I am trying to build GAN from scratch for super resolution. I am using Robust U-net for GAN’s generator and ResNet34 for discriminator.
I have seen a lot of examples where people are using simple linear layers for discriminator. Which should be for discriminator and why ?
Thanks in advance.