Community Forensics: Using Thousands of Generators to Train Fake Image Detectors
Paper • 2411.04125 • Published • 1
This is a data processor for the Community Forensics classifiers. Detailed instructions will be updated at the Community Forensics GitHub_repo in the future.
- Code: GitHub Repository
- Paper: arXiv
Simple usage example.
dataprocessor_hf.py is included in this repository.
import dataprocessor_hf as dphf
from PIL import Image
data_processor = dphf.CommForImageProcessor.from_pretrained('OwensLab/commfor-data-preprocessor', size=384)
img = Image.open('path_to_img.png')
processed_image = data_processor(img, mode='test')
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("OwensLab/commfor-data-preprocessor", dtype="auto")