d4rl_mujoco_hopper

  • Descripción :

D4RL es un punto de referencia de código abierto para el aprendizaje por refuerzo fuera de línea. Proporciona entornos y conjuntos de datos estandarizados para algoritmos de entrenamiento y evaluación comparativa.

Los conjuntos de datos siguen el formato RLDS para representar pasos y episodios.

@misc{fu2020d4rl,
    title
={D4RL: Datasets for Deep Data-Driven Reinforcement Learning},
    author
={Justin Fu and Aviral Kumar and Ofir Nachum and George Tucker and Sergey Levine},
    year
={2020},
    eprint
={2004.07219},
    archivePrefix
={arXiv},
    primaryClass
={cs.LG}
}

d4rl_mujoco_hopper/v0-expert (configuración por defecto)

  • Tamaño de la descarga : 51.56 MiB

  • Tamaño del conjunto de datos : 64.10 MiB

  • Almacenamiento automático en caché ( documentación ): Sí

  • Divisiones :

Separar Ejemplos
'train' 1,029
  • Estructura de características :
FeaturesDict({
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v0-medio

  • Tamaño de la descarga : 51.74 MiB

  • Tamaño del conjunto de datos : 64.68 MiB

  • Almacenamiento automático en caché ( documentación ): Sí

  • Divisiones :

Separar Ejemplos
'train' 3,064
  • Estructura de características :
FeaturesDict({
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v0-medio-experto

  • Tamaño de la descarga : 62.01 MiB

  • Tamaño del conjunto de datos : 77.25 MiB

  • Almacenamiento automático en caché ( documentación ): Sí

  • Divisiones :

Separar Ejemplos
'train' 2,277
  • Estructura de características :
FeaturesDict({
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v0-mixto

  • Tamaño de descarga : 10.48 MiB

  • Tamaño del conjunto de datos : 13.15 MiB

  • Almacenamiento automático en caché ( documentación ): Sí

  • Divisiones :

Separar Ejemplos
'train' 1,250
  • Estructura de características :
FeaturesDict({
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v0-aleatorio

  • Tamaño de la descarga : 51.83 MiB

  • Tamaño del conjunto de datos : 66.06 MiB

  • Almacenamiento automático en caché ( documentación ): Sí

  • Divisiones :

Separar Ejemplos
'train' 8,793
  • Estructura de características :
FeaturesDict({
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v1-experto

  • Tamaño de la descarga : 93.19 MiB

  • Tamaño del conjunto de datos : 608.03 MiB

  • Almacenamiento automático en caché ( documentación ): No

  • Divisiones :

Separar Ejemplos
'train' 1,836
  • Estructura de características :
FeaturesDict({
   
'algorithm': string,
   
'iteration': int32,
   
'policy': FeaturesDict({
       
'fc0': FeaturesDict({
           
'bias': Tensor(shape=(256,), dtype=float32),
           
'weight': Tensor(shape=(256, 11), dtype=float32),
       
}),
       
'fc1': FeaturesDict({
           
'bias': Tensor(shape=(256,), dtype=float32),
           
'weight': Tensor(shape=(256, 256), dtype=float32),
       
}),
       
'last_fc': FeaturesDict({
           
'bias': Tensor(shape=(3,), dtype=float32),
           
'weight': Tensor(shape=(3, 256), dtype=float32),
       
}),
       
'last_fc_log_std': FeaturesDict({
           
'bias': Tensor(shape=(3,), dtype=float32),
           
'weight': Tensor(shape=(3, 256), dtype=float32),
       
}),
       
'nonlinearity': string,
       
'output_distribution': string,
   
}),
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'infos': FeaturesDict({
           
'action_log_probs': float32,
           
'qpos': Tensor(shape=(6,), dtype=float32),
           
'qvel': Tensor(shape=(6,), dtype=float32),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
algoritmo Tensor cuerda
iteración Tensor int32
política CaracterísticasDict
política/fc0 CaracterísticasDict
política/fc0/sesgo Tensor (256,) flotar32
política/fc0/peso Tensor (256, 11) flotar32
política/fc1 CaracterísticasDict
política/fc1/sesgo Tensor (256,) flotar32
política/fc1/peso Tensor (256, 256) flotar32
política/last_fc CaracterísticasDict
política/last_fc/sesgo Tensor (3,) flotar32
política/last_fc/peso Tensor (3, 256) flotar32
política/last_fc_log_std CaracterísticasDict
política/last_fc_log_std/sesgo Tensor (3,) flotar32
política/last_fc_log_std/peso Tensor (3, 256) flotar32
política/no linealidad Tensor cuerda
política/distribución_de_salida Tensor cuerda
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar32
pasos/información/qpos Tensor (6,) flotar32
pasos/información/qvel Tensor (6,) flotar32
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v1-medio

