Course 1 Week 4 Exercise 7

Course Q&A Deep Learning Specialization Neural Networks and Deep Learning
1

Hi,

I’m working on writing up the linear_backward function and I have the following:

{moderator edit - solution code removed}

I am facing the error of TypeError: tuple indices must be integers or slices, not str.

Am I calling the cache wrongly?

Thank you.

2

Yes, the cache is a list or “tuple”, but you are treating it like a python dictionary with string keys.

Note that they already gave you the code in the template to extract the contents of the cache:

A_prev, W, b = cache

I suggest you read the template code a bit more carefully. :nerd_face:

1 Like
3

Hi @paulinpaloalto

Thank you for getting back to me. I must say my knowledge of python is really showing in this exercise. Are caches more like lists?

Also, in the previous example (example 5), why were parameters called via parameters[“W” + str(l)] and not just parameters[“W”]?

4

There are many different data structures in python. You have to look at the code and understand which one is being used. Did you read the template code that I pointed out? Just because parameters is a dictionary, that doesn’t many that any other particular variable is also a dictionary.

The particular code that references parameters is retrieving the weight matrix for a particular layer. The keys for those are the string names W1, W2 and so forth. You have the layer number in a variable l and you want to create a string that starts with W and includes the layer number. One easy way to do that in python is “W” + str(l). What the str() function does is convert whatever the argument is to a string. And what the + operator means for strings is concatenation.

1 Like
5

I see!!! I was wondering about that. Thank you.

6

@paulinpaloalto, very well explained sir. Thank you for this clarification. These short inputs and clarifications are so useful for us as well. Keep shining!

7

The other point here is that python is an interpreted language. It is interactive. You don’t have to wonder what something does: you can try it and see. For example, if you want to know what type caches is, do this:

print(f"type(caches) = {type(caches)}")

If you want to understand what “W + str(l)” does try some examples:

myStr = "Fred"
myNumber = 3.0
print(myStr)
longStr = myStr + "  " + "Flintstone" + "  " + str(myNumber)
print(longStr)