drange of the generative model

mit-han-lab / data-efficient-gans

[NeurIPS 2020] Differentiable Augmentation for Data-Efficient GAN Training

https://arxiv.org/abs/2006.10738

BSD 2-Clause "Simplified" License

1.27k stars 175 forks source link

drange of the generative model #72

Open slala2121 opened 3 years ago

slala2121 commented 3 years ago

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., does this make training easier?

zsyzzsoft commented 3 years ago

This is borrowed from recent GAN implementations like StyleGAN2, but I think maybe they do not make much difference.