自学教程：Keras在mnist上的CNN实践,并且自定义loss函数曲线图操作

from keras.datasets import mnistfrom keras.models import Sequentialfrom keras.layers import Dense, Dropout, Activation, Flattenfrom keras.layers import Convolution2D, MaxPooling2Dfrom keras.utils import np_utilsfrom keras import backend as K class LossHistory(keras.callbacks.Callback):    def on_train_begin(self, logs={}):        self.losses = {'batch':[], 'epoch':[]}        self.accuracy = {'batch':[], 'epoch':[]}        self.val_loss = {'batch':[], 'epoch':[]}        self.val_acc = {'batch':[], 'epoch':[]}     def on_batch_end(self, batch, logs={}):        self.losses['batch'].append(logs.get('loss'))        self.accuracy['batch'].append(logs.get('acc'))        self.val_loss['batch'].append(logs.get('val_loss'))        self.val_acc['batch'].append(logs.get('val_acc'))     def on_epoch_end(self, batch, logs={}):        self.losses['epoch'].append(logs.get('loss'))        self.accuracy['epoch'].append(logs.get('acc'))        self.val_loss['epoch'].append(logs.get('val_loss'))        self.val_acc['epoch'].append(logs.get('val_acc'))     def loss_plot(self, loss_type):        iters = range(len(self.losses[loss_type]))        plt.figure()        # acc        plt.plot(iters, self.accuracy[loss_type], 'r', label='train acc')        # loss        plt.plot(iters, self.losses[loss_type], 'g', label='train loss')        if loss_type == 'epoch':            # val_acc            plt.plot(iters, self.val_acc[loss_type], 'b', label='val acc')            # val_loss            plt.plot(iters, self.val_loss[loss_type], 'k', label='val loss')        plt.grid(True)        plt.xlabel(loss_type)        plt.ylabel('acc-loss')        plt.legend(loc="upper right")        plt.show() history = LossHistory() batch_size = 128nb_classes = 10nb_epoch = 20img_rows, img_cols = 28, 28nb_filters = 32pool_size = (2,2)kernel_size = (3,3)(X_train, y_train), (X_test, y_test) = mnist.load_data()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') Y_train = np_utils.to_categorical(y_train, nb_classes)Y_test = np_utils.to_categorical(y_test, nb_classes) model3 = Sequential() model3.add(Convolution2D(nb_filters, kernel_size[0] ,kernel_size[1],                        border_mode='valid',                        input_shape=input_shape))model3.add(Activation('relu')) model3.add(Convolution2D(nb_filters, kernel_size[0], kernel_size[1]))model3.add(Activation('relu')) model3.add(MaxPooling2D(pool_size=pool_size))model3.add(Dropout(0.25)) model3.add(Flatten()) model3.add(Dense(128))model3.add(Activation('relu'))model3.add(Dropout(0.5)) model3.add(Dense(nb_classes))model3.add(Activation('softmax')) model3.summary() model3.compile(loss='categorical_crossentropy',              optimizer='adadelta',              metrics=['accuracy']) model3.fit(X_train, Y_train, batch_size=batch_size, epochs=nb_epoch,          verbose=1, validation_data=(X_test, Y_test),callbacks=[history]) score = model3.evaluate(X_test, Y_test, verbose=0)print('Test score:', score[0])print('Test accuracy:', score[1]) #acc-losshistory.loss_plot('epoch')

from keras.layers import Dense,Flatten,Dropoutfrom keras.datasets import mnistfrom keras import Sequentialimport matplotlib.pyplot as pltimport numpy as np

# 训练集       训练集标签       测试集      测试集标签(train_image,train_label),(test_image,test_label) = mnist.load_data()print('shape:',train_image.shape)   #查看训练集的shapeplt.imshow(train_image[0])    #查看第一张图片print('label:',train_label[0])      #查看第一张图片对应的标签plt.show()

train_image = train_image.astype('float32')test_image = test_image.astype('float32')train_image /= 255.0test_image /= 255.0

#初始化模型（模型的优化 ---> 增大网络容量，直到过拟合）model = Sequential()model.add(Flatten(input_shape=(28,28)))    #将二维扁平化为一维（60000,28,28）---> (60000,28*28)输入28*28个神经元model.add(Dropout(0.1))model.add(Dense(1024,activation='relu'))   #全连接层 输出64个神经元 ,kernel_regularizer=l2(0.0003)model.add(Dropout(0.1))model.add(Dense(512,activation='relu'))    #全连接层model.add(Dropout(0.1))model.add(Dense(256,activation='relu'))    #全连接层model.add(Dropout(0.1))model.add(Dense(10,activation='softmax'))  #输出层，10个类别，用softmax分类

