The code block for checking split_data function is taking a huge amount of time. I stopped the process after it ran for more than 20 minutes and yet did not complete execution.
For the root directory grader cell, it clearly mentioned don’t hard-code the path. There is a error log for your split dataset grader cell.
This is your notebook error log 
which has happened because the interpreter throws a KeyboardInterrupt exception when the user/programmer presses the ctrl – c or del key either accidentally or intentionally. Syntax: As observed in the example above, the KeyboardInterrupt exception is a standard exception used to handle keyboard-related issues. But while going through your notebook I found some code error, so refer to the link below
Kindly refer this to make correction to these cells
Next in the GRADED FUNCTION: train_val_generators
def train_val_generators(TRAINING_DIR, VALIDATION_DIR):
why have you used different batch_size for training and validation directory? please use same batch size. also remember using too small batch_size also creates problems but creating too large batch size can also create issue in training.
Next error is in
grader-required-cell
GRADED FUNCTION: create_model
def create_model():
see this image from your notebook
Although you are using the correct optimizer, this assignment doesn’t support this RMSprop( your learning has become earning). Use instead ‘rmsprop’ without learning rate.
Do these corrections and run your model. Just let me know once you have clear the test
Happy Learning!!!
Regards
DP