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

1 points by adulau 1mo ago ↗ HN

0 comments

[ 2.6 ms ] story [ 9.9 ms ] thread

No comments yet.

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

Source · Live DB stats