Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
tokens/sec tuning ideas with TensorRT-LLM
#3
Quote:Good point. The pattern around tail latency has been pretty similar for me.

What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with DeepSeek and compare from there. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.
Reply


Messages In This Thread
Re: tokens/sec tuning ideas with TensorRT-LLM - by oxidemixer - 10-05-2026, 10:48 PM

Forum Jump:


Users browsing this thread: 1 Guest(s)