Search for a command to run...
VAST: A Vision-Audio-Subtitle-Text Omni-Modality Foundation Model and Dataset