Align-KD: Distilling Cross-Modal Alignment Knowledge for Mobile Vision-Language Model

Open in new window