dagger / README.md
Haoze1235's picture
Publish verified DAgger6 dataset
38ce075 verified
|
Raw
History Blame Contribute Delete
2.18 kB
metadata
license: other
task_categories:
  - robotics
  - other
language:
  - en
tags:
  - embodied-ai
  - tracking
  - dagger
  - habitat
pretty_name: OpenTrackVLA DAgger Recovery Dataset

OpenTrackVLA DAgger Recovery Dataset

Training-only DAgger recovery data generated in Habitat for OpenTrackVLA target tracking. This release contains V6-policy failure recovery data for AT, DT, and STT, using seeds 101 and 102.

The uploaded payload contains only the final training-ready datasets:

  • JSONL trajectory labels;
  • referenced RGB frames;
  • precomputed DINOv3 + SigLIP fine/coarse visual-token caches;
  • conversion, filtering, and integrity metadata.

Source rollouts, raw recovery videos, evaluation-scene data, model checkpoints, and EVT-Bench results are not included.

Layout

packed/v6_failures/train/<task>/seed_<seed>/
  jsonl.tar.zst.part-0000...
  frames.tar.zst.part-0000...
  vision_cache.tar.zst.part-0000...
metadata/v6_failures/train/<task>/seed_<seed>/
  dataset_stats.json
  dataset_inspection.json
  dagger_selection_plan.json
  referenced_frames.txt
  *.manifest.json

Each component is a Zstandard-compressed tar stream split into numbered parts. Reconstruct one component with:

cat <component>.tar.zst.part-* | tar --zstd -xf - -C /path/to/dataset

Every component manifest records the ordered part names, byte sizes, and SHA256 checksums. Verify all parts before extraction.

Data policy

This dataset is derived from simulated Habitat/HM3D tracking episodes. Users are responsible for complying with the licenses and terms of the upstream scene, avatar, simulator, and model assets. The other license marker is intentional: it does not replace those upstream terms.

Generation

  • Split: training scenes and training avatars only
  • Tasks: AT, DT, STT
  • Seeds: 101, 102
  • History: 31 frames
  • Prediction horizon: 8 waypoints
  • DAgger target fraction: 0.25
  • Failure window: 12 steps
  • Intervention pre/post windows: 12/8 steps
  • Maximum retained samples per episode: 26

Deterministic Habitat native-process crashes are explicitly audited and excluded; they are never converted into synthetic success/failure labels.