AlphaFlow: Understanding and Improving MeanFlow Models
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_ | 1866914109564387328 |
|---|---|
| author | Zhang, Huijie Siarohin, Aliaksandr Menapace, Willi Vasilkovsky, Michael Tulyakov, Sergey Qu, Qing Skorokhodov, Ivan |
| author_facet | Zhang, Huijie Siarohin, Aliaksandr Menapace, Willi Vasilkovsky, Michael Tulyakov, Sergey Qu, Qing Skorokhodov, Ivan |
| contents | MeanFlow has recently emerged as a powerful framework for few-step generative modeling trained from scratch, but its success is not yet fully understood. In this work, we show that the MeanFlow objective naturally decomposes into two parts: trajectory flow matching and trajectory consistency. Through gradient analysis, we find that these terms are strongly negatively correlated, causing optimization conflict and slow convergence. Motivated by these insights, we introduce $α$-Flow, a broad family of objectives that unifies trajectory flow matching, Shortcut Model, and MeanFlow under one formulation. By adopting a curriculum strategy that smoothly anneals from trajectory flow matching to MeanFlow, $α$-Flow disentangles the conflicting objectives, and achieves better convergence. When trained from scratch on class-conditional ImageNet-1K 256x256 with vanilla DiT backbones, $α$-Flow consistently outperforms MeanFlow across scales and settings. Our largest $α$-Flow-XL/2+ model achieves new state-of-the-art results using vanilla DiT backbones, with FID scores of 2.58 (1-NFE) and 2.15 (2-NFE). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_20771 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | AlphaFlow: Understanding and Improving MeanFlow Models Zhang, Huijie Siarohin, Aliaksandr Menapace, Willi Vasilkovsky, Michael Tulyakov, Sergey Qu, Qing Skorokhodov, Ivan Computer Vision and Pattern Recognition Machine Learning MeanFlow has recently emerged as a powerful framework for few-step generative modeling trained from scratch, but its success is not yet fully understood. In this work, we show that the MeanFlow objective naturally decomposes into two parts: trajectory flow matching and trajectory consistency. Through gradient analysis, we find that these terms are strongly negatively correlated, causing optimization conflict and slow convergence. Motivated by these insights, we introduce $α$-Flow, a broad family of objectives that unifies trajectory flow matching, Shortcut Model, and MeanFlow under one formulation. By adopting a curriculum strategy that smoothly anneals from trajectory flow matching to MeanFlow, $α$-Flow disentangles the conflicting objectives, and achieves better convergence. When trained from scratch on class-conditional ImageNet-1K 256x256 with vanilla DiT backbones, $α$-Flow consistently outperforms MeanFlow across scales and settings. Our largest $α$-Flow-XL/2+ model achieves new state-of-the-art results using vanilla DiT backbones, with FID scores of 2.58 (1-NFE) and 2.15 (2-NFE). |
| title | AlphaFlow: Understanding and Improving MeanFlow Models |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2510.20771 |