BERT-Whole Word Masked(WWM)

记录一下BERT的WWM实现和中文的WWM

• 上代码: 英文 Bert-WWM数据创建方法

def create_masked_lm_predictions(tokens, masked_lm_prob,
                                 max_predictions_per_seq, vocab_words, rng):
    """
    Creates the predictions for the masked LM objective.
    params:
      tokens: input ids;
      masked_lm_prob: masked prob(how many tokens will be masked, default 0.15)
      max_predictions_per_seq: a number
      vocab_words: vocab
      rng: random module
    """

    cand_indexes = []
    for (i, token) in enumerate(tokens):
        if token == "[CLS]" or token == "[SEP]":
            continue
        # Whole Word Masking means that if we mask all of the wordpieces
        # corresponding to an original word. When a word has been split into
        # WordPieces, the first token does not have any marker and any subsequence
        # tokens are prefixed with ##. So whenever we see the ## token, we
        # append it to the previous set of word indexes.
        #
        # Note that Whole Word Masking does *not* change the training code
        # at all -- we still predict each WordPiece independently, softmaxed
        # over the entire vocabulary.
        if (FLAGS.do_whole_word_mask and len(cand_indexes) >= 1 and
                token.startswith("##")):
            # 如果WWM，则记录每个单词的所有sub token，如此达到对整个单词所有sub token进行广义mask
            cand_indexes[-1].append(i)
        else:
            cand_indexes.append([i])

    rng.shuffle(cand_indexes)  # shuffle再取前15%的词进行8-1-1处理，表示随机取15%的词进行mask

    output_tokens = list(tokens)

    num_to_predict = min(max_predictions_per_seq,
                         max(1, int(round(len(tokens) * masked_lm_prob))))

    masked_lms = []
    covered_indexes = set()
    for index_set in cand_indexes:  # 对整个词或者单个字符
        if len(masked_lms) >= num_to_predict:
            break
        # If adding a whole-word mask would exceed the maximum number of
        # predictions, then just skip this candidate.
        if len(masked_lms) + len(index_set) > num_to_predict:
            continue
        is_any_index_covered = False
        for index in index_set:
            if index in covered_indexes:
                is_any_index_covered = True
                break
        if is_any_index_covered:
            continue
        for index in index_set:
            covered_indexes.add(index)
            masked_token = None
            # 80% of the time, replace with [MASK] 8-1-1mask策略
            if rng.random() < 0.8:
                masked_token = "[MASK]"
            else:
                # 10% of the time, keep original
                if rng.random() < 0.5:
                    masked_token = tokens[index]
                # 10% of the time, replace with random word
                else:
                    masked_token = vocab_words[rng.randint(0, len(vocab_words) - 1)]

            output_tokens[index] = masked_token
            # 记录label
            masked_lms.append(MaskedLmInstance(index=index, label=tokens[index]))
    assert len(masked_lms) <= num_to_predict
    masked_lms = sorted(masked_lms, key=lambda x: x.index)

    masked_lm_positions = []
    masked_lm_labels = []
    for p in masked_lms:
        masked_lm_positions.append(p.index)
        masked_lm_labels.append(p.label)

    return (output_tokens, masked_lm_positions, masked_lm_labels)

• 中文Bert-WWM预训练数据创建方法（from ymcui)