traceTorch Documentation

traceTorch is a PyTorch-based library built on the principles of spiking neural networks, replacing the PyTorch default backpropagation through time with lightweight, per-layer input traces, enabling biologically inspired, constant time and memory consumption learning on arbitrarily long or even streaming sequences.

The docs cover the following:

• Introduction
  • 1. Biological Plausibility in Machine Learning
  • 2. What is an SNN?
  • 3. What is a trace?
  • 4. How does traceTorch work?
  • 5. The advantages of traceTorch
  • 6. The disadvantages of traceTorch
  • 7. What is the purpose of traceTorch?
• Tutorials
  • 0. Terminology used throughout traceTorch
  • 1. Memorizing an input pattern to output pattern mapping
  • 2. Training a traceTorch model on MNIST
  • 3. Poisson clicks
• Documentation
  • Functional modules
  • Loss modules
  • NN modules
  • Plotting modules