NumPy

NumPy is a Python library that adds support for matrices and matrix operations. It’s comparable in some ways to operations supported in MATLAB.

We import with import numpy as np.

Arrays

The main feature of NumPy are arrays and matrices, which expand the functionality of standard lists in Python. We wrap rows in square brackets and separate rows with commas.

x = np.array([[1, 2], [3, 4]])

NumPy arrays have several attributes.

• x.shape gives the dimensions of the array.
• x.where(cond) allows us to conditionally find indices of data that meet a certain condition.
  • We can also access the data themselves with a conditional inside the index.

Interfaces with other packages

• To convert to a PyTorch tensor, use torch.from_numpy(arr).