+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Evaluation & Testing (https://www.datasculpting.ai/inference/forumdisplay.php?fid=16)
+--- Forum: Evals (https://www.datasculpting.ai/inference/forumdisplay.php?fid=66)
+--- Thread: quality tests for local assistants for Mistral (/showthread.php?tid=1286)
Pages:
1
2
quality tests for local assistants for Mistral - audioforge - 10-15-2026
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 vLLM and 4090.
Simple and boring is a valid answer.
For context, I'm not chasing leaderboard numbers. I mostly want the system to feel consistent day to day.
Re: quality tests for local assistants for Mistral - ggufvault - 10-15-2026
Interesting, I hadn't tried that. The pattern around VRAM headroom has been pretty similar for me. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.
Re: quality tests for local assistants for Mistral - plaingarden18 - 10-16-2026
A couple things I'd check first. On M3 Max, the easiest win is often to back off the aggressive settings and see whether the problem survives.
Re: quality tests for local assistants for Mistral - bytesupply76 - 10-16-2026
That matches what I was seeing. I ended up in about the same place using PyTorch. The biggest change for me was simplifying the stack instead of adding one more optimization layer.
Re: quality tests for local assistants for Mistral - edgequill - 10-17-2026
Quote:That matches what I was seeing. I ended up in about the same place using PyTorch. The biggest change for me was simplifying the stack instead of addin
A couple things I'd check first. Before rebuilding anything, I'd test one clean baseline with Gemma and compare from there. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.
Re: quality tests for local assistants for Mistral - cacheworks - 10-17-2026
Appreciate the detail. I don't think your conclusion is far off. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
Re: quality tests for local assistants for Mistral - sparrowshed - 10-18-2026
Interesting, I hadn't tried that. That matches what I saw on M3 Max.
Re: quality tests for local assistants for Mistral - cacheline - 10-18-2026
What worked for me was pretty simple. If you're on vLLM, I'd confirm the version first and then reduce variables until the behavior is obvious. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.
Re: quality tests for local assistants for Mistral - tinycontext - 10-18-2026
Did a small comparison this afternoon. The issue went away once I changed one setting and restarted the stack cleanly. The main thing that moved was latency, not quality.
Re: quality tests for local assistants for Mistral - marbleyard32 - 10-18-2026
Good point. I ended up in about the same place using SGLang. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.