Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
loading thread…
4 ms
·
Sequoia: Speculative decoding boosting LLM inference by 8-10x
3 points
by
fgfm
3y ago