Notes

- For templates/forcefield learning algorithm data augmentation might be important - one can have random transitions/structures generated from existing templates, and ground truth values calculated.

- reevaluate cuo2 props. w tersoff in ifftt

- barrier list unit test of copper atom on sio2