Pro Gan Pytorch Versions Save

Unofficial PyTorch implementation of the paper titled "Progressive growing of GANs for improved Quality, Stability, and Variation"

v_1.2.0

5 years ago

Release v_1.2.0:

New features:

  1. Support for other loss functions: "LSGAN", "LSGAN-WITH-SIGMOID" along with "WGAN" and "WGAN-GP".
  2. MinibatchStd layer numerically stablized.
  3. Loss interface can be extended to create your own loss.

Todo for next release

  1. Add support for Relativistic GAN loss, Sliced Wasserstein loss.

Todo for a future release

  1. Add samples
  2. support for Conditional GAN