ucsd_pick_and_place_dataset_converted_externally_to_rlds

  • 説明

xArm ディストラクタを使用してオブジェクトを選択および配置する

スプリット
'train' 1,355
  • 機能の構造:
FeaturesDict({
    'episode_metadata': FeaturesDict({
        'disclaimer': Text(shape=(), dtype=string),
        'file_path': Text(shape=(), dtype=string),
        'n_transitions': Scalar(shape=(), dtype=int32),
        'success': Scalar(shape=(), dtype=bool),
        'success_labeled_by': Text(shape=(), dtype=string),
    }),
    'steps': Dataset({
        'action': Tensor(shape=(4,), 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({
            'image': Image(shape=(224, 224, 3), dtype=uint8),
            'state': Tensor(shape=(7,), 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/免責事項文章特定のエピソードに関する免責事項。
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。
エピソードメタデータ/n_transitionsスカラーint32エピソード内のトランジションの数。
エピソード_メタデータ/成功スカラーブールエピソードの最後の状態が成功状態の場合は True、それ以外の場合は False。
エピソード_メタデータ/成功_ラベル付け_by文章エピソードの成功 (そしてそれによる報酬) にラベルを付けたのは誰ですか。 [人間、分類子] のいずれかになります。
ステップデータセット
ステップ/アクションテンソル(4,) float32ロボット動作は、[グリッパー速度 3 倍、グリッパー開閉トルク 1 倍] 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
手順・観察・イメージ画像(224, 224, 3) uint8カメラRGB観察。
ステップ/観察/状態テンソル(7、) float32ロボットの状態は、[3x グリッパー位置、3x グリッパー方向、1x 指の距離] 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。
@preprint{Feng2023Finetuning,
    title={Finetuning Offline World Models in the Real World},
    author={Yunhai Feng, Nicklas Hansen, Ziyan Xiong, Chandramouli Rajagopalan, Xiaolong Wang},
    year={2023}
}