সম্পূর্ণ

  • বর্ণনা :

ফ্রাঙ্কা স্কুপিং এবং ঢালা কাজ

  • হোমপেজ : https://toto-benchmark.org/

  • সোর্স কোড : tfds.robotics.rtx.Toto

  • সংস্করণ :

    • 0.1.0 (ডিফল্ট): প্রাথমিক প্রকাশ।
  • ডাউনলোড আকার : Unknown size

  • ডেটাসেটের আকার : Unknown size

  • স্বয়ংক্রিয় ক্যাশে ( ডকুমেন্টেশন ): অজানা

  • বিভাজন :

বিভক্ত উদাহরণ
  • বৈশিষ্ট্য গঠন :
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),
    }),
})
  • বৈশিষ্ট্য ডকুমেন্টেশন :
বৈশিষ্ট্য ক্লাস আকৃতি ডিটাইপ বর্ণনা
ফিচারসডিক্ট
পদক্ষেপ ডেটাসেট
পদক্ষেপ/ক্রিয়া ফিচারসডিক্ট
পদক্ষেপ/অ্যাকশন/ওপেন_গ্রিপার টেনসর bool
পদক্ষেপ/ক্রিয়া/ঘূর্ণন_ডেল্টা টেনসর (৩,) float32
steps/action/terminate_episode টেনসর float32
পদক্ষেপ/ক্রিয়া/বিশ্ব_ভেক্টর টেনসর (৩,) float32
steps/is_first টেনসর bool
ধাপ/শেষ_শেষ টেনসর bool
steps/is_terminal টেনসর bool
পদক্ষেপ/পর্যবেক্ষণ ফিচারসডিক্ট
পদক্ষেপ/পর্যবেক্ষণ/চিত্র ছবি (480, 640, 3) uint8
পদক্ষেপ/পর্যবেক্ষণ/প্রাকৃতিক_ভাষা_এম্বেডিং টেনসর (512,) float32
পদক্ষেপ/পর্যবেক্ষণ/প্রাকৃতিক_ভাষা_নির্দেশ টেনসর স্ট্রিং
পদক্ষেপ/পর্যবেক্ষণ/রাষ্ট্র টেনসর (৭,) float32 আকৃতির নমপি অ্যারে (7,)। প্রতিটি টাইমস্টেপে রোবট জয়েন্ট স্টেট (পরম যৌথ কোণ হিসাবে) ধারণ করে
পদক্ষেপ/পুরস্কার স্কেলার float32
  • তত্ত্বাবধান করা কী (দেখুন as_supervised doc ): None

  • চিত্র ( tfds.show_examples ): সমর্থিত নয়।

  • উদাহরণ ( tfds.as_dataframe ): অনুপস্থিত।

  • উদ্ধৃতি :

@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},
 }