Skip to main content

scorebook.metrics.accuracy

Accuracy metric implementation for Scorebook.

Accuracy Objects​

@MetricRegistry.register()
class Accuracy(MetricBase)

Accuracy metric for evaluating model predictions of any type.

Accuracy = correct predictions / total predictions

score​

@staticmethod
def score(outputs: List[Any],
          labels: List[Any]) -> Tuple[Dict[str, Any], List[Any]]

Calculate accuracy score between predictions and references.

Arguments:

  • outputs - A list of inference outputs.
  • labels - A list of ground truth labels.

Returns:

The aggregate accuracy score for all items (correct predictions / total predictions). The item scores for each output-label pair (true/false).