In [1]:

    

import os
import keras
import numpy as np
import tensorflow as tf
from keras.datasets import cifar10
from keras.preprocessing.image import ImageDataGenerator
from keras.models import Sequential
from keras.layers import Dense, Dropout, Activation, Flatten
from keras.layers import Conv2D, MaxPooling2D, GlobalAveragePooling2D, AveragePooling2D
from keras.initializers import he_normal
from keras import optimizers
from keras.callbacks import LearningRateScheduler, TensorBoard
from keras.layers.normalization import BatchNormalization
from keras.utils.data_utils import get_file


    

    




Using TensorFlow backend.

In [2]:

    

os.environ["CUDA_VISIBLE_DEVICES"] = "3"
from keras.backend.tensorflow_backend import set_session
config = tf.ConfigProto()
config.gpu_options.per_process_gpu_memory_fraction = 0.4
set_session(tf.Session(config=config))


    

In [3]:

    

num_classes  = 10
batch_size   = 128
epochs       = 170
iterations   = 391
dropout      = 0.5
log_filepath = r'./vgg19_retrain/'


    

In [4]:

    

def color_preprocessing(x_train,x_test):
    x_train = x_train.astype('float32')
    x_test = x_test.astype('float32')
    # data preprocessing 
    x_train[:,:,:,0] = (x_train[:,:,:,0]-123.680)
    x_train[:,:,:,1] = (x_train[:,:,:,1]-116.779)
    x_train[:,:,:,2] = (x_train[:,:,:,2]-103.939)
    x_test[:,:,:,0] = (x_test[:,:,:,0]-123.680)
    x_test[:,:,:,1] = (x_test[:,:,:,1]-116.779)
    x_test[:,:,:,2] = (x_test[:,:,:,2]-103.939)

    return x_train, x_test


    

In [5]:

    

def scheduler(epoch):
  learning_rate_init = 0.01
  if epoch > 80:
    learning_rate_init = 0.001
  if epoch > 120:
    learning_rate_init = 0.0001
  return learning_rate_init


    

In [6]:

    

def VGG19():
    model = Sequential()

    # Block 1
    model.add(Conv2D(64, (3, 3), padding='same', name='block1_conv1', input_shape=x_train.shape[1:]))
    model.add(Activation('relu'))
    model.add(Conv2D(64, (3, 3), padding='same', name='block1_conv2'))
    model.add(Activation('relu'))
    model.add(MaxPooling2D((2, 2), strides=(2, 2), name='block1_pool'))

    # Block 2
    model.add(Conv2D(128, (3, 3), padding='same', name='block2_conv1'))
    model.add(Activation('relu'))
    model.add(Conv2D(128, (3, 3), padding='same', name='block2_conv2'))
    model.add(Activation('relu'))
    model.add(MaxPooling2D((2, 2), strides=(2, 2), name='block2_pool'))

    # Block 3
    model.add(Conv2D(256, (3, 3), padding='same', name='block3_conv1'))
    model.add(Activation('relu'))
    model.add(Conv2D(256, (3, 3), padding='same', name='block3_conv2'))
    model.add(Activation('relu'))
    model.add(Conv2D(256, (3, 3), padding='same', name='block3_conv3'))
    model.add(Activation('relu'))
    model.add(Conv2D(256, (3, 3), padding='same', name='block3_conv4'))
    model.add(Activation('relu'))
    model.add(MaxPooling2D((2, 2), strides=(2, 2), name='block3_pool'))

    # Block 4
    model.add(Conv2D(512, (3, 3), padding='same', name='block4_conv1'))
    model.add(Activation('relu'))
    model.add(Conv2D(512, (3, 3), padding='same', name='block4_conv2'))
    model.add(Activation('relu'))
    model.add(Conv2D(512, (3, 3), padding='same', name='block4_conv3'))
    model.add(Activation('relu'))
    model.add(Conv2D(512, (3, 3), padding='same', name='block4_conv4'))
    model.add(Activation('relu'))
    model.add(MaxPooling2D((2, 2), strides=(2, 2), name='block4_pool'))

    # Block 5
    model.add(Conv2D(512, (3, 3), padding='same', name='block5_conv1'))
    model.add(Activation('relu'))
    model.add(Conv2D(512, (3, 3), padding='same', name='block5_conv2'))
    model.add(Activation('relu'))
    model.add(Conv2D(512, (3, 3), padding='same', name='block5_conv3'))
    model.add(Activation('relu'))
    model.add(Conv2D(512, (3, 3), padding='same', name='block5_conv4'))
    model.add(Activation('relu'))

    # model modification for cifar-10
    model.add(Flatten(name='flatten'))
    model.add(Dense(4096, use_bias = True, name='fc_cifa10'))
    model.add(Activation('relu'))
    model.add(Dropout(dropout))
    model.add(Dense(4096, name='fc2'))  
    model.add(Activation('relu'))
    model.add(Dropout(dropout))      
    model.add(Dense(10, name='predictions_cifa10'))        
    model.add(Activation('softmax'))
    
    return model


    

In [7]:

    

(x_train, y_train), (x_test, y_test) = cifar10.load_data()
y_train = keras.utils.to_categorical(y_train, num_classes)
y_test = keras.utils.to_categorical(y_test, num_classes)
# color preprocessing
x_train, x_test = color_preprocessing(x_train, x_test)

# build network
WEIGHTS_PATH = 'https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg19_weights_tf_dim_ordering_tf_kernels.h5'
filepath = get_file('vgg19_weights_tf_dim_ordering_tf_kernels.h5', WEIGHTS_PATH, cache_subdir='models')

model = VGG19()
print(model.summary())

# load pretrained weight from VGG19 by name      
model.load_weights(filepath, by_name=True)

# -------- optimizer setting -------- #
sgd = optimizers.SGD(lr=.1, momentum=0.9, nesterov=True)
model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy'])


    

    




_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
block1_conv1 (Conv2D)        (None, 32, 32, 64)        1792      
_________________________________________________________________
activation_1 (Activation)    (None, 32, 32, 64)        0         
_________________________________________________________________
block1_conv2 (Conv2D)        (None, 32, 32, 64)        36928     
_________________________________________________________________
activation_2 (Activation)    (None, 32, 32, 64)        0         
_________________________________________________________________
block1_pool (MaxPooling2D)   (None, 16, 16, 64)        0         
_________________________________________________________________
block2_conv1 (Conv2D)        (None, 16, 16, 128)       73856     
_________________________________________________________________
activation_3 (Activation)    (None, 16, 16, 128)       0         
_________________________________________________________________
block2_conv2 (Conv2D)        (None, 16, 16, 128)       147584    
_________________________________________________________________
activation_4 (Activation)    (None, 16, 16, 128)       0         
_________________________________________________________________
block2_pool (MaxPooling2D)   (None, 8, 8, 128)         0         
_________________________________________________________________
block3_conv1 (Conv2D)        (None, 8, 8, 256)         295168    
_________________________________________________________________
activation_5 (Activation)    (None, 8, 8, 256)         0         
_________________________________________________________________
block3_conv2 (Conv2D)        (None, 8, 8, 256)         590080    
_________________________________________________________________
activation_6 (Activation)    (None, 8, 8, 256)         0         
_________________________________________________________________
block3_conv3 (Conv2D)        (None, 8, 8, 256)         590080    
_________________________________________________________________
activation_7 (Activation)    (None, 8, 8, 256)         0         
_________________________________________________________________
block3_conv4 (Conv2D)        (None, 8, 8, 256)         590080    
_________________________________________________________________
activation_8 (Activation)    (None, 8, 8, 256)         0         
_________________________________________________________________
block3_pool (MaxPooling2D)   (None, 4, 4, 256)         0         
_________________________________________________________________
block4_conv1 (Conv2D)        (None, 4, 4, 512)         1180160   
_________________________________________________________________
activation_9 (Activation)    (None, 4, 4, 512)         0         
_________________________________________________________________
block4_conv2 (Conv2D)        (None, 4, 4, 512)         2359808   
_________________________________________________________________
activation_10 (Activation)   (None, 4, 4, 512)         0         
_________________________________________________________________
block4_conv3 (Conv2D)        (None, 4, 4, 512)         2359808   
_________________________________________________________________
activation_11 (Activation)   (None, 4, 4, 512)         0         
_________________________________________________________________
block4_conv4 (Conv2D)        (None, 4, 4, 512)         2359808   
_________________________________________________________________
activation_12 (Activation)   (None, 4, 4, 512)         0         
_________________________________________________________________
block4_pool (MaxPooling2D)   (None, 2, 2, 512)         0         
_________________________________________________________________
block5_conv1 (Conv2D)        (None, 2, 2, 512)         2359808   
_________________________________________________________________
activation_13 (Activation)   (None, 2, 2, 512)         0         
_________________________________________________________________
block5_conv2 (Conv2D)        (None, 2, 2, 512)         2359808   
_________________________________________________________________
activation_14 (Activation)   (None, 2, 2, 512)         0         
_________________________________________________________________
block5_conv3 (Conv2D)        (None, 2, 2, 512)         2359808   
_________________________________________________________________
activation_15 (Activation)   (None, 2, 2, 512)         0         
_________________________________________________________________
block5_conv4 (Conv2D)        (None, 2, 2, 512)         2359808   
_________________________________________________________________
activation_16 (Activation)   (None, 2, 2, 512)         0         
_________________________________________________________________
flatten (Flatten)            (None, 2048)              0         
_________________________________________________________________
fc_cifa10 (Dense)            (None, 4096)              8392704   
_________________________________________________________________
activation_17 (Activation)   (None, 4096)              0         
_________________________________________________________________
dropout_1 (Dropout)          (None, 4096)              0         
_________________________________________________________________
fc2 (Dense)                  (None, 4096)              16781312  
_________________________________________________________________
activation_18 (Activation)   (None, 4096)              0         
_________________________________________________________________
dropout_2 (Dropout)          (None, 4096)              0         
_________________________________________________________________
predictions_cifa10 (Dense)   (None, 10)                40970     
_________________________________________________________________
activation_19 (Activation)   (None, 10)                0         
=================================================================
Total params: 45,239,370
Trainable params: 45,239,370
Non-trainable params: 0
_________________________________________________________________
None

In [8]:

    

tb_cb = TensorBoard(log_dir=log_filepath, histogram_freq=0)
change_lr = LearningRateScheduler(scheduler)
cbks = [change_lr,tb_cb]


    

In [9]:

    

print('Using real-time data augmentation.')
datagen = ImageDataGenerator(horizontal_flip=True,
        width_shift_range=0.125,height_shift_range=0.125,fill_mode='constant',cval=0.)

datagen.fit(x_train)


    

    




Using real-time data augmentation.

In [10]:

    

model.fit_generator(datagen.flow(x_train, y_train,batch_size=batch_size),
                    steps_per_epoch=iterations,
                    epochs=epochs,
                    callbacks=cbks,
                    validation_data=(x_test, y_test))
model.save('vgg19_retrain.h5')


    

    




Epoch 1/170
391/391 [==============================] - 32s - loss: 1.4998 - acc: 0.4619 - val_loss: 0.8923 - val_acc: 0.7058
Epoch 2/170
391/391 [==============================] - 29s - loss: 0.7768 - acc: 0.7419 - val_loss: 0.6509 - val_acc: 0.7868
Epoch 3/170
391/391 [==============================] - 29s - loss: 0.6148 - acc: 0.7978 - val_loss: 0.5877 - val_acc: 0.8053
Epoch 4/170
391/391 [==============================] - 30s - loss: 0.5344 - acc: 0.8241 - val_loss: 0.5552 - val_acc: 0.8177
Epoch 5/170
391/391 [==============================] - 30s - loss: 0.4701 - acc: 0.8463 - val_loss: 0.4700 - val_acc: 0.8444
Epoch 6/170
391/391 [==============================] - 30s - loss: 0.4239 - acc: 0.8590 - val_loss: 0.4514 - val_acc: 0.8486
Epoch 7/170
391/391 [==============================] - 32s - loss: 0.3835 - acc: 0.8724 - val_loss: 0.4556 - val_acc: 0.8543
Epoch 8/170
391/391 [==============================] - 59s - loss: 0.3516 - acc: 0.8828 - val_loss: 0.4382 - val_acc: 0.8636
Epoch 9/170
391/391 [==============================] - 59s - loss: 0.3260 - acc: 0.8924 - val_loss: 0.4342 - val_acc: 0.8628
Epoch 10/170
391/391 [==============================] - 59s - loss: 0.2983 - acc: 0.8993 - val_loss: 0.4373 - val_acc: 0.8661
Epoch 11/170
391/391 [==============================] - 59s - loss: 0.2837 - acc: 0.9049 - val_loss: 0.3983 - val_acc: 0.8678
Epoch 12/170
391/391 [==============================] - 59s - loss: 0.2612 - acc: 0.9134 - val_loss: 0.4134 - val_acc: 0.8698
Epoch 13/170
391/391 [==============================] - 60s - loss: 0.2456 - acc: 0.9183 - val_loss: 0.3810 - val_acc: 0.8824
Epoch 14/170
391/391 [==============================] - 59s - loss: 0.2340 - acc: 0.9223 - val_loss: 0.3993 - val_acc: 0.8778
Epoch 15/170
391/391 [==============================] - 60s - loss: 0.2186 - acc: 0.9285 - val_loss: 0.3656 - val_acc: 0.8847
Epoch 16/170
391/391 [==============================] - 60s - loss: 0.2012 - acc: 0.9330 - val_loss: 0.3736 - val_acc: 0.8827
Epoch 17/170
391/391 [==============================] - 60s - loss: 0.1923 - acc: 0.9357 - val_loss: 0.3772 - val_acc: 0.8842
Epoch 18/170
391/391 [==============================] - 60s - loss: 0.1800 - acc: 0.9404 - val_loss: 0.4177 - val_acc: 0.8770
Epoch 19/170
391/391 [==============================] - 60s - loss: 0.1751 - acc: 0.9416 - val_loss: 0.3891 - val_acc: 0.8867
Epoch 20/170
391/391 [==============================] - 60s - loss: 0.1600 - acc: 0.9454 - val_loss: 0.3982 - val_acc: 0.8788
Epoch 21/170
391/391 [==============================] - 60s - loss: 0.1550 - acc: 0.9466 - val_loss: 0.3833 - val_acc: 0.8898
Epoch 22/170
391/391 [==============================] - 60s - loss: 0.1470 - acc: 0.9507 - val_loss: 0.3829 - val_acc: 0.8865
Epoch 23/170
391/391 [==============================] - 60s - loss: 0.1410 - acc: 0.9534 - val_loss: 0.3653 - val_acc: 0.8917
Epoch 24/170
391/391 [==============================] - 60s - loss: 0.1347 - acc: 0.9554 - val_loss: 0.4055 - val_acc: 0.8877
