AlekseyKorshuk/soda-chatml

该数据集包含一个名为conversation的特征，该特征是一个列表，列表中包含三个字段：content（内容，数据类型为字符串）、do_train（是否用于训练，数据类型为布尔值）和role（角色，数据类型为字符串）。数据集包含一个训练集（train），其大小为1405384981字节，包含1486896个样本。数据集的下载大小为780354430字节，总大小为1405384981字节。

The dataset contains a feature named conversation, which is a list containing three fields: content (content, data type is string), do_train (whether to use for training, data type is boolean), and role (role, data type is string). The dataset includes a training set (train) with a size of 1405384981 bytes, containing 1486896 samples. The download size of the dataset is 780354430 bytes, and the total size is 1405384981 bytes.