News
Newest
Show
Jobs
Show HN: TrnRT – An open-source LLM inference engine for AWS Trainium
(github.com)
19 points
by
sanchitmonga22
2 hours ago
2 comments
bookernath
1 hour ago
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.
badmonster0
36 minutes ago
how does this compare with sglang