MegaTrain Full Precision Training of 100B+ Parameter LLMs on a Single GPU (github.com)

1 points by adulau 4mo ago ↗ HN

0 comments

[ 2.7 ms ] story [ 10.6 ms ] thread

No comments yet.

A showcase of shard-db. Data from the Hacker News API. Refreshed every 15 minutes.

Source · Live DB stats