Convert torch vision maskrcnn to IR format

[Abdulsathar]

Hi,

Is it possible to convert a trained torch vision maskrcnn model to OpenVINO model. I am encountering different errors. if not, which maskrcnn is supported by OpenVINO currently?

[IntelSupport]

Hi Abdul Sathar,

Thanks for reaching out.

Pytorch framework is supported in OpenVINO by exporting the model into ONNX format. Refer Converting a PyTorch* Model for more details on the conversion.

Apart from that, you can find the list of Mask R-CNN models from OpenVINO™ Toolkit Public Pre-Trained Models.

Below is the example of converting Mask R-CNN Inception ResNet V2 1024x1024 from TensorFlow 2 Detection Model Zoo. Here is the command that I used:

mo --saved_model_dir <PATH_TO_SAVED_MODEL>\saved_model --transformations_config <INSTALL_DIR>\openvino\tools\mo\front\tf\mask_rcnn_support_api_v2.0.json --tensorflow_object_detection_api_pipeline_config <PATH_TO_PIPELINE_CONFIG>\pipeline.config --reverse_input_channels

Regards,

Aznie

[IntelSupport]

Hi Abdul Sathar,

This thread will no longer be monitored since we have provided the information. If you need any additional information from Intel, please submit a new question.

Regards,

Aznie