#编译模型model.compile(    optimizer='adam',      #优化器使用默认adam    loss='sparse_categorical_crossentropy', #损失函数使用sparse_categorical_crossentropy    metrics=['acc']       #评价指标)

#训练模型history = model.fit(    x=train_image,                          #训练的图片    y=train_label,                          #训练的标签    epochs=10,                              #迭代10次    batch_size=512,                         #划分批次    validation_data=(test_image,test_label) #验证集)

#绘制loss acc图plt.figure()plt.plot(history.history['acc'],label='training acc')plt.plot(history.history['val_acc'],label='val acc')plt.title('model acc')plt.ylabel('acc')plt.xlabel('epoch')plt.legend(loc='lower right')plt.figure()plt.plot(history.history['loss'],label='training loss')plt.plot(history.history['val_loss'],label='val loss')plt.title('model loss')plt.ylabel('loss')plt.xlabel('epoch')plt.legend(loc='upper right')plt.show()

print("前十个图片对应的标签：　",test_label[:10]) #前十个图片对应的标签print("取前十张图片测试集预测：",np.argmax(model.predict(test_image[:10]),axis=1)) #取前十张图片测试集预测

model.save('./mnist_model.h5')

from keras.layers import Dense,Flatten,Dropoutfrom keras.datasets import mnistfrom keras import Sequentialimport matplotlib.pyplot as pltimport numpy as np# 训练集       训练集标签       测试集      测试集标签(train_image,train_label),(test_image,test_label) = mnist.load_data()# print('shape:',train_image.shape)   #查看训练集的shape# plt.imshow(train_image[0]) #查看第一张图片# print('label:',train_label[0])      #查看第一张图片对应的标签# plt.show()#归一化（收敛）train_image = train_image.astype('float32')test_image = test_image.astype('float32')train_image /= 255.0test_image /= 255.0#初始化模型（模型的优化 ---> 增大网络容量，直到过拟合）model = Sequential()model.add(Flatten(input_shape=(28,28)))   #将二维扁平化为一维（60000,28,28）---> (60000,28*28)输入28*28个神经元model.add(Dropout(0.1))model.add(Dense(1024,activation='relu'))    #全连接层 输出64个神经元 ,kernel_regularizer=l2(0.0003)model.add(Dropout(0.1))model.add(Dense(512,activation='relu'))    #全连接层model.add(Dropout(0.1))model.add(Dense(256,activation='relu'))    #全连接层model.add(Dropout(0.1))model.add(Dense(10,activation='softmax')) #输出层，10个类别，用softmax分类#编译模型model.compile(    optimizer='adam',    loss='sparse_categorical_crossentropy',    metrics=['acc'])#训练模型history = model.fit(    x=train_image,                          #训练的图片    y=train_label,                          #训练的标签    epochs=10,                              #迭代10次    batch_size=512,                         #划分批次    validation_data=(test_image,test_label) #验证集)#绘制loss acc 图plt.figure()plt.plot(history.history['acc'],label='training acc')plt.plot(history.history['val_acc'],label='val acc')plt.title('model acc')plt.ylabel('acc')plt.xlabel('epoch')plt.legend(loc='lower right')plt.figure()plt.plot(history.history['loss'],label='training loss')plt.plot(history.history['val_loss'],label='val loss')plt.title('model loss')plt.ylabel('loss')plt.xlabel('epoch')plt.legend(loc='upper right')plt.show()print("前十个图片对应的标签：　",test_label[:10]) #前十个图片对应的标签print("取前十张图片测试集预测：",np.argmax(model.predict(test_image[:10]),axis=1)) #取前十张图片测试集预测#优化前（一个全连接层（隐藏层））#- 1s 12us/step - loss: 1.8765 - acc: 0.8825# [7 2 1 0 4 1 4 3 5 4]# [7 2 1 0 4 1 4 9 5 9]#优化后（三个全连接层（隐藏层））#- 1s 14us/step - loss: 0.0320 - acc: 0.9926 - val_loss: 0.2530 - val_acc: 0.9655# [7 2 1 0 4 1 4 9 5 9]# [7 2 1 0 4 1 4 9 5 9]model.save('./model_nameALL.h5')