Lecture 1

You might also like

You are on page 1of 1

ML Libraries: TensorFlow made by Google brain team (

Library:

Open source:

NumPy: Numerical Python

 fast rich expensive library


 only for numbers data types
Pandas:
 any time of data types (basic difference between numpy and pandas)

data should be preprocessed before machine learning by using numpy and pandas. All data must be in
the form of numbers whatever pics, videos, audios.

Vectorized Computation: fast (less time consuming 10-100 times faster), less number of steps

randn

Array

0 dimension is a scalar number have 0 dimension and no shape.

1D array (vector) is free of x, y and z axis. The elements of vectors is equal to the part of its shape (3, ),
(10, ), …

2D matrix (tensor)

Range in python = arange in numpy

Ravel

Shape of matrix must be same for mirror operation (element wise addition, subtraction, multiplication,
division)

Like = shape and data type

When Particular value is required we use indexing

Lecture 2

Fancy indexing:

You might also like