DeePMD-kit v3: A Multiple-Backend Framework for Machine Learning Potentials
University of Science and Technology of China · Peking University · +33 more institutions
Abstract
In recent years, machine learning potentials (MLPs) have become indispensable tools in physics, chemistry, and materials science, driving the development of software packages for molecular dynamics (MD) simulations and related applications. These packages, typically built on specific machine learning frameworks, such as TensorFlow, PyTorch, or JAX, face integration challenges when advanced applications demand communication across different frameworks. The previous TensorFlow-based implementation of the DeePMD-kit exemplified these limitations. In this work, we introduce DeePMD-kit version 3, a significant update featuring a multibackend framework that supports TensorFlow, PyTorch, JAX, and PaddlePaddle…
Citation impact
- FWCI
- 25.30
- Percentile
- 100%
- References
- 61
Authors
47- JZJinzhe ZengCorresponding
University of Science and Technology of China
- DZDuo Zhang
Peking University, Art Institute of Portland, Center for Interdisciplinary Studies, SP Technology (South Korea)
- APAnyang Peng
Art Institute of Portland
- XZXiangyu Zhang
Chinese Academy of Sciences, Institute of Computing Technology, University of Chinese Academy of Sciences
- SZS Z He
Baidu (China)
Topics & keywords
- Computer science
- Human–computer interaction
- Data science
Funding
- EHEuropean High Performance Computing Joint UndertakingAward: EHPC-REG-2023R02-088
- NSNational Science FoundationAward: 2209718
- NNNational Natural Science Foundation of ChinaAwards: 12122103, 92270206
- CSChina Scholarship Council
- NFNorges ForskningsrådAwards: 262695, 344993
- XUXiamen University
- SAScience and Technology Program of Hunan ProvinceAward: 2021RC4026
- NKNational Key Research and Development Program of ChinaAward: 2022YFA1004300
- NINational Institute of General Medical SciencesAward: GM107485