mnist 全连接单隐层

'''
sess
input,hidden,output,w,b
hidden:relu,dropout
relu(wx+b)
output:softmax(wx+b)
label   y_
cross-entropy
train
initial
accuracy
test
save
restore
'''

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


file = './MNIST/MNIST_data/'
mnist = input_data.read_data_sets(file,one_hot=True)

sess = tf.InteractiveSession()


input_units = 784
hidden_units = 300
x = tf.placeholder(dtype=tf.float32,shape=[None,input_units],name='x')
# W1 = tf.Variable(tf.zeros(shape=[input_units,hidden_units]))
W1 = tf.Variable(tf.truncated_normal(shape=[input_units,hidden_units],stddev=0.1))

b1 = tf.Variable(tf.zeros(shape=[hidden_units]))

W2 = tf.Variable(tf.zeros(shape=[hidden_units,10]))

b2 = tf.Variable(tf.zeros(shape=[10]))

keep_prob = tf.placeholder(tf.float32)


hidden1 = tf.nn.relu(tf.matmul(x, W1) + b1)

hidden1_drop = tf.nn.dropout(hidden1,keep_prob=keep_prob)



y = tf.nn.softmax(tf.matmul(hidden1_drop,W2) + b2)


y_ = tf.placeholder(dtype=tf.float32,shape=[None,10],name='y_')


cross_entropy = tf.reduce_mean(-tf.reduce_sum(input_tensor=y_*tf.log(y),reduction_indices=[1]))

# train_step = tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)#0.977
# train_step = tf.train.AdamOptimizer(0.001).minimize(cross_entropy)#0.975
train_step = tf.train.AdagradOptimizer(0.3).minimize(cross_entropy)#0.98

saver = tf.train.Saver()


tf.global_variables_initializer().run()

for i in range(3000):

batch_xs, batch_ys = mnist.train.next_batch(100)

train_step.run(feed_dict={x:batch_xs, y_:batch_ys, keep_prob:0.75})



correct_prediction = tf.equal(tf.argmax(input=y, axis=1),tf.argmax(input=y_, axis=1))


accuracy = tf.reduce_mean(tf.cast(x=correct_prediction,dtype=tf.float32))

save_path = saver.save(sess=sess, save_path='./model_mnist.ckpt')


test_accuracy = accuracy.eval({x:mnist.test.images, y_:mnist.test.labels, keep_prob:1.0})
print(test_accuracy)


saver.restore(sess,'./model_mnist.ckpt')

result = accuracy.eval({x:mnist.test.images, y_:mnist.test.labels, keep_prob:1.0})

print(result)