TensorFlow Example for Beginners

mnist_beginner.py

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import gzip
import os

import tensorflow.python.platform

import numpy
from six.moves import urllib
from six.moves import xrange  # pylint: disable=redefined-builtin

import tensorflow as tf

## SECTION FOR INPUT DATA

SOURCE_URL = 'http://yann.lecun.com/exdb/mnist/'

def maybe_download(filename, work_directory):
  """Download the data from Yann's website, unless it's already here."""
  if not os.path.exists(work_directory):
    os.mkdir(work_directory)
  filepath = os.path.join(work_directory, filename)
  if not os.path.exists(filepath):
    filepath, _ = urllib.request.urlretrieve(SOURCE_URL + filename, filepath)
    statinfo = os.stat(filepath)
    print('Successfully downloaded', filename, statinfo.st_size, 'bytes.')
  return filepath


def _read32(bytestream):
  dt = numpy.dtype(numpy.uint32).newbyteorder('>')
  return numpy.frombuffer(bytestream.read(4), dtype=dt)[0]


def extract_images(filename):
  """Extract the images into a 4D uint8 numpy array [index, y, x, depth]."""
  print('Extracting', filename)
  with gzip.open(filename) as bytestream:
    magic = _read32(bytestream)
    if magic != 2051:
      raise ValueError(
          'Invalid magic number %d in MNIST image file: %s' %
          (magic, filename))
    num_images = _read32(bytestream)
    rows = _read32(bytestream)
    cols = _read32(bytestream)
    buf = bytestream.read(rows * cols * num_images)
    data = numpy.frombuffer(buf, dtype=numpy.uint8)
    data = data.reshape(num_images, rows, cols, 1)
    return data


def dense_to_one_hot(labels_dense, num_classes=10):
  """Convert class labels from scalars to one-hot vectors."""
  num_labels = labels_dense.shape[0]
  index_offset = numpy.arange(num_labels) * num_classes
  labels_one_hot = numpy.zeros((num_labels, num_classes))
  labels_one_hot.flat[index_offset + labels_dense.ravel()] = 1
  return labels_one_hot


def extract_labels(filename, one_hot=False):
  """Extract the labels into a 1D uint8 numpy array [index]."""
  print('Extracting', filename)
  with gzip.open(filename) as bytestream:
    magic = _read32(bytestream)
    if magic != 2049:
      raise ValueError(
          'Invalid magic number %d in MNIST label file: %s' %
          (magic, filename))
    num_items = _read32(bytestream)
    buf = bytestream.read(num_items)
    labels = numpy.frombuffer(buf, dtype=numpy.uint8)
    if one_hot:
      return dense_to_one_hot(labels)
    return labels


class DataSet(object):

  def __init__(self, images, labels, fake_data=False, one_hot=False,
               dtype=tf.float32):
    """Construct a DataSet.

    one_hot arg is used only if fake_data is true.  `dtype` can be either
    `uint8` to leave the input as `[0, 255]`, or `float32` to rescale into
    `[0, 1]`.
    """
    dtype = tf.as_dtype(dtype).base_dtype
    if dtype not in (tf.uint8, tf.float32):
      raise TypeError('Invalid image dtype %r, expected uint8 or float32' %
                      dtype)
    if fake_data:
      self._num_examples = 10000
      self.one_hot = one_hot
    else:
      assert images.shape[0] == labels.shape[0], (
          'images.shape: %s labels.shape: %s' % (images.shape,
                                                 labels.shape))
      self._num_examples = images.shape[0]

      # Convert shape from [num examples, rows, columns, depth]
      # to [num examples, rows*columns] (assuming depth == 1)
      assert images.shape[3] == 1
      images = images.reshape(images.shape[0],
                              images.shape[1] * images.shape[2])
      if dtype == tf.float32:
        # Convert from [0, 255] -> [0.0, 1.0].
        images = images.astype(numpy.float32)
        images = numpy.multiply(images, 1.0 / 255.0)
    self._images = images
    self._labels = labels
    self._epochs_completed = 0
    self._index_in_epoch = 0

  @property
  def images(self):
    return self._images

  @property
  def labels(self):
    return self._labels

  @property
  def num_examples(self):
    return self._num_examples

  @property
  def epochs_completed(self):
    return self._epochs_completed

  def next_batch(self, batch_size, fake_data=False):
    """Return the next `batch_size` examples from this data set."""
    if fake_data:
      fake_image = [1] * 784
      if self.one_hot:
        fake_label = [1] + [0] * 9
      else:
        fake_label = 0
      return [fake_image for _ in xrange(batch_size)], [
          fake_label for _ in xrange(batch_size)]
    start = self._index_in_epoch
    self._index_in_epoch += batch_size
    if self._index_in_epoch > self._num_examples:
      # Finished epoch
      self._epochs_completed += 1
      # Shuffle the data
      perm = numpy.arange(self._num_examples)
      numpy.random.shuffle(perm)
      self._images = self._images[perm]
      self._labels = self._labels[perm]
      # Start next epoch
      start = 0
      self._index_in_epoch = batch_size
      assert batch_size <= self._num_examples
    end = self._index_in_epoch
    return self._images[start:end], self._labels[start:end]


def read_data_sets(train_dir, fake_data=False, one_hot=False, dtype=tf.float32):
  class DataSets(object):
    pass
  data_sets = DataSets()

  if fake_data:
    def fake():
      return DataSet([], [], fake_data=True, one_hot=one_hot, dtype=dtype)
    data_sets.train = fake()
    data_sets.validation = fake()
    data_sets.test = fake()
    return data_sets

