CGTec3/searchengine-kie

搜索引擎截图VQA资料集是一个包含搜索引擎截图和对应问题的数据集，用于视觉问答任务。数据集中的每个条目包括一个查询问题和一张截图，以及一个JSON格式的答案，其中包含了搜索内容和搜索结果的描述。为了保证数据质量，数据集只标记了搜索结果和第一笔相关结果。关键字是随机生成的，并且JSON文件的结构不遵循常见的标记结构，需要根据文件路径进行匹配。使用时需谨慎，因为数据不一定完全准确。Search Engine Screenshot VQA Dataset is a collection of search engine screenshots and corresponding questions for visual question answering tasks. Each entry in the dataset includes a query question, a screenshot, and a JSON-formatted answer containing the search content and a description of the search results. To ensure data quality, the dataset only marks the search results and the first relevant result. The keywords are randomly generated, and the JSON file structure does not follow the common marking structure, requiring matching based on the file path. Caution is advised when using the dataset as the data may not be entirely accurate.