Fetching the paper…

VALOR: Vision-Audio-Language Omni-Perception Pretraining Model and Dataset · Around