HotPP
is an open-source package designed for constructing message passing network interatomic potentials. It facilitates the utilization of arbitrary order Cartesian tensors as messages while maintaining equivalence maintenance.
- Building machine learning potentials for molecular and periodic systems;
- Learning dipole moments and polarizability tensors;
- Interface to LAMMPS and ASE;
- An overview of code documentation and tutorials for getting started with
HotPP
can be found here folder.
You can use https:
$ pip install git+https://gitlab.com/bigd4/hotpp.git
or use ssh
$ pip install git+ssh://git@gitlab.com/bigd4/hotpp.git
Your may need to add --user
if you do not have the root permission. Or use --force-reinstall
if you already have HotPP
(add --no-dependencies
if you do not want to reinstall the dependencies).
- Use git clone to get the source code:
$ git clone https://gitlab.com/bigd4/hotpp.git
Alternatively, you can download the source code from website.
- Go into the directory and install with pip:
$ pip install -e .
pip will read setup.py in your current directory and install. The -e
option means python will directly import the module from the current path, but not copy the codes to the default lib path and import the module there, which is convenient for modifying in the future. If you do not have the need, you can remove the option.
You can use
$ hotpp -v
to check if you have installed successfully
If you installed by pip, use:
$ hotpp update
If you installed from source, use:
$ cd <path-to-magus-package>
$ git pull origin master
HotPP
now support ASE and lammps.
HotPP is developed by Prof. Jian Sun's group at the School of Physics at Nanjing University.
The contributors are:
- Jian Sun
- Junjie Wang
- Yong Wang
- Haoting Zhang
- Ziyang Yang
- Zhixin Liang
- Jiuyang Shi
Reference | cite for what |
---|---|
[1] | for any work that used HotPP |
[1] 1. Wang, J. et al. E(n)-Equivariant Cartesian Tensor Passing Potential. Preprint at http://arxiv.org/abs/2402.15286 (2024). (https://arxiv.org/abs/2402.15286)