alexnet_multi_gpu.py

from six.moves import xrange  # pylint: disable=redefined-builtin
from datetime import datetime
import math
import re
import time

import tensorflow.python.platform
import tensorflow as tf

FLAGS = tf.app.flags.FLAGS

tf.app.flags.DEFINE_integer('batch_size', 64,
                            """Batch size.""")
tf.app.flags.DEFINE_integer('num_batches', 100,
                            """Number of batches to run.""")
tf.app.flags.DEFINE_string('data_format', 'NCHW',
                           """The data format for Convnet operations.
                           Can be either NHWC or NCHW.
                           """)

tf.app.flags.DEFINE_string('train_dir', '/train_model',
                           """Directory where to write event logs """
                           """and checkpoint.""")
tf.app.flags.DEFINE_integer('num_gpus', 4,
                            """How many GPUs to use.""")
tf.app.flags.DEFINE_boolean('log_device_placement', False,
                            """Whether to log device placement.""")

NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN=50000
NUM_EPOCHS_PER_DECAY=50
INITIAL_LEARNING_RATE = 0.1 
LEARNING_RATE_DECAY_FACTOR = 0.1
TOWER_NAME = 'tower'


def _conv(name, inpOp, nIn, nOut, kH, kW, dH, dW, padType, wd=0.005):
    with tf.name_scope(name) as scope:
        kernel = tf.get_variable(name + '_w',[kH, kW, nIn, nOut],
          initializer=tf.truncated_normal_initializer(stddev=0.01, dtype=tf.float32),
          dtype=tf.float32)

        if wd is not None:
            weight_decay = tf.mul(tf.nn.l2_loss(kernel), wd, name='weight_loss')
            tf.add_to_collection('losses', weight_decay)

        if FLAGS.data_format == 'NCHW':
          strides = [1, 1, dH, dW]
        else:
          strides = [1, dH, dW, 1]
        conv = tf.nn.conv2d(inpOp, kernel, strides, padding=padType,
                            data_format=FLAGS.data_format)

        biases = tf.get_variable(name=name + '_b', shape=[nOut], 
            initializer=tf.constant_initializer(value=0.0, dtype=tf.float32),
            dtype=tf.float32)

        bias = tf.reshape(
            tf.nn.bias_add(conv, biases, data_format=FLAGS.data_format),
            conv.get_shape())

        conv1 = tf.nn.relu(bias, name=scope)
        return conv1

def _affine(name, inpOp, nIn, nOut, wd=0.005, act=True):
    with tf.name_scope(name) as scope:
        kernel = tf.get_variable(name + '_w', [nIn, nOut],
            initializer=tf.truncated_normal_initializer(stddev=0.01, dtype=tf.float32),
            dtype=tf.float32)

        if wd is not None:
            weight_decay = tf.mul(tf.nn.l2_loss(kernel), wd, name='weight_loss')
            tf.add_to_collection('losses', weight_decay)

        biases = tf.get_variable(name + '_b', [nOut],
            initializer=tf.constant_initializer(value=0.0, dtype=tf.float32),
            dtype=tf.float32,trainable=True)

        affine1 = tf.nn.relu_layer(inpOp, kernel, biases, name=name) if act else \
                  tf.matmul(inpOp, kernel) + biases

        return affine1

def _mpool(name, inpOp, kH, kW, dH, dW):
    if FLAGS.data_format == 'NCHW':
      ksize = [1, 1, kH, kW]
      strides = [1, 1, dH, dW]
    else:
      ksize = [1, kH, kW, 1]
      strides = [1, dH, dW, 1]
    return tf.nn.max_pool(inpOp,
                          ksize=ksize,
                          strides=strides,
                          padding='VALID',
                          data_format=FLAGS.data_format,
                          name=name)

def _norm(name, l_input, lsize=4):
    return tf.nn.lrn(l_input, lsize, bias=1.0,
                     alpha=0.001 / 9.0,
                     beta=0.75, name=name)

def loss(logits, labels):
    labels = tf.cast(labels, tf.int64)
    cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(
                    logits, labels, name='cross_entropy_per_example')
    cross_entropy_mean = tf.reduce_mean(cross_entropy, name='cross_entropy')
    tf.add_to_collection('losses', cross_entropy_mean)

    # The total loss is defined as the cross entropy loss plus all of the weight
    # decay terms (L2 loss).
    return tf.add_n(tf.get_collection('losses'), name='total_loss')


def get_incoming_shape(incoming):
    """ Returns the incoming data shape """
    if isinstance(incoming, tf.Tensor):
        return incoming.get_shape().as_list()
    elif type(incoming) in [np.array, list, tuple]:
        return np.shape(incoming)
    else:
        raise Exception("Invalid incoming layer.")

def inference(images):
    conv1 = _conv ('conv1', images, 3, 96, 11, 11, 4, 4, 'VALID')
    pool1 = _mpool('pool1', conv1,  3, 3, 2, 2)
    norm1 = _norm ('norm1', pool1, lsize=5)
    conv2 = _conv ('conv2', norm1,  96, 256, 5, 5, 1, 1, 'SAME')
    pool2 = _mpool('pool2', conv2,  3, 3, 2, 2)
    norm2 = _norm ('norm2', pool2, lsize=5)
    conv3 = _conv ('conv3', norm2,  256, 384, 3, 3, 1, 1, 'SAME')
    conv4 = _conv ('conv4', conv3,  384, 384, 3, 3, 1, 1, 'SAME')
    conv5 = _conv ('conv5', conv4,  384, 256, 3, 3, 1, 1, 'SAME')
    pool5 = _mpool('pool5', conv5,  3, 3, 2, 2)
    resh1 = tf.reshape(pool5, [-1, 256 * 6 * 6])
    affn1 = _affine('fc6', resh1, 256 * 6 * 6, 4096)
    affn2 = _affine('fc7', affn1, 4096, 4096)
    affn3 = _affine('fc8', affn2, 4096, 1000, wd=None, act=False) # last fc

    return affn3

def tower_loss(scope):
    """Calculate the total loss on a single tower running the model.
    Args:
        scope: unique prefix string identifying the tower, e.g. 'tower_0'
    Returns:
        Tensor of shape [] containing the total loss for a batch of data
    """
    image_size = 224
    if FLAGS.data_format == 'NCHW':
        image_shape = [FLAGS.batch_size, 3, image_size + 3, image_size + 3]
    else:
        image_shape = [FLAGS.batch_size, image_size + 3, image_size + 3, 3]
    images = tf.get_variable('image', image_shape, 
                             initializer=tf.truncated_normal_initializer(stddev=0.1, dtype=tf.float32),
                             dtype=tf.float32,
                             trainable=False)

    labels = tf.get_variable('label', [FLAGS.batch_size],
                             initializer=tf.constant_initializer(1),
                             dtype=tf.int32,
                             trainable=False)

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

    # Build the portion of the Graph calculating the losses. Note that we will
    # assemble the total_loss using a custom function below.
    _ = loss(last_layer, labels)
    
    # Assemble all of the losses for the current tower only.
    losses = tf.get_collection('losses', scope)

    # Calculate the total loss for the current tower.
    total_loss = tf.add_n(losses, name='total_loss')

    # Compute the moving average of all individual losses and the total loss.
    loss_averages = tf.train.ExponentialMovingAverage(0.9, name='avg')
    loss_averages_op = loss_averages.apply(losses + [total_loss])

    # Attach a scalar summary to all individual losses and the total loss; do the
    # same for the averaged version of the losses.
    for l in losses + [total_loss]:
        # Remove 'tower_[0-9]/' from the name in case this is a multi-GPU training
        # session. This helps the clarity of presentation on tensorboard.
        loss_name = re.sub('%s_[0-9]*/' % TOWER_NAME, '', l.op.name)
        # Name each loss as '(raw)' and name the moving average version of the loss
        # as the original loss name.
        tf.scalar_summary(loss_name +' (raw)', l)
        tf.scalar_summary(loss_name, loss_averages.average(l))

    with tf.control_dependencies([loss_averages_op]):
        total_loss = tf.identity(total_loss)
    return total_loss


def average_gradients(tower_grads):
  """Calculate the average gradient for each shared variable across all towers.
  Note that this function provides a synchronization point across all towers.
  Args:
    tower_grads: List of lists of (gradient, variable) tuples. The outer list
      is over individual gradients. The inner list is over the gradient
      calculation for each tower.
  Returns:
     List of pairs of (gradient, variable) where the gradient has been averaged
     across all towers.
  """
  average_grads = []
  for grad_and_vars in zip(*tower_grads):
    # Note that each grad_and_vars looks like the following:
    #   ((grad0_gpu0, var0_gpu0), ... , (grad0_gpuN, var0_gpuN))
    grads = []
    for g, _ in grad_and_vars:
      # Add 0 dimension to the gradients to represent the tower.
      expanded_g = tf.expand_dims(g, 0)

      # Append on a 'tower' dimension which we will average over below.
      grads.append(expanded_g)

    # Average over the 'tower' dimension.
    grad = tf.concat(0, grads)
    grad = tf.reduce_mean(grad, 0)

    # Keep in mind that the Variables are redundant because they are shared
    # across towers. So .. we will just return the first tower's pointer to
    # the Variable.
    v = grad_and_vars[0][1]
    grad_and_var = (grad, v)
    average_grads.append(grad_and_var)
  return average_grads

def time_tensorflow_run(session, target):
    num_steps_burn_in = 50
    total_duration = 0.0
    total_duration_squared = 0.0
    for i in xrange(FLAGS.num_batches + num_steps_burn_in):
      start_time = time.time()
      _, loss_value = session.run(target)
      duration = time.time() - start_time
      if i > num_steps_burn_in:
        if not i % 10:
          num_examples_per_step = FLAGS.batch_size * FLAGS.num_gpus
          examples_per_sec = num_examples_per_step / duration
          sec_per_batch = duration
           
          format_str = ('%s: step %d, loss = %.2f (%.1f examples/sec; %.3f '
                      'sec/batch batch_size = %d)')
          print (format_str %
                (datetime.now(), i - num_steps_burn_in, 
                 loss_value, duration, sec_per_batch, num_examples_per_step))

        total_duration += duration
        total_duration_squared += duration * duration

    mn = total_duration / FLAGS.num_batches
    vr = total_duration_squared / FLAGS.num_batches - mn * mn
    sd = math.sqrt(vr)
    print ('%s: FwdBwd across %d steps, %.3f +/- %.3f sec / batch' %
          (datetime.now(), FLAGS.num_batches, mn, sd))

def run_benchmark():
  with tf.Graph().as_default(), tf.device('/cpu:0'):
    # Create a variable to count the number of train() calls. This equals the
    # number of batches processed * FLAGS.num_gpus.
    global_step = tf.get_variable(
        'global_step', [],
        initializer=tf.constant_initializer(0), trainable=False)

    # Calculate the learning rate schedule.
    num_batches_per_epoch = (NUM_EXAMPLES_PER_EPOCH_FOR_TRAIN /
                             FLAGS.batch_size)
    decay_steps = int(num_batches_per_epoch * NUM_EPOCHS_PER_DECAY)

    # Decay the learning rate exponentially based on the number of steps.
    lr = tf.train.exponential_decay(INITIAL_LEARNING_RATE,
                                    global_step,
                                    decay_steps,
                                    LEARNING_RATE_DECAY_FACTOR,
                                    staircase=True)

    # Create an optimizer that performs gradient descent.
    # opt = tf.train.GradientDescentOptimizer(lr)
    opt = tf.train.MomentumOptimizer(lr, 0.9)

    # Calculate the gradients for each model tower.
    tower_grads = []
    for i in xrange(FLAGS.num_gpus):
      with tf.device('/gpu:%d' % i):
        with tf.name_scope('%s_%d' % (TOWER_NAME, i)) as scope:
          # Calculate the loss for one tower of the model. This function
          # constructs the entire model but shares the variables across
          # all towers.
          loss = tower_loss(scope)

          # Reuse variables for the next tower.
          tf.get_variable_scope().reuse_variables()

          # Retain the summaries from the final tower.
          summaries = tf.get_collection(tf.GraphKeys.SUMMARIES, scope)

          # Calculate the gradients for the batch of data on this tower.
          grads = opt.compute_gradients(loss)
          
          # Keep track of the gradients across all towers.
          tower_grads.append(grads)

    # We must calculate the mean of each gradient. Note that this is the
    # synchronization point across all towers.
    grads = average_gradients(tower_grads)

    # Apply the gradients to adjust the shared variables.
    apply_gradient_op = opt.apply_gradients(grads, global_step=global_step)

    # Group all updates to into a single train op.
    train_op = tf.group(apply_gradient_op)

    # Build an initialization operation.
    init = tf.initialize_all_variables()

    # Start running operations on the Graph. allow_soft_placement must be set to
    # True to build towers on GPU, as some of the ops do not have GPU
    # implementations.
    sess = tf.Session(config=tf.ConfigProto(
        allow_soft_placement=True,
        log_device_placement=FLAGS.log_device_placement))
    sess.run(init)
    time_tensorflow_run(sess, [train_op, loss])


def main(_):
  run_benchmark()


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