Jiazi Bu

DiCache: Let Diffusion Model Determine Its Own Cache

Pref-GRPO: Pairwise Preference Reward-based GRPO for Stable Text-to-Image Reinforcement Learning

$\text{G}^2$RPO: Granular GRPO for Precise Reward in Flow Models

HiFlow: Training-free High-Resolution Image Generation with Flow-Aligned Guidance

MotionClone: Training-Free Motion Cloning for Controllable Video Generation

BroadWay: Boost Your Text-to-Video Generation Model in a Training-free Way