> "498 linear_37"  [label="(197, 768)", style=solid];
+"498 linear_37" -> "499 view_79"  [label="(197, 768)", style=solid];
+"499 view_79" -> "500 transpose_59"  [label="(197, 1, 768)", style=solid];
+"500 transpose_59" -> "501 dropout_28"  [label="(1, 197, 768)", style=solid];
+"501 dropout_28" -> "502 add_19"  [label="(1, 197, 768)", style=solid];
+"502 add_19" -> "503 layer_norm_19"  [label="(1, 197, 768)", style=solid];
+"502 add_19" -> "509 add_20"  [label="(1, 197, 768)", style=solid];
+"503 layer_norm_19" -> "504 linear_38"  [label="(1, 197, 768)", style=solid];
+"504 linear_38" -> "505 gelu_9"  [label="(1, 197, 3072)", style=solid];
+"505 gelu_9" -> "506 dropout_29"  [label="(1, 197, 3072)", style=solid];
+"506 dropout_29" -> "507 linear_39"  [label="(1, 197, 3072)", style=solid];
+"507 linear_39" -> "508 dropout_30"  [label="(1, 197, 768)", style=solid];
+"508 dropout_30" -> "509 add_20"  [label="(1, 197, 768)", style=solid];
+"509 add_20" -> "510 layer_norm_20"  [label="(1, 197, 768)", style=solid];
+"509 add_20" -> "537 add_21"  [label="(1, 197, 768)", style=solid];
+"510 layer_norm_20" -> "511 transpose_60"  [label="(1, 197, 768)", style=solid];
+"511 transpose_60" -> "512 linear_40"  [label="(197, 1, 768)", style=solid];
+"512 linear_40" -> "513 unflatten_10"  [label="(197, 1, 2304)", style=solid];
+"513 unflatten_10" -> "514 unsqueeze_10"  [label="(197, 1, 3, 768)", style=solid];
+"514 unsqueeze_10" -> "515 transpose_61"  [label="(1, 197, 1, 3, 768)", style=solid];
+"515 transpose_61" -> "516 squeeze_10"  [label="(3, 197, 1, 1, 768)", style=solid];
+"516 squeeze_10" -> "517 contiguous_10"  [label="(3, 197, 1, 768)", style=solid];
+"517 contiguous_10" -> "518 select_30"  [label="(3, 197, 1, 768)", style=solid];
+"517 contiguous_10" -> "519 select_31"  [label="(3, 197, 1, 768)", style=solid];
+"517 contiguous_10" -> "520 select_32"  [label="(3, 197, 1, 768)", style=solid];
+"518 select_30" -> "521 view_80"  [label="(197, 1, 768)", style=solid];
+"519 select_31" -> "523 view_81"  [label="(197, 1, 768)", style=solid];
+"520 select_32" -> "525 view_82"  [label="(197, 1, 768)", style=solid];
+"521 view_80" -> "522 transpose_62"  [label="(197, 12, 64)", style=solid];
+"522 transpose_62" -> "527 view_83"  [label="(12, 197, 64)", style=solid];
+"523 view_81" -> "524 transpose_63"  [label="(197, 12, 64)", style=solid];
+"524 transpose_63" -> "528 view_84"  [label="(12, 197, 64)", style=solid];
+"525 view_82" -> "526 transpose_64"  [label="(197, 12, 64)", style=solid];
+"526 transpose_64" -> "529 view_85"  [label="(12, 197, 64)", style=solid];
+"527 view_83" -> "530 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
+"528 view_84" -> "530 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
+"529 view_85" -> "530 scaled_dot_product_attention_10"  [label="(1, 12, 197, 64)", style=solid];
+"530 scaled_dot_product_attention_10" -> "531 permute_11"  [label="(1, 12, 197, 64)", style=solid];
+"531 permute_11" -> "532 view_86"  [label="(197, 1, 12, 64)", style=solid];
+"532 view_86" -> "533 linear_41"  [label="(197, 768)", style=solid];
+"533 linear_41" -> "534 view_87"  [label="(197, 768)", style=solid];
+"534 view_87" -> "535 transpose_65"  [label="(197, 1, 768)", style=solid];
+"535 transpose_65" -> "536 dropout_31"  [label="(1, 197, 768)", style=solid];
+"536 dropout_31" -> "537 add_21"  [label="(1, 197, 768)", style=solid];
+"537 add_21" -> "538 layer_norm_21"  [label="(1, 197, 768)", style=solid];
+"537 add_21" -> "544 add_22"  [label="(1, 197, 768)", style=solid];
+"538 layer_norm_21" -> "539 linear_42"  [label="(1, 197, 768)", style=solid];
+"539 linear_42" -> "540 gelu_10"  [label="(1, 197, 3072)", style=solid];
+"540 gelu_10" -> "541 dropout_32"  [label="(1, 197, 3072)", style=solid];
+"541 dropout_32" -> "542 linear_43"  [label="(1, 197, 3072)", style=solid];
+"542 linear_43" -> "543 dropout_33"  [label="(1, 197, 768)", style=solid];
+"543 dropout_33" -> "544 add_22"  [label="(1, 197, 768)", style=solid];
+"544 add_22" -> "545 layer_norm_22"  [label="(1, 197, 768)", style=solid];
+"544 add_22" -> "572 add_23"  [label="(1, 197, 768)", style=solid];
+"545 layer_norm_22" -> "546 transpose_66"  [label="(1, 197, 768)", style=solid];
+"546 transpose_66" -> "547 linear_44"  [label="(197, 1, 768)", style=solid];
+"547 linear_44" -> "548 unflatten_11"  [label="(197, 1, 2304)", style=solid];
+"548 unflatten_11" -> "549 unsqueeze_11"  [label="(197, 1, 3, 768)", style=solid];
+"549 unsqueeze_11" -> "550 transpose_67"  [label="(1, 197, 1, 3, 768)", style=solid];
+"550 transpose_67" -> "551 squeeze_11"  [label="(3, 197, 1, 1, 768)", style=solid];
+"551 squeeze_11" -> "552 contiguous_11"  [label="(3, 197, 1, 768)", style=solid];
+"552 contiguous_11" -> "553 select_33"  [label="(3, 197, 1, 768)", style=solid];
+"552 contiguous_11" -> "554 select_34"  [label="(3, 197, 1, 768)", style=solid];
+"552 contiguous_11" -> "555 select_35"  [label="(3, 197, 1, 768)", style=solid];
+"553 select_33" -> "556 view_88"  [label="(197, 1, 768)", style=solid];
+"554 select_34" -> "558 view_89"  [label="(197, 1, 768)", style=solid];
+"555 select_35" -> "560 view_90"  [label="(197, 1, 768)", style=solid];
+"556 view_88" -> "557 transpose_68"  [label="(197, 12, 64)", style=solid];
+"557 transpose_68" -> "562 view_91"  [label="(12, 197, 64)", style=solid];
+"558 view_89" -> "559 transpose_69"  [label="(197, 12, 64)", style=solid];
+"559 transpose_69" -> "563 view_92"  [label="(12, 197, 64)", style=solid];
+"560 view_90" -> "561 transpose_70"  [label="(197, 12, 64)", style=solid];
+"561 transpose_70" -> "564 view_93"  [label="(12, 197, 64)", style=solid];
+"562 view_91" -> "565 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
+"563 view_92" -> "565 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
+"564 view_93" -> "565 scaled_dot_product_attention_11"  [label="(1, 12, 197, 64)", style=solid];
+"565 scaled_dot_product_attention_11" -> "566 permute_12"  [label="(1, 12, 197, 64)", style=solid];
+"566 permute_12" -> "567 view_94"  [label="(197, 1, 12, 64)", style=solid];
+"567 view_94" -> "568 linear_45"  [label="(197, 768)", style=solid];
+"568 linear_45" -> "569 view_95"  [label="(197, 768)", style=solid];
+"569 view_95" -> "570 transpose_71"  [label="(197, 1, 768)", style=solid];
+"570 transpose_71" -> "571 dropout_34"  [label="(1, 197, 768)", style=solid];
+"571 dropout_34" -> "572 add_23"  [label="(1, 197, 768)", style=solid];
+"572 add_23" -> "573 layer_norm_23"  [label="(1, 197, 768)", style=solid];
+"572 add_23" -> "579 add_24"  [label="(1, 197, 768)", style=solid];
+"573 layer_norm_23" -> "574 linear_46"  [label="(1, 197, 768)", style=solid];
+"574 linear_46" -> "575 gelu_11"  [label="(1, 197, 3072)", style=solid];
+"575 gelu_11" -> "576 dropout_35"  [label="(1, 197, 3072)", style=solid];
+"576 dropout_35" -> "577 linear_47"  [label="(1, 197, 3072)", style=solid];
+"577 linear_47" -> "578 dropout_36"  [label="(1, 197, 768)", style=solid];
+"578 dropout_36" -> "579 add_24"  [label="(1, 197, 768)", style=solid];
+"579 add_24" -> "580 layer_norm_24"  [label="(1, 197, 768)", style=solid];
+"580 layer_norm_24" -> "581 slice_1"  [label="(1, 197, 768)", style=solid];
+"581 slice_1" -> "582 select_36"  [label="(1, 197, 768)", style=solid];
+"582 select_36" -> "583 linear_48"  [label="(1, 768)", style=solid];
+"583 linear_48" -> "584 output_1"  [label="(1, 1000)", style=solid];
 }
diff --git a/tests/torch/fx/__init__.py b/tests/torch/fx/__init__.py
index 138fbd9bf31..2e49d63977d 100644
--- a/tests/torch/fx/__init__.py
+++ b/tests/torch/fx/__init__.py
@@ -8,10 +8,3 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-import sys
-
-import pytest
-
-# Remove after fixing ticket: 156781
-if sys.platform.startswith("win"):
-    pytest.skip("capture_pre_autograd is not supported in windows", allow_module_level=True)
diff --git a/tests/torch/fx/helpers.py b/tests/torch/fx/helpers.py
index a54a83ec14f..a63f8f76076 100644
--- a/tests/torch/fx/helpers.py
+++ b/tests/torch/fx/helpers.py
@@ -19,7 +19,6 @@
 import torchvision.datasets as datasets
 import torchvision.transforms as transforms
 from fastdownload import FastDownload
-from torch._export import capture_pre_autograd_graph
 from torch.fx.passes.graph_drawer import FxGraphDrawer
 
