Dark Hacker News
new
|
best
|
ask
|
show
|
jobs
pidtom | Dark Hacker News
user:
pidtom
created:
March 27, 2026
karma:
5
submissions
comments
1.
Skipping 90% of KV dequant work speeds up LLM decode by 22%
(github.com)
1 points
by
pidtom
52 days ago
|
0 comments