The first version of code fork from here and add some useful function I will remove some ugly function and re-coding later... I prefer pytorch framework, so the final version will support pytorch 1.x. Anyway, thanks to the authors.
-
Tensorflow-gpu >= 1.10
-
opencv >= 2.4.5
-
dlib
- Linux and Window platform
- Chinese name
- CPU and GPU (请等待...)
- run the face recognition
python main.py
- train model
python train.py --epochs 1000
TODO:
- Pytorch CNN version
- Train Backend network using recently face network
- Using more fast face detector
- Store face into encoder-format