Room-to-Room (R2R) vision-and-language navigation - View it on GitHub