Keras model gets "corrupted" when cloning it from my GitHub repository

Question

I have a Keras model which I use to recognize hand gestures, and it works well. However, when I push my program WITH the files of the model on my GitHub, and then clone the code onto a new machine (or even re-clone it onto my main machine for testing purposes), the model simply doesn't work. I get an error message saying :

C:\Users\user\AppData\Local\Programs\Python\Python38\lib\site-packages\keras\saving\saved_model\load.py:115: RuntimeWarning: Unexpected end-group tag: Not all data was converted
  metadata.ParseFromString(file_content)

and the program simply crashes when it tries to recognize a hand. However, if I simply copy the exact same files I pushed on the GitHub but that I kept as a backup on my PC in my program to replace the ones I cloned from GitHub, then it works. They should be the exact same files (only difference : one was pulled from GitHub, and the other one was a backup on my machine), and yet only one of them work. Any reasons why this might happen?

Answer 1

You can see from the log that all files are updated

[ Screenshot ]: