Open jiehwa opened 5 months ago
Yeah - I was just about to open an issue about this. It appears that PhotoMakerStableDiffusionXLPipeline
does not support a local safetensors
format. We would have to convert the model to diffusers I think. However, the diffusers StableDiffusionXLPipeline
does support the safetensors
format, so not sure what is going on here.
For example, this works:
pipe = StableDiffusionXLPipeline.from_single_file(
base_model_path, revision="fp16", variant="fp16", torch_dtype=torch.float16, scheduler="dpm"
)
pipe.scheduler = UniPCMultistepScheduler.from_config(pipe.scheduler.config)
pipe.to("cuda")
This does not:
pipe = PhotoMakerStableDiffusionXLPipeline.from_pretrained(
base_model_path, # can change to any base model based on SDXL
torch_dtype=torch.bfloat16,
use_safetensors=True,
variant="fp16"
).to("cuda")
and returns this error
ValueError: The provided pretrained_model_name_or_path "E:/github/stable-diffusion-webui/models/Stable-diffusion/sdxl/sd_xl_base_1.0_0.9vae.safetensors" is neither a valid local path nor a valid repo id. Please check the parameter.
The file exists.
Download corresponding safetensors
file to ./civitai_models