torch_geometric.datasets.Wikidata5M

class Wikidata5M(root: str, setting: str = 'transductive', transform: Optional[Callable] = None, pre_transform: Optional[Callable] = None, force_reload: bool = False)[source]

Bases: InMemoryDataset

The Wikidata-5M dataset from the “KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representation” paper, containing 4,594,485 entities, 822 relations, 20,614,279 train triples, 5,163 validation triples, and 5,133 test triples.

Wikidata-5M is a large-scale knowledge graph dataset with aligned corpus extracted form Wikidata.

Parameters:

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

  • setting (str, optional) – If "transductive", loads the transductive dataset. If "inductive", loads the inductive dataset. (default: "transductive")

  • 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)

  • force_reload (bool, optional) – Whether to re-process the dataset. (default: False)