Can we create a fast ML-trained QM-based energy/force calculator for proteins and ligand systems?
The purpose of this notebook is to document steps taken to implement a QM-accurate calculator for energy and gradients (forces) on atoms (interatomic forces and energy potential) in varying conformations and length. The potential energy represents the interactions between particles, such as bonded and non-bonded interactions. The force can then be derived from the potential energy through differentiation. I am hoping to expand the applications' scope of such calculator in various ML projects.
The repository also contains the scripts to run the training and inferrence. If you want to obtain the full training directory with pretrained model (as per the notebook): -> download it from GDrive