Documentation about superresolution #9566
Description
Describe the bug
Reading this documentation which is not possible.
Indeed, super resolution pipeline works with an image as input in the image input parameter.
So it's not possible to use output_type="latent" for image_2.
Otherwise you'll get an error :
ValueError: Incorrect configuration settings! The config of pipeline.unet expects 7 but received num_channels_latents: 4 + num_channels_image: 4 = 8. Please verify the config of pipeline.unet or your image input.
Because it's placing a latent instead of an image.
Please also note that "variant="fp16" does not work for stabilityai/sd-x2-latent-upscaler" for me. The error message I get is :
ValueError: The deprecation tuple ('no variant default', '0.24.0', "You are trying to load the model files of the `variant=fp16`, but no such modeling files are available.The default model files: {'unet/diffusion_pytorch_model.bin', 'text_encoder/pytorch_model.bin', 'text_encoder/model.safetensors', 'vae/diffusion_pytorch_model.safetensors', 'vae/diffusion_pytorch_model.bin', 'unet/diffusion_pytorch_model.safetensors'} will be loaded instead. Make sure to not load from `variant=fp16`if such variant modeling files are not available. Doing so will lead to an error in v0.24.0 as defaulting to non-variantmodeling files is deprecated.") should be removed since diffusers' version 0.30.3 is >= 0.24.0
Reproduction
simple snippet of code
Logs
No response
System Info
ubuntu