GDPy builds MLIP-driven workflows for atomistic simulation using advanced and newly developed structure-exploration methods. It integrates sampling, dataset construction, and model training into adaptive workflows.
Focused systems, with an emphasis on heterogeneous catalysis:
- metal oxides
- supported clusters
- disorder and amorphous surfaces
- solid–liquid interfaces
Highlighted methods:
- molecular dynamics (accelerated dynamics and enhanced sampling)
- monte carlo (canonical, semi-grand canonical, and grand canonical)
- global optimisation (structures, and reactions)
The package name is gdpx.
Install the nightly version from Git:
python -m pip install git+https://github.com/hsulab/GDPy.gitInstall a release with conda:
conda install gdpx -c conda-forge- Jiayan Xu (jxu15@qub.ac.uk)
- Prof. P. Hu (supervisor, Queen's University Belfast)