 from nncf.experimental.torch.fx.transformations import apply_quantization_transformations
@@ -123,15 +122,37 @@ def visualize_fx_model(model: torch.fx.GraphModule, output_svg_path: str):
     g.get_dot_graph().write_svg(output_svg_path)
 
 
-def get_torch_fx_model(model: torch.nn.Module) -> torch.fx.GraphModule:
-    device = next(model.named_parameters())[1].device
-    input_shape = model.INPUT_SIZE
-    if input_shape is None:
-        input_shape = [1, 3, 32, 32]
-    ex_input = torch.ones(input_shape).to(device)
+def get_torch_fx_model(model: torch.nn.Module, ex_input: torch.Tensor) -> torch.fx.GraphModule:
+    """
+    Converts given module to GraphModule.
+
+    :param model: Given torch Module.
+    :return: Exported GraphModule.
+    """
+    try:
+        named_param = next(model.named_parameters())
+    except StopIteration:
+        named_param = None
+    if named_param is None:
+        device = torch.device("cpu")
+    else:
+        device = named_param[1].device
+
+    ex_input = ex_input.to(device)
     model.eval()
-    with disable_patching():
-        fx_model = capture_pre_autograd_graph(model, args=(ex_input,))
+    with torch.no_grad():
+        with disable_patching():
+            return torch.export.export_for_training(model, args=(ex_input,)).module()
+
+
+def get_torch_fx_model_q_transformed(model: torch.nn.Module, ex_input: torch.Tensor) -> torch.fx.GraphModule:
+    """
+    Converts given module to GraphModule and applies required quantization transformations to it.
+
+    :param model: Given torch Module.
+    :return: Exported GraphModule.
+    """
+    fx_model = get_torch_fx_model(model, ex_input)
     apply_quantization_transformations(fx_model)
     return fx_model
 
diff --git a/tests/torch/fx/test_bias_correction.py b/tests/torch/fx/test_bias_correction.py
index 94bb2066fd3..06db212fe79 100644
--- a/tests/torch/fx/test_bias_correction.py
+++ b/tests/torch/fx/test_bias_correction.py
@@ -28,7 +28,7 @@
 from tests.cross_fw.test_templates.helpers import SplittedModel
 from tests.cross_fw.test_templates.helpers import TransposeConvTestModel
 from tests.cross_fw.test_templates.test_bias_correction import TemplateTestBCAlgorithm
-from tests.torch.fx.helpers import get_torch_fx_model
+from tests.torch.fx.helpers import get_torch_fx_model_q_transformed
 
 
 class TestFXBCAlgorithm(TemplateTestBCAlgorithm):
@@ -43,7 +43,7 @@ def get_backend() -> FXBiasCorrectionAlgoBackend:
 
