N
Hacker Next
new
past
show
ask
show
jobs
submit
login
▲
Show HN: UL-SMF – Open-source linear-complexity ~300x KV-cache compression
(
github.com
)
13 points by
liventruth
2 days ago
|
1 comment
add comment
Rendered at 08:19:39 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
colingauvin 1 days ago
[-]
This should be trivially demonstrable if it actually works. Cosine similarity is not a good metric, however, for attention compression.
liventruth 14 hours ago
[-]
[flagged]
liventruth 12 hours ago
[-]
[flagged]
liventruth 16 hours ago
[-]
[dead]
liventruth 2 days ago
[-]
[flagged]