README
fann2
Python bindings for Fast Artificial Neural Networks 2.2.0 (FANN >= 2.2.0). These are the original python bindings included with FANN 2.1.0 beta and updated to include support for python 2.6-3.6.
DESCRIPTION
This is a python binding for Fast Artificial Neural Network Library (FANN >= 2.2.0) that implements multi-layer artificial neural networks with support for both fully-connected and sparsely-connected networks. It includes a framework for easily handling training data sets. It is easy to use, versatile, well- documented, and fast.
FANN 2.2.0 source
INSTALLATION
You can install fann2 from pkgsrc or from pypi, using either pip or easy_install:
wheels
$ pip install ${wheel_file}
pypi
$ pip install fann2
or
$ easy_install fann2
pkgsrc
Source installation
Get and install pkgsrc. See pkgsrc documentation. for platform-specific information.
cd ${PKGSRCDIR}/devel/py-fann2
bmake install
From binaries
Get and install pkgsrc. See pkgsrc quickstart. for platform-specific information.
pkgin -y install py-fann2
Windows considerations
Source installation
- Install Visual C++ Build Tools;
- Install FANN source code, using cmake;
- Copy "fanndouble.lib" from FANN installed files to ${python_libs_directory} as "doublefann.lib";
- Install swig for Windows (you will need to set an Enviroment Variable for "swig.exe");
- Run > python setup.py install from PowerShell/Command Prompt.
USAGE
Just
>> from fann2 import libfann
and then create libfann.neural_net and libfann.training_data objects
>> ann = libfann.neural_net()
>> train_data = libfann.training_data()
Look at the examples in the FANN documentation and its C++ bindings for further reference.
LICENSE
As with the original python bindings, this package is distributed under the terms of the GNU Lesser General Public License, Version 2.1. See LICENSE for full terms and conditions.
LINKS
FANN.
CONTACT
Send us your patches and pull requests! We will release as often as these changes are received and integrated. There's no reason to have countless branches of this package. Consider this the official one and that it's being maintained!
The pkgsrc package is maintained by us as well. We are active users of FANN and fann2. If you don't have or want a github account, send your patches for this package or the pkgsrc version to pkgsrc@futurelinkcorporation.com.
