Skip to main content

An implementation of indexed convolution and pooling

Project description

Indexed Convolution

The indexed operations allow the user to perform convolution and pooling on non-Euclidian grids of data given that the neighbors pixels of each pixel is known and provided.

It gives an alternative to masking or resampling the data in order to apply standard Euclidian convolution. This solution has been developed in order to apply convolutional neural networks to data from physics experiments that propose specific pixels arrangements.

It is used in the GammaLearn project <>_ for the Cherenkov Telescope Array.

Here you will find the code for the indexed operations as well as applied examples. The current implementation has been done for pytorch.

Documentation may be found online. <>_

.. image:: :target: .. image:: :target: :alt: Documentation Status .. image:: :target:


Install from IndexedConv folder:

.. code-block:: bash

python install

Install with pip:

.. code-block:: bash

pip install indexedconv

Install with conda:

first add conda-forge channel for tensorboardX if needed

.. code-block:: bash

conda config --append channels conda-forge

.. code-block:: bash

conda install -c gammalearn indexedconv


.. code-block:: bash


Running an experiment

For example, to train the network with indexed convolution on the CIFAR10 dataset transformed to hexagonal:

.. code-block:: bash

python examples/ main_folder data_folder experiment_name --hexa --batch 125 --epochs 300 --seeds 1 2 3 4 --device cpu

In order to train on the AID dataset, it must be downloaded and can be found here <>_.


The development of the indexed convolution is born from a collaboration between physicists and computer scientists.

  • Luca Antiga, Orobix
  • Mikael Jacquemont, LAPP (CNRS), LISTIC (USMB)
  • Thomas Vuillaume, LAPP (CNRS)


If you want to use IndexedConv, please cite:

.. image:: :target:


All contributions are welcome.

Start by contacting the authors, either directly by email or by creating a GitHub issue.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for IndexedConv, version 1.1
Filename, size & hash File type Python version Upload date
IndexedConv-1.1-py3-none-any.whl (16.7 kB) View hashes Wheel py3

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page