span

class SpanEvaluator(limit=0.5)

Bases: paddle.metric.metrics.Metric

SpanEvaluator computes the precision, recall and F1-score for span detection.

compute(start_probs, end_probs, gold_start_ids, gold_end_ids)

Computes the precision, recall and F1-score for span detection.

update(num_correct_spans, num_infer_spans, num_label_spans)

This function takes (num_infer_spans, num_label_spans, num_correct_spans) as input, to accumulate and update the corresponding status of the SpanEvaluator object.

eval_span(predict_start_ids, predict_end_ids, label_start_ids, label_end_ids)

evaluate position extraction (start, end) return num_correct, num_infer, num_label input: [1, 2, 10] [4, 12] [2, 10] [4, 11] output: (1, 2, 2)

accumulate()

This function returns the mean precision, recall and f1 score for all accumulated minibatches.

Returns

Returns tuple (precision, recall, f1 score).

Return type

tuple

reset()

Reset function empties the evaluation memory for previous mini-batches.

name()

Return name of metric instance.