We cannot emphasize enough how important it is to understand the underlying dataset. In the current scenario, we are dealing with a visual dataset consisting of over 10,000 samples spread across 120 classes (dog breeds). Readers can refer to all the steps related to exploratory analysis in the IPython Notebook titled dog_breed_eda.ipynb.
Since this is a visual dataset, let's first visualize a few samples from the dataset. There are multiple ways to ingest and visualize image data in Python; we will be relying on SciPy and matplotlib-related utilities to do so. The following snippet imports the required libraries:
In [1]: import os ...: import scipy as sp ...: import numpy as np ...: import pandas as pd ...: ...: import ...
