distributed tf

distributed.py

from __future__ import print_function

import tensorflow as tf
import sys
import time
import os
import json
import logging

# cluster specification


cluster_spec = json.loads(os.environ["CLUSTER_SPEC"])
cluster = tf.train.ClusterSpec(cluster_spec)
logging.getLogger().setLevel(logging.INFO)


# input flags
tf.app.flags.DEFINE_string("job_name", "", "Either 'ps' or 'worker'")
tf.app.flags.DEFINE_integer("task_index", 0, "Index of task within the job")
FLAGS = tf.app.flags.FLAGS
FLAGS.job_name = os.environ.get('JOB_NAME')
FLAGS.task_index = int(os.environ.get('TASK_INDEX', '0'))
logging.info(FLAGS.job_name)

# start a server for a specific task
server = tf.train.Server(
    cluster,
    job_name=FLAGS.job_name,
    task_index=FLAGS.task_index)

# config
batch_size = 100
learning_rate = 0.0005
training_epochs = 20
logs_path = "/tmp/mnist/1"

# load mnist data set
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets('MNIST_data', one_hot=True)

if FLAGS.job_name == "ps":
    server.join()
elif FLAGS.job_name == "worker":

	# Between-graph replication
	with tf.device(tf.train.replica_device_setter(
		worker_device="/job:worker/task:%d" % FLAGS.task_index,
		cluster=cluster)):

		# count the number of updates
		global_step = tf.get_variable(
            'global_step',
            [],
            initializer = tf.constant_initializer(0),
			trainable = False)

		# input images
		with tf.name_scope('input'):
		  # None -> batch size can be any size, 784 -> flattened mnist image
		  x = tf.placeholder(tf.float32, shape=[None, 784], name="x-input")
		  # target 10 output classes
		  y_ = tf.placeholder(tf.float32, shape=[None, 10], name="y-input")

		# model parameters will change during training so we use tf.Variable
		tf.set_random_seed(1)
		with tf.name_scope("weights"):
			W1 = tf.Variable(tf.random_normal([784, 100]))
			W2 = tf.Variable(tf.random_normal([100, 10]))

		# bias
		with tf.name_scope("biases"):
			b1 = tf.Variable(tf.zeros([100]))
			b2 = tf.Variable(tf.zeros([10]))

		# implement model
		with tf.name_scope("softmax"):
			# y is our prediction
			z2 = tf.add(tf.matmul(x,W1),b1)
			a2 = tf.nn.sigmoid(z2)
			z3 = tf.add(tf.matmul(a2,W2),b2)
			y  = tf.nn.softmax(z3)

		# specify cost function
		with tf.name_scope('cross_entropy'):
			# this is our cost
			cross_entropy = tf.reduce_mean(
                -tf.reduce_sum(y_ * tf.log(y), reduction_indices=[1]))

		# specify optimizer
		with tf.name_scope('train'):
			# optimizer is an "operation" which we can execute in a session
			grad_op = tf.train.GradientDescentOptimizer(learning_rate)
			'''
			rep_op = tf.train.SyncReplicasOptimizer(
                grad_op,
			    replicas_to_aggregate=len(workers),
 				replica_id=FLAGS.task_index, 
 			    total_num_replicas=len(workers),
 				use_locking=True)
 			train_op = rep_op.minimize(cross_entropy, global_step=global_step)
 			'''
			train_op = grad_op.minimize(cross_entropy, global_step=global_step)
			
		'''
		init_token_op = rep_op.get_init_tokens_op()
		chief_queue_runner = rep_op.get_chief_queue_runner()
		'''

		with tf.name_scope('Accuracy'):
			# accuracy
			correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(y_,1))
			accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

		# create a summary for our cost and accuracy
		tf.summary.scalar("cost", cross_entropy)
		tf.summary.scalar("accuracy", accuracy)

		# merge all summaries into a single "operation" which we can execute in a session 
		summary_op = tf.summary.merge_all()
		init_op = tf.global_variables_initializer()
		logging.info("Variables initialized ...")

	sv = tf.train.Supervisor(is_chief=(FLAGS.task_index == 0),
														global_step=global_step,
														init_op=init_op)

	begin_time = time.time()
	frequency = 100
	with sv.prepare_or_wait_for_session(server.target) as sess:
		'''
		# is chief
		if FLAGS.task_index == 0:
			sv.start_queue_runners(sess, [chief_queue_runner])
			sess.run(init_token_op)
		'''
		# create log writer object (this will log on every machine)
		writer = tf.summary.FileWriter(logs_path, graph=tf.get_default_graph())
				
		# perform training cycles
		start_time = time.time()
		for epoch in range(training_epochs):

			# number of batches in one epoch
			batch_count = int(mnist.train.num_examples/batch_size)

			count = 0
			for i in range(batch_count):
				batch_x, batch_y = mnist.train.next_batch(batch_size)
				
				# perform the operations we defined earlier on batch
				_, cost, summary, step = sess.run(
												[train_op, cross_entropy, summary_op, global_step], 
												feed_dict={x: batch_x, y_: batch_y})
				writer.add_summary(summary, step)

				count += 1
				if count % frequency == 0 or i+1 == batch_count:
					elapsed_time = time.time() - start_time
					start_time = time.time()
					logging.info("Step: {}, Epoch: {}, Batch: {} of {}, Cost: {}, AvgTime: {}".format(step, epoch, i + 1, batch_count, cost, float(elapsed_time*1000/frequency)))
					count = 0


		logging.info("Test-Accuracy: %2.2f" % sess.run(accuracy, feed_dict={x: mnist.test.images, y_: mnist.test.labels}))
		logging.info("Total Time: %3.2fs" % float(time.time() - begin_time))
		logging.info("Final Cost: %.4f" % cost)

	sv.stop()
	logging.info("done")