Datasetpublic

thewisp/aloha_sim_selfplay_wide_1k

This dataset was created using LeRobot. Dataset Description 1,000 episodes of scripted self-play in gym-aloha (AlohaInsertion-v0 scene), generated for world-model video pretraining (VLA-JEPA-style teaching of an ACT student). Peg AND socket are spawned over wide ranges every episode (x in [0,0.3]/[-0.3,0], y in [0.35,0.65]). Episodes alternate three scripted Jacobian behaviors: right-arm peg push, left-arm socket push, bimanual reach (84-100% measured contact for push… See the full description on the dataset page: https://huggingface.co/datasets/thewisp/aloha_sim_selfplay_wide_1k.

sourceHugging Faceapache-2.0updated 4mo agoView on Hugging Face

0likes39downloads

Dataset Card

This dataset was created using LeRobot.

Dataset Description

1,000 episodes of scripted self-play in gym-aloha (AlohaInsertion-v0 scene), generated for world-model video pretraining (VLA-JEPA-style teaching of an ACT student). Peg AND socket are spawned over wide ranges every episode (x in [0,0.3]/[-0.3,0], y in [0.35,0.65]). Episodes alternate three scripted Jacobian behaviors: right-arm peg push, left-arm socket push, bimanual reach (84-100% measured contact for push modes, ~7cm object displacement; outcomes deliberately mixed - failures are valid world-model data). 150 frames/ep @50fps, 224x224 top camera, 14-dim state/action (scripted commands included; the associated research used video only in stage 1). SPLIT: episodes 0-969 were the stage-1 training corpus; episodes 970-999 were the held-out wide-scene validation gate (never trained on). Research context: github.com/TheWisp/lerobot branch proto/self-play-embodiment, research/selfplayembodiment/.

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

Dataset Structure

meta/info.json:

json

{
    "codebase_version": "v3.0",
    "robot_type": "bimanual_viperx_aloha_sim",
    "total_episodes": 1000,
    "total_frames": 150000,
    "total_tasks": 3,
    "chunks_size": 1000,
    "data_files_size_in_mb": 100,
    "video_files_size_in_mb": 200,
    "fps": 50,
    "splits": {
        "train": "0:1000"
    },
    "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.top": {
            "dtype": "video",
            "shape": [
                224,
                224,
                3
            ],
            "names": [
                "height",
                "width",
                "channel"
            ],
            "info": {
                "video.height": 224,
                "video.width": 224,
                "video.codec": "av1",
                "video.pix_fmt": "yuv420p",
                "video.is_depth_map": false,
                "video.fps": 50,
                "video.channels": 3,
                "has_audio": false
            }
        },
        "observation.state": {
            "dtype": "float32",
            "shape": [
                14
            ],
            "names": [
                "joint_0",
                "joint_1",
                "joint_2",
                "joint_3",
                "joint_4",
                "joint_5",
                "joint_6",
                "joint_7",
                "joint_8",
                "joint_9",
                "joint_10",
                "joint_11",
                "joint_12",
                "joint_13"
            ]
        },
        "action": {
            "dtype": "float32",
            "shape": [
                14
            ],
            "names": [
                "joint_0",
                "joint_1",
                "joint_2",
                "joint_3",
                "joint_4",
                "joint_5",
                "joint_6",
                "joint_7",
                "joint_8",
                "joint_9",
                "joint_10",
                "joint_11",
                "joint_12",
                "joint_13"
            ]
        },
        "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
        }
    }
}

Citation

BibTeX:

bibtex

[More Information Needed]