Convolutional Neural Network (CNN)

View on TensorFlow.org Run in Google Colab View source on GitHub Download notebook

This tutorial demonstrates training a simple Convolutional Neural Network (CNN) to classify CIFAR images. Because this tutorial uses the Keras Sequential API, creating and training our model will take just a few lines of code.

Import TensorFlow

from __future__ import absolute_import, division, print_function, unicode_literals

import tensorflow as tf

from tensorflow.keras import datasets, layers, models
import matplotlib.pyplot as plt

Download and prepare the CIFAR10 dataset

The CIFAR10 dataset contains 60,000 color images in 10 classes, with 6,000 images in each class. The dataset is divided into 50,000 training images and 10,000 testing images. The classes are mutually exclusive and there is no overlap between them.

(train_images, train_labels), (test_images, test_labels) = datasets.cifar10.load_data()

# Normalize pixel values to be between 0 and 1
train_images, test_images = train_images / 255.0, test_images / 255.0
Downloading data from https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz
170500096/170498071 [==============================] - 11s 0us/step

Verify the data

To verify that the dataset looks correct, let's plot the first 25 images from the training set and display the class name below each image.

class_names = ['airplane', 'automobile', 'bird', 'cat', 'deer',
               'dog', 'frog', 'horse', 'ship', 'truck']

plt.figure(figsize=(10,10))
for i in range(25):
    plt.subplot(5,5,i+1)
    plt.xticks([])
    plt.yticks([])
    plt.grid(False)
    plt.imshow(train_images[i], cmap=plt.cm.binary)
    # The CIFAR labels happen to be arrays, 
    # which is why you need the extra index
    plt.xlabel(class_names[train_labels[i][0]])
plt.show()

png

Create the convolutional base

The 6 lines of code below define the convolutional base using a common pattern: a stack of Conv2D and MaxPooling2D layers.

As input, a CNN takes tensors of shape (image_height, image_width, color_channels), ignoring the batch size. If you are new to these dimensions, color_channels refers to (R,G,B). In this example, you will configure our CNN to process inputs of shape (32, 32, 3), which is the format of CIFAR images. You can do this by passing the argument input_shape to our first layer.

model = models.Sequential()
model.add(layers.Conv2D(32, (3, 3), activation='relu', input_shape=(32, 32, 3)))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))

Let's display the architecture of our model so far.

model.summary()
Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d (Conv2D)              (None, 30, 30, 32)        896       
_________________________________________________________________
max_pooling2d (MaxPooling2D) (None, 15, 15, 32)        0         
_________________________________________________________________
conv2d_1 (Conv2D)            (None, 13, 13, 64)        18496     
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 6, 6, 64)          0         
_________________________________________________________________
conv2d_2 (Conv2D)            (None, 4, 4, 64)          36928     
=================================================================
Total params: 56,320
Trainable params: 56,320
Non-trainable params: 0
_________________________________________________________________

Above, you can see that the output of every Conv2D and MaxPooling2D layer is a 3D tensor of shape (height, width, channels). The width and height dimensions tend to shrink as you go deeper in the network. The number of output channels for each Conv2D layer is controlled by the first argument (e.g., 32 or 64). Typically, as the width and height shrink, you can afford (computationally) to add more output channels in each Conv2D layer.

Add Dense layers on top

To complete our model, you will feed the last output tensor from the convolutional base (of shape (3, 3, 64)) into one or more Dense layers to perform classification. Dense layers take vectors as input (which are 1D), while the current output is a 3D tensor. First, you will flatten (or unroll) the 3D output to 1D, then add one or more Dense layers on top. CIFAR has 10 output classes, so you use a final Dense layer with 10 outputs and a softmax activation.

model.add(layers.Flatten())
model.add(layers.Dense(64, activation='relu'))
model.add(layers.Dense(10, activation='softmax'))

Here's the complete architecture of our model.

model.summary()
Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d (Conv2D)              (None, 30, 30, 32)        896       
_________________________________________________________________
max_pooling2d (MaxPooling2D) (None, 15, 15, 32)        0         
_________________________________________________________________
conv2d_1 (Conv2D)            (None, 13, 13, 64)        18496     
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 6, 6, 64)          0         
_________________________________________________________________
conv2d_2 (Conv2D)            (None, 4, 4, 64)          36928     
_________________________________________________________________
flatten (Flatten)            (None, 1024)              0         
_________________________________________________________________
dense (Dense)                (None, 64)                65600     
_________________________________________________________________
dense_1 (Dense)              (None, 10)                650       
=================================================================
Total params: 122,570
Trainable params: 122,570
Non-trainable params: 0
_________________________________________________________________

As you can see, our (3, 3, 64) outputs were flattened into vectors of shape (576) before going through two Dense layers.

Compile and train the model

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

history = model.fit(train_images, train_labels, epochs=10, 
                    validation_data=(test_images, test_labels))
Train on 50000 samples, validate on 10000 samples
Epoch 1/10
50000/50000 [==============================] - 11s 222us/sample - loss: 1.5684 - accuracy: 0.4266 - val_loss: 1.2367 - val_accuracy: 0.5543
Epoch 2/10
50000/50000 [==============================] - 7s 130us/sample - loss: 1.1648 - accuracy: 0.5879 - val_loss: 1.1616 - val_accuracy: 0.5853
Epoch 3/10
50000/50000 [==============================] - 6s 130us/sample - loss: 1.0314 - accuracy: 0.6379 - val_loss: 1.0190 - val_accuracy: 0.6374
Epoch 4/10
50000/50000 [==============================] - 6s 129us/sample - loss: 0.9382 - accuracy: 0.6711 - val_loss: 0.9656 - val_accuracy: 0.6617
Epoch 5/10
50000/50000 [==============================] - 6s 129us/sample - loss: 0.8650 - accuracy: 0.6970 - val_loss: 0.9005 - val_accuracy: 0.6870
Epoch 6/10
50000/50000 [==============================] - 6s 129us/sample - loss: 0.8054 - accuracy: 0.7183 - val_loss: 0.9217 - val_accuracy: 0.6748
Epoch 7/10
50000/50000 [==============================] - 7s 131us/sample - loss: 0.7557 - accuracy: 0.7356 - val_loss: 0.9336 - val_accuracy: 0.6858
Epoch 8/10
50000/50000 [==============================] - 7s 132us/sample - loss: 0.7145 - accuracy: 0.7514 - val_loss: 0.9107 - val_accuracy: 0.6906
Epoch 9/10
50000/50000 [==============================] - 7s 134us/sample - loss: 0.6758 - accuracy: 0.7647 - val_loss: 0.8696 - val_accuracy: 0.7077
Epoch 10/10
50000/50000 [==============================] - 6s 130us/sample - loss: 0.6410 - accuracy: 0.7757 - val_loss: 0.8684 - val_accuracy: 0.7076

Evaluate the model

plt.plot(history.history['accuracy'], label='accuracy')
plt.plot(history.history['val_accuracy'], label = 'val_accuracy')
plt.xlabel('Epoch')
plt.ylabel('Accuracy')
plt.ylim([0.5, 1])
plt.legend(loc='lower right')

test_loss, test_acc = model.evaluate(test_images,  test_labels, verbose=2)
10000/1 - 1s - loss: 0.8927 - accuracy: 0.7076

png

print(test_acc)
0.7076

Our simple CNN has achieved a test accuracy of over 70%. Not bad for a few lines of code! For another CNN style, see an example using the Keras subclassing API and a tf.GradientTape here.