Dynamic batch inferencing in Python using Openvino 2021.3

[nnain1]

I like to have dynamic batch input for my Tensorflow model.

When I convert  in model optimization for Tensorflow model, I have error for having 

--input_shape [-1,24,94,3] .

So input_shape is set to batch size 1.

Then inside the program, I tried to set dynamic batch as

rec_exec_net = rec_ie.load_network(rec_net, args.device, {"DYN_BATCH_ENABLED": "YES"})

But when I run program, I have error as

ValueError: could not broadcast input array from shape (2,3,24,94) into shape (1,3,24,94) 

 at line 

request_wrap.execute("async", {rec_input_blob: pimages})

How can I have dynamic batch using OpenVino in python?

[Vladimir_Dudnik]

please refer to Using Dynamic Batching article in OpenVINO online documentation

[nnain1]

Yes I know that one. That is C sample and I am looking for Python sample.

[IntelSupport]

Hi nnain1,

Thanks for reaching out. Unfortunately, we do not have anything yet for dynamic batching in Pyhton and only available for C configuration only at the moment.

Regards,

Aznie

[IntelSupport]

Hi nnain1,

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

Regards,

Aznie