/gLund

Generative models for jet substructure.

Primary LanguagePython

DOI

gLund

This repository contains the code and results presented in arXiv:1909.01359.

About

gLund is a framework using the Lund jet plane to construct generative models for jet substructure.

Install gLund

gLund is tested and supported on 64-bit systems running Linux.

Install gLund with Python's pip package manager:

git clone https://github.com/JetsGame/gLund.git
cd gLund
pip install .

To install the package in a specific location, use the "--target=PREFIX_PATH" flag.

This process will copy the glund program to your environment python path.

We recommend the installation of the gLund package using a miniconda3 environment with the configuration specified here.

gLund requires the following packages:

  • python3
  • numpy
  • fastjet (compiled with --enable-pyext)
  • matplotlib
  • pandas
  • keras
  • tensorflow
  • json
  • gzip
  • argparse
  • scikit-image
  • scikit-learn
  • hyperopt (optional)

Pre-trained models

The final models presented in arXiv:1909.01359 are stored in:

  • results/lsgan: gLund LSGAN model trained on QCD jets (Pythia 8 + Delphes v3.4.1 fast detector simulation).
  • results/vae: gLund VAE model trained on QCD jets (Pythia 8 + Delphes v3.4.1 fast detector simulation).
  • results/wgangp: gLund WGAN-GP model trained on QCD jets (Pythia 8 + Delphes v3.4.1 fast detector simulation).

Input data

All data used for the final models can be downloaded from the git-lfs repository at https://github.com/JetsGame/data.

Running the code

In order to launch the code run:

glund --output <output_folder>  <runcard.yaml>

This will create a folder containing the result of the fit.

To create new samples from an existing model, as well as some diagnostic plots, use

glund_generate --save --ngen <number_to_generate> --output <result_file.npy> <model>

References

  • S. Carrazza and F. A. Dreyer, "Lund jet images from generative and cycle-consistent adversarial networks," arXiv:1909.01359