NHacker Next
  • new
  • past
  • show
  • ask
  • show
  • jobs
  • submit
▲Show HN: TrnRT – An open-source LLM inference engine for AWS Trainium (github.com)
bookernath 1 hours 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 20 minutes ago [-]
how does this compare with sglang
Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact
Rendered at 18:42:23 GMT+0000 (Coordinated Universal Time) with Vercel.