Sinabs (Sinabs Is Not A Brain Simulator) is a python library for the development and implementation of Spiking Convolutional Neural Networks (SCNNs). The library implements several layers that are spiking equivalents of CNN layers. In addition it provides support to import CNN models implemented in torch conveniently to test their spiking equivalent implementation. This project is managed by SynSense (former aiCTX AG).

The sinabs-dynapcnn was incorporated to this project, and it enables porting sinabs models to chips and dev-kits with DYNAP-CNN technology.

Installation

For the stable release on the main branch:

pip install sinabs

or (thanks to @Tobias-Fischer)

conda install -c conda-forge sinabs

For the latest pre-release on the develop branch that passed the tests:

pip install sinabs --pre

The package has been tested on the following configurations

Documentation and Examples

https://sinabs.readthedocs.io/

Questions? Feedback?

Please join us on the #sinabs Discord channel!

• If you would like to report bugs or push any changes, you can do this on our github repository.

License

Sinabs is published under AGPL v3.0. See the LICENSE file for details.

Contributing to Sinabs

Checkout the contributing page for more info.

Citation

In case you find this software library useful for your work please consider citing it as follows:

@software{sinabs,
author = {Sheik, Sadique and Lenz, Gregor  and Bauer, Felix and Kuepelioglu, Nogay },
doi = {10.5281/zenodo.8385545},
license = {AGPL-3.0},
title = {{SINABS: A simple Pytorch based SNN library specialised for Speck}},
url = {https://github.com/synsense/sinabs}
}