search

rosinality / stylegan2-pytorch

Implementation of Analyzing and Improving the Image Quality of StyleGAN (StyleGAN 2) in PyTorch
MIT License
2.74k stars 623 forks source link

argument ada_every is not used #282

Open dhinz94 opened 2 years ago

dhinz94 commented 2 years ago

Thank you for your clean implementation! I just tried training a model with adaptive augmentation and noticed, that the argument --ada_every is not used when initializing the AdaptiveAugment object. Instead of using the argument, 8 is set as the update interval for augmentation. Best regards! https://github.com/rosinality/stylegan2-pytorch/blob/bef283a1c24087da704d16c30abc8e36e63efa0e/train.py#L424-L429

https://github.com/rosinality/stylegan2-pytorch/blob/bef283a1c24087da704d16c30abc8e36e63efa0e/train.py#L157

anguoyang commented 1 year ago

I have the same question, do you find the answer?