VC-Attention: Faster Low-Bit Attention Without Retraining (nunchux.ai)

3 points by lmxyy 10h ago ↗ HN

0 comments

[ 0.33 ms ] story [ 7.7 ms ] thread

No comments yet.

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

Source · Live DB stats