Question with C2W3 assignment

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

In the week3 assignment - Tensorflow Introduction

Why need to tf.transpose(labels), tf.transpose(logits) in the compute_total_loss(logits, labels) function?

2

The short answer is, you have to get the labels in the correct shape for TensorFlow to use them in this function.