Linux stops training while windows doesen't

1

Hi!

I am new to Linux, trying too learn it while practicing some ML.
I am training a model, the size of which you can select before training.
When working in Linux (Mint), if I set the depth of the model higher then 27, the training stops with no alarms/messages. This does not happen on windows.
Also, I create a separate thread for the training so it does not freeze up the UI.

I have latest python and pytorch I believe on both operating systems.

How do I go about debugging this?

2

What kind of UI are you using?

3

I built it with tkinter, ttkbootstrep for better looks.

4

removing the ui fixed the issue, so now I know where to keep poking. Thanks for the pointer!

5

Good to hear removing the UI helped and thanks for the update.