Stable Doodle is a new sketch-to-image tool from Stability AI, the creators of the powerful image synthesis model Stable Diffusion. With Stable Doodle, anyone can turn a simple drawing into a realistic and dynamic image without artistic skills or training.
Whether you want to create a building, a landscape, a portrait, or anything else, Stable Doodle can help you bring your imagination to life. Stable Doodle is designed for both professionals and hobbyists who want to explore the limitless possibilities of AI image generation.
Available for free on the Clipdrop by Stability AI website, users can access the user-friendly tool along with the latest Stable diffusion model SDXL 0.9.
With Stable Doodle, anyone with basic drawing skills and internet access can quickly generate high-quality original images. Sketches can be immediately implemented into designs for clients, presentation decks, websites, or logos.
Stable Doodle uses Stability AI’s Stable Diffusion XL, which makes images from text. It also uses T2I-Adapter, a feature from Tencent ARC (license) that lets users control how the images look.
T2I-Adapter works by adding some extra parts to the existing text-to-image models so that they can use different kinds of inputs, such as drawings, colors, or poses. This way, users can have more options to customize their images. For Stable Doodle, T2I-Adapter helps the text-to-image model (SDXL) to understand the shapes of sketches and make pictures based on words and drawings.
Some other recent breakthroughs from Stability AI include the AI image generation algorithm DeepFloyd and StableVicuna, a large-scale open-source chatbot trained with human feedback.
Read more:
Source: mPost
