Using AutoGluon for image data

Up to this point, we have been exploring AutoML methodologies on an Artificial Neural Network (ANN) algorithm. However, many use cases might require more complicated algorithms, such as Convolutional Neural Networks (CNNs) for image classification and image recognition, or Long-Short-Term Memory (LSTM) networks, for speech recognition and text data. Due to the complexity of these algorithms, many ML practitioners may have to leverage multiple machines for distributed training and potentially multiple GPUs to handle the multi-dimensional matrix calculations. In this section, we are going to segue from the Age Calculator use case to explore how AutoGluon can be used to apply an AutoML methodology to an image classification use case.

Note

Since we will be utilizing GPU-based AWS instances, running this example will exceed the usage limits of the AWS Free Tier and, therefore, incur additional costs.

Prerequisites

As was the case with the previous...