ChavyvAkvar/Combined-SEA-Dataset-16384

该数据集包含三个主要特征：JSON元数据、令牌数量和会话信息（包括内容和角色）。会话信息由对话内容和对话角色组成。数据集划分为训练集，大小为约9.1GB，包含2,687,111个示例。数据集的配置信息表明，训练数据可以通过指定的路径进行访问。

The dataset includes three main features: JSON metadata, number of tokens, and conversation information (including content and role). The conversation information consists of dialogue content and the role of the speaker. The dataset is split into a training set, which is approximately 9.1GB in size and contains 2,687,111 examples. The configuration information of the dataset indicates that the training data can be accessed via specified paths.