NVFP4 / INT4 tradeoffs

[northhall88]

Thought I'd post a small recipe because a lot of threads ask similar things.

Maybe useful as a starting point: LM Studio on M3 Max, keeping structured output in mind from the start.

If someone wants the exact settings I can clean them up and post them.

[stackarch]

same here on my side.

[visionforge]

My first guess would be this: On mini PC, the easiest win is often to back off the aggressive settings and see whether the problem survives. Once you know the plain path works, then add the faster path back in piece by piece.

[cachebench]

Maybe, but I had the opposite result. I wouldn't optimize around that first. On my side the bigger swing came from the runtime and not the model. Worth testing the simpler explanation before adopting the more interesting one.

[turboscope]

Maybe, but I had the opposite result. I wouldn't optimize around that first. On my side the bigger swing came from the runtime and not the model. Wort

Need a little more detail here. Are you testing this on B200 or something else?

[quietcraft]

I tested this tonight. My result was mixed: first run looked bad, second run looked normal. The main thing that moved was latency, not quality.

[audiodesk]

Appreciate the detail. That matches what I saw on 7900 XTX. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[relaysupply90]

Appreciate the detail. That matches what I saw on 7900 XTX. Worth rechecking your defaults though, because one bad assumption can make the rest of the

Can you fill in a bit more context? What model, what runtime, and roughly what context size?

[kernelhouse70]

that fixed it for me too.

[rocmcrate26]

What exact setup are you running? Are you testing this on Threadripper box or something else?