Block Flow: Learning Straight Flow on Data Blocks
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910790428131328 |
|---|---|
| author | Wang, Zibin Ouyang, Zhiyuan Zhang, Xiangyun |
| author_facet | Wang, Zibin Ouyang, Zhiyuan Zhang, Xiangyun |
| contents | Flow-matching models provide a powerful framework for various applications, offering efficient sampling and flexible probability path modeling. These models are characterized by flows with low curvature in learned generative trajectories, which results in reduced truncation error at each sampling step. To further reduce curvature, we propose block matching. This novel approach leverages label information to partition the data distribution into blocks and match them with a prior distribution parameterized using the same label information, thereby learning straighter flows. We demonstrate that the variance of the prior distribution can control the curvature upper bound of forward trajectories in flow-matching models. By designing flexible regularization strategies to adjust this variance, we achieve optimal generation performance, effectively balancing the trade-off between maintaining diversity in generated samples and minimizing numerical solver errors. Our results demonstrate competitive performance with models of the same parameter scale.Code is available at \url{https://github.com/wpp13749/block_flow}. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_11361 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Block Flow: Learning Straight Flow on Data Blocks Wang, Zibin Ouyang, Zhiyuan Zhang, Xiangyun Machine Learning Computer Vision and Pattern Recognition Flow-matching models provide a powerful framework for various applications, offering efficient sampling and flexible probability path modeling. These models are characterized by flows with low curvature in learned generative trajectories, which results in reduced truncation error at each sampling step. To further reduce curvature, we propose block matching. This novel approach leverages label information to partition the data distribution into blocks and match them with a prior distribution parameterized using the same label information, thereby learning straighter flows. We demonstrate that the variance of the prior distribution can control the curvature upper bound of forward trajectories in flow-matching models. By designing flexible regularization strategies to adjust this variance, we achieve optimal generation performance, effectively balancing the trade-off between maintaining diversity in generated samples and minimizing numerical solver errors. Our results demonstrate competitive performance with models of the same parameter scale.Code is available at \url{https://github.com/wpp13749/block_flow}. |
| title | Block Flow: Learning Straight Flow on Data Blocks |
| topic | Machine Learning Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2501.11361 |