gnn-protmd

Deep learning and Graph Neural Networks for proteins and molecular dynamics simulations

Contents

• The main script to run is the train.py
• For doing inference on data use the eval.py script
• The sub_dist.lcf shows has to run in data parallel the training

Contents for the branches

• The LSTM branch uses the same GNN models as the master branch but has temporal component using an LSTM
• The MPNN-Toy branch uses the same models as the master branch but the data loaders are different

The basic GNN operations were used based on pytorch_geometric and pytorch_geometric_temporal