第一步：下載數(shù)據(jù)集

from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("MNIST_data",one_hot=True)

第二步：創(chuàng)建輸入，輸出，權(quán)重和偏差，并定義softmax回歸算法

x = tf.placeholder(tf.float32,[None,784])
y_ = tf.placeholder("float",[None,10])
W = tf.Variable(tf.zeros([784,10]))
b = tf.Variable(tf.zeros([10]))
y = tf.nn.softmax(tf.matmul(x,W)+b)

第三步：定義損失函數(shù)和梯度下降算法

cross_entropy = tf.reduce_mean(-tf.reduce_sum(y_*tf.log(y),reduction_indices=[1]))
train_step = tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)

第四步：開啟對話，進(jìn)行訓(xùn)練

#在session中啟動模型
sess = tf.Session()
sess.run(init)
#開始訓(xùn)練
for i in range(1000):
    batch_xs,batch_ys = mnist.train.next_batch(100)
    sess.run(train_step,feed_dict={x:batch_xs,y_:batch_ys})
    if i % 100 == 0:
        print(i)

第五步：驗證

correct_prediction = tf.equal(tf.argmax(y,1),tf.argmax(y_,1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32))
print(sess.run(accuracy,feed_dict={x:mnist.test.images,y_:mnist.test.labels}))

完整代碼

import tensorflow as tf
#下載數(shù)據(jù)集
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("MNIST_data",one_hot=True)
#定義輸入輸出
x = tf.placeholder(tf.float32,[None,784])
y_ = tf.placeholder("float",[None,10])
W = tf.Variable(tf.zeros([784,10]))
b = tf.Variable(tf.zeros([10]))
y = tf.nn.softmax(tf.matmul(x,W)+b)
#定義損失函數(shù)
cross_entropy = tf.reduce_mean(-tf.reduce_sum(y_*tf.log(y),reduction_indices=[1]))
#梯度下降
train_step = tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)
#初始化所有變量
init = tf.initialize_all_variables()
#在session中啟動模型
sess = tf.Session()
sess.run(init)
#開始訓(xùn)練
for i in range(1000):
    batch_xs,batch_ys = mnist.train.next_batch(100)
    sess.run(train_step,feed_dict={x:batch_xs,y_:batch_ys})
    if i % 100 == 0:
        print(i)
#驗證
correct_prediction = tf.equal(tf.argmax(y,1),tf.argmax(y_,1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32))
print(sess.run(accuracy,feed_dict={x:mnist.test.images,y_:mnist.test.labels}))