desktop apps around local AI

[rocmriley]

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

Just sharing a setup that ended up working reasonably well: Qwen, vLLM, dual 3090, with a lot of attention on structured output.

Maybe useful as a starting point for newer folks. I keep promising myself this is the last rebuild.

[cachebench]

Ran a quick check on my box. My result was mixed: first run looked bad, second run looked normal. The main thing that moved was latency, not quality.

[vectorport]

What worked for me was pretty simple. If you're on Ollama, I'd confirm the version first and then reduce variables until the behavior is obvious. Once you know the plain path works, then add the faster path back in piece by piece.

[rasterbench]

What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at NVFP4. I keep promising myself this is the last rebuild.

[promptspark58]

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

[mintstation99]

What exact setup are you running? Is this local-only or are you routing anything externally?

[relayforge]

What exact setup are you running? Is this local-only or are you routing anything externally?

worth checking the driver first.

[rocmcrate26]

My first guess would be this: If you're on llama.cpp, I'd confirm the version first and then reduce variables until the behavior is obvious. Once you know the plain path works, then add the faster path back in piece by piece.

[mlxgarden]

A couple things I'd check first. Before rebuilding anything, I'd test one clean baseline with Mistral and compare from there. At this point the real bottleneck is my patience.

[rocmriley]

That matches what I was seeing. I ended up in about the same place using LM Studio. The biggest change for me was simplifying the stack instead of adding one more optimization layer.