Epoch 25/170
391/391 [==============================] - 60s - loss: 0.1282 - acc: 0.9568 - val_loss: 0.3890 - val_acc: 0.8887
Epoch 26/170
391/391 [==============================] - 60s - loss: 0.1230 - acc: 0.9586 - val_loss: 0.3851 - val_acc: 0.8918
Epoch 27/170
391/391 [==============================] - 60s - loss: 0.1133 - acc: 0.9630 - val_loss: 0.4082 - val_acc: 0.8899
Epoch 28/170
391/391 [==============================] - 60s - loss: 0.1116 - acc: 0.9624 - val_loss: 0.3777 - val_acc: 0.8928
Epoch 29/170
391/391 [==============================] - 60s - loss: 0.1094 - acc: 0.9630 - val_loss: 0.4308 - val_acc: 0.8896
Epoch 30/170
391/391 [==============================] - 60s - loss: 0.1077 - acc: 0.9642 - val_loss: 0.4133 - val_acc: 0.8920
Epoch 31/170
391/391 [==============================] - 60s - loss: 0.1007 - acc: 0.9662 - val_loss: 0.3906 - val_acc: 0.8932
Epoch 32/170
391/391 [==============================] - 60s - loss: 0.0961 - acc: 0.9676 - val_loss: 0.3690 - val_acc: 0.8999
Epoch 33/170
391/391 [==============================] - 60s - loss: 0.0921 - acc: 0.9692 - val_loss: 0.4521 - val_acc: 0.8909
Epoch 34/170
391/391 [==============================] - 60s - loss: 0.0917 - acc: 0.9706 - val_loss: 0.4052 - val_acc: 0.8968
Epoch 35/170
391/391 [==============================] - 60s - loss: 0.0891 - acc: 0.9700 - val_loss: 0.3931 - val_acc: 0.8976
Epoch 36/170
391/391 [==============================] - 60s - loss: 0.0789 - acc: 0.9741 - val_loss: 0.3902 - val_acc: 0.8956
Epoch 37/170
391/391 [==============================] - 60s - loss: 0.0738 - acc: 0.9757 - val_loss: 0.4377 - val_acc: 0.8966
Epoch 38/170
391/391 [==============================] - 60s - loss: 0.0765 - acc: 0.9745 - val_loss: 0.4491 - val_acc: 0.8890
Epoch 39/170
391/391 [==============================] - 60s - loss: 0.0694 - acc: 0.9767 - val_loss: 0.4250 - val_acc: 0.8963
Epoch 40/170
391/391 [==============================] - 60s - loss: 0.0720 - acc: 0.9763 - val_loss: 0.3926 - val_acc: 0.9011
Epoch 41/170
391/391 [==============================] - 60s - loss: 0.0656 - acc: 0.9777 - val_loss: 0.4813 - val_acc: 0.8964
Epoch 42/170
391/391 [==============================] - 60s - loss: 0.0666 - acc: 0.9784 - val_loss: 0.4421 - val_acc: 0.8945
Epoch 43/170
391/391 [==============================] - 60s - loss: 0.0651 - acc: 0.9786 - val_loss: 0.4410 - val_acc: 0.8949
Epoch 44/170
391/391 [==============================] - 60s - loss: 0.0628 - acc: 0.9792 - val_loss: 0.4352 - val_acc: 0.8976
Epoch 45/170
391/391 [==============================] - 60s - loss: 0.0625 - acc: 0.9795 - val_loss: 0.4385 - val_acc: 0.8984
Epoch 46/170
391/391 [==============================] - 60s - loss: 0.0634 - acc: 0.9790 - val_loss: 0.4120 - val_acc: 0.8986
Epoch 47/170
391/391 [==============================] - 60s - loss: 0.0549 - acc: 0.9821 - val_loss: 0.4520 - val_acc: 0.9030
Epoch 48/170
391/391 [==============================] - 60s - loss: 0.0555 - acc: 0.9824 - val_loss: 0.4363 - val_acc: 0.9024
Epoch 49/170
391/391 [==============================] - 60s - loss: 0.0510 - acc: 0.9830 - val_loss: 0.4404 - val_acc: 0.8970
Epoch 50/170
391/391 [==============================] - 59s - loss: 0.0540 - acc: 0.9821 - val_loss: 0.4509 - val_acc: 0.8966
Epoch 51/170
391/391 [==============================] - 60s - loss: 0.0518 - acc: 0.9831 - val_loss: 0.4697 - val_acc: 0.8940
Epoch 52/170
391/391 [==============================] - 60s - loss: 0.0488 - acc: 0.9840 - val_loss: 0.4651 - val_acc: 0.9011
Epoch 53/170
391/391 [==============================] - 60s - loss: 0.0521 - acc: 0.9830 - val_loss: 0.4067 - val_acc: 0.9069
Epoch 54/170
