This keeps coming up in conversation so I figured I'd start a thread.

The more I work with local setups, the more it seems like structured output decides everything downstream.

Curious whether that lines up with what other people are seeing.

For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.

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

Two paths both seem viable, but they break down differently once HBM constraints becomes important.

What would you pick if you had to live with it for a few months?

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

I'm not trying to build a perfect architecture diagram. I just want a sane starting point around PyTorch and 7900 XTX.

I'm especially interested in mistakes worth avoiding. The logs were somehow calmer than I was.

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

I've gone back and forth between llama.cpp and LM Studio for Mixtral. The interesting difference for me has been consumer GPUs.

Interested in tradeoffs more than fan loyalty.

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: FLUX, TensorRT-LLM, A6000, with a lot of attention on router policy.

Feel free to tear it apart if you see something dumb.

For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.

Maybe I'm missing something obvious.

I'm trying to decide whether to keep the current stack or change course. Main question is how people here are handling small local stacks.

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

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

I'm not trying to build a perfect architecture diagram. I just want a sane starting point around PyTorch and Threadripper box.

Simple and boring is a valid answer.

This is mostly for local work rather than public serving.

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: Phi, SGLang, B200, with a lot of attention on prefix caching.

Feel free to tear it apart if you see something dumb.

If I had to summarize the lesson so far, it would be that clean setup and observability beat clever guesses.

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.

Did a bit of testing and thought I'd post the result here.

I tried this with Gemma using PyTorch on Threadripper box. The headline result was better than I expected, but KV cache pressure still ended up being the limiting factor.

Nothing scientific here, just practical notes from a real setup.