Nervos

A Spiking Neural Network Simulation Framework

Nervos is a flexible and customizable tool designed to simulate simple Spiking Neural Networks (SNNs). Developed under the supervision of Prof. Sandip Lashkare at IIT Gandhinagar, it provides a platform for experimenting with biological neuron models, learning rules, and hardware constraints.

Documentation: https://nervos.readthedocs.io/

Installation

The simplest way to install nervos is using pip:

pip install nervos

Features

• Biologically-Inspired Models: Implements the Leaky Integrate-and-Fire (LIF) neuron model.
• STDP Learning: Built-in support for Spike-Timing-Dependent Plasticity (STDP) for unsupervised learning.
• Hardware Simulation:
  • Synaptic Variability: Simulate read noise in synaptic weights (enable_synaptic_noise).
  • Endurance Modeling: Model cycle-dependent weight updates to simulate device aging or non-idealities (enable_cycle_dependent_weights).
• Customizable: Fully configurable parameters for neurons, synapses, and training loops.
• Model Persistence: Save and load trained models and training states.
• Data Loaders: Built-in loaders for MNIST, Iris, and Circles datasets with SNN preprocessing.

Examples

Check out the lib_examples directory for usage scripts and notebooks:

1. MNIST Classification: Standard SNN training on MNIST.
2. Hardware Variability: Simulating synaptic noise during inference/training.
3. Device Endurance: Simulating weight update degradation over time.
4. Simple Datasets: Examples for Iris and Circles datasets.
5. Single Neuron: Demonstrations of current injection and spiking behavior.

Poster & Paper

• Poster: ./Project_Files/Poster.pdf
• Paper: arXiv | Springer

Citation

If you use this library or adapt code from it, please cite the following paper:

@InProceedings{maskeen_unified_2026,
author="Maskeen, Jaskirat Singh
and Lashkare, Sandip",
title="A Unified Platform to Evaluate STDP Learning Rule and Synapse Model Using Pattern Recognition in a Spiking Neural Network",
booktitle="Artificial Neural Networks and Machine Learning -- ICANN 2025",
year="2026",
publisher="Springer Nature Switzerland",
doi={10.1007/978-3-032-04558-4_41},
address="Cham",
pages="509--520",
isbn="978-3-032-04558-4"
}

Feel free to fork and contribute!