sgnlp.models.span_extraction.utils.get_final_text

get_final_text(pred_text, orig_text, do_lower_case, verbose_logging=False)

Project the tokenized prediction back to the original text.