Slimmable Generative Adversarial Networks

Abstract

Generative adversarial networks (GANs) have achieved remarkable progress in recent years, but the continuously growing scale of models makes them challenging to deploy widely in practical applications. In particular, for real-time generation tasks, different devices require generators of different sizes due to varying computing power. In this paper, we introduce slimmable GANs (SlimGANs), which can flexibly switch the width of the generator to accommodate various quality-efficiency trade-offs at runtime. Specifically, we leverage multiple discriminators that share partial parameters to train the slimmable generator. To facilitate the \textit{consistency} between generators of different widths, we present a stepwise inplace distillation technique that encourages narrow generators to learn from wide ones. As for class-conditional generation, we propose a sliceable conditional batch normalization that incorporates the label information into different widths. Our methods are validated, both quantitatively and qualitatively, by extensive experiments and a detailed ablation study.

Accepted to AAAI 2021

References in corpus (6)

• GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
• AutoSlim: Towards One-Shot Architecture Search for Channel Numbers
• Improving Generalization and Stability of Generative Adversarial Networks
• Compressing GANs using Knowledge Distillation
• Mimicry: Towards the Reproducibility of GAN Research
• GAN Slimming: All-in-One GAN Compression by A Unified Optimization Framework