The original neural network from Torch7, 'nn', contains stable and widely used modules. 'nnx' contains more experimental, unproven modules, and optimizations. Eventually, modules that become stable enough will make their way into 'nn' (some already have).

1/ Torch7 is required:

Dependencies, on Linux (Ubuntu > 9.04):

$ apt-get install gcc g++ git libreadline5-dev cmake wget libqt4-core libqt4-gui libqt4-dev

Dependencies, on Mac OS (Leopard, or more), using Homebrew:

$ brew install git readline cmake wget qt

Then on both platforms:

$ git clone https://github.com/andresy/torch
$ cd torch
$ mkdir build; cd build
$ cmake ..
$ make
$ [sudo] make install

2/ Once Torch7 is available, install this package:

$ [sudo] torch-pkg install nnx

First run torch, and load nnx:

$ torch

> require 'nnx'

Once loaded, tab-completion will help you navigate through the library (note that most function are added directly to nn):

> nnx. + TAB
...
> nn. + TAB

In particular, it's good to verify that all modules provided pass their tests:

> nnx.test_all()
> nnx.test_omp()