Spectrally-Guided Diffusion Noise Schedules

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Esteves, Carlos, Makadia, Ameesh
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909029087838208
author Esteves, Carlos
Makadia, Ameesh
author_facet Esteves, Carlos
Makadia, Ameesh
contents Denoising diffusion models are widely used for high-quality image and video generation. Their performance depends on noise schedules, which define the distribution of noise levels applied during training and the sequence of noise levels traversed during sampling. Noise schedules are typically handcrafted and require manual tuning across different resolutions. In this work, we propose a principled way to design per-instance noise schedules for pixel diffusion, based on the image's spectral properties. By deriving theoretical bounds on the efficacy of minimum and maximum noise levels, we design ``tight'' noise schedules that eliminate redundant steps. During inference, we propose to conditionally sample such noise schedules. Experiments show that our noise schedules improve generative quality of single-stage pixel diffusion models, particularly in the low-step regime.
format Preprint
id arxiv_https___arxiv_org_abs_2603_19222
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Spectrally-Guided Diffusion Noise Schedules
Esteves, Carlos
Makadia, Ameesh
Computer Vision and Pattern Recognition
Machine Learning
Denoising diffusion models are widely used for high-quality image and video generation. Their performance depends on noise schedules, which define the distribution of noise levels applied during training and the sequence of noise levels traversed during sampling. Noise schedules are typically handcrafted and require manual tuning across different resolutions. In this work, we propose a principled way to design per-instance noise schedules for pixel diffusion, based on the image's spectral properties. By deriving theoretical bounds on the efficacy of minimum and maximum noise levels, we design ``tight'' noise schedules that eliminate redundant steps. During inference, we propose to conditionally sample such noise schedules. Experiments show that our noise schedules improve generative quality of single-stage pixel diffusion models, particularly in the low-step regime.
title Spectrally-Guided Diffusion Noise Schedules
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2603.19222