Project sensei is developed for course Machine Learning at faculty of Mathematics, University of Belgrade. Since it's written in C++ it's only purpose is to enable the author to get in-depth understanding of underlaying concept of Restricted Boltzmann Machines. Beside RBM, sensei provides framework for deep learning so that further development is still open.
- armadillo
From root folder execute build process
make
In order to run example for training RBM on MNIST dataset, first you need a dataset. I've prepaired
MNIST dataset in CSV format that's readable by sensei (although, any CSV should be readable). This
file should be placed inside test/dataset
folder. One can download the dataset from here
After the dataset has been downloaded, issue this command from the project root to train RBM model:
make run
- Uros Stegic - [email protected] - uros-stegic
This project is licensed under the GNU GPL-3 License - see the LICENSE file for details