CADEvolve：通过程序演化生成逼真的CAD模型

摘要

计算机辅助设计（CAD）为工程制造领域提供了可快速编辑的建模方案。随着人工智能技术的进步，各类CAD任务现已能够实现全自动化。然而，数据瓶颈制约了发展进程：现有公共数据集大多仅包含草图拉伸序列，缺乏复杂操作、多操作组合及设计意图，导致模型微调效果受限。当前尝试通过冻结视觉语言模型规避此问题的方法，由于基础模型对三维空间理解有限，往往只能生成简单或无效程序。我们提出CADEvolve——一种基于演化机制的流程与数据集，该方法从简单几何基元出发，通过VLM引导的编辑验证机制，逐步构建出具备工业级复杂度的CAD程序。最终生成8,000个以可执行CadQuery参数化生成器表达的复杂零件。经过多阶段后处理与数据增强，我们获得了包含130万条脚本的统一数据集，每条脚本均配有渲染几何体并完整覆盖CadQuery操作集。基于CADEvolve微调的VLM在Image2CAD任务中，于DeepCAD、Fusion 360和MCB三大基准测试上均取得了最先进的性能表现。

English

Computer-Aided Design (CAD) delivers rapid, editable modeling for engineering and manufacturing. Recent AI progress now makes full automation feasible for various CAD tasks. However, progress is bottlenecked by data: public corpora mostly contain sketch-extrude sequences, lack complex operations, multi-operation composition and design intent, and thus hinder effective fine-tuning. Attempts to bypass this with frozen VLMs often yield simple or invalid programs due to limited 3D grounding in current foundation models. We present CADEvolve, an evolution-based pipeline and dataset that starts from simple primitives and, via VLM-guided edits and validations, incrementally grows CAD programs toward industrial-grade complexity. The result is 8k complex parts expressed as executable CadQuery parametric generators. After multi-stage post-processing and augmentation, we obtain a unified dataset of 1.3m scripts paired with rendered geometry and exercising the full CadQuery operation set. A VLM fine-tuned on CADEvolve achieves state-of-the-art results on the Image2CAD task across the DeepCAD, Fusion 360, and MCB benchmarks.

CADEvolve：通过程序演化生成逼真的CAD模型

CADEvolve: Creating Realistic CAD via Program Evolution

摘要

Support