Unable to Convert the Custom TensorFlow* 2 Model to Intermediate Representation (IR) Using OpenVINO™ 2020.4.582

• Converted custom TensorFlow 2 model mask_rcnn_inception_resnet_v2_1024x1024_coco17 to IR using the Model Optimizer command:
   

  Note	The link above may not download the file correctly. It may be necessary for you to copy (right-click and copy link address) and paste the link directly into your browser in order to open the file.

  "Exception: Exception occurred during running replacer "REPLACEMENT_ID (<class'extensions.front.tf.tensorflow_custom_operations_config_update.TensorflowCustomOperationsConfigUpdate'>)": The function 'update_custom_layer_attributes' must be implemented in the sub-class."

  python mo_tf.py

  --saved_model_dir C:\DATASETS\mask50000\exports\saved_model

  --output_dir C:\DATASETS\mask50000

  --reverse_input_channels

  --tensorflow_custom_operations_config extensions\front\tf\mask_rcnn_support_api_v2.4.json

  --tensorflow_object_detection_api_pipeline_config C:\DATASETS\mask50000\exports\pipeline.config

  --log_level=DEBUG

• Received error:

  "Exception: Exception occurred during running replacer "REPLACEMENT_ID (<class'extensions.front.tf.tensorflow_custom_operations_config_update.TensorflowCustomOperationsConfigUpdate'>)": The function 'update_custom_layer_attributes' must be implemented in the sub-class."

Support for TensorFlow 2.0 Object Detection API models was fully enabled only starting from OpenVINO 2021.3.

Upgrade to OpenVINO 2021.3 or later versions.