Efficient RLHF: Reducing the Memory Usage of PPO (arxiv.org)

2 points by sestinj 3y ago ↗ HN

0 comments

[ 3.3 ms ] story [ 5.5 ms ] thread

No comments yet.

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

Source · Live DB stats