Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
best local runtime for mixed models
#11
What exact setup are you running? Is this local-only or are you routing anything externally?
Reply
#12
Did a small comparison this afternoon. My result was mixed: first run looked bad, second run looked normal.
Reply
#13
I'd start with the simpler stack.
mostly local
Reply
#14
Thanks, that helps. The pattern around INT4 has been pretty similar for me. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
Reply
#15
Good point. That matches what I saw on dual 3090.
Reply
#16
Interesting, I hadn't tried that. The pattern around batch size has been pretty similar for me. The biggest change for me was simplifying the stack instead of adding one more optimization layer.
small models can still be good
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)