On the Effects of Batch and Weight Normalization in Generative Adversarial Networks
Sitao Xiang, Hao Li
TL;DR
This work critically examines Batch Normalization in GANs, showing it can destabilize training and degrade sample quality despite faster early convergence. It proposes Weight Normalization with a novel TReLU variant and residual-block adaptations, plus a reconstruction-based objective to quantify speed, stability, and quality. Extensive experiments on DCGAN and ResNet architectures across CelebA, LSUN, and CIFAR-10 demonstrate that WN delivers lower reconstruction error and more stable training than BN, while BN can induce artifacts and mode collapse. The findings advocate adopting Weight Normalization as a preferred normalization strategy for GANs and provide a practical evaluation framework to compare GAN training methods.
Abstract
Generative adversarial networks (GANs) are highly effective unsupervised learning frameworks that can generate very sharp data, even for data such as images with complex, highly multimodal distributions. However GANs are known to be very hard to train, suffering from problems such as mode collapse and disturbing visual artifacts. Batch normalization (BN) techniques have been introduced to address the training. Though BN accelerates the training in the beginning, our experiments show that the use of BN can be unstable and negatively impact the quality of the trained model. The evaluation of BN and numerous other recent schemes for improving GAN training is hindered by the lack of an effective objective quality measure for GAN models. To address these issues, we first introduce a weight normalization (WN) approach for GAN training that significantly improves the stability, efficiency and the quality of the generated samples. To allow a methodical evaluation, we introduce squared Euclidean reconstruction error on a test set as a new objective measure, to assess training performance in terms of speed, stability, and quality of generated samples. Our experiments with a standard DCGAN architecture on commonly used datasets (CelebA, LSUN bedroom, and CIFAR-10) indicate that training using WN is generally superior to BN for GANs, achieving 10% lower mean squared loss for reconstruction and significantly better qualitative results than BN. We further demonstrate the stability of WN on a 21-layer ResNet trained with the CelebA data set. The code for this paper is available at https://github.com/stormraiser/gan-weightnorm-resnet