---
title: SeFi-Image-5B-Base-demo
canonical_url: "https://www.modelscope.cn/studios/AI-ModelScope/SeFi-Image-5B-Base-demo"
md_url: "https://www.modelscope.cn/studios/AI-ModelScope/SeFi-Image-5B-Base-demo.md"
repository: AI-ModelScope/SeFi-Image-5B-Base-demo
last_updated: 2026-07-06
sdk_type: gradio
sdk_version: 6.17.3
downloads: 0
stars: 43
---

# SeFi-Image-5B-Base-demo

> SeFi-Image-5B-Base-demo - AI-ModelScope 在 ModelScope 创建的在线 Demo。A Gradio demo for SeFi-Image/SeFi-Image-5B-Base, a 5B-parameter text-to-image foundation model built with Semantic-First Diffusion. It separates semantic and texture latent streams and denoises…

AI-ModelScope/SeFi-Image-5B-Base-demo 是 ModelScope 魔搭社区上的在线可交互 Demo（创空间），基于 gradio 6.17.3 构建。

- **Repository**: AI-ModelScope/SeFi-Image-5B-Base-demo
- **SDK**: gradio
- **SDK version**: 6.17.3
- **Downloads**: 0
- **Stars**: 43
- **Last updated**: 2026-07-06

Source: https://www.modelscope.cn/studios/AI-ModelScope/SeFi-Image-5B-Base-demo

---

# SeFi-Image-5B-Base

A Gradio demo for [SeFi-Image/SeFi-Image-5B-Base](https://huggingface.co/SeFi-Image/SeFi-Image-5B-Base),
a 5B-parameter text-to-image foundation model built with **Semantic-First Diffusion**.
It separates semantic and texture latent streams and denoises semantic structure
slightly ahead of texture details, giving the texture stream a cleaner structural
anchor during generation.

Runs on ZeroGPU using the official `sefi` inference package
([code](https://github.com/jmliu206/SeFi-Image)).