HDP-SC

November 9, 2024 · View on GitHub

Project Overview

HDP-SC is a project designed for building coarse-grained potentials for polymer.

Project Structure

HDP-SC/
├── README.md
├── LICENSE
├── data/
│   ├── AA/
│   │   ├── AA.data
│   │   └── AA.lammpstrj
│   ├── CG/
│   │   ├── atom_types.npy
│   │   ├── box.npy
│   │   ├── build.json
│   │   ├── CG.data
│   │   ├── CG.lammpstrj
│   │   ├── force.npy
│   │   ├── pos.npy
│   │   ├── prior.json
│   │   └── structure.psf
│   └── DP/
│       ├── input.json
│       ├── training_data/
│       │   ├── set.000/
│       │   │   ├── box.npy
│       │   │   ├── coord.npy
│       │   │   └── force.npy
│       │   └── type.raw
│       └── validation_data/
│           ├── set.000/
│           │   ├── box.npy
│           │   ├── coord.npy
│           │   └── force.npy
│           └── type.raw
├── model/
│   └── graph.pb
└── scripts/
    ├── do_cg_map.ipynb
    ├── model_train.ipynb
    ├── prior_fit.ipynb
    ├── rdf_analysis.ipynb
    └── utils.py

Guide

  1. CG Mapping

    Run do_cg_map.ipynb to generate the coarse-grained mapping.

  2. Prior Energy Fitting

    Run prior_fit.ipynb to generate prior parameters.

  3. Deep Potential Training

    Run model_train.ipynb to train the deep potential model.

  4. RDF Analysis

    Run rdf_analysis.ipynb to perform RDF analysis.

Dependencies

All project dependencies are listed in requirements.txt. The main dependencies include:

  • deepmd-kit
  • numpy
  • scipy
  • mdtraj
  • moleculekit
  • matplotlib
  • scikit-learn

License

This project is licensed under the Apache License. See the LICENSE file for details.

Contact

If you have any questions or suggestions, please contact me via email: huangqi@mail.sim.ac.cn.