utokyo_xarm_pick_and_place_converted_externally_to_rlds

  • 説明

xArm でオブジェクトを選択および配置する

スプリット
'train' 92
'val' 10
  • 機能の構造:
FeaturesDict({
    'episode_metadata': FeaturesDict({
        'file_path': Text(shape=(), dtype=string),
    }),
    'steps': Dataset({
        'action': Tensor(shape=(7,), dtype=float32),
        'discount': Scalar(shape=(), dtype=float32),
        'is_first': bool,
        'is_last': bool,
        'is_terminal': bool,
        'language_embedding': Tensor(shape=(512,), dtype=float32),
        'language_instruction': Text(shape=(), dtype=string),
        'observation': FeaturesDict({
            'end_effector_pose': Tensor(shape=(6,), dtype=float32),
            'hand_image': Image(shape=(224, 224, 3), dtype=uint8),
            'image': Image(shape=(224, 224, 3), dtype=uint8),
            'image2': Image(shape=(224, 224, 3), dtype=uint8),
            'joint_state': Tensor(shape=(14,), dtype=float32),
            'joint_trajectory': Tensor(shape=(21,), dtype=float32),
        }),
        'reward': Scalar(shape=(), dtype=float32),
    }),
})
  • 機能ドキュメント:
特徴クラスDタイプ説明
特徴辞書
エピソード_メタデータ特徴辞書
エピソードメタデータ/ファイルパス文章元のデータ ファイルへのパス。
ステップデータセット
ステップ/アクションテンソル(7、) float32ロボットのアクションは、[3x EEF 位置、3x EEF 方向ヨー/ピッチ/ロール、1x グリッパー開閉位置] で構成されます。
歩数/割引スカラーfloat32割引が指定されている場合、デフォルトは 1 です。
ステップ/is_firstテンソルブール
ステップ/is_lastテンソルブール
ステップ/is_terminalテンソルブール
ステップ/言語_埋め込みテンソル(512,) float32コナ言語の埋め込み。 https://tfhub.dev/google/universal-sentence-encoder-large/5を参照してください。
ステップ/言語説明文章言語指導。
ステップ/観察特徴辞書
ステップ/観察/end_effect_poseテンソル(6,) float32ロボット エンド エフェクターのポーズは、[3x EEF 位置、3x EEF 方向のヨー/ピッチ/ロール] で構成されます。
歩数/観察/手の画像画像(224, 224, 3) uint8ハンドカメラRGB観察。
手順・観察・イメージ画像(224, 224, 3) uint8メインカメラRGB観察。
手順/観察/画像2画像(224, 224, 3) uint8別のカメラの別の視点からの RGB 観察。
ステップ/観察/joint_stateテンソル(14,) float32ロボット ジョイントの状態は、[7x ロボット ジョイント角度、7x ロボット ジョイント速度] で構成されます。
ステップ/観察/関節軌跡テンソル(21,) float32ロボット ジョイントの軌道は、[7x ロボット ジョイント角度、7x ロボット ジョイント速度、7x ロボット ジョイント加速度] で構成されます。
歩数/報酬スカラーfloat32提供されている場合は報酬、デモの最終ステップで 1。
@misc{matsushima2023weblab,
  title={Weblab xArm Dataset},
  author={Tatsuya Matsushima and Hiroki Furuta and Yusuke Iwasawa and Yutaka Matsuo},
  year={2023},
}