About training on FFHQ

[Suimingzhe]

Thanks for your work.

I have two questions for training p2-weighting on FFHQ dataset.

1. Is it correct to set p2_gamma = 1 and p2_k = 1 for training face dataset (FFHQ) ?
2. I notice you choose a lightweight unet based on ADM. Do you think using a larger-scale unet will further improve the quality of generated fake face images? If so, which hyperparameters should I modify？

Thanks again and look forward to your reply.

Kind Regards

[jychoi118]

1. We used p2_gamma = 0.5 and p2_k = 1 for FFHQ. However, it is okay to use p2_gamma = 1.
2. Self-attention seems to be effective in datasets (FFHQ) where global consistency is important. You may try self-attention at multiple resolutions: --attention_resolutions 8,16,32

[Suimingzhe]

1. We used p2_gamma = 0.5 and p2_k = 1 for FFHQ. However, it is okay to use p2_gamma = 1.
2. Self-attention seems to be effective in datasets (FFHQ) where global consistency is important. You may try self-attention at multiple resolutions: --attention_resolutions 8,16,32

Thanks for your advice. I will try it as you said.

Kind Regards