QLoRA resource reality check with LM Studio

[quietcraft]

Has anyone else run into this?

My current setup is LM Studio with GB200, and I keep running into token cost.

If you've tested something similar, what did you keep and what did you stop bothering with?

My use case is pretty ordinary: local chat, some document work, and occasional batch jobs.

[cudascope]

What worked for me was pretty simple. On Threadripper box, the easiest win is often to back off the aggressive settings and see whether the problem survives. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[kernelarch]

Ran a quick check on my box. I couldn't reproduce the slowdown until I pushed context much higher.

[cacheline]

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. I keep promising myself this is the last rebuild.

[cudascope]

Maybe, but I had the opposite result. I get why that seems right, but on 7900 XTX the bottleneck felt elsewhere. Worth testing the simpler explanation before adopting the more interesting one.

[northgrid33]

Need a little more detail here. Is this local-only or are you routing anything externally?

[southtinker]

Need a little more detail here. Is this local-only or are you routing anything externally?

Thanks, that helps. I ended up in about the same place using PyTorch. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[sparrowworks88]

My first guess would be this: Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at tool calling.

[paperport64]

Thanks, that helps. The pattern around continuous batching has been pretty similar for me. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[benchplease]

Thanks, that helps. The pattern around continuous batching has been pretty similar for me. I got better results once I stopped chasing one metric and

I'd start with the simpler stack.