+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Announcements (https://www.datasculpting.ai/inference/forumdisplay.php?fid=5)
+--- Forum: Forum Announcements (https://www.datasculpting.ai/inference/forumdisplay.php?fid=6)
+--- Thread: Please keep off topic stuff in the Off Topic forum (/showthread.php?tid=3)
Pages:
1
2
Please keep off topic stuff in the Off Topic forum - Data Sculpting - 08-20-2026
Quick reminder: please keep off-topic conversations in the Off Topic forum.
That helps the technical sections stay useful for people searching for hardware, models, engines, and troubleshooting threads later.
Re: Please keep off topic stuff in the Off Topic forum - vramstream - 08-22-2026
Need a little more detail here. Is this local-only or are you routing anything externally?
Re: Please keep off topic stuff in the Off Topic forum - turboscope - 08-22-2026
Can you fill in a bit more context? Are you testing this on 5090 or something else?
Re: Please keep off topic stuff in the Off Topic forum - Data Sculpting - 08-22-2026
following this because I'm in the same boat.
Re: Please keep off topic stuff in the Off Topic forum - vramghost - 08-24-2026
Appreciate the detail. The pattern around prefill/decode split 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.
If you post your exact settings, people here can probably narrow it down pretty fast.
Re: Please keep off topic stuff in the Off Topic forum - Data Sculpting - 08-25-2026
What worked for me was pretty simple. On 7900 XTX, 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.
Re: Please keep off topic stuff in the Off Topic forum - deltapoint17 - 08-25-2026
Interesting, I hadn't tried that. I don't think your conclusion is far off.
Re: Please keep off topic stuff in the Off Topic forum - batchyard - 08-25-2026
What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Gemma and compare from there. Not glamorous advice, but repeatable baselines save a lot of time.
Re: Please keep off topic stuff in the Off Topic forum - Data Sculpting - 08-25-2026
Need a little more detail here. Do you see the same thing with a smaller batch or shorter prompt?
Re: Please keep off topic stuff in the Off Topic forum - Data Sculpting - 08-26-2026
Maybe, but I had the opposite result. I wouldn't optimize around that first. On my side the bigger swing came from the runtime and not the model. Not saying you're wrong, just that I wouldn't lock that in as the lesson yet.