  • Tamaño de la descarga : 92.03 MiB

  • Tamaño del conjunto de datos : 1.78 GiB

  • Almacenamiento automático en caché ( documentación ): No

  • Divisiones :

Separar Ejemplos
'train' 6,328
  • Estructura de características :
FeaturesDict({
   
'algorithm': string,
   
'iteration': int32,
   
'policy': FeaturesDict({
       
'fc0': FeaturesDict({
           
'bias': Tensor(shape=(256,), dtype=float32),
           
'weight': Tensor(shape=(256, 11), dtype=float32),
       
}),
       
'fc1': FeaturesDict({
           
'bias': Tensor(shape=(256,), dtype=float32),
           
'weight': Tensor(shape=(256, 256), dtype=float32),
       
}),
       
'last_fc': FeaturesDict({
           
'bias': Tensor(shape=(3,), dtype=float32),
           
'weight': Tensor(shape=(3, 256), dtype=float32),
       
}),
       
'last_fc_log_std': FeaturesDict({
           
'bias': Tensor(shape=(3,), dtype=float32),
           
'weight': Tensor(shape=(3, 256), dtype=float32),
       
}),
       
'nonlinearity': string,
       
'output_distribution': string,
   
}),
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'infos': FeaturesDict({
           
'action_log_probs': float32,
           
'qpos': Tensor(shape=(6,), dtype=float32),
           
'qvel': Tensor(shape=(6,), dtype=float32),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
algoritmo Tensor cuerda
iteración Tensor int32
política CaracterísticasDict
política/fc0 CaracterísticasDict
política/fc0/sesgo Tensor (256,) flotar32
política/fc0/peso Tensor (256, 11) flotar32
política/fc1 CaracterísticasDict
política/fc1/sesgo Tensor (256,) flotar32
política/fc1/peso Tensor (256, 256) flotar32
política/last_fc CaracterísticasDict
política/last_fc/sesgo Tensor (3,) flotar32
política/last_fc/peso Tensor (3, 256) flotar32
política/last_fc_log_std CaracterísticasDict
política/last_fc_log_std/sesgo Tensor (3,) flotar32
política/last_fc_log_std/peso Tensor (3, 256) flotar32
política/no linealidad Tensor cuerda
política/distribución_de_salida Tensor cuerda
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar32
pasos/información/qpos Tensor (6,) flotar32
pasos/información/qvel Tensor (6,) flotar32
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v1-medio-experto

  • Tamaño de la descarga : 184.59 MiB

  • Tamaño del conjunto de datos : 230.24 MiB

  • Auto-caché ( documentación ): Solo cuando shuffle_files=False (tren)

  • Divisiones :

Separar Ejemplos
'train' 8,163
  • Estructura de características :
FeaturesDict({
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'infos': FeaturesDict({
           
'action_log_probs': float32,
           
'qpos': Tensor(shape=(6,), dtype=float32),
           
'qvel': Tensor(shape=(6,), dtype=float32),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar32
pasos/información/qpos Tensor (6,) flotar32
pasos/información/qvel Tensor (6,) flotar32
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v1-medium-replay

  • Tamaño de la descarga : 55.65 MiB

  • Tamaño del conjunto de datos : 34.78 MiB

  • Almacenamiento automático en caché ( documentación ): Sí

  • Divisiones :

Separar Ejemplos
'train' 1,151
  • Estructura de características :
FeaturesDict({
   
'algorithm': string,
   
'iteration': int32,
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float64),
       
'discount': float64,
       
'infos': FeaturesDict({
           
'action_log_probs': float64,
           
'qpos': Tensor(shape=(6,), dtype=float64),
           
'qvel': Tensor(shape=(6,), dtype=float64),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float64),
       
'reward': float64,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
algoritmo Tensor cuerda
iteración Tensor int32
pasos conjunto de datos
pasos/acción Tensor (3,) flotar64
pasos/descuento Tensor flotar64
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar64
pasos/información/qpos Tensor (6,) flotar64
pasos/información/qvel Tensor (6,) flotar64
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar64
pasos/recompensa Tensor flotar64

d4rl_mujoco_hopper/v1-full-replay

  • Tamaño de la descarga : 183.32 MiB

  • Tamaño del conjunto de datos : 114.78 MiB

  • Almacenamiento automático en caché ( documentación ): Sí

  • Divisiones :

Separar Ejemplos
'train' 2,907
  • Estructura de características :
FeaturesDict({
   
'algorithm': string,
   
'iteration': int32,
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float64),
       
'discount': float64,
       
'infos': FeaturesDict({
           
'action_log_probs': float64,
           
'qpos': Tensor(shape=(6,), dtype=float64),
           
'qvel': Tensor(shape=(6,), dtype=float64),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float64),
       
