sionic-ai/arxivqa

该数据集包含图像和对话信息，每个样本具有一个唯一标识符id，一张图片image，一段对话conversations（包括对话内容和角色），以及一些元数据metadata（如语言和来源）。数据集被划分为训练集train，共有99999个示例，大小为18801亿字节。数据集的下载大小为15761亿字节。

The dataset contains images and conversation information, with each sample having a unique identifier id, an image image, a conversation conversations (including dialogue content and role), and some metadata metadata (such as language and source). The dataset is split into a training set train, with a total of 99999 examples and a size of 18,801 billion bytes. The download size of the dataset is 15,761 billion bytes.