Closed
Description
Is your feature request related to a problem? Please describe.
In auto1111 I can use img2img with inpainting in combination with controlnet canny edge detection to generate a new image.
With current tools in diffusers it appears I can't do that. I can only do text2img with controlnet
Describe the solution you'd like
A new pipeline StableDiffusionImg2ImgControlNetPipeline
or something similar which will allow img2img with inpainting and controlnet
Describe alternatives you've considered
Tried with auto1111 and works perfectly but can't use diffusers library for it