---
language:
- en
license: mit
size_categories:
- 10K<n<100K
task_categories:
- text-to-image
- image-feature-extraction
tags:
- diffusion models
- image copy detection
dataset_info:
  features:
  - name: Name
    dtype: string
  - name: Level
    dtype: int64
  - name: generated_images
    dtype: image
  - name: real_images
    dtype: image
  splits:
  - name: Test
    num_bytes: 2538590040.0
    num_examples: 4000
  - name: Train
    num_bytes: 22265208436.0
    num_examples: 36000
  download_size: 24773596239
  dataset_size: 24803798476.0
configs:
- config_name: default
  data_files:
  - split: Test
    path: data/Test-*
  - split: Train
    path: data/Train-*
---

<p align="center">
  <img src="https://huggingface.co/datasets/WenhaoWang/D-Rep/resolve/main/D-Rep.png" width="800">
</p>

# Summary
This is the dataset proposed in our paper [**Image Copy Detection for Diffusion Models**](https://arxiv.org/abs/2410.xxxxx) (NeurIPS 2024).

D-Rep consists of 40, 000 image-replica pairs, in which each replica is generated by a diffusion model. The 40, 000 image-replica pairs are manually labeled with 6 replication levels ranging from 0 (no replication) to 5 (total replication). We divide D-Rep into a training set with 90% (36, 000) pairs and a test set with the remaining 10% (4, 000) pairs.