• Hamiltonian Monte Carlo (hmc)
• Multivariable Hamiltonian Monte Carlo (mhmc)
• DropOut Hamiltonian Monte Carlo (dhmc)
• Multivariable DropOut Hamiltonian Monte Carlo (dmhmc)
• Logistic Regression (lr)
• Softmax Regression (sr)
• Utils
• Others

unzip datasets:

cat data.zip* > data.zip
unzip data.zip

compile programs:

mkdir build
cd build/
cmake ..
make

run programs:

cd build/
./<program>

For the creation of the features through Keras VGG_Face, first it is necessary to download the ADIENCE faces database from the following links:

https://drive.google.com/drive/folders/1A0EDo0oYH3pBEZyq6zfk_jVg8ZvYM2cE?usp=sharing

or

https://www.openu.ac.il/home/hassner/Adience/data.html

Download "aligned.tar.gz" archive, then:

mv download_path/aligned.tar.gz dropout-hmc_cpp/data/

cd dropout-hmc_cpp/data/

tar -xvf aligned.tar.gz 

Run python scripts:

cd dropout-hmc_python/python/

python keras_vgg_face_features.py

• CMake

• C++ 4.8 or later, C++ 11

• OpenCV 3.4 or later

• Eigen 3.3 or later

• Python 2.7

• Tensorflow 1.3 or later, Tensorflow-gpu (alternative)

• Edward 1.3 or later

• Keras 2.1 or later, keras-vggface

• SKlearn, Numpy, Scipy, Pandas, Seaborn, Matplotlib (according to dependence on previous packages)