neuromorphic-paris/peregrine

Machine learning library for fast and efficient Gaussian mixture models

This repository contains code which implements the Stochastic Gaussian Mixture Model (S-GMM) for event-based datasets

Dependencies

• CMake
• Premake4
• Blaze
• Intel TBB
• pytorch 1.4+
• Tonic

Running a classification task

1. Generate time-surfaces using: time_surfaces.ipynb (look at tonic for more details on what datasets are support)
2. Compile C++14 source code

premake4 gmake && cd build && make

3. Main entry point for clustering and classification: variational-gmm.ipynb

Installation

Tonic

pip install tonic

CMake and Premake4

On Mac

You can install cmake and premake via the homebrew package manager

/usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
brew install cmake premake

On Linux

sudo apt-get install cmake premake4

Blaze

Make sure you have BLAS and LAPACK installed first

sudo apt-get install libopenblas-dev

Proceed to the blaze installation

git clone https://bitbucket.org/blaze-lib/blaze.git
cd blaze && cmake -DCMAKE_INSTALL_PREFIX=/usr/local/
sudo make install

Intel TBB

On Mac

You can install intel TBB via the homebrew package manager

brew install tbb

On Linux

sudo apt install libtbb-dev