Gh actions codecov

.github/workflows/post_pr_merge.yml

@@ -24,6 +24,7 @@ jobs:
       merge_commit_sha: ${{ github.event.pull_request.merge_commit_sha }}
       last_sha_in_pr: ${{ github.event.pull_request.head.sha }}
       coverage_artifact_name_in_pr: coverage_common
+      coverage_flags: COMMON
     secrets:
       CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
   upload-coverage-onnx:
@@ -33,5 +34,6 @@ jobs:
       merge_commit_sha: ${{ github.event.pull_request.merge_commit_sha }}
       last_sha_in_pr: ${{ github.event.pull_request.head.sha }}
       coverage_artifact_name_in_pr: coverage_onnx
+      coverage_flags: ONNX
     secrets:
       CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}

.github/workflows/precommit.yml

@@ -33,6 +33,7 @@ jobs:
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
           name: coverage_common
+          flags: COMMON
   onnx:
     runs-on: ubuntu-20.04
     steps:
@@ -58,4 +59,5 @@ jobs:
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
           name: coverage_onnx
+          flags: ONNX
 

.github/workflows/upload_coverage_for_develop.yml

@@ -12,6 +12,9 @@ on:
       coverage_artifact_name_in_pr:
         required: true
         type: string
+      coverage_flags: 
+        required: true
+        type: string     
     secrets:
       CODECOV_TOKEN:
         required: true
@@ -37,4 +40,4 @@ jobs:
 
           # github.event.pull_request.merge_commit_sha is the fresh commit in the develop,
           # provided that github.event.pull_request.merged == true
-          ./codecov -f ./coverage.xml -t ${{ secrets.CODECOV_TOKEN }}  -C ${{ inputs.merge_commit_sha }} -B develop -n "${{ inputs.coverage_artifact_name_in_pr }}"
+          ./codecov -f ./coverage.xml -t ${{ secrets.CODECOV_TOKEN }} -F ${{ inputs.coverage_flags }} -C ${{ inputs.merge_commit_sha }} -B develop -n "${{ inputs.coverage_artifact_name_in_pr }}"

.gitignore

@@ -108,6 +108,9 @@ ENV/
 # snapshots
 *.tar
 
+# pylint config is left at dev's discretion, CI uses ruff/isort/black instead
+.pylintrc
+
 # VSCode
 .vscode/
 

CONTRIBUTING.md

@@ -34,9 +34,12 @@ If your testing code is more extensive than unit tests (in terms of test executi
 
 Changes to NNCF Python code should conform to [Python Style Guide](./docs/styleguide/PyGuide.md)
 
-Pylint is used throughout the project to ensure code cleanliness and quality.
-A Pylint run is also done as part of the pre-commit scope - the pre-commit `pytest` scope will not be run if your code fails the Pylint checks.
-The Pylint rules and exceptions for this repository are described in the standard [.pylintrc](./.pylintrc) format - make sure your local linter uses these.
+Basic code style and static checks are enforced using a `pre-commit` Github action.
+The exact checks that are run are described in the corresponding [config file](./.pre-commit-config.yaml).
+The checks can be run locally using `make pre-commit`.
+Most of these checks do in-place fixes at the same time they are run using `make pre-commit`.
+Static analysis is done via `ruff` which does not do in-place fixes by default - run `ruff --fix` locally to autofix errors for which it is possible to do so.
+Developers may use other tools locally (such as `pylint`) as long as no tool-specific control code is introduced and the pre-commit checks pass.
 
 ## Binary files
 

Makefile

@@ -17,9 +17,6 @@ endif
 install-pre-commit:
 	pip install pre-commit==3.2.2
 
-install-pylint:
-	pip install pylint==2.13.9
-	pip install pylintfileheader==0.3.2
 
 ###############################################################################
 # ONNX backend
@@ -31,15 +28,12 @@ install-onnx-test:
 	pip install -r tests/cross_fw/examples/requirements.txt
 	pip install -r tests/onnx/benchmarking/requirements.txt
 
-install-onnx-dev: install-onnx-test install-pre-commit install-pylint
+install-onnx-dev: install-onnx-test install-pre-commit
 	pip install -r examples/post_training_quantization/onnx/mobilenet_v2/requirements.txt
 
 test-onnx:
 	pytest ${COVERAGE_ARGS} tests/onnx $(DATA_ARG) --junitxml ${JUNITXML_PATH}
 
