“> Apologies if this came across the wrong way I was uncertain but your other statements made me think that sentiment was unintentional. I just want to push back against it because it is too common and misused even with good intentions. I hope you don't see this as me saying anything about your character. Honestly, impressions are that you do care. > It s…”
Muon is Scalable for LLM Training
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.
候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。
HN 书评
推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及“A lot of people point to the Muon optimizer that Moonshot (the creators of Kimi) pioneered. Compared to the standard optimizer AdamW, Muon amplifies low-magnitude gradient directions which makes the model learn faster (and maybe gives Kimi its unique qualities). Muon paper: https://arxiv.org/abs/2502.16982”