A PyTorch implementation of EventProp [https://arxiv.org/abs/2009.08378], a method to train Spiking Neural Networks

Overview

Spiking Neural Network training with EventProp

This is an unofficial PyTorch implemenation of EventProp, a method to compute exact gradients for Spiking Neural Networks. The repo currently contains code to train a 1-layer Spiking Neural Network with leaky integrate-and-fire (LIF) neurons for 10-way digit classification on MNIST.

Implementation Details

The implementation of EventProp itself is in models.py, in form of the forward and backward methods of the SpikingLinear module, which compute the forward passes of a spiking layer and its adjoint layer.

In particular, the manual_forward method computes the discretized dynamics of a spiking layer:

While the manual_backward method computes the discretized dynamics of the adjoint model, used to compute exact gradients for the weight parameters:

The network is run for a fixed amount of time and discrete time steps are used to approximate the continuous dynamics. These can be set through the T and dt arguments when running main.py (default values are T=40ms and dt=1ms, so a total of 40 forward passes are executed for each mini-batch).

To encode the MNIST dataset as spikes, images were first binarized and black/white pixels were encoded as spikes at times 10/20ms, respectively. The dynamics of one of the 10 output neurons are as follows, for a randomly-initialized network:

where vertical black lines indicate spike times.

Usage

The code was tested with Python 2.7 + PyTorch 0.4 and Python 3.8 + PyTorch 1.4, producing similar results.

To train the SNN with default settings, just run

python main.py

which will automatically download MNIST and train a SNN for 40 epochs with Adam, on gpu.

Check out the available args in main.py to change training settings such as the learning rate, batch size, and SNN-specific parameters such as membrane/synaptic constants and time discretization.

The default hyperparameters result in stable training, reaching around 85% train/test accuracy in under 10 epochs:

Extensions

If there is enough interest, I can try to extend the EventProp implementation to handle hidden layers / convolutions. If you'd like to extend it yourself, feel free to submit a pull request.

Owner
Pedro Savarese
PhD student at TTIC
Pedro Savarese
An SMPC companion library for Syft

SyMPC A library that extends PySyft with SMPC support SyMPC /ˈsɪmpəθi/ is a library which extends PySyft ≥0.3 with SMPC support. It allows computing o

Arturo Marquez Flores 0 Oct 13, 2021
Official PyTorch implementation of Retrieve in Style: Unsupervised Facial Feature Transfer and Retrieval.

Retrieve in Style: Unsupervised Facial Feature Transfer and Retrieval PyTorch This is the PyTorch implementation of Retrieve in Style: Unsupervised Fa

60 Oct 12, 2022
Official Pytorch Implementation of 'Learning Action Completeness from Points for Weakly-supervised Temporal Action Localization' (ICCV-21 Oral)

Learning-Action-Completeness-from-Points Official Pytorch Implementation of 'Learning Action Completeness from Points for Weakly-supervised Temporal A

Pilhyeon Lee 67 Jan 03, 2023
A new codebase for Group Activity Recognition. It contains codes for ICCV 2021 paper: Spatio-Temporal Dynamic Inference Network for Group Activity Recognition and some other methods.

Spatio-Temporal Dynamic Inference Network for Group Activity Recognition The source codes for ICCV2021 Paper: Spatio-Temporal Dynamic Inference Networ

40 Dec 12, 2022
[AAAI-2021] Visual Boundary Knowledge Translation for Foreground Segmentation

Trans-Net Code for (Visual Boundary Knowledge Translation for Foreground Segmentation, AAAI2021). [https://ojs.aaai.org/index.php/AAAI/article/view/16

ZJU-VIPA 2 Mar 04, 2022
Simple streamlit app to demonstrate HERE Tour Planning

Table of Contents About the Project Built With Getting Started Prerequisites Installation Usage Roadmap Contributing License Acknowledgements About Th

Amol 8 Sep 05, 2022
Realtime YOLO Monster Detection With Non Maximum Supression

Realtime-YOLO-Monster-Detection-With-Non-Maximum-Supression Table of Contents In

5 Oct 07, 2022
MADE (Masked Autoencoder Density Estimation) implementation in PyTorch

pytorch-made This code is an implementation of "Masked AutoEncoder for Density Estimation" by Germain et al., 2015. The core idea is that you can turn

Andrej 498 Dec 30, 2022
Implementations of paper Controlling Directions Orthogonal to a Classifier

Classifier Orthogonalization Implementations of paper Controlling Directions Orthogonal to a Classifier , ICLR 2022, Yilun Xu, Hao He, Tianxiao Shen,

Yilun Xu 33 Dec 01, 2022
Implementing SYNTHESIZER: Rethinking Self-Attention in Transformer Models using Pytorch

Implementing SYNTHESIZER: Rethinking Self-Attention in Transformer Models using Pytorch Reference Paper URL Author: Yi Tay, Dara Bahri, Donald Metzler

Myeongjun Kim 66 Nov 30, 2022
Magic tool for managing internet connection in local network by @zalexdev

Megacut ✂️ A new powerful Python3 tool for managing internet on a local network Installation git clone https://github.com/stryker-project/megacut cd m

Stryker 12 Dec 15, 2022
Rendering Point Clouds with Compute Shaders

Compute Shader Based Point Cloud Rendering This repository contains the source code to our techreport: Rendering Point Clouds with Compute Shaders and

Markus Schütz 460 Jan 05, 2023
A package related to building quasi-fibration symmetries

qf A package related to building quasi-fibration symmetries. If you'd like to learn more about how it works, see the brief explanation and References

Paolo Boldi 1 Dec 01, 2021
Adaptive Pyramid Context Network for Semantic Segmentation (APCNet CVPR'2019)

Adaptive Pyramid Context Network for Semantic Segmentation (APCNet CVPR'2019) Introduction Official implementation of Adaptive Pyramid Context Network

21 Nov 09, 2022
LEDNet: A Lightweight Encoder-Decoder Network for Real-time Semantic Segmentation

LEDNet: A Lightweight Encoder-Decoder Network for Real-time Semantic Segmentation Table of Contents: Introduction Project Structure Installation Datas

Yu Wang 492 Dec 02, 2022
Direct application of DALLE-2 to video synthesis, using factored space-time Unet and Transformers

DALLE2 Video (wip) ** only to be built after DALLE2 image is done and replicated, and the importance of the prior network is validated ** Direct appli

Phil Wang 105 May 15, 2022
Inferring Lexicographically-Ordered Rewards from Preferences

Inferring Lexicographically-Ordered Rewards from Preferences Code author: Alihan Hüyük ([e

Alihan Hüyük 1 Feb 13, 2022
O2O-Afford: Annotation-Free Large-Scale Object-Object Affordance Learning (CoRL 2021)

O2O-Afford: Annotation-Free Large-Scale Object-Object Affordance Learning Object-object Interaction Affordance Learning. For a given object-object int

Kaichun Mo 26 Nov 04, 2022
Context-Aware Image Matting for Simultaneous Foreground and Alpha Estimation

Context-Aware Image Matting for Simultaneous Foreground and Alpha Estimation This is the inference codes of Context-Aware Image Matting for Simultaneo

Qiqi Hou 125 Oct 22, 2022