I'm using auto-py-to-exe [settings are: One File, Window Based (hide the console)] to convert myPyQt5Code.py to an executable Windows file myApp.exe. It works fine and I can run myApp.exe, but if I import load_model (to read my model from the same directory) and then run the auto-py-to-exe again on the same myPyQt5Code.py file (with the new import line), it gives the following error (when I run myApp.exe):
I read many related questions/answers, some of them say that I should downgrade my tensorflow from 2.x to 1.14 because the PyInstaller cannot deal with higher versions of tensorflow, I downgraded but I still have the same error.
My current version of tensorflow is downgraded to 1.14 (I can upgrade it again to 2.0 if needed)
Update:
In sum, I need a set of compatible versions of PyQt5, Tensorflow, Keras and PyInstaller work correctly all together to deploy my project as an executable Windows application.
I would appreciate it if someone suggests any valid set of versions.