Stable Diffusion-Powered Tool for Prompt-Based Inpainting

Have a look at this amazing new Stable Diffusion-powered tool, presented recently by an AI enthusiast known as ammrs via GitHub. Utilizing SD's Inpainting feature in combination with ClipSeg, the tool allows its users to change photos, replacing various details like hair, clothes, etc., and tweak the generated results using such parameters as Mask Precision and Stable Diffusion Generation Strength to your liking.

According to the creator, all you need to do is upload an image, input a prompt of the part in the input image that you want to replace, and type what you want to replace the existing part with to generate a new image. Please note that the tool is currently a notebook-based project but can be converted into a Gradio Web UI.

And here's an in-depth tutorial that will help you better understand how the tool works:

Click here to access the code and try it out. Also, don't forget to join our Reddit page and our Telegram channel, follow us on Instagram and Twitter, where we share breakdowns, the latest news, awesome artworks, and more.