Table 2.
Description of the data structure used in each molecular dataset to store the quantum chemical properties collected from the Gaussian 09 outputs for the 120,000 geometry configurations.
| No. | Quantity | Units | Shape | Description |
|---|---|---|---|---|
| 1 | Z | (n_atoms,) | Atomic numbers of nuclei | |
| 2 | R | Å | (120000, n_atoms, 3) | Cartesian coordinates |
| 3 | F | kcal mol−1 Å−1 | (120000, n_atoms, 3) | Atomic forces |
| 4 | Q | e | (120000, n_atoms, 1) | Mulliken charges |
| 5 | D | D | (120000, 3) | Dipole moment |
| 6 | P | (120000, 6) | Isotropic polarizability | |
| 7 | RC | GHz | (120000, 3) | Rotational constants |
| 8 | HL | eV | (120000, 2) | HOMO and LUMO energies |
| 9 | E | kcal mol−1 | (120000, 1) | Potential energy |
| 10 | R2 | (120000, 1) | Electronic spatial extent | |
| 11 | CONF | — | (120000, 1) | Conformation identifier |
The quantity column provides the list of acronyms used as keys of the Python dictionary to access the corresponding properties of the dataset. n_atoms is the number of atoms in a molecule.