ViTGAN: Training GANs with Vision Transformers

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lee, Kwonjoon, Chang, Huiwen, Jiang, Lu, Zhang, Han, Tu, Zhuowen, Liu, Ce
Format: Preprint
Veröffentlicht: 2021
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909212378923008
author Lee, Kwonjoon
Chang, Huiwen
Jiang, Lu
Zhang, Han
Tu, Zhuowen
Liu, Ce
author_facet Lee, Kwonjoon
Chang, Huiwen
Jiang, Lu
Zhang, Han
Tu, Zhuowen
Liu, Ce
contents Recently, Vision Transformers (ViTs) have shown competitive performance on image recognition while requiring less vision-specific inductive biases. In this paper, we investigate if such performance can be extended to image generation. To this end, we integrate the ViT architecture into generative adversarial networks (GANs). For ViT discriminators, we observe that existing regularization methods for GANs interact poorly with self-attention, causing serious instability during training. To resolve this issue, we introduce several novel regularization techniques for training GANs with ViTs. For ViT generators, we examine architectural choices for latent and pixel mapping layers to facilitate convergence. Empirically, our approach, named ViTGAN, achieves comparable performance to the leading CNN-based GAN models on three datasets: CIFAR-10, CelebA, and LSUN bedroom.
format Preprint
id arxiv_https___arxiv_org_abs_2107_04589
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle ViTGAN: Training GANs with Vision Transformers
Lee, Kwonjoon
Chang, Huiwen
Jiang, Lu
Zhang, Han
Tu, Zhuowen
Liu, Ce
Computer Vision and Pattern Recognition
Machine Learning
Image and Video Processing
Recently, Vision Transformers (ViTs) have shown competitive performance on image recognition while requiring less vision-specific inductive biases. In this paper, we investigate if such performance can be extended to image generation. To this end, we integrate the ViT architecture into generative adversarial networks (GANs). For ViT discriminators, we observe that existing regularization methods for GANs interact poorly with self-attention, causing serious instability during training. To resolve this issue, we introduce several novel regularization techniques for training GANs with ViTs. For ViT generators, we examine architectural choices for latent and pixel mapping layers to facilitate convergence. Empirically, our approach, named ViTGAN, achieves comparable performance to the leading CNN-based GAN models on three datasets: CIFAR-10, CelebA, and LSUN bedroom.
title ViTGAN: Training GANs with Vision Transformers
topic Computer Vision and Pattern Recognition
Machine Learning
Image and Video Processing
url https://arxiv.org/abs/2107.04589