arnaudstiegler/vizdoom-500-episodes-skipframe-4-lvl5

该数据集包含多个特征字段，包括episode_id（剧集ID）、frames（帧数据，二进制格式）、actions（动作）、health（健康状态）和step_ids（步骤ID）。数据集被分割为训练集，包含2,470,229个样本，总大小为48,381,470,150字节。这些特征表明数据集可能与某种序列或时间序列数据相关，如视频游戏记录或模拟环境中的交互数据。

This dataset includes several feature fields such as episode_id, frames (in binary format), actions, health, and step_ids. The dataset is split into a training set containing 2,470,229 samples, with a total size of 48,381,470,150 bytes. These features suggest that the dataset may be related to some form of sequential or time-series data, such as video game recordings or interaction data in a simulated environment.