3. Image Classification with Convolutional Neural Networks (CNNs)

Activity 3.01: Building a Multiclass Classifier Based on the Fashion MNIST Dataset

Solution

1. Open a new Jupyter Notebook.
2. Import tensorflow.keras.datasets.fashion_mnist:

   from tensorflow.keras.datasets import fashion_mnist

3. Load the Fashion MNIST dataset using fashion_mnist.load_data() and save the results to (features_train, label_train), (features_test, label_test):

   (features_train, label_train), (features_test, label_test) = \
   fashion_mnist.load_data()

4. Print the shape of the training set:

   features_train.shape

   The output will be as follows:

   (60000, 28, 28)

   The training set is composed of 60000 images of size 28 by 28. We will need to reshape it and add the channel dimension.

5. Print the shape of the testing set:

   features_test.shape

   The output will be as follows:

   (10000, 28, 28)

   The testing set is composed of 10000 images of size 28 by 28. We will need to reshape it and add the channel dimension

...