Also, I believe the newer fast.ai course uses Pytorch, and the prior used Keras?

I would like to know more from the article about setting x,y,m1, and m2. Any explanation is appreciated.