bitmind/conceptual_captions

该数据集包含两个主要特征：caption（描述）和image_url（图片链接），数据类型均为字符串。数据集分为训练集和验证集，训练集包含3,318,333个样本，验证集包含15,840个样本。总下载大小为375,258,690字节，数据集总大小为587,216,210字节。配置信息指定了训练集和验证集的数据文件路径。

This dataset contains text captions and corresponding image URLs, divided into a training set and a validation set. The training set has 3,318,333 samples, and the validation set has 15,840 samples. The total download size of the dataset is 375,258,690 bytes, and the total size is 587,216,210 bytes.