robosuite_panda_pick_place_can

  • คำอธิบาย :

ชุดข้อมูลเหล่านี้ถูกสร้างขึ้นด้วยสภาพแวดล้อม PickPlaceCan ของ เครื่องจำลองแขนหุ่นยนต์ robosuite ชุดข้อมูลของมนุษย์ถูกบันทึกโดยผู้ปฏิบัติงานคนเดียวโดยใช้ RLDS Creator และตัวควบคุมเกมแพด

ชุดข้อมูลสังเคราะห์ได้รับการบันทึกโดยใช้ ไลบรารี EnvLogger

ชุดข้อมูลเป็นไปตาม รูปแบบ RLDS เพื่อแสดงขั้นตอนและตอนต่างๆ

ตอนประกอบด้วย 400 ขั้นตอน ในแต่ละตอน แท็กจะถูกเพิ่มเมื่องานเสร็จสิ้น แท็กนี้จะถูกเก็บไว้เป็นส่วนหนึ่งของข้อมูลเมตาของขั้นตอนที่กำหนดเอง

โปรดทราบว่า เนื่องจากการพึ่งพา EnvLogger การสร้างชุดข้อมูลนี้ในปัจจุบันได้รับการสนับสนุนบนสภาพแวดล้อม Linux เท่านั้น

@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 (การกำหนดค่าเริ่มต้น)

  • คำอธิบาย การกำหนดค่า : ชุดข้อมูลที่มนุษย์สร้างขึ้น (50 ตอน)

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

  • ขนาดการดาวน์โหลด : 96.67 MiB

  • ขนาดชุดข้อมูล : 407.24 MiB

  • แยก :

แยก ตัวอย่าง
'train' 50
  • โครงสร้างคุณลักษณะ :
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,
    }),
})
  • เอกสารคุณสมบัติ :
ลักษณะเฉพาะ ระดับ รูปร่าง Dประเภท คำอธิบาย
คุณสมบัติDict
รหัสตัวแทน เทนเซอร์ สตริง
ตอนที่_id เทนเซอร์ สตริง
ตอนที่_ดัชนี เทนเซอร์ int32
ขั้นตอน ชุดข้อมูล
ขั้นตอน/การดำเนินการ เทนเซอร์ (7,) float64
ขั้นตอน/ส่วนลด เทนเซอร์ float64
ขั้นตอน/ภาพ ภาพ (ไม่มี, ไม่มี, 3) uint8
ขั้นตอน/is_first เทนเซอร์ บูล
ขั้นตอน/is_last เทนเซอร์ บูล
ขั้นตอน/is_terminal เทนเซอร์ บูล
ขั้นตอน/ข้อสังเกต คุณสมบัติDict
ขั้นตอน/ข้อสังเกต/Can_pos เทนเซอร์ (3,) float64
ขั้นตอน/ข้อสังเกต/Can_quat เทนเซอร์ (4,) float64
ขั้นตอน/ข้อสังเกต/Can_to_robot0_eef_pos เทนเซอร์ (3,) float64
ขั้นตอน/ข้อสังเกต/Can_to_robot0_eef_quat เทนเซอร์ (4,) ลอย32
ขั้นตอน / การสังเกต / สถานะวัตถุ เทนเซอร์ (14,) float64
ขั้นตอน/การสังเกต/robot0_eef_pos เทนเซอร์ (3,) float64
ขั้นตอน/การสังเกต/robot0_eef_quat เทนเซอร์ (4,) float64
ขั้นตอน/การสังเกต/robot0_gripper_qpos เทนเซอร์ (2,) float64
ขั้นตอน/การสังเกต/robot0_gripper_qvel เทนเซอร์ (2,) float64
ขั้นตอน/ข้อสังเกต/robot0_joint_pos_cos เทนเซอร์ (7,) float64
ขั้นตอน/ข้อสังเกต/robot0_joint_pos_sin เทนเซอร์ (7,) float64
ขั้นตอน/ข้อสังเกต/robot0_joint_vel เทนเซอร์ (7,) float64
ขั้นตอน/การสังเกต/robot0_proprio-state เทนเซอร์ (32,) float64
ขั้นตอน / รางวัล เทนเซอร์ float64
ขั้นตอน/แท็ก: วาง เทนเซอร์ บูล

robosuite_panda_pick_place_can/human_images_dc29b40a

  • คำอธิบาย การกำหนดค่า : ชุดข้อมูลที่มนุษย์สร้างขึ้น รวมถึงภาพที่มีมุมกล้องต่างกันในการสังเกต โปรดทราบว่าอาจใช้เวลาสักครู่ในการสร้าง

  • หน้าแรก : https://www.tensorflow.org/datasets/catalog/robosuite_panda_pick_place_can

  • ขนาดการดาวน์โหลด : 10.95 GiB

  • ขนาดชุดข้อมูล : 7.53 GiB

  • แยก :

