Dynamic Dual-Output Diffusion Models

TL;DR

This paper introduces a dynamic dual-output approach for diffusion models that alternates between predicting noise and images, significantly improving generation quality with fewer iterations.

Contribution

It proposes a novel method that dynamically switches between two denoising equations, enhancing image quality in diffusion models during low-iteration sampling.

Findings

Improves image quality in diffusion models with fewer iterations.

Applicable to various state-of-the-art architectures.

Achieves these improvements with minimal additional complexity.

Abstract

Iterative denoising-based generation, also known as denoising diffusion models, has recently been shown to be comparable in quality to other classes of generative models, and even surpass them. Including, in particular, Generative Adversarial Networks, which are currently the state of the art in many sub-tasks of image generation. However, a major drawback of this method is that it requires hundreds of iterations to produce a competitive result. Recent works have proposed solutions that allow for faster generation with fewer iterations, but the image quality gradually deteriorates with increasingly fewer iterations being applied during generation. In this paper, we reveal some of the causes that affect the generation quality of diffusion models, especially when sampling with few iterations, and come up with a simple, yet effective, solution to mitigate them. We consider two opposite equations for the iterative denoising, the first predicts the applied noise, and the second predicts the image directly. Our solution takes the two options and learns to dynamically alternate between them through the denoising process. Our proposed solution is general and can be applied to any existing diffusion model. As we show, when applied to various SOTA architectures, our solution immediately improves their generation quality, with negligible added complexity and parameters. We experiment on multiple datasets and configurations and run an extensive ablation study to support these findings.