feat: new benchmark with folmsbee conformer dataset

[lwehrhan]

Summary

The Folmsbee dataset of low-energy conformers of drug-like molecules. The differences in energy are smaller compared to the Wiggle500 dataset and it features a greater number of molecules. The highest available level of theory for energy evaluations to be used as ground-truth is DLPNO-CCSD(T).

Interactive features

For now just a table with MAE.

Category

Conformers

Data availability

Assessing conformer energies using electronic structure and
machine learning methods

Dakota Folmsbee, Geoffrey Hutchinson
International Journal of Quantum Chemistry 2020 121 (1) e26381
DOI: 10.1002/qua.26381

Computational cost

Low: Only inference of 693 molecules (varying number of conformers) is needed.

Additional context

This is a test for moving the benchmarks of mlip-audit into this repository. I have included an analysis script for this benchmark, however would like to kindly ask for assistance with building and harmonizing the Dash layout.