xcodemind/webcode2m

WebCode2M是一个包含256万个实例的数据集，旨在服务于从网页设计到代码的自动生成任务。每个实例由网页的截图、对应的网页代码和布局细节组成。该数据集从真实世界的网页中收集，并经过评分模型筛选，确保了数据的高质量。它包含了丰富的特征，如布局信息、代码文本、缩放比例、页面显示的主要语言以及代码的标记计数。

WebCode2M is a dataset comprising 2.56 million instances designed for the task of automatic code generation from webpage designs. Each instance consists of a screenshot of the webpage, the corresponding webpage code, and layout details. The dataset is sourced from real-world web pages and filtered through a scoring model to ensure high quality. It features rich attributes such as layout information, code text, scaling ratio, the primary language displayed on the rendered page, and token counts of the code.