7. 使用cifar10数据集利用keras进行训练

# cifar10数据集共有60000张32*32的彩色图像，分为10个类，每类6000张图.50000张用于训练,10000张用于测试
# TODO: Build a model
from keras.models import Sequential
from keras.utils.vis_utils import plot_model
from keras.layers.core import Dense, Activation, Flatten, Dropout
from keras.layers.convolutional import Convolution2D
from keras.layers.pooling import MaxPooling2D
# TODO: 独热编码标签One Hot encode the labels to the variable y_one_hot
from sklearn.preprocessing import LabelBinarizer
from keras.datasets import cifar10
(X_train, y_train), (X_test, y_test) = cifar10.load_data()
# print(np.min(X_train), np.max(X_train))  result:0,255
# LabelBinarizer（）标签二值化函数
label_binarizer = LabelBinarizer()
# 将y_train转换为标签二值化（one-hot）
y_train_one_hot = label_binarizer.fit_transform(y_train)
y_test_one_hot = label_binarizer.fit_transform(y_test)
# print(y_train_one_hot[:10])


# TODO: 标准化数据，Normalize the data features to the variable X_normalized
def normalize_grayscale(image_data):
    a = -0.5
    b = 0.5
    grayscale_min = 0
    grayscale_max = 255
    # for line in range(image_data.shape[0]):
    #     return a + (((image_data[line] - grayscale_min) * (b - a)) / (grayscale_max - grayscale_min))
    return a + (((image_data - grayscale_min)*(b - a))/(grayscale_max - grayscale_min))


X_train_norm = normalize_grayscale(X_train)
# 测试 print(X_train_norm[:10])
X_test_norm = normalize_grayscale(X_test)

model = Sequential()
# shape[30,30,32]
model.add(Convolution2D(32, 3, 3, input_shape=(32, 32, 3)))
# [15,15,32]
model.add(MaxPooling2D((2, 2)))
model.add(Dropout(0.5))
model.add(Activation('relu'))
# [13,13,32]
model.add(Convolution2D(32, 3, 3))
# [6,6,32]
model.add(MaxPooling2D((2, 2)))
model.add(Dropout(0.5))
model.add(Activation('relu'))
# 1152
model.add(Flatten())
# 128
model.add(Dense(128))
model.add(Activation('relu'))
# 10
model.add(Dense(10))
model.add(Activation('softmax'))

# 通过pydot画出模型图
plot_model(model, to_file='model1.png', show_shapes=True)
# TODO: Compile and train the model
model.compile('adam', 'categorical_crossentropy', ['accuracy'])
history = model.fit(X_train_norm, y_train_one_hot, nb_epoch=10, validation_split=0.2)

# TODO: 测试机评估模型Evaluate model on test data
metrics = model.evaluate(X_test_norm, y_test_one_hot)
# model.metrics_names为标量输出的显示标签：loss、acc
for metric_i in range(len(model.metrics_names)):
    metric_name = model.metrics_names[metric_i]
    metric_value = metrics[metric_i]
    print('{}: {}'.format(metric_name, metric_value))