chenhegu/geo3k_imgurl

该数据集包含提示(prompt)、图像(images)和标签(label)三个特征。提示特征由内容(content)和角色(role)组成，均为字符串类型。图像特征是一系列字符串，可能指向图像文件。标签特征是字符串类型，可能用于表示图像的分类。数据集分为训练集和测试集，共有2101个训练示例和601个测试示例。

The dataset consists of three features: prompt, images, and label. The prompt feature includes content and role, both of which are string types. The images feature is a sequence of strings, possibly pointing to image files. The label feature is a string type, which may be used to represent the classification of the images. The dataset is split into a training set and a test set, with a total of 2101 training examples and 601 test examples.