{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/on-the-effects-of-batch-and-weight","title":"On the Effects of Batch and Weight Normalization in Generative Adversarial Networks","arxiv_id":"1704.03971","date":"2017-04-13","proceeding":null,"authors":["Sitao Xiang","Hao Li"],"abstract":"Generative adversarial networks (GANs) are highly effective unsupervised\nlearning frameworks that can generate very sharp data, even for data such as\nimages with complex, highly multimodal distributions. However GANs are known to\nbe very hard to train, suffering from problems such as mode collapse and\ndisturbing visual artifacts. Batch normalization (BN) techniques have been\nintroduced to address the training. Though BN accelerates the training in the\nbeginning, our experiments show that the use of BN can be unstable and\nnegatively impact the quality of the trained model. The evaluation of BN and\nnumerous other recent schemes for improving GAN training is hindered by the\nlack of an effective objective quality measure for GAN models. To address these\nissues, we first introduce a weight normalization (WN) approach for GAN\ntraining that significantly improves the stability, efficiency and the quality\nof the generated samples. To allow a methodical evaluation, we introduce\nsquared Euclidean reconstruction error on a test set as a new objective\nmeasure, to assess training performance in terms of speed, stability, and\nquality of generated samples. Our experiments with a standard DCGAN\narchitecture on commonly used datasets (CelebA, LSUN bedroom, and CIFAR-10)\nindicate that training using WN is generally superior to BN for GANs, achieving\n10% lower mean squared loss for reconstruction and significantly better\nqualitative results than BN. We further demonstrate the stability of WN on a\n21-layer ResNet trained with the CelebA data set. The code for this paper is\navailable at https://github.com/stormraiser/gan-weightnorm-resnet","url_abs":"http://arxiv.org/abs/1704.03971v4","url_pdf":"http://arxiv.org/pdf/1704.03971v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"on-the-effects-of-batch-and-weight","repo_url":"https://github.com/stormraiser/gan-weightnorm-resnet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"on-the-effects-of-batch-and-weight","repo_url":"https://github.com/fyumoto/MHGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"on-the-effects-of-batch-and-weight","repo_url":"https://github.com/nardeas/MHGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"bottleneck-residual-block","method_name":"Bottleneck Residual Block"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"kaiming-initialization","method_name":"Kaiming Initialization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-block","method_name":"Residual Block"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"weight-normalization","method_name":"Weight Normalization"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1704.03971","atlas_url":"https://app.syntology.ai/?focus=1704.03971","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}