project1：手写数字识别

import tensorflow as tf
import matplotlib.pyplot as plt
mnist=tf.keras.datasets.mnist
(x_train,y_train),(x_test,y_test)=mnist.load_data()
x_train,x_test=x_train/255.0,x_test/255.0#归一化，把输入的值变小更适合神经网络吸收
plt.figure()#看看训练集，没什么毛病
for i in range(10):
    plt.subplot(1,10,i+1)
    plt.imshow(x_train[i])
    print(y_train[i])
model=tf.keras.models.Sequential([
    tf.keras.layers.Flatten(),#把输入特征拉直为一维数组
    tf.keras.layers.Dense(50,activation='relu'),
    tf.keras.layers.Dense(100,activation='softmax')
])#sequential是连续的意思，也就是可以一层一层堆叠神经网络，事实上，sequential可以表示全连接神经网络，卷积神经网络，循环神经网络等
model.compile(optimizer='adam',
             loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),#输出满足概率分布，因此这里的from_logits是False
             metrics=['sparse_categorical_accuracy'])#衡量指标
model.fit(x_train,y_train,batch_size=32,epochs=50,validation_data=(x_test,y_test),validation_freq=1)#validation_freq是测试的频率
model.summary()#打印出网络结构和参数统计    

import tensorflow as tf
from tensorflow.keras.layers import Dense,Flatten
from tensorflow.keras import Model
mnist=tf.keras.datasets.mnist
(x_train,y_train),(x_test,y_test)=mnist.load_data()
x_train,x_test=x_train/255.0,x_test/255.0#跟前面没什么区别，只是把layer里的Dense，Flatten单独提了出来
class MnistModel(Model):
    def __init__(self):
        super(MnistModel,self).__init__()#重新定义__init__()会覆盖父类的属性，但是通过super函数可以同时继承父类的属性，并且也会有重新添加的属性
        self.flatten=Flatten()
        self.d1=Dense(128,activation='relu')
        self.d2=Dense(10,activation='softmax')
    def call(self,x):#前面说过sequential相当于走了一遍前向传播，与这里的call函数相似
        x=self.flatten(x)
        x=self.d1(x)
        y=self.d2(x)
        return y
model=MnistModel()
model.compile(optimizer='adam',
             loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
             metrics=['sparse_categorical_accuracy'])
model.fit(x_train,y_train,batch_size=32,epochs=5,validation_data=(x_test,y_test),validation_freq=1)
model.summary()