QLoRA resource reality check on A6000

[deltastation]

I'm trying to sanity-check my setup.

I'm trying to decide whether to keep the current stack or change course. Main question is how people here are handling small local stacks.

Would appreciate practical advice rather than theory.

[cacheline]

I tested this tonight. My result was mixed: first run looked bad, second run looked normal. The surprise was that memory pressure showed up before throughput collapsed.

[flintriver50]

Tried to reproduce it just now. I couldn't reproduce the slowdown until I pushed context much higher.

[cudascope]

What exact setup are you running? Are you testing this on M3 Max or something else?

[macropoint]

I tested this tonight. The issue went away once I changed one setting and restarted the stack cleanly.

[batchsize]

that fixed it for me too.

If you post your exact settings, people here can probably narrow it down pretty fast.

[kernelghost]

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

[cudascope]

Ran a quick check on my box. The issue went away once I changed one setting and restarted the stack cleanly. The main thing that moved was latency, not quality.

[cacheline]

Appreciate the detail. I ended up in about the same place using Ollama. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[circuitworks90]

Maybe, but I had the opposite result. I've seen people blame INT4 when the real issue was much more mundane. Worth testing the simpler explanation before adopting the more interesting one.