An official implementation for " UniVL: A Unified Video and Language Pre-Training Model for Multimodal Understanding and Generation" - View it on GitHub