CAISE

该数据集名为CAISE，包含了用户与助手之间的对话内容，其中包括了对话语句、搜索或编辑的图片以及执行的命令。该数据集对用户请求进行了分类，包括直接请求、隐含请求和对象参照请求，而助手的回应则具有不同的长度。该数据集的规模为1,611个对话和6,173个任务实例，其任务是进行对话式的图像搜索与编辑。

This dataset is named CAISE. It contains conversational content between users and assistants, including dialogue utterances, images searched or edited, and executed commands. This dataset classifies user requests into three categories: direct requests, implicit requests, and object-referring requests, while assistant responses have varying lengths. The dataset comprises 1,611 conversations and 6,173 task instances, and its targeted task is conversational image search and editing.