In [1]:
import pickle
pickle_file = '-images_6.pickle'
with open(pickle_file, 'rb') as f:
save = pickle.load(f)
X = save['X']
y = save['y']
del save # hint to help gc free up memory
print('Training set', X.shape, y.shape)
In [2]:
# Add additional imports
import numpy as np
np.random.seed(1337) # for reproducibility
from keras.datasets import mnist
from keras.models import Sequential
from keras.layers import Dense, Dropout, Activation, Flatten
from keras.layers import Convolution2D, MaxPooling2D
from keras.utils import np_utils
from keras import backend as K
from keras.callbacks import ModelCheckpoint
from keras.datasets import mnist
print "using ordering:", K.image_dim_ordering()
In [3]:
# number of classes
num_classes = np.max(y) + 1 # number of categories
print "number of classes:", num_classes
# image dimensions
img_rows, img_cols = X.shape[1], X.shape[2]
if K.image_dim_ordering() == 'th':
X = X.reshape(X.shape[0], 1, img_rows, img_cols)
input_shape = (1, img_rows, img_cols)
else:
X = X.reshape(X.shape[0], img_rows, img_cols, 1)
input_shape = (img_rows, img_cols, 1)
y_ohe = np_utils.to_categorical(y, num_classes)
print X.shape
print y_ohe.shape
In [4]:
# Plot to make sure we have imported correctly..
%matplotlib inline
from matplotlib.pyplot import imshow
import matplotlib.pyplot as plt
img_num = 0
if K.image_dim_ordering() == 'th':
img = X[img_num][0,:,:]
else:
img = X[img_num][:,:,0]
print img.shape
print y_ohe[img_num]
imshow(img, cmap = plt.get_cmap('gray'), vmin = 0, vmax = 1, interpolation='nearest')
Out[4]:
In [5]:
## Construct CNN
# model hyperparameters
batch_size = 32
nb_epoch = 50
# network architecture
patch_size_1 = 3
patch_size_2 = 3
patch_size_3 = 3
patch_size_4 = 3
patch_size_5 = 3
depth_1 = 64
depth_2 = 128
depth_3 = 128
depth_4 = 256
depth_5 = 256
pool_size = 2
num_hidden_1 = 512
num_hidden_2 = 1024
dropout = 0.25
##---------------------------
# create new Keras Sequential model
model = Sequential()
model.add(Convolution2D(depth_1, patch_size_1, patch_size_1,
border_mode='same',
input_shape=input_shape))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(pool_size, pool_size)))
model.add(Convolution2D(depth_2, patch_size_2, patch_size_2,
border_mode='same'))
model.add(Activation('relu'))
model.add(Convolution2D(depth_3, patch_size_3, patch_size_3,
border_mode='same'))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(pool_size, pool_size)))
model.add(Convolution2D(depth_4, patch_size_4, patch_size_4,
border_mode='same'))
model.add(Activation('relu'))
model.add(Convolution2D(depth_5, patch_size_5, patch_size_5,
border_mode='same'))
model.add(Activation('relu'))
model.add(Flatten())
model.add(Dense(num_hidden_1))
model.add(Activation('relu'))
model.add(Dropout(dropout))
model.add(Dense(num_hidden_2))
model.add(Activation('relu'))
model.add(Dropout(dropout))
model.add(Dense(num_classes))
model.add(Activation('softmax'))
print "Model Created"
##-------------------------------------------
# compile model
checkpoint_name = "-model.hdf5"
checkpointer = ModelCheckpoint(checkpoint_name, verbose=0, save_best_only=True)
model.compile(loss='categorical_crossentropy', optimizer='adadelta', metrics=['accuracy'])
print "Compiled.."
In [6]:
# Train Model
history = model.fit(X, y_ohe, validation_split=0.25, batch_size=batch_size, nb_epoch=nb_epoch,
verbose=1, callbacks=[checkpointer])