FreqCa: Accelerating Diffusion Models via Frequency-Aware Caching
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912640220004352 |
|---|---|
| author | Liu, Jiacheng Cai, Peiliang Zhou, Qinming Lin, Yuqi Kong, Deyang Huang, Benhao Pan, Yupei Xu, Haowen Zou, Chang Tang, Junshu Zheng, Shikang Zhang, Linfeng |
| author_facet | Liu, Jiacheng Cai, Peiliang Zhou, Qinming Lin, Yuqi Kong, Deyang Huang, Benhao Pan, Yupei Xu, Haowen Zou, Chang Tang, Junshu Zheng, Shikang Zhang, Linfeng |
| contents | The application of diffusion transformers is suffering from their significant inference costs. Recently, feature caching has been proposed to solve this problem by reusing features from previous timesteps, thereby skipping computation in future timesteps. However, previous feature caching assumes that features in adjacent timesteps are similar or continuous, which does not always hold in all settings. To investigate this, this paper begins with an analysis from the frequency domain, which reveal that different frequency bands in the features of diffusion models exhibit different dynamics across timesteps. Concretely, low-frequency components, which decide the structure of images, exhibit higher similarity but poor continuity. In contrast, the high-frequency bands, which decode the details of images, show significant continuity but poor similarity. These interesting observations motivate us to propose Frequency-aware Caching (FreqCa)
which directly reuses features of low-frequency components based on their similarity, while using a second-order Hermite interpolator to predict the volatile high-frequency ones based on its continuity.
Besides, we further propose to cache Cumulative Residual Feature (CRF) instead of the features in all the layers, which reduces the memory footprint of feature caching by 99%.
Extensive experiments on FLUX.1-dev, FLUX.1-Kontext-dev, Qwen-Image, and Qwen-Image-Edit demonstrate its effectiveness in both generation and editing. Codes are available in the supplementary materials and will be released on GitHub. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_08669 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FreqCa: Accelerating Diffusion Models via Frequency-Aware Caching Liu, Jiacheng Cai, Peiliang Zhou, Qinming Lin, Yuqi Kong, Deyang Huang, Benhao Pan, Yupei Xu, Haowen Zou, Chang Tang, Junshu Zheng, Shikang Zhang, Linfeng Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition The application of diffusion transformers is suffering from their significant inference costs. Recently, feature caching has been proposed to solve this problem by reusing features from previous timesteps, thereby skipping computation in future timesteps. However, previous feature caching assumes that features in adjacent timesteps are similar or continuous, which does not always hold in all settings. To investigate this, this paper begins with an analysis from the frequency domain, which reveal that different frequency bands in the features of diffusion models exhibit different dynamics across timesteps. Concretely, low-frequency components, which decide the structure of images, exhibit higher similarity but poor continuity. In contrast, the high-frequency bands, which decode the details of images, show significant continuity but poor similarity. These interesting observations motivate us to propose Frequency-aware Caching (FreqCa) which directly reuses features of low-frequency components based on their similarity, while using a second-order Hermite interpolator to predict the volatile high-frequency ones based on its continuity. Besides, we further propose to cache Cumulative Residual Feature (CRF) instead of the features in all the layers, which reduces the memory footprint of feature caching by 99%. Extensive experiments on FLUX.1-dev, FLUX.1-Kontext-dev, Qwen-Image, and Qwen-Image-Edit demonstrate its effectiveness in both generation and editing. Codes are available in the supplementary materials and will be released on GitHub. |
| title | FreqCa: Accelerating Diffusion Models via Frequency-Aware Caching |
| topic | Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2510.08669 |