error using stable diffusion 2

[omidreza-amrollahi]

Hi, I get an error as below when I try to use stable diffusion 2:

/usr/local/lib/python3.8/dist-packages/diffusers/pipeline_utils.py in from_pretrained(cls, pretrained_model_name_or_path, **kwargs)
672 elif len(missing_modules) > 0:
673 passed_modules = set(list(init_kwargs.keys()) + list(passed_class_obj.keys())) - optional_kwargs
--> 674 raise ValueError(
675 f"Pipeline {pipeline_class} expected {expected_modules}, but only {passed_modules} were passed."
676 )

ValueError: Pipeline <class 'stable_diffusion_videos.stable_diffusion_pipeline.StableDiffusionWalkPipeline'> expected {'safety_checker', 'vae', 'tokenizer', 'unet', 'text_encoder', 'feature_extractor', 'scheduler'}, but only {'vae', 'tokenizer', 'unet', 'text_encoder', 'scheduler'} were passed.

any ideas how it can be solved? Thanks

[nateraw]

I was able to load it like this (using latest version of this library, 0.7.1)

from stable_diffusion_videos import StableDiffusionWalkPipeline
from diffusers import EulerDiscreteScheduler
import torch

model_id = "stabilityai/stable-diffusion-2-base"

# Use the Euler scheduler here instead
scheduler = EulerDiscreteScheduler.from_pretrained(model_id, subfolder="scheduler")
pipe = StableDiffusionWalkPipeline.from_pretrained(
    model_id,
    scheduler=scheduler,
    feature_extractor=None,
    safety_checker=None,
    revision="fp16",
    torch_dtype=torch.float16,
).to("cuda")

[omidreza-amrollahi]

works! thank you very much. Is there also a way to create videos between two custom images? (rather than images generated by stable diffusion)

[nateraw]

Nope it only works for images generated from the model. You can read a detailed explanation of the logic here in the music videos post to understand why that is.