แยก ตัวอย่าง
'train' 50
  • โครงสร้างคุณลักษณะ :
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,
    }),
})
  • เอกสารคุณสมบัติ :
ลักษณะเฉพาะ ระดับ รูปร่าง Dประเภท คำอธิบาย
คุณสมบัติDict
รหัสตัวแทน เทนเซอร์ สตริง
ตอนที่_id เทนเซอร์ สตริง
ตอนที่_ดัชนี เทนเซอร์ int32
ขั้นตอน ชุดข้อมูล
ขั้นตอน/การดำเนินการ เทนเซอร์ (7,) float64
ขั้นตอน/ส่วนลด เทนเซอร์ float64
ขั้นตอน/ภาพ ภาพ (ไม่มี, ไม่มี, 3) uint8
ขั้นตอน/is_first เทนเซอร์ บูล
ขั้นตอน/is_last เทนเซอร์ บูล
ขั้นตอน/is_terminal เทนเซอร์ บูล
ขั้นตอน/ข้อสังเกต คุณสมบัติDict
ขั้นตอน/ข้อสังเกต/Can_pos เทนเซอร์ (3,) float64
ขั้นตอน/ข้อสังเกต/Can_quat เทนเซอร์ (4,) float64
ขั้นตอน/ข้อสังเกต/Can_to_robot0_eef_pos เทนเซอร์ (3,) float64
ขั้นตอน/ข้อสังเกต/Can_to_robot0_eef_quat เทนเซอร์ (4,) ลอย32
ขั้นตอน/ข้อสังเกต/agentview_image ภาพ (256, 256, 3) uint8
ขั้นตอน/ข้อสังเกต/birdview_image ภาพ (256, 256, 3) uint8
ขั้นตอน / การสังเกต / สถานะวัตถุ เทนเซอร์ (14,) float64
ขั้นตอน/การสังเกต/robot0_eef_pos เทนเซอร์ (3,) float64
ขั้นตอน/การสังเกต/robot0_eef_quat เทนเซอร์ (4,) float64
ขั้นตอน/การสังเกต/robot0_eye_in_hand_image ภาพ (256, 256, 3) uint8
ขั้นตอน/การสังเกต/robot0_gripper_qpos เทนเซอร์ (2,) float64
ขั้นตอน/การสังเกต/robot0_gripper_qvel เทนเซอร์ (2,) float64
ขั้นตอน/ข้อสังเกต/robot0_joint_pos_cos เทนเซอร์ (7,) float64
ขั้นตอน/ข้อสังเกต/robot0_joint_pos_sin เทนเซอร์ (7,) float64
ขั้นตอน/ข้อสังเกต/robot0_joint_vel เทนเซอร์ (7,) float64
ขั้นตอน/การสังเกต/robot0_proprio-state เทนเซอร์ (32,) float64
ขั้นตอน/ข้อสังเกต/robot0_robotview_image ภาพ (256, 256, 3) uint8
ขั้นตอน / รางวัล เทนเซอร์ float64
ขั้นตอน/แท็ก: วาง เทนเซอร์ บูล

robosuite_panda_pick_place_can/synthetic_stochastic_sac_afe13968

  • คำอธิบาย การกำหนดค่า : ชุดข้อมูลสังเคราะห์ที่สร้างโดยตัวแทนสุ่มที่ได้รับการฝึกฝนด้วย SAC (200 ตอน)

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

  • ขนาดการดาวน์โหลด : 144.44 MiB

  • ขนาดชุดข้อมูล : 622.86 MiB

  • แยก :

แยก ตัวอย่าง
'train' 200
  • โครงสร้างคุณลักษณะ :
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,
    }),
})
  • เอกสารคุณสมบัติ :
ลักษณะเฉพาะ ระดับ รูปร่าง Dประเภท คำอธิบาย
คุณสมบัติDict
รหัสตัวแทน เทนเซอร์ สตริง
ตอนที่_id เทนเซอร์ สตริง
ตอนที่_ดัชนี เทนเซอร์ int32
ขั้นตอน ชุดข้อมูล
ขั้นตอน/การดำเนินการ เทนเซอร์ (7,) ลอย32
ขั้นตอน/ส่วนลด เทนเซอร์ float64
ขั้นตอน/ภาพ ภาพ (ไม่มี, ไม่มี, 3) uint8
ขั้นตอน/is_first เทนเซอร์ บูล
ขั้นตอน/is_last เทนเซอร์ บูล
ขั้นตอน/is_terminal เทนเซอร์ บูล
ขั้นตอน/ข้อสังเกต คุณสมบัติDict
ขั้นตอน/ข้อสังเกต/Can_pos เทนเซอร์ (3,) ลอย32
ขั้นตอน/ข้อสังเกต/Can_quat เทนเซอร์ (4,) ลอย32
ขั้นตอน/ข้อสังเกต/Can_to_robot0_eef_pos เทนเซอร์ (3,) ลอย32
ขั้นตอน/ข้อสังเกต/Can_to_robot0_eef_quat เทนเซอร์ (4,) ลอย32
ขั้นตอน / การสังเกต / สถานะวัตถุ เทนเซอร์ (14,) ลอย32
ขั้นตอน/การสังเกต/robot0_eef_pos เทนเซอร์ (3,) ลอย32
ขั้นตอน/การสังเกต/robot0_eef_quat เทนเซอร์ (4,) ลอย32
ขั้นตอน/การสังเกต/robot0_gripper_qpos เทนเซอร์ (2,) ลอย32
ขั้นตอน/การสังเกต/robot0_gripper_qvel เทนเซอร์ (2,) ลอย32
ขั้นตอน/ข้อสังเกต/robot0_joint_pos_cos เทนเซอร์ (7,) ลอย32
ขั้นตอน/ข้อสังเกต/robot0_joint_pos_sin เทนเซอร์ (7,) ลอย32
ขั้นตอน/ข้อสังเกต/robot0_joint_vel เทนเซอร์ (7,) ลอย32
ขั้นตอน/การสังเกต/robot0_proprio-state เทนเซอร์ (32,) ลอย32
ขั้นตอน / รางวัล เทนเซอร์ float64
ขั้นตอน/แท็ก: วาง เทนเซอร์ บูล