Datasets:

io-intelligence
/

arx_pick_block

license: apache-2.0
task_categories:
  - robotics
tags:
  - LeRobot
  - manipulation
  - dual-arm
  - pick-and-place
configs:
  - config_name: default
    data_files: data/*/*.parquet
language:
  - en

ARX Pick Block Dataset

This dataset was created using LeRobot and contains dual-arm robotic manipulation demonstrations for pick-and-place tasks.

Dataset Description

Homepage: https://io-ai.tech
License: Apache-2.0

Task Overview

This dataset contains demonstrations of a dual-arm robotic manipulation task where the robot picks up a block from a desk and places it into a plate.

Task Details

Task Objective: Pick up the block from the desk and place into plate
Operational Objects: Block
Operation Duration: Each operation takes approximately 8 to 15 seconds
Recording Frequency: 15 Hz
Robot Type: 7-DOF dual-arm ARX R5A
End Effector: Gripper
Dual-Arm Operation: Yes
Image Resolution: 640x480
Camera Positions: High; Low
Data Content: • Robot's current state (8 joint positions) • Robot's next action (8 joint commands) • Current camera view images (dual camera setup) • Task success score

Dataset Structure

meta/info.json:

{
    "codebase_version": "v3.0",
    "robot_type": "Arx_R5A",
    "total_episodes": 162,
    "total_frames": 35733,
    "total_tasks": 1,
    "chunks_size": 1000,
    "data_files_size_in_mb": 100,
    "video_files_size_in_mb": 500,
    "fps": 15,
    "splits": {
        "train": "0:162"
    },
    "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",
    "features": {
        "observation.images.camera_high": {
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channels"
            ],
            "info": {
                "video.height": 480,
                "video.width": 640,
                "video.codec": "av1",
                "video.pix_fmt": "yuv420p",
                "video.is_depth_map": false,
                "video.fps": 15,
                "video.channels": 3,
                "has_audio": false
            }
        },
        "observation.images.camera_low": {
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channels"
            ],
            "info": {
                "video.height": 480,
                "video.width": 640,
                "video.codec": "av1",
                "video.pix_fmt": "yuv420p",
                "video.is_depth_map": false,
                "video.fps": 15,
                "video.channels": 3,
                "has_audio": false
            }
        },
        "observation.state": {
            "dtype": "float64",
            "shape": [
                8
            ],
            "names": [
                "joint1",
                "joint2",
                "joint3",
                "joint4",
                "joint5",
                "joint6",
                "joint7",
                "joint8"
            ]
        },
        "action": {
            "dtype": "float64",
            "shape": [
                8
            ],
            "names": [
                "joint1",
                "joint2",
                "joint3",
                "joint4",
                "joint5",
                "joint6",
                "joint7",
                "joint8"
            ]
        },
        "observation.score": {
            "dtype": "float32",
            "shape": [
                1
            ],
            "names": [
                "score"
            ]
        },
        "timestamp": {
            "dtype": "float32",
            "shape": [
                1
            ],
            "names": []
        },
        "frame_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": []
        },
        "episode_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": []
        },
        "index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": []
        },
        "task_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": []
        }
    }
}

Data Fields

Field	Type	Shape	Description
`observation.images.camera_high`	video	(480, 640, 3)	High camera view
`observation.images.camera_low`	video	(480, 640, 3)	Low camera view
`observation.state`	float64	(8,)	Joint positions (joint1-joint8)
`action`	float64	(8,)	Joint commands (joint1-joint8)
`observation.score`	float32	(1,)	Task success score
`timestamp`	float32	(1,)	Timestamp
`frame_index`	int64	(1,)	Frame index within episode
`episode_index`	int64	(1,)	Episode index
`task_index`	int64	(1,)	Task index

Dataset Statistics

Total Episodes: 162
Total Frames: 35,733
Total Tasks: 1
Frame Rate: 15 FPS
Data Size: ~100 MB (parquet files)
Video Size: ~500 MB (MP4 files)
Chunk Size: 1000 frames per chunk

Citation

If you use this dataset in your research, please cite:

@dataset{arx_pick_block_2024,
  title={ARX Pick Block Dataset},
  author={IO Intelligence},
  year={2024},
  url={https://huggingface.co/datasets/io-intelligence/arx_pick_block},
  license={Apache-2.0}
}

License

This dataset is released under the Apache License 2.0. See the LICENSE file for more details.

Contributing

Contributions to improve this dataset are welcome! Please feel free to submit issues or pull requests.

Contact

For questions or support, please contact us at https://io-ai.tech.