-pylint-onnx:
-	pylint --rcfile .pylintrc               \
-		$(shell python3 tools/collect_pylint_input_files_for_backend.py onnx)
 
 test-install-onnx:
 	pytest tests/cross_fw/install -s       \
@@ -60,7 +54,7 @@ install-openvino-test:
 	pip install -r tests/cross_fw/install/requirements.txt
 	pip install -r tests/cross_fw/examples/requirements.txt
 
-install-openvino-dev: install-openvino-test install-pre-commit install-pylint
+install-openvino-dev: install-openvino-test install-pre-commit
 	pip install -r examples/experimental/openvino/bert/requirements.txt
 	pip install -r examples/experimental/openvino/yolo_v5/requirements.txt
 	pip install -r examples/post_training_quantization/openvino/mobilenet_v2/requirements.txt
@@ -71,10 +65,6 @@ install-openvino-dev: install-openvino-test install-pre-commit install-pylint
 test-openvino:
 	pytest ${COVERAGE_ARGS} tests/openvino $(DATA_ARG) --junitxml ${JUNITXML_PATH}
 
-pylint-openvino:
-	pylint --rcfile .pylintrc               \
-		$(shell  python3 tools/collect_pylint_input_files_for_backend.py openvino)
-
 test-install-openvino:
 	pytest tests/cross_fw/install -s        \
 		--backend openvino                  \
@@ -95,18 +85,14 @@ install-tensorflow-test:
 	pip install -r tests/cross_fw/examples/requirements.txt
 	pip install -r examples/tensorflow/requirements.txt
 
-install-tensorflow-dev: install-tensorflow-test install-pre-commit install-pylint
+install-tensorflow-dev: install-tensorflow-test install-pre-commit
 	pip install -r examples/post_training_quantization/tensorflow/mobilenet_v2/requirements.txt
 
 test-tensorflow:
-	pytest ${COVERAGE_ARGS} tests/common tests/tensorflow    \
+	pytest ${COVERAGE_ARGS} tests/tensorflow    \
 		--junitxml ${JUNITXML_PATH}         \
 		$(DATA_ARG)
 
-pylint-tensorflow:
-	pylint --rcfile .pylintrc               \
-		$(shell python3 tools/collect_pylint_input_files_for_backend.py tensorflow)
-
 test-install-tensorflow:
 	pytest tests/cross_fw/install -s --backend tf --junitxml ${JUNITXML_PATH}
 
@@ -123,25 +109,19 @@ install-torch-test:
 	pip install -r tests/cross_fw/examples/requirements.txt
 	pip install -r examples/torch/requirements.txt --index-url https://download.pytorch.org/whl/cu118 --extra-index-url=https://pypi.org/simple
 
-install-torch-dev: install-torch-test install-pre-commit install-pylint
+install-torch-dev: install-torch-test install-pre-commit
 	pip install -r examples/post_training_quantization/torch/mobilenet_v2/requirements.txt
 	pip install -r examples/post_training_quantization/torch/ssd300_vgg16/requirements.txt
 
 test-torch:
-	pytest ${COVERAGE_ARGS} tests/common tests/torch -m "not weekly and not nightly" --junitxml ${JUNITXML_PATH} $(DATA_ARG)
+	pytest ${COVERAGE_ARGS} tests/torch -m "not weekly and not nightly" --junitxml ${JUNITXML_PATH} $(DATA_ARG)
 
 test-torch-nightly:
 	pytest ${COVERAGE_ARGS} tests/torch -m nightly --junitxml ${JUNITXML_PATH} $(DATA_ARG)
 
 test-torch-weekly:
 	pytest ${COVERAGE_ARGS} tests/torch -m weekly --junitxml ${JUNITXML_PATH} $(DATA_ARG) ${WEEKLY_MODELS_ARG}
 
-COMMON_PYFILES := $(shell python3 tools/collect_pylint_input_files_for_backend.py common)
-pylint-torch:
-	pylint --rcfile .pylintrc   \
-		$(COMMON_PYFILES)       \
-		$(shell python3 tools/collect_pylint_input_files_for_backend.py torch)
-
 test-install-torch-cpu:
 	pytest tests/cross_fw/install -s       \
 		--backend torch                     \
@@ -167,10 +147,6 @@ install-common-test:
 	pip install -r tests/cross_fw/install/requirements.txt
 	pip install -r tests/cross_fw/examples/requirements.txt
 
