running local models on a phone

[batchpeak]

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

I'm helping a small team set up a local workflow. We don't need a giant stack, but we do care about tokens/sec.

I'm especially interested in mistakes worth avoiding.

[vramstream]

I don't think that's the whole story. The reported numbers can look compelling, but in ordinary usage I didn't see the same benefit.

[cozysmith]

I don't think that's the whole story. The reported numbers can look compelling, but in ordinary usage I didn't see the same benefit.

Appreciate the detail. I don't think your conclusion is far off.

[macrotinker]

Not sure I agree there. I've seen people blame tool calling when the real issue was much more mundane. Worth testing the simpler explanation before adopting the more interesting one.

There are a few different conversations hiding inside this thread, so being explicit about the goal helps.

[northgrid33]

Good point. The pattern around router policy has been pretty similar for me.

[promptspark58]

Did a small comparison this afternoon. The issue went away once I changed one setting and restarted the stack cleanly. The surprise was that memory pressure showed up before throughput collapsed.

[turbosupply83]

Tried to reproduce it just now. I couldn't reproduce the slowdown until I pushed context much higher. I expected the model choice to matter more than it did.

[fastquill74]

Tried to reproduce it just now. I couldn't reproduce the slowdown until I pushed context much higher. I expected the model choice to matter more than

Did a small comparison this afternoon. My result was mixed: first run looked bad, second run looked normal. I expected the model choice to matter more than it did.

There are a few different conversations hiding inside this thread, so being explicit about the goal helps.

[marbleyard32]

Appreciate the detail. The pattern around workstation thermals has been pretty similar for me. The biggest change for me was simplifying the stack instead of adding one more optimization layer.

There are a few different conversations hiding inside this thread, so being explicit about the goal helps.

[promptspark58]

Not sure I agree there. I get why that seems right, but on 5090 the bottleneck felt elsewhere. Worth testing the simpler explanation before adopting the more interesting one.