C2_W2_Assignment Exercise 6

Course Q&A Math for Machine Learning Calculus for Machine Learning and Data Science
1

I am getting errors with this code:

def gradient_descent(dEdm, dEdb, m, b, X, Y, learning_rate = 0.001, num_iterations = 1000, print_cost=False):
for iteration in range(num_iterations):
### START CODE HERE ### (~ 2 lines of code)
m_new = m-(learning_ratedEdm)
b_new = b-(learning_ratedEdb)
### END CODE HERE ###
m = m_new
b = b_new
if print_cost:
print (f"Cost after iteration {iteration}: {E(m, b, X, Y)}")

return m, b

The error is:

TypeError Traceback (most recent call last)
Cell In [28], line 1
----> 1 print(gradient_descent(dEdm, dEdb, 0, 0, X_norm, Y_norm))
2 print(gradient_descent(dEdm, dEdb, 1, 5, X_norm, Y_norm, learning_rate = 0.01, num_iterations = 10))

Cell In [27], line 4, in gradient_descent(dEdm, dEdb, m, b, X, Y, learning_rate, num_iterations, print_cost)
1 def gradient_descent(dEdm, dEdb, m, b, X, Y, learning_rate = 0.001, num_iterations = 1000, print_cost=False):
2 for iteration in range(num_iterations):
3 ### START CODE HERE ### (~ 2 lines of code)
----> 4 m_new = m-(learning_ratedEdm)
5 b_new = b-(learning_ratedEdb)
6 ### END CODE HERE ###

TypeError: unsupported operand type(s) for *: ‘float’ and ‘function’

Could you provide me some help or hint?

2

The error here might be in the learning_ratedEdm if that is function you need to call it like function(parameters)…and will produce some output that you can use further on.

1 Like
4

I tried many ways (np.dot(_)) but still got an error. Please tell me how to fix it

5

@gent.spah Please help me how to figure out

6

You should not publish code solution here as its against code of conduct.

Your error seems to be the same as above, make sure that dEdm which seems to be a function is assigned to a variable and multiplication can happen.

Other than this I dont have access to this course’s assignment.

7

Dear Carlos,

Let me see if I can help.

According to what I see on your code. There is an issue with m_new and b_new.

Remember that for gradient descend you need to do the following equation:
m_new = m - learning_rate*dEdm
but dEdm is a function defined on python which means you need to call it as a function and inser the parameters over there.

In order to call a function you need to use parenthesis and include the parameters on the parenthesis something like this:

dEdm(m, b, X, Y)

The same applies for the function dEdb(m, b, X, Y)

I hope this helps!

2 Likes