Project
DragNUWA
DragNUWA is a video generation model that utilizes text, images, and trajectory as three essential control factors to facilitate highly controllable video generation. DragNUWA is a video generation model that utilizes text, images, and trajectory…
Publication
LightGlue: Local Feature Matching at Light Speed
Publication