totò

  • Descrizione :

Franka compiti di raccolta e versamento

Diviso Esempi
  • Struttura delle caratteristiche :
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),
    }),
})
  • Documentazione delle funzionalità :
Caratteristica Classe Forma Tipo D Descrizione
CaratteristicheDict
passi Set di dati
passi/azione CaratteristicheDict
passi/azione/open_gripper Tensore bool
passi/azione/rotazione_delta Tensore (3,) float32
passaggi/azione/termina_episodio Tensore float32
passi/azione/vettore_mondo Tensore (3,) float32
passi/è_primo Tensore bool
passi/è_ultimo Tensore bool
passi/è_terminale Tensore bool
passi/osservazione CaratteristicheDict
passi/osservazione/immagine Immagine (480, 640, 3) uint8
passaggi/osservazione/incorporamento_linguaggio_naturale Tensore (512,) float32
passi/osservazione/istruzione_linguaggio_naturale Tensore corda
passi/osservazione/stato Tensore (7,) float32 array di forma numpy (7,). Contiene gli stati dei giunti del robot (come angoli dei giunti assoluti) in ogni fase temporale
passi/ricompensa Scalare float32
@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},
 }