Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/electronicvisions/norse

Deep learning for spiking neural networks
https://github.com/electronicvisions/norse

deep-learning machine-learning neuromorphic-computing python pytorch spiking-neural-networks tensors

Last synced: 3 months ago
JSON representation

Deep learning for spiking neural networks

Awesome Lists containing this project

README 

        







A [deep learning](https://en.wikipedia.org/wiki/Deep_learning) library for [spiking neural networks](https://en.wikipedia.org/wiki/Spiking_neural_network).





    

        Test status

    

        

    

    

        

    

    

        

    

    

        chat on Discord

    

    

    DOI




Norse aims to exploit the advantages of bio-inspired neural components, which are sparse and event-driven - a fundamental difference from artificial neural networks.

Norse expands [PyTorch](https://pytorch.org/) with primitives for bio-inspired neural components, 

bringing you two advantages: a modern and proven infrastructure based on PyTorch and deep learning-compatible spiking neural network components.



**Documentation**: [norse.github.io/norse/](https://norse.github.io/norse/)



## 1. Getting started



The fastest way to try Norse is via the [jupyter notebooks on Google collab](https://github.com/norse/notebooks/tree/master/).



Alternatively, [you can install Norse locally](#installation) and run one of the [included tasks](https://norse.github.io/norse/tasks.html) such as [MNIST](https://en.wikipedia.org/wiki/MNIST_database):

```bash

python -m norse.task.mnist

```



## 2. Using Norse



Norse presents plug-and-play components for deep learning with spiking neural networks.

Here, we describe how to install Norse and start to apply it in your own work.

[Read more in our documentation](https://norse.github.io/norse/working.html).



### 2.1. Installation




We assume you are using **Python version 3.7+** and have **installed PyTorch version 1.9 or higher**. 

[Read more about the prerequisites in our documentation](https://norse.github.io/norse/installing.html).



MethodInstructionsPrerequisites



From PyPi

pip install norse

Pip



From source

pip install -qU git+https://github.com/norse/norse

Pip, PyTorch



With Docker

docker pull quay.io/norse/norse

Docker



From Conda 

conda install -c norse norse

Anaconda or Miniconda



For troubleshooting, please refer to our [installation guide](https://norse.github.io/norse/pages/installing.html#installation-troubleshooting), [create an issue on GitHub](https://github.com/norse/norse/issues) or [write us on Discord](https://discord.gg/7fGN359).



### 2.2. Running examples



Norse is bundled with a number of example tasks, serving as short, self contained, correct examples ([SSCCE](http://www.sscce.org/)).

They can be run by invoking the `norse` module from the base directory.

More information and tasks are available [in our documentation](https://norse.github.io/norse/tasks.html) and in your console by typing: `python -m norse.task. --help`, where `` is one of the task names.



- To train an MNIST classification network, invoke

    ```bash

    python -m norse.task.mnist

    ```

- To train a CIFAR classification network, invoke

    ```bash

    python -m norse.task.cifar10

    ```

- To train the cartpole balancing task with Policy gradient, invoke

    ```bash

    python -m norse.task.cartpole

    ```



Norse is compatible with [PyTorch Lightning](https://pytorch-lightning.readthedocs.io/en/stable/),

as demonstrated in the [PyTorch Lightning MNIST task variant](https://github.com/norse/norse/blob/master/norse/task/mnist_pl.py) (requires PyTorch lightning):



```bash

python -m norse.task.mnist_pl --gpus=4

```



### 2.3. Example: Spiking convolutional classifier 



[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/norse/notebooks/blob/master/mnist_classifiers.ipynb)



This classifier is a taken from our [tutorial on training a spiking MNIST classifier](https://github.com/norse/notebooks#level-intermediate) and achieves >99% accuracy.



```python

import torch, torch.nn as nn

from norse.torch import LICell             # Leaky integrator

from norse.torch import LIFCell            # Leaky integrate-and-fire

from norse.torch import SequentialState    # Stateful sequential layers



model = SequentialState(

    nn.Conv2d(1, 20, 5, 1),      # Convolve from 1 -> 20 channels

    LIFCell(),                   # Spiking activation layer

    nn.MaxPool2d(2, 2),

    nn.Conv2d(20, 50, 5, 1),     # Convolve from 20 -> 50 channels

    LIFCell(),

    nn.MaxPool2d(2, 2),

    nn.Flatten(),                # Flatten to 800 units

    nn.Linear(800, 10),

    LICell(),                    # Non-spiking integrator layer

)



data = torch.randn(8, 1, 28, 28) # 8 batches, 1 channel, 28x28 pixels

output, state = model(data)      # Provides a tuple (tensor (8, 10), neuron state)

