Saved in:
Bibliographic Details
Main Authors: Su, Duo, Hou, Junjie, Gao, Weizhi, Tian, Yingjie, Tang, Bowen
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.15138
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913439504400384
author Su, Duo
Hou, Junjie
Gao, Weizhi
Tian, Yingjie
Tang, Bowen
author_facet Su, Duo
Hou, Junjie
Gao, Weizhi
Tian, Yingjie
Tang, Bowen
contents Dataset distillation offers a lightweight synthetic dataset for fast network training with promising test accuracy. To imitate the performance of the original dataset, most approaches employ bi-level optimization and the distillation space relies on the matching architecture. Nevertheless, these approaches either suffer significant computational costs on large-scale datasets or experience performance decline on cross-architectures. We advocate for designing an economical dataset distillation framework that is independent of the matching architectures. With empirical observations, we argue that constraining the consistency of the real and synthetic image spaces will enhance the cross-architecture generalization. Motivated by this, we introduce Dataset Distillation via Disentangled Diffusion Model (D$^4$M), an efficient framework for dataset distillation. Compared to architecture-dependent methods, D$^4$M employs latent diffusion model to guarantee consistency and incorporates label information into category prototypes. The distilled datasets are versatile, eliminating the need for repeated generation of distinct datasets for various architectures. Through comprehensive experiments, D$^4$M demonstrates superior performance and robust generalization, surpassing the SOTA methods across most aspects.
format Preprint
id arxiv_https___arxiv_org_abs_2407_15138
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle D$^4$M: Dataset Distillation via Disentangled Diffusion Model
Su, Duo
Hou, Junjie
Gao, Weizhi
Tian, Yingjie
Tang, Bowen
Computer Vision and Pattern Recognition
Dataset distillation offers a lightweight synthetic dataset for fast network training with promising test accuracy. To imitate the performance of the original dataset, most approaches employ bi-level optimization and the distillation space relies on the matching architecture. Nevertheless, these approaches either suffer significant computational costs on large-scale datasets or experience performance decline on cross-architectures. We advocate for designing an economical dataset distillation framework that is independent of the matching architectures. With empirical observations, we argue that constraining the consistency of the real and synthetic image spaces will enhance the cross-architecture generalization. Motivated by this, we introduce Dataset Distillation via Disentangled Diffusion Model (D$^4$M), an efficient framework for dataset distillation. Compared to architecture-dependent methods, D$^4$M employs latent diffusion model to guarantee consistency and incorporates label information into category prototypes. The distilled datasets are versatile, eliminating the need for repeated generation of distinct datasets for various architectures. Through comprehensive experiments, D$^4$M demonstrates superior performance and robust generalization, surpassing the SOTA methods across most aspects.
title D$^4$M: Dataset Distillation via Disentangled Diffusion Model
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2407.15138