Python bindings for Fast Artificial Neural Networks 2.2.0 (FANN >= 2.2.0). These are the original python bindings included with FANN 2.1.0beta and updated to include support for python 2.6-3.4.
This is a python binding for Fast Artificial Neural Network Library (FANN >= 2.2.0) that implements multilayer artificial neural networks with support for both fully-connected and sparsely-connected networks. It includes a framework for easy handling of training data sets. It is easy to use, versatile, well- documented, and fast.
You can install fann2 from pkgsrc or from pypi, using either pip or easy_install:
$ pip install fann2
or
$ easy_install fann2
Get and install pkgsrc. See pkgsrc documentation. for platform-specific information.
cd ${PKGSRCDIR}/devel/py-fann2
bmake install
Get and install pkgsrc. See pkgsrc quickstart. for platform-specific information.
pkgin -y install py-fann2
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 at the FANN documentation and its C++ bindings for further reference.
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.
FANN.
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.