Skip to content

DeepIceLearning - An Approach to use Deep Neural Networks for the Regression and Classification of IceCube quantities

License

Notifications You must be signed in to change notification settings

coenders/NN_Reco

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DeepIceLearning

DeepIceLearning - An Approach to use Deep Neural Networks for the Regression and Classification of IceCube quantities

Firstly: This version of the software is still in test-mode and might contain many bugs or inconsistencies. As input you can use any files that are produced with the Create_Data_Files.py script in the main directory. A set of muon neutrino events can be found in my user directory under /data/user/tglauch/ML_Reco/training_data/

How to train a neural network with this code? The goal of this software is to provide an easy usable framework for the training of neural network on IceCube data. In order to start the training, mainly two files have to be changed.

  1. The config.cfg defines all the variables for the local environment, as well as for the training. Here you can also set the path to the training files.

  2. In the folder 'Networks' a config file defining the NN structure has to be created. An example for the syntax can be found in the file 'test.cfg'

In order to run the NN's training you can either:

  1. directly run the script Neural_Network.py or
  2. submit a job to the condor or slurm cluster by running condor_submit/train_NN_network.py

About

DeepIceLearning - An Approach to use Deep Neural Networks for the Regression and Classification of IceCube quantities

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 72.0%
  • Python 24.5%
  • Shell 3.5%