Large Scale GAN Training for High Fidelity Natural Image SynthesisDownload PDF

Published: 21 Dec 2018, Last Modified: 22 Oct 2023ICLR 2019 Conference Blind SubmissionReaders: Everyone
Abstract: Despite recent progress in generative image modeling, successfully generating high-resolution, diverse samples from complex datasets such as ImageNet remains an elusive goal. To this end, we train Generative Adversarial Networks at the largest scale yet attempted, and study the instabilities specific to such scale. We find that applying orthogonal regularization to the generator renders it amenable to a simple "truncation trick", allowing fine control over the trade-off between sample fidelity and variety by reducing the variance of the Generator's input. Our modifications lead to models which set the new state of the art in class-conditional image synthesis. When trained on ImageNet at 128x128 resolution, our models (BigGANs) achieve an Inception Score (IS) of 166.3 and Frechet Inception Distance (FID) of 9.6, improving over the previous best IS of 52.52 and FID of 18.65.
Keywords: GANs, Generative Models, Large Scale Training, Deep Learning
TL;DR: GANs benefit from scaling up.
Code: [![Papers with Code](/images/pwc_icon.svg) 30 community implementations](https://paperswithcode.com/paper/?openreview=B1xsqj09Fm)
Data: [ArtBench-10 (32x32)](https://paperswithcode.com/dataset/artbench-10), [CIFAR-10](https://paperswithcode.com/dataset/cifar-10), [ImageNet](https://paperswithcode.com/dataset/imagenet), [JFT-300M](https://paperswithcode.com/dataset/jft-300m)
Community Implementations: [![CatalyzeX](/images/catalyzex_icon.svg) 19 code implementations](https://www.catalyzex.com/paper/arxiv:1809.11096/code)
26 Replies

Loading