     @staticmethod
     def backend_specific_model(model: torch.nn.Module, tmp_dir: str) -> torch.fx.GraphModule:
-        return get_torch_fx_model(model)
+        return get_torch_fx_model_q_transformed(model, torch.ones(model.INPUT_SIZE))
 
     @staticmethod
     def fn_to_type(tensor) -> np.ndarray:
@@ -201,19 +201,19 @@ def test__get_subgraph_data_for_node(self, quantized_test_model, layer_name, ref
                 SplittedModel,
                 {
                     ("conv2d", 0): ("concat", 0),
-                    ("concat", 1): ("arg0_1", 0),
+                    ("concat", 1): ("x", 0),
                 },
             ),
             (
                 MultipleConvTestModel,
                 {
-                    ("conv2d", 0): ("arg0_1", 0),
-                    ("conv2d_2", 0): ("arg0_1", 0),
+                    ("conv2d", 0): ("x", 0),
+                    ("conv2d_2", 0): ("x", 0),
                 },
             ),
-            (ConvTestModel, {("conv2d", 0): ("arg0_1", 0)}),
-            (DepthwiseConvTestModel, {("conv2d", 0): ("arg0_1", 0)}),
-            (TransposeConvTestModel, {("conv_transpose2d", 0): ("arg0_1", 0)}),
+            (ConvTestModel, {("conv2d", 0): ("x", 0)}),
+            (DepthwiseConvTestModel, {("conv2d", 0): ("x", 0)}),
+            (TransposeConvTestModel, {("conv_transpose2d", 0): ("x", 0)}),
         ),
     )
     def test_verify_collected_stat_inputs_map(self, model_cls, ref_stat_inputs_map, tmpdir):
diff --git a/tests/torch/fx/test_compress_weights.py b/tests/torch/fx/test_compress_weights.py
index 2caffb35edf..874a6c64d16 100644
--- a/tests/torch/fx/test_compress_weights.py
+++ b/tests/torch/fx/test_compress_weights.py
@@ -13,7 +13,6 @@
 
 import pytest
 import torch
-from torch._export import capture_pre_autograd_graph
 
 import nncf
 from nncf import BackupMode
@@ -24,6 +23,7 @@
 from nncf.quantization import compress_weights
 from nncf.quantization.advanced_parameters import AdvancedCompressionParameters
 from nncf.torch.dynamic_graph.patch_pytorch import disable_patching
+from tests.torch.fx.helpers import get_torch_fx_model
 from tests.torch.ptq.test_weights_compression import ALL_SENSITIVITY_METRICS
 from tests.torch.ptq.test_weights_compression import DATA_BASED_SENSITIVITY_METRICS
 from tests.torch.ptq.test_weights_compression import INT4_MODES
@@ -35,7 +35,7 @@
 from tests.torch.ptq.test_weights_compression import EmptyModel
 from tests.torch.ptq.test_weights_compression import FunctionalModel
 from tests.torch.ptq.test_weights_compression import MatMulModel
-from tests.torch.ptq.test_weights_compression import ShortTransformer
+from tests.torch.test_models.synthetic import ShortTransformer
 
 
 def get_model_size(model):
@@ -73,17 +73,11 @@ def get_compressed_modules_weights(
     return n_target_modules, n_compressed_weights
 
 
-def _capture_model(model, inputs):
-    with torch.no_grad():
-        with disable_patching():
-            return capture_pre_autograd_graph(model, (inputs,))
-
-
 @pytest.mark.parametrize("mode", SUPPORTED_MODES)
 def test_compress_weights(mode):
     model = ShortTransformer(8, 16)
     input_ids = torch.randint(0, 10, (8,))
-    exported_model = _capture_model(model, input_ids)
+    exported_model = get_torch_fx_model(model, input_ids)
     kwargs = {}
     if mode in [CompressWeightsMode.INT4_SYM, CompressWeightsMode.INT4_ASYM]:
         kwargs["group_size"] = 4
@@ -103,7 +97,7 @@ def test_compress_weights(mode):
 def test_compress_weights_graph_edge(mode):
     model = ShortTransformer(5, 10)
     input_ids = torch.randint(0, 10, (5,))
-    exported_model = _capture_model(model, input_ids)
+    exported_model = get_torch_fx_model(model, input_ids)
     compressed_model = compress_weights(exported_model, mode=mode)
     nncf_graph = NNCFGraphFactory.create(compressed_model)
     for node in nncf_graph.get_all_nodes():
@@ -118,7 +112,7 @@ def test_compress_weights_shared_weights(mocker, mode):
     with disable_patching():
         model = ShortTransformer(8, 16, share_weights=True)
         input_ids = torch.randint(0, 10, (8,))
-        exported_model = _capture_model(model, input_ids)
+        exported_model = get_torch_fx_model(model, input_ids)
         kwargs = {}
         if mode in [CompressWeightsMode.INT4_SYM, CompressWeightsMode.INT4_ASYM]:
             kwargs["group_size"] = 4
@@ -154,7 +148,7 @@ def test_compressed_model_inference(mode):
     torch.manual_seed(42)
     model = ShortTransformer(8, 16, share_weights=True)
     input_ids = torch.randint(0, 10, (8,))
-    exported_model = _capture_model(model, input_ids)
+    exported_model = get_torch_fx_model(model, input_ids)
     exported_model_output = exported_model(input_ids)
     kwargs = {}
     if mode in [CompressWeightsMode.INT4_SYM, CompressWeightsMode.INT4_ASYM]:
@@ -174,7 +168,7 @@ def test_compress_weights_model_size_conv(mode):
     model = ConvolutionModel()
 
     input_ids = torch.randint(0, 10, [1, 3, 256, 256])
-    exported_model = _capture_model(model, input_ids)
+    exported_model = get_torch_fx_model(model, input_ids)
     model_size = get_model_size(exported_model)
     compressed_model = compress_weights(exported_model, mode=mode)
     compressed_model_size = get_model_size(compressed_model)
@@ -199,7 +193,7 @@ def test_compress_weights_functional_model(mode):
     )
 
     input_ids = torch.randint(0, 10, [1, 3, 256, 256])
-    exported_model = _capture_model(model, input_ids)
+    exported_model = get_torch_fx_model(model, input_ids)
     compressed_model = compress_weights(exported_model, mode=mode)
 
     n_compressed_weights = 0
@@ -233,7 +227,7 @@ def test_compress_weights_functional_model(mode):
 def test_raise_error_with_unsupported_params_for_int8(mode, params):
     dummy_torch_model = EmptyModel()
     dummy_input = torch.Tensor()
-    exported_model = _capture_model(dummy_torch_model, dummy_input)
+    exported_model = get_torch_fx_model(dummy_torch_model, dummy_input)
     with pytest.raises(nncf.ParameterNotSupportedError):
         compress_weights(exported_model, mode=mode, **params)
 
