abICS is a software framework for training a machine learning model to reproduce first-principles energies and then using the model to perform configurational sampling in disordered systems. Specific emphasis is placed on multi-component solid state systems such as metal and oxide alloys. The current version of abics can use neural network models implemented in aenet to be used as the machine learning model. As of this moment, abICS can also generate Quantum Espresso, VASP, and OpenMX input files for obtaining the reference training data for the machine learning model.
- python3 (>=3.9)
- numpy
- scipy
- toml (for parsing input files)
- mpi4py (for parallel tempering)
- This requires one of the MPI implementation
- pymatgen (>=2022.1.20) (for using Structure as a configuration)
- This requires Cython
- qe-tools (for parsing QE I/O)
Pymatgen requires Cython but Cython will not be installed automatically, please make sure that this is installed,
$ python3 -m pip install Cythonmpi4py requires one of the MPI implementations such as OpenMPI, please make sure that this is also installed. In the case of using homebrew on macOS, for example,
$ brew install open-mpiAfter installing Cython and MPI,
$ python3 -m pip install abicswill install abICS and dependencies.
If you want to change the directory where abICS is installed,
add --user option or --prefix=DIRECTORY option to the above command as
$ python3 -m pip install --user abicsFor details of pip , see the manual of pip by python3 -m pip help install
If you want to install abICS from source, see wiki page
The distribution of the program package and the source codes follow GNU General Public License version 3 (GPL v3).
We hope that you cite the following article when you publish the results using abICS.
Shusuke Kasamatsu, Yuichi Motoyama, Kazuyoshi Yoshimi, Tatsumi Aoyama, “Configuration sampling in multi-component multi-sublattice systems enabled by ab Initio Configuration Sampling Toolkit (abICS)”, accepted in STAM: Methods (arXiv:2309.04769.)
Bibtex:
@article{kasamatsu2023configuration,
author = {Shusuke Kasamatsu, Yuichi Motoyama, Kazuyoshi Yoshimi and Tatsumi Aoyama},
title = {Configuration sampling in multi-component multi-sublattice systems enabled by ab initio Configuration sampling toolkit ({abICS})},
journal = {Science and Technology of Advanced Materials: Methods},
volume = {0},
number = {ja},
pages = {2284128},
year = {2023},
publisher = {Taylor & Francis},
doi = {10.1080/27660400.2023.2284128},
URL = {https://doi.org/10.1080/27660400.2023.2284128},
eprint = {https://doi.org/10.1080/27660400.2023.2284128}
}
https://www.pasums.issp.u-tokyo.ac.jp/abics
Shusuke Kasamatsu, Yuichi Motoyama, Tatsumi Aoyama, Kazuyoshi Yoshimi