+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Drivers / Frameworks / Toolkits (https://www.datasculpting.ai/inference/forumdisplay.php?fid=8)
+--- Forum: PyTorch / JAX / Other Frameworks (https://www.datasculpting.ai/inference/forumdisplay.php?fid=34)
+--- Thread: when a plain Python stack is enough (/showthread.php?tid=1447)
when a plain Python stack is enough - cudacrate - 10-18-2026
Posting this in case it saves someone else a little time.
I'm helping a small team set up a local workflow. We don't need a giant stack, but we do care about reranking.
I'm especially interested in mistakes worth avoiding.
I care more about stability than squeezing the last possible tokens/sec out of it.
Re: when a plain Python stack is enough - mangohouse - 10-18-2026
I'd approach it in a boring order. If you're on llama.cpp, I'd confirm the version first and then reduce variables until the behavior is obvious. Not glamorous advice, but repeatable baselines save a lot of time.
Re: when a plain Python stack is enough - localcraft17 - 10-18-2026
I think the bottleneck is somewhere else. I get why that seems right, but on 7900 XTX the bottleneck felt elsewhere.
Re: when a plain Python stack is enough - sanevault - 10-18-2026
Did a small comparison this afternoon. My result was mixed: first run looked bad, second run looked normal. I expected the model choice to matter more than it did.
Re: when a plain Python stack is enough - deltapoint17 - 10-18-2026
Need a little more detail here. Are you testing this on GB200 or something else?
Re: when a plain Python stack is enough - vramghost - 10-18-2026
Tried to reproduce it just now. The issue went away once I changed one setting and restarted the stack cleanly. The main thing that moved was latency, not quality.
I'm happy to test one more variant if you want a second data point.
Re: when a plain Python stack is enough - kernelrunner - 10-18-2026
Quote:Tried to reproduce it just now. The issue went away once I changed one setting and restarted the stack cleanly. The main thing that moved was latency,
My first guess would be this: If you're on PyTorch, I'd confirm the version first and then reduce variables until the behavior is obvious. Not glamorous advice, but repeatable baselines save a lot of time.