@@ -253,7 +247,7 @@ def test_raise_error_with_unsupported_params_for_int8(mode, params):
 def test_raise_error_with_unsupported_params_for_int4(mode, params):
     dummy_torch_model = EmptyModel()
     dummy_input = torch.Tensor()
-    exported_model = _capture_model(dummy_torch_model, dummy_input)
+    exported_model = get_torch_fx_model(dummy_torch_model, dummy_input)
     with pytest.raises(nncf.ParameterNotSupportedError):
         compress_weights(exported_model, mode=mode, **params)
 
@@ -262,7 +256,7 @@ def test_raise_error_with_unsupported_params_for_int4(mode, params):
 def test_raise_error_with_not_int8(mode):
     dummy_torch_model = EmptyModel()
     dummy_input = torch.Tensor()
-    exported_model = _capture_model(dummy_torch_model, dummy_input)
+    exported_model = get_torch_fx_model(dummy_torch_model, dummy_input)
     with pytest.raises(nncf.ParameterNotSupportedError):
         compress_weights(exported_model, mode=mode)
 
@@ -270,7 +264,7 @@ def test_raise_error_with_not_int8(mode):
 def test_raise_error_for_statistics_caching():
     dummy_torch_model = EmptyModel()
     dummy_input = torch.Tensor()
-    exported_model = _capture_model(dummy_torch_model, dummy_input)
+    exported_model = get_torch_fx_model(dummy_torch_model, dummy_input)
     with pytest.raises(nncf.ParameterNotSupportedError):
         compress_weights(exported_model, advanced_parameters=AdvancedCompressionParameters(statistics_path="anything"))
 
@@ -278,7 +272,7 @@ def test_raise_error_for_statistics_caching():
 def test_get_dtype_attribute_of_parameter():
     model = DTypeModel()
     dummy_input = torch.randint(0, 10, [3, 3])
-    exported_model = _capture_model(model, dummy_input)
+    exported_model = get_torch_fx_model(model, dummy_input)
     compressed_model = compress_weights(exported_model)
     assert compressed_model.matmul_updated_constant0.dtype == torch.uint8
     compressed_model(dummy_input)
@@ -296,7 +290,7 @@ def test_model_devices_and_precisions(use_cuda, dtype):
     if dtype == torch.float16:
         model.half()
     dummy_input = torch.rand((1, 256), dtype=dtype, device=device)
-    exported_model = _capture_model(model, dummy_input)
+    exported_model = get_torch_fx_model(model, dummy_input)
     compressed_model = compress_weights(exported_model)
     result = compressed_model(dummy_input)
 
diff --git a/tests/torch/fx/test_fast_bias_correction.py b/tests/torch/fx/test_fast_bias_correction.py
index 4fa7d0b64fc..8c94a32cafd 100644
--- a/tests/torch/fx/test_fast_bias_correction.py
+++ b/tests/torch/fx/test_fast_bias_correction.py
@@ -19,7 +19,7 @@
 from nncf.quantization.algorithms.fast_bias_correction.torch_fx_backend import FXFastBiasCorrectionAlgoBackend
 from nncf.torch.model_graph_manager import OPERATORS_WITH_BIAS_METATYPES
 from tests.cross_fw.test_templates.test_fast_bias_correction import TemplateTestFBCAlgorithm
-from tests.torch.fx.helpers import get_torch_fx_model
+from tests.torch.fx.helpers import get_torch_fx_model_q_transformed
 
 
 class TestTorchFXFBCAlgorithm(TemplateTestFBCAlgorithm):
@@ -33,7 +33,7 @@ def get_backend() -> FXFastBiasCorrectionAlgoBackend:
 
     @staticmethod
     def backend_specific_model(model: torch.nn.Module, tmp_dir: str) -> torch.fx.GraphModule:
-        fx_model = get_torch_fx_model(model)
+        fx_model = get_torch_fx_model_q_transformed(model, torch.ones(model.INPUT_SIZE))
         return fx_model
 
     @staticmethod
@@ -72,8 +72,7 @@ def list_to_backend_type(data: List) -> torch.Tensor:
 
     @staticmethod
     def backend_specific_model(model: bool, tmp_dir: str):
-        fx_cuda_model = get_torch_fx_model(model.cuda())
-        return fx_cuda_model
+        return get_torch_fx_model_q_transformed(model.cuda(), torch.ones(model.INPUT_SIZE))
 
     @staticmethod
     def fn_to_type(tensor):
diff --git a/tests/torch/fx/test_model_transformer.py b/tests/torch/fx/test_model_transformer.py
index 2c1926b0d39..d42d52f4d8d 100644
--- a/tests/torch/fx/test_model_transformer.py
+++ b/tests/torch/fx/test_model_transformer.py
@@ -18,7 +18,6 @@
 import torch
 import torch.ao.quantization
 import torch.fx
-from torch._export import capture_pre_autograd_graph
 from torch.ao.quantization.fx.utils import create_getattr_from_value
 from torch.ao.quantization.observer import MinMaxObserver
 from torch.ao.quantization.observer import PerChannelMinMaxObserver
@@ -27,7 +26,6 @@
 import nncf
 import nncf.common
 import nncf.common.factory
-from nncf.common.factory import NNCFGraph
 from nncf.common.factory import NNCFGraphFactory
 from nncf.common.graph.transformations.commands import TargetType
 from nncf.common.graph.transformations.layout import TransformationLayout
@@ -48,11 +46,9 @@
 from nncf.experimental.torch.fx.transformations import node_removal_transformation_builder
 from nncf.experimental.torch.fx.transformations import output_insertion_transformation_builder
 from nncf.experimental.torch.fx.transformations import qdq_insertion_transformation_builder
-from nncf.experimental.torch.fx.transformations import shared_constants_unification_transformation
-from nncf.torch import disable_patching
-from nncf.torch.graph.operator_metatypes import CONST_NOOP_METATYPES
 from nncf.torch.graph.transformations.commands import PTModelExtractionCommand
 from nncf.torch.graph.transformations.commands import PTTargetPoint
+from tests.torch.fx.helpers import get_torch_fx_model
 from tests.torch.test_compressed_graph import check_graph
 from tests.torch.test_models.synthetic import ConstantFoldingTestModel
 from tests.torch.test_models.synthetic import ConvolutionWithAllConstantInputsModel
@@ -111,15 +107,9 @@ def _target_point_to_str(target_point: PTTargetPoint) -> str:
     )
 
