अलोहा_मोबाइल

  • विवरण :

वास्तविक डेटासेट. मोबाइल हेरफेर कार्यों का अनुकरण करना जो द्वि-मैनुअल हैं और पूरे शरीर पर नियंत्रण की आवश्यकता होती है। प्रत्येक कार्य के लिए 50 प्रदर्शन।

विभाजित करना उदाहरण
'train' 276
  • फ़ीचर संरचना :
FeaturesDict({
    'episode_metadata': FeaturesDict({
        'file_path': string,
    }),
    'steps': Dataset({
        'action': Tensor(shape=(16,), dtype=float32),
        'discount': Scalar(shape=(), dtype=float32),
        'is_first': bool,
        'is_last': bool,
        'is_terminal': bool,
        'language_instruction': string,
        'observation': FeaturesDict({
            'cam_high': Image(shape=(480, 640, 3), dtype=uint8),
            'cam_left_wrist': Image(shape=(480, 640, 3), dtype=uint8),
            'cam_right_wrist': Image(shape=(480, 640, 3), dtype=uint8),
            'state': Tensor(shape=(14,), dtype=float32),
        }),
        'reward': Scalar(shape=(), dtype=float32),
    }),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीप्रकार विवरण
फीचर्सडिक्ट
एपिसोड_मेटाडेटा फीचर्सडिक्ट
एपिसोड_मेटाडेटा/फ़ाइल_पथ टेन्सर डोरी
कदम डेटासेट
कदम/कार्रवाई टेन्सर (16,) फ्लोट32
कदम/छूट अदिश फ्लोट32
चरण/पहला है टेन्सर बूल
चरण/अंतिम है टेन्सर बूल
चरण/is_terminal टेन्सर बूल
चरण/भाषा_निर्देश टेन्सर डोरी
चरण/अवलोकन फीचर्सडिक्ट
चरण/अवलोकन/cam_high छवि (480, 640, 3) uint8
चरण/अवलोकन/कैम_बाएं_कलाई छवि (480, 640, 3) uint8
चरण/अवलोकन/cam_right_wrist छवि (480, 640, 3) uint8
चरण/अवलोकन/स्थिति टेन्सर (14,) फ्लोट32
कदम/इनाम अदिश फ्लोट32
  • उद्धरण :
@inproceedings{fu2024mobile,author = {Fu, Zipeng and Zhao, Tony Z. and Finn, Chelsea},title = {Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperation},booktitle = {arXiv},year = {2024},}