robomimic_mg

  • وصف :

تم جمع مجموعات البيانات التي تم إنشاؤها بواسطة آلة Robomimic باستخدام وكيل Soft Actor Critic الذي تم تدريبه بمكافأة كثيفة. تتكون كل مجموعة بيانات من المخزن المؤقت لإعادة التشغيل الخاص بالوكيل.

تحتوي كل مهمة على نسختين: واحدة تحتوي على ملاحظات منخفضة الأبعاد ( low_dim ) وأخرى تحتوي على صور ( image ).

تتبع مجموعات البيانات تنسيق RLDS لتمثيل الخطوات والحلقات.

@inproceedings{robomimic2021,
  title={What Matters in Learning from Offline Human Demonstrations for Robot Manipulation},
  author={Ajay Mandlekar and Danfei Xu and Josiah Wong and Soroush Nasiriany
          and Chen Wang and Rohun Kulkarni and Li Fei-Fei and Silvio Savarese
          and Yuke Zhu and Roberto Mart\'{i}n-Mart\'{i}n},
  booktitle={Conference on Robot Learning},
  year={2021}
}

robomimic_mg/lift_mg_image (التكوين الافتراضي)

  • حجم التحميل : 18.04 GiB

  • حجم مجموعة البيانات : 2.73 GiB

  • التخزين المؤقت التلقائي ( الوثائق ): لا

  • الإنشقاقات :

ينقسم أمثلة
'train' 1500
  • هيكل الميزة :
FeaturesDict({
    'episode_id': string,
    'horizon': int32,
    'steps': Dataset({
        'action': Tensor(shape=(7,), dtype=float64),
        'discount': int32,
        'is_first': bool,
        'is_last': bool,
        'is_terminal': bool,
        'observation': FeaturesDict({
            'agentview_image': Image(shape=(84, 84, 3), dtype=uint8),
            'object': Tensor(shape=(10,), dtype=float64),
            'robot0_eef_pos': Tensor(shape=(3,), dtype=float64, description=End-effector position),
            'robot0_eef_quat': Tensor(shape=(4,), dtype=float64, description=End-effector orientation),
            'robot0_eef_vel_ang': Tensor(shape=(3,), dtype=float64, description=End-effector angular velocity),
            'robot0_eef_vel_lin': Tensor(shape=(3,), dtype=float64, description=End-effector cartesian velocity),
            'robot0_eye_in_hand_image': Image(shape=(84, 84, 3), dtype=uint8),
            'robot0_gripper_qpos': Tensor(shape=(2,), dtype=float64, description=Gripper position),
            'robot0_gripper_qvel': Tensor(shape=(2,), dtype=float64, description=Gripper velocity),
            'robot0_joint_pos': Tensor(shape=(7,), dtype=float64, description=7DOF joint positions),
            '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, description=7DOF joint velocities),
        }),
        'reward': float64,
        'states': Tensor(shape=(32,), dtype=float64),
    }),
})
  • وثائق الميزة :
ميزة فصل شكل نوع D وصف
المميزاتDict
معرف الحلقة الموتر خيط
الأفق الموتر int32
خطوات مجموعة البيانات
الخطوات/الإجراء الموتر (7،) float64
الخطوات/الخصم الموتر int32
الخطوات/is_first الموتر منطقي
الخطوات/is_last الموتر منطقي
الخطوات/is_terminal الموتر منطقي
الخطوات/الملاحظة المميزاتDict
الخطوات/الملاحظة/agentview_image صورة (84، 84، 3) uint8
الخطوات/الملاحظة/الكائن الموتر (10،) float64
الخطوات/الملاحظة/robot0_eef_pos الموتر (3،) float64 موقف المؤثر النهائي
الخطوات/الملاحظة/robot0_eef_quat الموتر (4،) float64 اتجاه المؤثر النهائي
الخطوات/الملاحظة/robot0_eef_vel_ang الموتر (3،) float64 السرعة الزاوية للمؤثر النهائي
الخطوات/الملاحظة/robot0_eef_vel_lin الموتر (3،) float64 السرعة الديكارتية للمؤثر النهائي
الخطوات/الملاحظة/robot0_eye_in_hand_image صورة (84، 84، 3) uint8
الخطوات/الملاحظة/robot0_gripper_qpos الموتر (2،) float64 موقف القابض
الخطوات/الملاحظة/robot0_gripper_qvel الموتر (2،) float64 سرعة القابض
الخطوات/الملاحظة/robot0_joint_pos الموتر (7،) float64 7DOF المواقف المشتركة
الخطوات/الملاحظة/robot0_joint_pos_cos الموتر (7،) float64
الخطوات/الملاحظة/robot0_joint_pos_sin الموتر (7،) float64
الخطوات/الملاحظة/robot0_joint_vel الموتر (7،) float64 7DOF السرعات المشتركة
خطوات/مكافأة الموتر float64
الخطوات/الدول الموتر (32،) float64

