Requirements
- Python >= 3.9
- CUDA-capable GPU (recommended for ML potential evaluation; CPU-only mode is also supported)
- Operating System: Linux (recommended), macOS, or Windows with WSL
Tip
A CUDA-capable GPU is strongly recommended for production workloads. ML potential evaluation on GPU is orders of magnitude faster than on CPU.
Installation Steps
1
Clone the Repository
Download the MAPLE source code from GitHub.
git clone https://github.com/ClickFF/MAPLE.git
cd MAPLE2
Install MAPLE
Install the package in development mode so you can easily update.
pip install -e .3
Install Dependencies
MAPLE requires PyTorch and several scientific computing packages. Install them in the correct order.
PyTorch (with CUDA support)
# For CUDA 11.8:
pip install torch --index-url https://download.pytorch.org/whl/cu118
# For CUDA 12.1:
pip install torch --index-url https://download.pytorch.org/whl/cu121
# For CPU only:
pip install torch --index-url https://download.pytorch.org/whl/cpufairchem-core (for AIMNet2, ANI, and UMA models)
pip install fairchem-coreCore scientific dependencies
conda install -c conda-forge numpy scipy matplotlib ase4
Verify Installation
Run a quick test to confirm that MAPLE is installed correctly and can access a GPU.
# Check that MAPLE is importable
python -c "import maple; print(maple.__version__)"
# Check GPU availability
python -c "import torch; print('CUDA available:', torch.cuda.is_available())"
Note
If you encounter issues during installation, please check the Report Issue page for guidance on how to get help.
Source Code
MAPLE is open source and hosted on GitHub. You can browse the source code, submit issues, or contribute to the project.