---
license: llama2
datasets:
- saberzl/SID_Set
language:
- en
metrics:
- accuracy
base_model:
- xinlai/LISA-13B-llama2-v1
pipeline_tag: image-segmentation
---

# SIDA Model Card

## Model details

**Model type:**
SIDA is a model fine-tuned from LISA, designed to detect and localize tampered regions in images.

**Model date:**
SIDA-13B was trained in Febuary 2025.

**Paper or resources for more information:**

Paper: https://arxiv.org/pdf/2412.04292

Resource: https://github.com/hzlsaber/SIDA

## License
Llama 2 is licensed under the LLAMA 2 Community License, 
Copyright (c) Meta Platforms, Inc. All Rights Reserved.

## Trained Data

SIDA was trained on SID_Set, which consists of real images, tampered images, and fully synthetic images. More information is available [here](https://huggingface.co/datasets/saberzl/SID_Set)

## Citation Information

If you find this dataset useful, please consider citing our paper:

```
@misc{huang2025sidasocialmediaimage,
      title={SIDA: Social Media Image Deepfake Detection, Localization and Explanation with Large Multimodal Model}, 
      author={Zhenglin Huang and Jinwei Hu and Xiangtai Li and Yiwei He and Xingyu Zhao and Bei Peng and Baoyuan Wu and Xiaowei Huang and Guangliang Cheng},
      year={2025},
      booktitle={Conference on Computer Vision and Pattern Recognition}
}
```