  TRAIN_IMAGES = 'train-images-idx3-ubyte.gz'
  TRAIN_LABELS = 'train-labels-idx1-ubyte.gz'
  TEST_IMAGES = 't10k-images-idx3-ubyte.gz'
  TEST_LABELS = 't10k-labels-idx1-ubyte.gz'
  VALIDATION_SIZE = 5000

  local_file = maybe_download(TRAIN_IMAGES, train_dir)
  train_images = extract_images(local_file)

  local_file = maybe_download(TRAIN_LABELS, train_dir)
  train_labels = extract_labels(local_file, one_hot=one_hot)

  local_file = maybe_download(TEST_IMAGES, train_dir)
  test_images = extract_images(local_file)

  local_file = maybe_download(TEST_LABELS, train_dir)
  test_labels = extract_labels(local_file, one_hot=one_hot)

  validation_images = train_images[:VALIDATION_SIZE]
  validation_labels = train_labels[:VALIDATION_SIZE]
  train_images = train_images[VALIDATION_SIZE:]
  train_labels = train_labels[VALIDATION_SIZE:]

  data_sets.train = DataSet(train_images, train_labels, dtype=dtype)
  data_sets.validation = DataSet(validation_images, validation_labels,
                                 dtype=dtype)
  data_sets.test = DataSet(test_images, test_labels, dtype=dtype)

  return data_sets
## END OF SECTION

# Download data to MNIST_data/
# There are 55000 records (28 * 28 pixels = 784)
mnist = read_data_sets("MNIST_data/", one_hot=True)

## Creating Model

# Describe interacting operations by manipulating symbolic variables
# x will be calculated by TensorFlow (None = dimension in any length)
x = tf.placeholder(tf.float32, [None, 784])

# Weight (W) & Bias (b)
# Initially, all of them are zeros
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([10]))

# Tensor Flow it!
# We can run it in CPU and GPU (let TensorFlow handle it)
# We flip Wx to (x, W), because we need to deal with x being a
# 2D tensor with multiple inputs
y = tf.nn.softmax(tf.matmul(x, W) + b)

## Training

# Placeholder to input the correct answers
y_ = tf.placeholder(tf.float32, [None, 10])

# Implement cross entropy
# 1e-10 is for smoothing
# See http://stackoverflow.com/a/34364526
cross_entropy = -tf.reduce_sum(y_ * tf.log(y + 1e-10))

# Debug line (print entropy every 100 images)
# cross_entropy = tf.Print(cross_entropy, [cross_entropy], "CrossE")

# Backpropagation algorithm
train_step = tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy)

# Another alternatives (you could also lower 0.01 to other learning rate)
# train_step = tf.train.AdagradOptimizer(0.01).minimize(cross_entropy)
# train_step = tf.train.AdamOptimizer().minimize(cross_entropy)
# train_step = tf.train.FtrlOptimizer(0.01).minimize(cross_entropy)
# train_step = tf.train.RMSPropOptimizer(0.01, 0.1).minimize(cross_entropy)

# Initialize
init = tf.initialize_all_variables()
sess = tf.Session()
sess.run(init)

# Run the training step 1000 times
# Stochastic training (gradient descent in this one)
for i in range(1000):
    batch_xs, batch_ys = mnist.train.next_batch(100) # take 100 images randomly
    sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys})

## Evaluation
correct_prediction = tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1))

# Cast to floating point and take the mean
# For example: [True, False, True, True] -> [1, 0, 1 , 1] -> 0.75
accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))

print(sess.run(accuracy, feed_dict={x: mnist.test.images, y_: mnist.test.labels}))

How to install and run (Python with Virtualenv)

Follow the guide at https://github.com/tensorflow/tensorflow/blob/master/tensorflow/g3doc/get_started/os_setup.md#virtualenv-installation.

Example Value in Image

The content of single randomly chosen row:

 [[  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [ 38]
  [254]
  [254]
  [ 77]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]
  [  0]]

Each image has 28 rows. There are 60000 images in train dataset (55000 for training, 5000 for validation).

Run result

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
I tensorflow/core/common_runtime/local_device.cc:40] Local device intra op parallelism threads: 8
I tensorflow/core/common_runtime/direct_session.cc:58] Direct session inter op parallelism threads: 8
0.9078

The result 0.9078 is random (because we use sampling in the stochastic training, with 100 random images per batch).