is there a fp32 model such that it can run on CPU
#188
by
chizh
- opened
is there a fp32 model such that it can run on CPU?
@chizh you can just manually move it to cpu and fp32 I believe.
pipe.to('cpu', dtype=torch.float32)
pipe = StableDiffusion3Pipeline.from_pretrained(
"stabilityai/stable-diffusion-3-medium-diffusers",
low_cpu_mem_usage=False,
text_encoder_3=None,
tokenizer_3=None,
torch_dtype=torch.float32,
use_safetensors=True
)