yikedi

Question about choice of loss and normalization techniques

2

The paper states that WGAN-GP increases the training stability so I am curious what other loss and normalization techniques you have tried. I am trying to add spectral normalization suggested...

Problem about generating random samples of arbitrary size

1

I can only generate random samples of arbitrary size with gen_start_scale=0. I would get runtime error if I use gen_start_scale bigger than 0.