Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
desk setup thread on 5090
#11
I'd approach it in a boring order. If you're on vLLM, I'd confirm the version first and then reduce variables until the behavior is obvious.
Reply
#12
same here on my side.
throughput matters, but tail latency matters too
Reply
#13
same here on my side.
small models can still be good
Reply
#14
Good point. I don't think your conclusion is far off.
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)