This repository contains the implementation of various deep learning models for deepfake detection, including experiments with different architectures and ensemble methods.
df*_*.ipynb: Jupyter notebooks containing different model implementations and experiments- ResNet-based models
- EfficientNet models (B0, B6, B7)
- ConvNeXt models
- Swin Transformer models
- Ensemble approaches
- ResNet-based models
- EfficientNet (B0, B6, B7)
- ConvNeXt
- Swin Transformer
- Ensemble models combining multiple architectures
- Face Forensics dataset
- Kaggle Deepfake Detection dataset
The project uses various deep learning frameworks and libraries. Key dependencies include:
- PyTorch
- torchvision
- timm
- scikit-learn
- pandas
- numpy
- matplotlib
- seaborn
Each notebook contains a complete implementation of a specific model or experiment. The notebooks are numbered sequentially and named according to the model architecture and dataset used.
The project includes various evaluation metrics and visualizations:
- ROC curves
- Confusion matrices
- Training logs
Model checkpoints and large data files are not included in this repository due to size limitations. Please refer to the respective notebooks for information on how to obtain or generate these files.