TensorFlow官方文档学习 Keras版MNIST Get Started with TensorFlow

import tensorflow as tf
mnist = tf.keras.datasets.mnist  #下载mnist图像的数据

(x_train, y_train),(x_test, y_test) = mnist.load_data() #划分训练集和测试集
x_train, x_test = x_train / 255.0, x_test / 255.0   #归一化处理[0,1]

#序贯（Sequential）模型
model = tf.keras.Sequential([                        
  tf.keras.layers.Flatten(input_shape=[28,28]),          #展平图像数据
  tf.keras.layers.Dense(512, activation=tf.nn.relu),     #全连接层512列
  tf.keras.layers.Dropout(0.2),                          #需要断开的神经元的比例
  tf.keras.layers.Dense(10, activation=tf.nn.softmax)    #全连接层10列
])
model.compile(optimizer='adam',                          #优化器
              loss='sparse_categorical_crossentropy',    #损失函数交叉熵
              metrics=['accuracy'])                      #评价指标列表

model.fit(x_train, y_train, epochs=5,batch_size=50)      #fit开始训练迭代5轮
model.evaluate(x_test, y_test)                           #对测试集进行测试

运行得到输出结果：

55900/60000 [==========================>...]55900/60000 [==========================>...] - ETA: 0s - loss: 0.0460 - acc: 0.9857
56400/60000 [===========================>..]56400/60000 [===========================>..] - ETA: 0s - loss: 0.0461 - acc: 0.9857
56900/60000 [===========================>..]56900/60000 [===========================>..] - ETA: 0s - loss: 0.0460 - acc: 0.9856
57400/60000 [===========================>..]57400/60000 [===========================>..] - ETA: 0s - loss: 0.0463 - acc: 0.9856
57900/60000 [===========================>..]57900/60000 [===========================>..] - ETA: 0s - loss: 0.0464 - acc: 0.9856
58400/60000 [============================>.]58400/60000 [============================>.] - ETA: 0s - loss: 0.0466 - acc: 0.9855
58900/60000 [============================>.]58900/60000 [============================>.] - ETA: 0s - loss: 0.0465 - acc: 0.9855
59400/60000 [============================>.]59400/60000 [============================>.] - ETA: 0s - loss: 0.0468 - acc: 0.9854
59900/60000 [============================>.]59900/60000 [============================>.] - ETA: 0s - loss: 0.0469 - acc: 0.9853
60000/60000 [==============================]60000/60000 [==============================] - 7s 116us/step - loss: 0.0469 - acc: 0.9853


   32/10000 [..............................]   32/10000 [..............................] - ETA: 8s
  928/10000 [=>............................]  928/10000 [=>............................] - ETA: 0s
 1856/10000 [====>.........................] 1856/10000 [====>.........................] - ETA: 0s
 2720/10000 [=======>......................] 2720/10000 [=======>......................] - ETA: 0s
 3584/10000 [=========>....................] 3584/10000 [=========>....................] - ETA: 0s
 4544/10000 [============>.................] 4544/10000 [============>.................] - ETA: 0s
 5376/10000 [===============>..............] 5376/10000 [===============>..............] - ETA: 0s
 6272/10000 [=================>............] 6272/10000 [=================>............] - ETA: 0s
 7136/10000 [====================>.........] 7136/10000 [====================>.........] - ETA: 0s
 8096/10000 [=======================>......] 8096/10000 [=======================>......] - ETA: 0s
 9056/10000 [==========================>...] 9056/10000 [==========================>...] - ETA: 0s
 9888/10000 [============================>.] 9888/10000 [============================>.] - ETA: 0s
10000/10000 [==============================]10000/10000 [==============================] - 1s 60us/step

[loss,accuracy]
[0.06254660903802142, 0.9796]