Installation

Installation#

Requirements#

  • Python 3.12

  • PyTorch >= 2.4.0

  • CUDA >= 12.4

  • A C++ compiler (g++ is used by default)

Setting Up the Environment#

Create a fresh conda/mamba environment:

mamba create -n torchff python=3.12
mamba activate torchff

Install PyTorch with CUDA support:

pip install torch==2.6.0 --index-url https://download.pytorch.org/whl/cu124

Install testing dependencies (optional):

pip install pytest

Installing torchff#

Clone the repository and install:

git clone https://github.com/THGLab/torchff-lib.git
cd torchff-lib
python setup.py install

For development mode (editable install):

pip install -e .

Verifying the Installation#

After installation, verify that the package is importable:

import torchff
print("torchff installed successfully")

To verify CUDA extensions are available (requires GPU):

import torch
from torchff.bond import HarmonicBond

bond_fn = HarmonicBond(use_customized_ops=True)
print("CUDA extensions loaded successfully")