BEGAN: Boundary Equilibrium Generative Adversarial Networks

Code

• github.com/Heumi/BEGAN-tensorflow
  tf★ 0
• github.com/davidismael/BEGAN
  tf★ 0
• github.com/artcg/BEGAN
  tf★ 0
• github.com/vbnmzxc9513/GAN_BEGAN_hw2
  pytorch★ 0
• github.com/mlvc-lab/BeGan_pytorch
  pytorch★ 0
• github.com/carpedm20/BEGAN-tensorflow
  tf★ 0
• github.com/timsainb/GAIA
  tf★ 0
• github.com/taey16/pix2pixBEGAN.pytorch
  pytorch★ 0
• github.com/evan11401/CS_IOC5008_0856043_HW2
  pytorch★ 0
• github.com/consequencesunintended/BEGAN
  tf★ 0

Abstract

We propose a new equilibrium enforcing method paired with a loss derived from the Wasserstein distance for training auto-encoder based Generative Adversarial Networks. This method balances the generator and discriminator during training. Additionally, it provides a new approximate convergence measure, fast and stable training and high visual quality. We also derive a way of controlling the trade-off between image diversity and visual quality. We focus on the image generation task, setting a new milestone in visual quality, even at higher resolutions. This is achieved while using a relatively simple model architecture and a standard training procedure.

Tasks

Reproductions