Handwriting recognition basics in Computer Vision - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load the MNIST dataset for handwriting recognition.

Computer Vision

from tensorflow.keras.datasets import [1]

(train_images, train_labels), (test_images, test_labels) = [1].load_data()

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

Amnist

Bcifar10

Cfashion_mnist

Dimdb

Attempts:

3 left

💡 Hint

Common Mistakes

Using CIFAR-10 which is for object images, not handwriting.

Using IMDB which is for text data.

Using Fashion MNIST which is for clothing images.

The MNIST dataset contains handwritten digit images commonly used for handwriting recognition tasks.

2fill in blank

medium

Complete the code to normalize the pixel values of the images between 0 and 1.

Computer Vision

train_images = train_images.astype('float32') / [1]
test_images = test_images.astype('float32') / [1]

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

A256

B100

C1

D255

Attempts:

3 left

💡 Hint

Common Mistakes

Dividing by 100 or 256 which does not scale correctly.

Dividing by 1 which does nothing.

Pixel values range from 0 to 255, so dividing by 255 scales them to 0-1.

3fill in blank

hard

Fix the error in the model definition by choosing the correct activation function for the output layer.

Computer Vision

from tensorflow.keras import models, layers

model = models.Sequential([
    layers.Flatten(input_shape=(28, 28)),
    layers.Dense(128, activation='relu'),
    layers.Dense(10, activation=[1])
])

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

A'sigmoid'

B'softmax'

C'relu'

D'tanh'

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'relu' or 'tanh' which are not suitable for output probabilities.

Using 'sigmoid' which is for binary classification.

For multi-class classification like digits 0-9, 'softmax' activation outputs probabilities for each class.

4fill in blank

hard

Fill both blanks to compile the model with the correct loss function and optimizer for handwriting recognition.

Computer Vision

model.compile(optimizer=[1], loss=[2], metrics=['accuracy'])

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

A'adam'

B'sgd'

C'sparse_categorical_crossentropy'

D'binary_crossentropy'

Attempts:

3 left

💡 Hint

Common Mistakes

Using binary crossentropy which is for binary classification.

Using SGD optimizer without adaptive learning rate.

Adam optimizer is popular for training neural networks. Sparse categorical crossentropy is used for multi-class classification with integer labels.

5fill in blank

hard

Fill all three blanks to train the model for 5 epochs with a batch size of 64 and validate on test data.

Computer Vision

history = model.fit(train_images, train_labels, epochs=[1], batch_size=[2], validation_data=([3], test_labels))

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

A32

B64

Ctest_images

D5

Attempts:

3 left

💡 Hint

Common Mistakes

Using batch size 32 instead of 64 as specified.

Using train_images instead of test_images for validation.

Training for 5 epochs with batch size 64 is common. Validation data uses test images and labels.