asu_table_top_converted_externally_to_rlds

  • বর্ণনা :

UR5 টেবিল-টপ পিক/প্লেস/ঘোরানো কাজগুলি সম্পাদন করছে

বিভক্ত উদাহরণ
'train' 110
  • বৈশিষ্ট্য গঠন :
FeaturesDict({
    'episode_metadata': FeaturesDict({
        'file_path': Text(shape=(), dtype=string),
    }),
    'steps': Dataset({
        'action': Tensor(shape=(7,), dtype=float32),
        'action_delta': Tensor(shape=(7,), dtype=float32),
        'action_inst': Text(shape=(), dtype=string),
        'discount': Scalar(shape=(), dtype=float32),
        'goal_object': Text(shape=(), dtype=string),
        'ground_truth_states': FeaturesDict({
            'EE': Tensor(shape=(6,), dtype=float32),
            'bottle': Tensor(shape=(6,), dtype=float32),
            'bread': Tensor(shape=(6,), dtype=float32),
            'coke': Tensor(shape=(6,), dtype=float32),
            'cube': Tensor(shape=(6,), dtype=float32),
            'milk': Tensor(shape=(6,), dtype=float32),
            'pepsi': Tensor(shape=(6,), dtype=float32),
        }),
        'is_first': bool,
        'is_last': bool,
        'is_terminal': bool,
        'language_embedding': Tensor(shape=(512,), dtype=float32),
        'language_instruction': Text(shape=(), dtype=string),
        'observation': FeaturesDict({
            'image': Image(shape=(224, 224, 3), dtype=uint8),
            'state': Tensor(shape=(7,), dtype=float32),
            'state_vel': Tensor(shape=(7,), dtype=float32),
        }),
        'reward': Scalar(shape=(), dtype=float32),
    }),
})
  • বৈশিষ্ট্য ডকুমেন্টেশন :
বৈশিষ্ট্য ক্লাস আকৃতি ডিটাইপ বর্ণনা
ফিচারসডিক্ট
episode_metadata ফিচারসডিক্ট
episode_metadata/file_path পাঠ্য স্ট্রিং মূল ডেটা ফাইলের পথ।
পদক্ষেপ ডেটাসেট
পদক্ষেপ/ক্রিয়া টেনসর (৭,) float32 রোবট কর্ম, [7x যৌথ বেগ, 2x গ্রিপার বেগ, 1x সমাপ্ত পর্ব] নিয়ে গঠিত।
পদক্ষেপ/অ্যাকশন_ডেল্টা টেনসর (৭,) float32 রোবট ডেল্টা অ্যাকশন, [7x জয়েন্ট বেগ, 2x গ্রিপার বেগ, 1x টার্মিনেট পর্ব] নিয়ে গঠিত।
steps/action_inst পাঠ্য স্ট্রিং কর্ম সঞ্চালিত করা.
পদক্ষেপ/ছাড় স্কেলার float32 ডিসকাউন্ট দেওয়া হলে, ডিফল্ট 1.
পদক্ষেপ/লক্ষ্য_বস্তু পাঠ্য স্ট্রিং সঙ্গে হেরফের করা বস্তু.
পদক্ষেপ/স্থল_সত্য_রাষ্ট্র ফিচারসডিক্ট
steps/ground_truth_states/EE টেনসর (6,) float32 xyzrpy
steps/ground_truth_states/বোতল টেনসর (6,) float32 xyzrpy
পদক্ষেপ/স্থল_সত্য_রাষ্ট্র/রুটি টেনসর (6,) float32 xyzrpy
পদক্ষেপ/গ্রাউন্ড_সত্য_রাষ্ট্র/কোক টেনসর (6,) float32 xyzrpy
steps/ground_truth_states/cube টেনসর (6,) float32 xyzrpy
পদক্ষেপ/স্থল_সত্য_রাষ্ট্র/দুধ টেনসর (6,) float32 xyzrpy
steps/ground_truth_states/pepsi টেনসর (6,) float32 xyzrpy
steps/is_first টেনসর bool
ধাপ/শেষ_শেষ টেনসর bool
steps/is_terminal টেনসর bool
পদক্ষেপ/ভাষা_এম্বেডিং টেনসর (512,) float32 কোন ভাষা এম্বেডিং. https://tfhub.dev/google/universal-sentence-encoder-large/5 দেখুন
পদক্ষেপ/ভাষা_নির্দেশ পাঠ্য স্ট্রিং ভাষার নির্দেশনা।
পদক্ষেপ/পর্যবেক্ষণ ফিচারসডিক্ট
পদক্ষেপ/পর্যবেক্ষণ/চিত্র ছবি (224, 224, 3) uint8 প্রধান ক্যামেরা আরজিবি পর্যবেক্ষণ।
পদক্ষেপ/পর্যবেক্ষণ/রাষ্ট্র টেনসর (৭,) float32 রোবট অবস্থা, [6x রোবট যুগ্ম কোণ, 1x গ্রিপার অবস্থান] নিয়ে গঠিত।
পদক্ষেপ/পর্যবেক্ষণ/state_vel টেনসর (৭,) float32 রোবট যুগ্ম বেগ, [6x রোবট যুগ্ম কোণ, 1x গ্রিপার অবস্থান] নিয়ে গঠিত।
পদক্ষেপ/পুরস্কার স্কেলার float32 প্রদান করা হলে পুরস্কার, ডেমোর জন্য চূড়ান্ত ধাপে 1।
  • উদ্ধৃতি :
@inproceedings{zhou2023modularity,
  title={Modularity through Attention: Efficient Training and Transfer of Language-Conditioned Policies for Robot Manipulation},
  author={Zhou, Yifan and Sonawani, Shubham and Phielipp, Mariano and Stepputtis, Simon and Amor, Heni},
  booktitle={Conference on Robot Learning},
  pages={1684--1695},
  year={2023},
  organization={PMLR}
}
@article{zhou2023learning,
  title={Learning modular language-conditioned robot policies through attention},
  author={Zhou, Yifan and Sonawani, Shubham and Phielipp, Mariano and Ben Amor, Heni and Stepputtis, Simon},
  journal={Autonomous Robots},
  pages={1--21},
  year={2023},
  publisher={Springer}
}