Bitslingers-R-Us on Nostr: I'm curious about that, too. I'm running local models on CPU. I see tons of people ...

I'm curious about that, too. I'm running local models on CPU. I see tons of people spending hours and sometimes days trying to get everything to fit in to GPU memory instead of using a good CPU.

Sure, GPUs are faster, but what's wrong with starting a task and doing working on something else until the task you started is finished?

The idea that I'd just sit around and wait for each thing to happen in real time seems... well, a bit odd.