 
-def _capture_model(model: torch.nn.Module, inputs: torch.Tensor) -> torch.fx.GraphModule:
-    with torch.no_grad():
-        with disable_patching():
-            return capture_pre_autograd_graph(model, (inputs,))
-
-
 @pytest.mark.parametrize("test_case", MODEL_EXTRACTION_CASES, ids=idfn)
 def test_model_extraction(test_case: ModelExtractionTestCase):
-    captured_model = _capture_model(test_case.model(), torch.ones(test_case.input_shape))
+    captured_model = get_torch_fx_model(test_case.model(), torch.ones(test_case.input_shape))
 
     layout = TransformationLayout()
     layout.register(test_case.command)
@@ -149,7 +139,7 @@ def forward(self, x):
     transformation = builder(test_module_instance, target_points, target_node_name)
 
     model = MultiBranchesConnectedModel()
-    captured_model = _capture_model(model, torch.ones((1, 3, 3, 3)))
+    captured_model = get_torch_fx_model(model, torch.ones((1, 3, 3, 3)))
     transformation(captured_model)
 
     nncf_graph = GraphConverter.create_nncf_graph(captured_model)
@@ -160,7 +150,7 @@ def forward(self, x):
 @pytest.mark.parametrize("bias", [True, False], ids=["bias", "constant"])
 def test_constant_update_transformation(bias: bool):
     model = MultiBranchesConnectedModel()
-    captured_model = _capture_model(model, torch.ones((1, 3, 3, 3)))
+    captured_model = get_torch_fx_model(model, torch.ones((1, 3, 3, 3)))
     nncf_graph = GraphConverter.create_nncf_graph(captured_model)
     target_node = nncf_graph.get_node_by_name("conv2d" if bias else "add_")
 
@@ -179,7 +169,7 @@ def test_constant_update_transformation(bias: bool):
 @pytest.mark.parametrize("bias", [True, False], ids=["bias", "constant"])
 def test_constant_update_transformation_no_constant(bias: bool):
     model = MultiBranchesConnectedModel()
-    captured_model = _capture_model(model, torch.ones((1, 3, 3, 3)))
+    captured_model = get_torch_fx_model(model, torch.ones((1, 3, 3, 3)))
     nncf_graph = GraphConverter.create_nncf_graph(captured_model)
     target_node = nncf_graph.get_node_by_name("add")
 
@@ -265,7 +255,7 @@ def test_one_target_point(
         transformation = qdq_insertion_transformation_builder(quantizer, [target_point])
 
         model = MultiBranchesConnectedModel()
-        captured_model = _capture_model(model, torch.ones((1, 3, 3, 3)))
+        captured_model = get_torch_fx_model(model, torch.ones((1, 3, 3, 3)))
         transformation(captured_model)
 
         self._check_qdq_params(captured_model, target_point, dtype, is_per_channel)
@@ -317,7 +307,7 @@ def test_shared_target_point(
         transformation = qdq_insertion_transformation_builder(quantizer, target_points)
 
         model = MultiBranchesConnectedModel()
-        captured_model = _capture_model(model, torch.ones((1, 3, 3, 3)))
+        captured_model = get_torch_fx_model(model, torch.ones((1, 3, 3, 3)))
         if not weights:
             with pytest.raises(nncf.InternalError):
                 transformation(captured_model)
@@ -343,7 +333,7 @@ def test_shared_target_point(
 
 def test_node_removal_transformation():
     model = MultiBranchesConnectedModel()
-    captured_model = _capture_model(model, torch.ones((1, 3, 3, 3)))
+    captured_model = get_torch_fx_model(model, torch.ones((1, 3, 3, 3)))
     nncf_graph = GraphConverter.create_nncf_graph(captured_model)
     node = nncf_graph.get_node_by_name("conv2d")
     transformation = node_removal_transformation_builder(node, input_port_id=0)
@@ -356,7 +346,7 @@ def test_node_removal_transformation():
 @pytest.mark.parametrize("target_point", MultiBranchesConnectedModel_TARGET_POINTS)
 def test_output_insertion_transformation(tuple_output: bool, target_point: PTTargetPoint):
     model = MultiBranchesConnectedModel()
-    captured_model = _capture_model(model, torch.ones((1, 3, 3, 3)))
+    captured_model = get_torch_fx_model(model, torch.ones((1, 3, 3, 3)))
 
     if not tuple_output:
         output_node = [node for node in captured_model.graph.nodes if node.op == "output"][0]
@@ -384,41 +374,12 @@ def count_constants(model: torch.fx.GraphModule) -> int:
     return num_constant_nodes
 
 
-def test_create_shared_constant_transformation():
-    model = MultiBranchesConnectedModel()
-    ex_inputs = torch.ones((1, 3, 3, 3))
-    captured_model = _capture_model(model, ex_inputs)
-    shared_constants_unification_transformation(captured_model)
-    nncf_graph = GraphConverter.create_nncf_graph(captured_model)
-    check_graph(
-        nncf_graph, "shared_constants_unification_transformation_test.dot", TRANSFORMED_GRAPH_DIR_NAME, extended=True
-    )
-
-
-def get_shared_constant_nodes(nncf_graph: NNCFGraph):
-    """
-    Gets a dict of constant nodes as key and consumer nodes as values which are shared in the model.
-    eg:
-          const
-          /   \
-    node1     node2
-
-    returns ({const:[node1, node2]})
-    """
-    shared_const_node_consumer_node = {}
-    for node in nncf_graph.get_all_nodes():
-        consumer_nodes = nncf_graph.get_next_nodes(node)
-        if node.metatype in CONST_NOOP_METATYPES and len(consumer_nodes) > 1:
-            shared_const_node_consumer_node[node] = consumer_nodes
-    return shared_const_node_consumer_node
-
-
 def insert_qdq_nodes(
     model: torch.fx.GraphModule,
     correct_pattern: bool,
     per_channel: bool,
     node_name: str = "conv2d",
-    w_const_node_name: str = "_param_constant0",
+    w_const_node_name: str = "conv_a_weight",
 ):
     const_node = get_graph_node_by_name(model.graph, w_const_node_name)
     if per_channel:
@@ -468,7 +429,7 @@ def test_compress_post_quantize_transformation(is_per_channel: bool):
     model = MultiBranchesConnectedModel()
     ex_input = torch.ones(1, 3, 224, 224)
 
