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

3 points by lmxyy 1mo ago ↗ HN

0 comments

[ 0.22 ms ] story [ 14.6 ms ] thread

No comments yet.

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

Source · Live DB stats