Search⌘ K
Home
Courses
Image Recognition with Machine Learning

Initialization

Explore the initialization process of SqueezeNet for image classification using the CIFAR-10 dataset. Understand how to set image dimensions, handle RGB input data, and configure sparse label representations in a Python class setup.

We'll cover the following...

Chapter Goals:

  • Learn about the CIFAR-10 dataset
  • Initialize the model with data dimensions

A. CIFAR-10

The CIFAR-10 (Canadian Institute for Advanced Research) dataset contains 60,000 color images with dimensions 32x32. The images are distributed evenly across 10 categories: airplane, automobile, bird, cat, deer, dog, frog, horse, ship, truck. We split the dataset into 50,000 images for training and 10,000 images for testing.

Example images from the CIFAR-10 dataset.
Example images from the CIFAR-10 dataset.

The CIFAR-10 dataset is available for download (along with the 100 category version, CIFAR-100), on Alex Krizhevsky's website.

(Fun fact: Alex Krizhevsky helped invent the AlexNet model mentioned in the previous chapter, which incidentally was named after him. The LeNet model from the CNN section was also named after its inventor, Yann LeCun. ...