-    model_with_correct_pattern = _capture_model(model, ex_input)
+    model_with_correct_pattern = get_torch_fx_model(model, ex_input)
     insert_qdq_nodes(model_with_correct_pattern, correct_pattern=True, per_channel=is_per_channel)
     compress_post_quantize_transformation(model_with_correct_pattern)
     graph_name = f"compress_post_quantize_{'per_channel' if is_per_channel else 'per_tensor'}_valid.dot"
@@ -479,7 +440,7 @@ def test_compress_post_quantize_transformation(is_per_channel: bool):
         extended=True,
     )
 
-    model_with_incorrect_pattern = _capture_model(model, ex_input)
+    model_with_incorrect_pattern = get_torch_fx_model(model, ex_input)
     insert_qdq_nodes(model_with_incorrect_pattern, correct_pattern=False, per_channel=is_per_channel)
     compress_post_quantize_transformation(model_with_incorrect_pattern)
     graph_name = f"compress_post_quantize_{'per_channel' if is_per_channel else 'per_tensor'}_invalid.dot"
@@ -491,46 +452,24 @@ def test_compress_post_quantize_transformation(is_per_channel: bool):
     )
 
 
-def test_update_shared_constant():
-    model = MultiBranchesConnectedModel()
-    ex_inputs = torch.ones((1, 3, 3, 3))
-    captured_model = _capture_model(model, ex_inputs)
-
-    shared_constants_unification_transformation(captured_model)
-    nncf_graph = NNCFGraphFactory.create(captured_model)
-    shared_constants_consumers_dict = get_shared_constant_nodes(nncf_graph)
-
-    # This returns all the constant nodes as keys and list of consumer as values
-    consumer_nodes = list(shared_constants_consumers_dict.values())[0]
-
-    constant_update_transformation_builder(consumer_nodes[0], torch.tensor([100]))(captured_model)
-
-    nncf_graph_updated_constant = NNCFGraphFactory.create(captured_model)
-    updated_const_node = nncf_graph_updated_constant.get_previous_nodes(consumer_nodes[0])[1]
-    fx_node_to_check_const = get_graph_node_by_name(captured_model.graph, updated_const_node.node_name)
-    fx_node_to_check_const_value = get_tensor_constant_from_node(fx_node_to_check_const, captured_model)
-
-    assert fx_node_to_check_const_value == torch.tensor([100])
-
-
 def test_get_connected_nodes():
     model = MultiBranchesConnectedModel()
     ex_inputs = torch.ones((1, 3, 3, 3))
-    captured_model = _capture_model(model, ex_inputs)
+    captured_model = get_torch_fx_model(model, ex_inputs)
     connected_nodes_list = _get_connected_nodes(captured_model.graph)
-    assert len(connected_nodes_list) == 18
+    assert len(connected_nodes_list) == 16
 
     add_node = get_graph_node_by_name(captured_model.graph, "add__1")
     conv_1_node = get_graph_node_by_name(captured_model.graph, "conv2d")
     conv_2_node = get_graph_node_by_name(captured_model.graph, "conv2d_1")
     add_node.replace_input_with(conv_2_node, conv_1_node)
     connected_nodes_list = _get_connected_nodes(captured_model.graph)
-    assert len(connected_nodes_list) == 15
+    assert len(connected_nodes_list) == 13
 
 
 def test_constant_folding():
     model = ConstantFoldingTestModel()
-    captured_model = _capture_model(model, torch.ones(model.INPUT_SIZE))
+    captured_model = get_torch_fx_model(model, torch.ones(model.INPUT_SIZE))
     folded_model = deepcopy(captured_model)
     constant_fold(folded_model)
     ex_input = torch.ones(model.INPUT_SIZE)
@@ -542,14 +481,14 @@ def test_constant_folding():
 
 def test_constant_folding_with_constraints(is_per_channel):
     model = ConstantFoldingTestModel()
-    model_with_correct_pattern = _capture_model(model, torch.ones(model.INPUT_SIZE))
+    model_with_correct_pattern = get_torch_fx_model(model, torch.ones(model.INPUT_SIZE))
 
     insert_qdq_nodes(
         model_with_correct_pattern,
         correct_pattern=True,
         per_channel=is_per_channel,
         node_name="linear_1",
-        w_const_node_name="_param_constant3",
+        w_const_node_name="linear_act_weight",
     )
     fold_constant_except_qdq(model_with_correct_pattern)
 
diff --git a/tests/torch/fx/test_models.py b/tests/torch/fx/test_models.py
index e0554af6719..a43c4fafb56 100644
--- a/tests/torch/fx/test_models.py
+++ b/tests/torch/fx/test_models.py
@@ -25,7 +25,6 @@
 import torch.utils.data
 import torch.utils.data.distributed
 import torchvision.models as models
-from torch._export import capture_pre_autograd_graph
 
 import nncf
 from nncf.common.graph.graph import NNCFNodeName
@@ -36,15 +35,14 @@
 from nncf.experimental.torch.fx.quantization.backend_parameters import FXBackendParameters
 from nncf.experimental.torch.fx.transformations import DEQUANTIZE_NODE_TARGETS
 from nncf.experimental.torch.fx.transformations import _get_node_inputs
-from nncf.experimental.torch.fx.transformations import shared_constants_unification_transformation
 from nncf.quantization.advanced_parameters import AdvancedQuantizationParameters
-from nncf.torch.dynamic_graph.patch_pytorch import disable_patching
 from tests.cross_fw.shared.paths import TEST_ROOT
 from tests.torch import test_models
+from tests.torch.fx.helpers import get_torch_fx_model
 from tests.torch.fx.test_sanity import count_q_dq
-from tests.torch.ptq.test_weights_compression import ShortTransformer
 from tests.torch.test_compressed_graph import check_graph
 from tests.torch.test_models.synthetic import MultiBranchesConnectedModel
+from tests.torch.test_models.synthetic import ShortTransformer
 
