llama icon indicating copy to clipboard operation
llama copied to clipboard

how to put model and tensor on mac m1 gpu?

Open srulik-ben-david opened this issue 1 year ago • 1 comments

Is it already implemented in huggingface? something like this?

mps_device = torch.device("mps")
model.to(mps_device)

srulik-ben-david avatar Mar 14 '23 07:03 srulik-ben-david

This would be a powerful addition.

okpatil4u avatar Mar 30 '23 06:03 okpatil4u