When users attempt to run the software, they desire to get a deeper insight into this question: “Numpy dot vs matmul in Python“.

It is a frequent concern that all developers make. So, what is this, and how can it be remedied? We’ll go through it all with you.

What is the difference between Numpy dot vs matmul in Python?

Arrays are handled as vectors in Python. 2-D arrays are indeed known as matrices. In Python, researchers have features that can perform amplification between them. 

The numpy.dot() feature and the @ operator (the array’s __matmul__ method) have been used. It may appear how they both undertake the same multiplication purpose. Nevertheless, there is a distinction between the two, which is discussed in this guide.

In Python, the numpy.dot() feature is used to conduct matrix multiplication. It also verifies the multiplication condition, which states that the column in the first matrix should be equivalent to the row in the second. 

It also tends to work with multi-dimensional arrays. We could also clearly state an alternative array as a parameter to place the result. The @ operator for multiplying calls the array’s matmul() component, which is used to undertake the same multiplication. Let’s look at these below example to understand more about the difference between numpy dot vs matmul in Python.

Example 1:

import numpy as np 
arr1 = np.array([[3,2],[5,3]]) 
arr2 = np.array(([4,6],[4,7])) 
print(np.dot(arr1,arr2)) 
print(arr1@arr2)

Output:

[[20 32] 
[32 51]] 
[[20 32] 
[32 51]]

Example 2

import numpy as np 
arr1 = np.array([[4,5,6], [3,2,1], [1, 2, 3]])
arr2 = np.array([[[6,7,8], [7,8,9], [1, 2, 3]]]) 
print("vector multiplication") 
print(np.dot(arr1, arr2))

Output:

[[[65 80 95]] 
[[33 39 45]] 
[[23 29 35]]]

Example 3: 

The result is slightly different when we handle multi-dimensional arrays (N-D arrays with N>2). You can observe the distinction through the following example:

a = np.random.rand(2,3,3)
b = np.random.rand(2,3,3)
c = (a@b)
d = np.dot(a,b)
print(c,c.shape)
print(d,d.shape)

Output:

[[[0.63629871 0.55054463 0.22289276]
[1.27578425 1.13950519 0.55370078]
[1.37809353 1.32313811 0.75460862]]

[[1.63546361 1.54607801 0.67134528]
[1.05906619 1.07509384 0.42526795]
[1.38932102 1.32829749 0.47240808]]] (2, 3, 3)
[[[[0.63629871 0.55054463 0.22289276]
[0.7938068 0.85668481 0.26504028]]

[[1.27578425 1.13950519 0.55370078]
[1.55589497 1.45794424 0.5335743 ]]

[[1.37809353 1.32313811 0.75460862]
[1.60564885 1.39494713 0.59370927]]]


[[[1.48529826 1.55580834 0.96142976]
[1.63546361 1.54607801 0.67134528]]
[[0.94601586 0.97181894 0.56701004]
[1.05906619 1.07509384 0.42526795]]

[[1.13268609 1.00262696 0.47226983]
[1.38932102 1.32829749 0.47240808]]]] (2, 3, 2, 3)

The array is broadcast by the matmul() method as elements located in final two indexes, correspondingly, as a stack of matrices. On the contrary, the numpy.dot() function multiplies like the sum total of products over the final axis of the 1st array and the next-to-last of the second array.

The numpy.dot function may multiply an array with scalar values, whereas the matmul() function cannot. This is also a difference between the two functions.

Conclusion

The examples shown in the figure above has proven to be the most useful for those who are still bothered by this question: “Numpy dot vs matmul in Python“.

If you still need assistance or have frequent questions, we have a vibrant group where everybody is often willing to assist. Finally, we wish you a wonderful session filled with creative code remedies.

Read more

→ How To Write Numpy Array to CSV in Python?