391/391 [==============================] - 60s - loss: 0.0456 - acc: 0.9853 - val_loss: 0.4457 - val_acc: 0.9027
Epoch 55/170
391/391 [==============================] - 60s - loss: 0.0445 - acc: 0.9859 - val_loss: 0.4092 - val_acc: 0.9064
Epoch 56/170
391/391 [==============================] - 60s - loss: 0.0445 - acc: 0.9850 - val_loss: 0.4368 - val_acc: 0.9012
Epoch 57/170
391/391 [==============================] - 60s - loss: 0.0427 - acc: 0.9859 - val_loss: 0.4330 - val_acc: 0.9014
Epoch 58/170
391/391 [==============================] - 60s - loss: 0.0455 - acc: 0.9853 - val_loss: 0.4081 - val_acc: 0.9009
Epoch 59/170
391/391 [==============================] - 60s - loss: 0.0424 - acc: 0.9861 - val_loss: 0.4043 - val_acc: 0.9062
Epoch 60/170
391/391 [==============================] - 60s - loss: 0.0407 - acc: 0.9867 - val_loss: 0.4508 - val_acc: 0.8989
Epoch 61/170
391/391 [==============================] - 61s - loss: 0.0396 - acc: 0.9874 - val_loss: 0.4511 - val_acc: 0.9039
Epoch 62/170
391/391 [==============================] - 60s - loss: 0.0393 - acc: 0.9874 - val_loss: 0.4344 - val_acc: 0.9056
Epoch 63/170
391/391 [==============================] - 60s - loss: 0.0383 - acc: 0.9879 - val_loss: 0.4306 - val_acc: 0.9026
Epoch 64/170
391/391 [==============================] - 60s - loss: 0.0399 - acc: 0.9873 - val_loss: 0.4541 - val_acc: 0.9021
Epoch 65/170
391/391 [==============================] - 60s - loss: 0.0388 - acc: 0.9872 - val_loss: 0.4749 - val_acc: 0.9053
Epoch 66/170
391/391 [==============================] - 60s - loss: 0.0331 - acc: 0.9899 - val_loss: 0.4197 - val_acc: 0.9062
Epoch 67/170
391/391 [==============================] - 60s - loss: 0.0332 - acc: 0.9894 - val_loss: 0.4713 - val_acc: 0.9034
Epoch 68/170
391/391 [==============================] - 60s - loss: 0.0351 - acc: 0.9890 - val_loss: 0.4457 - val_acc: 0.9075
Epoch 69/170
391/391 [==============================] - 60s - loss: 0.0312 - acc: 0.9897 - val_loss: 0.4721 - val_acc: 0.9025
Epoch 70/170
391/391 [==============================] - 60s - loss: 0.0310 - acc: 0.9900 - val_loss: 0.4882 - val_acc: 0.9042
Epoch 71/170
391/391 [==============================] - 60s - loss: 0.0338 - acc: 0.9887 - val_loss: 0.4365 - val_acc: 0.9032
Epoch 72/170
391/391 [==============================] - 60s - loss: 0.0276 - acc: 0.9908 - val_loss: 0.5012 - val_acc: 0.9005
Epoch 73/170
391/391 [==============================] - 60s - loss: 0.0344 - acc: 0.9885 - val_loss: 0.4497 - val_acc: 0.9046
Epoch 74/170
391/391 [==============================] - 60s - loss: 0.0298 - acc: 0.9908 - val_loss: 0.4682 - val_acc: 0.9048
Epoch 75/170
391/391 [==============================] - 60s - loss: 0.0298 - acc: 0.9908 - val_loss: 0.5020 - val_acc: 0.9024
Epoch 76/170
391/391 [==============================] - 60s - loss: 0.0304 - acc: 0.9906 - val_loss: 0.4416 - val_acc: 0.9069
Epoch 77/170
391/391 [==============================] - 60s - loss: 0.0297 - acc: 0.9908 - val_loss: 0.4562 - val_acc: 0.9092
Epoch 78/170
391/391 [==============================] - 60s - loss: 0.0283 - acc: 0.9914 - val_loss: 0.4444 - val_acc: 0.9048
Epoch 79/170
391/391 [==============================] - 60s - loss: 0.0292 - acc: 0.9907 - val_loss: 0.4850 - val_acc: 0.9015
Epoch 80/170
391/391 [==============================] - 60s - loss: 0.0277 - acc: 0.9914 - val_loss: 0.4236 - val_acc: 0.9077
Epoch 81/170
391/391 [==============================] - 60s - loss: 0.0291 - acc: 0.9910 - val_loss: 0.5120 - val_acc: 0.8969
Epoch 82/170
391/391 [==============================] - 60s - loss: 0.0169 - acc: 0.9947 - val_loss: 0.4215 - val_acc: 0.9121
Epoch 83/170
391/391 [==============================] - 60s - loss: 0.0076 - acc: 0.9976 - val_loss: 0.4331 - val_acc: 0.9129
Epoch 84/170
