model fallback rules that work: setup advice

[thetafield]

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

This isn't revolutionary, but the combination of Ollama and GB200 felt surprisingly usable once I tuned around router policy.

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

[quietcraft]

Did a small comparison this afternoon. 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.

[thetashed24]

Ran a quick check on my box. My result was mixed: first run looked bad, second run looked normal. The surprise was that memory pressure showed up before throughput collapsed.

[macrotinker]

That matches what I was seeing. I don't think your conclusion is far off. The biggest change for me was simplifying the stack instead of adding one more optimization layer.

[dockridge]

I think the bottleneck is somewhere else. I've seen people blame small local stacks when the real issue was much more mundane.

[bytestack]

What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Mixtral and compare from there.

[phasescope]

I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with Gemma and compare from there.

[dockkite69]

A couple things I'd check first. On mini PC, the easiest win is often to back off the aggressive settings and see whether the problem survives. Not glamorous advice, but repeatable baselines save a lot of time.

[oxideworks59]

That matches what I was seeing. I don't think your conclusion is far off. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[northgrid33]

I tested this tonight. I couldn't reproduce the slowdown until I pushed context much higher.