1 paper
Yunze Tong, Mushui Liu, Canyu Zhao +9
With large pretrained models, existing methods have effectively improved instruction-based video editing. However, most of them rely on an in-place editing assumption. They align t…