BOOT: Data-free Distillation of Denoising Diffusion Models with Bootstrapping

Vision running on diffusion

Denoising Diffusion Probabilistic Models

Diffusion models are a class of generative models that iteratively transform simple noise into complex data through a series of denoising steps. These models excel at producing high-quality images by modeling the reverse process of adding noise to data. Each step in the model gradually refines the data, ultimately generating realistic outputs. Use cases of diffusion models include image synthesis, restoration, and inpainting, enabling advancements in fields such as computer graphics, art creatio

Deep Generative Models