One-step and few-step text-to-image generation must accurately render the details specified in the prompt despite limited opportunities for refinement. Distribution matching distillation (DMD) uses score differences to bring the generated distribution closer to a reference distribution. Following prior work, we estimat...
De-Yuan Liu, Yan-Sen Han, Kaiwen Zheng et al.· 1 citation
Preference optimization is a standard alignment method for generative models, yet extending it to continuous-time dynamics remains non-trivial. In flow matching, reward-driven updates modify transport trajectories without an inherent constraint to the pretrained data manifold and can move terminal samples off the pretr...
Yan-Sen Han, Shengyi Liao, Yuan-Xing Zhang et al.· 0 citations
This work introduces FlowCPO, an offline forward-KL objective that uses both preferred and dispreferred samples without online rollouts and shows under explicit regularity conditions that the forward-KL objective is bounded by a contrastive flow matching loss, yielding a tractable surrogate on fixed data.
Yan-Sen Han, Shengyi Liao, Peng Sun et al.· 1 citation
The decomposition provides a theoretical basis for adapting likelihood-based LLM methods to flow matching, while distinguishing exact substitutions from controlled surrogates, while distinguishing exact substitutions from controlled surrogates.
Yan-Sen Han, Hong-Xing Sun, Tao Lin· 1 citation
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.