DeePMD
Model Introduction
DeePMD is a Deep Potential Molecular Dynamics model ecosystem for training machine-learning potentials for atomic systems. It provides minimal runnable training entry points based on PyTorch and TensorFlow backends.
Reference implementation: DeepMD-kit
Model Description
DeePMD uses a deep neural network architecture and atomic-system data to train interatomic potentials for molecular and materials systems and to run molecular dynamics simulations.
Use Cases
| Use case | Description |
|---|---|
| DeePMD water training | Train a water potential with configurations such as demo/water_se_e2_a_pt/input_torch.json |
| Multi-GPU SLURM jobs | See demo/water_se_e2_a_pt/submit_4card.sh and submit_8card.sh |
| Environment connectivity check | Run dp_install.sh to verify that DeepMD-kit can be installed |
| Custom data migration | Replace the systems path in the configuration with your own DeepMD NumPy data |
Usage
1. Using OneCode
Try intelligent, one-click AI4S programming in the OneCode online environment:
Try intelligent, one-click AI4S programming
2. Manual Installation and Usage
Hardware requirements
- A GPU or DCU is recommended for training.
- A CPU can be used for installation checks and small-data connectivity tests, but full training will be slow.
- DCU users must install DTK in advance. DTK 25.04.2 or later, or the OneScience-recommended version matching the current cluster, is recommended.
Download the Model Package
hf download --model OneScience-Sugon/DeePMD --local-dir ./deepmd
cd deepmd
Install the Runtime Environment
DCU environment
# Activate DTK and conda first
conda create -n onescience311 python=3.11 -y
conda activate onescience311
# uv installation is also supported
pip install onescience[matchem-dcu] -i http://mirrors.onescience.ai:3141/pypi/simple/ --trusted-host mirrors.onescience.ai
GPU environment
# Activate conda first
conda create -n onescience311 python=3.11 -y libstdcxx-ng=12 libgcc-ng=12 gcc_linux-64=12 gxx_linux-64=12
conda activate onescience311
# uv installation is also supported
pip install onescience[matchem-gpu] -i http://mirrors.onescience.ai:3141/pypi/simple/ --trusted-host mirrors.onescience.ai
Install DeepMD-kit
# The test_pip environment is used by default. To use another conda environment:
# export MATCHEM_CONDA_NAME=your_env
bash dp_install.sh
Training Data
Training data is not bundled with this repository. Using the DeePMD water dataset as an example, download it from Hugging Face to data/ in the repository root:
hf download --dataset OneScience-Sugon/DeePMD --local-dir ./data
After downloading, the data is located at data/DeePMD/water/data_0..3/.
Training
Single GPU:
cd demo/water_se_e2_a_pt
dp --pt train input_torch.json
Multiple GPUs:
cd demo/water_se_e2_a_pt
bash submit_4card.sh
Trained Weights
This repository does not currently include trained weights. You can obtain them by following the training procedure above.
Official OneScience Resources
| Platform | OneScience Main Repository | Skills Repository |
|---|---|---|
| Gitee | https://gitee.com/onescience-ai/onescience | https://gitee.com/onescience-ai/oneskills |
| GitHub | https://github.com/onescience-ai/OneScience | https://github.com/onescience-ai/oneskills |
Citation and License
- The DeePMD example code comes from the MatChem examples in the OneScience project and refers to the upstream DeepMD-kit project. See the official upstream repository for the DeepMD-kit license.
- If you use DeePMD training results in research, please cite DeepMD-kit, the relevant OneScience projects, and the datasets used.