- 説明:
これらのデータセットは、ロボスイート ロボット アーム シミュレーターの PickPlaceCan 環境で作成されています。人間のデータセットは、 RLDS Creatorとゲームパッド コントローラーを使用して 1 人のオペレーターによって記録されました。
合成データセットは、 EnvLogger ライブラリを使用して記録されています。
データセットはRLDS 形式に従って、ステップとエピソードを表します。
エピソードは 400 ステップで構成されます。各エピソードでは、タスクが完了するとタグが追加されます。このタグは、カスタム ステップ メタデータの一部として保存されます。
EnvLogger の依存関係により、このデータセットの生成は現在 Linux 環境でのみサポートされていることに注意してください。
ソースコード:
tfds.rlds.robosuite_panda_pick_place_can.RobosuitePandaPickPlaceCan
バージョン:
-
1.0.0
(デフォルト): 初期リリース。
-
自動キャッシュ(ドキュメント): いいえ
監視されたキー(
as_supervised
docを参照):None
図( tfds.show_examples ): サポートされていません。
引用:
@misc{ramos2021rlds,
title={RLDS: an Ecosystem to Generate, Share and Use Datasets in Reinforcement Learning},
author={Sabela Ramos and Sertan Girgin and Léonard Hussenot and Damien Vincent and Hanna Yakubovich and Daniel Toyama and Anita Gergely and Piotr Stanczyk and Raphael Marinier and Jeremiah Harmsen and Olivier Pietquin and Nikola Momchev},
year={2021},
eprint={2111.02767},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
robosuite_panda_pick_place_can/human_dc29b40a (デフォルト設定)
構成の説明: 人間が生成したデータセット (50 エピソード)。
ダウンロードサイズ:
96.67 MiB
データセットサイズ:
407.24 MiB
スプリット:
スプリット | 例 |
---|---|
'train' | 50 |
- 機能構造:
FeaturesDict({
'agent_id': string,
'episode_id': string,
'episode_index': int32,
'steps': Dataset({
'action': Tensor(shape=(7,), dtype=float64),
'discount': float64,
'image': Image(shape=(None, None, 3), dtype=uint8),
'is_first': bool,
'is_last': bool,
'is_terminal': bool,
'observation': FeaturesDict({
'Can_pos': Tensor(shape=(3,), dtype=float64),
'Can_quat': Tensor(shape=(4,), dtype=float64),
'Can_to_robot0_eef_pos': Tensor(shape=(3,), dtype=float64),
'Can_to_robot0_eef_quat': Tensor(shape=(4,), dtype=float32),
'object-state': Tensor(shape=(14,), dtype=float64),
'robot0_eef_pos': Tensor(shape=(3,), dtype=float64),
'robot0_eef_quat': Tensor(shape=(4,), dtype=float64),
'robot0_gripper_qpos': Tensor(shape=(2,), dtype=float64),
'robot0_gripper_qvel': Tensor(shape=(2,), dtype=float64),
'robot0_joint_pos_cos': Tensor(shape=(7,), dtype=float64),
'robot0_joint_pos_sin': Tensor(shape=(7,), dtype=float64),
'robot0_joint_vel': Tensor(shape=(7,), dtype=float64),
'robot0_proprio-state': Tensor(shape=(32,), dtype=float64),
}),
'reward': float64,
'tag:placed': bool,
}),
})
- 機能のドキュメント:
特徴 | クラス | 形 | Dtype | 説明 |
---|---|---|---|---|
特徴辞書 | ||||
エージェント_id | テンソル | ストリング | ||
episode_id | テンソル | ストリング | ||
episode_index | テンソル | int32 | ||
ステップ | データセット | |||
ステップ/アクション | テンソル | (7、) | float64 | |
ステップ/割引 | テンソル | float64 | ||
歩数/画像 | 画像 | (なし、なし、3) | uint8 | |
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/観察 | 特徴辞書 | |||
ステップ/観察/Can_pos | テンソル | (3) | float64 | |
ステップ/観察/Can_quat | テンソル | (4) | float64 | |
ステップ/観察/Can_to_robot0_eef_pos | テンソル | (3) | float64 | |
ステップ/観察/Can_to_robot0_eef_quat | テンソル | (4) | float32 | |
ステップ/観察/オブジェクトの状態 | テンソル | (14) | float64 | |
ステップ/観察/robot0_eef_pos | テンソル | (3) | float64 | |
ステップ/観察/robot0_eef_quat | テンソル | (4) | float64 | |
