TrackFormers - Machine Learning Pipelines
[Contributors: Sascha Caron, José D. Martin-Guerrero, Uraz Odyurt, Roberto Ruiz de Austri Bazan]
Abstract
TrackFormers is a machine learning framework for track reconstruction in particle physics
experiments. It leverages transformer- and U-Net-inspired deep learning architectures to predict
particle tracks from hit data.
This repository contains 4 directories corresponding to the 4 models described in the paper
TrackFormers: In Search of Transformer-Based Particle Tracking for the High-Luminosity LHC Era.
EncDec, EncCla, and EncReg are transformer-based models, whereas U-Net is, as the name suggests, a
U-Net model.
Refer to the provided README file for further details.
Metadata
Links
Licence
Artefacts shared as PDF are licenced under
CC BY 4.0.