W4_A2_Two_layer_Neural Network_Wrong output

Course Q&A Deep Learning Specialization Neural Networks and Deep Learning
3

Hello @yanivh ,

Without revealing my solution… the closest thing would be to say that I almost literally copied the sample lines of “The functions and their inputs are:” from the cell “Use the helper functions”

4 - Two-layer Neural Network

Exercise 1 - two_layer_model

Use the helper functions you have implemented in the previous assignment to build a 2-layer neural network with the following structure: LINEAR → RELU → LINEAR → SIGMOID. The functions and their inputs are:

def initialize_parameters(n_x, n_h, n_y):
    ...
    return parameters 
def linear_activation_forward(A_prev, W, b, activation):
    ...
    return A, cache
def compute_cost(AL, Y):
    ...
    return cost
def linear_activation_backward(dA, cache, activation):
    ...
    return dA_prev, dW, db
def update_parameters(parameters, grads, learning_rate):
    ...
    return parameters

but obviously replacing the AL with the appropriate variable, and assigning the proper “activation” according to the corresponding layers…

Note that the provided “initialize_parameters” function doesn’t accept the three dimensions as individual parameters, but seems to expect an array(list)…hence I used layers_dim

That said, training my implementation of this “two_layer_model” on the next cell:

4.1 - Train the model

parameters, costs = two_layer_model(. . . )

gave:

Cost after iteration 0: 0.6950464961800915
Cost after iteration 100: 0.5892596054583805
Cost after iteration 200: 0.5232609173622991
. . .
. . .
Cost after iteration 2300: 0.028387859212946117
Cost after iteration 2400: 0.026615212372776077
Cost after iteration 2499: 0.024821292218353375

Not the “expected” results, but similar ones…

I’ve repeatedly re-checked these lines:

   # Initialize parameters dictionary
   # parameters = ...
   # YOUR CODE STARTS HERE
   . . .
   # YOUR CODE ENDS HERE

        # Forward propagation:
        # A1, cache1 = ...
        # A2, cache2 = ...
        # YOUR CODE STARTS HERE
        . . .
        # YOUR CODE ENDS HERE

        # Compute cost
        # cost = ...
        # YOUR CODE STARTS HERE
        . . .
        # YOUR CODE ENDS HERE

        # Backward propagation.
        # dA1, dW2, db2 = ...
        # dA0, dW1, db1 = ...
        # YOUR CODE STARTS HERE
        . . .
        # YOUR CODE ENDS HERE

   # Update parameters.
   # parameters = ...
   # YOUR CODE STARTS HERE
   . . .
   # YOUR CODE ENDS HERE

and don’t see my error(s)…

Note that the problem happens upon executing “two_layer_model_test” evaluation cell:

parameters, costs = two_layer_model(train_x, train_y, layers_dims = (n_x, n_h, n_y), num_iterations = 2, print_cost=False)

print("Cost after first iteration: " + str(costs[0]))

two_layer_model_test(two_layer_model)

I get “exactly” the same “error” messages as image|690x437
presented by @tfu on post " Week 4 exercise 4 not passing both tests" but @tfu solution

made it work by replacing layers_dim with the three dimensions

doesn’t work for me…

I feel that any further than this would be revealing my implemented solution…