ステップ/観察/robot0_gripper_qpos | テンソル | (2) | float64 | |
ステップ/観察/robot0_gripper_qvel | テンソル | (2) | float64 | |
ステップ/観測/robot0_joint_pos_cos | テンソル | (7、) | float64 | |
ステップ/観測/robot0_joint_pos_sin | テンソル | (7、) | float64 | |
steps/observation/robot0_joint_vel | テンソル | (7、) | float64 | |
steps/observation/robot0_proprio-state | テンソル | (32) | float64 | |
ステップ/報酬 | テンソル | float64 | ||
ステップ/タグ:配置 | テンソル | ブール |
- 例( tfds.as_dataframe ):
robosuite_panda_pick_place_can/human_images_dc29b40a
構成の説明: 観測で異なるカメラ アングルの画像を含む、人間が生成したデータセット。生成には時間がかかる場合があります。
ホームページ: https://www.tensorflow.org/datasets/catalog/robosuite_panda_pick_place_can
ダウンロードサイズ:
10.95 GiB
データセットサイズ:
7.53 GiB
スプリット:
スプリット | 例 |
---|---|
'train' | 50 |
- 機能構造:
FeaturesDict({
'agent_id': string,
'episode_id': string,
'episode_index': int32,
'steps': Dataset({
'action': Tensor(shape=(7,), dtype=float64),
'discount': float64,
'image': Image(shape=(None, None, 3), dtype=uint8),
'is_first': bool,
'is_last': bool,
'is_terminal': bool,
'observation': FeaturesDict({
'Can_pos': Tensor(shape=(3,), dtype=float64),
'Can_quat': Tensor(shape=(4,), dtype=float64),
'Can_to_robot0_eef_pos': Tensor(shape=(3,), dtype=float64),
'Can_to_robot0_eef_quat': Tensor(shape=(4,), dtype=float32),
'agentview_image': Image(shape=(256, 256, 3), dtype=uint8),
'birdview_image': Image(shape=(256, 256, 3), dtype=uint8),
'object-state': Tensor(shape=(14,), dtype=float64),
'robot0_eef_pos': Tensor(shape=(3,), dtype=float64),
'robot0_eef_quat': Tensor(shape=(4,), dtype=float64),
'robot0_eye_in_hand_image': Image(shape=(256, 256, 3), dtype=uint8),
'robot0_gripper_qpos': Tensor(shape=(2,), dtype=float64),
'robot0_gripper_qvel': Tensor(shape=(2,), dtype=float64),
'robot0_joint_pos_cos': Tensor(shape=(7,), dtype=float64),
'robot0_joint_pos_sin': Tensor(shape=(7,), dtype=float64),
'robot0_joint_vel': Tensor(shape=(7,), dtype=float64),
'robot0_proprio-state': Tensor(shape=(32,), dtype=float64),
'robot0_robotview_image': Image(shape=(256, 256, 3), dtype=uint8),
}),
'reward': float64,
'tag:placed': bool,
}),
})
- 機能のドキュメント:
特徴 | クラス | 形 | Dtype | 説明 |
---|---|---|---|---|
特徴辞書 | ||||
エージェント_id | テンソル | ストリング | ||
episode_id | テンソル | ストリング | ||
episode_index | テンソル | int32 | ||
ステップ | データセット | |||
ステップ/アクション | テンソル | (7、) | float64 | |
ステップ/割引 | テンソル | float64 | ||
歩数/画像 | 画像 | (なし、なし、3) | uint8 | |
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/観察 | 特徴辞書 | |||
ステップ/観察/Can_pos | テンソル | (3) | float64 | |
ステップ/観察/Can_quat | テンソル | (4) | float64 | |
ステップ/観察/Can_to_robot0_eef_pos | テンソル | (3) | float64 | |
ステップ/観察/Can_to_robot0_eef_quat | テンソル | (4) | float32 | |
ステップ/観察/agentview_image | 画像 | (256, 256, 3) | uint8 | |
ステップ/観察/birdview_image | 画像 | (256, 256, 3) | uint8 | |
ステップ/観察/オブジェクトの状態 | テンソル | (14) | float64 | |
ステップ/観察/robot0_eef_pos | テンソル | (3) | float64 | |
ステップ/観察/robot0_eef_quat | テンソル | (4) | float64 | |
steps/observation/robot0_eye_in_hand_image | 画像 | (256, 256, 3) | uint8 | |
