torch_geometric.datasets.MD17

class MD17(root: str, name: str, train: Optional[bool] = None, transform: Optional[Callable] = None, pre_transform: Optional[Callable] = None, pre_filter: Optional[Callable] = None)

Bases: InMemoryDataset

A variety of ab-initio molecular dynamics trajectories from the authors of sGDML. This class provides access to the original MD17 datasets as well as all other datasets released by sGDML since then (15 in total).

For every trajectory, the dataset contains the Cartesian positions of atoms (in Angstrom), their atomic numbers, as well as the total energy (in kcal/mol) and forces (kcal/mol/Angstrom) on each atom. The latter two are the regression targets for this collection.

Note

Data objects contain no edge indices as these are most commonly constructed via the torch_geometric.transforms.RadiusGraph transform, with its cut-off being a hyperparameter.

Some of the trajectories were computed at different levels of theory, and for most molecules there exists two versions: a long trajectory on DFT level of theory and a short trajectory on coupled cluster level of theory. Check the table below for detailed information on the molecule, level of theory and number of data points contained in each dataset. Which trajectory is loaded is determined by the name argument. For the coupled cluster trajectories, the dataset comes with pre-defined training and testing splits which are loaded separately via the train argument.

When using these datasets, make sure to cite the appropriate publications listed on the sGDML website.

Molecule	Level of Theory	Name	#Examples
Benzene	DFT	benzene	49,863
Benzene	DFT FHI-aims	benzene FHI-aims	627,983
Benzene	CCSD(T)	benzene CCSD(T)	1,500
Uracil	DFT	uracil	133,770
Naphthalene	DFT	napthalene	326,250
Aspirin	DFT	aspirin	211,762
Aspirin	CCSD	aspirin CCSD	1,500
Salicylic acid	DFT	salicylic acid	320,231
Malonaldehyde	DFT	malonaldehyde	993,237
Malonaldehyde	CCSD(T)	malonaldehyde CCSD(T)	1,500
Ethanol	DFT	ethanol	555,092
Ethanol	CCSD(T)	ethanol CCSD(T)	2,000
Toluene	DFT	toluene	442,790
Toluene	CCSD(T)	toluene CCSD(T)	1,501
Paracetamol	DFT	paracetamol	106,490
Azobenzene	DFT	azobenzene	99,999

Parameters

• root (str) – Root directory where the dataset should be saved.

• name (str) – Keyword of the trajectory that should be loaded.

• train (bool, optional) – Determines whether the train or test split gets loaded for the coupled cluster trajectories. (default: None)

• transform (callable, optional) – A function/transform that takes in an torch_geometric.data.Data object and returns a transformed version. The data object will be transformed before every access. (default: None)

• pre_transform (callable, optional) – A function/transform that takes in an torch_geometric.data.Data object and returns a transformed version. The data object will be transformed before being saved to disk. (default: None)

• pre_filter (callable, optional) – A function that takes in an torch_geometric.data.Data object and returns a boolean value, indicating whether the data object should be included in the final dataset. (default: None)

STATS:

Name	#graphs	#nodes	#edges	#features	#classes
Benzene FHI-aims	49,863	12	0	0	0
Benzene	627,983	12	0	0	0
Benzene CCSD-T	1,500	12	0	0	0
Uracil	133,770	12	0	0	0
Naphthalene	326,250	10	0	0	0
Aspirin	211,762	21	0	0	0
Aspirin CCSD-T	1,500	21	0	0	0
Salicylic acid	320,231	16	0	0	0
Malonaldehyde	993,237	9	0	0	0
Malonaldehyde CCSD-T	1,500	9	0	0	0
Ethanol	555,092	9	0	0	0
Ethanol CCSD-T	2000	9	0	0	0
Toluene	442,790	15	0	0	0
Toluene CCSD-T	1,501	15	0	0	0
Paracetamol	106,490	20	0	0	0
Azobenzene	99,999	24	0	0	0