BindsNET

BindsNET is an open-source Python library for simulating spiking neural networks (SNNs) on CPUs and GPUs using PyTorch. Developed by the BindsNET team on GitHub, where it has collected more than 1,700 stars, it remains actively maintained, with the latest code pushed in October 2026, and is released under the AGPL-3.0 license.

Overview

BindsNET is a spiking neural network simulation library geared towards the development of biologically inspired algorithms for machine learning. It grew out of ongoing research in the Biologically Inspired Neural and Dynamical Systems (BINDS) lab at the University of Massachusetts Amherst and the Allen Discovery Center at Tufts University. Because PyTorch does not natively solve the differential equations behind spiking neuron dynamics, BindsNET converts them into difference equations solved at short, regular intervals using torch tensors, which makes it possible to reuse PyTorch operations such as convolution and pooling inside SNN architectures. The package supports unsupervised, supervised and reinforcement learning experiments, and its learning rules are validated against the equations of the papers they implement.

Key Features

• PyTorch-native simulation: build spiking networks from torch tensors and reuse torch.nn.functional operations in SNN architectures
• CPU and GPU execution for spiking neural network simulations
• Biologically inspired learning rules, including spike-timing-dependent plasticity (STDP), each validated against its source paper
• Multiple learning paradigms: unsupervised representation learning, supervised training and reinforcement learning with Gymnasium environments
• Ready-made examples, including a replication of the Diehl and Cook 2015 MNIST network and an Atari Breakout reinforcement learning experiment
• Reproducibility tooling: pinned dependency installs, a Dockerfile, declared dataset sources and step-by-step reproduction guides
• Full documentation on Read the Docs, with the package published on PyPI for a one-line pip install

💳 Pricing

BindsNET is completely free and open source under the GNU AGPL-3.0 license. There are no paid tiers, subscriptions or hosted versions: install it from PyPI with pip, through the pinned Poetry environment, or from source, and run simulations on your own CPU or GPU hardware.

Who Is It For

BindsNET is built for researchers and graduate students working on spiking neural networks, neuromorphic computing and biologically inspired machine learning, engineers who want to experiment with STDP-based learning or SNN agents for reinforcement learning problems, and educators who need a PyTorch-based SNN simulator with documented, reproducible examples.