Iterator

Set up an Iterator to extract data from an image dataset.

We'll cover the following

Chapter Goals:

  • Set up an Iterator to extract data from a pixel array dataset

A. Using Iterator

The way we can extract the decoded image data from our Dataset is through a tf.data.Iterator. For a more in-depth look at tf.data.Iterator, check out the Industry Case Study course on Educative.

We use the get_next function to obtain a next-element tensor, which is used for data extraction. Note that the next-element tensor doesn't have an actual value until we execute the iteration process using tf.compat.v1.Session (see next chapter).

Time to Code!

In this chapter you'll be working on the get_image_data function. This function uses an Iterator object to get decoded image data from a dataset.

Using the get_dataset function from the previous chapter (not shown), we can create our image pixel Dataset.

Set dataset equal to get_dataset with arguments image_paths, image_type, resize_shape, and channels.

We'll now make an Iterator for dataset.

Set iterator equal to tf.compat.v1.data.make_one_shot_iterator(dataset) with no arguments.

The one-shot iterator is a very simple iterator. It is associated with a particular dataset and only iterates through it once.

Finally, we set up the next-element tensor for extracting data from dataset.

Set next_image equal to iterator.get_next with no arguments.

Press + to interact
import tensorflow as tf
def get_image_data(image_paths, image_type=None, resize_shape=None, channels=0):
    # CODE HERE
    pass

Get hands-on with 1300+ tech skills courses.