- 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.
Descripción de la configuración : vea más detalles sobre la tarea y sus versiones en https://github.com/rail-berkeley/d4rl/wiki/Tasks#gym
Página de inicio : https://sites.google.com/view/d4rl-anonymous
Código fuente :
tfds.d4rl.d4rl_mujoco_hopper.D4rlMujocoHopperVersiones :
-
1.0.0: versión inicial. -
1.1.0: Se agregó is_last. -
1.2.0(predeterminado): actualizado para tener en cuenta la siguiente observación.
-
Claves supervisadas (ver documento
as_supervised):NoneFigura ( tfds.show_examples ): no compatible.
Citación :
@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 predeterminada)
Tamaño de descarga :
51.56 MiBTamaño del conjunto de datos :
64.10 MiBAlmacenamiento en caché automático ( documentación ): Sí
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v0-medium
Tamaño de descarga :
51.74 MiBTamaño del conjunto de datos :
64.68 MiBAlmacenamiento en caché automático ( documentación ): Sí
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v0-medio-experto
Tamaño de descarga :
62.01 MiBTamaño del conjunto de datos :
77.25 MiBAlmacenamiento en caché automático ( documentación ): Sí
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v0-mixto
Tamaño de descarga :
10.48 MiBTamaño del conjunto de datos :
13.15 MiBAlmacenamiento en caché automático ( documentación ): Sí
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v0-aleatorio
Tamaño de descarga :
51.83 MiBTamaño del conjunto de datos :
66.06 MiBAlmacenamiento en caché automático ( documentación ): Sí
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v1-experto
Tamaño de descarga :
93.19 MiBTamaño del conjunto de datos :
608.03 MiBAlmacenamiento en caché automático ( documentación ): No
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| algoritmo | Tensor | cadena | ||
| iteración | Tensor | int32 | ||
| política | FuncionesDict | |||
| política/fc0 | FuncionesDict | |||
| política/fc0/sesgo | Tensor | (256,) | flotador32 | |
| política/fc0/peso | Tensor | (256, 11) | flotador32 | |
| política/fc1 | FuncionesDict | |||
| política/fc1/sesgo | Tensor | (256,) | flotador32 | |
| política/fc1/peso | Tensor | (256, 256) | flotador32 | |
| política/último_fc | FuncionesDict | |||
| política/last_fc/sesgo | Tensor | (3,) | flotador32 | |
| política/last_fc/peso | Tensor | (3, 256) | flotador32 | |
| política/last_fc_log_std | FuncionesDict | |||
| política/last_fc_log_std/sesgo | Tensor | (3,) | flotador32 | |
| política/last_fc_log_std/peso | Tensor | (3, 256) | flotador32 | |
| política/no linealidad | Tensor | cadena | ||
| política/distribución_salida | Tensor | cadena | ||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador32 | ||
| pasos/información/qpos | Tensor | (6,) | flotador32 | |
| pasos/información/qvel | Tensor | (6,) | flotador32 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v1-medio
Tamaño de descarga :
92.03 MiBTamaño del conjunto de datos :
1.78 GiBAlmacenamiento en caché automático ( documentación ): No
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| algoritmo | Tensor | cadena | ||
| iteración | Tensor | int32 | ||
| política | FuncionesDict | |||
| política/fc0 | FuncionesDict | |||
| política/fc0/sesgo | Tensor | (256,) | flotador32 | |
| política/fc0/peso | Tensor | (256, 11) | flotador32 | |
| política/fc1 | FuncionesDict | |||
| política/fc1/sesgo | Tensor | (256,) | flotador32 | |
| póliza/fc1/peso | Tensor | (256, 256) | flotador32 | |
| política/último_fc | FuncionesDict | |||
| política/last_fc/sesgo | Tensor | (3,) | flotador32 | |
| política/last_fc/peso | Tensor | (3, 256) | flotador32 | |
| política/last_fc_log_std | FuncionesDict | |||
| política/last_fc_log_std/sesgo | Tensor | (3,) | flotador32 | |
| política/last_fc_log_std/peso | Tensor | (3, 256) | flotador32 | |
| política/no linealidad | Tensor | cadena | ||
| política/distribución_salida | Tensor | cadena | ||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador32 | ||
| pasos/información/qpos | Tensor | (6,) | flotador32 | |
| pasos/información/qvel | Tensor | (6,) | flotador32 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v1-medio-experto
Tamaño de descarga :
184.59 MiBTamaño del conjunto de datos :
230.24 MiBAlmacenamiento en caché automático ( documentación ): solo cuando
shuffle_files=False(entrenamiento)Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador32 | ||
| pasos/información/qpos | Tensor | (6,) | flotador32 | |
| pasos/información/qvel | Tensor | (6,) | flotador32 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v1-repetición media
Tamaño de descarga :
55.65 MiBTamaño del conjunto de datos :
34.78 MiBAlmacenamiento en caché automático ( documentación ): Sí
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| algoritmo | Tensor | cadena | ||
| iteración | Tensor | int32 | ||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador64 | |
| pasos/descuento | Tensor | flotador64 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador64 | ||
| pasos/información/qpos | Tensor | (6,) | flotador64 | |
| pasos/información/qvel | Tensor | (6,) | flotador64 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador64 | |
| pasos/recompensa | Tensor | flotador64 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v1-repetición completa
Tamaño de descarga :
183.32 MiBTamaño del conjunto de datos :
114.78 MiBAlmacenamiento en caché automático ( documentación ): Sí
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| algoritmo | Tensor | cadena | ||
| iteración | Tensor | int32 | ||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador64 | |
| pasos/descuento | Tensor | flotador64 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador64 | ||
