Compute the total loss

Course Q&A Deep Learning Specialization Improving Deep Neural Networks: Hyperparameter tun
,
1

Hi,

Is any mistake in the code that I am using. I am getting a assertion error.

image
image1000×421 15.6 KB

I have used the defined function compute_total_loss_test. It cannot be modified

def compute_total_loss_test(target, Y):
pred = tf.constant([[ 2.4048107, 5.0334096 ],
[-0.7921977, -4.1523376 ],
[ 0.9447198, -0.46802214],
[ 1.158121, 3.9810789 ],
[ 4.768706, 2.3220146 ],
[ 6.1481323, 3.909829 ]])
minibatches = Y.batch(2)
for minibatch in minibatches:
result = target(pred, tf.transpose(minibatch))
break

print("Test 1: ", result)
assert(type(result) == EagerTensor), "Use the TensorFlow API"
assert (np.abs(result - (0.50722074 + 1.1133534) / 2.0) < 1e-7), "Test 1 does not match. Did you get the reduce sum of your loss functions?"

### Test 2
labels = tf.constant([[1., 0., 0.], [0., 1., 0.], [0., 0., 1.]])
logits = tf.constant([[1., 0., 0.], [1., 0., 0.], [1., 0., 0.]])

result = compute_total_loss(logits, labels)
print("Test 2: ", result)
assert np.allclose(result, 3.295837 ), "Test 2 does not match."

print("\033[92mAll test passed")

compute_total_loss_test(compute_total_loss, new_y_train )

In the assignment part I have modified compute_total_loss function.

def compute_total_loss(logits, labels):
    """
 Computes the total loss
Arguments:
logits -- output of forward propagation (output of the last LINEAR unit), of shape (6, num_examples)
labels -- "true" labels vector, same shape as Z3

Returns:
total_loss - Tensor of the total loss value
"""

#(1 line of code)
# remember to set `from_logits=True`
# total_loss = ...
# YOUR CODE STARTS HERE

Moderator Edit: Solution Code Removed.

# YOUR CODE ENDS HERE
return total_loss

Please provide your insight what might be the reason for failure.

2

Yes. Here is the checklist.

The compute_total_loss_test code is correct and it is just testing your code.

3

  1. Make sure you did the transpose on the inputs. Yeah (result = target(pred, tf.transpose(minibatch)))

  2. Make sure you use the from_logits option to tell the cost function that you are giving it logits and not activation values. - Yeah It is logits

  3. Make sure you use reduce_sum and not reduce_mean to get the final scalar value. - I have used the reduce_sum function

  4. Make sure you use the loss function specified in the instructions. Used tf.keras.losses.categorical_crossentropy

  5. Make sure you specify the positional arguments to the loss function in the correct order.

Yeah the positional arguments are corrtect.

4

No. You don’t have to do your code here. The checklist mentions the part of your code that you have to do between # YOUR CODE STARTS HERE and # YOUR CODE ENDS HERE. I see your code for compute_total_loss is not using transpose.

1 Like
5

Please do not share your code on the forum. That’s not allowed by the Code of Conduct.

Posting your error messages is sufficient.