Dataset · competitions
AI image detection for the AI or Not competition
competitions/aiornot
Image and text label dataset with 10K to 100K examples for training neural network-generated image classifiers.
What's inside
Contains images and text data for image classification tasks. The dataset size ranges from 10,000 to 100,000 records. Created for the AI or Not competition, downloading requires Hugging Face authentication and acceptance of competition rules; the sample may contain NSFW content.
What you can do with it
- Training a binary classifier to distinguish real photos from neural network generations
- Testing the robustness of synthetic content detectors
- Tasks
- Image classification
- Size
- 10K–100K rows
- Downloads this month
- 1.2K
- Likes
- 33
- Access
- on request on Hugging Face
Models trained on it
- Nahrawy/AIorNotImage classification 3.3K
- prithivMLmods/AIorNot-SigLIP2Image classification 62
- RishiDarkDevil/ai-image-det-resnet152Embeddings 25
- Binyamin/Hybrid_1 24
- psyne/AIResnetCloneImage classification 24
- kgmann/ai-image-det-resnet18Image classification 23
- emanehabn/aiornot_emanImage classification 8
- ted52/aiornot_cnn 0
How to load it
from datasets import load_dataset
ds = load_dataset("competitions/aiornot")