InternSVG: Towards Unified SVG Tasks with Multimodal Large Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912889418285056 |
|---|---|
| author | Wang, Haomin Yin, Jinhui Wei, Qi Zeng, Wenguang Gu, Lixin Ye, Shenglong Gao, Zhangwei Wang, Yaohui Zhang, Yanting Li, Yuanqi Guo, Yanwen Wang, Wenhai Chen, Kai Qiao, Yu Zhang, Hongjie |
| author_facet | Wang, Haomin Yin, Jinhui Wei, Qi Zeng, Wenguang Gu, Lixin Ye, Shenglong Gao, Zhangwei Wang, Yaohui Zhang, Yanting Li, Yuanqi Guo, Yanwen Wang, Wenhai Chen, Kai Qiao, Yu Zhang, Hongjie |
| contents | General SVG modeling remains challenging due to fragmented datasets, limited transferability of methods across tasks, and the difficulty of handling structural complexity. In response, we leverage the strong transfer and generalization capabilities of multimodal large language models (MLLMs) to achieve unified modeling for SVG understanding, editing, and generation. We present the InternSVG family, an integrated data-benchmark-model suite. At its core is SAgoge, the largest and most comprehensive multimodal dataset for SVG tasks, encompassing both static graphics and dynamic animations. It covers icons, long-sequence illustrations, scientific diagrams, and dynamic animations, supporting tasks of varied difficulty levels and providing deeper hierarchies with richer attributes compared to previous datasets. Based on this resource, we introduce SArena, a companion benchmark with comprehensive task definitions and standardized evaluation that aligns with the domains and difficulty spectrum covered by SAgoge. Building on these foundations, we propose InternSVG, a unified MLLM for SVG understanding, editing, and generation with SVG-specific special tokens, subword-based embedding initialization, and a two-stage training strategy that progresses from short static SVGs to long-sequence illustrations and complex animations. This unified formulation induces positive transfer and improves overall performance. Experiments on SArena and prior benchmark confirm that InternSVG achieves substantial gains and consistently outperforms leading open and proprietary counterparts. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_11341 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | InternSVG: Towards Unified SVG Tasks with Multimodal Large Language Models Wang, Haomin Yin, Jinhui Wei, Qi Zeng, Wenguang Gu, Lixin Ye, Shenglong Gao, Zhangwei Wang, Yaohui Zhang, Yanting Li, Yuanqi Guo, Yanwen Wang, Wenhai Chen, Kai Qiao, Yu Zhang, Hongjie Computer Vision and Pattern Recognition General SVG modeling remains challenging due to fragmented datasets, limited transferability of methods across tasks, and the difficulty of handling structural complexity. In response, we leverage the strong transfer and generalization capabilities of multimodal large language models (MLLMs) to achieve unified modeling for SVG understanding, editing, and generation. We present the InternSVG family, an integrated data-benchmark-model suite. At its core is SAgoge, the largest and most comprehensive multimodal dataset for SVG tasks, encompassing both static graphics and dynamic animations. It covers icons, long-sequence illustrations, scientific diagrams, and dynamic animations, supporting tasks of varied difficulty levels and providing deeper hierarchies with richer attributes compared to previous datasets. Based on this resource, we introduce SArena, a companion benchmark with comprehensive task definitions and standardized evaluation that aligns with the domains and difficulty spectrum covered by SAgoge. Building on these foundations, we propose InternSVG, a unified MLLM for SVG understanding, editing, and generation with SVG-specific special tokens, subword-based embedding initialization, and a two-stage training strategy that progresses from short static SVGs to long-sequence illustrations and complex animations. This unified formulation induces positive transfer and improves overall performance. Experiments on SArena and prior benchmark confirm that InternSVG achieves substantial gains and consistently outperforms leading open and proprietary counterparts. |
| title | InternSVG: Towards Unified SVG Tasks with Multimodal Large Language Models |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2510.11341 |