391/391 [==============================] - 60s - loss: 0.0051 - acc: 0.9984 - val_loss: 0.4454 - val_acc: 0.9141
Epoch 85/170
391/391 [==============================] - 60s - loss: 0.0049 - acc: 0.9986 - val_loss: 0.4556 - val_acc: 0.9153
Epoch 86/170
391/391 [==============================] - 60s - loss: 0.0048 - acc: 0.9986 - val_loss: 0.4639 - val_acc: 0.9140
Epoch 87/170
391/391 [==============================] - 60s - loss: 0.0043 - acc: 0.9987 - val_loss: 0.4644 - val_acc: 0.9148
Epoch 88/170
391/391 [==============================] - 60s - loss: 0.0034 - acc: 0.9989 - val_loss: 0.4726 - val_acc: 0.9145
Epoch 89/170
391/391 [==============================] - 61s - loss: 0.0027 - acc: 0.9992 - val_loss: 0.4860 - val_acc: 0.9141
Epoch 90/170
391/391 [==============================] - 60s - loss: 0.0030 - acc: 0.9989 - val_loss: 0.4885 - val_acc: 0.9140
Epoch 91/170
391/391 [==============================] - 60s - loss: 0.0036 - acc: 0.9989 - val_loss: 0.4873 - val_acc: 0.9156
Epoch 92/170
391/391 [==============================] - 60s - loss: 0.0030 - acc: 0.9992 - val_loss: 0.5038 - val_acc: 0.9153
Epoch 93/170
391/391 [==============================] - 60s - loss: 0.0026 - acc: 0.9992 - val_loss: 0.4947 - val_acc: 0.9159
Epoch 94/170
391/391 [==============================] - 60s - loss: 0.0028 - acc: 0.9993 - val_loss: 0.4980 - val_acc: 0.9160
Epoch 95/170
391/391 [==============================] - 60s - loss: 0.0022 - acc: 0.9994 - val_loss: 0.5047 - val_acc: 0.9158
Epoch 96/170
391/391 [==============================] - 60s - loss: 0.0019 - acc: 0.9995 - val_loss: 0.5084 - val_acc: 0.9165
Epoch 97/170
391/391 [==============================] - 60s - loss: 0.0026 - acc: 0.9991 - val_loss: 0.5075 - val_acc: 0.9144
Epoch 98/170
391/391 [==============================] - 60s - loss: 0.0020 - acc: 0.9994 - val_loss: 0.5088 - val_acc: 0.9166
Epoch 99/170
391/391 [==============================] - 60s - loss: 0.0028 - acc: 0.9991 - val_loss: 0.5042 - val_acc: 0.9163
Epoch 100/170
391/391 [==============================] - 59s - loss: 0.0015 - acc: 0.9995 - val_loss: 0.5082 - val_acc: 0.9175
Epoch 101/170
391/391 [==============================] - 60s - loss: 0.0018 - acc: 0.9994 - val_loss: 0.5124 - val_acc: 0.9171
Epoch 102/170
391/391 [==============================] - 60s - loss: 0.0023 - acc: 0.9993 - val_loss: 0.5179 - val_acc: 0.9168
Epoch 103/170
391/391 [==============================] - 60s - loss: 0.0017 - acc: 0.9995 - val_loss: 0.5161 - val_acc: 0.9152
Epoch 104/170
391/391 [==============================] - 60s - loss: 0.0016 - acc: 0.9994 - val_loss: 0.5301 - val_acc: 0.9151
Epoch 105/170
391/391 [==============================] - 60s - loss: 0.0018 - acc: 0.9995 - val_loss: 0.5250 - val_acc: 0.9169
Epoch 106/170
391/391 [==============================] - 60s - loss: 0.0019 - acc: 0.9994 - val_loss: 0.5242 - val_acc: 0.9164
Epoch 107/170
391/391 [==============================] - 60s - loss: 0.0016 - acc: 0.9994 - val_loss: 0.5357 - val_acc: 0.9165
Epoch 108/170
391/391 [==============================] - 60s - loss: 0.0014 - acc: 0.9996 - val_loss: 0.5291 - val_acc: 0.9177
Epoch 109/170
391/391 [==============================] - 60s - loss: 0.0019 - acc: 0.9994 - val_loss: 0.5280 - val_acc: 0.9162
Epoch 110/170
391/391 [==============================] - 60s - loss: 0.0014 - acc: 0.9995 - val_loss: 0.5376 - val_acc: 0.9172
Epoch 111/170
391/391 [==============================] - 60s - loss: 0.0012 - acc: 0.9997 - val_loss: 0.5377 - val_acc: 0.9180
Epoch 112/170
391/391 [==============================] - 60s - loss: 0.0016 - acc: 0.9994 - val_loss: 0.5390 - val_acc: 0.9167
Epoch 113/170
391/391 [==============================] - 60s - loss: 0.0014 - acc: 0.9996 - val_loss: 0.5441 - val_acc: 0.9171
