modificaiton.md

修改了 data.py的 batch_yeild

原函数 like def batch_yield(data, batch_size, vocab, tag2label, shuffle=False): """

:param data:
:param batch_size:
:param vocab:
:param tag2label:
:param shuffle:
:return:
"""
if shuffle:
    random.shuffle(data)

seqs, labels = [], []
for (sent_, tag_) in data:
    sent_ = sentence2id(sent_, vocab)
    label_ = [tag2label[tag] for tag in tag_]

    if len(seqs) == batch_size:
        yield seqs, labels
        seqs, labels = [], []

    seqs.append(sent_)
    labels.append(label_)

if len(seqs) != 0:
    yield seqs, labels

要修改label_ like https://www.tensorflow.org/tutorials/text/text_generation

add pre_train mode in main.py