[ACL 2020] PyTorch code for TVQA+: Spatio-Temporal Grounding for Video Question Answering - View it on GitHub