TY  - RPRT
TI  - PE-OPSD: Internalizing Prompt Enhancement into Flow-matching Models via On-Policy Self-Distillation
AU  - Mingfeng Lin
AU  - Chengfei Cai
AU  - Lin Xu
AU  - Chengqian Ma
AU  - Yuxiang Wei
AU  - Liang Han
PY  - 2026
UR  - https://arxiv.org/abs/2609.36638
ID  - 2609.36638
ER  -