Share your thoughts, 1 month free Claude Pro on usSee more
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
Class-conditional Image GenerationImageNet 256x256
Inception Score (IS)225.5
967
Image GenerationImageNet 256x256
IS225.5
517
Class-conditional Image GenerationImageNet 256x256 (val)
Inception Score (IS)225.5
493
Image GenerationImageNet 256x256 (val)
FID3.45
399
Class-conditional Image GenerationImageNet 256x256 (train)
IS225.5
367
Class-conditional Image GenerationImageNet 256x256 (test)
FID3.45
223
Image GenerationImageNet 256x256 (train)
FID3.45
211
Class-conditional Image GenerationImageNet 256x256 (train val)
FID3.45
203
Class-conditional generationImageNet 256 x 256 1k (val)
FID3.45
104
Image GenerationImageNet
FID3.45
101
Showing 10 of 29 rows

Other info

Follow for update