data-efficient-gans

data-efficient-gans copied to clipboard

pre-trained discriminator

1

Hi, Some refinement methods use the trained discriminator for generating images. I am trying to apply these methods to DiffAugment to see whether it can improve the results. May I...

ngoc-nguyen-0

Question, during training score spike up

1

Hi! I'd like to know in your experience, what cause the sudden spike up during the training. Thank you. 

aprilycliu

Question of FID score

2

Hi, It seems like the reported baseline FID results are slightly different from the value reported in ADA (e.g. FID for FFHQ1k is 62.16 in your paper and 100.16 in...

cuikaiwen18

Regarding implementation of DiffAugment with BiGAN

6

Hi, Thanks a lot for your work. I just wanted to ask that as in paper, Diff Augment is only done in Discriminator. So if we want to try with...

shreejalt

drange of the generative model

1

Does the drange affect the quality of the generated images? I think the default is [-1,1] in the code. Is there a reason why this as opposed to [0,1]? e.g.,...

slala2121

Question about image sampling

1

Hi, Do you sample the 1k images from the FFHQ dataset randomly for training or is there a specific way to sample these images?

cuikaiwen18

FID scores spike up at the beginning of training

9

This is a plot of FID scores over time in ticks:  I'm wondering what is causing the initial spike upward. I've tried a bunch of things to avoid it...

ghost

In https://github.com/mit-han-lab/data-efficient-gans/blob/master/DiffAugment-stylegan2/training/training_loop.py#L92 it says: ```python # Learning rate. s.G_lrate = G_lrate_dict.get(s.resolution, G_lrate_base) s.D_lrate = D_lrate_dict.get(s.resolution, D_lrate_base) if lrate_rampup_kimg > 0: rampup = min(s.kimg / lrate_rampup_kimg, 1.0) s.G_lrate *= rampup s.D_lrate...

ghost

Disable path regularization and lazy regularization

3

Hi, In the paper, for FFHQ 1k 256x256 training with DiffAugment, it is written that path length regularization and lazy regularization is disabled. If I am not wrong, in the...

nupurkmr9

How to train StyleGAN2 with labeled dataset

2

Hello, I found that compared with the native stylegan2, your code specifically removes the label. I want to ask you whether you have tried to add the label for training,...

TccccD