GianniBot/Zero-To-CAD-1m

Zero-to-CAD 1M 是一个大规模数据集，包含100万个可执行的、可解释的CAD构造序列，完全无需真实世界数据合成。该数据集由LLM在反馈驱动的CAD环境中生成，覆盖广泛的CAD操作词汇，包括布尔运算、圆角、倒角、抽壳、放样、扫描和阵列等。每个样本都是有效的CadQuery Python代码，可运行并生成几何体，具有命名参数、逻辑构造顺序和人类可读的代码，并经过多阶段几何验证（拓扑、连续性、复杂性、导出）。数据集包含训练、验证和测试分割，提供源代码、渲染图像、STL和STEP文件，以及预计算的嵌入和FAISS索引，适用于CAD序列模型训练、图像到CAD重建等任务。Zero-to-CAD 1M is a large-scale dataset of 1,000,000 executable, interpretable CAD construction sequences synthesized entirely without real-world data. Generated by an LLM operating inside a feedback-driven CAD environment, it covers a broad vocabulary of CAD operations including booleans, fillets, chamfers, shells, lofts, sweeps, and patterns. Each sample is executable CadQuery Python code that runs and produces geometry, with named parameters, logical construction order, human-readable code, and multi-stage geometric validation (topology, contiguity, complexity, export). The dataset includes train, validation, and test splits, providing source code, rendered images, STL and STEP files, along with precomputed embeddings and a FAISS index, intended for training CAD sequence models, image-to-CAD reconstruction, and more.