Code of paper: A Recurrent Vision-and-Language BERT for Navigation - View it on GitHub