使用Transformers做情感分析时出现type_spec_registry导入错误求助

尝试使用Transformers库的pipeline函数执行情感分析任务时触发运行时错误，根源是无法从tensorflow.python.framework导入type_spec_registry，进而导致transformers.models.distilbert.modeling_tf_distilbert模块导入失败。

复现代码 #

from transformers import pipeline
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

from sklearn.model_selection import train_test_split
from sklearn.metrics import roc_auc_score, f1_score, confusion_matrix

# Basic usage
classifier = pipeline('sentiment-analysis')

完整报错堆栈 #

ImportError                               Traceback (most recent call last)
File ~\anaconda3\envs\Asiwajuflow\lib\site-packages\transformers\utils\import_utils.py:1099, in _LazyModule._get_module(self, module_name)
   1098 try:
-> 1099     return importlib.import_module("." + module_name, self.__name__)
   1100 except Exception as e:

File ~\anaconda3\envs\Asiwajuflow\lib\importlib\__init__.py:127, in import_module(name, package)
    126         level += 1
-> 127 return _bootstrap._gcd_import(name[level:], package, level)

...（中间堆栈内容省略）

File ~\anaconda3\envs\Asiwajuflow\lib\site-packages\keras\src\saving\legacy\saved_model\json_utils.py:38
     37 # isort: off
-> 38 from tensorflow.python.framework import type_spec_registry
     40 _EXTENSION_TYPE_SPEC = "_EXTENSION_TYPE_SPEC"

ImportError: cannot import name 'type_spec_registry' from 'tensorflow.python.framework' (C:\Users\hp\anaconda3\envs\Asiwajuflow\lib\site-packages\tensorflow\python\framework\__init__.py)

The above exception was the direct cause of the following exception:

RuntimeError                              Traceback (most recent call last)
Cell In[9], line 2
      1 #Basic usage
----> 2 classifier = pipeline('sentiment-analysis')

...（中间堆栈内容省略）

RuntimeError: Failed to import transformers.models.distilbert.modeling_tf_distilbert because of the following error (look up to see its traceback):
cannot import name 'type_spec_registry' from 'tensorflow.python.framework' (C:\Users\hp\anaconda3\envs\Asiwajuflow\lib\site-packages\tensorflow\python\framework\__init__.py)

• 升级TensorFlow到兼容版本
  type_spec_registry是TensorFlow 2.10+引入的API，旧版本不存在。执行以下命令升级：

  # 查看当前TensorFlow版本
  pip show tensorflow
  # 升级到稳定兼容版（如2.15.x系列）
  pip install --upgrade tensorflow
  

• 同步Transformers库版本
  确保Transformers与TensorFlow版本匹配，执行升级：

  pip install --upgrade transformers
  

• 强制使用PyTorch后端（无需TensorFlow时）
  创建pipeline时指定framework='pt'，跳过TensorFlow相关模块加载：

  classifier = pipeline('sentiment-analysis', framework='pt')
  

• 修复依赖冲突
  如果单独安装了Keras，可能与TensorFlow自带的Keras冲突，执行以下操作：

  # 卸载独立安装的Keras
  pip uninstall -y keras
  # 重新安装TensorFlow确保依赖完整
  pip install tensorflow --force-reinstall
  

内容的提问来源于stack exchange，提问作者Abdusalam Idris