This repository contains the official PyTorch implementation for "Forgery-Aware Adaptive Vision Transformer for Generalized Face Forgery Detection".
- timm == 0.5.4
- pytorch == 1.8.0
The trained model (on FF++ C23) can be accessed from here https://drive.google.com/file/d/1QYuFeLDlbQW1uzzID31e5c0o6LaBbtl4/view?usp=drive_link
We provide a demo in evaluate.py, which makes it very easy to evaluate the trained model on your customized dataset.