FineDiffusion: scaling up diffusion models for fine-grained image generation with 10,000 classes
Applied Intelligence · 2025-04 · 55 (5)
DOI: 10.1007/s10489-024-06215-1 ↗Kurz gefasst
Fachartikel in der Zeitschrift Applied Intelligence, online erschienen am 14.01.2025. Verfasst von Ziying Pan und 4 weiteren Autor:innen. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 49 Literaturangaben. Gefördert von: Natural Science Foundation of Fujian Province.
Abstract (Originalsprache)
The class-conditional image generation based on diffusion models is renowned for generating high-quality and diverse images. However, most prior efforts focus on generating images for general categories, e.g., 1000 classes in ImageNet-1k. A more challenging task, large-scale fine-grained image generation, remains the boundary to explore. In this work, we present a parameter-efficient strategy, called FineDiffusion, to fine-tune large pre-trained diffusion models scaling to large-scale fine-grained image generation with 10,000 categories. FineDiffusion significantly accelerates training and reduces storage overhead by only fine-tuning tiered class embedder, bias terms, and normalization layers’ parameters. To further improve the image generation quality of fine-grained categories, we propose a novel sampling method for fine-grained image generation, which utilizes superclass-conditioned guidance, specifically tailored for fine-grained categories, to replace the conventional classifier-free guidance sampling. Compared to full fine-tuning, FineDiffusion achieves a remarkable 1.56×\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\times $$\end{document} training speed-up and requires storing merely 1.77% of the total model parameters, while achieving state-of-the-art FID of 9.776 on image generation of 10,000 classes. Extensive qualitative and quantitative experiments demonstrate the superiority of our method compared to other parameter-efficient fine-tuning methods. The code and more generated results are available at our project website: https://finediffusion.github.io/.
Förderung
- Natural Science Foundation of Fujian Province
Zitieren
APA
Pan, Z., Wang, K., Li, G., He, F., & Lai, Y. (2025). FineDiffusion: scaling up diffusion models for fine-grained image generation with 10,000 classes. Applied Intelligence, 55(5). https://doi.org/10.1007/s10489-024-06215-1
Vancouver
Pan Z, Wang K, Li G, He F, Lai Y. FineDiffusion: scaling up diffusion models for fine-grained image generation with 10,000 classes. Applied Intelligence. 2025;55(5). doi:10.1007/s10489-024-06215-1
BibTeX
@article{Pan2025,
author = {Pan, Ziying and Wang, Kun and Li, Gang and He, Feihong and Lai, Yongxuan},
title = {FineDiffusion: scaling up diffusion models for fine-grained image generation with 10,000 classes},
journal = {Applied Intelligence},
year = {2025},
volume = {55},
number = {5},
doi = {10.1007/s10489-024-06215-1}
}
EUVIMED