```



### 2.4. Example: Long short-term spiking neural networks

The long short-term spiking neural networks from the paper by [G. Bellec, D. Salaj, A. Subramoney, R. Legenstein, and W. Maass (2018)](https://arxiv.org/abs/1803.09574) is another interesting way to apply norse: 

```python

import torch

from norse.torch import LSNNRecurrent

# Recurrent LSNN network with 2 input neurons and 10 output neurons

layer = LSNNRecurrent(2, 10)

# Generate data: 20 timesteps with 8 datapoints per batch for 2 neurons

data  = torch.zeros(20, 8, 2)

# Tuple of (output spikes of shape (20, 8, 2), layer state)

output, new_state = layer(data)

```



## 3. Why Norse?



Norse was created for two reasons: to 1) apply findings from decades of research in practical settings

and to 2) accelerate our own research within bio-inspired learning.



We are passionate about Norse: we strive to follow best practices and promise to maintain this library for the

simple reason that we depend on it ourselves.

We have implemented a number of neuron models, synapse dynamics, encoding and decoding algorithms, 

dataset integrations, tasks, and examples.

Combined with the PyTorch infrastructure and our high coding standards, we have found Norse to be an excellent tool for modelling *scaleable* experiments and [Norse is actively being used in research](https://norse.github.io/norse/papers.html).



Finally, we are working to keep Norse as performant as possible. 

Preliminary benchmarks suggest that Norse [achieves excellent performance on small networks of up to ~5000 neurons per layer](https://github.com/norse/norse/tree/master/norse/benchmark). 

Aided by the preexisting investment in scalable training and inference with PyTorch, Norse scales from a single laptop to several nodes on an HPC cluster with little effort.

As illustrated by our [PyTorch Lightning example task](https://norse.github.io/norse/tasks.html#mnist-in-pytorch-lightning).



[Read more about Norse in our documentation](https://norse.github.io/norse/about.html).



## 4. Similar work

The list of projects below serves to illustrate the state of the art, while explaining our own incentives to create and use norse.



* [BindsNET](https://github.com/BindsNET/bindsnet) also builds on PyTorch and is explicitly targeted at machine learning tasks. It implements a Network abstraction with the typical 'node' and 'connection' notions common in spiking neural network simulators like nest.

* [cuSNN](https://github.com/tudelft/cuSNN) is a C++ GPU-accelerated simulator for large-scale networks. The library focuses on CUDA and includes spike-time dependent plasicity (STDP) learning rules.

* [decolle](https://github.com/nmi-lab/decolle-public) implements an online learning algorithm described in the paper ["Synaptic Plasticity Dynamics for Deep Continuous Local Learning (DECOLLE)"](https://arxiv.org/abs/1811.10766) by J. Kaiser, M. Mostafa and E. Neftci.

* [GeNN](http://genn-team.github.io/genn/) compiles SNN network models to NVIDIA CUDA to achieve high-performing SNN model simulations.

* [Long short-term memory Spiking Neural Networks (LSNN)](https://github.com/IGITUGraz/LSNN-official) is a tool from the University of Graaz for modelling LSNN cells in [Tensorflow](https://www.tensorflow.org/). The library focuses on a single neuron and gradient model.

* [Nengo](https://www.nengo.ai/nengo-dl/introduction.html) is a neuron simulator, and Nengo-DL is a deep learning network simulator that optimised spike-based neural networks based on an approximation method suggested by [Hunsberger and Eliasmith (2016)](https://arxiv.org/abs/1611.05141).

* [Nengo PyTorch](https://github.com/nengo/pytorch-spiking) a thin wrapper for PyTorch that adds a single voltage-only spiking model. The approach is independent from the Nengo framework.

* [Neuron Simulation Toolkit (NEST)](https://nest-simulator.org) constructs and evaluates highly detailed simulations of spiking neural networks. This is useful in a medical/biological sense but maps poorly to large datasets and deep learning.

* [PyNN](http://neuralensemble.org/docs/PyNN/) is a Python interface that allows you to define and simulate spiking neural network models on different backends (both software simulators and neuromorphic hardware). It does not currently provide mechanisms for optimisation or arbitrary synaptic plasticity.

