Wn18rr

Wn18rr dataset.

Parameters

• batch_size

• shuffle – defaults to True

• pre_compute – defaults to True

• num_workers – defaults to 1

• seed – defaults to None

Attributes

• train (list): Training set.

  valid (list): Validation set. test (list): Testing set. entities (dict): Index of entities. relations (dict): Index of relations. n_entity (int): Number of entities. n_relation (int): Number of relations.

Examples

>>> from ckb import datasets

>>> dataset = datasets.Wn18rr(batch_size=1, pre_compute=True, shuffle=True, seed=42, num_workers = 0)

>>> dataset
Wn18rr dataset
    Batch size  1
    Entities  40943
    Relations  11
    Shuffle  True
    Train triples  69585
    Validation triples  11381
    Test triples  12037

Methods

fetch

get_train_loader

Initialize train dataset loader.

Parameters

• mode

mapping_entities

Construct mapping entities.

mapping_relations

Construct mapping relations.

test_dataset

test_stream

validation_dataset

References

---

1. Liang Yao, Chengsheng Mao, and Yuan Luo. 2019. Kg- bert: Bert for knowledge graph completion. arXiv preprint arXiv:1909.03193. ↩