tensorflowチュートリアルプログラム(cifar10)が動きません…

###前提・実現したいこと
初めまして。
独学で画像認識を勉強している者です。
tensorflowのチュートリアルプログラム(cifar10)を実行しようとしたのですが、
エラーが発生して動きません。
どうか、ご教授願います。

###発生している問題・エラーメッセージ

Traceback(most recent call last):
  File "cifar10_train.py",line 47, in <module>
    from tensorflow.models.image.cifar10 import cifar10
ImportError: No module named 'tensorflow.models'

###該当のソースコード

python
1# Copyright 2015 The TensorFlow Authors. All Rights Reserved.
2#
3# Licensed under the Apache License, Version 2.0 (the "License");
4# you may not use this file except in compliance with the License.
5# You may obtain a copy of the License at
6#
7#     http://www.apache.org/licenses/LICENSE-2.0
8#
9# Unless required by applicable law or agreed to in writing, software
10# distributed under the License is distributed on an "AS IS" BASIS,
11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12# See the License for the specific language governing permissions and
13# limitations under the License.
14# ==============================================================================
15
16"""A binary to train CIFAR-10 using a single GPU.
17
18Accuracy:
19cifar10_train.py achieves ~86% accuracy after 100K steps (256 epochs of
20data) as judged by cifar10_eval.py.
21
22Speed: With batch_size 128.
23
24System        | Step Time (sec/batch)  |     Accuracy
25------------------------------------------------------------------
261 Tesla K20m  | 0.35-0.60              | ~86% at 60K steps  (5 hours)
271 Tesla K40m  | 0.25-0.35              | ~86% at 100K steps (4 hours)
28
29Usage:
30Please see the tutorial and website for how to download the CIFAR-10
31data set, compile the program and train the model.
32
33http://tensorflow.org/tutorials/deep_cnn/
34"""
35from __future__ import absolute_import
36from __future__ import division
37from __future__ import print_function
38
39from datetime import datetime
40
41import os.path
42import time
43
44import numpy as np
45from six.moves import xrange  # pylint: disable=redefined-builtin
46import tensorflow as tf
47
48from tensorflow.models.image.cifar10 import cifar10
49
50FLAGS = tf.app.flags.FLAGS
51
52tf.app.flags.DEFINE_string('train_dir', '/tmp/cifar10_train',
53                           """Directory where to write event logs """
54                           """and checkpoint.""")
55tf.app.flags.DEFINE_integer('max_steps', 1000000,
56                            """Number of batches to run.""")
57tf.app.flags.DEFINE_boolean('log_device_placement', False,
58                            """Whether to log device placement.""")
59
60
61def train():
62  """Train CIFAR-10 for a number of steps."""
63  with tf.Graph().as_default():
64    global_step = tf.Variable(0, trainable=False)
65
66    # Get images and labels for CIFAR-10.
67    images, labels = cifar10.distorted_inputs()
68
69    # Build a Graph that computes the logits predictions from the
70    # inference model.
71    logits = cifar10.inference(images)
72
73    # Calculate loss.
74    loss = cifar10.loss(logits, labels)
75
76    # Build a Graph that trains the model with one batch of examples and
77    # updates the model parameters.
78    train_op = cifar10.train(loss, global_step)
79
80    # Create a saver.
81    saver = tf.train.Saver(tf.all_variables())
82
83    # Build the summary operation based on the TF collection of Summaries.
84    summary_op = tf.merge_all_summaries()
85
86    # Build an initialization operation to run below.
87    init = tf.initialize_all_variables()
88
89    # Start running operations on the Graph.
90    sess = tf.Session(config=tf.ConfigProto(
91        log_device_placement=FLAGS.log_device_placement))
92    sess.run(init)
93
94    # Start the queue runners.
95    tf.train.start_queue_runners(sess=sess)
96
97    summary_writer = tf.train.SummaryWriter(FLAGS.train_dir, sess.graph)
98
99    for step in xrange(FLAGS.max_steps):
100      start_time = time.time()
101      _, loss_value = sess.run([train_op, loss])
102      duration = time.time() - start_time
103
104      assert not np.isnan(loss_value), 'Model diverged with loss = NaN'
105
106      if step % 10 == 0:
107        num_examples_per_step = FLAGS.batch_size
108        examples_per_sec = num_examples_per_step / duration
109        sec_per_batch = float(duration)
110
111        format_str = ('%s: step %d, loss = %.2f (%.1f examples/sec; %.3f '
112                      'sec/batch)')
113        print (format_str % (datetime.now(), step, loss_value,
114                             examples_per_sec, sec_per_batch))
115
116      if step % 100 == 0:
117        summary_str = sess.run(summary_op)
118        summary_writer.add_summary(summary_str, step)
119
120      # Save the model checkpoint periodically.
121      if step % 1000 == 0 or (step + 1) == FLAGS.max_steps:
122        checkpoint_path = os.path.join(FLAGS.train_dir, 'model.ckpt')
123        saver.save(sess, checkpoint_path, global_step=step)
124
125
126def main(argv=None):  # pylint: disable=unused-argument
127  cifar10.maybe_download_and_extract()
128  if tf.gfile.Exists(FLAGS.train_dir):
129    tf.gfile.DeleteRecursively(FLAGS.train_dir)
130  tf.gfile.MakeDirs(FLAGS.train_dir)
131  train()
132
133
134if __name__ == '__main__':
135  tf.app.run()
136
137