+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Prompting & Context Engineering (https://www.datasculpting.ai/inference/forumdisplay.php?fid=15)
+--- Forum: Long Context / Memory (https://www.datasculpting.ai/inference/forumdisplay.php?fid=64)
+--- Thread: long context that stayed useful (/showthread.php?tid=33)
Pages:
1
2
long context that stayed useful - cozywire - 08-25-2026
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 LM Studio and mini PC.
Would rather learn from your scars than collect my own all at once.
Re: long context that stayed useful - sanepeak48 - 08-25-2026
same here on my side.
Re: long context that stayed useful - sanescope - 08-26-2026
That matches what I was seeing. The pattern around tokens/sec has been pretty similar for me. The biggest change for me was simplifying the stack instead of adding one more optimization layer.
Re: long context that stayed useful - thetanest - 08-26-2026
What worked for me was pretty simple. Before rebuilding anything, I'd test one clean baseline with Llama and compare from there.
Re: long context that stayed useful - stackarch - 08-26-2026
That matches what I was seeing. I ended up in about the same place using PyTorch. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
Re: long context that stayed useful - plainember52 - 08-28-2026
Thanks, that helps. The pattern around router policy has been pretty similar for me. The biggest change for me was simplifying the stack instead of adding one more optimization layer.
Re: long context that stayed useful - bytesupply76 - 08-28-2026
That matches what I was seeing. That matches what I saw on B200. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.
Re: long context that stayed useful - audiopoint18 - 08-28-2026
Quote:That matches what I was seeing. That matches what I saw on B200. I got better results once I stopped chasing one metric and just looked at whether the
same here on my side.
Re: long context that stayed useful - fusedhall56 - 08-28-2026
following this because I'm in the same boat.
Re: long context that stayed useful - audiopoint18 - 08-29-2026
worth checking the driver first.