-pylint-common:
-	pylint --rcfile .pylintrc   \
-		$(COMMON_PYFILES)
-
 test-common:
 	pytest ${COVERAGE_ARGS} tests/common $(DATA_ARG) --junitxml ${JUNITXML_PATH}
 

README.md

@@ -284,6 +284,10 @@ A collection of ready-to-run Jupyter* notebooks are available to demonstrate how
 - [NNCF Post-Training Optimization of BLIP Model](https://github.com/openvinotoolkit/openvino_notebooks/tree/main/notebooks/233-blip-visual-language-processing)
 - [NNCF Post-Training Optimization of DeepFloyd IF Model](https://github.com/openvinotoolkit/openvino_notebooks/tree/main/notebooks/238-deepfloyd-if)
 - [NNCF Post-Training Optimization of Grammatical Error Correction Model](https://github.com/openvinotoolkit/openvino_notebooks/tree/main/notebooks/214-grammar-correction)
+- [NNCF Post-Training Optimization of Dolly 2.0 Model](https://github.com/openvinotoolkit/openvino_notebooks/tree/main/notebooks/240-dolly-2-instruction-following)
+- [NNCF Post-Training Optimization of Massively Multilingual Speech Model](https://github.com/openvinotoolkit/openvino_notebooks/tree/main/notebooks/255-mms-massively-multilingual-speech)
+- [NNCF Post-Training Optimization of OneFormer Model](https://github.com/openvinotoolkit/openvino_notebooks/tree/main/notebooks/249-oneformer-segmentation)
+- [NNCF Post-Training Optimization of InstructPix2Pix Model](https://github.com/openvinotoolkit/openvino_notebooks/tree/main/notebooks/231-instruct-pix2pix-image-editing)
 - [Quantize a Segmentation Model and Show Live Inference](https://github.com/openvinotoolkit/openvino_notebooks/blob/main/notebooks/110-ct-segmentation-quantize)
 - [Training to Deployment with TensorFlow and OpenVINO](https://github.com/openvinotoolkit/openvino_notebooks/blob/main/notebooks/301-tensorflow-training-openvino)
 - [Migrate quantization from POT API to NNCF API](https://github.com/openvinotoolkit/openvino_notebooks/blob/main/notebooks/111-yolov5-quantization-migration)
@@ -375,7 +379,7 @@ You may also use one of the Dockerfiles in the [docker](./docker) directory to b
   - ONNX\* ~=1.13.1
   - OpenVINO\* >=2022.3.0
 
-This repository is tested on Python* 3.8.10, PyTorch* 2.0.1 (NVidia CUDA\* Toolkit 11.8) and TensorFlow* 2.12.1 (NVidia CUDA\* Toolkit 11.8).
+This repository is tested on Python* 3.8.10, PyTorch* 2.1.0 (NVidia CUDA\* Toolkit 11.8) and TensorFlow* 2.12.1 (NVidia CUDA\* Toolkit 11.8).
 
 ## NNCF Compressed Model Zoo
 

docs/Installation.md

@@ -67,8 +67,9 @@ Use one of the Dockerfiles in the [docker](../docker) directory to build an imag
 The following table lists the recommended corresponding versions of backend packages
 as well as the supported versions of Python:
 
-| NNCF    | OpenVINO   | PyTorch  | ONNX     | TensorFlow | Python  |
-| ------- | ---------- | -------- | -------- | ---------- | ------- |
-| `2.6.0` | `2023.1.0` | `2.0.1`  | `1.13.1` | `2.12.0`   | `3.8`   |
-| `2.5.0` | `2023.0.0` | `1.13.1` | `1.13.1` | `2.11.1`   | `3.8`   |
-| `2.4.0` | `2022.1.0` | `1.12.1` | `1.12.0` | `2.8.2`    | `3.8`   |
+| NNCF      | OpenVINO   | PyTorch  | ONNX     | TensorFlow | Python |
+|-----------|------------|----------|----------|------------|--------|
+| `develop` | `2023.1.0` | `2.1`    | `1.13.1` | `2.12.0`   | `3.8`  |
+| `2.6.0`   | `2023.1.0` | `2.0.1`  | `1.13.1` | `2.12.0`   | `3.8`  |
+| `2.5.0`   | `2023.0.0` | `1.13.1` | `1.13.1` | `2.11.1`   | `3.8`  |
+| `2.4.0`   | `2022.1.0` | `1.12.1` | `1.12.0` | `2.8.2`    | `3.8`  |

