+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Evaluation & Testing (https://www.datasculpting.ai/inference/forumdisplay.php?fid=16)
+--- Forum: Benchmarks (https://www.datasculpting.ai/inference/forumdisplay.php?fid=65)
+--- Thread: comparing models without fooling yourself with PyTorch (/showthread.php?tid=1010)
Pages:
1
2
Re: comparing models without fooling yourself with PyTorch - kernelghost - 10-13-2026
What exact setup are you running? Do you see the same thing with a smaller batch or shorter prompt?
Re: comparing models without fooling yourself with PyTorch - cacheline - 10-16-2026
Quote:What exact setup are you running? Do you see the same thing with a smaller batch or shorter prompt?
A couple things I'd check first. 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.
Re: comparing models without fooling yourself with PyTorch - cudascope - 10-16-2026
Thanks, that helps. That matches what I saw on Threadripper box. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
Re: comparing models without fooling yourself with PyTorch - localtinker - 10-17-2026
Good point. I don't think your conclusion is far off. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.
Re: comparing models without fooling yourself with PyTorch - batchsize - 10-17-2026
I tested this tonight. My result was mixed: first run looked bad, second run looked normal. I expected the model choice to matter more than it did.
If you post your exact settings, people here can probably narrow it down pretty fast.
Re: comparing models without fooling yourself with PyTorch - amberhall - 10-17-2026
Quote:I tested this tonight. My result was mixed: first run looked bad, second run looked normal. I expected the model choice to matter more than it did.
What exact setup are you running? What model, what runtime, and roughly what context size?