MMSVG-Icon
收藏魔搭社区2026-01-06 更新2025-04-12 收录
下载链接:
https://modelscope.cn/datasets/AI-ModelScope/MMSVG-Icon
下载链接
链接失效反馈官方服务:
资源简介:
<h1>OmniSVG: A Unified Scalable Vector Graphics Generation Model</h1>
[![Project Page]](https://omnisvg.github.io/)
# Dataset Card for MMSVG-Icon
## Dataset Description
This dataset contains SVG icon examples for training and evaluating SVG models for text-to-SVG and image-to-SVG task.
## Dataset Structure
### Features
The dataset contains the following fields:
| Field Name | Description |
| :--------- | :---------- |
| `id` | Unique ID for each SVG |
| `svg` | SVG code (resized to 200×200, simplified with picosvg) |
| `description` | Description of the SVG |
| `keywords` | Keywords associated with the SVG |
| `detail` | Detailed description of the SVG |
| `image` | PNG image preview (resized to 448×448) |
| `token_len` | Token length (OmniSVG tokenizer) |
## Changelog
### v2.0 (2025-12-22)
- **Data Volume**: Increased from 264,360 to 904,011 samples
- **Deduplication**: Removed duplicate SVGs based on MD5 hash (keeping only one instance per unique SVG)
- **Added PNG Previews**: Each SVG now includes a corresponding PNG image
- **Enhanced Captions**: Increased caption diversity with `description`, `keywords`, and `detail` fields
- **SVG Normalization**: All SVGs resized to 200×200 and simplified using [picosvg](https://github.com/nicothin/picosvg)
- **PNG Standardization**: All PNG images resized to 448×448 for uniform input size
## Citation
```bibtex
@article{yang2025omnisvg,
title={OmniSVG: A Unified Scalable Vector Graphics Generation Model},
author={Yiying Yang and Wei Cheng and Sijin Chen and Xianfang Zeng and Jiaxu Zhang and Liao Wang and Gang Yu and Xinjun Ma and Yu-Gang Jiang},
journal={arXiv preprint arxiv:2504.06263},
year={2025}
}
```
## Tags
- scalable vector graphics (SVG)
- vision language models
- multimodal
- Icon
# OmniSVG:统一化可缩放矢量图形生成模型
[项目页面](https://omnisvg.github.io/)
## MMSVG-Icon 数据集卡片
## 数据集说明
本数据集包含可缩放矢量图形(Scalable Vector Graphics,SVG)图标示例,用于训练和评估面向文本转SVG、图像转SVG任务的SVG模型。
## 数据集结构
### 特征字段
| 字段名 | 描述 |
| :--------- | :---------- |
| `id` | 每个SVG的唯一标识符 |
| `svg` | SVG代码(已调整尺寸至200×200,使用picosvg进行简化) |
| `description` | 该SVG的描述文本 |
| `keywords` | 与该SVG相关的关键词 |
| `detail` | 该SVG的详细描述文本 |
| `image` | PNG格式预览图(已调整尺寸至448×448) |
| `token_len` | Token长度(使用OmniSVG分词器计算) |
## 更新日志
### v2.0(2025年12月22日)
- **数据规模**:样本量从264,360条提升至904,011条
- **去重处理**:基于MD5哈希值移除重复SVG,每个唯一SVG仅保留一条实例
- **新增PNG预览图**:每个SVG现已附带对应的PNG格式预览图
- **增强文本标注**:新增`description`、`keywords`与`detail`字段,提升标注多样性
- **SVG标准化**:所有SVG均调整至200×200尺寸,并使用[picosvg](https://github.com/nicothin/picosvg)进行简化
- **PNG标准化**:所有PNG图片均调整至448×448尺寸,以保证输入尺寸统一
## 引用格式
bibtex
@article{yang2025omnisvg,
title={OmniSVG: A Unified Scalable Vector Graphics Generation Model},
author={Yiying Yang and Wei Cheng and Sijin Chen and Xianfang Zeng and Jiaxu Zhang and Liao Wang and Gang Yu and Xinjun Ma and Yu-Gang Jiang},
journal={arXiv preprint arxiv:2504.06263},
year={2025}
}
## 标签
- 可缩放矢量图形(Scalable Vector Graphics,SVG)
- 视觉语言模型
- 多模态
- 图标
提供机构:
maas
创建时间:
2025-04-11



