TY  - RPRT
TI  - TDM-R1: Reinforcing Few-Step Diffusion Models with Non-Differentiable Reward
AU  - Yihong Luo
AU  - Tianyang Hu
AU  - Weijian Luo
AU  - Jing Tang
PY  - 2026
UR  - https://arxiv.org/abs/2603.07700
ID  - 2603.07700
ER  -