#基于mnist数据集的手写数字识别
#构造了三层全连接层组成的多层感知机,最后一层为输出层
#基于Keras 2.1.1 Tensorflow 1.4.0
代码:
1 import keras 2 from keras.datasets import mnist 3 from keras.models import Sequential 4 from keras.layers import Dense,Dropout 5 from keras.optimizers import RMSprop 6 7 (x_train,y_train),(x_test,y_test) = mnist.load_data() 8 #载入数据,第一次运行时会从外部网络下载数据集到对应目录下 9 print(x_train.shape,y_train.shape) 10 print(x_test.shape,y_test.shape) 11 12 # import matplotlib.pyplot as plt 13 # im = plt.imshow(x_train[0],cmap='gray') 14 # plt.show() 15 # im2 = plt.imshow(x_train[1],cmap='gray') 16 # plt.show() 17 x_train = x_train.reshape(60000,784) 18 x_test = x_test.reshape(10000,784) 19 x_train = x_train.astype('float32') 20 x_train = x_train.astype('float32') 21 print(x_train.shape) 22 x_train = x_train/255 23 x_test = x_test/255 24 y_train = keras.utils.to_categorical(y_train,10) 25 y_test = keras.utils.to_categorical(y_test,10) 26 27 model = Sequential() 28 model.add(Dense(512,activation='relu',input_shape=(784,))) 29 model.add(Dropout(0.2)) 30 model.add(Dense(512,activation='relu')) 31 model.add(Dropout(0.2)) 32 model.add(Dense(10,activation='softmax')) 33 model.summary() 34 model.compile(loss='categorical_crossentropy',optimizer=RMSprop(),metrics=['accuracy']) 35 model.fit(x_train,y_train,batch_size=64,epochs=2,verbose=1,validation_data=(x_test,y_test)) 36 score = model.evaluate(x_test,y_test,verbose=1) 37 print('Test loss:',score[0]) 38 print('Test accuracy',score[1])
结果:
Test loss: 0.123420921481 Test accuracy 0.9682
本站文章如无特殊说明,均为本站原创,如若转载,请注明出处:Keras mlp 手写数字识别示例 - Python技术站