Muon outperforms AdamW in expanding the Pareto frontier for compute-time tradeoff, enhancing data efficiency at large batch sizes while enabling economical training through effective hyperparameter transfer.

https://arxiv.org/abs//2505.02222

YouTube: https://www.youtube.com/@ArxivPapers

TikTok: https://www.tiktok.com/@arxiv_papers

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers