cifar10.py

# Copyright 2015 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================

"""Builds the CIFAR-10 network.

Summary of available functions:

 # Compute input images and labels for training. If you would like to run
 # evaluations, use inputs() instead.
 inputs, labels = distorted_inputs()

 # Compute inference on the model inputs to make a prediction.
 predictions = inference(inputs)

"""
# pylint: disable=missing-docstring
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import gzip
import os
import re
import sys
import sys
import tarfile

from six.moves import urllib
import tensorflow as tf


from tensorflow.models.image.cifar10 import cifar10_input
FLAGS = tf.app.flags.FLAGS


tf.app.flags.DEFINE_boolean('use_outerp', False,
                                    """Use Outer product for computing convolutional filter.""")

# Global constants describing the CIFAR-10 data set.
IMAGE_SIZE = cifar10_input.IMAGE_SIZE
CIFAR10 = True
NUM_CLASSES = cifar10_input.NUM_CLASSES
NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN = cifar10_input.NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN
NUM_EXAMPLES_PER_EPOCH_FOR_EVAL = cifar10_input.NUM_EXAMPLES_PER_EPOCH_FOR_EVAL


# Constants describing the training process.
MOVING_AVERAGE_DECAY = 0.9999     # The decay to use for the moving average.
NUM_EPOCHS_PER_DECAY = 350.0      # Epochs after which learning rate decays.
LEARNING_RATE_DECAY_FACTOR = 0.1  # Learning rate decay factor.
INITIAL_LEARNING_RATE = 0.1       # Initial learning rate.

# If a model is trained with multiple GPUs, prefix all Op names with tower_name
# to differentiate the operations. Note that this prefix is removed from the
# names of the summaries when visualizing a model.
TOWER_NAME = 'tower'

DATA_URL = 'http://www.cs.toronto.edu/~kriz/cifar-10-binary.tar.gz'




def _variable_on_cpu(name, shape, initializer):
  """Helper to create a Variable stored on CPU memory.

  Args:
    name: name of the variable
    shape: list of ints
    initializer: initializer for Variable

  Returns:
    Variable Tensor
  """
  sys.stdout.flush()
  with tf.device('/cpu:0'):
    dtype = tf.float32
    if len(shape) != 4 or not FLAGS.use_outerp:
        var = tf.get_variable(name, shape, initializer=initializer, dtype=dtype)
    else:
        allw = []
        for i in xrange(shape[-1]):
            bw = []
            for j in range(shape[-2]):
                v1 = tf.get_variable(name+"v1_%d_%d"%(i,j), shape[0], initializer=initializer, dtype=dtype)
                v2 = tf.get_variable(name+"v2_%d_%d"%(i,j), shape[1], initializer=initializer, dtype=dtype)
                v1 = tf.reshape(v1, (shape[0], 1))
                v2 = tf.reshape(v2, (1, shape[1]))
                v3 = tf.mul(v1, v2)
                bw.append(v3)
            aw = tf.pack(bw)
            allw.append(aw)
        var = tf.pack(allw)
        var = tf.transpose(var, [2,3,1,0])
        print("My conv is of shape", var.get_shape())
        sys.stdout.flush()
  return var


def _variable_with_weight_decay(name, shape, stddev, wd):
  """Helper to create an initialized Variable with weight decay.

  Note that the Variable is initialized with a truncated normal distribution.
  A weight decay is added only if one is specified.

  Args:
    name: name of the variable
    shape: list of ints
    stddev: standard deviation of a truncated Gaussian
    wd: add L2Loss weight decay multiplied by this float. If None, weight
        decay is not added for this Variable.

  Returns:
    Variable Tensor
  """
  dtype = tf.float32
  var = _variable_on_cpu(
      name,
      shape,
      tf.truncated_normal_initializer(stddev=stddev, dtype=dtype))
  if wd is not None:
    weight_decay = tf.mul(tf.nn.l2_loss(var), wd, name='weight_loss')
    tf.add_to_collection('losses', weight_decay)
  return var


def distorted_inputs():
  """Construct distorted input for CIFAR training using the Reader ops.

  Returns:
    images: Images. 4D tensor of [batch_size, IMAGE_SIZE, IMAGE_SIZE, 3] size.
    labels: Labels. 1D tensor of [batch_size] size.

