Datasets:

nvidia
/

video_to_data_challenge

Modalities:

Size:

Tags:

human-object-interaction

Libraries:

License:

Dataset card Data Studio Files Files and versions

Replace Track 1 with 30 selected single-view sequences

#5

by dzou-nv - opened 8 days ago

base: refs/heads/main

←

from: refs/pr/5

Discussion Files changed

initial commita5304827

Add track_3: LeRobot v2.1 dataset (46 episodes, 276 videos, 11 object meshes)bff4258a

Add track_3: LeRobot v2.1 dataset (46 episodes, 276 videos, 11 object meshes) (part 2)d47a178a

Add track_3 READMEb11ef798

Update track_3 READMEf3c2ad6d

Add Track 2 whole-body reconstruction dataset (#2)f0a3e9ac

Updating assets0bcba466

adding URDF7d7db06e

dropping episode_000029a1488fb2

Split Track 3 into public and confidential holdout setsf5479f8f

Move episode 42 to Track 3 public split4af816dc

Add track_1: single-view LeRobot v2.1 dataset (67 episodes, 48846 frames, 33 object meshes, SOMA-X + MHR human params)a2854e73

Remove Track 3 exocentric videos553afe1b

updated track3 datacard8b8ef738

Remove Track 1 ground-truth annotations (#3)5ae9b066

Remove confidential Track 3 holdout from public releasee531d6f7

License Track 3 dataset and assets under CC BY-4.00581bb68

Update README.mdf6d39876

Organize Track 2 trajectory variants (#4)ba499c5e

Update track_3/v2d-challenge-track3.md1516b4b2

NVIDIA org 8 days ago

Retain 12 selected sequences, remove 25, and add 18 September camera videos. Rebuild LeRobot v2.1 indexes, task mappings, metadata, statistics, and documentation in selection order. Preserve source video bytes and the existing video-only metadata schema. Tracks 2 and 3 are unchanged.

Replace Track 1 with 30 selected single-view sequences97510bc2

zhuxh-nvidia changed pull request status to merged 8 days ago

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment