I am currently taking lesson 2 on Fast.AI, and I see they teach GAN later too. However, it seems to be what they are using is not Pytorch, but their library or a library on top of Pytorch. But I just want to learn Pytorch that I can do like GAN, Time series prediction, NPL and more. Is taking their course to allow me to learn Pytorch more deeply in these fields?
I’m kinda half way through fast.ai but I think I want to learn the proper steps as fast.ai I feel is the ‘wham bam’ type. What is good I think it’s the one-cyce-policy that is implemented but I did do a search and there are 3rd party codes for this in pytorch, my 2 cents