replicate/latent-consistency-model

is this running with cpu? just tested for the first time in my M2 8Gb ram

Flizyx opened this issue · 1 comments

Flizyx commented
image ran the commands in https://replicate.com/blog/run-latent-consistency-model-on-mac

It is running on the GPU, check for this line:

model.to(torch_device="cpu", torch_dtype=torch.float32).to('mps:0')

If I'm correct, that is telling the machine to use MPS (Metal Performance Shaders)

Those three spikes in my screenshot are my tests:

Screenshot 2023-10-27 at 5 08 25 PM