Convolutional Neural Networks

In this notebook, we train an MLP to classify images from the CIFAR-10 database.

1. Load CIFAR-10 Database



In [1]:

    
import keras
from keras.datasets import cifar10

# load the pre-shuffled train and test data
(x_train, y_train), (x_test, y_test) = cifar10.load_data()









    



Using TensorFlow backend.

2. Visualize the First 24 Training Images



In [2]:

    
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline

fig = plt.figure(figsize=(20,5))
for i in range(36):
    ax = fig.add_subplot(3, 12, i + 1, xticks=[], yticks=[])
    ax.imshow(np.squeeze(x_train[i]))

3. Rescale the Images by Dividing Every Pixel in Every Image by 255



In [3]:

    
# rescale [0,255] --> [0,1]
x_train = x_train.astype('float32')/255
x_test = x_test.astype('float32')/255

4. Break Dataset into Training, Testing, and Validation Sets



In [4]:

    
from keras.utils import np_utils

# one-hot encode the labels
num_classes = len(np.unique(y_train))
y_train = keras.utils.to_categorical(y_train, num_classes)
y_test = keras.utils.to_categorical(y_test, num_classes)

# break training set into training and validation sets
(x_train, x_valid) = x_train[5000:], x_train[:5000]
(y_train, y_valid) = y_train[5000:], y_train[:5000]

# print shape of training set
print('x_train shape:', x_train.shape)

# print number of training, validation, and test images
print(x_train.shape[0], 'train samples')
print(x_test.shape[0], 'test samples')
print(x_valid.shape[0], 'validation samples')









    



x_train shape: (45000, 32, 32, 3)
45000 train samples
10000 test samples
5000 validation samples

5. Define the Model Architecture



In [5]:

    
from keras.models import Sequential
from keras.layers import Dense, Dropout, Flatten

# define the model
model = Sequential()
model.add(Flatten(input_shape = x_train.shape[1:]))
model.add(Dense(1000, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(512, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(num_classes, activation='softmax'))

model.summary()









    



_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
flatten_1 (Flatten)          (None, 3072)              0         
_________________________________________________________________
dense_1 (Dense)              (None, 1000)              3073000   
_________________________________________________________________
dropout_1 (Dropout)          (None, 1000)              0         
_________________________________________________________________
dense_2 (Dense)              (None, 512)               512512    
_________________________________________________________________
dropout_2 (Dropout)          (None, 512)               0         
_________________________________________________________________
dense_3 (Dense)              (None, 10)                5130      
=================================================================
Total params: 3,590,642.0
Trainable params: 3,590,642.0
Non-trainable params: 0.0
_________________________________________________________________

6. Compile the Model



In [6]:

    
# compile the model
model.compile(loss='categorical_crossentropy', optimizer='rmsprop', 
                  metrics=['accuracy'])

7. Train the Model



In [7]:

    
from keras.callbacks import ModelCheckpoint   

# train the model
checkpointer = ModelCheckpoint(filepath='MLP.weights.best.hdf5', verbose=1, 
                               save_best_only=True)
hist = model.fit(x_train, y_train, batch_size=32, epochs=20,
          validation_data=(x_valid, y_valid), callbacks=[checkpointer], 
          verbose=2, shuffle=True)









    



