torch-admp
Automatic Differentiable Multipolar Polarizable (ADMP) in PyTorch backend
Overview
torch-admp is a PyTorch implementation of the ADMP (Automatic Differentiable Multipolar Polarizable) module in DMFF (Differentiable Molecular Force Field) package. This package provides efficient implementations of various molecular dynamics force calculations including:
- Particle Mesh Ewald (PME) for electrostatic interactions
- Charge Equilibration (QEq) methods
- Polarizable electrode models
- Neighbor list management
- Optimization algorithms
Installation
For development:
Features
- GPU Accelerated: Built on PyTorch for efficient GPU computation
- JIT Compilation: Support for TorchScript compilation
- Modular Design: Clean separation of different force components
- Extensible: Easy to add new force modules
Documentation
- API Reference - Complete API documentation
- Examples - Usage examples and tutorials
Citation
If you use torch-admp in your research, please cite:
@software{torch_admp,
author = {ChiahsinChu},
title = {torch-admp: ADMP in PyTorch backend},
url = {https://github.com/ChiahsinChu/torch-admp},
year = {2024}
}
License
This project is licensed under the LGPL-3.0-or-later License - see the LICENSE file for details.