neuralmonkey.evaluators.accuracy module

class neuralmonkey.evaluators.accuracy.AccuracyEvaluator(name: str = None) → None

Bases: neuralmonkey.evaluators.evaluator.SequenceEvaluator

Accuracy Evaluator.

This class uses the default SequenceEvaluator implementation, i.e. works on sequences of equal lengths (but can be used to others as well) and use == as the token scorer.

class neuralmonkey.evaluators.accuracy.AccuracySeqLevelEvaluator(name: str = None) → None

Bases: neuralmonkey.evaluators.evaluator.Evaluator

Sequence-level accuracy evaluator.

This class uses the default evaluator implementation. It gives 1.0 to equal sequences and 0.0 to others, averaging the scores over the batch.