data.ASEAtomsData

class data.ASEAtomsData(datapath: str, load_properties: List[str] | None = None, load_structure: bool = True, transforms: List[torch.nn.Module] | None = None, subset_idx: List[int] | None = None, property_units: Dict[str, str] | None = None, distance_unit: str | None = None)[source]

PyTorch dataset for atomistic data. The raw data is stored in the specified ASE database.

Parameters:

  • datapath – Path to ASE DB.

  • load_properties – Set of properties to be loaded and returned. If None, all properties in the ASE dB will be returned.

  • load_structure – If True, load structure properties.

  • transforms – preprocessing torch.nn.Module (see schnetpack.data.transforms)

  • subset_idx – List of data indices.

  • units – property-> unit string dictionary that overwrites the native units of the dataset. Units are converted automatically during loading.