FaceMandara searches similar face to you in real time from over 200,000 face data(celebA).
OS: MacOS10.13.6 or Ubuntu16.04
dlib
face_recognition
faiss
opencv-python
-
At first, install Anaconda.
-
make virtual env by using
conda
$ conda create -n face_mandara python=3.5 dlib opencv-python face_recognition numpy
$ source activate face_mandara
- install faiss which is the package which can do Product Quantization from FaceBook
# CPU version only
$ conda install faiss-cpu -c pytorch
# Make sure you have CUDA installed before installing faiss-gpu, otherwise it falls back to CPU version
$ conda install faiss-gpu -c pytorch # [DEFAULT]For CUDA8.0
$ conda install faiss-gpu cuda90 -c pytorch # For CUDA9.0
$ conda install faiss-gpu cuda91 -c pytorch # For CUDA9.1
# cuda90/cuda91 shown above is a feature, it doesn't install CUDA for you.
-
download CelebA dataset and put the data to "big_database_def" directory. CelebA dataset
-
Rename the directory "big_database_def" to "big_database"
-
Then put the command terminal like bellow
$ python image2vector.py
$ python face_mandara.py
- You don't have to do this command from next execution
$ python image2vector.py
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent