Skip to content

Dataset · Zitacron

Large corpus for AI image detection

Zitacron/real-vs-ai-corpus

Collection of tens of millions of images for training generated image detectors with commercially clean licenses.

What's inside

The dataset, with a volume ranging from 10 million to 100 million records, is designed for binary classification of real and neural network-generated images. It was assembled by streaming and merging 17 open datasets from Hugging Face without intermediate local storage. All source parts are distributed under CC BY 4.0, Apache 2.0, or MIT licenses, permitting commercial use of trained models with attribution.

What you can do with it

  • Training classifiers to recognize deepfakes and AI-generated content
  • Fine-tuning commercial visual content moderation systems
Tasks
Image classification
Languages
English
Size
10M–100M rows
Downloads this month
32.9K
Likes
4

How to load it

from datasets import load_dataset

ds = load_dataset("Zitacron/real-vs-ai-corpus")