Episodes Preview Franka Panda Visualizer
200 episodes · 50 fps

This dataset was created using LeRobot.

Dataset Description

Expert demonstrations of stack the red block on the yellow block, stack the red block on the other, put the red block on top of the yellow block, put the red block on the other, stack the yellow block on the red block, stack the yellow block on the other, put the yellow block on top of the red block, put the yellow block on the other in the mj_pick_block MuJoCo Franka Panda scene.

  • Source: dataset recipe stack_red_yellow (record_expert_dataset.py)

    • stack_red_on_yellow: 100 episodes
    • stack_yellow_on_red: 100 episodes
  • Control rate: 50 Hz

  • Frames: action.ee.delta_xyz is a gripper-site translation in the MuJoCo world frame (meters). action.ee.orientation is the first two rows of the gripper site's world-frame rotation matrix (site axes expressed in world).

  • Observations: observation.image.front, observation.image.wrist, observation.image.side (224×224 RGB, PNG bytes in the episode parquet files); observation.state (cartesian_position: x, y, z, gripper) is world-frame end-effector xyz (gripper site, meters) plus gripper opening in [0, 1] (0 = closed, 1 = open)

  • Scene init: repeated each frame; MuJoCo free-joint qpos [x, y, z, qw, qx, qy, qz] at episode start:

    • scene_init.block_red.pose: shape (7,), names ['x', 'y', 'z', 'qw', 'qx', 'qy', 'qz']
    • scene_init.block_yellow.pose: shape (7,), names ['x', 'y', 'z', 'qw', 'qx', 'qy', 'qz']
  • Gripper: target opening in [0, 1] (0 = closed, 1 = open)

  • Recorded actions:

    • action.ee.delta_xyz: shape (3,), names ['dx', 'dy', 'dz']
    • action.ee.orientation: shape (6,), names ['r00', 'r01', 'r02', 'r10', 'r11', 'r12']
    • action.joint.delta_angles: shape (7,), names ['joint1', 'joint2', 'joint3', 'joint4', 'joint5', 'joint6', 'joint7']
    • action.gripper.target: shape (1,), names ['gripper']
  • Homepage: [More Information Needed]

  • Paper: [More Information Needed]

  • License: apache-2.0

Dataset Structure

meta/info.json:

{
    "codebase_version": "v3.0",
    "fps": 50,
    "features": {
        "observation.image.front": {
            "dtype": "image",
            "shape": [
                3,
                224,
                224
            ],
            "names": [
                "C",
                "H",
                "W"
            ]
        },
        "observation.image.wrist": {
            "dtype": "image",
            "shape": [
                3,
                224,
                224
            ],
            "names": [
                "C",
                "H",
                "W"
            ]
        },
        "observation.image.side": {
            "dtype": "image",
            "shape": [
                3,
                224,
                224
            ],
            "names": [
                "C",
                "H",
                "W"
            ]
        },
        "observation.state": {
            "dtype": "float32",
            "shape": [
                4
            ],
            "names": [
                "x",
                "y",
                "z",
                "gripper"
            ]
        },
        "action.ee.delta_xyz": {
            "dtype": "float32",
            "shape": [
                3
            ],
            "names": [
                "dx",
                "dy",
                "dz"
            ]
        },
        "action.ee.orientation": {
            "dtype": "float32",
            "shape": [
                6
            ],
            "names": [
                "r00",
                "r01",
                "r02",
                "r10",
                "r11",
                "r12"
            ]
        },
        "action.joint.delta_angles": {
            "dtype": "float32",
            "shape": [
                7
            ],
            "names": [
                "joint1",
                "joint2",
                "joint3",
                "joint4",
                "joint5",
                "joint6",
                "joint7"
            ]
        },
        "action.gripper.target": {
            "dtype": "float32",
            "shape": [
                1
            ],
            "names": [
                "gripper"
            ]
        },
        "scene_init.block_red.pose": {
            "dtype": "float32",
            "shape": [
                7
            ],
            "names": [
                "x",
                "y",
                "z",
                "qw",
                "qx",
                "qy",
                "qz"
            ]
        },
        "scene_init.block_yellow.pose": {
            "dtype": "float32",
            "shape": [
                7
            ],
            "names": [
                "x",
                "y",
                "z",
                "qw",
                "qx",
                "qy",
                "qz"
            ]
        },
        "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": 200,
    "total_frames": 94985,
    "total_tasks": 8,
    "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": "panda",
    "splits": {
        "train": "0:200"
    }
}

Citation

BibTeX:

[More Information Needed]
Downloads last month
65

Models trained or fine-tuned on arkojit1/mj_stack_red_yellow_3cam