Error while computing compute_total_loss Week 3 assignment

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

Hi,
I am doing the assignment for week 3 and passed all test cases before the function “compute_total_loss”. I am using tf.reduce_sum but the test case gives the following error message “Test does not match. Did you get the reduce sum of your loss functions?”.
While from_logits=False I am getting tf.Tensor(0.88275003, shape=(), dtype=float32)

Screenshot_12
Screenshot_12917×392 13.8 KB

when from_logits=True I am getting tf.Tensor(0.17102128, shape=(), dtype=float32)

Screenshot_13
Screenshot_13994×304 13 KB

2

Ok, you’ve correctly handled 2 out of the 3 issues mentioned on this thread. The only one you are missing is the transpose.

3

If you are surprised by the need for the transpose, it was mentioned in the instructions but only briefly. They didn’t really say much about why: here’s a thread which discusses that.

4

Thank you
I was able to fix this with the matrix transpose