Things spotted in C3_W1_Assignment (shape error and compute accuracy denominator)

[Anna_Sztyber]

Hi everyone,
this is what I noticed doing C3_W1_Assignment:

1. Exercise 2 - data_generator - shape error
   The easy mistake is not converting example_weights to np.array.
   In that case, w1_unittest.test_data_generator is giving a confusing message because all tests for output shapes are printing result[0].shape

2. Exercise 7 - compute_accuracy
   # Sum up the weighted correct predictions (of type np.float32), to go in the # denominator. → should be nominator

[arvyzukai]

Hi @Anna_Sztyber

@Anna_Sztyber

# Sum up the weighted correct predictions (of type np.float32), to go in the # denominator. → should be nominator

Actually it should be numerator Thank you for noticing this mistake

@Anna_Sztyber

Exercise 2 - data_generator - shape error
The easy mistake is not converting example_weights to np.array.
In that case, w1_unittest.test_data_generator is giving a confusing message because all tests for output shapes are printing result[0].shape

Yes, the docstring asks for the output of:

example_weights - An array specifying the importance of each example

It could have been stated clearer, that the array is numpy array.

Cheers