FlowConsist: Make Your Flow Consistent with Real Trajectory
Tianyi Zhang, Chengcheng Liu, Jinwei Chen, Chun-Le Guo, Chongyi Li, Ming-Ming Cheng, Bo Li, Peng-Tao Jiang · 2026
Read FlowConsist: Make Your Flow Consistent with Real Trajectory by Tianyi Zhang, Chengcheng Liu, Jinwei Chen, Chun-Le Guo, Chongyi Li, Ming-Ming Cheng, Bo Li, Peng-Tao Jiang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-06346.txt.
Description
Flow models make pictures by following a learned path from noise to image, and fast versions try to jump the path in one or few steps. Tianyi Zhang and 7 co-authors find two flaws in how those fast models are trained. Random pairings of noise and data bend the target path, and small errors pile up over long jumps.
Their framework, FlowConsist, trains the model on velocities from its own marginal path instead, keeping it on a consistent trajectory. A rectification step lines up the distributions of made and real samples at every point along the path. On a standard image benchmark, the method set a new quality record for one-step generation.
Catalog record
| CREATOR | Tianyi Zhang, Chengcheng Liu, Jinwei Chen, Chun-Le Guo, Chongyi Li, Ming-Ming Cheng, Bo Li, Peng-Tao Jiang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 55 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-06346