example.py

import numpy as np
import time
import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data
from tqdm import tqdm
from spsa import SimultaneousPerturbationOptimizer

logdir = 'data/log/run_{}'.format(time.time())
mnist = input_data.read_data_sets("data/mnist", one_hot=True)

# common training parameters
training_epochs = 5
batch_size = 100

# Gradient Descent parameters
learning_rate = 0.01

# SPSA parameters
a = 0.01
c = 0.01
alpha = 1.0
gamma = 0.4

x = tf.placeholder(tf.float32, [None, 784], name='Input')
y = tf.placeholder(tf.float32, [None, 10], name='Labels')

# model
with tf.name_scope('Model'):
    W = tf.Variable(tf.random_uniform([784, 10]), name='Weights')
    b = tf.Variable(tf.random_uniform([10]), name='Bias')
    p = tf.nn.softmax(tf.matmul(x, W) + b)

# objective
with tf.name_scope('Loss'):
    cost = tf.reduce_mean(-tf.reduce_sum(y*tf.log(p), reduction_indices=1))

with tf.name_scope('Optimizer'):
    #optimizer = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost)
    optimizer = SimultaneousPerturbationOptimizer().minimize(cost)

with tf.name_scope('Accuracy'):
    acc = tf.equal(tf.argmax(p, 1), tf.argmax(y, 1))
    acc = tf.reduce_mean(tf.cast(acc, tf.float32))

# Variables for Tensorboard
tf.summary.scalar("loss", cost)
tf.summary.scalar("accuracy", acc)
merged_summary_op = tf.summary.merge_all()

init = tf.global_variables_initializer()

# Launch the graph
with tf.Session() as sess:
    sess.run(init)
    summary_writer = tf.summary.FileWriter(logdir, graph=tf.get_default_graph())

    # Training Loop
    for epoch in range(training_epochs):
        avg_cost = 0.
        total_batch = int(mnist.train.num_examples/batch_size)
        # Loop over all batches
        for i in range(total_batch):
            batch_xs, batch_ys = mnist.train.next_batch(batch_size)
            # Run optimization op, cost op (to get loss value)
            # and summary nodes
            _, c, summary = sess.run([optimizer, cost, merged_summary_op],
                                     feed_dict={x: batch_xs, y: batch_ys})
            # Write logs at every iteration
            summary_writer.add_summary(summary, epoch * total_batch + i)
            # Compute average loss
            avg_cost += c / total_batch
        # Display logs per epoch step
        print("Epoch:", '%04d' % (epoch+1), "cost=", "{:.9f}".format(avg_cost))

    print("Finished!")

    # Test model
    # Calculate accuracy
    print("Accuracy:", acc.eval({x: mnist.test.images, y: mnist.test.labels}))