Skip to content

Latest commit

 

History

History
7 lines (6 loc) · 487 Bytes

README.md

File metadata and controls

7 lines (6 loc) · 487 Bytes

VAE-WGAN-GP

  • This is a Variational Auto-encoder Wasserstein Generative Adverserial Networks Gradient Penalty implementation in tensorflow.
  • It is based on the architecture proposed in original VAEGAN paper: https://arxiv.org/abs/1512.09300 .
  • WGAN and GP implementations are based on: https://keras.io/examples/generative/wgan_gp/ .
  • Encoder, Decoder and Discriminator networks exists in the code to be an example.
  • You can simply initialize VAEWGANGP class and .fit() .