8000 GitHub - huang-yh/SpectralAR: [ICCV 25]SpectralAR: Spectral Autoregressive Visual Generation
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

huang-yh/SpectralAR

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 

Repository files navigation

SpectralAR: Spectral Autoregressive Visual Generation

SpectralAR: Spectral Autoregressive Visual Generation

Yuanhui Huang, Weiliang Chen, Wenzhao Zheng$\dagger$, Yueqi Duan, Jie Zhou, Jiwen Lu

$\dagger$ Project leader

We propose a Spectral AutoRegressive (SpectralAR) visual generation framework, which realizes causality for visual sequences from the spectral perspective.

teaser

News.

  • [2025/06/13] Paper released on arXiv.
  • [2025/06/13] Demo release.

Demo

demo

demo

Overview

Autoregressive visual generation has garnered increasing attention due to its scalability and compatibility with other modalities compared with diffusion models. Most existing methods construct visual sequences as spatial patches for autoregressive generation. However, image patches are inherently parallel, contradicting the causal nature of autoregressive modeling. To address this, we propose a Spectral AutoRegressive (SpectralAR) visual generation framework, which realizes causality for visual sequences from the spectral perspective. Specifically, we first transform an image into ordered spectral tokens with Nested Spectral Tokenization, representing lower to higher frequency components. We then perform autoregressive generation in a coarse-to-fine manner with the sequences of spectral tokens. By considering different levels of detail in images, our SpectralAR achieves both sequence causality and token efficiency without bells and whistles. We conduct extensive experiments on ImageNet-1K for image reconstruction and autoregressive generation, and SpectralAR achieves 3.02 gFID with only 64 tokens and 310M parameters.

overview

Performance

We train and evaluate our SpectralAR on the large scale ImageNet-1K dataset for both reconstruction and generation.

overview

Getting Started

We will release the code soon.

Citation

If you find this project helpful, please consider citing the following paper:

@misc{huang2025spectralarspectralautoregressivevisual,
      title={SpectralAR: Spectral Autoregressive Visual Generation}, 
      author={Yuanhui Huang and Weiliang Chen and Wenzhao Zheng and Yueqi Duan and Jie Zhou and Jiwen Lu},
      year={2025},
      eprint={2506.10962},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2506.10962}, 
}

About

[ICCV 25]SpectralAR: Spectral Autoregressive Visual Generation

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published
0