We introduce TwinFlow, a framework that realizes high-quality 1-step and few-step generation without the pipeline bloat.
Instead of relying on external discriminators or frozen teachers, TwinFlow creates an internal "twin trajectory". By extending the time interval to $t\in[−1,1]$, we utilize the negative time branch to map noise to "fake" data, creating a self-adversarial signal directly within the model.
Then, the model can rectify itself by minimizing the difference of the velocity fields between real trajectory and fake trajectory, i.e. the $\Delta_\mathrm{v}$. The rectification performs distribution matching as velocity matching, which gradually transforms the model into a 1-step/few-step generator.
1@article{cheng2025twinflow,
2 title={TwinFlow: Realizing One-step Generation on Large Models with Self-adversarial Flows},
3 author={Cheng, Zhenglin and Sun, Peng and Li, Jianguo and Lin, Tao},
4 journal={arXiv preprint arXiv:2512.05150},
5 year={2025}
6}