docs/api/source/conf.py

@@ -56,7 +56,6 @@ def __init__(self):
         self.canonical_name_vs_fqn: Dict[str, str] = {}
 
 
-# pylint: disable=too-many-branches
 def collect_api_entities() -> APIInfo:
     """
     Collects the fully qualified names of symbols in NNCF package that contain a special attribute (set via
@@ -73,21 +72,21 @@ def collect_api_entities() -> APIInfo:
     for _, modname, _ in pkgutil.walk_packages(path=nncf.__path__, prefix=nncf.__name__ + ".", onerror=lambda x: None):
         try:
             modules[modname] = importlib.import_module(modname)
-        except Exception as e:  # pylint: disable=broad-except
+        except Exception as e:
             skipped_modules[modname] = str(e)
 
     from nncf.common.utils.api_marker import api
 
     canonical_imports_seen = set()
-    # pylint: disable=too-many-nested-blocks
+
     for modname, module in modules.items():
         print(f"{modname}")
         for obj_name, obj in inspect.getmembers(module):
             objects_module = getattr(obj, "__module__", None)
             if objects_module == modname:
                 if inspect.isclass(obj) or inspect.isfunction(obj):
                     if hasattr(obj, api.API_MARKER_ATTR):
-                        marked_object_name = obj._nncf_api_marker  # pylint: disable=protected-access
+                        marked_object_name = obj._nncf_api_marker
                         # Check the actual name of the originally marked object
                         # so that the classes derived from base API classes don't
                         # all automatically end up in API

docs/compression_algorithms/CompressWeights.md

@@ -6,17 +6,44 @@
 
 The Weights Compression algorithm is aimed at compressing the weights of the models and can be used to optimize the model footprint and performance of large models where the size of weights is relatively larger than the size of activations, for example, Large Language Models (LLM). The algorithm compresses weights only for Linear and Embedding layers.
 
+##### INT8 and NF4 modes
+
+By default, weights are compressed to 8-bit integer data type - "INT8" mode.
+OpenVINO backend has also an experimental support for "NF4" mode - compression to [nf4](https://arxiv.org/pdf/2305.14314v1.pdf) data type.
+It goes with a grouped quantization, when small group of weights (e.g. 128) in the channel dimension share quantization parameters (scale).
+First embedding and last linear layers are always compressed to 8-bit integer data type in the "NF4" mode.
+Percent of the rest layers compressed to NF4 can be configured by "ratio" parameter.
+E.g. ratio=0.9 means 90% of layers compressed to nf4 and the rest to 8-bit integer data type.
+
 #### User guide
 
-- Compress weights of linear layers and embeddings to int8
+- Compress weights to 8-bit integer data type.
 
 ```python
 from nncf import compress_weights
 compressed_model = compress_weights(model)
 ```
 
+- Compress weights to nf4 data type with group size = 128, except first embedding and last linear layers - they are compressed to 8-bit integer data type.
+
+```python
+from nncf import compress_weights
+from nncf import CompressWeightsMode
+compressed_model = compress_weights(model, mode=CompressWeightsMode.NF4)
+```
+
+- Compress weights of 90% of layers to nf4 with the group size 64, and the rest of layers to 8-bit integer data type.
+
+```python
+from nncf import compress_weights
+from nncf import CompressWeightsMode
+compressed_model = compress_weights(model, mode=CompressWeightsMode.NF4, group_size=64, ratio=0.9)
+```
+
 ##### Limitations
 
 - The algorithm is supported for OpenVINO and PyTorch models.
 - The compression applies in-place.
 - The compressed model is not trainable.
+- NF4 mode, grouped quantization and mixed nf4-int8 precision selection is available for OpenVINO backend only.
+- NF4 support is experimental - models quantized to nf4 should not be faster models quantized to 8-bit integer.