MathematicianNLPer/TULU-3-50k-darija-english

该数据集是一个包含对话信息的文本数据集，每个对话包含角色和对话内容。数据集还包含对话的来源、唯一标识符、ifd分数（包括版本和内容分数）、注释信息（包括instag内容、对话轮数、CaR值和deita信息，其中deita包含质量评分和复杂性评分）以及文本内容。数据集分为训练集和验证集，提供了各自的字节数和示例数。

This dataset is a text dataset containing dialogue information, with each dialogue including role and content. The dataset also includes source of the dialogues, unique identifier, ifd score (including version and content score), annotation information (including instag content, dialogue round, CaR value, and deita information, where deita contains quality scores and complexity scores), and text content. The dataset is split into training and validation sets, with byte size and number of examples provided for each.