VATT: Transformers for Multimodal Self-Supervised Learning from Raw Video, Audio and Text

Open in new window