---
title: ARForensics
canonical_url: "https://www.modelscope.cn/datasets/YanranZhang/ARForensics"
md_url: "https://www.modelscope.cn/datasets/YanranZhang/ARForensics.md"
repository: YanranZhang/ARForensics
last_updated: 2026-07-31
license: "MIT License"
storage_size: "53 GB"
downloads: 16586
stars: 2
---

# ARForensics

> ARForensics - YanranZhang 在 ModelScope 开源的数据集。[ICCV 2025] D³QE: Learning Discrete Distribution Discrepancy-aware Quantization Error for Autoregressive-Generated Image Detection

YanranZhang/ARForensics 是 ModelScope 魔搭社区上的数据集，存储大小 53 GB，采用 MIT License 许可。

- **Repository**: YanranZhang/ARForensics
- **License**: MIT License
- **Storage size**: 53 GB
- **Downloads**: 16586
- **Stars**: 2
- **Last updated**: 2026-07-31

Source: https://www.modelscope.cn/datasets/YanranZhang/ARForensics

---

# [ICCV 2025] D³QE: Learning Discrete Distribution Discrepancy-aware Quantization Error for Autoregressive-Generated Image Detection


<div align='center' style='margin-bottom:20px'>
  <a href='http://arxiv.org/abs/2510.05891'><img src='https://img.shields.io/badge/ArXiv-red?logo=arxiv'></a>  &nbsp;
  <a href='https://ivg-yanranzhang.github.io/D3QE/'><img src='https://img.shields.io/badge/Visualization-green?logo=github'></a> &nbsp;
  <a href="https://github.com/Zhangyr2022/D3QE"><img src="https://img.shields.io/badge/Code-9E95B7?logo=github"></a>
</div>

`ARForensics` dataset from [ICCV 2025] D³QE: Learning Discrete Distribution Discrepancy-aware Quantization Error for Autoregressive-Generated Image Detection

<div align=center>
<img src='dataset.png'>
</div>

## Introduction

The `ARForensics` dataset is the first benchmark for **visual autoregressive model detection**. It comprises 304,000 images (152,000 real from ImageNet, 152,000 synthetic). The dataset features 7 diverse autoregressive models: LlamaGen<small>[![Star](https://img.shields.io/github/stars/FoundationVision/LlamaGen.svg?style=social&label=Star)](https://github.com/FoundationVision/LlamaGen)</small>, VAR<small>[![Star](https://img.shields.io/github/stars/FoundationVision/VAR.svg?style=social&label=Star)](https://github.com/FoundationVision/VAR)</small>, Infinity<small>[![Star](https://img.shields.io/github/stars/FoundationVision/Infinity.svg?style=social&label=Star)](https://github.com/FoundationVision/Infinity)</small>, Janus-Pro<small>[![Star](https://img.shields.io/github/stars/FoundationVision/Infinity.svg?style=social&label=Star)](https://github.com/deepseek-ai/Janus)</small>, RAR<small>[![Star](https://img.shields.io/github/stars/bytedance/1d-tokenizer.svg?style=social&label=Star)](https://github.com/bytedance/1d-tokenizer/tree/main)</small>, Switti<small>[![Star](https://img.shields.io/github/stars/yandex-research/switti.svg?style=social&label=Star)](https://github.com/yandex-research/switti)</small>, and Open-MAGVIT2<small>[![Star](https://img.shields.io/github/stars/TencentARC/SEED-Voken.svg?style=social&label=Star)](https://github.com/TencentARC/SEED-Voken)</small>, which encompasses both token-based and scale-based architectures. It is structured into training (100k LlamaGen), validation (10k), and a comprehensive test set (6k samples from each of the 7 models), ensuring a balanced and technically varied evaluation benchmark for AI-generated image detection.

## Usage

After downloading all the files in this directory, you can run the following command to merge them into a single zip file:

```bash
cat ARForensics_part.* > ARForensics_part.zip
```

**Download:** The dataset `ARForensics` is uploaded and available at: [🤗 HuggingFace](https://huggingface.co/datasets/Yanran21/ARForensics)   |   [🤖 ModelScope](https://www.modelscope.cn/datasets/YanranZhang/ARForensics).

**Folder structure (expected):**

```
ARForensics/
├─ train/
│  ├─ 0_real/
│  └─ 1_fake/
├─ val/
│  ├─ 0_real/
│  └─ 1_fake/
└─ test/
   ├─ Infinity/
   │  ├─ 0_real/
   │  └─ 1_fake/
   ├─ Janus_Pro/
   │  ├─ ..
   ├─ RAR/
   ├─ Switti/
   ├─ VAR/
   ├─ LlamaGen/
   └─ Open_MAGVIT2/
```

## Cite

If you find this repository useful for your research, please consider citing this bibtex.