search

kxhit / zero123-hf

A diffuser implementation of Zero123. Zero-1-to-3: Zero-shot One Image to 3D Object (ICCV23)
https://zero123.cs.columbia.edu/
MIT License
135 stars 10 forks source link

convert_zero123_to_diffusers error #15

Closed YuzhiChen001 closed 5 months ago

YuzhiChen001 commented 5 months ago

After run diffusers/scripts/convert_zero123_to_diffusers.py , i got a dir like this. 1715287875682 But when i want to use it in diffusers like this, it is report an Error. pipe = StableDiffusionControlNetImg2ImgPipeline.from_pretrained( "./model/zero1to3", controlnet=controlnet, torch_dtype=torch.float16, local_files_only = True )

ValueError: Pipeline <class 'diffusers.pipelines.controlnet.pipeline_controlnet_img2img.StableDiffusionControlNetImg2ImgPipeline'> expected {'text_encoder', 'scheduler', 'feature_extractor', 'safety_checker', 'vae', 'controlnet', 'unet', 'tokenizer'}, but only {'scheduler', 'feature_extractor', 'controlnet', 'vae', 'unet'} were passed.

Any help would be greatly appreciated!

YuzhiChen001 commented 5 months ago

Can this be loaded only by pipeline_zero1to3?

kxhit commented 5 months ago

Hi, yes it should be loaded by pipeline_zero1to3

YuzhiChen001 commented 5 months ago

Thank you!