Scaling a dataset to improve model accuracy

Scaling a dataset is the process of ensuring that the variables are confined to a finite range. In this section, we will confine the independent variables' values to values between 0 and 1 by dividing each input value by the maximum possible value in the dataset. This is a value of 255, which corresponds to white pixels:

The following code is available as Scaling_the_dataset.ipynb in the Chapter03 folder of this book's GitHub repository - https://tinyurl.com/mcvp-packt

1. Fetch the dataset, as well as the training images and targets, as we did in the previous section:

from torchvision import datasets
from torch.utils.data import Dataset, DataLoader
import torch
import torch.nn as nn
device = "cuda" if torch.cuda.is_available() else "cpu"
import numpy as np
data_folder = '~/data/FMNIST' # This can be any directory you 
# want to download FMNIST to
fmnist = datasets.FashionMNIST(data_folder, download=True, \
   ...