MNIST的学习资料

mnist_example.py

import tensorflow as tf
import tensorflow.examples.tutorials.mnist.input_data as input_data

minist = input_data.read_data_sets("MNIST_data/", one_hot=True)

#None代表任意个数
#输入是个 batch数＊像素数 的矩阵
x = placeholder(tf.float32, [None, 784])

#注意W，b的维度
W = tf.Variable(tf.zeros([784,10]))
b = tf.Variable(tf.zeros([10]))

#y是 batch数＊label数 的矩阵
y = tf.nn.softmax(tf.matmul(x,W) + b)

y_ = tf.placeholder("float", [None, 10])

cross_entropy =	-tf.reduce_sum(y_*tf.log(y))

train_step = tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy)

#初始化变量
init = tf.initialize_all_variables()

sess = tf.Session()
sess.run(init)

#每次训练随机选择100个图像，训练1000次
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})

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