traceTorch Documentation

A strict, ergonomic, and powerful Spiking Neural Network (SNN) library for PyTorch.

This documentation is written to help you familiarize with SNNs and traceTorch. By the end of it, you should have all the background knowledge necessary to understand and make your own networks, as well as a solid understanding of the library and good practices for creating powerful models. It is recommended to at least briefly read the introduction section to familiarize yourself with some background theory and terminology before proceeding with the library.

Contents:

• Introduction
  • 1. What is an SNN?
  • 2. The Ethos of traceTorch
  • 3. The traceTorch Layers
  • 4. The Distributed Hidden State
  • 5. Monotonic and Nonmonotonic Functions
  • 6. Surrogate Derivatives
• Installation
• Tutorials
  • T0. Memorizing Patterns
  • T1. Monotonic MNIST
  • T2. Nonmonotonic MNIST
• Documentation
  • Functional modules
  • Loss modules
  • NN modules
  • Plotting modules