[Solved] Faster rcnn runs train.py and reports an error: AssertionError: Duplicate registrations for type ‘optimizer’

The error message is displayed as follows:

Traceback (most recent call last):
File “D:/Pycharm/try/Bearing-fault-Diagnosis-based-on-deep-learning-main/Bearing-fault-Diagnosis-based-on-deep-learning-main/sign/cnn.py”, line 3 , in
from tensorflow import keras
File “D:\Anaconda\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\__init__.py”, line 473, in
keras._load()
File “D:\Anaconda\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\util\lazy_loader.py”, line 41, in _load
module = importlib.import_module(self.__name__)
File “D:\Anaconda\Anaconda\envs\tensorflow\lib\importlib\__init__.py”, line 127, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File “D:\Anaconda\Anaconda\envs\tensorflow\lib\site-packages\keras\__init__.py”, line 25, in
from keras import models
File “D:\Anaconda\Anaconda\envs\tensorflow\lib\site-packages\keras\models\__init__.py”, line 18, in
from keras.engine.functional import Functional
File “D:\Anaconda\Anaconda\envs\tensorflow\lib\site-packages\keras\engine\functional.py”, line 25, in
from keras.engine import base_layer
File “D:\Anaconda\Anaconda\envs\tensorflow\lib\site-packages\keras\engine\base_layer.py”, line 40, in
from keras.mixed_precision import loss_scale_optimizer
File “D:\Anaconda\Anaconda\envs\tensorflow\lib\site-packages\keras\mixed_precision\loss_scale_optimizer.py”, line 20, in
from keras.optimizer_v2 import optimizer_v2
File “D:\Anaconda\Anaconda\envs\tensorflow\lib\site-packages\keras\optimizer_v2\optimizer_v2.py”, line 1469, in
setter=RestoredOptimizer._set_hyper # pylint: disable=protected-access
File “D:\Anaconda\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\saved_model\revived_types.py”, line 133, in register_revived_type
raise AssertionError(f”Duplicate registrations for type ‘{identifier}'”)

My solution:

Uninstall keras to run normally. The uninstall statement is as follows:

pip uninstall keras -y 

Successfully solved! (I am running under pytorch, I don’t know if this method can solve the problem in tensorflow)