Skip to content

Orkis-Research/Pytorch-Quaternion-Neural-Networks

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

34 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PyTorch-Quaternion-Neural-Networks

This repository offers up-to-date PyTorch implementations of various quaternion neural networks (QNN), such as QRNNs, QLSTMs, QCNNs, or QCAEs. Core components are situated within the core_qnn package, and can be reused to create custom QNNs, based on easy to customize PyTorch modules like QuaternionLinear, QuaternionConv, or QuaternionTransposeConv.

Requirements

  1. A GPU with a running CUDA installation is preferable. Please be certain that CUDA is correctly configured if you plan to use GPUs.
  2. Install PyTorch and torchvision depending on your environment: PyTorch
  3. Install others needed packages:
pip install imageio numpy scipy    
  1. Finally, install our core_qnn package:
pip install .   

Usage

Please navigate through the exp directory. Different running examples are proposed based on published papers. If you use this code or part of it, please cite the following paper:

Titouan Parcollet, Mirco Ravanelli, Mohamed Morchid, Georges Linarès, Chiheb Trabelsi, Renato De Mori, Yoshua Bengio - "Quaternion Recurrent Neural Networks", OpenReview

@inproceedings{
parcollet2018quaternion,
title={Quaternion Recurrent Neural Networks},
author={Titouan Parcollet and Mirco Ravanelli and Mohamed Morchid and Georges Linarès and Chiheb Trabelsi and Renato De Mori and Yoshua Bengio},
booktitle={International Conference on Learning Representations},
year={2019},
url={https://openreview.net/forum?id=ByMHvs0cFQ},
}

About

This repository is an update to all previous repositories with implementations of various Quaternion-valued Neural Networks in PyTorch

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages