Adding Trained PyTorch Model to App Inventor using extensions

I have trained a pretty simple ML model using PyTorch, I wonder if it is possible to create an extension related to that. I know that PyTorch model can definitely be moved to Java, but i don't know whether the App Inventor will run correctly with this model. Are there any limitations? Does it guarantee to work as long as the language and code are fine?

Are you making any progress with this? It seems no-one here has a definitive answer for you...

maybe TensorFlow.js can help.

1 Like