d4rl_antmaze

  • Description:

D4RL is an open-source benchmark for offline reinforcement learning. It provides standardized environments and datasets for training and benchmarking algorithms.

The datasets follow the RLDS format to represent steps and episodes.

FeaturesDict({
    'steps': Dataset({
        'action': Tensor(shape=(8,), dtype=tf.float32),
        'discount': tf.float32,
        'infos': FeaturesDict({
            'goal': Tensor(shape=(2,), dtype=tf.float32),
            'qpos': Tensor(shape=(15,), dtype=tf.float32),
            'qvel': Tensor(shape=(14,), dtype=tf.float32),
        }),
        'is_first': tf.bool,
        'is_last': tf.bool,
        'is_terminal': tf.bool,
        'observation': Tensor(shape=(29,), dtype=tf.float32),
        'reward': tf.float32,
    }),
})
  • Feature documentation:
Feature Class Shape Dtype Description
FeaturesDict
steps Dataset
steps/action Tensor (8,) tf.float32
steps/discount Tensor tf.float32
steps/infos FeaturesDict
steps/infos/goal Tensor (2,) tf.float32
steps/infos/qpos Tensor (15,) tf.float32
steps/infos/qvel Tensor (14,) tf.float32
steps/is_first Tensor tf.bool
steps/is_last Tensor tf.bool
steps/is_terminal Tensor tf.bool
steps/observation Tensor (29,) tf.float32
steps/reward Tensor tf.float32
@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_antmaze/umaze-v0 (default config)

  • Download size: 221.76 MiB

  • Dataset size: 274.83 MiB

  • Splits:

Split Examples
'train' 10,154

d4rl_antmaze/umaze-diverse-v0

  • Download size: 220.16 MiB

  • Dataset size: 270.23 MiB

  • Splits:

Split Examples
'train' 1,154

d4rl_antmaze/medium-play-v0

  • Download size: 220.40 MiB

  • Dataset size: 275.20 MiB

  • Splits:

Split Examples
'train' 10,695

d4rl_antmaze/medium-diverse-v0

  • Download size: 220.39 MiB

  • Dataset size: 271.18 MiB

  • Splits:

Split Examples
'train' 2,924

d4rl_antmaze/large-diverse-v0

  • Download size: 220.47 MiB

  • Dataset size: 273.36 MiB

  • Splits:

Split Examples
'train' 7,141

d4rl_antmaze/large-play-v0

  • Download size: 220.19 MiB

  • Dataset size: 276.61 MiB

  • Splits:

Split Examples
'train' 13,458