SuperNN is an artificial neural network library, with support to arbitrary feed-forward topologies, that implements state-of-the-art supervised training algorithms such as Improved Resilient Backpropagation (iRprop) [Igel and Hüsken, 2000] and Neuron by Neuron (NBN) [Wilamowski, Yu, 2010].
SuperNN is released under the LGPLv3+ license. It was tested on modern GNU/Linux distributions and on MS-Windows 8 (by using MingW, VS2010, and VS2013).
Provides an exporter tool that allows one to generate standalone Java or C++ code to use trained neural networks without requiring SuperNN at runtime.
SuperNN depends on Eigen (compile time only).
$ mkdir build $ cd build $ cmake -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/usr/local .. $ make -j # make install
To generate the API documentation to doc/html:
$ doxygen doc/Doxyfile
To include SuperNN into your cmake project you can use the provided module located at cmake/FindSuperNN.cmake.
You can also generate standalone Java or C++ code to run a trained neural network without requiring supernn at runtime. The generators are located on the tools folder.
There are some usage examples in the tests folder.
Lucas Hermann Negri - email@example.com - OProj