从python中的nltk模块执行bigramtagger时出错

Question

我正在尝试使用从CMU ARK Twitter获得的训练集训练bigramtagger词性标签v0.3.2 http://www.ark.cs.cmu.edu/TweetNLP/训练集包含标记的推文但是当我尝试训练时我得到一个错误：

File "/Applications/PyCharm Edu.app/Contents/helpers/pycharm/pycharm_setup_runner.py", line 26, in <module>
    exec (fh.read(), globals(), locals())
  File "<string>", line 41, in <module>
  File "/Users/ravinkohli/Library/Python/3.5/lib/python/site-packages/nltk/tag/sequential.py", line 341, in __init__
    backoff, cutoff, verbose)
  File "/Users/ravinkohli/Library/Python/3.5/lib/python/site-packages/nltk/tag/sequential.py", line 288, in __init__
    self._train(train, cutoff, verbose)
  File "/Users/ravinkohli/Library/Python/3.5/lib/python/site-packages/nltk/tag/sequential.py", line 179, in _train
    tokens, tags = zip(*sentence)
ValueError: not enough values to unpack (expected 2, got 0)

我的代码是

from nltk import tag
import nltk

def read_from_file(file):
    f = open(file, "r")
    data = []
    row = []
    for line in f:
        if line != "\n":
            row.append(line)
        elif line == "\n":
            data.append(row)
            row = []

    list_item = []
    data_final = []
    for item in data:
        for row1 in item:
            row1 = row1.replace("\n","")
            row1 = row1.split(" ")
            list_item.append(row1)
        data_final.append(list_item)
        list_item=[]
    return data_final

data = read_from_file("pos_training.txt")

list_item = []
final_tags = []

for row in data:
    for item in row:
        print(item)
        tagged_token = tag.str2tuple(item[0] + "/" + item[1])
        list_item.append(tagged_token)
    final_tags.append(list_item)
    list_item = []

bigram_tagger = nltk.BigramTagger(final_tags)

link to the training set