Open vfdev-5 opened 4 years ago
Yeah, I think it would be great :)
I would say, the repo is ~70% ready. It's somewhat useful (I hope), but there are a few things that are still missing from the original StyleGAN2 work. I also saw several papers about improving various aspects of StyleGAN, that I would like to try and test.
The original work contains optimized AMP settings (like the loss scaling coefficients), but I didn't study them closely: https://github.com/NVlabs/stylegan2/blob/master/dnnlib/tflib/optimizer.py
Hi @dmitry-vorobiev
I would like to add your project to https://github.com/pytorch/ignite#others and optionally even create a working Colab with wandb online logs and generated images by trained model. How would you estimate the readiness of this repository ?
I took a look on your "Projects" kanban and I'd also be interested in adding AMP training.
What do you think about all this ?
PS: optionally you can also ping me on ODS slack (@vfdev-5) if you would like to discuss more.