Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- 1. Install https://github.com/AUTOMATIC1111/stable-diffusion-webui (version >= 1.8.0)
- 2. Install https://github.com/continue-revolution/sd-webui-segment-anything into the extensions folder. This extension enables automatic image segmentation, and is needed to generate image masks in sd-webui-replacer.
- 3. Install a segmentation model (e.g. download https://huggingface.co/lkeab/hq-sam/resolve/main/sam_hq_vit_l.pth and put it inside extensions/sd-webui-segment-anything/models/sam)
- 4. Install https://github.com/light-and-ray/sd-webui-replacer into the extensions folder.
- [Launch the webui and make sure the replacer extension works]
- 5. Install https://github.com/Mikubill/sd-webui-controlnet into the extensions folder.
- 6. Install controlnet models (e.g. download all *.pth/*.yaml files from https://huggingface.co/lllyasviel/ControlNet-v1-1/tree/main and put them inside extensions/sd-webui-controlnet/models). Technically, the only model required for inpainting is control_v11p_sd15_inpaint.pth.
- 7. Install https://github.com/light-and-ray/sd-webui-animatediff- into the extensions folder.
- 8. Install motion models (e.g. download https://huggingface.co/guoyww/animatediff/resolve/main/v3_sd15_mm.ckpt and put it inside extensions/sd-webui-animatediff/model)
- You should see "AnimateDiff mode" under the Video tab in the replacer extension tab. Under models, make sure a control net inpainting model and motion module is auto-selected. You should use a non-inpainting model when using animatediff (I recommend epicphotogasm). You can also increase the context batch size to 16, if you have the VRAM for it. Otherwise, just hit run.
- INDEX OUT OF BOUNDS ERROR: Settings > Optimizations > check Pad prompt/negative prompt
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement