CNN using VGG19 architecture

vgg19.py

### TODO: Define your architecture.
from keras.layers import Conv2D, MaxPooling2D, GlobalAveragePooling2D
from keras.layers import Dropout, Flatten, Dense
from keras.models import Sequential

vgg19_model = Sequential()
vgg19_model.add(GlobalAveragePooling2D(input_shape=train_vgg19.shape[1:]))
vgg19_model.add(Dense(512, activation='relu'))
vgg19_model.add(Dropout(0.5))
vgg19_model.add(Dense(512, activation='relu'))
vgg19_model.add(Dense(133, activation='softmax'))
vgg19_model.summary()

### TODO: Train the model.
checkpointer = ModelCheckpoint(filepath='saved_models/weights.best.VGG19.hdf5', 
                               verbose=1, save_best_only=True)

vgg19_model.fit(train_vgg19, train_targets, 
          validation_data=(valid_vgg19, valid_targets),validation_split=0.2,
          epochs=20, batch_size=120, callbacks=[checkpointer], verbose=1)