Stable Diffusion
ℹ️
This model is very large, please specify its name at startup so that the model can be fully downloaded on the first launch. This way, there will be no issues when switching on the website in the future.
--model sd1.5: https://github.com/runwayml/stable-diffusion (opens in a new tab)--model sd2: https://github.com/Stability-AI/stablediffusion (opens in a new tab)--model anything4: huggingface.co/andite/anything-v4.0--model realisticVision1.4: huggingface.co/SG161222/Realistic_Vision_V1.4
You can use Stable Diffusion to replace objects in the picture base on a text prompt.
Erase object
Stable Diffusion can also be used for "normal" inpainting(erase object from picture), here is some example using background as prompt:
Load local ckpt/safetensors
You can learn how to create a inpainting model from any diffusion model in AUTO1111's webui here (opens in a new tab).
Then you can load it into Lama Cleaner with following command:
--model sd1.5 --sd-local-model-path /path/to/model-inpaint.ckpt