Unrolled Generative Adversarial Networks

Code

• github.com/poolio/unrolled_gan
  OfficialIn papertf★ 0
• github.com/mangoubi/Min-max-optimization-algorithm-for-training-GANs
  tf★ 11
• github.com/andrewliao11/unrolled-gans
  pytorch★ 0
• github.com/locuslab/gradient_regularized_gan
  tf★ 0
• github.com/apaszke/pytorch-dist
  pytorch★ 0
• github.com/chameleonTK/continual-learning-for-HAR
  pytorch★ 0
• github.com/lyken17/pytorch
  pytorch★ 0
• github.com/MarisaKirisame/unroll_gan
  pytorch★ 0
• github.com/alex98chen/testGAN
  tf★ 0

Abstract

We introduce a method to stabilize Generative Adversarial Networks (GANs) by defining the generator objective with respect to an unrolled optimization of the discriminator. This allows training to be adjusted between using the optimal discriminator in the generator's objective, which is ideal but infeasible in practice, and using the current value of the discriminator, which is often unstable and leads to poor solutions. We show how this technique solves the common problem of mode collapse, stabilizes training of GANs with complex recurrent generators, and increases diversity and coverage of the data distribution by the generator.

Tasks

Reproductions