[ICCV 2021] Official implementation of the paper "TRAR: Routing the Attention Spans in Transformers for Visual Question Answering"
visualization pytorch transformer attention official multi-modal clevr visual-question-answering vision-and-language dynamic-network multi-modality multi-modal-learning multi-scale-features vqav2 iccv2021 local-and-global
-
Updated
Oct 11, 2021 - Python