  Raises:
    ValueError: If no data_dir
  """

  if not FLAGS.data_dir:
    raise ValueError('Please supply a data_dir')
  if CIFAR10:
    data_dir = os.path.join(FLAGS.data_dir, 'cifar-10-batches-bin')
  else:
    data_dir = os.path.join(FLAGS.data_dir, 'cifar-100-binary')
  images, labels = cifar10_input.distorted_inputs(data_dir=data_dir,
                                                  batch_size=FLAGS.batch_size)
  return images, labels


def inference(images):
  """Build the CIFAR-10 model.

  Args:
    images: Images returned from distorted_inputs() or inputs().

  Returns:
    Logits.
  """
  # We instantiate all variables using tf.get_variable() instead of
  # tf.Variable() in order to share variables across multiple GPU training runs.
  # If we only ran this model on a single GPU, we could simplify this function
  # by replacing all instances of tf.get_variable() with tf.Variable().
  #
  
  # attention_conv1
  with tf.variable_scope('conv1') as scope:
    kernel = _variable_with_weight_decay('conv1_weights',
                                         shape=[5, 5, 3, 64],
                                         stddev=5e-2,
                                         wd=0.0)
    conv = tf.nn.conv2d(images, kernel, [1, 1, 1, 1], padding='SAME')
    biases = _variable_on_cpu('biases', [64], tf.constant_initializer(0.0))
    bias = tf.nn.bias_add(conv, biases)
    conv1 = tf.nn.relu(bias, name=scope.name)

  # pool1
  pool1 = tf.nn.max_pool(conv1, ksize=[1, 3, 3, 1], strides=[1, 2, 2, 1],
                         padding='SAME', name='pool1')
  # norm1
  norm1 = tf.nn.lrn(pool1, 4, bias=1.0, alpha=0.001 / 9.0, beta=0.75,
                    name='norm1')

  # conv2
  with tf.variable_scope('conv2') as scope:
    kernel = _variable_with_weight_decay('conv2_weights',
                                         shape=[5, 5, 64, 64],
                                         stddev=5e-2,
                                         wd=0.0)
    conv = tf.nn.conv2d(norm1, kernel, [1, 1, 1, 1], padding='SAME')
    biases = _variable_on_cpu('biases', [64], tf.constant_initializer(0.1))
    bias = tf.nn.bias_add(conv, biases)
    conv2 = tf.nn.relu(bias, name=scope.name)

  # norm2
  norm2 = tf.nn.lrn(conv2, 4, bias=1.0, alpha=0.001 / 9.0, beta=0.75,
                    name='norm2')
  # pool2
  pool2 = tf.nn.max_pool(norm2, ksize=[1, 3, 3, 1],
                         strides=[1, 2, 2, 1], padding='SAME', name='pool2')

  # local3
  with tf.variable_scope('local3') as scope:
    # Move everything into depth so we can perform a single matrix multiply.
    reshape = tf.reshape(pool2, [FLAGS.batch_size, -1])
    dim = reshape.get_shape()[1].value
    weights = _variable_with_weight_decay('weights', shape=[dim, 384],
                                          stddev=0.04, wd=0.004)
    biases = _variable_on_cpu('biases', [384], tf.constant_initializer(0.1))
    local3 = tf.nn.relu(tf.matmul(reshape, weights) + biases, name=scope.name)
  # local4
  with tf.variable_scope('local4') as scope:
    weights = _variable_with_weight_decay('weights', shape=[384, 192],
                                          stddev=0.04, wd=0.004)
    biases = _variable_on_cpu('biases', [192], tf.constant_initializer(0.1))
    local4 = tf.nn.relu(tf.matmul(local3, weights) + biases, name=scope.name)

  # linear layer(WX + b),
  # We don't apply softmax here because 
  # tf.nn.sparse_softmax_cross_entropy_with_logits accepts the unscaled logits 
  # and performs the softmax internally for efficiency.
  with tf.variable_scope('softmax_linear') as scope:
    weights = _variable_with_weight_decay('weights', [192, NUM_CLASSES],
                                          stddev=1/192.0, wd=0.0)
    biases = _variable_on_cpu('biases', [NUM_CLASSES],
                              tf.constant_initializer(0.0))
    print("Final pre softmax hain")
    sys.stdout.flush()
    softmax_linear = tf.add(tf.matmul(local4, weights), biases, name=scope.name)

  return softmax_linear


def maybe_download_and_extract():
  """Download and extract the tarball from Alex's website."""
  dest_directory = FLAGS.data_dir
  if not os.path.exists(dest_directory):
    os.makedirs(dest_directory)
  filename = DATA_URL.split('/')[-1]
  filepath = os.path.join(dest_directory, filename)
  if not os.path.exists(filepath):
    def _progress(count, block_size, total_size):
      sys.stdout.write('\r>> Downloading %s %.1f%%' % (filename,
          float(count * block_size) / float(total_size) * 100.0))
      sys.stdout.flush()
    filepath, _ = urllib.request.urlretrieve(DATA_URL, filepath, _progress)
    print()
    statinfo = os.stat(filepath)
    print('Successfully downloaded', filename, statinfo.st_size, 'bytes.')
  
  tarfile.open(filepath, 'r:gz').extractall(dest_directory)

run.py

# Copyright 2015 The TensorFlow Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================

"""A binary to train CIFAR-10 using a single GPU.

Accuracy:
cifar10_train.py achieves ~86% accuracy after 100K steps (256 epochs of
data) as judged by cifar10_eval.py.

Speed: With batch_size 128.

System        | Step Time (sec/batch)  |     Accuracy
------------------------------------------------------------------
1 Tesla K20m  | 0.35-0.60              | ~86% at 60K steps  (5 hours)
1 Tesla K40m  | 0.25-0.35              | ~86% at 100K steps (4 hours)

Usage:
Please see the tutorial and website for how to download the CIFAR-10
data set, compile the program and train the model.

http://tensorflow.org/tutorials/deep_cnn/
"""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

from datetime import datetime
import os.path
import time
import sys

import numpy as np
from six.moves import xrange  # pylint: disable=redefined-builtin
import tensorflow as tf

import cifar10

FLAGS = tf.app.flags.FLAGS

tf.app.flags.DEFINE_string('train_dir', './cifar10_train',
                           """Directory where to write event logs """
                           """and checkpoint.""")
tf.app.flags.DEFINE_integer('max_steps', 1000000,
                            """Number of batches to run.""")
tf.app.flags.DEFINE_boolean('log_device_placement', False,
                            """Whether to log device placement.""")


def train():
  """Train CIFAR-10 for a number of steps."""
  with tf.Graph().as_default():
    global_step = tf.Variable(0, trainable=False)

    # Get images and labels for CIFAR-10.
    images, _ = cifar10.distorted_inputs()

    # Build a Graph that computes the logits predictions from the
    # inference model.
    logits = cifar10.inference(images)


    # Build a Graph that trains the model with one batch of examples and
    # updates the model parameters.

    # Create a saver.
    saver = tf.train.Saver(tf.all_variables())


    # Build an initialization operation to run below.
    #init = tf.global_variables_initializer()
    init = tf.initialize_all_variables()

    # Start running operations on the Graph.
    sess = tf.Session(config=tf.ConfigProto(
        log_device_placement=FLAGS.log_device_placement))
    sess.run(init)

    # Start the queue runners.
    tf.train.start_queue_runners(sess=sess)

    summary_writer = tf.train.SummaryWriter(FLAGS.train_dir, sess.graph)

    for step in xrange(FLAGS.max_steps):
      start_time = time.time()
      _ = sess.run([logits])
      duration = time.time() - start_time


      if step % 10 == 0:
        num_examples_per_step = FLAGS.batch_size
        examples_per_sec = num_examples_per_step / duration
        sec_per_batch = float(duration)

        format_str = ('%s: step %d, (%.1f examples/sec; %.3f '
                      'sec/batch)')
        print (format_str % (datetime.now(), step,
                             examples_per_sec, sec_per_batch))
        sys.stdout.flush()

      # Save the model checkpoint periodically.
      if step % 1000 == 0 or (step + 1) == FLAGS.max_steps:
        checkpoint_path = os.path.join(FLAGS.train_dir, 'model.ckpt')
        saver.save(sess, checkpoint_path, global_step=step)


def main(argv=None):  # pylint: disable=unused-argument
  cifar10.maybe_download_and_extract()
  if tf.gfile.Exists(FLAGS.train_dir):
    tf.gfile.DeleteRecursively(FLAGS.train_dir)
  tf.gfile.MakeDirs(FLAGS.train_dir)
  train()


if __name__ == '__main__':
  tf.app.run()