'reward': float64,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
algoritmo Tensor cuerda
iteración Tensor int32
pasos conjunto de datos
pasos/acción Tensor (3,) flotar64
pasos/descuento Tensor flotar64
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar64
pasos/información/qpos Tensor (6,) flotar64
pasos/información/qvel Tensor (6,) flotar64
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar64
pasos/recompensa Tensor flotar64

d4rl_mujoco_hopper/v1-aleatorio

  • Tamaño de la descarga : 91.11 MiB

  • Tamaño del conjunto de datos : 130.73 MiB

  • Auto-caché ( documentación ): Solo cuando shuffle_files=False (tren)

  • Divisiones :

Separar Ejemplos
'train' 45,265
  • Estructura de características :
FeaturesDict({
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'infos': FeaturesDict({
           
'action_log_probs': float32,
           
'qpos': Tensor(shape=(6,), dtype=float32),
           
'qvel': Tensor(shape=(6,), dtype=float32),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar32
pasos/información/qpos Tensor (6,) flotar32
pasos/información/qvel Tensor (6,) flotar32
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v2-experto

  • Tamaño de la descarga : 145.37 MiB

  • Tamaño del conjunto de datos : 390.40 MiB

  • Almacenamiento automático en caché ( documentación ): No

  • Divisiones :

Separar Ejemplos
'train' 1,028
  • Estructura de características :
FeaturesDict({
   
'algorithm': string,
   
'iteration': int32,
   
'policy': FeaturesDict({
       
'fc0': FeaturesDict({
           
'bias': Tensor(shape=(256,), dtype=float32),
           
'weight': Tensor(shape=(256, 11), dtype=float32),
       
}),
       
'fc1': FeaturesDict({
           
'bias': Tensor(shape=(256,), dtype=float32),
           
'weight': Tensor(shape=(256, 256), dtype=float32),
       
}),
       
'last_fc': FeaturesDict({
           
'bias': Tensor(shape=(3,), dtype=float32),
           
'weight': Tensor(shape=(3, 256), dtype=float32),
       
}),
       
'last_fc_log_std': FeaturesDict({
           
'bias': Tensor(shape=(3,), dtype=float32),
           
'weight': Tensor(shape=(3, 256), dtype=float32),
       
}),
       
'nonlinearity': string,
       
'output_distribution': string,
   
}),
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'infos': FeaturesDict({
           
'action_log_probs': float64,
           
'qpos': Tensor(shape=(6,), dtype=float64),
           
'qvel': Tensor(shape=(6,), dtype=float64),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
algoritmo Tensor cuerda
iteración Tensor int32
política CaracterísticasDict
política/fc0 CaracterísticasDict
política/fc0/sesgo Tensor (256,) flotar32
política/fc0/peso Tensor (256, 11) flotar32
política/fc1 CaracterísticasDict
política/fc1/sesgo Tensor (256,) flotar32
política/fc1/peso Tensor (256, 256) flotar32
política/last_fc CaracterísticasDict
política/last_fc/sesgo Tensor (3,) flotar32
política/last_fc/peso Tensor (3, 256) flotar32
política/last_fc_log_std CaracterísticasDict
política/last_fc_log_std/sesgo Tensor (3,) flotar32
política/last_fc_log_std/peso Tensor (3, 256) flotar32
política/no linealidad Tensor cuerda
política/distribución_de_salida Tensor cuerda
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar64
pasos/información/qpos Tensor (6,) flotar64
pasos/información/qvel Tensor (6,) flotar64
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v2-full-replay

  • Tamaño de la descarga : 179.29 MiB

  • Tamaño del conjunto de datos : 115.04 MiB

  • Almacenamiento automático en caché ( documentación ): Sí

  • Divisiones :

Separar Ejemplos
'train' 3,515
  • Estructura de características :
FeaturesDict({
   
'algorithm': string,
   
'iteration': int32,
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'infos': FeaturesDict({
           
'action_log_probs': float64,
           
'qpos': Tensor(shape=(6,), dtype=float64),
           
'qvel': Tensor(shape=(6,), dtype=float64),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
algoritmo Tensor cuerda
iteración Tensor int32
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar64
pasos/información/qpos Tensor (6,) flotar64
pasos/información/qvel Tensor (6,) flotar64
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v2-medio

  • Tamaño de la descarga : 145.68 MiB

  • Tamaño del conjunto de datos : 702.57 MiB

  • Almacenamiento automático en caché ( documentación ): No

  • Divisiones :

Separar Ejemplos
'train' 2,187
  • Estructura de características :
FeaturesDict({
   
