Issues converting a Pytorch model into openvino directly

[Aji]

Hello folks,

I have been tasked with converting one of our action recognition pytorch model we use inhouse into an openvino model. 

I modeled our pytorch model conversion code from the Slowfast fast pytorch model conversion example https://docs.openvino.ai/2023.3/notebooks/210-slowfast-video-recognition-with-output.html#convert-model-to-openvino-intermediate-representation I ran into some issues with "example_input" param of the openvino.convert_model() syntax with our model. 

Can you point me to someone internally with an expertise in the model conversion that could help with my issue or have them reach out to me on teams or email?

Thanks,

Aji 

[Wan_Intel]

Hi Aji,

Thanks for reaching out to us.

Could you please share the following information with us?

• Steps to converting PyTorch model into IR
• Conversion command (PyTorch to IR or PyTorch to Onnx to IR)
• Error messages that you encountered
• Input shape of the model
• Input names of the model
• Output names of the model

On the other hand, please share your model to waix.fook.wan@intel.com.

Regards,

Wan

[Aji]

Hi Wan,

I have sent you an email, please take a look. 

Thanks,

Aji 

[Wan_Intel]

Hi Aji,

Thanks for the information.

I've received your email and model file. For your information, I’ve tried to create a wrapper model class to initialize the “epoch-250.model” file, however, I encountered different errors when converting the model into Intermediate Representation.

Could you please share the command or script that you used to initialize the model with us to further investigate the issue?

Regards,

Wan

[Wan_Intel]

Hi Aji,

If you need additional information from Intel, please submit a new question as this thread will no longer be monitored.

Regards,

Wan