Code

Framework

Source code openly available at: https://gitlab.in2p3.fr/gammalearn/gammalearn

To ease developments and experiments running, we have developed a complete environment allowing the user to reproduce previous results test new networks test many hyperparameters compare the results of these experiments. It allows the user to easily:

load datasets
pre-process data (filter, augment, transform)
train, validate and test networks
monitor the training process
visualize training results

You can learn more about it in:

Mikaël Jacquemont, Thomas Vuillaume, A Benoit, Gilles Maurin, Patrick Lambert, et al - GammaLearn: a Deep Learning framework for IACT data. 36th International Cosmic Ray Conference, Jul 2019, Madison, United States. pp.705, https://pos.sissa.it/358/705/

GammaLearn framework

Indexed Convolution

Standard convolution in frameworks such as Pytorch or Tensorflow has been developed for Cartesian lattices adapted to pixel grids in standard images. However, in many scientific experiments, sensors are arranged with different lattices. This is also the case for the Cherenkov Telescope Array cameras. We developed convolution and pooling kernels to be used in Pytorch in order to be able to apply these operations on any data given that all pixel neighbours are known and provided.

Paper: Jacquemont, M.; Antiga, L.; Vuillaume, T.; Silvestri, G.; Benoit, A.; Lambert, P. and Maurin, G. (2019). Indexed Operations for Non-rectangular Lattices Applied to Convolutional Neural Networks. In Proceedings of the 14th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 5 VISAPP: VISAPP, ISBN 978-989-758-354-4, ISSN 2184-4321, pages 362-371. DOI: 10.5220/0007364303620371

Code: Mikaël Jacquemont, & Thomas Vuillaume. (2021, January 5). IndexedConv/IndexedConv: v1.3 (Version v1.3). Zenodo.