How can I use my custom trained controlnet with diffusers img-to-img pipeline?
After I trained controlnet using the tutorial provided in the repository, I got the weights in .ckpt file. I wanted to use this controlnet in diffusers img-to-img pipeline which expects weights in .safetensors format and has different configuration file. How can I use my own trained controlnet with the img-to-img pipeline?
Use this script https://github.com/huggingface/diffusers/blob/main/scripts/convert_original_controlnet_to_diffusers.py
@kopyl thanks for the reply. It worked. Is there any similar script for converting base model on which controlnet is trained into diffusers format?
@Bilal143260 have no idea. Why do you need other scripts if it already worked?