pixelou/nnbox

Yet another Neural Network toolbox for quick development and flexible architectures

# NNBox
NNBox is a Matlab toolbox for neural networks. Many other toolboxes are
already available for matlab and may either offer more models, a higher levels
of support, better optimization, or simply a bigger user community... This
toolbox has been concieved with two main objectives:
- Providing very clear and simple implementations of some neural networks
models and architectures.
- Providing a flexible interface where building blocks can be arranged
together easily.
In particular, this library provides support for Restricted Boltzmann Machines
(RBM), Convolutional Neural Networks (CNN), simple perceptrons models. It
allows to arrange these models in parallel, as stacked multiple layers, or even
in a Siamese network architecture.
This library does not focus on completeness though, because attempting to do so
rarely gives satisfying results. Instead it tries to provide simple and
flexible architectural fundations to help you implement your own model quickly.

For your information, here is a list of other existing libraries:

- Matlab Neural Network toolbox (http://fr.mathworks.com/help/nnet/index.html)
- DeepLearnToolbox (https://github.com/rasmusbergpalm/DeepLearnToolbox)
A popular deep learning toolbox
- MEDAL (https://github.com/dustinstansbury/medal) Similarily provides
implementations for several sorts of Deep Learning models.
- MatConvNet (http://www.vlfeat.org/matconvnet/) Provides awrapper to a C++
implementation of convolutional neural networks. It is actually used here
for the CNN model.

## Requirements

As far as I can tell, any version of matlab above R2011a should work, R2014a is
known to work. Octave is not supported because classes are not yet fully
supported.

## Installation

Just add nnbox subfolders to your path:

> addpath('nnbox/utils:nnbox/networks:nnbox/costfun:nnbox/distances');

CNN implementation requires the [MatConvNet](http://www.vlfeat.org/matconvnet/)
library as a backend, follow installation instructions and add the matlab
bindings to the path.

## Examples

> X = [0 1 0 1;
> 0 0 1 1];
> Y = [0 .5 .5 1];
> net = Perceptron(2, 1, struct('lRate', 0.5));
> trainOpts = struct('nIter', 100, 'displayEvery', 10);
> train(net, SquareCost(), X, Y, trainOpts);

- MNIST figure recognition using a Deep belief network :
examples/MNIST_DNN.m (https://github.com/pixelou/nnbox/blob/master/examples/MNIST_DNN.m)

## Documentation

Refer to DOCUMENTATION.md (https://github.com/pixelou/nnbox/blob/master/DOCUMENTATION.md)
The upstream source repository can be accessed at https://github.com/pixelou/nnbox where bugs should be reported.
Please, feel free to open a bug report for feature or documentation requests as well.