keras 手写数字识别

简单的手写识别模型（具有一个隐藏层的）

import keras
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
import keras.datasets.mnist as mnist


(train_image,train_label),(test_image,test_label) =mnist.load_data()
train_image.shape

训练数据集的shape结果：
(60000, 28, 28)

plt.imshow(train_image[1000])

可视化训练数据集中的第1000图片

建立模型

model =keras.Sequential()
model.add(layers.Flatten())  #(6000,28,28)=>(6000,28*28)
model.add(layers.Dense(64,activation='relu'))
model.add(layers.Dense(10,activation='softmax'))

model.compile(optimizer='adam',loss='sparse_categorical_crossentropy',metrics=['acc'])

model.fit(train_image,train_label,epochs=50,batch_size=512)#每一次输入512

训练过程;

模型评价;

model.evaluate(test_image,test_label)

结果：
10000/10000 [==============================] - 0s 42us/step
[3.4554689140319823, 0.784]
测试：

np.argmax(model.predict(test_image[:10]),axis=1)

结果：
array([7, 2, 1, 0, 4, 1, 4, 4, 8, 7], dtype=int64)
实际的结果:

test_label[:10]

array([7, 2, 1, 0, 4, 1, 4, 9, 5, 9], dtype=uint8)

优化模型

模型:

model =keras.Sequential()
model.add(layers.Flatten())  #(6000,28,28)=>(6000,28*28)
model.add(layers.Dense(64,activation='relu'))
model.add(layers.Dense(64,activation='relu'))
model.add(layers.Dense(64,activation='relu'))
model.add(layers.Dense(10,activation='softmax'))
model.compile(optimizer='adam',loss='sparse_categorical_crossentropy',metrics=['acc'])
model.fit(train_image,train_label,epochs=50,batch_size=512,validation_data=(test_image,test_label))#每一次输入512

结果：