GAN已死；GAN萬歲！一個現代GAN基準。

The GAN is dead; long live the GAN! A Modern GAN Baseline

January 9, 2025

作者: Yiwen Huang, Aaron Gokaslan, Volodymyr Kuleshov, James Tompkin

cs.AI

摘要

有一個廣泛流傳的說法是，生成對抗網絡（GANs）很難訓練，文獻中的GAN架構充斥著經驗性技巧。我們提供證據反駁這一說法，並以更為原則性的方式建立一個現代GAN基線。首先，我們推導出一個行為良好的正則化相對論GAN損失，解決了以往通過一堆臨時技巧來應對的模式丟失和不收斂問題。我們從數學上分析我們的損失並證明它具有局部收斂保證，這與大多數現有的相對論損失不同。其次，我們的新損失使我們能夠放棄所有臨時技巧，並將常見GAN中使用的過時骨幹替換為現代架構。以StyleGAN2為例，我們提出了一個簡化和現代化的路線圖，形成了一個新的極簡主義基線 - R3GAN。儘管簡單，我們的方法在FFHQ、ImageNet、CIFAR和Stacked MNIST數據集上超越了StyleGAN2，並與最先進的GAN和擴散模型相比表現出色。

English

There is a widely-spread claim that GANs are difficult to train, and GAN architectures in the literature are littered with empirical tricks. We provide evidence against this claim and build a modern GAN baseline in a more principled manner. First, we derive a well-behaved regularized relativistic GAN loss that addresses issues of mode dropping and non-convergence that were previously tackled via a bag of ad-hoc tricks. We analyze our loss mathematically and prove that it admits local convergence guarantees, unlike most existing relativistic losses. Second, our new loss allows us to discard all ad-hoc tricks and replace outdated backbones used in common GANs with modern architectures. Using StyleGAN2 as an example, we present a roadmap of simplification and modernization that results in a new minimalist baseline -- R3GAN. Despite being simple, our approach surpasses StyleGAN2 on FFHQ, ImageNet, CIFAR, and Stacked MNIST datasets, and compares favorably against state-of-the-art GANs and diffusion models.