Skip to content

Getting Started

spikeDE empowers you to build Fractional-Order Spiking Neural Networks (f-SNNs) using an API that aligns closely with PyTorch. This design ensures a seamless transition for PyTorch users, allowing you to leverage existing skills while exploring advanced fractional dynamics with minimal learning curve.

We support a diverse range of modern neural architectures, including Multilayer Perceptrons (MLPs), Convolutional Neural Networks (CNNs), Residual Networks (ResNets) and Transformers.

Whether you are researching neuromorphic vision, modeling complex time-series, or developing energy-efficient AI, spikeDE provides the robust tools needed to construct, train, and deploy high-performance spiking models.

In this section, you will find step-by-step guides ranging from installation to building and training your first functional network. Choose the path that best fits your current needs:

  • Installation

    Get up and running quickly. Install spikeDE via pip or from source, with detailed instructions for setting up all necessary dependencies.

    Go to Installation Guide

  • Introduction by Example

    Dive straight into code. Follow a complete walkthrough to define a fractional spiking model, encode inputs, train on a dataset, and evaluate performance.

    Try Your First f-SNN

New to Spiking Neural Networks?

If you are unfamiliar with SNN concepts, start with the Introduction by Example. It assumes only basic familiarity with PyTorch and gently introduces core concepts such as spike encoding, fractional leaky integrate-and-fire (f-LIF) neurons, and surrogate gradients.

Happy spiking!