09-23-2026, 05:50 AM
Interesting, I hadn't tried that. That matches what I saw on mini PC.
learning in public
|
CUDA update broke my throughput
|
|
09-23-2026, 05:50 AM
Interesting, I hadn't tried that. That matches what I saw on mini PC.
learning in public
09-24-2026, 06:20 AM
Interesting, I hadn't tried that. I ended up in about the same place using SGLang.
09-24-2026, 06:55 AM
Quote:Interesting, I hadn't tried that. I ended up in about the same place using SGLang. My first guess would be this: Before rebuilding anything, I'd test one clean baseline with Qwen and compare from there. Not glamorous advice, but repeatable baselines save a lot of time. If you post your exact settings, people here can probably narrow it down pretty fast.
09-24-2026, 09:38 AM
What worked for me was pretty simple. On dual 3090, the easiest win is often to back off the aggressive settings and see whether the problem survives. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.
09-24-2026, 11:18 AM
Thanks, that helps. The pattern around quant drift has been pretty similar for me.
|
|
« Next Oldest | Next Newest »
|