LOGAN: Latent Optimisation for Generative Adversarial Networks

TL;DR

This paper introduces LOGAN, a latent optimization method using natural gradients that enhances GAN training stability and performance, achieving state-of-the-art results on ImageNet with significant improvements in Inception Score and FID.

Contribution

The paper proposes a novel latent optimization technique for GANs that improves training dynamics and performance, outperforming previous models on large-scale datasets.

Findings

Achieves an Inception Score of 148 on ImageNet 128x128

Reduces FID to 3.4, outperforming baseline models

Demonstrates improved training stability and mode coverage

Abstract

Training generative adversarial networks requires balancing of delicate adversarial dynamics. Even with careful tuning, training may diverge or end up in a bad equilibrium with dropped modes. In this work, we improve CS-GAN with natural gradient-based latent optimisation and show that it improves adversarial dynamics by enhancing interactions between the discriminator and the generator. Our experiments demonstrate that latent optimisation can significantly improve GAN training, obtaining state-of-the-art performance for the ImageNet ($128 \times 128$) dataset. Our model achieves an Inception Score (IS) of $148$ and an Fr\'echet Inception Distance (FID) of $3.4$, an improvement of $17\%$ and $32\%$ in IS and FID respectively, compared with the baseline BigGAN-deep model with the same architecture and number of parameters.