 FX_DIR_NAME = Path("fx")
 FX_QUANTIZED_DIR_NAME = Path("fx") / "quantized"
@@ -88,12 +86,6 @@ def get_full_path_to_json(model_json_name: str, attributes: bool = False) -> str
     return path_to_json
 
 
-def _capture_model(model: torch.nn.Module, inputs: torch.Tensor) -> torch.fx.GraphModule:
-    with torch.no_grad():
-        with disable_patching():
-            return capture_pre_autograd_graph(model, (inputs,))
-
-
 def get_ref_from_json(
     model_name: str, model_metatypes: Dict[NNCFNodeName, Union[Type[OperatorMetatype], bool]], attributes=False
 ) -> Dict[NNCFNodeName, Union[Type[OperatorMetatype], bool]]:
@@ -120,11 +112,9 @@ def test_model(test_case: ModelCase):
     model = test_case.model_builder()
     model.to(device)
 
-    with torch.no_grad():
-        dtype = torch.int32 if test_case.model_id == "synthetic_transformer" else torch.float32
-        ex_input = torch.ones(test_case.input_shape, dtype=dtype)
-        model.eval()
-        exported_model = _capture_model(model, ex_input)
+    dtype = torch.int32 if test_case.model_id == "synthetic_transformer" else torch.float32
+    ex_input = torch.ones(test_case.input_shape, dtype=dtype)
+    exported_model = get_torch_fx_model(model, ex_input)
     nncf_graph = GraphConverter.create_nncf_graph(exported_model)
 
     # Check NNCFGrpah
@@ -185,9 +175,7 @@ def test_quantized_model(model_case: ModelCase, quantization_parameters, compres
     dtype = torch.int32 if model_case.model_id == "synthetic_transformer" else torch.float32
     example_input = torch.ones(model_case.input_shape, dtype=dtype)
 
-    with torch.no_grad():
-        model.eval()
-        fx_model = _capture_model(model, example_input)
+    fx_model = get_torch_fx_model(model, example_input)
 
     def transform_fn(data_item):
         return data_item.to("cpu")
@@ -240,16 +228,12 @@ def check_compressed_post_quantized(quantized_model):
         assert result.dtype == torch.float32
 
 
-@pytest.mark.parametrize("unification", [False, True])
-def test_is_shared_attribute(unification):
+def test_is_shared_attribute_default():
     model = MultiBranchesConnectedModel()
     ex_inputs = torch.ones((1, 3, 3, 3))
-    captured_model = _capture_model(model, ex_inputs)
-    file_prefix = "not_unified"
-    if unification:
-        file_prefix = "unified"
-        shared_constants_unification_transformation(captured_model)
-    nncf_graph = GraphConverter.create_nncf_graph(captured_model)
+    fx_model = get_torch_fx_model(model, ex_inputs)
+    nncf_graph = GraphConverter.create_nncf_graph(fx_model)
+
     shared_attributes = {n.node_name: n.is_shared() for n in nncf_graph.get_all_nodes()}
-    ref_attributes = get_ref_from_json(f"{file_prefix}_shared_attribute_test_model", shared_attributes, attributes=True)
+    ref_attributes = get_ref_from_json("default_shared_attribute_test_model", shared_attributes, attributes=True)
     assert shared_attributes == ref_attributes
diff --git a/tests/torch/fx/test_ptq_params.py b/tests/torch/fx/test_ptq_params.py
index e16fc6fdde4..04a852a8a71 100644
--- a/tests/torch/fx/test_ptq_params.py
+++ b/tests/torch/fx/test_ptq_params.py
@@ -35,7 +35,7 @@
 from tests.common.quantization.metatypes import SoftmaxTestMetatype
 from tests.cross_fw.test_templates.test_ptq_params import TemplateTestPTQParams
 from tests.torch.fx.helpers import get_single_conv_nncf_graph
-from tests.torch.fx.helpers import get_torch_fx_model
+from tests.torch.fx.helpers import get_torch_fx_model_q_transformed
 from tests.torch.ptq.helpers import get_single_no_weight_matmul_nncf_graph
 from tests.torch.test_models.synthetic import LinearPTQParamsTestModel
 
@@ -98,7 +98,8 @@ def nncf_graph_cls(self):
 
     @pytest.fixture(scope="session")
     def test_params(self):
-        linear_model = get_torch_fx_model(LinearPTQParamsTestModel())
+        linear_model = LinearPTQParamsTestModel()
+        linear_model = get_torch_fx_model_q_transformed(linear_model, torch.ones((1, 3, 32, 32)))
 
         return {
             "test_range_estimator_per_tensor": {
diff --git a/tests/torch/fx/test_sanity.py b/tests/torch/fx/test_sanity.py
index ef176692971..eaa5b6c0530 100644
--- a/tests/torch/fx/test_sanity.py
+++ b/tests/torch/fx/test_sanity.py
@@ -23,7 +23,6 @@
 import torch.utils.data
 import torch.utils.data.distributed
 import torchvision.models as models
-from torch._export import capture_pre_autograd_graph
 
 import nncf
 from nncf.common.logging.track_progress import track
@@ -133,7 +132,7 @@ def transform_fn(data_item):
         with torch.no_grad():
             ex_input = next(iter(calibration_dataset.get_inference_data()))
             model.eval()
-            exported_model = capture_pre_autograd_graph(model, args=(ex_input,))
+            exported_model = torch.export.export_for_training(model, args=(ex_input,)).module()
             quantized_model = nncf.quantize(exported_model, calibration_dataset)
             quantized_model = torch.compile(quantized_model, backend="openvino")
 
diff --git a/tests/torch/fx/test_smooth_quant.py b/tests/torch/fx/test_smooth_quant.py
index a6b468fc58e..9caf968d1fd 100644
--- a/tests/torch/fx/test_smooth_quant.py
+++ b/tests/torch/fx/test_smooth_quant.py
@@ -15,10 +15,8 @@
 import openvino.runtime as ov
 import pytest
 import torch
-from torch._export import capture_pre_autograd_graph
 
