Improved Training of Wasserstein GANs

TL;DR

This paper introduces a gradient penalty method for training Wasserstein GANs, improving stability and sample quality over traditional weight clipping approaches, and enabling effective training of deep and diverse models.

Contribution

It proposes a gradient penalty technique as an alternative to weight clipping, enhancing the stability and quality of Wasserstein GAN training.

Findings

Gradient penalty outperforms weight clipping in WGANs.

Stable training achieved across various architectures including deep ResNets.

High-quality samples generated on CIFAR-10 and LSUN datasets.

Abstract

Generative Adversarial Networks (GANs) are powerful generative models, but suffer from training instability. The recently proposed Wasserstein GAN (WGAN) makes progress toward stable training of GANs, but sometimes can still generate only low-quality samples or fail to converge. We find that these problems are often due to the use of weight clipping in WGAN to enforce a Lipschitz constraint on the critic, which can lead to undesired behavior. We propose an alternative to clipping weights: penalize the norm of gradient of the critic with respect to its input. Our proposed method performs better than standard WGAN and enables stable training of a wide variety of GAN architectures with almost no hyperparameter tuning, including 101-layer ResNets and language models over discrete data. We also achieve high quality generations on CIFAR-10 and LSUN bedrooms.