quant choices for 70B local

[paperrunner]

Appreciate the detail. I ended up in about the same place using Ollama. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

[vrambandit]

I hit almost the same issue last night.

[nanobrook]

My first guess would be this: On M3 Max, 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.

[fastbench]

Wanted to compare notes with people doing similar things.

I'm mostly working with FLUX in TensorRT-LLM on GB200. The part I'm trying to understand is workstation thermals.

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

This is mostly for local work rather than public serving.

[sparrowghost]

Appreciate the detail. The pattern around speculative decoding has been pretty similar for me. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

If you post your exact settings, people here can probably narrow it down pretty fast.

[cacheworks]

What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Mistral and compare from there. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[paperrunner]

subscribed, curious where this lands.

[fastbench]

Ran a quick check on my box. The issue went away once I changed one setting and restarted the stack cleanly. I expected the model choice to matter more than it did.

[edgecoast]

Need a little more detail here. Are you testing this on GB200 or something else?

[turboscope]

I'd approach it in a boring order. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at prefix caching. Not glamorous advice, but repeatable baselines save a lot of time.