ntudlcv/dlcv_2024_final1

该数据集是一个视觉问答（Visual Question Answering）任务的数据集，主要涉及交通相关的主题。数据集包含id、image和conversations三个主要字段，其中conversations字段是一个包含from和value子字段的列表。数据集分为训练集、验证集和测试集，分别包含28810、8716和900个示例。数据集的总大小为29904366282.25字节，下载大小为29874007166字节。

This dataset is for the Visual Question Answering task, primarily focusing on traffic-related topics. The dataset includes three main fields: id, image, and conversations, where the conversations field is a list containing subfields from and value. The dataset is divided into training, validation, and test sets, containing 28,810, 8,716, and 900 examples respectively. The total size of the dataset is 29,904,366,282.25 bytes, with a download size of 29,874,007,166 bytes.