👩💻Stable diffusion is a generative model that learns a probability distribution of data and can generate new instances.
🧠The model includes a forward process (adding noise) and a reverse process (removing noise) that can be trained using neural networks.
🔑Training the reverse process involves learning parameters to denoise images by predicting the amount of noise present.
🏞️Stable diffusion can be used for text-to-image generation, image-to-image manipulation, and inpainting.
💡To control the generative process, a prompt or conditioning signal can be introduced during the noise removal chain.
