Stable Diffusion is a deep learning-basedAI text-to-image generationmodel released in 2022. This model is mainly used to generate detailed images and illustrations conditioned on text descriptions, and was developed by the startup StabilityAI in collaboration with academic researchers and non-profit organizations.
Stable Diffusion is a Latent Diffusion Model—a type of deep generative neural network. Stable Diffusion's code and model weights have been publicly released and can run on most computer hardware equipped with a moderate GPU, whereas previous proprietary text-to-image models (such as DALL-E and Midjourney) could only be accessed through cloud computing services.
The online version of Stable Diffusion can mainly be accessed through the following official websites:
- DreamStudio version(officially launched by StabilityAI)
- Stable Diffusion XL (launched by ClipDrop under StabilityAI)
- Stable Artisan(the Discord version service launched by StabilityAI)
Stable Diffusion local deployment methods can be used through the following tools:
- Hui Shi Launcher(the desktop version of the Stable Diffusion-WebUI launcher created and freely released by Bilibili uploader @秋葉aaaki)
- Stable Diffusion integrated package(see How to Use Stable Diffusion – Method 2 for details)
