adityx23/icl-dataset

ICL Dataset

Bimanual robot manipulation data collected by human teleoperation on the YOR robot, in LeRobot v2.1 format.

3,149 episodes · 2,385,871 frames · 22.1 hours · 36 tasks · 3 camera views

Each episode is a single teleoperated attempt at a manipulation task, recorded at 30 fps with synchronised proprioception, commanded actions and three camera streams.

---

⚠️ Read this before training on it

This dataset contains failures and unusable episodes, mixed in with the successes. They are not separated on disk — every episode lives in the same data/ and videos/ directories, distinguished only by the success field in meta/episodes.jsonl.

success	episodes	frames	meaning
success	2,715	2,069,799	task completed
fail	329	268,438	genuine attempt, goal not achieved
invalid	105	47,634	not a usable demonstration — the robot never moved, the scene was already in the goal state, or a human hand is resetting the scene

A naive load gives you all 3,149. To train only on successes:

import json

with open("meta/episodes.jsonl") as f:
    episodes = [json.loads(l) for l in f if l.strip()]

train = [e["episode_index"] for e in episodes if e["success"] == "success"]
# -> 2,715 episode indices

The invalid episodes should not be treated as failure demonstrations — in most of them nothing happens at all, so they are neither a positive nor a useful negative.

---

Structure

meta/
  info.json             LeRobot v2.1 header (features, shapes, path templates)
  episodes.jsonl        one row per episode: index, task, length, outcome, provenance
  episodes_stats.jsonl  per-episode feature statistics
  tasks.jsonl           36 task instructions, indexed
data/chunk-XXX/episode_NNNNNN.parquet                          tabular, one row per frame
videos/chunk-XXX/observation.images.<cam>/episode_NNNNNN.mp4   h264, 640x480, 30 fps

Chunking is episode_index // 100. Cameras are zed (scene), fish0 (left gripper), fish1 (right gripper).

Per-frame features

feature	shape	description
observation.state	[15]	measured proprioception: lj0..lj6 (left arm joints), rj0..rj6 (right arm joints), lift
action.left_ee	[7]	commanded left end-effector pose, [qw,qx,qy,qz,x,y,z] (quaternion first)
action.right_ee	[7]	commanded right end-effector pose, same convention
action.gripper	[2]	commanded [left, right] gripper, 0.0 = closed, 1.0 = open
action.base_vel	[3]	commanded base velocity [vx, vy, omega]
action.lift_cmd	[1]	commanded lift, -1 / 0 / 1 as float
observation.images.zed	480x640x3	scene view (ZED, left eye)
observation.images.fish0	480x640x3	left-hand gripper fisheye
observation.images.fish1	480x640x3	right-hand gripper fisheye

Plus the standard bookkeeping columns: timestamp, frame_index, episode_index, index, task_index.

observation.state is measured joint state; everything under action.* is the operator's commanded target. Gripper state is therefore in the action space.

---

Tasks

id	task (instruction)	episodes	minutes	success	fail	invalid
26	put the circle on the peg	322	73	279	14	29
15	pick up the orange cube and place it on the plate with the left arm on an uncluttered table	223	32	197	16	10
32	stack the colored octagons	196	166	140	50	6
4	open the gatorade bottle	131	44	99	31	1
31	sort the items into their containers	122	274	100	19	3
5	open the notebook	121	40	109	12	0
34	uncap the blue marker	117	36	103	14	0
33	stack the three cups into a tower	115	147	98	12	5
35	uncap the red marker	108	27	102	6	0
1	clean the plate	104	32	103	0	1
0	assemble the two gears	79	53	55	16	8
2	hit the yellow cube with the mallet using the left arm	76	26	63	10	3
25	put the bottle on the plate with the right arm	71	10	67	1	3
30	put the red chilli into the grocery bag	68	32	57	10	1
8	pass the salt shaker from the left arm to the right arm	67	34	49	16	2
29	put the green pepper into the grocery bag	66	28	55	9	2
17	pick up the orange cube and place it on the plate with the right arm on an uncluttered table	66	8	49	15	2
18	pick up the pumpkin and place it on the plate with the left arm	64	22	48	14	2
19	pick up the pumpkin and place it on the plate with the right arm	64	19	51	9	4
28	put the eggplant into the box with the right arm	63	7	57	0	6
23	place the red chilli in the cup with the left arm	61	17	50	10	1
12	pick up the eggplant and place it on the plate with the left arm	60	9	57	3	0
13	pick up the eggplant and place it on the plate with the right arm	60	9	57	3	0
21	pick up the red chilli and place it on the plate with the right arm	58	11	48	8	2
6	pass the red chilli from the left arm to the right arm	57	14	55	0	2
10	pass the yellow cube from the left arm to the right arm	57	17	52	4	1
20	pick up the red chilli and place it on the plate with the left arm	57	18	51	4	2
14	pick up the orange cube and place it on the plate with the left arm on a cluttered table	56	14	48	6	2
7	pass the red chilli from the right arm to the left arm	56	13	55	1	0
9	pass the salt shaker from the right arm to the left arm	56	23	52	4	0
11	pass the yellow cube from the right arm to the left arm	56	17	55	1	0
24	put the bottle on the plate	56	7	51	1	4
22	place the red chilli in the cup	55	12	49	6	0
27	put the eggplant into the box with the left arm	55	8	52	3	0
3	hit the yellow cube with the mallet using the right arm	53	9	52	1	0
16	pick up the orange cube and place it on the plate with the right arm on a cluttered table	53	16	50	0	3

