A Comprehensive Review on Noise Control of Diffusion Model

TL;DR

This paper reviews various noise schedules in diffusion models, emphasizing their impact on image quality and training, and discusses their design features and performance implications.

Contribution

It provides a comprehensive analysis of different noise schedules in diffusion models, highlighting their effects on generative performance and training stability.

Findings

Different noise schedules significantly influence image quality.

Certain noise schedules improve training stability.

Performance varies notably across different noise schedule designs.

Abstract

Diffusion models have recently emerged as powerful generative frameworks for producing high-quality images. A pivotal component of these models is the noise schedule, which governs the rate of noise injection during the diffusion process. Since the noise schedule substantially influences sampling quality and training quality, understanding its design and implications is crucial. In this discussion, various noise schedules are examined, and their distinguishing features and performance characteristics are highlighted.