running local models on a phone for Phi

[fusedriver]

Posting this in case it saves someone else a little time.

If you were starting from scratch today for this use case, what would you choose first and what would you ignore?

I'm especially interested in mistakes worth avoiding.

[routeline23]

Appreciate the detail. That matches what I saw on M3 Max. The biggest change for me was simplifying the stack instead of adding one more optimization layer.

[cudacrate]

I'd approach it in a boring order. On 4090, 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.

[relaysupply90]

My first guess would be this: On mini PC, 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.

[sparrowworks88]

Need a little more detail here. What model, what runtime, and roughly what context size?

[rocmmixer]

Good point. I ended up in about the same place using TensorRT-LLM. The biggest change for me was simplifying the stack instead of adding one more optimization layer.

[amberrunner]

What worked for me was pretty simple. If you're on MLX, I'd confirm the version first and then reduce variables until the behavior is obvious.

Thanks, that helps. I ended up in about the same place using LM Studio. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[plainline]

Good point. 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.

[batchpeak]

subscribed, curious where this lands.

[llmpeak]

What worked for me was pretty simple. If you're on MLX, I'd confirm the version first and then reduce variables until the behavior is obvious.