The Pytorch implemention of Deepfake Detection based on Faceforensics++
We also reproduced the MesoNet with pytorch version, and you can use the mesonet network in this project.
The code has been tested on pytorch=1.3.1 and python 3.6, please refer to requirements.txt for more details.
python -m pip install -r requirements.txt
Although you can install all dependencies at a time. But it is easy to install dlib via conda install -c conda-forge dlib
If you want to use the opensource dataset Faceforensics++, you can use the script './download-FaceForensics_v3.py' to download the dataset accroding the instructions of download section.
You can train the model with full images, but we suggest you take only face region as input.
The model provided just be used to test the effectiveness of our code. We suggest you train you own models based on your dataset.
And we will upload models which have better performance as soon as possible.
we provide some pretrained model based on FaceForensics++
- FF++_c23.pth
- FF++_c40.pth
model choice
Base CORE SelfBlend
To test with videos
python detect_from_video.py --video_path ./videos/003_000.mp4 --model_name Base/CORE/SelfBlend -o ./output --cuda
To test with images
python detect_from_image.py --image_path ./data_list/Deepfakes_c0_299.txt --model_name Base/CORE/SelfBlend -o ./output --cuda
To train a model
python train_CNN.py
(Please set the arguments after read the code)
If our project is helpful to you, we hope you can star and fork it. If there are any questions and suggestions, please feel free to contact us.
Thanks for your support.
The provided implementation is strictly for academic purposes only. Should you be interested in using our technology for any commercial use, please feel free to contact us.