I made major changes including new models, refactoring variables and unifying training frameworks.
Major changes:
I reimplemented TensorNet and SO3Net in MatGL by taking core components with modifications from the original repo. Moreover, I refactored M3GNetCalculator and M3GNetDataset into PESCalculator and MGLDataset for other models. I also unified the training framework for all models. Finally, I included ZBL repulsive potentials for better extrapolations when that atoms are getting too close.
Checklist
[x] Google format doc strings added. Check with ruff.
[x] Type annotations included. Check with mypy.
[x] Tests added for new features/fixes.
[x] If applicable, new classes/functions/modules have duecredit@due.dcite decorators to reference relevant papers by DOI (example)
Tip: Install pre-commit hooks to auto-check types and linting before every commit:
Summary
I made major changes including new models, refactoring variables and unifying training frameworks. Major changes: I reimplemented TensorNet and SO3Net in MatGL by taking core components with modifications from the original repo. Moreover, I refactored M3GNetCalculator and M3GNetDataset into PESCalculator and MGLDataset for other models. I also unified the training framework for all models. Finally, I included ZBL repulsive potentials for better extrapolations when that atoms are getting too close.
Checklist
ruff.mypy.duecredit@due.dcitedecorators to reference relevant papers by DOI (example)Tip: Install
pre-commithooks to auto-check types and linting before every commit: