{"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/equi-normalization-of-neural-networks","title":"Equi-normalization of Neural Networks","arxiv_id":"1902.10416","date":"2019-02-27","proceeding":"ICLR 2019 5","authors":["Pierre Stock","Benjamin Graham","Rémi Gribonval","Hervé Jégou"],"abstract":"Modern neural networks are over-parametrized. In particular, each rectified\nlinear hidden unit can be modified by a multiplicative factor by adjusting\ninput and output weights, without changing the rest of the network. Inspired by\nthe Sinkhorn-Knopp algorithm, we introduce a fast iterative method for\nminimizing the L2 norm of the weights, equivalently the weight decay\nregularizer. It provably converges to a unique solution. Interleaving our\nalgorithm with SGD during training improves the test accuracy. For small\nbatches, our approach offers an alternative to batch-and group-normalization on\nCIFAR-10 and ImageNet with a ResNet-18.","url_abs":"http://arxiv.org/abs/1902.10416v1","url_pdf":"http://arxiv.org/pdf/1902.10416v1.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":"equi-normalization-of-neural-networks","repo_url":"https://github.com/facebookresearch/enorm","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[],"methods":[{"method_slug":"sgd","method_name":"SGD"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1902.10416","atlas_url":"https://app.syntology.ai/?focus=1902.10416","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}