'algorithm': string,
   
'iteration': int32,
   
'policy': FeaturesDict({
       
'fc0': FeaturesDict({
           
'bias': Tensor(shape=(256,), dtype=float32),
           
'weight': Tensor(shape=(256, 11), dtype=float32),
       
}),
       
'fc1': FeaturesDict({
           
'bias': Tensor(shape=(256,), dtype=float32),
           
'weight': Tensor(shape=(256, 256), dtype=float32),
       
}),
       
'last_fc': FeaturesDict({
           
'bias': Tensor(shape=(3,), dtype=float32),
           
'weight': Tensor(shape=(3, 256), dtype=float32),
       
}),
       
'last_fc_log_std': FeaturesDict({
           
'bias': Tensor(shape=(3,), dtype=float32),
           
'weight': Tensor(shape=(3, 256), dtype=float32),
       
}),
       
'nonlinearity': string,
       
'output_distribution': string,
   
}),
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'infos': FeaturesDict({
           
'action_log_probs': float64,
           
'qpos': Tensor(shape=(6,), dtype=float64),
           
'qvel': Tensor(shape=(6,), dtype=float64),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
algoritmo Tensor cuerda
iteración Tensor int32
política CaracterísticasDict
política/fc0 CaracterísticasDict
política/fc0/sesgo Tensor (256,) flotar32
política/fc0/peso Tensor (256, 11) flotar32
política/fc1 CaracterísticasDict
política/fc1/sesgo Tensor (256,) flotar32
política/fc1/peso Tensor (256, 256) flotar32
política/last_fc CaracterísticasDict
política/last_fc/sesgo Tensor (3,) flotar32
política/last_fc/peso Tensor (3, 256) flotar32
política/last_fc_log_std CaracterísticasDict
política/last_fc_log_std/sesgo Tensor (3,) flotar32
política/last_fc_log_std/peso Tensor (3, 256) flotar32
política/no linealidad Tensor cuerda
política/distribución_de_salida Tensor cuerda
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar64
pasos/información/qpos Tensor (6,) flotar64
pasos/información/qvel Tensor (6,) flotar64
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v2-medio-experto

  • Tamaño de la descarga : 290.43 MiB

  • Tamaño del conjunto de datos : 228.28 MiB

  • Auto-caché ( documentación ): Solo cuando shuffle_files=False (tren)

  • Divisiones :

Separar Ejemplos
'train' 3,214
  • Estructura de características :
FeaturesDict({
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'infos': FeaturesDict({
           
'action_log_probs': float64,
           
'qpos': Tensor(shape=(6,), dtype=float64),
           
'qvel': Tensor(shape=(6,), dtype=float64),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar64
pasos/información/qpos Tensor (6,) flotar64
pasos/información/qvel Tensor (6,) flotar64
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v2-medium-replay

  • Tamaño de la descarga : 72.34 MiB

  • Tamaño del conjunto de datos : 46.51 MiB

  • Almacenamiento automático en caché ( documentación ): Sí

  • Divisiones :

Separar Ejemplos
'train' 2,041
  • Estructura de características :
FeaturesDict({
   
'algorithm': string,
   
'iteration': int32,
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'infos': FeaturesDict({
           
'action_log_probs': float64,
           
'qpos': Tensor(shape=(6,), dtype=float64),
           
'qvel': Tensor(shape=(6,), dtype=float64),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
algoritmo Tensor cuerda
iteración Tensor int32
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar64
pasos/información/qpos Tensor (6,) flotar64
pasos/información/qvel Tensor (6,) flotar64
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32

d4rl_mujoco_hopper/v2-aleatorio

  • Tamaño de la descarga : 145.46 MiB

  • Tamaño del conjunto de datos : 130.72 MiB

  • Auto-caché ( documentación ): Solo cuando shuffle_files=False (tren)

  • Divisiones :

Separar Ejemplos
'train' 45,240
  • Estructura de características :
FeaturesDict({
   
'steps': Dataset({
       
'action': Tensor(shape=(3,), dtype=float32),
       
'discount': float32,
       
'infos': FeaturesDict({
           
'action_log_probs': float64,
           
'qpos': Tensor(shape=(6,), dtype=float64),
           
'qvel': Tensor(shape=(6,), dtype=float64),
       
}),
       
'is_first': bool,
       
'is_last': bool,
       
'is_terminal': bool,
       
'observation': Tensor(shape=(11,), dtype=float32),
       
'reward': float32,
   
}),
})
  • Documentación de características :
Rasgo Clase Forma Tipo D Descripción
CaracterísticasDict
pasos conjunto de datos
pasos/acción Tensor (3,) flotar32
pasos/descuento Tensor flotar32
pasos/información CaracterísticasDict
pasos/infos/action_log_probs Tensor flotar64
pasos/información/qpos Tensor (6,) flotar64
pasos/información/qvel Tensor (6,) flotar64
pasos/es_primero Tensor bool
pasos/es_último Tensor bool
pasos/es_terminal Tensor bool
pasos/observación Tensor (11,) flotar32
pasos/recompensa Tensor flotar32