CoolFace
Datasetpublic

zmfldlwx/community-v1-so101-pick_and_place-converted

This dataset was created using LeRobot. Dataset Description Public LeRobot v3 derivative of the high-confidence pick-and-place, stacking, sorting, and chess-relocation subset of HuggingFaceVLA/community_dataset_v1. The source SO-100 six-joint action/state coordinates were converted into the SO-101 coordinate convention measured in the author's 150-episode reference set. The body-joint mapping is pan=x, shoulder_lift=180-x, elbow_flex=x-180, wrist_flex=x, and wrist_roll=x;… See the full description on the dataset page: https://huggingface.co/datasets/zmfldlwx/community-v1-so101-pick_and_place-converted.

sourceHugging Faceapache-2.0updated 24d agoView on Hugging Face
1likes1.1kdownloads
Dataset Card

This dataset was created using LeRobot.

<a class="flex" href="https://huggingface.co/spaces/lerobot/visualize_dataset?path=zmfldlwx/community-v1-so101-pickplace-chess-v3"> <img class="block dark:hidden" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-xl.svg"/> <img class="hidden dark:block" src="https://huggingface.co/datasets/huggingface/badges/resolve/main/visualize-this-dataset-xl-dark.svg"/> </a>

Dataset Description

Public LeRobot v3 derivative of the high-confidence pick-and-place, stacking, sorting, and chess-relocation subset of HuggingFaceVLA/communitydatasetv1.

The source SO-100 six-joint action/state coordinates were converted into the SO-101 coordinate convention measured in the author's 150-episode reference set. The body-joint mapping is pan=x, shoulder_lift=180-x, elbow_flex=x-180, wrist_flex=x, and wrist_roll=x; action and observation.state use the same mapping. Gripper direction and robust q01/q99 endpoints were inferred per source leaf and linearly mapped to the reference SO-101 gripper endpoints. This is a coordinate normalization for direct joint-action pretraining, not a claim that the source demonstrations were physically replayed on the author's SO-101.

Because Community-v1 camera extrinsics are not standardized, source camera observation.images.image is exposed as observation.images.primary and observation.images.image2 as observation.images.secondary; the names do not claim a fixed physical mounting position. No synthetic view was added. Every included leaf passed structural, Parquet, timestamp, finite-signal, video FPS/resolution, and cross-camera frame-count checks after merge. Minority H.264 source shards were transcoded to AV1 with exact frame-count verification so the merged feature schema is uniform. Automated QC passed; manual success review of all 6288 demonstrations remains PENDING.

Exact source leaves, rejected leaves, action transforms, gripper calibration endpoints, and QC results are stored under meta/source_selection_manifest.json and meta/source_conversion_results.json.

Episodes: 6288. Frames: 2821930. FPS: 30. License: Apache-2.0.

  • —Homepage: https://huggingface.co/datasets/HuggingFaceVLA/communitydatasetv1
  • —Paper: https://arxiv.org/abs/2506.01844
  • —License: apache-2.0

Dataset Structure

meta/info.json:

json
{
    "codebase_version": "v3.0",
    "fps": 30,
    "features": {
        "action": {
            "dtype": "float32",
            "shape": [
                6
            ],
            "names": [
                "shoulder_pan.pos",
                "shoulder_lift.pos",
                "elbow_flex.pos",
                "wrist_flex.pos",
                "wrist_roll.pos",
                "gripper.pos"
            ],
            "fps": 30
        },
        "observation.state": {
            "dtype": "float32",
            "shape": [
                6
            ],
            "names": [
                "shoulder_pan.pos",
                "shoulder_lift.pos",
                "elbow_flex.pos",
                "wrist_flex.pos",
                "wrist_roll.pos",
                "gripper.pos"
            ],
            "fps": 30
        },
        "observation.images.secondary": {
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channels"
            ],
            "info": {
                "video.fps": 30.0,
                "video.height": 480,
                "video.width": 640,
                "video.channels": 3,
                "video.codec": "av1",
                "video.pix_fmt": "yuv420p",
                "video.is_depth_map": false,
                "has_audio": false,
                "video.g": null,
                "video.video_backend": "pyav",
                "video.extra_options": null,
                "video.fast_decode": null,
                "video.preset": null,
                "video.crf": null
            }
        },
        "observation.images.primary": {
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channels"
            ],
            "info": {
                "video.fps": 30.0,
                "video.height": 480,
                "video.width": 640,
                "video.channels": 3,
                "video.codec": "av1",
                "video.pix_fmt": "yuv420p",
                "video.is_depth_map": false,
                "has_audio": false,
                "video.g": null,
                "video.video_backend": "pyav",
                "video.extra_options": null,
                "video.fast_decode": null,
                "video.preset": null,
                "video.crf": null
            }
        },
        "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": 6288,
    "total_frames": 2821930,
    "total_tasks": 60,
    "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": "so_follower",
    "splits": {
        "train": "0:6288"
    }
}

Citation

BibTeX:

bibtex
@misc{community_v1_so101_pickplace_chess_v3,
  title={Community Dataset v1 SO-101 Pick-Place and Chess v3},
  author={zmfldlwx},
  year={2026},
  url={https://huggingface.co/datasets/zmfldlwx/community-v1-so101-pickplace-chess-v3}
}