[Bug]: Check 'm_element_type != element::undefined && m_element_type.is_static()' failed

[jikechao]

OpenVINO Version

2023.1.0-12185-9e6b00e51cd-releases/2023/1

Operating System

Ubuntu 18.04 (LTS)

Framework

Tensorflow/Keras

Model used

Given in the following script

Issue description

When the input layer`s dtype="uint16" or "uint32", the conversion process from the given Keras model to OpenVINO will crash unexpectedly.

As a comparison, other input types (e.g., int16, 'uint8') can converted very well.

Step-by-step reproduction

import tensorflow as tf
from tensorflow import keras as keras
from tensorflow.keras import layers, models
import numpy as np

import openvino as ov

layer = keras.layers.ReLU()
input_shape = [1, 2, 3, 4]
input_data = np.random.randint(2, size=input_shape)
# weights = layer.get_weights()
# layer.set_weights(weights)

x = layers.Input(shape=input_shape[1:], dtype="uint16")
y = layer(x)
model = models.Model(x, y)
# model.summary()
res_keras = model(input_data)

tf2_model_path = f"_temp_model"
tf.saved_model.save(model, tf2_model_path)
ov_model = ov.convert_model(tf2_model_path,  input=input_shape)

Relevant log output

Traceback (most recent call last):
  File "test.py", line 24, in <module>
    ov_model = ov.convert_model(tf2_model_path,  input=input_shape)
  File "C:\software\conda\envs\torch\lib\site-packages\openvino\tools\ovc\convert.py", line 101, in convert_model
    ov_model, _ = _convert(cli_parser, params, True)
  File "C:\software\conda\envs\torch\lib\site-packages\openvino\tools\ovc\convert_impl.py", line 524, in _convert
    raise e
  File "C:\software\conda\envs\torch\lib\site-packages\openvino\tools\ovc\convert_impl.py", line 476, in _convert
    ov_model = driver(argv, {"conversion_parameters": non_default_params})
  File "C:\software\conda\envs\torch\lib\site-packages\openvino\tools\ovc\convert_impl.py", line 226, in driver
    ov_model = moc_emit_ir(prepare_ir(argv), argv)
  File "C:\software\conda\envs\torch\lib\site-packages\openvino\tools\ovc\convert_impl.py", line 172, in prepare_ir
    ov_model = moc_pipeline(argv, moc_front_end)
  File "C:\software\conda\envs\torch\lib\site-packages\openvino\tools\ovc\moc_frontend\pipeline.py", line 247, in moc_pipeline
    ov_model = moc_front_end.convert(input_model)
  File "C:\software\conda\envs\torch\lib\site-packages\openvino\frontend\frontend.py", line 18, in convert
    converted_model = super().convert(model)
RuntimeError: Check 'm_element_type != element::undefined && m_element_type.is_static()' failed at src\inference\src\dev\make_tensor.cpp:34


Process finished with exit code 1

### Issue submission checklist

- [X] I'm reporting an issue. It's not a question.
- [X] I checked the problem with the documentation, FAQ, open issues, Stack Overflow, etc., and have not found a solution.
- [X] There is reproducer code and related data files such as images, videos, models, etc.

[rkazants]

Hi @jikechao, thanks for reporting this bug. Indeed, this bug can still exists according the current state of type mapping: https://github.com/openvinotoolkit/openvino/blob/master/src/frontends/tensorflow/src/tf_utils.cpp#L154

I will fix it.

Best regards,
Roman

[jikechao]

BTW, "uint16", "uint32", and "uint64" have the same crash message.

[rkazants]

BTW, "uint16", "uint32", and "uint64" have the same crash message.

Yes, these types are also absent in that type map.

Thanks,
Roman

[rkazants]

Hi @jikechao,

#21439 is solving the problem. Please check and close if no concern.

Best regards,
Roman

[jikechao]

@rkazants Thanks for your fixing.

[andrei-kochin]

Re-opening original case as fix caused regression and we have to temporary revert #21656 to further improve the fix.

[rkazants]

The problem was with incorrect unpacking of compressed fp16 constant and PRelu fusing transformation was not applied due to incorrect fp16 Constant. Now I recover this PR and apply fix for this.
PR: #21864

Best regards,
Roman