Improved Denoising Diffusion Probabilistic Models
Citations Over Time
Abstract
Denoising diffusion probabilistic models (DDPM) are a class of generative models which have recently been shown to produce excellent samples. We show that with a few simple modifications, DDPMs can also achieve competitive log-likelihoods while maintaining high sample quality. Additionally, we find that learning variances of the reverse diffusion process allows sampling with an order of magnitude fewer forward passes with a negligible difference in sample quality, which is important for the practical deployment of these models. We additionally use precision and recall to compare how well DDPMs and GANs cover the target distribution. Finally, we show that the sample quality and likelihood of these models scale smoothly with model capacity and training compute, making them easily scalable. We release our code at https://github.com/openai/improved-diffusion
Related Papers
- Score-Based Generative Modeling through Stochastic Differential Equations(2021)
- → Denoising Diffusion Probabilistic Models(2020)5,576 cited
- → GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium(2017)4,490 cited
- → DiffWave: A Versatile Diffusion Model for Audio Synthesis(2020)121 cited
- → Deep unsupervised learning using nonequilibrium thermodynamics(2024)62 cited