Room-across-Room (RxR) is a large-scale, multilingual dataset for Vision-and-Language Navigation (VLN) in Matterport3D environments. It contains 126k navigation instructions in English, Hindi and Telugu, and 126k navigation following demonstrations. Both annotation types include dense spatiotemporal alignments between the text and the visual perceptions of the annotators
Creative Commons Attribution 4.0 International
113
stars
12
forks
source link
How to load the dataset and visualize in habitat environment? #11
Hi. Thanks for this dataset. May I know how I can load the dataset and visualize it in habitat environment?