टोटो

  • विवरण :

फ्रेंका स्कूपिंग और डालना कार्य

विभाजित करना उदाहरण
  • फ़ीचर संरचना :
FeaturesDict({
    'steps': Dataset({
        'action': FeaturesDict({
            'open_gripper': bool,
            'rotation_delta': Tensor(shape=(3,), dtype=float32),
            'terminate_episode': float32,
            'world_vector': Tensor(shape=(3,), dtype=float32),
        }),
        'is_first': bool,
        'is_last': bool,
        'is_terminal': bool,
        'observation': FeaturesDict({
            'image': Image(shape=(480, 640, 3), dtype=uint8),
            'natural_language_embedding': Tensor(shape=(512,), dtype=float32),
            'natural_language_instruction': string,
            'state': Tensor(shape=(7,), dtype=float32),
        }),
        'reward': Scalar(shape=(), dtype=float32),
    }),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीप्रकार विवरण
फीचर्सडिक्ट
कदम डेटासेट
कदम/कार्रवाई फीचर्सडिक्ट
चरण/कार्रवाई/ओपन_ग्रिपर टेन्सर बूल
चरण/क्रिया/रोटेशन_डेल्टा टेन्सर (3,) फ्लोट32
चरण/कार्रवाई/समाप्त_एपिसोड टेन्सर फ्लोट32
चरण/कार्रवाई/विश्व_वेक्टर टेन्सर (3,) फ्लोट32
चरण/पहला है टेन्सर बूल
चरण/अंतिम है टेन्सर बूल
चरण/is_terminal टेन्सर बूल
चरण/अवलोकन फीचर्सडिक्ट
चरण/अवलोकन/छवि छवि (480, 640, 3) uint8
चरण/अवलोकन/प्राकृतिक_भाषा_एम्बेडिंग टेन्सर (512,) फ्लोट32
चरण/अवलोकन/प्राकृतिक_भाषा_निर्देश टेन्सर डोरी
चरण/अवलोकन/स्थिति टेन्सर (7,) फ्लोट32 आकार की सुन्न सरणी (7,)। प्रत्येक टाइमस्टेप पर रोबोट संयुक्त स्थिति (पूर्ण संयुक्त कोण के रूप में) शामिल है
कदम/इनाम अदिश फ्लोट32
  • पर्यवेक्षित कुंजियाँ ( as_supervised doc देखें): None

  • चित्र ( tfds.show_examples ): समर्थित नहीं है।

  • उदाहरण ( tfds.as_dataframe ): गुम है।

  • उद्धरण :

@inproceedings{zhou2023train,
  author={Zhou, Gaoyue and Dean, Victoria and Srirama, Mohan Kumar and Rajeswaran, Aravind and Pari, Jyothish and Hatch, Kyle and Jain, Aryan and Yu, Tianhe and Abbeel, Pieter and Pinto, Lerrel and Finn, Chelsea and Gupta, Abhinav},
  booktitle={2023 IEEE International Conference on Robotics and Automation (ICRA)},
  title={Train Offline, Test Online: A Real Robot Learning Benchmark},
  year={2023},
 }