Augment remains at 0.000 and stats.jsonl reports scores/fake as NaN

autonomousvision / stylegan-xl

[SIGGRAPH'22] StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasets

MIT License

961 stars 113 forks source link

Augment remains at 0.000 and stats.jsonl reports scores/fake as NaN #91

Closed nom57 closed 2 years ago

nom57 commented 2 years ago

is this expected behavior?

I had started a model at 32x32 , started looking good at 100kimg , collapsed after , and now even at 500kimg it is not looking very good should I try from scratch again?

nom57 commented 2 years ago

found the solution it was misconfig