keras mnist cnn example

# encoding:utf-8from __future__ import print_functionimport kerasfrom keras.datasets import mnistfrom keras.models import Sequentialfrom keras.layers import Dense, Dropout, Flattenfrom keras.layers import Conv2D, MaxPooling2Dfrom keras import backend as K# batch_size 太小会导致训练慢，过拟合等问题，太大会导致欠拟合。所以要适当选择batch_size = 128# 0-9手写数字一个有10个类别num_classes = 10# 12次完整迭代，差不多够了epochs = 2# input image dimensions# 输入的图片是28*28像素的灰度图img_rows, img_cols = 28, 28# the data, shuffled and split between train and test sets(x_train, y_train), (x_test, y_test) = mnist.load_data()# keras输入数据有两种格式，一种是通道数放在前面，一种是通道数放在后面# 在如何表示一组彩色图片的问题上，Theano和TensorFlow发生了分歧，'th'模式，也即Theano模式会把100张RGB三通道的16×32（高为16宽为32）彩色图表示为下面这种形式（100,3,16,32），Caffe采取的也是这种方式。# 第0个维度是样本维，代表样本的数目，第1个维度是通道维，代表颜色通道数。后面两个就是高和宽了。这种theano风格的数据组织方法，称为“channels_first”，即通道维靠前。# 而TensorFlow，的表达形式是（100,16,32,3），即把通道维放在了最后，这种数据组织方式称为“channels_last”。# Keras默认的数据组织形式在~/.keras/keras.json中规定，可查看该文件的image_data_format一项查看，也可在代码中通过K.image_data_format()函数返回，请在网络的训练和测试中保持维度顺序一致。if K.image_data_format() == 'channels_first':    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)    input_shape = (1, img_rows, img_cols)else:    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)    input_shape = (img_rows, img_cols, 1)x_train = x_train.astype('float32')x_test = x_test.astype('float32')x_train /= 255x_test /= 255print('x_train shape:', x_train.shape)print(x_train.shape[0], 'train samples')print(x_test.shape[0], 'test samples')# convert class vectors to binary class matrices # 把类别0-9变成2进制，方便训练y_train = keras.utils.to_categorical(y_train, num_classes)y_test = keras.utils.to_categorical(y_test, num_classes)# 牛逼的Sequential类可以让我们灵活地插入不同的神经网络层model = Sequential()# 加上一个2D卷积层， 32个输出（也就是卷积通道），激活函数选用relu，# 卷积核的窗口选用3*3像素窗口model.add(Conv2D(32, kernel_size=(3, 3),                 activation='relu',                 input_shape=input_shape))# 64个通道的卷积层model.add(Conv2D(64, (3, 3), activation='relu'))# 池化层是2*2像素的model.add(MaxPooling2D(pool_size=(2, 2)))# 对于池化层的输出，采用0.25概率的Dropoutmodel.add(Dropout(0.25))# 展平所有像素，比如[28*28] -> [784]model.add(Flatten())# 对所有像素使用全连接层，输出为128，激活函数选用relumodel.add(Dense(128, activation='relu'))model.add(Dropout(0.5))# 对刚才Dropout的输出采用softmax激活函数，得到最后结果0-9model.add(Dense(num_classes, activation='softmax'))# 模型我们使用交叉熵损失函数，最优化方法选用Adadeltamodel.compile(loss=keras.losses.categorical_crossentropy,              optimizer=keras.optimizers.Adadelta(),              metrics=['accuracy'])model.fit(x_train, y_train,          batch_size=batch_size,          epochs=epochs,          verbose=1,          validation_data=(x_test, y_test))score = model.evaluate(x_test, y_test, verbose=0)print('Test loss:', score[0])print('Test accuracy:', score[1])