This repository contains the code for training and evaluating message passing neural networks (MPNNs) for predicting molecular properties.
The included data is a subset from Sereina Riniker.
The notebook run_and_evaluate.ipynb
contains the code for training and evaluating the MPNN.
The required python packages are
pip install dgl -f https://data.dgl.ai/wheels/cu117/repo.html
pip install dglgo -f https://data.dgl.ai/wheels-test/repo.html
pip install dgllife
pip install torch torchvision torchaudio
pip install rdkit
Machine Learning of Partial Charges Derived from High-Quality Quantum-Mechanical Calculations