Ariadne: PyTorch library for particle track reconstruction using deep learning

Particle tracking is a fundamental part of the event analysis in high energy and nuclear physics. Events multiplicity increases each year along with drastic growth experimental data which modern HENP detectors produce, so classical algorithms such as well-known Kalman filter cannot satisfy speed scaling requirements. At same time, breakthroughs study deep learning open an opportunity for application high-performance neural networks solving problems dense environment experiments heavy ions. However, there are no well-documented software libraries track reconstruction yet. We introduce Ariadne, first open-source library particle based on PyTorch framework. The goal our to provide simple interface that allows one prepare train test datasets evaluate models implemented from your specific experiment. user experience greatly facilitated because system gin-configurations. modular structure abstract classes let develop his processing pipeline model easily. proposed facilitate academic research field learning.