# deepmodeling/deepmd-kit

A deep learning package for many-body potential energy representation and molecular dynamics

Repository: https://github.com/deepmodeling/deepmd-kit
Canonical: https://ross.abutalabs.com/products/deepmd-kit
Homepage: https://docs.deepmodeling.com/projects/deepmd/
Language: Python
License: LGPL-3.0
License Family: copyleft
Topics: deep-learning, molecular-dynamics, deepmd, lammps, potential-energy, python, tensorflow, cpp, cuda, rocm, ipi, ase, computational-chemistry, materials-science, c, nodejs, pytorch, jax, paddle, machine-learning-potential
Last push: 2026-08-24T20:49:42+00:00

## Health v2 (maintenance only)
Score: 95/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 86, longevity 100
- inputs: {"age_days": 3186, "days_push": 9, "days_rel": 14, "gap_med": 77.0, "n_releases_24m": 9}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2021, forks 646 (observed 2026-08-28T04:06:06.328268+00:00)

## What it is
DeePMD-kit is a deep learning package for building many-body potential energy representations and running molecular dynamics simulations. It converts quantum-mechanical reference data into fast, scalable interatomic potentials, with pretrained models, fine-tuning, and deployment via LAMMPS, i-PI, and ASE.

## Use cases
- train machine learning interatomic potentials from DFT data
- run molecular dynamics simulations with neural network potentials
- fine-tune a pretrained foundation model for my chemical system
- simulate periodic solids, metals, and molecules at scale
- export a trained potential to LAMMPS for MPI-parallel MD
- model potential energy surfaces for materials science

## When to choose
- you need accurate, fast interatomic potentials for molecular or materials simulations
- you want to fine-tune pretrained models like DPA4 instead of training from scratch
- you need GPU-accelerated training and MPI-parallel MD deployment
- you work in computational chemistry or materials science with Python, TensorFlow, PyTorch, JAX, or Paddle

## When to avoid
- you need general-purpose deep learning frameworks rather than physics-based potentials
- your simulations don't involve atomic systems or potential energy surfaces
- you need classical force fields without machine learning

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, deep-learning, llm-training, simulation, sdk
- domain: machine-learning, chemistry, simulation, gpu-computing
- platform: python, cpp, windows, cross-platform
- tags: molecular-dynamics, interatomic-potentials, computational-chemistry, lammps, tensorflow, pytorch, jax, paddlepaddle, materials-science, mpi, cuda, linux, macos

## Member repositories
- deepmodeling/deepmd-kit (main) score 95

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:06.328268+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:00:10.428661+00:00, confidence not recorded.
  - readme: https://github.com/deepmodeling/deepmd-kit (fetched 2026-08-28T04:06:06.328268+00:00, sha 151d2085932f)
  - registry_pypi: https://pypi.org/pypi/deepmd-kit/json (fetched 2026-08-29T10:40:42.640508+00:00, sha 0fdc7ac359e2)
- Data as of 2026-08-30T08:39:29.467469+00:00.