KeyError: ('val_acc', None)

Course Q&A TensorFlow Developer Professional Convolutional Neural Networks in TensorFlow
1 
#-----------------------------------------------------------
acc=history.history['accuracy']
val_acc=history.history['val_acc', None]
loss=history.history['loss']
val_loss=history.history['val_loss', None]

epochs=range(len(acc)) # Get number of epochs

#------------------------------------------------

and I am getting this error


KeyError                                  Traceback (most recent call last)
<ipython-input-21-1b2da82d7eef> in <module>
      4 #-----------------------------------------------------------
      5 acc=history.history['accuracy']
----> 6 val_acc=history.history['val_acc', None]
      7 loss=history.history['loss']
      8 val_loss=history.history['val_loss', None]
KeyError: ('val_acc', None)

and this following block
`print (history.history.keys())`
outputs this:
'dict_keys(['loss', 'accuracy'])'
the dict_keys=> val_acc and val_loss is missing ....
2

The metric you’ve specified as part of your model training process is accuracy and not acc. Tensorflow adds val_ at the start to track performance on validation dataset. Please use val_accuracy instead of val_acc.

On a related note, the starter code correctly specifies the string as val_accuracy. So, there was no need to change it to val_acc.