Update: Everything before was awful and terrible, so I needed to fix it.
.
├── file_preperation - Folder containing some clever CSV parsing
├── neural_networks - Folder containing the neural network configurations
└── README.md - This file, duh.
Getting this project set up is a fairly simple task that will only take a few minutes of your time.
The project is now built on the keras theano-based deep learning library. It's built entirely on simplicity and ease of implementation.
I found the easiest way to set this up was via pip. pip is a package management system for the python programming language.
To install pip on a Ubuntu system, simply run the following command in the command line.
sudo apt-get install python-pip python-dev build-essential
Note: This project was made using Python 2.7.6.
Now that this is set up, you're going to need to install a few dependencies first.
- numpy and scipy:
sudo pip install scipy
- PIL:
sudo pip install pillow
- pyaml:
sudo pip install pyyaml
- theano:
sudo pip install theano
And finally, you can install keras using the following command.
sudo pip install keras
After that, load the software/neuralNetworkSoftware
project into Pycharm, or whatever IDE/plain text editor you use and everything should be set up. If not, feel free to email me or create a new issue.
Provided that you have CUDA set up correctly, running tests on a GPU is as easy as running a line in terminal.
THEANO_FLAGS=mode=FAST_RUN,device=gpu,floatX=float32 filename.py