cadevolve
收藏资源简介:
CADEvolve 是一个用于 CAD 逆向工程的合成数据集,基于可执行的 CadQuery 程序构建。该数据集通过 CADEvolve 流水线生成,该流水线将一组基本图元演化为 8000 个复杂的参数化生成器,并进一步扩展为约 130 万个 CAD 程序脚本,涵盖了广泛的 CadQuery 操作集。数据集仓库包含三个主要文件夹:CADEvolve-G/ 存储参数化生成器程序和预计算的嵌入向量;CADEvolve-P/ 和 CADEvolve-C/ 则分别包含不同类别的 .py 格式 CAD 程序。该数据集适用于 CAD 逆向工程、程序合成和参数化设计等研究领域。
CADEvolve is a synthetic dataset for CAD reverse engineering, built upon executable CadQuery programs. This dataset is generated via the CADEvolve pipeline, which evolves a set of basic primitives into 8,000 complex parametric generators and further scales them into approximately 1.3 million CAD program scripts covering a wide range of CadQuery operations. The dataset repository contains three main folders: CADEvolve-G/ stores parametric generator programs and pre-computed embeddings; CADEvolve-P/ and CADEvolve-C/ respectively contain categorized .py-format CAD program scripts. This dataset is applicable to research fields such as CAD reverse engineering, program synthesis, and parametric design.



