I am also getting the same error. And interestingly enough, if this model is trained on the data, it gives exactly the same accuracy as the expected output. The same thing also happens in the Exercise 2 (L layer Neural Network)
Can someone please help!