Running without MLP

[mrgloom]

I was curious how crucial are MLP layers, so I trained 64p model without them (i.e. comment out this section https://github.com/rosinality/style-based-gan-pytorch/blob/master/model.py#L458-L460)

Here is comparison:

320k with mlp layers:

320k without mlp layers:

Seems it's better with mlp layers, but it's not that bad without them.