Skip to content
This repository has been archived by the owner on Apr 21, 2021. It is now read-only.

what data would you plan to finish?? #2

Open
Johnson-yue opened this issue Jun 12, 2020 · 2 comments
Open

what data would you plan to finish?? #2

Johnson-yue opened this issue Jun 12, 2020 · 2 comments

Comments

@Johnson-yue
Copy link

pytorch version is very very import!!! I test some repo for re-implement stylegan-v2 but all failed

and then they never update。。。。or fixed bugs

I wish you code will working and if you have any problem for help, you can say in readme or call me

@huangzh13
Copy link
Owner

huangzh13 commented Jun 12, 2020

Thank you for your attention to my work, in fact I am doing this recently.
Regarding the implementation of stylegan2, I recommend https://github.com/genforce/mganprior to you. It includes a stylegan2 generator implemented in pure pytorch, but does not include the discriminator and training framework. The reason I recommend it is that its implementation is elegant and pure pytorch, so it is easy to understand and modify.
I tried to add this implementation to the training framework, but now I can't train normally. I am looking for the reason.
If you have any good suggestions, please let me know.

Best regard,
@huangzh13

@Johnson-yue
Copy link
Author

thank you for your recommend, I will check it,

About training, https://github.com/lucidrains/stylegan2-pytorch this repo seems to get good result , and someone can get too, but I failed! This training process is much more different from NVlabs,
but maybe it work, you can try it

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants