Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
continuous batching wins and losses with SGLang
#7
My first guess would be this: If you're on llama.cpp, I'd confirm the version first and then reduce variables until the behavior is obvious. Not glamorous advice, but repeatable baselines save a lot of time.

If you post your exact settings, people here can probably narrow it down pretty fast.
team red, team patient
Reply


Messages In This Thread
Re: continuous batching wins and losses with SGLang - by cozykite - 10-19-2026, 03:05 AM

Forum Jump:


Users browsing this thread: 1 Guest(s)