DiTCtrl: Exploring Attention Control in Multi-Modal Diffusion Transformer for Tuning-Free Multi-Prompt Longer Video Generation

SHIAE, CUHK (b) "Frosty pine: close-up shot medium shot forest vista, cinematic" (a) "Athlete glides across ocean waters snow mountain sand dunes" Figure 1. Our method DiTCtrl takes multiple text prompts as input and demonstrates superior capability in generating longer videos with multiple events, long-range coherence and smooth transitions as output.