Drop normalisation layers in critic #98
Description
In translation of book about deep learning and generative modelling it is said that in WGAN-GP models we should cut normalisation layers. This is strange because i can still see them on github page is it a mistake or not?