SVGen: Interpretable Vector Graphics Generation with Large Language Models
This addresses a time-consuming problem for front-end developers and UI/UX designers by enabling faster creation of scalable and editable vector graphics from text descriptions.
The paper tackles the challenge of generating precise vector graphics from natural language by introducing SVGen, an end-to-end model that uses a large-scale dataset (SVG-1M) with text-SVG pairs and Chain of Thought annotations, resulting in outperforming general large models and traditional rendering methods in effectiveness and efficiency.
Scalable Vector Graphics (SVG) is widely used in front-end development and UI/UX design due to its scalability, editability, and rendering efficiency. However, turning creative ideas into precise vector graphics remains a time-consuming challenge. To address this, we introduce SVG-1M, a large-scale dataset of high-quality SVGs paired with natural language descriptions. Through advanced data augmentation and annotation, we create well-aligned Text to SVG training pairs, including a subset with Chain of Thought annotations for enhanced semantic guidance. Based on this dataset, we propose SVGen, an end-to-end model that generates SVG code from natural language inputs. Our approach ensures semantic accuracy and structural completeness, supported by curriculum learning and reinforcement learning optimization. Experiments show that SVGen outperforms general large models and traditional rendering methods in both effectiveness and efficiency. Code, model, and dataset are available on GitHub.