Stork

Library designed for training spiking neural networks with backpropagation through time (BPTT).

Overview

Stork is a library designed for the training of spiking neural networks. It extends PyTorch’s auto-differentiation capabilities with surrogate gradients to enable the training of SNNs with backpropagation through time (BPTT). Stork supports leaky integrate-and-fire (LIF) neurons including adaptive LIF neurons and different kinds of synaptic connections allowing to use Dalian and Convolutional layers as well as constructing network architectures with recurrent or skip connections.

For each neuron group customizable activity regularizers are available to for example apply homeostatic plasticity. Stork uses per default initialization in the fluctuation-driven regime that enhances SNN training in deep networks. Stork can be used with Tonic.

Stork

GitHub Release GitHub Stars
Maintainer(s):
FMI Zenke Lab
Language:
Python
License:
MIT
Application:
Neuromorphic Hardware, Computational Neuroscience, Spiking Neural Networks, Local Plasticity
Dependencies:
PyTorch

Help Improve this Software Guide

Our software guide is maintained by the community. If you have updates, see an error, or want to suggest a new tool, please let us know by opening an issue on our GitHub repository.