Data augmentation in pipeline in TensorFlow - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to add random flipping to the image augmentation pipeline.

TensorFlow

data_augmentation = tf.keras.Sequential([
    tf.keras.layers.RandomFlip([1])
])

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

A'none'

B'horizontal'

C'vertical'

D'horizontal_and_vertical'

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'none' disables flipping, so no augmentation happens.

Choosing 'vertical' flips images upside down, which is less common.

The RandomFlip layer flips images horizontally when set to 'horizontal'.

2fill in blank

medium

Complete the code to add random rotation to the augmentation pipeline.

TensorFlow

data_augmentation = tf.keras.Sequential([
    tf.keras.layers.RandomRotation([1])
])

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

A0.2

B1.0

C2.0

D0.5

Attempts:

3 left

💡 Hint

Common Mistakes

Using values greater than 1 causes errors.

Using 0.5 rotates images too much, which may distort data.

The RandomRotation layer takes a float representing the fraction of 2π radians for max rotation. 0.2 means up to 20% of 360° rotation.

3fill in blank

hard

Fix the error in the code to correctly normalize images in the pipeline.

TensorFlow

normalization_layer = tf.keras.layers.Rescaling([1])

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

A1/255

B255

C0.255

D255.0

Attempts:

3 left

💡 Hint

Common Mistakes

Using 255 multiplies pixel values, making them larger.

Using 0.255 is not the correct scale factor.

The Rescaling layer expects a scale factor to multiply the input. Dividing by 255 normalizes pixel values to [0,1].

4fill in blank

hard

Fill both blanks to create a pipeline that randomly flips and normalizes images.

TensorFlow

data_augmentation = tf.keras.Sequential([
    tf.keras.layers.RandomFlip([1]),
    tf.keras.layers.Rescaling([2])
])

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

A'horizontal'

B'vertical'

C1/255

D255

Attempts:

3 left

💡 Hint

Common Mistakes

Using 255 for rescaling causes pixel values to increase.

Using 'vertical' flips images upside down, which is less common.

RandomFlip with 'horizontal' flips images left-right. Rescaling with 1/255 normalizes pixel values.

5fill in blank

hard

Fill all three blanks to build a pipeline that flips, rotates, and rescales images.

TensorFlow

data_augmentation = tf.keras.Sequential([
    tf.keras.layers.RandomFlip([1]),
    tf.keras.layers.RandomRotation([2]),
    tf.keras.layers.Rescaling([3])
])

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

A'horizontal'

B0.1

C1/255

D'vertical'

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'vertical' flip instead of 'horizontal'.

Using rotation values greater than 1 causes errors.

Using 255 instead of 1/255 for rescaling.

This pipeline flips images horizontally, rotates them up to 10% of 360°, and normalizes pixel values.