robomimic_mg/lift_mg_low_dim

  • حجم التحميل : 302.25 MiB

  • حجم مجموعة البيانات : 195.10 MiB

  • التخزين المؤقت التلقائي ( الوثائق ): فقط عندما shuffle_files=False (القطار)

  • الإنشقاقات :

ينقسم أمثلة
'train' 1500
  • هيكل الميزة :
FeaturesDict({
    'episode_id': string,
    'horizon': int32,
    'steps': Dataset({
        'action': Tensor(shape=(7,), dtype=float64),
        'discount': int32,
        'is_first': bool,
        'is_last': bool,
        'is_terminal': bool,
        'observation': FeaturesDict({
            'object': Tensor(shape=(10,), dtype=float64),
            'robot0_eef_pos': Tensor(shape=(3,), dtype=float64, description=End-effector position),
            'robot0_eef_quat': Tensor(shape=(4,), dtype=float64, description=End-effector orientation),
            'robot0_eef_vel_ang': Tensor(shape=(3,), dtype=float64, description=End-effector angular velocity),
            'robot0_eef_vel_lin': Tensor(shape=(3,), dtype=float64, description=End-effector cartesian velocity),
            'robot0_gripper_qpos': Tensor(shape=(2,), dtype=float64, description=Gripper position),
            'robot0_gripper_qvel': Tensor(shape=(2,), dtype=float64, description=Gripper velocity),
            'robot0_joint_pos': Tensor(shape=(7,), dtype=float64, description=7DOF joint positions),
            '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, description=7DOF joint velocities),
        }),
        'reward': float64,
        'states': Tensor(shape=(32,), dtype=float64),
    }),
})
  • وثائق الميزة :
ميزة فصل شكل نوع D وصف
المميزاتDict
معرف الحلقة الموتر خيط
الأفق الموتر int32
خطوات مجموعة البيانات
الخطوات/الإجراء الموتر (7،) float64
الخطوات/الخصم الموتر int32
الخطوات/is_first الموتر منطقي
الخطوات/is_last الموتر منطقي
الخطوات/is_terminal الموتر منطقي
الخطوات/الملاحظة المميزاتDict
الخطوات/الملاحظة/الكائن الموتر (10،) float64
الخطوات/الملاحظة/robot0_eef_pos الموتر (3،) float64 موقف المؤثر النهائي
الخطوات/الملاحظة/robot0_eef_quat الموتر (4،) float64 اتجاه المؤثر النهائي
الخطوات/الملاحظة/robot0_eef_vel_ang الموتر (3،) float64 السرعة الزاوية للمؤثر النهائي
الخطوات/الملاحظة/robot0_eef_vel_lin الموتر (3،) float64 السرعة الديكارتية للمؤثر النهائي
الخطوات/الملاحظة/robot0_gripper_qpos الموتر (2،) float64 موقف القابض
الخطوات/الملاحظة/robot0_gripper_qvel الموتر (2،) float64 سرعة القابض
الخطوات/الملاحظة/robot0_joint_pos الموتر (7،) float64 7DOF المواقف المشتركة
الخطوات/الملاحظة/robot0_joint_pos_cos الموتر (7،) float64
الخطوات/الملاحظة/robot0_joint_pos_sin الموتر (7،) float64
الخطوات/الملاحظة/robot0_joint_vel الموتر (7،) float64 7DOF السرعات المشتركة
خطوات/مكافأة الموتر float64
الخطوات/الدول الموتر (32،) float64

robomimic_mg/can_mg_image

  • حجم التحميل : 47.14 GiB

  • حجم مجموعة البيانات : 11.15 GiB

  • التخزين المؤقت التلقائي ( الوثائق ): لا

  • الإنشقاقات :

ينقسم أمثلة
'train' 3900
  • هيكل الميزة :
FeaturesDict({
    'episode_id': string,
    'horizon': int32,
    'steps': Dataset({
        'action': Tensor(shape=(7,), dtype=float64),
        'discount': int32,
        'is_first': bool,
        'is_last': bool,
        'is_terminal': bool,
        'observation': FeaturesDict({
            'agentview_image': Image(shape=(84, 84, 3), dtype=uint8),
            'object': Tensor(shape=(14,), dtype=float64),
            'robot0_eef_pos': Tensor(shape=(3,), dtype=float64, description=End-effector position),
            'robot0_eef_quat': Tensor(shape=(4,), dtype=float64, description=End-effector orientation),
            'robot0_eef_vel_ang': Tensor(shape=(3,), dtype=float64, description=End-effector angular velocity),
            'robot0_eef_vel_lin': Tensor(shape=(3,), dtype=float64, description=End-effector cartesian velocity),
            'robot0_eye_in_hand_image': Image(shape=(84, 84, 3), dtype=uint8),
            'robot0_gripper_qpos': Tensor(shape=(2,), dtype=float64, description=Gripper position),
            'robot0_gripper_qvel': Tensor(shape=(2,), dtype=float64, description=Gripper velocity),
            'robot0_joint_pos': Tensor(shape=(7,), dtype=float64, description=7DOF joint positions),
            '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, description=7DOF joint velocities),
        }),
        'reward': float64,
        'states': Tensor(shape=(71,), dtype=float64),
    }),
})
  • وثائق الميزة :
ميزة فصل شكل نوع D وصف
المميزاتDict
معرف الحلقة الموتر خيط
الأفق الموتر int32
خطوات مجموعة البيانات
الخطوات/الإجراء الموتر (7،) float64
الخطوات/الخصم الموتر int32
الخطوات/is_first الموتر منطقي
الخطوات/is_last الموتر منطقي
الخطوات/is_terminal الموتر منطقي
الخطوات/الملاحظة المميزاتDict
الخطوات/الملاحظة/agentview_image صورة (84، 84، 3) uint8
الخطوات/الملاحظة/الكائن الموتر (14،) float64
الخطوات/الملاحظة/robot0_eef_pos الموتر (3،) float64 موقف المؤثر النهائي
الخطوات/الملاحظة/robot0_eef_quat الموتر (4،) float64 اتجاه المؤثر النهائي
الخطوات/الملاحظة/robot0_eef_vel_ang الموتر (3،) float64 السرعة الزاوية للمؤثر النهائي
الخطوات/الملاحظة/robot0_eef_vel_lin الموتر (3،) float64 السرعة الديكارتية للمؤثر النهائي
الخطوات/الملاحظة/robot0_eye_in_hand_image صورة (84، 84، 3) uint8
الخطوات/الملاحظة/robot0_gripper_qpos الموتر (2،) float64 موقف القابض
الخطوات/الملاحظة/robot0_gripper_qvel الموتر (2،) float64 سرعة القابض
الخطوات/الملاحظة/robot0_joint_pos الموتر (7،) float64 7DOF المواقف المشتركة
الخطوات/الملاحظة/robot0_joint_pos_cos الموتر (7،) float64
الخطوات/الملاحظة/robot0_joint_pos_sin الموتر (7،) float64
الخطوات/الملاحظة/robot0_joint_vel الموتر (7،) float64 7DOF السرعات المشتركة
خطوات/مكافأة الموتر float64
الخطوات/الدول الموتر (71،) float64

