EpicPinkPenguin/dm_lab

该数据集包含由Dreamer V3强化学习代理生成的专家轨迹，这些轨迹是在DMLab-30中的30个环境中训练的。每个环境都有自己的配置、特征和数据分割，用于训练和测试。数据集特征包括图像观测、动作、奖励和终止标志。README还提供了关于代理的性能指标、训练步数以及如何使用Python中的load_dataset函数使用数据集的信息。

This dataset contains expert trajectories generated by a Dreamer V3 reinforcement learning agent, trained on each of the 30 environments in DMLab-30. Each environment has its own configuration, features, and data splits for training and testing. The dataset features include image observations, actions, rewards, and termination flags. The README also provides information about the agents performance metrics, training step sizes, and how to use the dataset in Python using the load_dataset function.