Why long context eats your VRAM: the KV cache explained (vettedconsumer.com)

3 points by ermantrout 1mo ago ↗ HN

0 comments

[ 0.23 ms ] story [ 7.8 ms ] thread

No comments yet.

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

Source · Live DB stats