Stable Diffusion
Stable Diffusion is an AI-powered image generator that can create high-quality, photorealistic images from textual descriptions. Textual inputs, such as captions, keywords, or even full paragraphs of text are used as input. The results generated by Stable Diffusion are often impressively realistic, making it a useful tool for a variety of applications, from digital art and design, to video game and virtual reality content creation.
It was developed Robin Rombach (Stability AI) and Patrick Esser (Runway ML) from the CompVis Group at LMU Munich headed by Prof. Dr. Björn Ommer, who led the original Stable Diffusion V1 release.
Stability AI also have a separate product that is powered by Stable Diffusion called DreamStudio – you can think of it like a Photoshop for image generations. You buy credits and use their computational power, hence you don’t need a crazy powerful graphics card to run it.
Features
- Image generation from textual descriptions
Pros and Cons of Stable Diffusion
- Very universal, can be used to generate almost any type of imagery
- Free and Open source
- Can be run locally on your machine if you got a NVIDIA graphics card with 6GB+ VRAM
- You can train your own model with Hypernetwork
- Installation is not easy process
- Human faces look deformed (v 2.1)
- Human hands look deformed (v 2.1)
- You need a very powerful graphics card to run it locally
Related Tools
Related Blog Posts
Coming soon ...