SquadExample

SquadExample为预处理语音数据的类。

example = SquadExample(
                qas_id=qas_id,
                question_text=question_text,
                context_text=context_text,
                answer_text=answer_text,
                start_position_character=start_position_character,  # eg.0
                title=None,
                is_impossible=is_impossible,
                answers=answers,
            )

类似上述生成，针对QA任务，提供context，question，answer。其中start_position为answer在question中的字符序号（包括空格）。

class SquadExample:
    """
    A single training/test example for the Squad dataset, as loaded from disk.

    Args:
        qas_id: The example's unique identifier
        question_text: The question string
        context_text: The context string
        answer_text: The answer string
        start_position_character: The character position of the start of the answer
        title: The title of the example
        answers: None by default, this is used during evaluation. Holds answers as well as their start positions.
        is_impossible: False by default, set to True if the example has no possible answer.
    """

    def __init__(
        self,
        qas_id,
        question_text,
        context_text,
        answer_text,
        start_position_character,
        title,
        answers=[],
        is_impossible=False,
    ):
        self.qas_id = qas_id
        self.question_text = question_text
        self.context_text = context_text
        self.answer_text = answer_text
        self.title = title
        self.is_impossible = is_impossible
        self.answers = answers

        self.start_position, self.end_position = 0, 0

        doc_tokens = []
        char_to_word_offset = []
        prev_is_whitespace = True

        # Split on whitespace so that different tokens may be attributed to their original position.
        for c in self.context_text:
            if _is_whitespace(c):
                prev_is_whitespace = True
            else:
                if prev_is_whitespace:
                    doc_tokens.append(c)
                else:
                    doc_tokens[-1] += c
                prev_is_whitespace = False
            char_to_word_offset.append(len(doc_tokens) - 1)

        self.doc_tokens = doc_tokens
        self.char_to_word_offset = char_to_word_offset

        # Start and end positions only has a value during evaluation.
        if start_position_character is not None and not is_impossible:
            self.start_position = char_to_word_offset[start_position_character]
            self.end_position = char_to_word_offset[
                min(start_position_character + len(answer_text) - 1, len(char_to_word_offset) - 1)
            ]

上述doc_tokens为去除空格和一些特殊字符的列表（还未进行分词）。

同时，根据answer及context，start_position确定end_position。