Skip to content

Dataset · zr-zhang

Dataset for detecting images from GPT Image2 and Nano Banana2

zr-zhang/MLLM-Generated-Image-Detection-Dataset

A collection of tens of thousands of real and generated images for training and evaluating AI image detectors.

What's inside

An image binary classification benchmark ranging from 10,000 to 100,000 examples, including real photos and images from GPT Image2 and Nano Banana2 models. Examples are grouped into three categories: texture-dominant, structure-dominant, and hybrid. Created to test detector robustness against new generative models.

What you can do with it

  • Training binary classifiers to detect AI-generated images
  • Testing deepfake and synthetic content detectors on images from new models
Tasks
Image classification
Size
10K–100K rows
Downloads this month
38.6K
Likes
1

How to load it

from datasets import load_dataset

ds = load_dataset("zr-zhang/MLLM-Generated-Image-Detection-Dataset")