CoolFace
Datasetpublic

DAVIAN-Robotics/metra-alpha-ma-medium-a02-n3000

This dataset was created using LeRobot. Dataset Description Successful BananaInBowl demonstrations collected from an RFCL-trained SAC policy in Isaac Lab (RoboLab), for behaviour-cloning research on strategy diversity. Lane: metra-alpha — a sweep of the METRA intrinsic-reward scale alpha, with every other axis held fixed (banana task, 50 demos, sf=0.5, z_dim=3, phi_space=full, z_unit=true). Each dataset is one (difficulty level, alpha) cell. Difficulty here: medium. Collection:… See the full description on the dataset page: https://huggingface.co/datasets/DAVIAN-Robotics/metra-alpha-ma-medium-a02-n3000.

sourceHugging Faceapache-2.0updated 1mo agoView on Hugging Face
0likes46downloads
Dataset Card

This dataset was created using LeRobot.

Dataset Description

Successful BananaInBowl demonstrations collected from an RFCL-trained SAC policy in Isaac Lab (RoboLab), for behaviour-cloning research on strategy diversity.

Lane: metra-alpha — a sweep of the METRA intrinsic-reward scale alpha, with every other axis held fixed (banana task, 50 demos, sf=0.5, zdim=3, phispace=full, z_unit=true). Each dataset is one (difficulty level, alpha) cell. Difficulty here: medium.

Collection: deterministic rollouts (tanh(mean), no action noise), success episodes only. Success is defined by the environment (K-step settle streak), so the hold segment is already inside every trajectory. Initial states are drawn from the curated pose bank (init_pose_source=dataset, train split) — the same distribution the policy was trained and evaluated on.

Size: exactly 3000 episodes, physically subset from a larger merged dataset with a fixed permutation. Do not use LeRobot's runtime episodes= argument to subsample further — combined with delta_timestamps it collapses the action chunk and sets action_is_pad everywhere, and it is silent (training loss goes down).

Video: H.264, constant frame rate (CFR-normalised after merge). Non-CFR merged video breaks the torchcodec backend used by pi0.5 fine-tuning.

Actions: absolute joint positions (7) + binary gripper (1). Delta conversion, if any, happens at training time — it is not baked in here.

Provenance: meta/provenance.json records the exact policy checkpoint, sampling mode and initial-state sampler that produced this data.

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": 215633,
    "total_tasks": 1,
    "chunks_size": 1000,
    "data_files_size_in_mb": 100,
    "video_files_size_in_mb": 200,
    "fps": 15,
    "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": 15
        },
        "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": 15
        },
        "observation.velocity": {
            "dtype": "float32",
            "shape": [
                8
            ],
            "names": null,
            "fps": 15
        },
        "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": 15
        },
        "observation.images.over_shoulder_left_camera": {
            "dtype": "video",
            "shape": [
                270,
                480,
                3
            ],
            "names": [
                "height",
                "width",
                "channels"
            ],
            "info": {
                "video.height": 270,
                "video.width": 480,
                "video.codec": "avc1",
                "video.pix_fmt": "yuv420p",
                "video.is_depth_map": false,
                "video.fps": 15.0,
                "video.channels": 3,
                "has_audio": false
            }
        },
        "observation.images.wrist_cam": {
            "dtype": "video",
            "shape": [
                270,
                480,
                3
            ],
            "names": [
                "height",
                "width",
                "channels"
            ],
            "info": {
                "video.height": 270,
                "video.width": 480,
                "video.codec": "avc1",
                "video.pix_fmt": "yuv420p",
                "video.is_depth_map": false,
                "video.fps": 15.0,
                "video.channels": 3,
                "has_audio": false
            }
        }
    }
}

Citation

BibTeX:

bibtex
[More Information Needed]