* [PySNN](https://github.com/BasBuller/PySNN/) is a PyTorch extension similar to Norse. Its approach to model building is slightly different than Norse in that the neurons are stateful.

* [Rockpool](https://gitlab.com/aiCTX/rockpool) is a Python package developed by SynSense for training, simulating and deploying spiking neural networks. It offers both JAX and PyTorch primitives.

* [Sinabs](https://gitlab.com/synsense/sinabs) is a PyTorch extension by [SynSense](https://www.synsense-neuromorphic.com/). It mainly focuses on convolutions and translation to neuromorphic hardware.

* [SlayerPyTorch](https://github.com/bamsumit/slayerPytorch) is a **S**pike **LAY**er **E**rror **R**eassignment library, that focuses on solutions for the temporal credit problem of spiking neurons and a probabilistic approach to backpropagation errors. It includes support for the [Loihi chip](https://en.wikichip.org/wiki/intel/loihi).

* [SNN toolbox](https://snntoolbox.readthedocs.io/en/latest/guide/intro.html) automates the conversion of pre-trained analog to spiking neural networks. The tool is solely for already trained networks and omits the (possibly platform specific) training.

* [snnTorch](https://snntorch.readthedocs.io/en/latest/) is a simulator built on PyTorch, featuring several [introduction tutorials](https://snntorch.readthedocs.io/en/latest/tutorials/index.html) on deep learning with SNNs.

* [SpikingJelly](https://github.com/fangwei123456/spikingjelly) is another PyTorch-based spiking neural network simulator. SpikingJelly uses stateful neurons. [Example of training a network on MNIST](https://spikingjelly.readthedocs.io/zh_CN/latest/clock_driven_en/3_fc_mnist.html).

* [SpyTorch](https://github.com/fzenke/spytorch) presents a set of tutorials for training SNNs with the surrogate gradient approach SuperSpike by [F. Zenke, and S. Ganguli (2017)](https://arxiv.org/abs/1705.11146). Norse [implements the SuperSpike surrogate gradient function](https://github.com/norse/norse/blob/master/norse/torch/functional/superspike.py), but allows for other surrogate gradients and training approaches.

* [s2net](https://github.com/romainzimmer/s2net) is based on the implementation presented in [SpyTorch](https://github.com/fzenke/spytorch), but implements convolutional layers as well. It also contains a demonstration how to use those primitives to train a model on the [Google Speech Commands dataset](https://arxiv.org/abs/1804.03209).



## 5. Contributing



Contributions are warmly encouraged and always welcome. However, we also have high expectations around the code base so if you wish to contribute, please refer to our [contribution guidelines](contributing.md).



## 6. Credits



Norse is created by

* [Christian Pehle](https://www.kip.uni-heidelberg.de/people/10110) (@GitHub [cpehle](https://github.com/cpehle/)), PostDoc at University of Heidelberg, Germany.

* [Jens E. Pedersen](https://www.kth.se/profile/jeped) (@GitHub [jegp](https://github.com/jegp/)), doctoral student at KTH Royal Institute of Technology, Sweden.



More information about Norse can be found [in our documentation](https://norse.github.io/norse/about.html). The research has received funding from the EC Horizon 2020 Framework Programme under Grant Agreements 785907 and 945539 (HBP) and by the Deutsche Forschungsgemeinschaft (DFG, German Research Fundation) under Germany's Excellence Strategy EXC 2181/1 - 390900948 (the Heidelberg STRUCTURES Excellence Cluster).



## 7. Citation



If you use Norse in your work, please cite it as follows:



```BibTex

@software{norse2021,

  author       = {Pehle, Christian and

                  Pedersen, Jens Egholm},

  title        = {{Norse -  A deep learning library for spiking 

                   neural networks}},

  month        = jan,

  year         = 2021,

  note         = {Documentation: https://norse.ai/docs/},

  publisher    = {Zenodo},

  version      = {0.0.7},

  doi          = {10.5281/zenodo.4422025},

  url          = {https://doi.org/10.5281/zenodo.4422025}

}

```



Norse is actively applied and cited in the literature. We are keeping track of the papers cited by Norse [in our documentation](https://norse.github.io/norse/papers.html).



## 8. License



LGPLv3. See [LICENSE](LICENSE) for license details.