openvino: [Bug] Could not run model optimizer with ONNX model

System information (version)

• OpenVINO => 2022.1 (docker container dlstreamer/dlstreamer)
• Operating System / Platform => Ubuntu 22.04 LTS
• Problem classification: Model Conversion
• Framework: TensorFlow
• Model name: Custom RNN model

Detailed description

Hello. I’m running dlstreamer/dlstreamer on a fitlet2 device https://fit-iot.com/web/products/fitlet2/ running Ubuntu 22.04. My goal is to run my custom model exported from Tensorflow (keras backend) to ONNX format and integrate it into the dlstreamer framework in order to compile custom C++ code that users gstreamer plugins and run inference on this custom model. Model is exported directly from tensorflow 1.15.5 running on Nvidia GPU after training for about 62 epochs. There is a problem when parsing a certain convolution layer. It says that the dilation size is greater than the padding size. Could you provide some help? What do I have to do by now? Do I have to keep the model and change export settings or do I have to use another exporting technique such as frozen model? Thank you in advance.

mo -w engine_file.onnx --input_shape [1,3,48,96] --input image_input --output tf_op_layer_ArgMax
Model Optimizer arguments:
Common parameters:
	- Path to the Input Model: 	/home/engine_file.onnx
	- Path for generated IR: 	/home/.
	- IR output name: 	engine_file
	- Log level: 	ERROR
	- Batch: 	Not specified, inherited from the model
	- Input layers: 	image_input
	- Output layers: 	tf_op_layer_ArgMax
	- Input shapes: 	[1,3,48,96]
	- Source layout: 	Not specified
	- Target layout: 	Not specified
	- Layout: 	Not specified
	- Mean values: 	Not specified
	- Scale values: 	Not specified
	- Scale factor: 	Not specified
	- Precision of IR: 	FP32
	- Enable fusing: 	True
	- User transformations: 	Not specified
	- Reverse input channels: 	False
	- Enable IR generation for fixed input shape: 	False
	- Use the transformations config file: 	None
Advanced parameters:
	- Force the usage of legacy Frontend of Model Optimizer for model conversion into IR: 	False
	- Force the usage of new Frontend of Model Optimizer for model conversion into IR: 	False
OpenVINO runtime found in: 	/usr/local/lib/python3.8/dist-packages/openvino
OpenVINO runtime version: 	2022.1.0-7019-cdb9bec7210-releases/2022/1
Model Optimizer version: 	2022.1.0-7019-cdb9bec7210-releases/2022/1
[ ERROR ]  -------------------------------------------------
[ ERROR ]  ----------------- INTERNAL ERROR ----------------
[ ERROR ]  Unexpected exception happened.
[ ERROR ]  Please contact Model Optimizer developers and forward the following information:
[ ERROR ]  While validating ONNX node '<Node(Conv): res3a_branch2a>':
Check 'window_dilated_dim <= data_padded_dilated_dim' failed at core/shape_inference/include/convolution_shape_inference.hpp:209:
While validating node 'v1::Convolution Convolution_460 (re_lu_4/Relu:0[0]:f32{1,64,1,1}, res3a_branch2a_W_new[0]:f32{128,64,3,3}) -> (dynamic...)' with friendly_name 'Convolution_460':
Window after dilation has dimension (dim: 3) larger than the data shape after padding (dim: 2) at axis 0.

[ ERROR ]  Traceback (most recent call last):
  File "/usr/local/lib/python3.8/dist-packages/openvino/tools/mo/main.py", line 533, in main
    ret_code = driver(argv)
  File "/usr/local/lib/python3.8/dist-packages/openvino/tools/mo/main.py", line 489, in driver
    graph, ngraph_function = prepare_ir(argv)
  File "/usr/local/lib/python3.8/dist-packages/openvino/tools/mo/main.py", line 394, in prepare_ir
    ngraph_function = moc_pipeline(argv, moc_front_end)
  File "/usr/local/lib/python3.8/dist-packages/openvino/tools/mo/moc_frontend/pipeline.py", line 147, in moc_pipeline
    ngraph_function = moc_front_end.convert(input_model)
RuntimeError: While validating ONNX node '<Node(Conv): res3a_branch2a>':
Check 'window_dilated_dim <= data_padded_dilated_dim' failed at core/shape_inference/include/convolution_shape_inference.hpp:209:
While validating node 'v1::Convolution Convolution_460 (re_lu_4/Relu:0[0]:f32{1,64,1,1}, res3a_branch2a_W_new[0]:f32{128,64,3,3}) -> (dynamic...)' with friendly_name 'Convolution_460':
Window after dilation has dimension (dim: 3) larger than the data shape after padding (dim: 2) at axis 0.


[ ERROR ]  ---------------- END OF BUG REPORT --------------
[ ERROR ]  -------------------------------------------------

Steps to reproduce

Here is the link to the model: Model

Issue submission checklist

• I report the issue, it’s not a question
• I checked the problem with documentation, FAQ, open issues, Stack Overflow, etc and have not found solution
• There is reproducer code and related data files: images, videos, models, etc.