Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
|
ArtRichards's favorites
login
submissions
|
comments
1.
Lossless LLM compression for efficient GPU inference via dynamic-length float
(
arxiv.org
)
411 points
by
CharlesW
7 months ago
|
117 comments
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
