polarisai-robots/bento_ur7e_v3

This dataset was created using LeRobot.

Dataset Structure

meta/info.json:

{
    "codebase_version": "v3.0",
    "fps": 60,
    "features": {
        "action": {
            "dtype": "float32",
            "names": [
                "joint_0.pos",
                "joint_1.pos",
                "joint_2.pos",
                "joint_3.pos",
                "joint_4.pos",
                "joint_5.pos",
                "gripper.pos"
            ],
            "shape": [
                7
            ]
        },
        "observation.state": {
            "dtype": "float32",
            "names": [
                "joint_0.pos",
                "joint_1.pos",
                "joint_2.pos",
                "joint_3.pos",
                "joint_4.pos",
                "joint_5.pos",
                "gripper.pos"
            ],
            "shape": [
                7
            ]
        },
        "observation.images.top": {
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channels"
            ],
            "info": {
                "is_depth_map": false,
                "video.height": 480,
                "video.width": 640,
                "video.codec": "av1",
                "video.pix_fmt": "yuv420p",
                "video.fps": 60,
                "video.channels": 3,
                "has_audio": false,
                "video.g": 2,
                "video.crf": 30,
                "video.preset": 12,
                "video.fast_decode": 0,
                "video.video_backend": "pyav",
                "video.extra_options": {}
            }
        },
        "observation.images.wrist": {
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channels"
            ],
            "info": {
                "is_depth_map": false,
                "video.height": 480,
                "video.width": 640,
                "video.codec": "av1",
                "video.pix_fmt": "yuv420p",
                "video.fps": 60,
                "video.channels": 3,
                "has_audio": false,
                "video.g": 2,
                "video.crf": 30,
                "video.preset": 12,
                "video.fast_decode": 0,
                "video.video_backend": "pyav",
                "video.extra_options": {}
            }
        },
        "timestamp": {
            "dtype": "float32",
            "shape": [
                1
            ],
            "names": null
        },
        "frame_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "episode_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "task_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        }
    },
    "total_episodes": 449,
    "total_frames": 849750,
    "total_tasks": 2,
    "chunks_size": 1000,
    "data_files_size_in_mb": 100,
    "video_files_size_in_mb": 200,
    "data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
    "video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4",
    "robot_type": "ur7e",
    "splits": {
        "train": "0:449"
    }
}

Citation

BibTeX:

[More Information Needed]

v3 composition and collection notes

This dataset contains 449 episodes and 849,750 frames at 60 fps.

Source	Episodes in v3 (zero-based)	Count	Frames
bento_ur7e_v1	0-400	401	835,852
Human corrections collected with DAgger	401-448	48	13,898

The additional corrections address reaching for distant or overlapping fish, motions intended to avoid robot safety stops (including sausage manipulation), and occasional fish grasp failures. These are collection targets reported by the operator, not measured improvements from training.

All 48 retained correction episodes use the instruction: Pack me a lunchbox with two fish fillet and three sausages.

Original correction episodes 0-20 (retained v3 episodes 401-419) were originally collected under Pack me a lunchbox according to its color. with a white box and fish/sausage manipulation. Their labels were subsequently changed at the operator's request. The original rollout instruction and the final training label therefore differ for these 19 retained episodes.

The collection policy was pi0.5 v4; the operator used 30k and 40k checkpoints during collection. Exact checkpoint attribution for each correction episode is not available in this dataset.

The intervention indicator was removed from a training copy to match v1's feature schema. Source recordings were retained locally. Each correction remains a separate episode. No correction oversampling or dataset rebalancing was applied. Videos were copied without concatenation or re-encoding during the merge.

For review, episodes 401-448 are the new corrections; episodes 0-400 are the original v1 data. Dataset version v3 is distinct from the LeRobot format tag v3.0.

Original v3 episode indices 405 and 419 were removed at the operator's request after video/data alignment issues. Remaining episodes were renumbered. New 405 corresponds to old 406; new 418 corresponds to old 420. Total: 401 v1 episodes plus 48 DAgger corrections. Shared video files containing the removed episodes were filtered and re-encoded by the installed LeRobot delete_episodes operation.