Why Spectral Normalization Stabilizes GANs: Analysis and Improvements
Spectral normalization (SN) is a widely-used technique for improving the stability and sample quality of Generative Adversarial Networks (GANs). However, there is currently limited understanding of why SN is effective. In this work, we show that SN controls two important failure modes of GAN training: exploding and vanishing gradients. Our proofs illustrate a (perhaps unintentional) connection with the successful LeCun initialization. This connection helps to explain why the most popular implementation of SN for GANs requires no hyper-parameter tuning, whereas stricter implementations of SN have poor empirical performance out-of-the-box. Unlike LeCun initialization which only controls gradient vanishing at the beginning of training, SN preserves this property throughout training. Building on this theoretical understanding, we propose a new spectral normalization technique: Bidirectional Scaled Spectral Normalization (BSSN), which incorporates insights from later improvements to LeCun initialization: Xavier initialization and Kaiming initialization. Theoretically, we show that BSSN gives better gradient control than SN. Empirically, we demonstrate that it outperforms SN in sample quality and training stability on several benchmark datasets.
Code (1)
Methods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Towards Efficient and Unbiased Implementation of Lipschitz Continuity in GANs
Lipschitz continuity recently becomes popular in generative adversarial networks (GANs). It was observed that the Lipschitz regularized discriminator leads to improved training stability and sample quality. The mainstrea…
Spectral Normalization for Generative Adversarial Networks
One of the challenges in the study of generative adversarial networks is the instability of its training. In this paper, we propose a novel weight normalization technique called spectral normalization to stabilize the tr…
Image GenerationA Uniform Generalization Error Bound for Generative Adversarial Networks
This paper focuses on the theoretical investigation of unsupervised generalization theory of generative adversarial networks (GANs). We first formulate a more reasonable definition of general error and generalization bo…
Generalization BoundsGradient Normalization for Generative Adversarial Networks
In this paper, we propose a novel normalization method called gradient normalization (GN) to tackle the training instability of Generative Adversarial Networks (GANs) caused by the sharp gradient space. Unlike existing w…
Controllable Orthogonalization in Training DNNs
Orthogonality is widely used for training deep neural networks (DNNs) due to its ability to maintain all singular values of the Jacobian close to 1 and reduce redundancy in representation. This paper proposes a computati…
image-classificationImage Classification