In this paper, we propose a novel normalization method called penalty gradient normalization (PGN) to tackle the training instability of Generative Adversarial Networks (GANs) caused by the sharp gradient space. Unlike existing work such as gradient penalty and spectral normalization, the proposed PGN only imposes a penalty gradient norm constraint on the discriminator function, which increases the capacity of the discriminator. Moreover, the proposed penalty gradient normalization can be applied to different GAN architectures with little modification. Extensive experiments on three datasets show that GANs trained with penalty gradient normalization outperform existing methods in terms of both Frechet Inception and Distance and Inception Score.

翻译：本文提出了一种名为惩罚梯度归一化（PGN）的新型归一化方法，用于解决因梯度空间陡峭导致的生成对抗网络（GANs）训练不稳定性问题。与梯度惩罚和谱归一化等现有工作不同，所提出的PGN方法仅对判别器函数施加惩罚梯度范数约束，从而提升了判别器的容量。此外，该惩罚梯度归一化方法几乎无需改动即可应用于不同的GAN架构。在三个数据集上的大量实验表明，采用惩罚梯度归一化训练的GANs在Fréchet初始距离（FID）和初始分数（IS）两项指标上均优于现有方法。