Deep MNIST for Experts - 전체 코드 (2)

텐서플로우 첫걸음에 나오는 전체 코드입니다. 구글 홈페이지 튜토리얼과는 다소 다르네요.

In [1]:

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

Extracting MNIST_data/train-images-idx3-ubyte.gz
Extracting MNIST_data/train-labels-idx1-ubyte.gz
Extracting MNIST_data/t10k-images-idx3-ubyte.gz
Extracting MNIST_data/t10k-labels-idx1-ubyte.gz

In [2]:

import tensorflow as tf

In [3]:

x = tf.placeholder("float", [None, 784])

In [4]:

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

In [5]:

x_image = tf.reshape(x, [-1,28,28,1])
print "x_image="
print x_image

x_image=
Tensor("Reshape:0", shape=(?, 28, 28, 1), dtype=float32)

In [6]:

def weight_variable(shape):
  initial = tf.truncated_normal(shape, stddev=0.1)
  return tf.Variable(initial)
 
def bias_variable(shape):
  initial = tf.constant(0.1, shape=shape)
  return tf.Variable(initial)

In [7]:

def conv2d(x, W):

    return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')
 
def max_pool_2x2(x):

    return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],
                        strides=[1, 2, 2, 1], padding='SAME')

In [8]:

W_conv1 = weight_variable([5, 5, 1, 32])
b_conv1 = bias_variable([32])

In [9]:

h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)
h_pool1 = max_pool_2x2(h_conv1)

In [10]:

W_conv2 = weight_variable([5, 5, 32, 64])
b_conv2 = bias_variable([64])

In [11]:

h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
h_pool2 = max_pool_2x2(h_conv2)

In [12]:

W_fc1 = weight_variable([7 * 7 * 64, 1024])
b_fc1 = bias_variable([1024])

In [13]:

h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)

In [14]:

keep_prob = tf.placeholder("float")
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

In [15]:

W_fc2 = weight_variable([1024, 10])
b_fc2 = bias_variable([10])

In [16]:

y_conv = tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)

In [17]:

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

In [18]:

train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)

In [19]:

correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1))

In [20]:

accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))

In [21]:

sess = tf.Session()

In [22]:

sess.run(tf.initialize_all_variables())

In [ ]:

for i in range(300):
    batch = mnist.train.next_batch(100)
    if i%100 == 0:
        train_accuracy = sess.run(accuracy, feed_dict={x:batch[0], y_: batch[1], keep_prob: 1.0})
        print ("step %d, training accuracy %g" % (i, train_accuracy))
    sess.run(train_step, feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5})

step 0, training accuracy 0.09

In [ ]:

print ("test accuracy %g" % sess.run(accuracy, feed_dict={x: mnist.test.images, y_: mnist.test.labels, keep_prob: 1.0}))