ステップ/観察/robot0_gripper_qpos | テンソル | (2) | float64 | |
ステップ/観察/robot0_gripper_qvel | テンソル | (2) | float64 | |
ステップ/観測/robot0_joint_pos_cos | テンソル | (7、) | float64 | |
ステップ/観測/robot0_joint_pos_sin | テンソル | (7、) | float64 | |
steps/observation/robot0_joint_vel | テンソル | (7、) | float64 | |
steps/observation/robot0_proprio-state | テンソル | (32) | float64 | |
ステップ/観察/robot0_robotview_image | 画像 | (256, 256, 3) | uint8 | |
ステップ/報酬 | テンソル | float64 | ||
ステップ/タグ:配置 | テンソル | ブール |
- 例( tfds.as_dataframe ):
robosuite_panda_pick_place_can/synthetic_stochastic_sac_afe13968
構成の説明: SAC でトレーニングされた確率的エージェントによって生成された合成データセット (200 エピソード)。
ダウンロードサイズ:
144.44 MiB
データセットのサイズ:
622.86 MiB
スプリット:
スプリット | 例 |
---|---|
'train' | 200 |
- 機能構造:
FeaturesDict({
'agent_id': string,
'episode_id': string,
'episode_index': int32,
'steps': Dataset({
'action': Tensor(shape=(7,), dtype=float32),
'discount': float64,
'image': Image(shape=(None, None, 3), dtype=uint8),
'is_first': bool,
'is_last': bool,
'is_terminal': bool,
'observation': FeaturesDict({
'Can_pos': Tensor(shape=(3,), dtype=float32),
'Can_quat': Tensor(shape=(4,), dtype=float32),
'Can_to_robot0_eef_pos': Tensor(shape=(3,), dtype=float32),
'Can_to_robot0_eef_quat': Tensor(shape=(4,), dtype=float32),
'object-state': Tensor(shape=(14,), dtype=float32),
'robot0_eef_pos': Tensor(shape=(3,), dtype=float32),
'robot0_eef_quat': Tensor(shape=(4,), dtype=float32),
'robot0_gripper_qpos': Tensor(shape=(2,), dtype=float32),
'robot0_gripper_qvel': Tensor(shape=(2,), dtype=float32),
'robot0_joint_pos_cos': Tensor(shape=(7,), dtype=float32),
'robot0_joint_pos_sin': Tensor(shape=(7,), dtype=float32),
'robot0_joint_vel': Tensor(shape=(7,), dtype=float32),
'robot0_proprio-state': Tensor(shape=(32,), dtype=float32),
}),
'reward': float64,
'tag:placed': bool,
}),
})
- 機能のドキュメント:
特徴 | クラス | 形 | Dtype | 説明 |
---|---|---|---|---|
特徴辞書 | ||||
エージェント_id | テンソル | ストリング | ||
episode_id | テンソル | ストリング | ||
episode_index | テンソル | int32 | ||
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 | |
ステップ/割引 | テンソル | float64 | ||
歩数/画像 | 画像 | (なし、なし、3) | uint8 | |
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/観察 | 特徴辞書 | |||
ステップ/観察/Can_pos | テンソル | (3) | float32 | |
ステップ/観察/Can_quat | テンソル | (4) | float32 | |
ステップ/観察/Can_to_robot0_eef_pos | テンソル | (3) | float32 | |
ステップ/観察/Can_to_robot0_eef_quat | テンソル | (4) | float32 | |
ステップ/観察/オブジェクトの状態 | テンソル | (14) | float32 | |
ステップ/観察/robot0_eef_pos | テンソル | (3) | float32 | |
ステップ/観察/robot0_eef_quat | テンソル | (4) | float32 | |
ステップ/観察/robot0_gripper_qpos | テンソル | (2) | float32 | |
ステップ/観察/robot0_gripper_qvel | テンソル | (2) | float32 | |
ステップ/観測/robot0_joint_pos_cos | テンソル | (7、) | float32 | |
ステップ/観測/robot0_joint_pos_sin | テンソル | (7、) | float32 | |
steps/observation/robot0_joint_vel | テンソル | (7、) | float32 | |
steps/observation/robot0_proprio-state | テンソル | (32) | float32 | |
ステップ/報酬 | テンソル | float64 | ||
ステップ/タグ:配置 | テンソル | ブール |
- 例( tfds.as_dataframe ):