JensenYuan/DexTacWAM_unscrew_bottle_cap

Unscrew Bottle Cap

One of the six evaluation tasks in DexTacWAM. 96 episodes, 58,064 frames at 30 fps, on a Dexmate torso with two Sharpa Wave dexterous hands.

Unscrew the cap off the bottle and set it down.

Once the hand closes around the cap, its thread state is no longer visible, so the policy has to judge from contact when to regrasp.

Format

LeRobot v2.1. Parquet under data/chunk-000/, metadata under meta/. 49 GB.

feature	dtype	shape
head_img	image	(192, 256, 3)	PNG-encoded
left_wrist_img	image	(192, 256, 3)	PNG-encoded
right_wrist_img	image	(192, 256, 3)	PNG-encoded
tactile	uint8	(2, 5, 192, 256)	two hands x five fingers
state	float32	(90)
actions	float32	(150)
tactile_flow	float32	(2, 5, 24, 32, 4)	per-taxel flow, precomputed
deform	uint8	(2, 5, 192, 256)	deformation rendering

episode_provenance.json records, per output episode, which raw episode and frame range it came from and why it was split there.

Intended use

Stage 2 (world model) and stage 3 (action expert) training. The configs live under configs/bottle_cap/ in the code repository, and the normalization statistics committed there are computed against exactly this data. Statistics from a different conversion will de-normalize actions incorrectly without raising an error.

The configs address this corpus by directory name, so unpack it as data/datasets_lerobot/20260724_unscrew_bottle_cap_v2/ and keep the published episode ordering; the validation split is selected by episode index.

The released tactile encoder was trained on the separate 488 diverse episodes corpus, not on this data.

We do not release a trained action expert for this dataset. In our experiments, the stage 3 action expert is randomly initialized and trained from scratch using this data.

License

Apache 2.0, matching the DexTacWAM code. Parts of that repository are additionally CC BY-NC-SA 4.0 where they derive from Genie-Envisioner; that restriction applies to those source files, not to this data.

Citation

@article{dextacwam2026,
  title   = {DexTacWAM: A Visuo-Tactile World-Action Model for Dexterous Manipulation},
  author  = {Yuan, Haoran and Wang, Zekai and Shao, Boning and Lu, Haoran and
             Darrell, Trevor and Lourentzou, Ismini and Zhan, Wei},
  journal = {arXiv preprint arXiv:2609.24976},
  year    = {2026}
}