Training an image classifier in Computer Vision - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the necessary library for building a neural network model.

Computer Vision

from tensorflow import [1]

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

Akeras

Bpandas

Cnumpy

Dmatplotlib

Attempts:

3 left

💡 Hint

Common Mistakes

Importing unrelated libraries like pandas or matplotlib.

Trying to import tensorflow.keras as a separate package.

We use keras from TensorFlow to build and train neural network models.

2fill in blank

medium

Complete the code to load the CIFAR-10 dataset for training and testing.

Computer Vision

(train_images, train_labels), (test_images, test_labels) = keras.datasets.cifar10.[1]()

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

Aget_dataset

Bfetch_data

Cload_data

Ddownload

Attempts:

3 left

💡 Hint

Common Mistakes

Using non-existent functions like fetch_data or get_dataset.

Trying to call download() which is not a valid method here.

The load_data() function loads the CIFAR-10 dataset as training and testing sets.

3fill in blank

hard

Fix the error in the model compilation step by completing the optimizer argument.

Computer Vision

model.compile(optimizer='[1]', loss='sparse_categorical_crossentropy', metrics=['accuracy'])

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

Aadam

Brmsprop

Csgd

Drandom

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'random' which is not a valid optimizer.

Choosing 'sgd' without adjusting learning rate may slow training.

The 'adam' optimizer is commonly used for training image classifiers and works well here.

4fill in blank

hard

Fill both blanks to normalize the image pixel values between 0 and 1.

Computer Vision

train_images = train_images [1] 255.0
 test_images = test_images [2] 255.0

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

A/

B*

C-

D+

Attempts:

3 left

💡 Hint

Common Mistakes

Using multiplication or subtraction instead of division.

Normalizing only train or only test images.

Dividing pixel values by 255.0 scales them from 0-255 to 0-1, which helps model training.

5fill in blank

hard

Fill all three blanks to define a simple CNN model with Conv2D, MaxPooling2D, and Dense layers.

Computer Vision

model = keras.Sequential([
    keras.layers.Conv2D(32, (3, 3), activation='relu', input_shape=[1]),
    keras.layers.MaxPooling2D([2]),
    keras.layers.Flatten(),
    keras.layers.Dense([3], activation='softmax')
])

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

A(32, 32, 3)

B(2, 2)

C10

D(3, 3)

Attempts:

3 left

💡 Hint

Common Mistakes

Using wrong input shape like (28, 28, 1).

Using incorrect pool size or missing it.

Setting output units to a wrong number.

The input shape for CIFAR-10 images is (32, 32, 3). MaxPooling2D commonly uses a pool size of (2, 2). The Dense output layer has 10 units for the 10 classes.