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
Hi, I am using the RxR dataset for a project and need to use the time stamped poses for a downstream task.
I noticed that {instruction_id:06}_guide_pose_trace.npz contains snapshots of the guide trajectory but does not include continuous pose information (Please correct me if I am wrong here!).
On the other hand, the RxR_VLNCE_v0/{data_type}/{data_type}_guide_gt.json.gz file has the ground truth ‘locations’. However, this data is not time stamped but is aligned with the forward steps of the guide.
I was wondering if you have access to, and can possibly share, the ground truth time stamped continuous pose information for the dataset. I would really appreciate the help!
Hi, I am using the RxR dataset for a project and need to use the time stamped poses for a downstream task.
I noticed that
{instruction_id:06}_guide_pose_trace.npz
contains snapshots of the guide trajectory but does not include continuous pose information (Please correct me if I am wrong here!).On the other hand, the
RxR_VLNCE_v0/{data_type}/{data_type}_guide_gt.json.gz
file has the ground truth ‘locations’. However, this data is not time stamped but is aligned with the forward steps of the guide.I was wondering if you have access to, and can possibly share, the ground truth time stamped continuous pose information for the dataset. I would really appreciate the help!
Thanks so much for your time!