Epoch 114/170
391/391 [==============================] - 60s - loss: 0.0015 - acc: 0.9995 - val_loss: 0.5464 - val_acc: 0.9164
Epoch 115/170
391/391 [==============================] - 60s - loss: 0.0012 - acc: 0.9997 - val_loss: 0.5465 - val_acc: 0.9163
Epoch 116/170
391/391 [==============================] - 60s - loss: 0.0012 - acc: 0.9997 - val_loss: 0.5505 - val_acc: 0.9172
Epoch 117/170
391/391 [==============================] - 60s - loss: 0.0014 - acc: 0.9995 - val_loss: 0.5504 - val_acc: 0.9175
Epoch 118/170
391/391 [==============================] - 60s - loss: 0.0012 - acc: 0.9996 - val_loss: 0.5489 - val_acc: 0.9165
Epoch 119/170
391/391 [==============================] - 60s - loss: 9.1770e-04 - acc: 0.9996 - val_loss: 0.5543 - val_acc: 0.9168
Epoch 120/170
391/391 [==============================] - 60s - loss: 0.0012 - acc: 0.9995 - val_loss: 0.5490 - val_acc: 0.9176
Epoch 121/170
391/391 [==============================] - 60s - loss: 0.0012 - acc: 0.9995 - val_loss: 0.5497 - val_acc: 0.9176
Epoch 122/170
391/391 [==============================] - 60s - loss: 0.0012 - acc: 0.9997 - val_loss: 0.5505 - val_acc: 0.9168
Epoch 123/170
391/391 [==============================] - 60s - loss: 9.0148e-04 - acc: 0.9997 - val_loss: 0.5509 - val_acc: 0.9171
Epoch 124/170
391/391 [==============================] - 60s - loss: 8.6409e-04 - acc: 0.9998 - val_loss: 0.5511 - val_acc: 0.9172
Epoch 125/170
391/391 [==============================] - 60s - loss: 7.9260e-04 - acc: 0.9998 - val_loss: 0.5509 - val_acc: 0.9172
Epoch 126/170
391/391 [==============================] - 60s - loss: 4.8098e-04 - acc: 0.9999 - val_loss: 0.5510 - val_acc: 0.9170
Epoch 127/170
391/391 [==============================] - 60s - loss: 9.1096e-04 - acc: 0.9997 - val_loss: 0.5512 - val_acc: 0.9171
Epoch 128/170
391/391 [==============================] - 60s - loss: 8.7066e-04 - acc: 0.9997 - val_loss: 0.5515 - val_acc: 0.9169
Epoch 129/170
391/391 [==============================] - 60s - loss: 7.7703e-04 - acc: 0.9998 - val_loss: 0.5519 - val_acc: 0.9174
Epoch 130/170
391/391 [==============================] - 60s - loss: 6.2042e-04 - acc: 0.9998 - val_loss: 0.5517 - val_acc: 0.9177
Epoch 131/170
391/391 [==============================] - 60s - loss: 6.8525e-04 - acc: 0.9998 - val_loss: 0.5516 - val_acc: 0.9171
Epoch 132/170
391/391 [==============================] - 60s - loss: 8.4728e-04 - acc: 0.9998 - val_loss: 0.5519 - val_acc: 0.9169
Epoch 133/170
391/391 [==============================] - 60s - loss: 8.6710e-04 - acc: 0.9997 - val_loss: 0.5520 - val_acc: 0.9176
Epoch 134/170
391/391 [==============================] - 60s - loss: 6.2568e-04 - acc: 0.9998 - val_loss: 0.5519 - val_acc: 0.9173
Epoch 135/170
391/391 [==============================] - 60s - loss: 4.6392e-04 - acc: 0.9999 - val_loss: 0.5525 - val_acc: 0.9171
Epoch 136/170
391/391 [==============================] - 60s - loss: 7.2151e-04 - acc: 0.9997 - val_loss: 0.5534 - val_acc: 0.9174
Epoch 137/170
391/391 [==============================] - 60s - loss: 6.8707e-04 - acc: 0.9998 - val_loss: 0.5537 - val_acc: 0.9175
Epoch 138/170
391/391 [==============================] - 60s - loss: 9.7959e-04 - acc: 0.9997 - val_loss: 0.5523 - val_acc: 0.9173
Epoch 139/170
391/391 [==============================] - 60s - loss: 6.2659e-04 - acc: 0.9998 - val_loss: 0.5534 - val_acc: 0.9167
Epoch 140/170
391/391 [==============================] - 60s - loss: 0.0010 - acc: 0.9997 - val_loss: 0.5532 - val_acc: 0.9169
Epoch 141/170
391/391 [==============================] - 60s - loss: 7.0019e-04 - acc: 0.9998 - val_loss: 0.5537 - val_acc: 0.9170
Epoch 142/170
