Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
loading thread…
3 ms
·
Eagle-3 Speculative Decoding for LLM Inference (5.6x speedup)
2 points
by
summarity
2y ago