Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
_0ffh
50 days ago
|
parent
|
context
|
favorite
| on:
DSpark: Speculative decoding accelerates LLM infer...
Lookahead Sparse Attention should be playing a big role as well, as it dramatically slashes memory consumption.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: