Home » Python » python – How to test your trained model and detect your own class-Exceptionshub

python – How to test your trained model and detect your own class-Exceptionshub

Posted by: admin February 24, 2020 Leave a comment

Questions:

I have used the following post in Github:
https://github.com/YunYang1994/tensorflow-yolov3/tree/eb02b0731dfb1c2614e216c619ac7b22db57d994

With that I have been able to re-train the network from COCO weights with my DataSets. Therefore, I have obtain the following files:
.data
.meta
.index

Now I need to generate the .pb file in order to execute ‘image_demo.py’ and test the network with my weights, one test image and detect the new added class.

Questions:

  1. How can I generate .pb file from this three files? .data // .meta // .index

  2. Once I have the .pb, how can I add my new trained class to be detected with the ‘image_demo.py’ code?

Thank you very much in advance!

Best regards.

How to&Answers: