1
DynMuon: A Dynamic Spectral Shaping View of Muon
这篇论文提出了DynMuon，从动态谱成形视角重新理解Muon优化器，为训练大模型提供理论洞察与方法改进。
arXiv:2605.17109v1 Announce Type: new Abstract: In recent years, Muon has emerged as the dominant method for training large language models, and trans…