llama
llama copied to clipboard
how to put model and tensor on mac m1 gpu?
Is it already implemented in huggingface? something like this?
mps_device = torch.device("mps")
model.to(mps_device)
This would be a powerful addition.