What is the difference between StableDiffusionPipeline and DiffusionPipeline

🧨 Diffusers
1

Hi guys, I would like to enquire that what is the difference between StableDiffusionPipeline and DiffusionPipeline. Can we loading stable diffusion weight using DiffusionPipeline, such as

pipeline = DiffusionPipeline.from_pretrained(
“runwayml/stable-diffusion-v1-5”
)

Thanks

2

DiffusionPipeline is a generic class that can be extended for specific tasks. StableDiffusionPipeline inherits and extends the DiffusionPipeline class specifically for text-2-image tasks.

You can call DiffusionPipeline.from_pretrained which will recognize and load a StableDiffusionPipeline based on the model that is passed to it. Or you can call StableDiffusionPipeline.from_pretrained directly also.

3

Thanks for your kind reply. So basically they will go to the same file.py in diffusers. The DiffusionPipeline finds the StableDiffusionPipeline class in the src dir automatically right?