Spiking neural network conversion toolbox¶
The SNN conversion toolbox (SNN-TB) is a framework to transform rate-based artificial neural networks into spiking neural networks, and to run them using various spike encodings. A unique feature about SNN-TB is that it accepts input models from many different deep-learning libraries (Keras / TF, pytorch, …) and provides an interface to several backends for simulation (pyNN, brian2, …) or deployment (SpiNNaker, Loihi). The source code can be found on GitHub. See also the accompanying articles [Rueckauer et al., 2017] and [Rueckauer and Liu, 2018].
These sections guide you through the installation, configuration and running of the toolbox. Examples are included.
Here you find detailed descriptions of specific functions and classes.