AttributeError: 'NoneType' object has no attribute 'items'

[Nemesis9450]

Train the model

cnnhistory=model.fit(x_traincnn, y_train, batch_size=20, epochs=500, validation_data=(x_testcnn, y_test))

after running this i am getting issue as:

Epoch 1/500

AttributeError Traceback (most recent call last)
Cell In[50], line 2
1 # Train the model
----> 2 cnnhistory=model.fit(x_traincnn, y_train, batch_size=20, epochs=500, validation_data=(x_testcnn, y_test))

File C:\Users\samenanc\Lib\site-packages\keras\src\utils\traceback_utils.py:123, in filter_traceback..error_handler(*args, **kwargs)
120 filtered_tb = _process_traceback_frames(e.traceback)
121 # To get the full stack trace, call:
122 # keras.config.disable_traceback_filtering()
--> 123 raise e.with_traceback(filtered_tb) from None
124 finally:
125 del filtered_tb

File C:\Users\samenanc\Lib\site-packages\keras\src\trainers\trainer.py:854, in Trainer._pythonify_logs(self, logs)
852 def _pythonify_logs(self, logs):
853 result = {}
--> 854 for key, value in sorted(logs.items()):
855 if isinstance(value, dict):
856 result.update(self._pythonify_logs(value))

AttributeError: 'NoneType' object has no attribute 'items'

Please resolve issue as ASAP.

[achal-khanna]

I encountered the same issue when using the following optimizer configuration:

opt = keras.optimizers.RMSprop(lr=0.00005, rho=0.9, epsilon=None, decay=0.0)

The problem was caused by setting epsilon=None. The solution is to set a small value for epsilon (e.g., 1e-07). This resolved the error during training (AttributeError: 'NoneType' object has no attribute 'items').