hanguyen1907/VN_Traffic_SFT_Messages

这是一个包含文本和图像数据的对话数据集。数据集的主要特征包括消息内容和角色，其中消息内容包含列名、图像（可能为空）、文本和类型等字段。数据集分为训练集和验证集，共有9392个训练样本和2348个验证样本。数据集的总大小约为991MB。

This is a dialogue dataset containing both text and image data. The main features of the dataset include message content and role, where the message content consists of fields such as column name, image (which may be null), text, and type. The dataset is split into training and validation sets, with a total of 9392 training samples and 2348 validation samples. The total size of the dataset is approximately 991MB.