Improved Consistency Regularization for GANs

Abstract

Recent work has increased the performance of Generative Adversarial Networks (GANs) by enforcing a consistency cost on the discriminator. We improve on this technique in several ways. We first show that consistency regularization can introduce artifacts into the GAN samples and explain how to fix this issue. We then propose several modifications to the consistency regularization procedure designed to improve its performance. We carry out extensive experiments quantifying the benefit of our improvements. For unconditional image synthesis on CIFAR-10 and CelebA, our modifications yield the best known FID scores on various GAN architectures. For conditional image synthesis on CIFAR-10, we improve the state-of-the-art FID score from 11.48 to 9.21. Finally, on ImageNet-2012, we apply our technique to the original BigGAN model and improve the FID from 6.66 to 5.38, which is the best score at that model size.

arXiv commentAAAI 2021

References in corpus (24)

• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Spectral Normalization for Generative Adversarial Networks
• Improved Regularization of Convolutional Neural Networks with Cutout
• Self-Attention Generative Adversarial Networks
• Conditional Image Synthesis With Auxiliary Classifier GANs
• Unsupervised Data Augmentation for Consistency Training
• Temporal Ensembling for Semi-Supervised Learning
• Improved Techniques for Training GANs
• Training Generative Adversarial Networks with Limited Data
• Generating Natural Adversarial Examples
• Learning with Pseudo-Ensembles
• On Convergence and Stability of GANs
• Diversity-Sensitive Conditional Generative Adversarial Networks
• Image Augmentations for GAN Training
• Differentiable Augmentation for Data-Efficient GAN Training
• Hierarchical Implicit Models and Likelihood-Free Variational Inference
• Geometric GAN
• LOGAN: Latent Optimisation for Generative Adversarial Networks
• Improving the Improved Training of Wasserstein GANs: A Consistency Term and Its Dual Effect
• Is Generator Conditioning Causally Related to GAN Performance?
• Small-GAN: Speeding Up GAN Training Using Core-sets
• Skill Rating for Generative Models

Cited by in corpus (14)

• Training Generative Adversarial Networks with Limited Data
• Freeze the Discriminator: a Simple Baseline for Fine-Tuning GANs
• Image Augmentations for GAN Training
• ContraGAN: Contrastive Learning for Conditional Image Generation
• Top-k Training of GANs: Improving GAN Performance by Throwing Away Bad Samples
• Regularizing Generative Adversarial Networks under Limited Data
• Triple Generative Adversarial Networks
• AlphaGAN: Fully Differentiable Architecture Search for Generative Adversarial Networks
• Data-Efficient GAN Training Beyond (Just) Augmentations: A Lottery Ticket Perspective
• Self-Supervised GANs with Label Augmentation
• Dual Contrastive Loss and Attention for GANs
• Learning Temporally Invariant and Localizable Features via Data Augmentation for Video Recognition
• Temporal-Aware Self-Supervised Learning for 3D Hand Pose and Mesh Estimation in Videos
• MVHM: A Large-Scale Multi-View Hand Mesh Benchmark for Accurate 3D Hand Pose Estimation