DeePMD-kit is a package written in Python/C++, designed to minimize the effort required to build deep learning-based models of interatomic potential energy and force field and to perform molecular dynamics (MD). This brings new hopes to addressing the accuracy-versus-efficiency dilemma in molecular simulations. Applications of DeePMD-kit span from finite molecules to extended systems and from metallic systems to chemically bonded systems.
Important
The project DeePMD-kit is licensed under GNU LGPLv3.0. If you use this code in any future publications, please cite this using Han Wang, Linfeng Zhang, Jiequn Han, and Weinan E. "DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics." Computer Physics Communications 228 (2018): 178-184.
getting-started/index
install/index data/index model/index train/index freeze/index test/index inference/index cli third-party/index nvnmd/index troubleshooting/index
Tutorials <https://tutorials.deepmodeling.com/> Publications <https://deepmodeling.com/blog/papers/deepmd-kit/>
development/* api_py/api_py api_op API_CC/api_cc api_c/api_c api_core/api_core
license credits logo
genindexmodindexsearch