Preprint Open access
MeqMuon: Matrix-Equilibrating Muon for LLM Pretraining
The success of large language models (LLMs) has been accompanied by continued growth in model size and pretraining costs. Muon offers high accuracy and training efficiency in LLM pretraining. Recent work introduces row-wise normalization into Muon to balance update magnitudes and improve pretraining performance. Howeve …