vinniesun / snntorch-ipu

Deep and online learning with spiking neural networks in Python for Graphcore IPU

Home Page:https://snntorch.readthedocs.io/en/latest/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Introduction

image

Documentation Status

Discord

image

image

image

Accelerating spiking neural networks on Graphcore's Intelligent Processing Units (IPUs). This fork runs parallel with the snnTorch project.

The brain is the perfect place to look for inspiration to develop more efficient neural networks. One of the main differences with modern deep learning is that the brain encodes information in spikes rather than continuous activations. snnTorch is a Python package for performing gradient-based learning with spiking neural networks. It extends the capabilities of PyTorch, taking advantage of its GPU accelerated tensor computation and applying it to networks of spiking neurons. Pre-designed spiking neuron models are seamlessly integrated within the PyTorch framework and can be treated as recurrent activation units.

image

If you like this project, please consider starring ⭐ this repo as it is the easiest and best way to support it.

If you have issues, comments, or are looking for advice on training spiking neural networks, you can open an issue, a discussion, or chat in our discord channel.

snnTorch Structure

snnTorch contains the following components:

Component Description
snntorch a spiking neuron library like torch.nn, deeply integrated with autograd
snntorch.backprop variations of backpropagation commonly used with SNNs
snntorch.functional common arithmetic operations on spikes, e.g., loss, regularization etc.
snntorch.spikegen a library for spike generation and data conversion
snntorch.spikeplot visualization tools for spike-based data using matplotlib and celluloid
snntorch.spikevision contains popular neuromorphic datasets
snntorch.surrogate optional surrogate gradient functions
snntorch.utils dataset utility functions

snnTorch is designed to be intuitively used with PyTorch, as though each spiking neuron were simply another activation in a sequence of layers. It is therefore agnostic to fully-connected layers, convolutional layers, residual connections, etc.

At present, the neuron models are represented by recursive functions which removes the need to store membrane potential traces for all neurons in a system in order to calculate the gradient. The lean requirements of snnTorch enable small and large networks to be viably trained on CPU, where needed. Provided that the network models and tensors are loaded onto CUDA, snnTorch takes advantage of GPU acceleration in the same way as PyTorch.

Citation

If you find snnTorch and the IPU-based build useful in your work, please cite the following sources:

Pao-Sheng Sun, Alexander Titterton, Anjlee Gopiani, Tim Santos, Arindam Basu, Wei D. Lu, and Jason K. Eshraghian “Intelligence Processing Units Accelerate Neuromorphic Learning”. arXiv preprint arXiv:2211.10725, November 2022.

Jason K. Eshraghian, Max Ward, Emre Neftci, Xinxin Wang, Gregor Lenz, Girish Dwivedi, Mohammed Bennamoun, Doo Seok Jeong, and Wei D. Lu “Training Spiking Neural Networks Using Lessons From Deep Learning”. arXiv preprint arXiv:2109.12894, September 2021.

@article{sun2022intelligence,
         title    =   {Intelligence Processing Units Accelerate Neuromorphic Learning},
         author   =   {Sun, Pao-Sheng Vincent and Titterton, Alexander and 
                       Gopiani, Anjlee and Santos, Tim and Basu, Arindam and 
                       Lu, Wei D and Eshraghian, Jason K},
         journal  =   {arXiv preprint arXiv:2211.10725},
         year     =   {2022}
}
@article{eshraghian2021training,
        title   =  {Training spiking neural networks using lessons from deep learning},
        author  =  {Eshraghian, Jason K and Ward, Max and Neftci, Emre and Wang, Xinxin 
                    and Lenz, Gregor and Dwivedi, Girish and Bennamoun, Mohammed and 
                   Jeong, Doo Seok and Lu, Wei D},
        journal = {arXiv preprint arXiv:2109.12894},
        year    = {2021}
}

Let us know if you are using snnTorch in any interesting work, research or blogs, as we would love to hear more about it! Reach out at snntorch@gmail.com.

