machine learning 界的hello world-使用高级API Keras

import tensorflow as tf
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

model = tf.keras.models.Sequential([
  tf.keras.layers.Flatten(),
  tf.keras.layers.Dense(512, activation=tf.nn.relu),
  tf.keras.layers.Dropout(0.2),
  tf.keras.layers.Dense(10, activation=tf.nn.softmax)
])

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

model.fit(x_train,y_train,epochs=5)
model.evaluate(x_test,y_test)