Hacker News
new
past
comments
ask
show
jobs
Show HN: TrnRT – An open-source LLM inference engine for AWS Trainium
(github.com)
18 points
by
sanchitmonga22
2 hours ago
|
2 comments
by
bookernath
1 hours ago
|
next
[-]
Tokens per minute is a strange metric, I usually see tok/s. A per-second figure would be more accessible in comparing this to alternatives.
reply
by
badmonster0
20 minutes ago
|
prev
|
[-]
how does this compare with sglang
reply