TensorFlow目标检测训练报错：无法导入typing_extensions的Any解决方法

问题场景 #

数据训练时触发如下导入错误，虽已安装所有依赖但问题仍存在：

Traceback (most recent call last):
File “C:\Users\dhrub\BDSLP-REALTIME\RealTimeObjectDetection\Tensorflow\models\research\object_detection\model_main_tf2.py”, line 31, in
from object_detection import model_lib_v2
File “C:\Users\dhrub\anaconda3\lib\site-packages\object_detection\model_lib_v2.py”, line 31, in
from object_detection import model_lib
File “C:\Users\dhrub\anaconda3\lib\site-packages\object_detection\model_lib.py”, line 35, in
from object_detection.builders import optimizer_builder
File “C:\Users\dhrub\anaconda3\lib\site-packages\object_detection\builders\optimizer_builder.py”, line 25, in
from official.modeling.optimization import ema_optimizer
File “C:\Users\dhrub\anaconda3\lib\site-packages\official\modeling\optimization_init_.py”, line 23, in
from official.modeling.optimization.optimizer_factory import OptimizerFactory
File “C:\Users\dhrub\anaconda3\lib\site-packages\official\modeling\optimization\optimizer_factory.py”, line 20, in
import tensorflow_addons.optimizers as tfa_optimizers
File “C:\Users\dhrub\anaconda3\lib\site-packages\tensorflow_addons_init_.py”, line 22, in
from tensorflow_addons import callbacks
File “C:\Users\dhrub\anaconda3\lib\site-packages\tensorflow_addons\callbacks_init_.py”, line 17, in
from tensorflow_addons.callbacks.average_model_checkpoint import AverageModelCheckpoint
File “C:\Users\dhrub\anaconda3\lib\site-packages\tensorflow_addons\callbacks\average_model_checkpoint.py”, line 17, in
from typeguard import typechecked
File “C:\Users\dhrub\anaconda3\lib\site-packages\typeguard_init_.py”, line 4, in
from ._checkers import TypeCheckerCallable as TypeCheckerCallable
File “C:\Users\dhrub\anaconda3\lib\site-packages\typeguard_checkers.py”, line 68, in
from typing_extensions import Any as SubclassableAny
ImportError: cannot import name ‘Any’ from ‘typing_extensions’ (C:\Users\dhrub\anaconda3\lib\site-packages\typing_extensions.py)

解决方案 #

• 检查当前typing_extensions版本
  执行命令查看已安装版本：

  pip show typing_extensions
  

  版本过低或过高都可能引发兼容性问题。

• 升级/降级到兼容版本
  优先尝试升级到最新稳定版：

  pip install --upgrade typing_extensions
  

  若升级后仍报错，可降级到已知兼容版本（如4.5.0）：

  pip install typing_extensions==4.5.0
  

• 调整typeguard版本
  错误由typeguard调用typing_extensions触发，若上述步骤无效，可尝试固定typeguard版本：

  pip install typeguard==2.13.3
  

• 确认环境一致性
  确保激活的虚拟环境（如Anaconda环境）是训练使用的环境，避免依赖安装到错误环境。使用conda的话可执行：

  conda update typing_extensions
  

内容的提问来源于stack exchange，提问作者TAHSINUL HAQUE DHRUBO