在指定环境下无法运行代码

[rogerxl-brennanhs]

您好，感谢您的开源🙏。
我按照您的 requirements 配制了环境 torch==1.7.1+cu110、stanza==1.5.1，但是在运行 data_prepro_getGraph.py 代码会报错 UnboundLocalError: local variable 'doc' referenced before assignment, 且其原因是Exception: module 'torch' has no attribute 'take_along_dim'。请问应该如何解决？

torch.take_along_dim 方法似乎是在 PyTorch 1.10.0 的版本中才引入的？ 我是需要更换 PyTorch 的版本么？还是更换 stanza 的版本？

def get_dependency(tokens):
    nlp = stanza.Pipeline(lang='en',
                          processors='tokenize,mwt,pos,lemma,depparse,constituency',
                          tokenize_pretokenized=True,
                          dir='/path/to/stanza_resources',
                          download_method=None)

    result = []
    result2 = []
    POS = []
    Pre_head = []
    for idx,token in enumerate(tokens):
        # sente =' '.join(['Pairing', 'it', 'with', 'an', 'iPhone', 'is', 'a', 'pure', 'pleasure', '-', 'talk', 'about', 'painless', 'syncing', '-', 'used', 'to', 'take', 'me', 'forever', '-', 'now', 'it', "'s", 'a', 'snap', '.'])
        # sente =' '.join(['she', 'is', 'a', 'beautiful', 'woman','.'])
        # sente = "It is really easy to use and it is quick to start up ."
        # sente = "great food but the service was dreadful!"


        sente = ' '.join(token)
        # if idx==627:
        #     print(idx)
        # token_list= ['The', 'wait', 'staff', 'is', 'pleasant', ',', 'fun', ',', 'and', 'for', 'the', 'most', 'part', 'gorgeous', '(', 'in', 'the', 'wonderful', 'aesthetic', 'beautification', 'way', ',', 'not', 'in', 'that', "she's-way-cuter-than-me-that-b", '@', '']
        # token_dicts = [{'id': str(i + 1), 'text': token} for i, token in enumerate(token)]
        try:
            doc = nlp(sente)
        except Exception as e:
            print(f"Error processing sentence at index {idx}: {token}")
            print(f"Exception: {e}")
        # break
        # 获取第一个句子的依存分析结果
        sent = doc.sentences
        dependencies = sent[0].dependencies
        dd = []
        pos = []
        for dependency in dependencies:
            this_word = dependency[2]

            token_id = this_word.id
            token_head_id = this_word.head
            token_dependency_label = this_word.deprel

            # 将依存关系转换成['root',1,2]形式
            if token_head_id == 0:
                dd.append(['root', token_head_id,token_id])
            else:
                dd.append([token_dependency_label, token_head_id, token_id])
            pos.append(this_word.pos)
        result.append(dd)
        dd2 = [""+e[0] for e in dd]
        POS.append(pos)
        result2.append(dd2)
        prehead = [e[1] for e in dd]
        Pre_head.append(prehead)
    return result2,result,POS,Pre_head

2024-12-26 12:40:24 WARNING: Can not find mwt: default from official model list. Ignoring it.
2024-12-26 12:40:24 INFO: Loading these models for language: en (English):
======================================
| Processor    | Package             |
--------------------------------------
| tokenize     | combined            |
| pos          | combined_charlm     |
| lemma        | combined_nocharlm   |
| constituency | ptb3-revised_charlm |
| depparse     | combined_charlm     |
======================================

2024-12-26 12:40:24 INFO: Using device: cuda
2024-12-26 12:40:24 INFO: Loading: tokenize
2024-12-26 12:40:24 INFO: Loading: pos
2024-12-26 12:40:28 INFO: Loading: lemma
2024-12-26 12:40:28 INFO: Loading: constituency
2024-12-26 12:40:29 INFO: Loading: depparse
2024-12-26 12:40:29 INFO: Done loading processors!
Error processing sentence at index 0: ['first', 'one', 'that', 'they', 'shipped', 'was', 'obviously', 'defective', ',', 'super', 'slow', 'and', 'speakers', 'were', 'garbled', '.']
Exception: module 'torch' has no attribute 'take_along_dim'
Traceback (most recent call last):
  File "/path/to/pycharm/EPMEI/data_prepro_getGraph.py", line 123, in <module>
    data_tackled = tackle_dataset(data_list)
  File "/path/to/pycharm/EPMEI/data_prepro_getGraph.py", line 65, in tackle_dataset
    predicted_dependencies, dependencies,POS,prehead = get_dependency(sentences)
  File "/path/to/pycharm/EPMEI/data_prepro_getGraph.py", line 35, in get_dependency
    sent = doc.sentences
UnboundLocalError: local variable 'doc' referenced before assignment