2024-02-12 Chemical Perception meeting notes

Tobias

• training energies md from 1000 K and fit ff from scratch

• triple bonds have been problematic and training on energies appear to make the results bad

• TG: what do the types look like qualitatively compared to e.g. sage?

  • informed types: many more than uninformed, but haven’t compared to the 1.3.0 types

What are the differences between fitting and and chemical perception?

1. audience?

2. chemper for more-or-less automated methods; “science” of chemical perception

3. fitting is more for targeted problems