Datasetpublic

DAVIAN-Robotics/fr3-ins-sqcirc-v4-metra-a0p3-s0-F1cm-3000-merged

This dataset was created using LeRobot. Dataset Description Simulated Franka FR3 manipulation demonstrations for the FMB-SqCircle task, generated by the Ours data source. RL rollouts from the METRA skill-discovery policy at diversity weight alpha=0.3. Summary Task: FMB-SqCircle Data source: Ours Episodes: 3000 Frames: 268145 Control / video rate: 20 Hz LeRobot codebase format: v3.0 Cameras: wrist_cam, over_shoulder_left_camera Action convention… See the full description on the dataset page: https://huggingface.co/datasets/DAVIAN-Robotics/fr3-ins-sqcirc-v4-metra-a0p3-s0-F1cm-3000-merged.

sourceHugging Faceapache-2.0updated 7d agoView on Hugging Face

0likes89downloads

Dataset Card

This dataset was created using LeRobot.

Dataset Description

Simulated Franka FR3 manipulation demonstrations for the FMB-SqCircle task, generated by the Ours data source.

RL rollouts from the METRA skill-discovery policy at diversity weight alpha=0.3.

Summary

—Task: FMB-SqCircle
—Data source: Ours
—Episodes: 3000
—Frames: 268145
—Control / video rate: 20 Hz
—LeRobot codebase format: v3.0
—Cameras: wristcam, overshoulderleftcamera

Action convention

Read this before training on the data — it is the most common source of silent error.

—Layout: absolute_panda_joint1..7_target_droid_frame,gripper_close_target
—Recorded format: RECORDED FORMAT (always, regardless of policy): arm = ABSOLUTE droid-frame joint target (arm_proc, j7 -pi/4); gripper = binary {0,1} (raw>0 -> close)
—Label convention: v2:arm=applied_this_step,gripper=effective_requests
—Action repeat: 1

Actions are stored absolute. Delta-action training must derive the delta at load time against the chunk-start proprioceptive state; do not apply a second delta transform.

Simulation plant

—Canonical env: fr3-peg-insert-sqcirc-v3
—Env fingerprint: 61daec5f5a052e07441b6814d49c5f6acb3b8cd57f3004b09f2071d210fb73c6
—Gripper: panda
—Initial-state distribution: pregrasp:0.03,15
—Frame convention: env_relative_pos+world_quat;world_frame_velocity

Provenance

—Generator: none
—Policy checkpoint: /home/nas_main/minhopark/repos/dg-rl/libs/FlashSAC/models/fr3-peg-insert/_snapshots/fr3-ins-sqcirc-v4-metra-a0p3-s0/step15860
—Episodes attempted / accepted: 1024 / 1024
—Instruction: 'Insert the two-pronged peg (a cylinder and a square) into its matching hole on the board'

The full provenance record ships inside the dataset at meta/provenance.json.

Notes

—This is simulation data rendered in Isaac Sim, not real-robot capture.
—Episode lengths vary by data source; check meta/episodes before fixing a chunk size.
—The v3.0 tag is the LeRobot codebase format version, not a dataset version.

Visualize

Open in the LeRobot dataset visualizer

—Homepage: [More Information Needed]
—Paper: [More Information Needed]
—License: apache-2.0

Dataset Structure

meta/info.json:

json

{
    "codebase_version": "v3.0",
    "robot_type": "unknown",
    "total_episodes": 3000,
    "total_frames": 268145,
    "total_tasks": 1,
    "chunks_size": 1000,
    "data_files_size_in_mb": 100,
    "video_files_size_in_mb": 200,
    "fps": 20,
    "splits": {
        "train": "0:3000"
    },
    "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": {
        "action": {
            "dtype": "float32",
            "shape": [
                8
            ],
            "names": {
                "motors": [
                    "j0",
                    "j1",
                    "j2",
                    "j3",
                    "j4",
                    "j5",
                    "j6",
                    "gripper"
                ]
            },
            "fps": 20
        },
        "observation.state": {
            "dtype": "float32",
            "shape": [
                8
            ],
            "names": {
                "motors": [
                    "joint_0",
                    "joint_1",
                    "joint_2",
                    "joint_3",
                    "joint_4",
                    "joint_5",
                    "joint_6",
                    "joint_7"
                ]
            },
            "fps": 20
        },
        "observation.velocity": {
            "dtype": "float32",
            "shape": [
                8
            ],
            "names": null,
            "fps": 20
        },
        "observation.skill_z": {
            "dtype": "float32",
            "shape": [
                3
            ],
            "names": null,
            "fps": 20
        },
        "episode_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "frame_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "task_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "timestamp": {
            "dtype": "float32",
            "shape": [
                1
            ],
            "names": null
        },
        "next.done": {
            "dtype": "bool",
            "shape": [
                1
            ],
            "names": null,
            "fps": 20
        },
        "observation.images.wrist_cam": {
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channels"
            ],
            "info": {
                "video.height": 480,
                "video.width": 640,
                "video.codec": "avc1",
                "video.pix_fmt": "yuv420p",
                "video.is_depth_map": false,
                "video.fps": 20.0,
                "video.channels": 3,
                "has_audio": false
            }
        },
        "observation.images.over_shoulder_left_camera": {
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channels"
            ],
            "info": {
                "video.height": 480,
                "video.width": 640,
                "video.codec": "avc1",
                "video.pix_fmt": "yuv420p",
                "video.is_depth_map": false,
                "video.fps": 20.0,
                "video.channels": 3,
                "has_audio": false
            }
        }
    }
}

Citation

BibTeX:

bibtex

[More Information Needed]