In [1]:

    

import matplotlib.pyplot as plt
from matplotlib import ticker

import seaborn as sns

from keras.preprocessing.image import ImageDataGenerator
from keras.models import Sequential
from keras.layers import Convolution2D, MaxPooling2D
from keras.layers import Activation, Dropout, Flatten, Dense
from keras.callbacks import ModelCheckpoint, Callback, EarlyStopping, TensorBoard


    

    




Using TensorFlow backend.

In [2]:

    

# dimensions of our images.
img_width, img_height = 50, 50

train_data_dir = 'data/symlinks/train'
validation_data_dir = 'data/symlinks/validation'
test_data_dir = 'data/symlinks/test'
nb_epoch = 10
batch_size=128


    

In [3]:

    

# this is the augmentation configuration we will use for training
train_datagen = ImageDataGenerator(
        samplewise_center=True,
        rescale=1./255,
        shear_range=0.2,
        zoom_range=0.2,
        horizontal_flip=True)

# this is the augmentation configuration we will use for testing:
# only rescaling
test_datagen = ImageDataGenerator(
        samplewise_center=True,
        rescale=1./255
)


    

In [4]:

    

train_generator = train_datagen.flow_from_directory(
        train_data_dir,
        target_size=(img_width, img_height),
        batch_size=batch_size,
        class_mode='binary',
        follow_links=True)

validation_generator = test_datagen.flow_from_directory(
        validation_data_dir,
        target_size=(img_width, img_height),
        batch_size=batch_size,
        class_mode='binary',
        follow_links=True)

test_generator = test_datagen.flow_from_directory(
        test_data_dir,
        target_size=(img_width, img_height),
        batch_size=batch_size,
        shuffle=False,
        class_mode='binary',
        follow_links=True)


    

    




Found 325298 images belonging to 2 classes.
Found 219720 images belonging to 2 classes.
Found 150833 images belonging to 2 classes.

In [5]:

    

def get_model():
    model = Sequential()
    model.add(Convolution2D(32, (3, 3), input_shape=(img_width, img_height, 3)))
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))
    
    model.add(Convolution2D(32, (3, 3)))
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))
    
    model.add(Convolution2D(64, (3, 3)))
    model.add(Activation('relu'))
    model.add(MaxPooling2D(pool_size=(2, 2)))
    
    model.add(Flatten())
    model.add(Dense(64))
    model.add(Activation('relu'))
    model.add(Dropout(0.5))
    model.add(Dense(1))
    model.add(Activation('sigmoid'))
    
    return model

model = get_model()


    

In [6]:

    

model.summary()


    

    




_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d_1 (Conv2D)            (None, 48, 48, 32)        896       
_________________________________________________________________
activation_1 (Activation)    (None, 48, 48, 32)        0         
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 24, 24, 32)        0         
_________________________________________________________________
conv2d_2 (Conv2D)            (None, 22, 22, 32)        9248      
_________________________________________________________________
activation_2 (Activation)    (None, 22, 22, 32)        0         
_________________________________________________________________
max_pooling2d_2 (MaxPooling2 (None, 11, 11, 32)        0         
_________________________________________________________________
conv2d_3 (Conv2D)            (None, 9, 9, 64)          18496     
_________________________________________________________________
activation_3 (Activation)    (None, 9, 9, 64)          0         
_________________________________________________________________
max_pooling2d_3 (MaxPooling2 (None, 4, 4, 64)          0         
_________________________________________________________________
flatten_1 (Flatten)          (None, 1024)              0         
_________________________________________________________________
dense_1 (Dense)              (None, 64)                65600     
_________________________________________________________________
activation_4 (Activation)    (None, 64)                0         
_________________________________________________________________
dropout_1 (Dropout)          (None, 64)                0         
_________________________________________________________________
dense_2 (Dense)              (None, 1)                 65        
_________________________________________________________________
activation_5 (Activation)    (None, 1)                 0         
=================================================================
Total params: 94,305.0
Trainable params: 94,305.0
Non-trainable params: 0.0
_________________________________________________________________

In [7]:

    

## Callback for loss logging per epoch
class LossHistory(Callback):
    def on_train_begin(self, logs={}):
        self.losses = []
        self.val_losses = []

    def on_epoch_end(self, batch, logs={}):
        self.losses.append(logs.get('loss'))
        self.val_losses.append(logs.get('val_loss'))
        
history = LossHistory()


    

In [8]:

    

tensorboard = TensorBoard(log_dir='./logs', histogram_freq=1, write_graph=True, write_images=False)
checkpointer = ModelCheckpoint(filepath="weights.hdf5", verbose=0, save_best_only=True)


    

In [ ]:

    

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

model.fit_generator(
        train_generator,
        steps_per_epoch=train_generator.samples/batch_size,
        epochs=nb_epoch,
        validation_data=validation_generator,
        verbose=1,
        validation_steps=validation_generator.samples/batch_size,
        workers=8,
        pickle_safe=True,
        callbacks=[history, tensorboard, checkpointer])


    

    




Epoch 1/10
2542/2541 [==============================] - 144s - loss: 0.0697 - acc: 0.9773 - val_loss: 0.0529 - val_acc: 0.9825
Epoch 2/10
2542/2541 [==============================] - 161s - loss: 0.0343 - acc: 0.9901 - val_loss: 0.0371 - val_acc: 0.9876
Epoch 3/10
2542/2541 [==============================] - 183s - loss: 0.0287 - acc: 0.9920 - val_loss: 0.0381 - val_acc: 0.9881
Epoch 4/10
2542/2541 [==============================] - 201s - loss: 0.0232 - acc: 0.9936 - val_loss: 0.0298 - val_acc: 0.9901
Epoch 5/10
2542/2541 [==============================] - 202s - loss: 0.0215 - acc: 0.9941 - val_loss: 0.0328 - val_acc: 0.9898
Epoch 6/10
2542/2541 [==============================] - 195s - loss: 0.0204 - acc: 0.9947 - val_loss: 0.0307 - val_acc: 0.9906
Epoch 7/10
2412/2541 [===========================>..] - ETA: 8s - loss: 0.0189 - acc: 0.9950

In [10]:

    

loss = history.losses
val_loss = history.val_losses

plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.title('Loss Trend')
plt.plot(loss, 'blue', label='Training Loss')
plt.plot(val_loss, 'green', label='Validation Loss')
plt.xticks(range(0,nb_epoch)[0::2])
plt.legend()

plt.show()


    

In [11]:

    

test_loss, test_acc = model.evaluate_generator(
         test_generator,
         workers=8,
         pickle_safe=True,
         steps=test_generator.samples/batch_size)

print("test_loss: %.4f - test_acc: %.4f"%(test_loss, test_acc))


    

    




test_loss: 0.0122 - test_acc: 0.9961

In [12]:

    

from IPython.display import SVG
from keras.utils.vis_utils import model_to_dot

SVG(model_to_dot(model,show_shapes=True, show_layer_names=True).create(prog='dot', format='svg'))


    

    
Out[12]:

In [ ]: