Muon Is Scalable for LLM Training

[yorwba]

For people who want to know more about the Muon optimizer: https://kellerjordan.github.io/posts/muon/