clean API wrappers for local agents

[mellowstream]

Spent some time on this today and wanted to compare notes.

Two paths both seem viable, but they break down differently once tool calling becomes important.

Interested in tradeoffs more than fan loyalty.

The part that surprised me was how much the boring details mattered more than the model choice.

[batchwire]

Maybe, but I had the opposite result. I've seen people blame speculative decoding when the real issue was much more mundane.

[quietarch]

I'd approach it in a boring order. If you're on SGLang, I'd confirm the version first and then reduce variables until the behavior is obvious.

[routecraft]

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

[kernelmike]

subscribed, curious where this lands.

[signaltrack44]

That matches what I was seeing. I don't think your conclusion is far off. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[quantsupply]

A couple questions before I guess too hard. Do you see the same thing with a smaller batch or shorter prompt?

[quantnest]

A couple questions before I guess too hard. Do you see the same thing with a smaller batch or shorter prompt?

What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at consumer GPUs. Once you know the plain path works, then add the faster path back in piece by piece.

[batchwire]

That matches what I was seeing. That matches what I saw on A6000.

[sparrowkite53]

That matches what I was seeing. That matches what I saw on A6000.

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.