search

junyanz / iGAN

Interactive Image Generation via Generative Adversarial Networks
MIT License
3.97k stars 587 forks source link

PyTorch Version #27

Closed XiaohangZhan closed 2 years ago

XiaohangZhan commented 5 years ago

Thank you for your excellent work. I believe this code will be more readable and extensible if in PyTorch. So will the PyTorch version be released? A link of third-party implementation is also fine, though I cannot find one for now. Specifically, I'm interested in latent vector prediction and projection section.

junyanz commented 5 years ago

We are working on a PyTorch version. Hopefully, It will be released in summer.

zzzzzzian commented 5 years ago

Thank you in advance! It would be really great if PyTorch version is released! I am trying to apply iGAN to a project and the reading of code from Theano is painful. I find it hard to understand the constraint & manifold gradient descent part from the code. Is it possible to enlighten me a little about that part? I am also using PyTorch

junyanz commented 5 years ago

Yeah, I also forgot how to write Theano code. This function implements the loss function.

Lifedecoder commented 2 years ago

Two years and a half passed and the pytorch version of iGAN still not released?😂

junyanz commented 2 years ago

Sorry that it keeps getting delayed. If you want to use it for face editing, you can find it in photoshop "smart portrait". We have some new work with open-source PyTorch code that improves iGAN: pix2latent (compatible with BigGAN/StyleGAN.) and ganseeing (compatible with StyleGAN).

Lifedecoder commented 2 years ago

Got it, thanks!