Why efficient data loading prevents bottlenecks in TensorFlow - Test Your Understanding

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a TensorFlow dataset from a list of file paths.

TensorFlow

dataset = tf.data.Dataset.from_tensor_slices([1])

Drag options to blanks, or click blank then click option'

Anum_epochs

Bfile_paths

Cshuffle_buffer

Dbatch_size

Attempts:

3 left

💡 Hint

Common Mistakes

Using batch size or shuffle buffer instead of the data source.

Confusing dataset creation parameters.

The from_tensor_slices method creates a dataset from the given list or array, here file_paths.

2fill in blank

medium

Complete the code to apply a map function to preprocess images in the dataset.

TensorFlow

dataset = dataset.map([1])

Drag options to blanks, or click blank then click option'

Ashuffle

Bbatch

Cload_and_preprocess_image

Drepeat

Attempts:

3 left

💡 Hint

Common Mistakes

Using batch or shuffle methods inside map.

Passing a method name that is not a function.

The map method applies a function to each element in the dataset. Here, load_and_preprocess_image is the function to load and preprocess images.

3fill in blank

hard

Fix the error in the code to enable parallel data loading.

TensorFlow

dataset = dataset.map(load_and_preprocess_image, [1]=tf.data.AUTOTUNE)

Drag options to blanks, or click blank then click option'

Arepeat_count

Bbatch_size

Cshuffle_buffer_size

Dnum_parallel_calls

Attempts:

3 left

💡 Hint

Common Mistakes

Using batch_size or shuffle_buffer_size as map arguments.

Not enabling parallel calls causes slow data loading.

The num_parallel_calls argument allows parallel processing of dataset elements, improving loading speed.

4fill in blank

hard

Fill both blanks to batch the dataset and prefetch data for efficiency.

TensorFlow

dataset = dataset.[1](32).[2](tf.data.AUTOTUNE)

Drag options to blanks, or click blank then click option'

Abatch

Bshuffle

Cprefetch

Drepeat

Attempts:

3 left

💡 Hint

Common Mistakes

Using shuffle instead of batch.

Not using prefetch causes slower training.

Batching groups data into batches of 32, and prefetching overlaps data preparation with model training to prevent bottlenecks.

5fill in blank

hard

Fill all three blanks to shuffle, batch, and prefetch the dataset correctly.

TensorFlow

dataset = dataset.[1](1000).[2](64).[3](tf.data.AUTOTUNE)

Drag options to blanks, or click blank then click option'

Ashuffle

Bbatch

Cprefetch

Drepeat

Attempts:

3 left

💡 Hint

Common Mistakes

Incorrect order of operations.

Skipping prefetch causes bottlenecks.

Shuffling randomizes data order, batching groups data, and prefetching loads data in the background to keep training smooth.