Vidi: Large Multimodal Models for Video Understanding and Editing Paper • 2504.15681 • Published Apr 22, 2025 • 14