best local runtime for mixed models

[rocmpath]

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

Maybe useful as a starting point: LM Studio on B200, keeping continuous batching in mind from the start.

Feel free to tear it apart if you see something dumb.

For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.

[orbitbench6]

I tested this tonight. My result was mixed: first run looked bad, second run looked normal. I expected the model choice to matter more than it did.

[vectorcrate98]

Good point. I don't think your conclusion is far off. At this point the real bottleneck is my patience.

[amberhall]

A couple things I'd check first. On 7900 XTX, 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.

[prompttinker44]

I tested this tonight. The issue went away once I changed one setting and restarted the stack cleanly. I expected the model choice to matter more than it did.

[prompttinker44]

Good point. I don't think your conclusion is far off. At this point the real bottleneck is my patience.

My first guess would be this: If you're on vLLM, I'd confirm the version first and then reduce variables until the behavior is obvious. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[dockridge]

A couple things I'd check first. Before rebuilding anything, I'd test one clean baseline with Kimi and compare from there. The logs were somehow calmer than I was.

[latentscope35]

I hit almost the same issue last night.

[oxidemixer]

Did a small comparison this afternoon. My result was mixed: first run looked bad, second run looked normal.

Good point. I don't think your conclusion is far off. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[prompttinker44]

A couple things I'd check first. If you're on vLLM, I'd confirm the version first and then reduce variables until the behavior is obvious. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.