Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Scaling up GANs for Text-to-Image Synthesis

About

The recent success of text-to-image synthesis has taken the world by storm and captured the general public's imagination. From a technical standpoint, it also marked a drastic change in the favored architecture to design generative image models. GANs used to be the de facto choice, with techniques like StyleGAN. With DALL-E 2, auto-regressive and diffusion models became the new standard for large-scale generative models overnight. This rapid shift raises a fundamental question: can we scale up GANs to benefit from large datasets like LAION? We find that na\"Ively increasing the capacity of the StyleGAN architecture quickly becomes unstable. We introduce GigaGAN, a new GAN architecture that far exceeds this limit, demonstrating GANs as a viable option for text-to-image synthesis. GigaGAN offers three major advantages. First, it is orders of magnitude faster at inference time, taking only 0.13 seconds to synthesize a 512px image. Second, it can synthesize high-resolution images, for example, 16-megapixel pixels in 3.66 seconds. Finally, GigaGAN supports various latent space editing applications such as latent interpolation, style mixing, and vector arithmetic operations.

Minguk Kang, Jun-Yan Zhu, Richard Zhang, Jaesik Park, Eli Shechtman, Sylvain Paris, Taesung Park• 2023

Related benchmarks

TaskDatasetResultRank
Image GenerationImageNet 256x256 (val)
FID3.45
307
Class-conditional Image GenerationImageNet 256x256 (train)
IS225.5
305
Class-conditional Image GenerationImageNet 256x256 (val)
FID3.45
293
Image GenerationImageNet 256x256
FID3.45
243
Class-conditional Image GenerationImageNet 256x256 (train val)
FID3.45
178
Class-conditional Image GenerationImageNet 256x256 (test)
FID3.45
167
Image GenerationImageNet 256x256 (train)
FID3.45
91
Class-conditional Image GenerationImageNet-1k (val)
FID3.45
68
Class-conditional generationImageNet 256 x 256 1k (val)
FID3.45
67
Class-conditional image synthesisImageNet 256x256 (val)
FID3.45
61
Showing 10 of 27 rows

Other info

Follow for update