OmniSVG
Unified framework for text/image-to-high-quality-SVG — logo/icon generation
A unified framework converting text/images to high-quality SVG vector graphics — for logo and icon generation.
Image generation and computer vision — detection, editing, 3D.
Key features
- First family of end-to-end multimodal SVG generators leveraging pre-trained VLMs
- Generates complex detailed SVGs from simple icons to intricate anime characters
- MMSVG-2M dataset: two million richly annotated SVG assets with standardized evaluation
- Model variants OmniSVG-3B, 1.1_4B, 1.1_8B with inference code and weights on Hugging Face
- Includes MMSVGBench benchmark and evaluation code; ComfyUI plugin by community
Good for
- Generating vector icons, illustrations, and anime-style SVG graphics
- Researchers benchmarking conditional SVG generation
Stack & tags
PythonPyTorchsvgvector-graphicsimage-generationvlmopen-weightsiconsneurips2025
Quick start
git clone https://github.com/OmniSVG/OmniSVG.git
conda create -n omnisvg python=3.10
conda activate omnisvg
pip install -r requirements.txtFrom the project README on GitHub.
Similar open-source projects
FLUX.2 klein
★ 25,996Black Forest Labs' sub-second image generation — generate and edit in <0.5s
DiffSynth-Studio
★ 13,200The framework behind Qwen-Image-i2L — the first "Image→LoRA" visual generation system