README.md

---
license: apache-2.0
language:
- en
base_model:
- stabilityai/stable-diffusion-3.5-medium
pipeline_tag: image-to-image
library_name: transformers
---

# CLUE Model Card
## Model details

**Model type:** This model, named CLUE, is an advanced image forgery localization framework that synergistically combines a LoRA-tuned Stable Diffusion 3 (SD3) and Segment Anything Model (SAM). It uniquely leverages SD3's Rectified Flow mechanism to amplify subtle forensic artifacts, which enables state-of-the-art localization of forged image regions.

**Project page:** https://github.com/youqiwong/CLUE.git