| pasos/información/qpos | Tensor | (6,) | flotador64 | |
| pasos/información/qvel | Tensor | (6,) | flotador64 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador64 | |
| pasos/recompensa | Tensor | flotador64 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v1-aleatorio
Tamaño de descarga :
91.11 MiBTamaño del conjunto de datos :
130.73 MiBAlmacenamiento en caché automático ( documentación ): solo cuando
shuffle_files=False(entrenamiento)Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador32 | ||
| pasos/información/qpos | Tensor | (6,) | flotador32 | |
| pasos/información/qvel | Tensor | (6,) | flotador32 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v2-experto
Tamaño de descarga :
145.37 MiBTamaño del conjunto de datos :
390.40 MiBAlmacenamiento en caché automático ( documentación ): No
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| algoritmo | Tensor | cadena | ||
| iteración | Tensor | int32 | ||
| política | FuncionesDict | |||
| política/fc0 | FuncionesDict | |||
| política/fc0/sesgo | Tensor | (256,) | flotador32 | |
| política/fc0/peso | Tensor | (256, 11) | flotador32 | |
| política/fc1 | FuncionesDict | |||
| política/fc1/sesgo | Tensor | (256,) | flotador32 | |
| póliza/fc1/peso | Tensor | (256, 256) | flotador32 | |
| política/último_fc | FuncionesDict | |||
| política/last_fc/sesgo | Tensor | (3,) | flotador32 | |
| política/last_fc/peso | Tensor | (3, 256) | flotador32 | |
| política/last_fc_log_std | FuncionesDict | |||
| política/last_fc_log_std/sesgo | Tensor | (3,) | flotador32 | |
| política/last_fc_log_std/peso | Tensor | (3, 256) | flotador32 | |
| política/no linealidad | Tensor | cadena | ||
| política/distribución_salida | Tensor | cadena | ||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador64 | ||
| pasos/información/qpos | Tensor | (6,) | flotador64 | |
| pasos/información/qvel | Tensor | (6,) | flotador64 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v2-repetición completa
Tamaño de descarga :
179.29 MiBTamaño del conjunto de datos :
115.04 MiBAlmacenamiento en caché automático ( documentación ): Sí
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| algoritmo | Tensor | cadena | ||
| iteración | Tensor | int32 | ||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador64 | ||
| pasos/información/qpos | Tensor | (6,) | flotador64 | |
| pasos/información/qvel | Tensor | (6,) | flotador64 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v2-medium
Tamaño de descarga :
145.68 MiBTamaño del conjunto de datos :
702.57 MiBAlmacenamiento en caché automático ( documentación ): No
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| algoritmo | Tensor | cadena | ||
| iteración | Tensor | int32 | ||
| política | FuncionesDict | |||
| política/fc0 | FuncionesDict | |||
| política/fc0/sesgo | Tensor | (256,) | flotador32 | |
| política/fc0/peso | Tensor | (256, 11) | flotador32 | |
| política/fc1 | FuncionesDict | |||
| política/fc1/sesgo | Tensor | (256,) | flotador32 | |
| política/fc1/peso | Tensor | (256, 256) | flotador32 | |
| política/último_fc | FuncionesDict | |||
| política/last_fc/sesgo | Tensor | (3,) | flotador32 | |
| política/last_fc/peso | Tensor | (3, 256) | flotador32 | |
| política/last_fc_log_std | FuncionesDict | |||
| política/last_fc_log_std/sesgo | Tensor | (3,) | flotador32 | |
| política/last_fc_log_std/peso | Tensor | (3, 256) | flotador32 | |
| política/no linealidad | Tensor | cadena | ||
| política/distribución_salida | Tensor | cadena | ||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador64 | ||
| pasos/información/qpos | Tensor | (6,) | flotador64 | |
| pasos/información/qvel | Tensor | (6,) | flotador64 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v2-medio-experto
Tamaño de descarga :
290.43 MiBTamaño del conjunto de datos :
228.28 MiBAlmacenamiento en caché automático ( documentación ): solo cuando
shuffle_files=False(entrenamiento)Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador64 | ||
| pasos/información/qpos | Tensor | (6,) | flotador64 | |
| pasos/información/qvel | Tensor | (6,) | flotador64 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v2-repetición media
Tamaño de descarga :
72.34 MiBTamaño del conjunto de datos :
46.51 MiBAlmacenamiento en caché automático ( documentación ): Sí
Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| algoritmo | Tensor | cadena | ||
| iteración | Tensor | int32 | ||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador64 | ||
| pasos/información/qpos | Tensor | (6,) | flotador64 | |
| pasos/información/qvel | Tensor | (6,) | flotador64 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):
d4rl_mujoco_hopper/v2-aleatorio
Tamaño de descarga :
145.46 MiBTamaño del conjunto de datos :
130.72 MiBAlmacenamiento en caché automático ( documentación ): solo cuando
shuffle_files=False(entrenamiento)Divisiones :
| Dividir | 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 funciones :
| Característica | Clase | Forma | tipo D | Descripción |
|---|---|---|---|---|
| FuncionesDict | ||||
| pasos | Conjunto de datos | |||
| pasos/acción | Tensor | (3,) | flotador32 | |
| pasos/descuento | Tensor | flotador32 | ||
| pasos/información | FuncionesDict | |||
| pasos/información/action_log_probs | Tensor | flotador64 | ||
| pasos/información/qpos | Tensor | (6,) | flotador64 | |
| pasos/información/qvel | Tensor | (6,) | flotador64 | |
| pasos/es_primero | Tensor | booleano | ||
| pasos/es_último | Tensor | booleano | ||
| pasos/es_terminal | Tensor | booleano | ||
| pasos/observación | Tensor | (11,) | flotador32 | |
| pasos/recompensa | Tensor | flotador32 |
- Ejemplos ( tfds.as_dataframe ):