OGBEvaluator

class OGBEvaluator(filtered: bool = False, **kwargs)[source]

Bases: SampledRankBasedEvaluator

A sampled, rank-based evaluator that applies a custom OGB evaluation.

Initialize the evaluator.

Parameters:

  • evaluation_factory – the factory with evaluation triples

  • additional_filter_triples – additional true triples to use for filtering; only relevant if not explicit negatives are given. cf. pykeen.evaluation.rank_based_evaluator.sample_negatives()

  • num_negatives – the number of negatives to sample; only relevant if not explicit negatives are given. cf. pykeen.evaluation.rank_based_evaluator.sample_negatives()

  • head_negatives – shape: (num_triples, num_negatives) the entity IDs of negative samples for head prediction for each evaluation triple

  • tail_negatives – shape: (num_triples, num_negatives) the entity IDs of negative samples for tail prediction for each evaluation triple

  • kwargs – additional keyword-based arguments passed to pykeen.evaluation.rank_based_evaluator.RankBasedEvaluator.__init__()

  • filtered (bool)

Raises:

ValueError – if only a single side’s negatives are given, or the negatives are in wrong shape

Methods Summary

evaluate(model, mapped_triples[, ...])

Run evaluate_ogb() with this evaluator.

Methods Documentation

evaluate(model: Model, mapped_triples: Tensor, batch_size: int | None = None, slice_size: int | None = None, device: device | None = None, use_tqdm: bool = True, tqdm_kwargs: Mapping[str, str] | None = None, restrict_entities_to: Collection[int] | None = None, restrict_relations_to: Collection[int] | None = None, do_time_consuming_checks: bool = True, additional_filter_triples: None | Tensor | list[Tensor] = None, pre_filtered_triples: bool = True, targets: Collection[Literal['head', 'relation', 'tail']] = ('head', 'tail')) → MetricResults[source]

Run evaluate_ogb() with this evaluator.

Parameters:
Return type:

MetricResults