Adaptive speculative decoding: picking draft lengths at runtime (fergusfinn.com)

5 points by hasheddan 21d ago ↗ HN

0 comments

[ 3.8 ms ] story [ 12.2 ms ] thread

No comments yet.

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

Source · Live DB stats