{"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/muon-is-scalable-for-llm-training","title":"Muon is Scalable for LLM Training","arxiv_id":"2502.16982","date":"2025-02-24","proceeding":null,"authors":["Jingyuan Liu","Jianlin Su","Xingcheng Yao","Zhejun Jiang","Guokun Lai","Yulun Du","Yidao Qin","Weixin Xu","Enzhe Lu","Junjie Yan","Yanru Chen","Huabin Zheng","Yibo Liu","Shaowei Liu","Bohong Yin","Weiran He","Han Zhu","Yuzhi Wang","Jianzhou Wang","Mengnan Dong","Zheng Zhang","Yongsheng Kang","Hao Zhang","Xinran Xu","Yutao Zhang","Yuxin Wu","Xinyu Zhou","Zhilin Yang"],"abstract":"Recently, the Muon optimizer based on matrix orthogonalization has demonstrated strong results in training small-scale language models, but the scalability to larger models has not been proven. We identify two crucial techniques for scaling up Muon: (1) adding weight decay and (2) carefully adjusting the per-parameter update scale. These techniques allow Muon to work out-of-the-box on large-scale training without the need of hyper-parameter tuning. Scaling law experiments indicate that Muon achieves $\\sim\\!2\\times$ computational efficiency compared to AdamW with compute optimal training. Based on these improvements, we introduce Moonlight, a 3B/16B-parameter Mixture-of-Expert (MoE) model trained with 5.7T tokens using Muon. Our model improves the current Pareto frontier, achieving better performance with much fewer training FLOPs compared to prior models. We open-source our distributed Muon implementation that is memory optimal and communication efficient. We also release the pretrained, instruction-tuned, and intermediate checkpoints to support future research.","url_abs":"https://arxiv.org/abs/2502.16982v1","url_pdf":"https://arxiv.org/pdf/2502.16982v1.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":"muon-is-scalable-for-llm-training","repo_url":"https://github.com/KellerJordan/Muon","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"computational-efficiency","task_name":"Computational Efficiency"}],"methods":[{"method_slug":"adamw","method_name":"AdamW"},{"method_slug":"weight-decay","method_name":"Weight Decay"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2502.16982","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}