Train on 45000 samples, validate on 5000 samples
Epoch 1/20
Epoch 00000: val_loss improved from inf to 1.91876, saving model to MLP.weights.best.hdf5
52s - loss: 3.2886 - acc: 0.2462 - val_loss: 1.9188 - val_acc: 0.3000
Epoch 2/20
Epoch 00001: val_loss did not improve
54s - loss: 1.8626 - acc: 0.3242 - val_loss: 1.9204 - val_acc: 0.3082
Epoch 3/20
Epoch 00002: val_loss improved from 1.91876 to 1.78092, saving model to MLP.weights.best.hdf5
52s - loss: 1.8230 - acc: 0.3438 - val_loss: 1.7809 - val_acc: 0.3588
Epoch 4/20
Epoch 00003: val_loss improved from 1.78092 to 1.72077, saving model to MLP.weights.best.hdf5
54s - loss: 1.7887 - acc: 0.3575 - val_loss: 1.7208 - val_acc: 0.3640
Epoch 5/20
Epoch 00004: val_loss did not improve
55s - loss: 1.7777 - acc: 0.3651 - val_loss: 1.7357 - val_acc: 0.3500
Epoch 6/20
Epoch 00005: val_loss improved from 1.72077 to 1.71538, saving model to MLP.weights.best.hdf5
53s - loss: 1.7641 - acc: 0.3675 - val_loss: 1.7154 - val_acc: 0.3818
Epoch 7/20
Epoch 00006: val_loss did not improve
52s - loss: 1.7616 - acc: 0.3700 - val_loss: 1.7708 - val_acc: 0.3670
Epoch 8/20
Epoch 00007: val_loss did not improve
52s - loss: 1.7641 - acc: 0.3729 - val_loss: 1.7766 - val_acc: 0.3738
Epoch 9/20
Epoch 00008: val_loss improved from 1.71538 to 1.70597, saving model to MLP.weights.best.hdf5
52s - loss: 1.7709 - acc: 0.3672 - val_loss: 1.7060 - val_acc: 0.3840
Epoch 10/20
Epoch 00009: val_loss did not improve
51s - loss: 1.7635 - acc: 0.3744 - val_loss: 1.8535 - val_acc: 0.3260
Epoch 11/20
Epoch 00010: val_loss did not improve
54s - loss: 1.7551 - acc: 0.3780 - val_loss: 1.7249 - val_acc: 0.3758
Epoch 12/20
Epoch 00011: val_loss did not improve
55s - loss: 1.7617 - acc: 0.3757 - val_loss: 1.7308 - val_acc: 0.3660
Epoch 13/20
Epoch 00012: val_loss did not improve
52s - loss: 1.7694 - acc: 0.3745 - val_loss: 1.9086 - val_acc: 0.3150
Epoch 14/20
Epoch 00013: val_loss did not improve
53s - loss: 1.7654 - acc: 0.3711 - val_loss: 1.7625 - val_acc: 0.3684
Epoch 15/20
Epoch 00014: val_loss did not improve
52s - loss: 1.7691 - acc: 0.3726 - val_loss: 1.7753 - val_acc: 0.3778
Epoch 16/20
Epoch 00015: val_loss did not improve
52s - loss: 1.7780 - acc: 0.3688 - val_loss: 1.7723 - val_acc: 0.3592
Epoch 17/20
Epoch 00016: val_loss did not improve
52s - loss: 1.7757 - acc: 0.3675 - val_loss: 1.7359 - val_acc: 0.3644
Epoch 18/20
Epoch 00017: val_loss did not improve
54s - loss: 1.7868 - acc: 0.3676 - val_loss: 1.7861 - val_acc: 0.3538
Epoch 19/20
Epoch 00018: val_loss did not improve
53s - loss: 1.7797 - acc: 0.3717 - val_loss: 1.7431 - val_acc: 0.3698
Epoch 20/20
Epoch 00019: val_loss improved from 1.70597 to 1.70173, saving model to MLP.weights.best.hdf5
52s - loss: 1.7857 - acc: 0.3670 - val_loss: 1.7017 - val_acc: 0.3926






    Out[7]:





<keras.callbacks.History at 0x10935a4a8>

8. Load the Model with the Best Classification Accuracy on the Validation Set



In [8]:

    
# load the weights that yielded the best validation accuracy
model.load_weights('MLP.weights.best.hdf5')

9. Calculate Classification Accuracy on Test Set



In [9]:

    
# evaluate and print test accuracy
score = model.evaluate(x_test, y_test, verbose=0)
print('\n', 'Test accuracy:', score[1])









    



 Test accuracy: 0.4