Sampling-Aware Quantization for Diffusion Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zeng, Qian, Song, Jie, Wan, Yuanyu, Wang, Huiqiong, Song, Mingli
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915947956142080
author Zeng, Qian
Song, Jie
Wan, Yuanyu
Wang, Huiqiong
Song, Mingli
author_facet Zeng, Qian
Song, Jie
Wan, Yuanyu
Wang, Huiqiong
Song, Mingli
contents Diffusion models have recently emerged as the dominant approach in visual generation tasks. However, the lengthy denoising chains and the computationally intensive noise estimation networks hinder their applicability in low-latency and resource-limited environments. Previous research has endeavored to address these limitations in a decoupled manner, utilizing either advanced samplers or efficient model quantization techniques. In this study, we uncover that quantization-induced noise disrupts directional estimation at each sampling step, further distorting the precise directional estimations of higher-order samplers when solving the sampling equations through discretized numerical methods, thereby altering the optimal sampling trajectory. To attain dual acceleration with high fidelity, we propose a sampling-aware quantization strategy, wherein a Mixed-Order Trajectory Alignment technique is devised to impose a more stringent constraint on the error bounds at each sampling step, facilitating a more linear probability flow. Extensive experiments on sparse-step fast sampling across multiple datasets demonstrate that our approach preserves the rapid convergence characteristics of high-speed samplers while maintaining superior generation quality. Code is publicly available at: https://github.com/TaylorJocelyn/Sampling-aware-Quantization.
format Preprint
id arxiv_https___arxiv_org_abs_2505_02242
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Sampling-Aware Quantization for Diffusion Models
Zeng, Qian
Song, Jie
Wan, Yuanyu
Wang, Huiqiong
Song, Mingli
Computer Vision and Pattern Recognition
Diffusion models have recently emerged as the dominant approach in visual generation tasks. However, the lengthy denoising chains and the computationally intensive noise estimation networks hinder their applicability in low-latency and resource-limited environments. Previous research has endeavored to address these limitations in a decoupled manner, utilizing either advanced samplers or efficient model quantization techniques. In this study, we uncover that quantization-induced noise disrupts directional estimation at each sampling step, further distorting the precise directional estimations of higher-order samplers when solving the sampling equations through discretized numerical methods, thereby altering the optimal sampling trajectory. To attain dual acceleration with high fidelity, we propose a sampling-aware quantization strategy, wherein a Mixed-Order Trajectory Alignment technique is devised to impose a more stringent constraint on the error bounds at each sampling step, facilitating a more linear probability flow. Extensive experiments on sparse-step fast sampling across multiple datasets demonstrate that our approach preserves the rapid convergence characteristics of high-speed samplers while maintaining superior generation quality. Code is publicly available at: https://github.com/TaylorJocelyn/Sampling-aware-Quantization.
title Sampling-Aware Quantization for Diffusion Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2505.02242