Error while converting Faster R-CNN Inception V2 through model optimizer

[Saikat]

Trying to convert a Faster R-CNN with Inception v2, configured for Oxford-IIIT Pets Dataset is giving error.

Command:

python3 /opt/intel/openvino/deployment_tools/model_optimizer/mo_tf.py

--input_model /home/ubuntu/Model_and_files/output/frozen_inference_graph.pb 

--tensorflow_use_custom_operations_config /opt/intel/openvino/deployment_tools

/model_optimizer/extensions/front/tf/faster_rcnn_support_api_v1.14.json

--tensorflow_object_detection_api_pipeline_config=

/home/ubuntu/Model_and_files/output/pipeline.config 

--reverse_input_channels

Error:

[ ERROR ]  Failed to match nodes from custom replacement description with id 'ObjectDetectionAPIProposalReplacement': 

It means model and custom replacement description are incompatible. 

Try to correct custom replacement description according to documentation with respect to model node names 

[ ANALYSIS INFO ]  Your model looks like TensorFlow Object Detection API Model. 

Check if all parameters are specified: 

    --tensorflow_use_custom_operations_config 

    --tensorflow_object_detection_api_pipeline_config 

    --input_shape (optional) 

    --reverse_input_channels (if you convert a model to use with the Inference Engine sample applications) 

Detailed information about conversion of this model can be found at 

https://docs.openvinotoolkit.org/latest/_docs_MO_DG_prepare_model_convert_model_tf_specific_Convert_Object_Detection_API_Models.html 

[ ERROR ]  Exception occurred during running replacer "ObjectDetectionAPIDetectionOutputReplacement" (<class 'extensions.front.tf.ObjectDetectionAPI.ObjectDetectionAPIDetectionOutputReplacement'>): Found the following nodes '[]' with name 'crop_proposals' but there should be exactly 1. Looks like ObjectDetectionAPIProposalReplacement replacement didn't work. 

If there is any resolution to this please let me know.

[Munesh_Intel]

Hi Saikat,

Please share your environment details (versions of OpenVINO, OS, TensorFlow, Python, CMake, etc.).

If possible, please share the trained model files for us to reproduce your issue.

Regards,

Munesh

[Munesh_Intel]

Hi Saikat,

Thank you for your question. If you need any additional information from Intel, please submit a new question as this thread is no longer being monitored.

Regards,

Munesh