Uses some code from https://github.com/mohammadpz/pytorch_forward_forward (mostly a few utilities)

The goal of this project is to:

(1) write some generalizable code for forward-forward experiments (i.e. treat FWFW as a nn module and let the other modules vary)

(2) try to apply it to imagenet

Very unsure whether I will complete this due to other schoolwork at the moment