Unified Controllable and Faithful Text-to-CAD Generation with LLMs
5 hours ago
- #Large Language Models
- #Controllable AI
- #CAD Generation
- PR-CAD is a progressive refinement framework that unifies generation and editing for text-to-CAD modeling using large language models (LLMs).
- A high-fidelity interaction dataset is curated, covering the full CAD lifecycle with multiple representations and qualitative/quantitative descriptions.
- The framework integrates intent understanding, parameter estimation, and precise edit localization into a single reinforcement learning-enhanced agent.
- Experiments show strong mutual reinforcement between generation and editing, achieving state-of-the-art controllability and faithfulness on benchmarks.
- PR-CAD improves CAD modeling efficiency and serves as an 'all-in-one' solution for design creation and refinement.