Task labels are natural-language instructions: one instruction per task, one task_index per instruction, used consistently in meta/tasks.jsonl, meta/episodes.jsonl and every parquet.

Left/right arm variants, the two passing directions, and the cluttered vs uncluttered orange-cube conditions are deliberately separate tasks.

---

Collection

Recorded on a bimanual YOR robot with an NVIDIA Jetson Orin capturing the cameras. An operator teleoperated the robot via VR controllers; the Orin subscribed to the teleop state/action stream and synchronised it with the cameras at 30 Hz.

• Cameras: one StereoLabs ZED (scene view, left eye) and two Innomaker fisheyes mounted on the grippers, all 640×480 h264.
• Action convention: end-effector poses are quaternion-first [qw,qx,qy,qz,x,y,z]. When an arm is not under VR control its *_ee target falls back to its measured pose, so the action fields are always well-defined.

Provenance

Merged from 50 separate collection runs. Every episode records where it came from:

{"episode_index": 0, "tasks": ["assemble the two gears"], "length": 1292,
 "success": "fail",
 "source_dataset": "2_gear_assembly_20260725_164450", "source_episode": 0}

Episodes reviewed by a human additionally carry valid, reviewed_at and reviewed_by; those whose outcome was never adjudicated carry merged_from_unknown.

---

Label quality — please read

The outcome labels are not uniformly reliable, and this is the main limitation.

• Labels were recorded by the operator at collection time. Some are wrong — most commonly an episode marked success where the robot never moved at all, or where a human hand reset the scene.
• An automated screening pass flagged 134 episodes as suspected mislabels; a human reviewed all of them and corrected 77. Those carry reviewed_at.
• The remaining ~3,000 episodes were not individually audited. The screening pass had roughly 34% recall, so additional mislabelled episodes almost certainly remain.
• Some invalid episodes were explicitly judged unusable by a human; others had no recorded outcome at all and carry merged_from_unknown: true.

If label fidelity matters, filter to episodes carrying reviewed_at for the subset a human has actually confirmed.

---

Known limitations

• Failures and unusable episodes are interleaved with successes (see the warning above).
• Outcome labels are only partially audited.
• No train/validation split is defined — splits declares everything as train.
• Camera extrinsics/intrinsics are not published, and the two gripper fisheyes are physically identical models, so left/right assignment relies on the USB port each is plugged into rather than anything recoverable from the files.
• The three pass X from … to … task pairs differ only in word order, so a bag-of-words text encoder may not distinguish the two directions. Key on task_index if that matters.

---

Revision history

2026-08-05 — full rebuild. This dataset was rebuilt from source and re-published. It supersedes an earlier version that had different episode indices, different task indices and 44 slug-style task labels ([pnp_left_arm]). Three under-populated tasks were dropped in this rebuild — open the book and scribble in it (5 episodes, 0 successes), stack the six cups into a tower (7 episodes, 1 success) and unscrew the bottle and water the plant (1 episode) — removing 13 episodes.

Any copy downloaded before 2026-08-05 is incompatible: episode and task indices both changed, so a cached copy will not error, it will simply be mislabelled. Re-clone rather than pulling incrementally.

License

Not specified. No licence has been chosen for this data, which means downstream use is legally ambiguous. If you intend to use it, check with the authors first.