ASpiReNN

ASpiReNN stands for Artificial SPIking REcurrent Neural Network and is a little library written in C which, as its name suggests, allows to simulates artificial spiking (or not) recurrent (or not) neural networks.

Features

ASpiReNN's current features are:

• leaky integrate-and-fire support
• possibility to add new neuron types
• Spike-Timing Dependent Plasticity learning rule
• possibility to add new learning rules
• fast multilayer networks
• best suited for highly recurrent neural nets
• Python bindings

Requirements

In order to install ASpiReNN you'll need:

• A C99 compatible compiler (such as GCC)
• ... and that should be enough, theoretically. In pratice it will be easier to install on an Unix platform with common utilities such as Makefiles, etc.

In order to install Aspy (the Python bindings for ASpiReNN), you'll need:

• Python, obviously (preferently Python2.4)
• SWIG

Downloads

Latest interface version is 0.4 ; you can download it here. You can also check the changelog. Note that 'aspirenn' corresponds to the library, and 'aspy' to the python bindings. Bindings version 'x.y.z' should be compatible with any library version 'x.y.w', but don't expect more compatibility between versions (to make that clear: aspy-0.2 and aspirenn-0.2.1 can work together, but aspy-0.1.1 and aspirenn-0.2 cannot).

Documentation

Documentation is included in the code and can be extracted with Doxygen; you can also browse it here.

Some examples may be available here.

License

GNU GPL, of course.

Contact

neryel AT reveries DOT info