Retro, comparable performance to GPT-3 using 25× fewer parameters (deepmind.com)

1 points by merqurio 4y ago ↗ HN

0 comments

[ 2.2 ms ] story [ 11.8 ms ] thread

No comments yet.

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

Source · Live DB stats