Is StableDiffusion3Pipeline.from_single_file not supported in local process?
After download .safetensors file which is uploaded in 'Files and versions', I wrote the code like:
pipe = StableDiffusion3Pipeline.from_single_file(
"./sd3_medium_incl_clips.safetensors",
torch_dtype=torch.float16,
)
but connection fail logs appeared 'Cannot access gated repo for url https://huggingface.co/api/models/stabilityai/stable-diffusion-3-medium-diffusers/revision/main. Access to model stabilityai/stable-diffusion-3-medium-diffusers is restricted. You must be authenticated to access it.'
Isnt is possible to use this model in local environment? or Did I miss something?
@shawn-kang its possible but you have to sign in, to do that simple do
huggingface-cli login
@YaTharThShaRma999
Thank you for your anwer:)
But the reason I raised this comment is wanna know why I need to do sign in.
Because other versions like sd 1,2 ,xl don't need to do that when I used 'from_single_file' in pipline class..
@shawn-kang stability ai just released this with a more strict license. Sdxl and sd were more free to use.
as @YaTharThShaRma999 mentioned, this model has a more restrictive License and requires auth token even when reading local weights.
pipe = StableDiffusion3Pipeline.from_single_file(
"./sd3_medium_incl_clips.safetensors",
torch_dtype=torch.float16,
token = "your_token_here"
)