Inpaint Anything for Stable Diffusion Web UI

Inpaint Anything extension performs stable diffusion inpainting on a browser UI using any mask selected from the output of Segment Anything.

Using Segment Anything enables users to specify masks by simply pointing to the desired areas, instead of manually filling them in. This can increase the efficiency and accuracy of the mask creation process, leading to potentially higher-quality inpainting results while saving time and effort.

Standalone version

Installation

To install the software, please follow these steps:

• Open the Extensions tab on the AUTOMATIC1111's Stable Diffusion Web UI.
• Select the Install from URL option.
• Enter https://github.com/Uminosachi/sd-webui-inpaint-anything.git in the URL for extension's git repository field.
• Click on the Install button.
• Once installation is complete, restart the Web UI.
• Note: This extension supports v1.3.0 or higher of AUTOMATIC1111's Stable Diffusion Web UI.

Running the application

• If you intend to use the memory-efficient xformers, please append the --xformers argument to your startup command. For example, run ./webui.sh --xformers or webui.bat --xformers
• Note: If you have a privacy protection extension enabled in your web browser, such as DuckDuckGo, you may not be able to retrieve the mask from your sketch.
• Note: In Gradio version 3.23.0 or older, the segmentation image may appear small on the Web UI.

Downloading the Model

• Navigate to the Inpaint Anything tab in the Web UI.
• Click on the Download model button, located next to the Segment Anything Model ID. This includes the Segment Anything in High Quality Model ID, Fast Segment Anything, and Faster Segment Anything (MobileSAM).
  • Please note that the SAM is available in three sizes: Base, Large, and Huge. Remember, larger sizes consume more VRAM.
• Wait for the download to complete.
• The downloaded model file will be stored in the models directory of this application's repository.

Usage

• Drag and drop your image onto the input image area.
  • Outpainting can be achieved by the Padding options, configuring the scale and balance, and then clicking on the Run Padding button.
  • The Anime Style checkbox enhances segmentation mask detection, particularly in anime style images, at the expense of a slight reduction in mask quality.
• Click on the Run Segment Anything button.
• Use sketching to point the area you want to inpaint. You can undo and adjust the pen size.
  • Hover over either the SAM image or the mask image and press the S key for Fullscreen mode, or the R key to Reset zoom.
• Click on the Create mask button. The mask will appear in the selected mask image area.

Mask Adjustment

• Expand mask region button: Use this to slightly expand the area of the mask for broader coverage.
• Trim mask by sketch button: Clicking this will exclude the sketched area from the mask.
• Add mask by sketch button: Clicking this will add the sketched area to the mask.

Inpainting Tab

• Enter your desired Prompt and Negative Prompt, then choose the Inpainting Model ID.
• Click on the Run Inpainting button (Please note that it may take some time to download the model for the first time).
  • In the Advanced options, you can adjust the Sampler, Sampling Steps, Guidance Scale, and Seed.
  • If you enable the Mask area Only option, modifications will be confined to the designated mask area only.
• Adjust the iteration slider to perform inpainting multiple times with different seeds.
• The inpainting process is powered by diffusers.

Tips

• You can directly drag and drop the inpainted image into the input image field on the Web UI. (useful with Chrome and Edge browsers)
• To load prompts saved in a PNG file, follow these steps:
  • Drag and drop the image into the 'PNG Info' tab on the Web UI, then click Send to txt2img (or img2img).
  • Navigate to the 'Inpainting' section within the 'Inpaint Anything' tab and click the Get prompt from: txt2img (or img2img) button.

Model Cache

• The inpainting model, which is saved in HuggingFace's cache and includes inpaint (case-insensitive) in its repo_id, will also be added to the Inpainting Model ID dropdown list.
  • If there's a specific model you'd like to use, you can cache it in advance using the following Python commands (venv/bin/python for Linux and MacOS):

venv\Scripts\python.exe

from diffusers import StableDiffusionInpaintPipeline
pipe = StableDiffusionInpaintPipeline.from_pretrained("Uminosachi/dreamshaper_5-inpainting")
exit()

• The model diffusers downloaded is typically stored in your home directory. You can find it at /home/username/.cache/huggingface/hub for Linux and MacOS users, or at C:\Users\username\.cache\huggingface\hub for Windows users.

Cleaner Tab

• Choose the Cleaner Model ID.
• Click on the Run Cleaner button (Please note that it may take some time to download the model for the first time).
• Cleaner process is performed using Lama Cleaner.

Inpainting webui Tab

• This tab becomes accessible when you have an inpainting model.
• The required model should include inpaint (case-insensitive) in its filename and must be located in the stable-diffusion-webui/models directory.
• Once the model is recognized, it becomes selectable from the Inpainting Model ID dropdown list.
• The process can be executed swiftly, without requiring model loading, when the Stable Diffusion checkpoint (located in the upper left corner of the Web UI) matches the selected Inpainting Model ID.

ControlNet Inpaint Tab

• To execute inpainting, use the Stable Diffusion checkpoint located in the upper left of the Web UI, and pair it with the ControlNet inpaint model.
• Enter your desired Prompt and Negative Prompt.
• Click on the Run ControlNet Inpaint button to start the process.
  • In the Advanced options, you can adjust the Sampler, Sampling Steps, Guidance Scale, Denoising Strength, and Seed.
• The Control Weight and Control Mode can be modified in the ControlNet options.
  • The Reference-Only Control can be utilized if the Multi ControlNet setting is configured to 2 or higher.
  • The IP-Adapter can be utilized if the IP-Adapter model is present in the extensions/sd-webui-controlnet/models directory, and the ControlNet version is updated.
• Make sure to install the ControlNet extension that supports the inpaint_only preprocessor and the ControlNet inpaint model.
• Requires: The sd-webui-controlnet extension and the ControlNet-v1-1 inpaint model in the extensions/sd-webui-controlnet/models directory.

Mask only Tab

• Gives ability to just save mask without any other processing, so it's then possible to use the mask in img2img's Inpaint upload with any model/extensions/tools you already have in your AUTOMATIC1111.
• Get mask as alpha of image button: Save the mask as RGBA image, with the mask put into the alpha channel of the input image.
• Get mask button: Save the mask as RGB image.
• After the Get mask button press you can use Send to img2img inpaint button under the mask image to send both input image and mask to the img2img tab.

Auto-saving images

• The inpainted image will be automatically saved in the folder that matches the current date within the outputs/inpaint-anything directory.
• You can switch to the outputs/img2img-images directory via the Inpaint Anything section found in the Settings tab on the Web UI.

Development

With the Inpaint Anything library, you can perform segmentation and create masks using sketches from other extensions.

License

The source code is licensed under the Apache 2.0 license.

References

• Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A. C., Lo, W-Y., Dollár, P., & Girshick, R. (2023). Segment Anything. arXiv:2304.02643.
• Ke, L., Ye, M., Danelljan, M., Liu, Y., Tai, Y-W., Tang, C-K., & Yu, F. (2023). Segment Anything in High Quality. arXiv:2306.01567.
• Zhao, X., Ding, W., An, Y., Du, Y., Yu, T., Li, M., Tang, M., & Wang, J. (2023). Fast Segment Anything. arXiv:2306.12156 [cs.CV].
• Zhang, C., Han, D., Qiao, Y., Kim, J. U., Bae, S-H., Lee, S., & Hong, C. S. (2023). Faster Segment Anything: Towards Lightweight SAM for Mobile Applications. arXiv:2306.14289.