Fetching the paper…

VIOLET : End-to-End Video-Language Transformers with Masked Visual-token Modeling · Around