10.6. Interfaces out of DeePMD-kit#
The codes of the following interfaces are not a part of the DeePMD-kit package and maintained by other repositories. We list these interfaces here for user convenience.
10.6.1. Plugins#
10.6.1.1. External GNN models (MACE/NequIP)#
DeePMD-GNN is DeePMD-kit plugin for various graph neural network (GNN) models. It has interfaced with MACE (PyTorch version) and NequIP (PyTorch version). It is also the first example to the DeePMD-kit plugin mechanism.
10.6.2. C/C++ interface used by other packages#
10.6.2.1. Third-party GROMACS interface to DeePMD-kit#
A third-party GROMACS interface to DeePMD-kit is available outside this repository at HuXioAn/gromacs/tree/deepmd-oneModel. It is based on the GROMACS Neural Network Potentials (NNPot) infrastructure and is described in Enabling AI Deep Potentials for Ab Initio-quality Molecular Dynamics Simulations in GROMACS.
According to that implementation and paper, this interface supports
DeePMD-kit inference through the C++/CUDA backend;
multiple DeePMD model families, including
se_e2_a,DPA,DPA2, andDPA3;hybrid workflows where DeePMD-kit is applied to selected atom groups inside a GROMACS simulation.
The reported examples use protein-in-water systems, where DeePMD-kit is applied to the protein internal interactions while water and protein-water interactions remain classical.
Users should also be aware of the current scope reported by the third-party project:
the published benchmarks enable DeePMD only in the production MD stage, not in EM/NVT/NPT;
the reported implementation uses single-rank inference in the current GROMACS NNPot workflow;
scalability and domain-decomposed inference are described as future optimization targets;
some DPA3 benchmark cases run out of GPU memory on the tested hardware.
This interface is maintained outside DeePMD-kit. Please refer to the corresponding third-party repository for installation instructions, supported GROMACS versions, and runtime details.
10.6.2.2. OpenMM plugin for DeePMD-kit#
An OpenMM plugin is provided from JingHuangLab/openmm_deepmd_plugin, written by the Huang Lab at Westlake University.
10.6.2.3. Amber interface to DeePMD-kit#
Starting from AmberTools24, sander includes an interface to the DeePMD-kit, which implements the Deep Potential Range Corrected (DPRc) correction. The DPRc model and the interface were developed by the York Lab from Rutgers University. More details are available in
Amber Reference Manuals, providing documentation for how to enable the interface and the
&dprcnamelist;GitLab RutgersLBSR/AmberDPRc, providing examples mdin files;
DP-Amber, a tiny tool to convert Amber trajectory to DPRc training data;
10.6.2.4. CP2K interface to DeePMD-kit#
CP2K v2024.2 adds an interface to the DeePMD-kit for molecular dynamics. Read the CP2K manual for details.
10.6.2.5. ABACUS#
ABACUS can run molecular dynamics with a DP model. User is required to build ABACUS with DeePMD-kit.
10.6.3. Command line interface used by other packages#
10.6.3.1. DP-GEN#
DP-GEN provides a workflow to generate accurate DP models by calling DeePMD-kit’s command line interface (CLI) in the local or remote server. Details can be found in this paper.
10.6.3.2. MLatom#
Mlatom provides an interface to the DeePMD-kit within MLatom’s workflow by calling DeePMD-kit’s CLI. Details can be found in this paper.