ERROR: can't pickle module objects about openvino2tensorflow

[PINTO0309]

Please read the log carefully to the end. It is coded to automatically recover from errors. The saved_model has been generated successfully.

Switch to the output of an optimized protocol buffer file (.pb).
　：
.pb output complete! - saved_model_320x320/model_float32.pb
　：
Optimized graph converted to SavedModel! - saved_model_320x320

[NaeemKhan333]

@PINTO0309 How I can check whether model is converted rightly or not?Does any script are provided by you?

[PINTO0309]

https://github.com/PINTO0309/PINTO_model_zoo/tree/main/061_U-2-Net/01_float32

[NaeemKhan333]

@PINTO0309 is it for testing Tensorflow Js format Converted Model?

[PINTO0309]

No. This is a good place to start.

https://github.com/w-okada/image-analyze-workers.git

[NaeemKhan333]

@PINTO0309 I have run this repository , It is running on the u2net_portrait model given by the owner.But when I tried to use my own converted u2net model.It is not giving any output.Model is converted successfully using your blog but I do not why it is not working in the web demo code.

[PINTO0309]

The output of u2net_portrait and u2net are different. Please read and analyze the code on your own. The logic should not be difficult. The results tested with tflite and JS will be the same.

[NaeemKhan333]

But u2net_portrait and u2net both are same network.I run the u2net model on the python portrait script,It is running and giving me output which is not a sktech but making mask.So it mean model are workable on the the python script but why not on the web demo.So my question is that is there is any thing missing in conversion . Because Both are same network although there output are not same in nature.