StreamFlow: Theory, Algorithm, and Implementation for High-Efficiency Rectified Flow Generation

Sen Fang, Hongbin Zhong, Yalin Feng, Yanxin Zhang, Dimitris N. Metaxas
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:29591-29600, 2026.

Abstract

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$\times$512 image generation speed to up to 611%, which is far beyond the current non-generalized acceleration methods. Project page at https://world-snapshot.github.io/StreamFlow/.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-fang26z, title = {{S}tream{F}low: Theory, Algorithm, and Implementation for High-Efficiency Rectified Flow Generation}, author = {Fang, Sen and Zhong, Hongbin and Feng, Yalin and Zhang, Yanxin and Metaxas, Dimitris N.}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {29591--29600}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/fang26z/fang26z.pdf}, url = {https://proceedings.mlr.press/v306/fang26z.html}, abstract = {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$\times$512 image generation speed to up to 611%, which is far beyond the current non-generalized acceleration methods. Project page at https://world-snapshot.github.io/StreamFlow/.} }
Endnote
%0 Conference Paper %T StreamFlow: Theory, Algorithm, and Implementation for High-Efficiency Rectified Flow Generation %A Sen Fang %A Hongbin Zhong %A Yalin Feng %A Yanxin Zhang %A Dimitris N. Metaxas %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-fang26z %I PMLR %P 29591--29600 %U https://proceedings.mlr.press/v306/fang26z.html %V 306 %X New technologies such as Rectified Flow and Flow Matching have significantly improved the performance of generative models in the past two years, especially in terms of control accuracy, generation quality, and generation efficiency. However, due to some differences in its theory, design, and existing diffusion models, the existing acceleration methods cannot be directly applied to the Rectified Flow model. In this article, we have comprehensively implemented an overall acceleration pipeline from the aspects of theory, design, and reasoning strategies. This pipeline uses new methods such as batch processing with a new velocity field, vectorization of heterogeneous time-step batch processing, and dynamic TensorRT compilation for the new methods to comprehensively accelerate related models based on flow models. Currently, the existing public methods usually achieve an acceleration of 18%, while experiments have proved that our new method can accelerate the 512$\times$512 image generation speed to up to 611%, which is far beyond the current non-generalized acceleration methods. Project page at https://world-snapshot.github.io/StreamFlow/.
APA
Fang, S., Zhong, H., Feng, Y., Zhang, Y. & Metaxas, D.N.. (2026). StreamFlow: Theory, Algorithm, and Implementation for High-Efficiency Rectified Flow Generation. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:29591-29600 Available from https://proceedings.mlr.press/v306/fang26z.html.

Related Material