 from nncf import IgnoredScope
-from nncf.experimental.torch.fx.transformations import apply_quantization_transformations
 from nncf.parameters import ModelType
 from nncf.quantization.advanced_parameters import AdvancedQuantizationParameters
 from nncf.quantization.advanced_parameters import AdvancedSmoothQuantParameters
@@ -26,13 +24,13 @@
 from nncf.quantization.algorithms.post_training.algorithm import PostTrainingQuantization
 from nncf.quantization.algorithms.smooth_quant.torch_fx_backend import FXSmoothQuantAlgoBackend
 from nncf.quantization.algorithms.smooth_quant.torch_fx_backend import FXSQMultiply
-from nncf.torch import disable_patching
 from nncf.torch.graph.operator_metatypes import PTConv2dMetatype
 from nncf.torch.graph.operator_metatypes import PTLinearMetatype
 from tests.cross_fw.test_templates.helpers import ConvTestModel
 from tests.cross_fw.test_templates.helpers import LinearMultiShapeModel
 from tests.cross_fw.test_templates.helpers import ShareWeghtsConvAndShareLinearModel
 from tests.cross_fw.test_templates.test_smooth_quant import TemplateTestSQAlgorithm
+from tests.torch.fx.helpers import get_torch_fx_model_q_transformed
 
 PT_LINEAR_MODEL_MM_MAP = {"Linear1": "linear_3", "Linear2": "linear_2", "Linear3": "linear", "Linear4": "linear_1"}
 
@@ -97,10 +95,7 @@ def get_backend() -> Type[FXSmoothQuantAlgoBackend]:
 
     @staticmethod
     def backend_specific_model(model: torch.nn.Module, tmp_dir: str) -> ov.Model:
-        with disable_patching():
-            captured_model = capture_pre_autograd_graph(model.eval(), (torch.rand(model.INPUT_SIZE),))
-            apply_quantization_transformations(captured_model)
-            return captured_model
+        return get_torch_fx_model_q_transformed(model, torch.ones(model.INPUT_SIZE))
 
     @staticmethod
     def check_scales(model: torch.nn.Module, reference_values: Dict[str, np.ndarray], model_cls) -> None:
diff --git a/tests/torch/fx/test_statistics_aggregator.py b/tests/torch/fx/test_statistics_aggregator.py
index d9e16b509e3..26e19d05608 100644
--- a/tests/torch/fx/test_statistics_aggregator.py
+++ b/tests/torch/fx/test_statistics_aggregator.py
@@ -15,7 +15,6 @@
 import pytest
 import torch
 from torch import nn
-from torch._export import capture_pre_autograd_graph
 
 from nncf import Dataset
 from nncf.common.graph.transformations.commands import TargetType
@@ -23,9 +22,9 @@
 from nncf.experimental.torch.fx.statistics.aggregator import FXStatisticsAggregator
 from nncf.quantization.algorithms.fast_bias_correction.torch_fx_backend import FXFastBiasCorrectionAlgoBackend
 from nncf.quantization.algorithms.min_max.torch_fx_backend import FXMinMaxAlgoBackend
-from nncf.torch.dynamic_graph.patch_pytorch import disable_patching
 from nncf.torch.graph.graph import PTTargetPoint
 from tests.common.test_statistics_aggregator import TemplateTestStatisticsAggregator
+from tests.torch.fx.helpers import get_torch_fx_model
 
 IDENTITY_NODE_NAME = "add"
 CONV_NODE_NAME = "conv2d"
@@ -60,9 +59,7 @@ def get_fast_bias_correction_algo_backend_cls(self) -> Type[FXFastBiasCorrection
     def get_backend_model(self, dataset_samples):
         sample = dataset_samples[0].reshape(INPUT_SHAPE[1:])
         conv_w = self.dataset_samples_to_conv_w(np.array(sample))
-        with disable_patching():
-            model = capture_pre_autograd_graph(IdentityConv(conv_w), args=(torch.ones(INPUT_SHAPE),))
-            return model
+        return get_torch_fx_model(IdentityConv(conv_w), torch.ones(INPUT_SHAPE))
 
     def get_statistics_aggregator(self, dataset):
         return FXStatisticsAggregator(dataset)
diff --git a/tests/torch/ptq/test_weights_compression.py b/tests/torch/ptq/test_weights_compression.py
index 3c3a424e2af..95c03013333 100644
--- a/tests/torch/ptq/test_weights_compression.py
+++ b/tests/torch/ptq/test_weights_compression.py
@@ -28,6 +28,7 @@
 from nncf.torch.quantization.quantize_functions import pack_uint4
 from nncf.torch.quantization.quantize_functions import unpack_int4
 from nncf.torch.quantization.quantize_functions import unpack_uint4
+from tests.torch.test_models.synthetic import ShortTransformer
 
 DATA_BASED_SENSITIVITY_METRICS = (
     SensitivityMetric.HESSIAN_INPUT_ACTIVATION,
@@ -44,23 +45,6 @@
 UNSUPPORTED_MODES = (CompressWeightsMode.NF4, CompressWeightsMode.E2M1)
 
 
-class ShortTransformer(torch.nn.Module):
-    def __init__(self, in_features, num_embeddings, share_weights=False):
-        super().__init__()
-        self.wte = torch.nn.Embedding(num_embeddings, in_features)
-        self.linear = torch.nn.Linear(in_features, in_features)
-        self.lm_head = torch.nn.Linear(in_features, num_embeddings)
-
-        if share_weights:
-            self.lm_head.weight = self.wte.weight
-
-    def forward(self, input_ids):
-        x = self.wte(input_ids)
-        x = self.linear(x)
-        res = self.lm_head(x)
-        return res
-
-
 class MatMulModel(torch.nn.Module):
     def __init__(self):
         super().__init__()
diff --git a/tests/torch/test_models/synthetic.py b/tests/torch/test_models/synthetic.py
index 6f4578a7b13..2074f1a31f2 100644
--- a/tests/torch/test_models/synthetic.py
+++ b/tests/torch/test_models/synthetic.py
@@ -592,3 +592,20 @@ def forward(self, x):
         y += 10
         x = self.linear_act(x)
         return x + y
+
+
+class ShortTransformer(torch.nn.Module):
+    def __init__(self, in_features, num_embeddings, share_weights=False):
+        super().__init__()
+        self.wte = torch.nn.Embedding(num_embeddings, in_features)
+        self.linear = torch.nn.Linear(in_features, in_features)
+        self.lm_head = torch.nn.Linear(in_features, num_embeddings)
+
+        if share_weights:
+            self.lm_head.weight = self.wte.weight
+
+    def forward(self, input_ids):
+        x = self.wte(input_ids)
+        x = self.linear(x)
+        res = self.lm_head(x)
+        return res