+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: LLM Models (https://www.datasculpting.ai/inference/forumdisplay.php?fid=12)
+--- Forum: Open Weight (https://www.datasculpting.ai/inference/forumdisplay.php?fid=49)
+---- Forum: Recipes & Configurations (https://www.datasculpting.ai/inference/forumdisplay.php?fid=52)
+---- Thread: simple router policy ideas with PyTorch (/showthread.php?tid=911)
Pages:
1
2
Re: simple router policy ideas with PyTorch - amberkite - 10-13-2026
My first guess would be this: Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at tokens/sec. Once you know the plain path works, then add the faster path back in piece by piece.
Re: simple router policy ideas with PyTorch - thetagrid15 - 10-14-2026
I don't think that's the whole story. I get why that seems right, but on M3 Max the bottleneck felt elsewhere. Worth testing the simpler explanation before adopting the more interesting one.
Re: simple router policy ideas with PyTorch - batchsize - 10-14-2026
I'd start with the simpler stack.
I'm happy to test one more variant if you want a second data point.
Re: simple router policy ideas with PyTorch - kernelghost - 10-15-2026
Quote:I'd start with the simpler stack.
following this because I'm in the same boat.
Re: simple router policy ideas with PyTorch - cobaltworks25 - 10-15-2026
Good point. I don't think your conclusion is far off. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.