MAchine-learning Potential for Landscape Exploration (MAPLE)

MAPLE is a machine-learning force-field (MLFF)–native molecular modeling platform designed for efficient and scalable exploration of potential energy landscapes, with a particular focus on geometry optimization, transition-state localization, and reaction pathway analysis under GPU-accelerated and parallel workflows.

---

Core Capabilities

Geometry Optimization

• Energy minimization using batch-parallel L-BFGS
• Saddle-point optimization via partitioned RFO variants (DS-P-RFO)

Transition-State and Reaction Path Search

• NEB / CI-NEB for minimum-energy pathway construction
• String-based and dimer-style TS localization (where applicable)
• Stable handling of non-equilibrium and reactive structures

Reaction Pathway Analysis

• Intrinsic Reaction Coordinate (IRC) calculations
• Extraction and refinement of highest-energy images (HEI)
• Mechanistic descriptor analysis along reaction coordinates

Vibrational and Hessian Analysis

• Mass-weighted Hessian construction
• Harmonic frequency analysis
• Imaginary-mode validation for transition-state characterization

---

Machine-Learning Force Fields

MAPLE interfaces with multiple general-purpose reactive ML force fields, including:

• ANI family
• AIMNet2 family
• Universal MLFFs (e.g., UMA)

These models are treated as interchangeable computational backends.
MAPLE itself remains model-agnostic, focusing on algorithmic robustness, physical consistency, and scalable execution.

MAPLE does not conceptually depend on any specific MLFF ecosystem.
Model providers and implementations are modular and replaceable by design.

---

Installation

Requirements

• Python ≥ 3.9
• CUDA-capable GPU strongly recommended for production workloads

Install MAPLE

git clone https://github.com/ClickFF/MAPLE.git
cd MAPLE
pip install -e .

## Installation

### Requirements

- Python ≥ 3.9
- CUDA-capable GPU (recommended for ML models)

### Install MAPLE

```bash
# Clone the repository
git clone https://github.com/ClickFF/maple.git
cd maple

# Install in development mode
pip install -e .

Install Dependencies

MAPLE requires several scientific computing and machine learning packages:

# Core dependencies (required)
conda install -c conda-forge numpy scipy matplotlib ase

# PyTorch (adjust for your CUDA version)
# For CUDA 11.8:
pip install torch --index-url https://download.pytorch.org/whl/cu118

# For CPU only:
pip install torch --index-url https://download.pytorch.org/whl/cpu

# Machine learning potentials
pip install fairchem-core  # For AIMNet2 and ANI models

Quick Start

Basic Usage

# Run optimization
maple input.inp

# Specify output file
maple input.inp output.out

Example Input File

#model=ANI-1xnr
#ts(method=neb,refine=nebts)
#device=gpu0

C  0.000  0.000  0.000
H  1.089  0.000  0.000
H -0.363  1.028  0.000
H -0.363 -0.514  0.890
H -0.363 -0.514 -0.890

Citation

If you use MAPLE in your research, please cite:

https://github.com/ClickFF/MAPLE

Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes with clear commit messages
4. Submit a pull request

Support

For questions, bug reports, or feature requests:

• Open an issue on GitHub
• Contact: xuw74@pitt.edu

---

Version: 0.1.0
Status: Active Development
Last Updated: December 2025