Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2508.03442 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915782884065280 |
|---|---|
| author | Zhu, Shangwen Peng, Qianyu Shu, Zhilei Hu, Yuting Yang, Zhantao Zhang, Han Pu, Zhao Zheng, Andy Cui, Xinyu Zhao, Jian Feng, Ruili Cheng, Fan |
| author_facet | Zhu, Shangwen Peng, Qianyu Shu, Zhilei Hu, Yuting Yang, Zhantao Zhang, Han Pu, Zhao Zheng, Andy Cui, Xinyu Zhao, Jian Feng, Ruili Cheng, Fan |
| contents | High-fidelity text-to-image and text-to-video generation typically relies on Classifier-Free Guidance (CFG), but achieving optimal results often demands computationally expensive sampling schedules. In this work, we propose MAMBO-G, a training-free acceleration framework that significantly reduces computational cost by dynamically optimizing guidance magnitudes. We observe that standard CFG schedules are inefficient, applying disproportionately large updates in early steps that hinder convergence speed. MAMBO-G mitigates this by modulating the guidance scale based on the update-to-prediction magnitude ratio, effectively stabilizing the trajectory and enabling rapid convergence. This efficiency is particularly vital for resource-intensive tasks like video generation. Our method serves as a universal plug-and-play accelerator, achieving up to 3x speedup on Stable Diffusion v3.5 (SD3.5) and 4x on Lumina. Most notably, MAMBO-G accelerates the 14B-parameter Wan2.1 video model by 2x while preserving visual fidelity, offering a practical solution for efficient large-scale video synthesis. Our implementation follows a mainstream open-source diffusion framework and is plug-and-play with existing pipelines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_03442 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MAMBO-G: Magnitude-Aware Mitigation for Boosted Guidance Zhu, Shangwen Peng, Qianyu Shu, Zhilei Hu, Yuting Yang, Zhantao Zhang, Han Pu, Zhao Zheng, Andy Cui, Xinyu Zhao, Jian Feng, Ruili Cheng, Fan Computer Vision and Pattern Recognition High-fidelity text-to-image and text-to-video generation typically relies on Classifier-Free Guidance (CFG), but achieving optimal results often demands computationally expensive sampling schedules. In this work, we propose MAMBO-G, a training-free acceleration framework that significantly reduces computational cost by dynamically optimizing guidance magnitudes. We observe that standard CFG schedules are inefficient, applying disproportionately large updates in early steps that hinder convergence speed. MAMBO-G mitigates this by modulating the guidance scale based on the update-to-prediction magnitude ratio, effectively stabilizing the trajectory and enabling rapid convergence. This efficiency is particularly vital for resource-intensive tasks like video generation. Our method serves as a universal plug-and-play accelerator, achieving up to 3x speedup on Stable Diffusion v3.5 (SD3.5) and 4x on Lumina. Most notably, MAMBO-G accelerates the 14B-parameter Wan2.1 video model by 2x while preserving visual fidelity, offering a practical solution for efficient large-scale video synthesis. Our implementation follows a mainstream open-source diffusion framework and is plug-and-play with existing pipelines. |
| title | MAMBO-G: Magnitude-Aware Mitigation for Boosted Guidance |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2508.03442 |