Preparing the datasets for image classification using the Apache MXNet Vision Datasets classes
In this recipe, we will set up the file and directory structure needed for the image classification experiments in this chapter. We will create five directories inside the tmp directory—train, validation, train_lst, validation_lst, and test. After that, we will use the Apache MXNet Vision Datasets classes to load the datasets required to train and test the image classification models in this chapter. We will perform the train-test split, store the loaded data as image files, and generate the .lst files that will be used for the training job.
Figure 8.17 – MNIST dataset
We have in Figure 8.17 a few sample image files that will be prepared in this recipe. In the recipe Training and deploying an image classifier using the built-in image classification algorithm in SageMaker, we will use these image files to train an image classifier model that can recognize...
