Not able to modify the numpy array in python

Question

I am a beginner in python . I am trying to modify a numpy array but somehow it is not getting modify. Here is my program

def test_numpy(x):
    count = 0
    for i in x:
        i-=np.max(i)
        i=(np.exp(i)/np.sum(np.exp(i)))
        print "The value of i is "
        print i
        x[count] = i
        count+=1
        print "the value of x is "
        print x

if __name__ == "__main__":
    test_numpy(np.array([[1,2],[3,4]])).

The output it prints is :

The value of i is [0.26894142 0.73105858] the value of x is [[0 0] [3 4]] The value of i is [0.26894142 0.73105858] the value of x is [[0 0] [0 0]]

I am assuming the value of x should be overridden by the value of i . So after looping twice the value of x should become ([0.26894142 0.73105858],[0.26894142 0.73105858]) But somehow the value is not getting overridden. Can anyone please point out my mistake here

Matthieu Brucher · Accepted Answer · 2018-12-16 12:57:59Z

2

When you do:

i=(np.exp(i)/np.sum(np.exp(i)))

You are creating a new variable, you are not changing the data in place. You forgot to tell Python to modify the data:

i[:]=(np.exp(i)/np.sum(np.exp(i)))

answered Dec 16, 2018 at 12:57

Matthieu Brucher

22.1k7 gold badges43 silver badges66 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Daniel · Accepted Answer · 2018-12-16 13:56:41Z

0

Your input array is of type int, so no floating point numbers can be saved. You have to provide a float array:

q=np.array([[1,2],[3,4]], dtype=float)
test_numpy(q)
print q

or you don't change the array, but create a new one, then the resulting type is automatically chosen correctly:

def test_numpy(x):
    x = np.exp(x - x.max(axis=1)[:, None])
    return x / x.sum(axis=1)[:, None]

q = np.array([[1,2],[3,4]])
a = test_numpy(q)
print(a)

edited Dec 16, 2018 at 13:56

answered Dec 16, 2018 at 13:49

Daniel

42.9k4 gold badges57 silver badges82 bronze badges

Collectives™ on Stack Overflow

Not able to modify the numpy array in python

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related