fixed compatibility issues with Python 3.5 and removed deprecated Ten…

…sorFlow code

data_helpers.py

@@ -1,13 +1,17 @@
 '''Imports CIFAR-10 data.'''
 
 import numpy as np
-import cPickle as pickle
+import pickle
+import sys
 
 def load_CIFAR10_batch(filename):
   '''load data from single CIFAR-10 file'''
 
   with open(filename, 'rb') as f:
-    dict = pickle.load(f)
+    if sys.version_info[0] < 3:
+      dict = pickle.load(f)
+    else:
+      dict = pickle.load(f, encoding='latin1')
     x = dict['data']
     y = dict['labels']
     x = x.astype(float)
@@ -19,7 +23,7 @@ def load_data():
 
   xs = []
   ys = []
-  for i in xrange(1, 6):
+  for i in range(1, 6):
     filename = 'cifar-10-batches-py/data_batch_' + str(i)
     X, Y = load_CIFAR10_batch(filename)
     xs.append(X)
@@ -69,7 +73,7 @@ def reshape_data(data_dict):
 def gen_batch(data, batch_size, num_iter):
   data = np.array(data)
   index = len(data)
-  for i in xrange(num_iter):
+  for i in range(num_iter):
     index += batch_size
     if (index + batch_size > len(data)):
       index = 0

run_fc_model.py

@@ -70,7 +70,7 @@
 accuracy = two_layer_fc.evaluation(logits, labels_placeholder)
 
 # Operation merging summary data for TensorBoard
-summary = tf.merge_all_summaries()
+summary = tf.summary.merge_all()
 
 # Define saver to save model state at checkpoints
 saver = tf.train.Saver()
@@ -81,12 +81,12 @@
 
 with tf.Session() as sess:
   # Initialize variables and create summary-writer
-  sess.run(tf.initialize_all_variables())
-  summary_writer = tf.train.SummaryWriter(logdir, sess.graph)
+  sess.run(tf.global_variables_initializer())
+  summary_writer = tf.summary.FileWriter(logdir, sess.graph)
 
   # Generate input data batches
   zipped_data = zip(data_sets['images_train'], data_sets['labels_train'])
-  batches = data_helpers.gen_batch(zipped_data, FLAGS.batch_size,
+  batches = data_helpers.gen_batch(list(zipped_data), FLAGS.batch_size,
     FLAGS.max_steps)
 
   for i in range(FLAGS.max_steps):

softmax.py

@@ -58,7 +58,7 @@
 
 with tf.Session() as sess:
   # Initialize variables
-  sess.run(tf.initialize_all_variables())
+  sess.run(tf.global_variables_initializer())
 
   # Repeat max_steps times
   for i in range(max_steps):

two_layer_fc.py

@@ -28,7 +28,6 @@ def inference(images, image_pixels, hidden_units, classes, reg_constant=0):
     weights = tf.get_variable(
       name='weights',
       shape=[image_pixels, hidden_units],
-      # initializer=tf.truncated_normal_initializer(stddev=1.0 / np.sqrt(float(image_pixels)), seed=1234),
       initializer=tf.truncated_normal_initializer(
         stddev=1.0 / np.sqrt(float(image_pixels))),
       regularizer=tf.contrib.layers.l2_regularizer(reg_constant)
@@ -53,7 +52,7 @@ def inference(images, image_pixels, hidden_units, classes, reg_constant=0):
     logits = tf.matmul(hidden, weights) + biases
 
     # Define summery-operation for 'logits'-variable
-    tf.histogram_summary('logits', logits)
+    tf.summary.histogram('logits', logits)
 
   return logits
 
@@ -80,7 +79,7 @@ def loss(logits, labels):
       tf.GraphKeys.REGULARIZATION_LOSSES))
 
     # Add a scalar summary for the loss
-    tf.scalar_summary('loss', loss)
+    tf.summary.scalar('loss', loss)
 
   return loss
 
@@ -128,6 +127,6 @@ def evaluation(logits, labels):
     accuracy =  tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
 
     # Summary operation for the accuracy
-    tf.scalar_summary('train_accuracy', accuracy)
+    tf.summary.scalar('train_accuracy', accuracy)
 
   return accuracy