StreamFlow: Theory, Algorithm, and Implementation for High-Efficiency Rectified Flow Generation
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_ | 1866918277500895232 |
|---|---|
| author | Fang, Sen Zhong, Hongbin Feng, Yalin Zhang, Yanxin Metaxas, Dimitris N. |
| author_facet | Fang, Sen Zhong, Hongbin Feng, Yalin Zhang, Yanxin Metaxas, Dimitris N. |
| contents | New technologies such as Rectified Flow and Flow Matching have significantly improved the performance of generative models in the past two years, especially in terms of control accuracy, generation quality, and generation efficiency. However, due to some differences in its theory, design, and existing diffusion models, the existing acceleration methods cannot be directly applied to the Rectified Flow model. In this article, we have comprehensively implemented an overall acceleration pipeline from the aspects of theory, design, and reasoning strategies. This pipeline uses new methods such as batch processing with a new velocity field, vectorization of heterogeneous time-step batch processing, and dynamic TensorRT compilation for the new methods to comprehensively accelerate related models based on flow models. Currently, the existing public methods usually achieve an acceleration of 18%, while experiments have proved that our new method can accelerate the 512*512 image generation speed to up to 611%, which is far beyond the current non-generalized acceleration methods. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_22009 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | StreamFlow: Theory, Algorithm, and Implementation for High-Efficiency Rectified Flow Generation Fang, Sen Zhong, Hongbin Feng, Yalin Zhang, Yanxin Metaxas, Dimitris N. Computer Vision and Pattern Recognition New technologies such as Rectified Flow and Flow Matching have significantly improved the performance of generative models in the past two years, especially in terms of control accuracy, generation quality, and generation efficiency. However, due to some differences in its theory, design, and existing diffusion models, the existing acceleration methods cannot be directly applied to the Rectified Flow model. In this article, we have comprehensively implemented an overall acceleration pipeline from the aspects of theory, design, and reasoning strategies. This pipeline uses new methods such as batch processing with a new velocity field, vectorization of heterogeneous time-step batch processing, and dynamic TensorRT compilation for the new methods to comprehensively accelerate related models based on flow models. Currently, the existing public methods usually achieve an acceleration of 18%, while experiments have proved that our new method can accelerate the 512*512 image generation speed to up to 611%, which is far beyond the current non-generalized acceleration methods. |
| title | StreamFlow: Theory, Algorithm, and Implementation for High-Efficiency Rectified Flow Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2511.22009 |