Hacker News new | ask | show | jobs
MegaTrain Full Precision Training of 100B+ Parameter LLMs on a Single GPU (github.com)
1 points by adulau 36 days ago