shijianS01/multi_templates_llava_vsft

该数据集包含两个主要特征：messages和images。messages是一个列表，包含content和role两个字段，其中content又包含index、text和type三个子字段。images字段的数据类型为图像。数据集分为训练集和测试集，训练集包含270,000个样本，测试集包含30,000个样本。数据集的下载大小为15,048,784,055字节，总大小为15,628,419,928字节。

The dataset contains two main features: messages and images. messages is a list containing two fields: content and role, where content further includes three subfields: index, text, and type. The images field has a data type of image. The dataset is divided into a training set and a test set, with the training set containing 270,000 samples and the test set containing 30,000 samples. The download size of the dataset is 15,048,784,055 bytes, and the total size is 15,628,419,928 bytes.