Requirements

The following packages need to be installed to use snnTorch:

  • torch >= 1.1.0
  • numpy >= 1.17
  • poptorch
  • pandas
  • matplotlib
  • math
  • The Poplar SDK

Refer to Graphcore's documentation for installation instructions of poptorch and the Poplar SDK.

Installation

You can clone the public repository:

$ git clone git://github.com/vinniesun/snntorch-ipu

Once you have a copy of the source, you can install it with:

$ python setup.py install

Alternatively, install from PyPi using the following to install:

$ python
$ pip install snntorch-ipu

Low-level custom operations for IPU compatibility will be automatically compiled when import snntorch is called for the first time. Therefore, we recommend installing from source.

When updating the Poplar SDK, these operations may need to be recompiled. This can be done by reinstalling snntorch-ipu, deleting files in the base directory with an .so extension.

The snntorch.backprop module, and several functions from snntorch.functional and snntorch.surrogate, are incompatible with IPUs, but can be recreated using PyTorch primitives.

API & Examples

A complete API is available here. Examples, tutorials and Colab notebooks are provided.

Quickstart

Open In Colab

Here are a few ways you can get started with snnTorch:

For a quick example to run snnTorch, see the following snippet, or test the quickstart notebook:

import torch, torch.nn as nn
import snntorch as snn
from snntorch import surrogate

num_steps = 25 # number of time steps
batch_size = 1 
beta = 0.5  # neuron decay rate 
spike_grad = surrogate.fast_sigmoid()

net = nn.Sequential(
      nn.Conv2d(1, 8, 5),
      nn.MaxPool2d(2),
      snn.Leaky(beta=beta, init_hidden=True, spike_grad=spike_grad),
      nn.Conv2d(8, 16, 5),
      nn.MaxPool2d(2),
      snn.Leaky(beta=beta, init_hidden=True, spike_grad=spike_grad),
      nn.Flatten(),
      nn.Linear(16 * 4 * 4, 10),
      snn.Leaky(beta=beta, init_hidden=True, spike_grad=spike_grad, output=True)
      )

# random input data
data_in = torch.rand(num_steps, batch_size, 1, 28, 28)

spike_recording = []

for step in range(num_steps):
    spike, state = net(data_in[step])
    spike_recording.append(spike)

For IPU acceleration, the model must be wrapped in a dedicated class. Refer to the "Accelerating snnTorch on IPUs" tutorial for an example of how to do this.

A Deep Dive into SNNs

If you wish to learn all the fundamentals of training spiking neural networks, from neuron models, to the neural code, up to backpropagation, the snnTorch tutorial series is a great place to begin. It consists of interactive notebooks with complete explanations that can get you up to speed.

Tutorial Title Colab Link
Tutorial 1 Spike Encoding with snnTorch Open In Colab
Tutorial 2 The Leaky Integrate and Fire Neuron Open In Colab
Tutorial 3 A Feedforward Spiking Neural Network Open In Colab
Tutorial 4 2nd Order Spiking Neuron Models (Optional) Open In Colab
Tutorial 5 Training Spiking Neural Networks with snnTorch Open In Colab
Tutorial 6 Surrogate Gradient Descent in a Convolutional SNN Open In Colab
Tutorial 7 Neuromorphic Datasets with Tonic + snnTorch Open In Colab
Advanced Tutorials Colab Link
Population Coding Open In Colab

Contributing

If you're ready to contribute to snnTorch, instructions to do so can be found here.

Acknowledgments

snnTorch was initially developed by Jason K. Eshraghian in the Lu Group (University of Michigan).

Additional contributions were made by Xinxin Wang, Vincent Sun, and Emre Neftci.

Several features in snnTorch were inspired by the work of Friedemann Zenke, Emre Neftci, Doo Seok Jeong, Sumit Bam Shrestha and Garrick Orchard.

snnTorch is licensed under the GNU General Public License v3.0: https://www.gnu.org/licenses/gpl-3.0.en.html.

About

Deep and online learning with spiking neural networks in Python for Graphcore IPU

https://snntorch.readthedocs.io/en/latest/

License:Other


Languages

Language:Python 87.2%Language:C++ 11.4%Language:Makefile 1.4%