code/run_mrqa.py [402:438]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        prelim_predictions = sorted(
            prelim_predictions,
            key=lambda x: (x.start_logit + x.end_logit),
            reverse=True)

        _NbestPrediction = collections.namedtuple(
            "NbestPrediction", ["text", "start_logit", "end_logit"])
        seen_predictions = {}
        nbest = []
        for pred in prelim_predictions:
            if len(nbest) >= n_best_size:
                break
            feature = features[pred.feature_index]
            if pred.start_index > 0:
                tok_tokens = feature.tokens[pred.start_index:(pred.end_index + 1)]
                orig_doc_start = feature.token_to_orig_map[pred.start_index]
                orig_doc_end = feature.token_to_orig_map[pred.end_index]
                orig_tokens = example.doc_tokens[orig_doc_start:(orig_doc_end + 1)]
                tok_text = " ".join(tok_tokens)
                tok_text = tok_text.replace(" ##", "")
                tok_text = tok_text.replace("##", "")
                tok_text = tok_text.strip()
                tok_text = " ".join(tok_text.split())
                orig_text = " ".join(orig_tokens)
                final_text = get_final_text(tok_text, orig_text, do_lower_case, verbose_logging)
                if final_text in seen_predictions:
                    continue
                seen_predictions[final_text] = True
            else:
                final_text = ""
                seen_predictions[final_text] = True

            nbest.append(
                _NbestPrediction(
                    text=final_text,
                    start_logit=pred.start_logit,
                    end_logit=pred.end_logit))
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



code/run_squad.py [436:472]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        prelim_predictions = sorted(
            prelim_predictions,
            key=lambda x: (x.start_logit + x.end_logit),
            reverse=True)

        _NbestPrediction = collections.namedtuple(
            "NbestPrediction", ["text", "start_logit", "end_logit"])
        seen_predictions = {}
        nbest = []
        for pred in prelim_predictions:
            if len(nbest) >= n_best_size:
                break
            feature = features[pred.feature_index]
            if pred.start_index > 0:
                tok_tokens = feature.tokens[pred.start_index:(pred.end_index + 1)]
                orig_doc_start = feature.token_to_orig_map[pred.start_index]
                orig_doc_end = feature.token_to_orig_map[pred.end_index]
                orig_tokens = example.doc_tokens[orig_doc_start:(orig_doc_end + 1)]
                tok_text = " ".join(tok_tokens)
                tok_text = tok_text.replace(" ##", "")
                tok_text = tok_text.replace("##", "")
                tok_text = tok_text.strip()
                tok_text = " ".join(tok_text.split())
                orig_text = " ".join(orig_tokens)
                final_text = get_final_text(tok_text, orig_text, do_lower_case, verbose_logging)
                if final_text in seen_predictions:
                    continue
                seen_predictions[final_text] = True
            else:
                final_text = ""
                seen_predictions[final_text] = True

            nbest.append(
                _NbestPrediction(
                    text=final_text,
                    start_logit=pred.start_logit,
                    end_logit=pred.end_logit))
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -