Minimal ab Initio Neural Network Interatomic Engine
MINNIE is a C++/Python software and library that can be used to train an atomic Neural Network to reproduce ab-inito total energy and force calculations.
The focus of this library is the easy-to-use and extendibility through python scripting. The aim of the project is to make it interfaced with tools like ASE (Atomic Simulation Environment), to provide a fast and reliable calculator for energy, forces and stress tensors of condensed matter system as well as big organic molecules
MINNIE is a free software, and comes as it is, without any warranty.
THIS IS AN EXPERIMENTAL VERSION, DO NOT USE IT IF YOU DO NOT KNOW WHAT YOU ARE DOING
The MINNIE software requires python (>= 3.6), numpy and C/C++ compilers. To compile the C/C++ code you need:
- libconfig. On ubuntu, it is installed with
sudo apt install libconfig++-dev. - Gnu Scientific Library (GSL), with developer headers
- Python.h headers (
sudo apt install python-dev)
On ubuntu, the ingredients to compile the code can be installed with
sudo apt install libboost-all-dev libconfig++-dev libgsl-dev