391/391 [==============================] - 60s - loss: 0.0011 - acc: 0.9997 - val_loss: 0.5543 - val_acc: 0.9173
Epoch 143/170
391/391 [==============================] - 60s - loss: 7.0663e-04 - acc: 0.9998 - val_loss: 0.5542 - val_acc: 0.9172
Epoch 144/170
391/391 [==============================] - 60s - loss: 6.9069e-04 - acc: 0.9997 - val_loss: 0.5541 - val_acc: 0.9171
Epoch 145/170
391/391 [==============================] - 59s - loss: 6.0200e-04 - acc: 0.9998 - val_loss: 0.5544 - val_acc: 0.9173
Epoch 146/170
391/391 [==============================] - 59s - loss: 6.8912e-04 - acc: 0.9998 - val_loss: 0.5549 - val_acc: 0.9173
Epoch 147/170
391/391 [==============================] - 59s - loss: 0.0010 - acc: 0.9997 - val_loss: 0.5550 - val_acc: 0.9172
Epoch 148/170
391/391 [==============================] - 60s - loss: 7.9533e-04 - acc: 0.9998 - val_loss: 0.5556 - val_acc: 0.9168
Epoch 149/170
391/391 [==============================] - 60s - loss: 7.7223e-04 - acc: 0.9997 - val_loss: 0.5546 - val_acc: 0.9171
Epoch 150/170
391/391 [==============================] - 60s - loss: 6.2405e-04 - acc: 0.9998 - val_loss: 0.5546 - val_acc: 0.9175
Epoch 151/170
391/391 [==============================] - 60s - loss: 6.9738e-04 - acc: 0.9999 - val_loss: 0.5557 - val_acc: 0.9177
Epoch 152/170
391/391 [==============================] - 60s - loss: 7.7314e-04 - acc: 0.9998 - val_loss: 0.5558 - val_acc: 0.9175
Epoch 153/170
391/391 [==============================] - 60s - loss: 9.1681e-04 - acc: 0.9997 - val_loss: 0.5553 - val_acc: 0.9169
Epoch 154/170
391/391 [==============================] - 60s - loss: 7.0240e-04 - acc: 0.9998 - val_loss: 0.5556 - val_acc: 0.9170
Epoch 155/170
391/391 [==============================] - 59s - loss: 6.1999e-04 - acc: 0.9998 - val_loss: 0.5566 - val_acc: 0.9172
Epoch 156/170
391/391 [==============================] - 60s - loss: 7.2874e-04 - acc: 0.9997 - val_loss: 0.5576 - val_acc: 0.9171
Epoch 157/170
391/391 [==============================] - 60s - loss: 7.2218e-04 - acc: 0.9998 - val_loss: 0.5575 - val_acc: 0.9173
Epoch 158/170
391/391 [==============================] - 60s - loss: 5.5770e-04 - acc: 0.9999 - val_loss: 0.5571 - val_acc: 0.9172
Epoch 159/170
391/391 [==============================] - 60s - loss: 5.4413e-04 - acc: 0.9999 - val_loss: 0.5575 - val_acc: 0.9170
Epoch 160/170
391/391 [==============================] - 60s - loss: 6.9955e-04 - acc: 0.9998 - val_loss: 0.5579 - val_acc: 0.9171
Epoch 161/170
391/391 [==============================] - 60s - loss: 5.2776e-04 - acc: 0.9998 - val_loss: 0.5577 - val_acc: 0.9169
Epoch 162/170
391/391 [==============================] - 60s - loss: 8.2080e-04 - acc: 0.9998 - val_loss: 0.5580 - val_acc: 0.9170
Epoch 163/170
391/391 [==============================] - 59s - loss: 6.6092e-04 - acc: 0.9998 - val_loss: 0.5574 - val_acc: 0.9170
Epoch 164/170
391/391 [==============================] - 59s - loss: 7.5002e-04 - acc: 0.9997 - val_loss: 0.5568 - val_acc: 0.9173
Epoch 165/170
391/391 [==============================] - 59s - loss: 4.6109e-04 - acc: 0.9998 - val_loss: 0.5564 - val_acc: 0.9168
Epoch 166/170
391/391 [==============================] - 59s - loss: 4.0144e-04 - acc: 1.0000 - val_loss: 0.5569 - val_acc: 0.9168
Epoch 167/170
391/391 [==============================] - 59s - loss: 5.7081e-04 - acc: 0.9999 - val_loss: 0.5580 - val_acc: 0.9170
Epoch 168/170
391/391 [==============================] - 59s - loss: 7.8736e-04 - acc: 0.9997 - val_loss: 0.5579 - val_acc: 0.9168
Epoch 169/170
391/391 [==============================] - 59s - loss: 7.2820e-04 - acc: 0.9998 - val_loss: 0.5579 - val_acc: 0.9161
Epoch 170/170
391/391 [==============================] - 59s - loss: 6.4590e-04 - acc: 0.9999 - val_loss: 0.5590 - val_acc: 0.9167

In [ ]: