robosuite_panda_pick_place_can

  • 설명 :

이러한 데이터 세트는 robosuite 로봇 팔 시뮬레이터 의 PickPlaceCan 환경으로 생성되었습니다. 인간 데이터 세트는 RLDS Creator 와 게임패드 컨트롤러를 사용하여 단일 운영자가 기록했습니다.

합성 데이터 세트는 EnvLogger 라이브러리 를 사용하여 기록되었습니다.

데이터 세트는 RLDS 형식 을 따라 단계와 에피소드를 나타냅니다.

에피소드는 400개의 단계로 구성됩니다. 각 에피소드에서 작업이 완료되면 태그가 추가되며 이 태그는 사용자 정의 단계 메타데이터의 일부로 저장됩니다.

EnvLogger 종속성으로 인해 이 데이터 세트 생성은 현재 Linux 환경에서만 지원됩니다.

@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(기본 구성)

  • 구성 설명 : 인간이 생성한 데이터 세트(50개 에피소드).

  • 홈페이지 : https://github.com/google-research/rlds

  • 다운로드 크기 : 96.67 MiB

  • 데이터 세트 크기 : 407.24 MiB

  • 분할 :

나뉘다
'train' 50
  • 기능 구조 :
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,
    }),
})
  • 기능 문서 :
특징 수업 모양 D타입 설명
풍모Dict
agent_id 텐서
episode_id 텐서
episode_index 텐서 int32
단계 데이터세트
단계/액션 텐서 (7,) float64
단계/할인 텐서 float64
단계/이미지 영상 (없음, 없음, 3) uint8
단계/is_first 텐서 부울
단계/is_last 텐서 부울
단계/is_terminal 텐서 부울
단계/관찰 풍모Dict
단계/관찰/Can_pos 텐서 (삼,) float64
단계/관찰/Can_quat 텐서 (4,) float64
단계/관찰/Can_to_robot0_eef_pos 텐서 (삼,) float64
단계/관찰/Can_to_robot0_eef_quat 텐서 (4,) float32
단계/관찰/객체 상태 텐서 (14,) float64
단계/관찰/robot0_eef_pos 텐서 (삼,) float64
단계/관찰/robot0_eef_quat 텐서 (4,) float64
단계/관찰/robot0_gripper_qpos 텐서 (2,) float64
단계/관찰/robot0_gripper_qvel 텐서 (2,) float64
단계/관찰/robot0_joint_pos_cos 텐서 (7,) float64
단계/관찰/robot0_joint_pos_sin 텐서 (7,) float64
단계/관찰/robot0_joint_vel 텐서 (7,) float64
단계/관찰/robot0_proprio-상태 텐서 (32,) float64
단계/보상 텐서 float64
단계/태그:배치됨 텐서 부울

robosuite_panda_pick_place_can/human_images_dc29b40a

나뉘다
'train' 50
  • 기능 구조 :
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,
    }),
})
  • 기능 문서 :
특징 수업 모양 D타입 설명
풍모Dict
agent_id 텐서
episode_id 텐서
episode_index 텐서 int32
단계 데이터세트
단계/액션 텐서 (7,) float64
단계/할인 텐서 float64
단계/이미지 영상 (없음, 없음, 3) uint8
단계/is_first 텐서 부울
단계/is_last 텐서 부울
단계/is_terminal 텐서 부울
단계/관찰 풍모Dict
단계/관찰/Can_pos 텐서 (삼,) float64
단계/관찰/Can_quat 텐서 (4,) float64
단계/관찰/Can_to_robot0_eef_pos 텐서 (삼,) float64
단계/관찰/Can_to_robot0_eef_quat 텐서 (4,) float32
단계/관찰/agentview_image 영상 (256, 256, 3) uint8
단계/관측/birdview_image 영상 (256, 256, 3) uint8
단계/관찰/객체 상태 텐서 (14,) float64
단계/관찰/robot0_eef_pos 텐서 (삼,) float64
단계/관찰/robot0_eef_quat 텐서 (4,) float64
단계/관찰/robot0_eye_in_hand_image 영상 (256, 256, 3) uint8
단계/관찰/robot0_gripper_qpos 텐서 (2,) float64
단계/관찰/robot0_gripper_qvel 텐서 (2,) float64
단계/관찰/robot0_joint_pos_cos 텐서 (7,) float64
단계/관찰/robot0_joint_pos_sin 텐서 (7,) float64
단계/관찰/robot0_joint_vel 텐서 (7,) float64
단계/관찰/robot0_proprio-상태 텐서 (32,) float64
단계/관찰/robot0_robotview_image 영상 (256, 256, 3) uint8
단계/보상 텐서 float64
단계/태그:배치됨 텐서 부울

robosuite_panda_pick_place_can/synthetic_stochastic_sac_afe13968

  • 구성 설명 : SAC(200개 에피소드)로 훈련된 확률적 에이전트가 생성한 합성 데이터 세트입니다.

  • 홈페이지 : https://github.com/google-research/rlds

  • 다운로드 크기 : 144.44 MiB

  • 데이터 세트 크기 : 622.86 MiB

  • 분할 :

나뉘다
'train' 200
  • 기능 구조 :
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,
    }),
})
  • 기능 문서 :
특징 수업 모양 D타입 설명
풍모Dict
agent_id 텐서
episode_id 텐서
episode_index 텐서 int32
단계 데이터세트
단계/액션 텐서 (7,) float32
단계/할인 텐서 float64
단계/이미지 영상 (없음, 없음, 3) uint8
단계/is_first 텐서 부울
단계/is_last 텐서 부울
단계/is_terminal 텐서 부울
단계/관찰 풍모Dict
단계/관찰/Can_pos 텐서 (삼,) float32
단계/관찰/Can_quat 텐서 (4,) float32
단계/관찰/Can_to_robot0_eef_pos 텐서 (삼,) float32
단계/관찰/Can_to_robot0_eef_quat 텐서 (4,) float32
단계/관찰/객체 상태 텐서 (14,) float32
단계/관찰/robot0_eef_pos 텐서 (삼,) float32
단계/관찰/robot0_eef_quat 텐서 (4,) float32
단계/관찰/robot0_gripper_qpos 텐서 (2,) float32
단계/관찰/robot0_gripper_qvel 텐서 (2,) float32
단계/관찰/robot0_joint_pos_cos 텐서 (7,) float32
단계/관찰/robot0_joint_pos_sin 텐서 (7,) float32
단계/관찰/robot0_joint_vel 텐서 (7,) float32
단계/관찰/robot0_proprio-상태 텐서 (32,) float32
단계/보상 텐서 float64
단계/태그:배치됨 텐서 부울