Skip to content

GANs for simulation of electromagnetic showers in the ATLAS calorimeter

Notifications You must be signed in to change notification settings

polklin/CERN_project

Repository files navigation

CERN_project

GANs for simulation of electromagnetic showers in the ATLAS calorimeter. Details of my project are in the file Report.pdf

  1. MNIST.py

First implementation of Improved Wasserstein GAN (WGAN-gp), on the MNIST dataset. See the article : https://arxiv.org/abs/1704.00028

  1. WGAN for electromagnetic showers in the ATLAS calorimeter A first script - PreProcessing.py - applies transformations to the former dataset (noise cuts, remove events>300 GeV), as describe in the file report.pdf.

Three Python files that work together :

  • config.py (contains training parameters, path to files, ...)
  • plot_functions.py (where all plot functions are defined)
  • training.py (main file to train the WGAN)

To launch a training, you should write in the terminal : python training.py Name (Name = name of your folder in which all plots and weights will be saved. Will create the folder if it doens't already exist)

Plots are generated automatically each 250 epochs, a folder is created each time.

  1. A Jupyter notebook called Physics_variates.ipynb

Interactive plots

About

GANs for simulation of electromagnetic showers in the ATLAS calorimeter

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published