Skip to content

jiwoogit/SeaCache

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

[CVPR 2026] SeaCache: Spectral-Evolution-Aware Cache for Accelerating Diffusion Models

arXiv Home Page

Method Overview

SeaCache is a training-free diffusion cache schedule that bases reuse decisions on a spectrally aligned representations. By modeling spectral evolution along the denoising trajectory, we derive a Spectral-Evolution-Aware (SEA) filter that preserves content-relevant components while suppressing noise components.

Practical key points:

  • No additional parameters to tune. SeaCache introduces no additional hyperparameters (e.g., retention ratios or the coefficients). You just apply the scheduler-based SEA filtering and run inference.
  • Minimal overhead. SEA filtering (including FFT/iFFT) is inexpensive yet highly effective, accounting only ~0.2% to total inference latency.
  • Compatibility. SeaCache is compatible with orthogonal acceleration techniques, such as efficient attention and block-wise caching.

Figure 1: Conceptual illustration of SeaCache. SeaCache applies timestep-aligned, frequency-aware filtering to guide cache scheduling, achieving strong acceleration with negligible overhead.

Latest News

Supported Models

ComfyUI support is planned for a future update.

Text-to-Video

Image-to-Video

Text-to-Image

Community Contributions

  • More open-source model support will be added over time.
  • If you use SeaCache in your project and want it listed here, contact us at jiwoo.jg@gmail.com.

Acknowledgement

This repository is built on top of TeaCache, VBench, Wan2.1, Diffusers, HunyuanVideo, and FLUX. We thank the authors and contributors for their efforts.

Citation

If SeaCache is useful for your research or applications, please consider starring this repository and citing:

@article{chung2026seacache,
  title={SeaCache: Spectral-Evolution-Aware Cache for Accelerating Diffusion Models},
  author={Chung, Jiwoo and Hyun, Sangeek and Lee, MinKyu and Han, Byeongju and Cha, Geonho and Wee, Dongyoon and Hong, Youngjun and Heo, Jae-Pil},
  journal={arXiv preprint arXiv:2602.18993},
  year={2026}
}

About

[CVPR 2026] SeaCache: Spectral-Evolution-Aware Cache for Accelerating Diffusion Models

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages