search

junyanz / iGAN

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

Regarding the upcoming variational autoencoder feature #4

Closed kaihuchen closed 7 years ago

kaihuchen commented 7 years ago

This is great stuff!

The README.md mentioned that there is a plan to support variational autoencoder, which I very much look forward to. My understanding is that the vanilla VAE tends to create blurry images (perhaps due to the disentangled representation that it learns), and I wonder if the planned VAE feature will be based on the vanilla VAE, or something like the autoencoding beyond pixels, or adversarial autoencoders?

junyanz commented 7 years ago

We are looking at different architectures right now such as VAE-GAN or PPGAN. I am busy with a paper deadline right now, and plan to add new architectures by the late Jan or early Feb. If you would like to help add a particular network, send us a pull request.