Training the model on a single GPU

[shuvam-creditmate]

Hi! I'm aware that the question has been asked previously, but could you guide how to modify jigclu to remove the distributeddataparallel depedency?

Thanks!

[akuxcw]

Actually, I didn't try to run on a single GPU. In my opinion, the most important part you need to change lies in jigclu.py. This file contains gather multi-GPU results to calculate the loss, and you need to delete these parts. A good way is trying to run this code on a single GPU and see the errors, then fix the bugs accordingly.

[rnangia]

@shuvam-creditmate Have you managed to run it on a single GPU?