Data Preprocessing (Part 1)
Explore the tf.data API to preprocess TensorFlow datasets through transformations such as shuffle to randomize data, repeat to duplicate datasets, and batch to group data samples. Understand how to chain these operations for efficient dataset management, preparing data effectively for deep learning on Android.
The TF framework provides us with the tf.data API to construct TF datasets. This API allows us to preprocess TF datasets using various transformations, such as the shuffle, repeat, batch, and unbatch transformations. Let’s explore the use of these transformations.
The shuffle transformation
The shuffle transformation randomly mixes the elements of tf.data.Dataset. The following code applies this ...