robomimic_mg/can_mg_low_dim

  • حجم التحميل : 1.01 GiB

  • حجم مجموعة البيانات : 697.71 MiB

  • التخزين المؤقت التلقائي ( الوثائق ): لا

  • الإنشقاقات :

ينقسم أمثلة
'train' 3900
  • هيكل الميزة :
FeaturesDict({
    'episode_id': string,
    'horizon': int32,
    'steps': Dataset({
        'action': Tensor(shape=(7,), dtype=float64),
        'discount': int32,
        'is_first': bool,
        'is_last': bool,
        'is_terminal': bool,
        'observation': FeaturesDict({
            'object': Tensor(shape=(14,), dtype=float64),
            'robot0_eef_pos': Tensor(shape=(3,), dtype=float64, description=End-effector position),
            'robot0_eef_quat': Tensor(shape=(4,), dtype=float64, description=End-effector orientation),
            'robot0_eef_vel_ang': Tensor(shape=(3,), dtype=float64, description=End-effector angular velocity),
            'robot0_eef_vel_lin': Tensor(shape=(3,), dtype=float64, description=End-effector cartesian velocity),
            'robot0_gripper_qpos': Tensor(shape=(2,), dtype=float64, description=Gripper position),
            'robot0_gripper_qvel': Tensor(shape=(2,), dtype=float64, description=Gripper velocity),
            'robot0_joint_pos': Tensor(shape=(7,), dtype=float64, description=7DOF joint positions),
            '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, description=7DOF joint velocities),
        }),
        'reward': float64,
        'states': Tensor(shape=(71,), dtype=float64),
    }),
})
  • وثائق الميزة :
ميزة فصل شكل نوع D وصف
المميزاتDict
معرف الحلقة الموتر خيط
الأفق الموتر int32
خطوات مجموعة البيانات
الخطوات/الإجراء الموتر (7،) float64
الخطوات/الخصم الموتر int32
الخطوات/is_first الموتر منطقي
الخطوات/is_last الموتر منطقي
الخطوات/is_terminal الموتر منطقي
الخطوات/الملاحظة المميزاتDict
الخطوات/الملاحظة/الكائن الموتر (14،) float64
الخطوات/الملاحظة/robot0_eef_pos الموتر (3،) float64 موقف المؤثر النهائي
الخطوات/الملاحظة/robot0_eef_quat الموتر (4،) float64 اتجاه المؤثر النهائي
الخطوات/الملاحظة/robot0_eef_vel_ang الموتر (3،) float64 السرعة الزاوية للمؤثر النهائي
الخطوات/الملاحظة/robot0_eef_vel_lin الموتر (3،) float64 السرعة الديكارتية للمؤثر النهائي
الخطوات/الملاحظة/robot0_gripper_qpos الموتر (2،) float64 موقف القابض
الخطوات/الملاحظة/robot0_gripper_qvel الموتر (2،) float64 سرعة القابض
الخطوات/الملاحظة/robot0_joint_pos الموتر (7،) float64 7DOF المواقف المشتركة
الخطوات/الملاحظة/robot0_joint_pos_cos الموتر (7،) float64
الخطوات/الملاحظة/robot0_joint_pos_sin الموتر (7،) float64
الخطوات/الملاحظة/robot0_joint_vel الموتر (7،) float64 7DOF السرعات المشتركة
خطوات/مكافأة الموتر float64
الخطوات/الدول الموتر (71،) float64