This is an main version of the model and can be used for txt2img and img2img. It is recommended to use the inpainting version for inpainting, as typically it should work better.
The orangemix VAE is recommended and is baked all versions of the model here (but not those on HuggingFace), so be sure to set SD VAE to automatic to use it (if you changed it), OR you may manually select the orangemix VAE or the VAE of your choice should you have it or others downloaded.
This is an main version of the model and can be used for txt2img and img2img. It is recommended to use the inpainting version for inpainting, as typically it should work better.
The orangemix VAE is recommended and is baked all versions of the model here (but not those on HuggingFace), so be sure to set SD VAE to automatic to use it (if you changed it), OR you may manually select the orangemix VAE or the VAE of your choice should you have it or others downloaded.
Additional notes - Using the SD MSE VAE (https://huggingface.co/stabilityai/sd-vae-ft-mse-original/tree/main) and the Obsidian Skin LoCon (LyCORIS), available at https://civitai.com/models/18584/obsidian-skin for the dark skinned preview image.