Merge branch 'master' into master

research/differential_privacy/multiple_teachers/analysis.py

@@ -41,6 +41,7 @@ python analysis.py
 import os
 import math
 import numpy as np
+from six.moves import xrange
 import tensorflow as tf
 
 from differential_privacy.multiple_teachers.input import maybe_download
@@ -139,7 +140,7 @@ def logmgf_exact(q, priv_eps, l):
     try:
       log_t = math.log(t)
     except ValueError:
-      print "Got ValueError in math.log for values :" + str((q, priv_eps, l, t))
+      print("Got ValueError in math.log for values :" + str((q, priv_eps, l, t)))
       log_t = priv_eps * l
   else:
     log_t = priv_eps * l
@@ -171,7 +172,7 @@ def sens_at_k(counts, noise_eps, l, k):
   """
   counts_sorted = sorted(counts, reverse=True)
   if 0.5 * noise_eps * l > 1:
-    print "l too large to compute sensitivity"
+    print("l too large to compute sensitivity")
     return 0
   # Now we can assume that at k, gap remains positive
   # or we have reached the point where logmgf_exact is
@@ -268,8 +269,8 @@ def main(unused_argv):
   # Solving gives eps = (alpha - ln (delta))/l
   eps_list_nm = (total_log_mgf_nm - math.log(delta)) / l_list
 
-  print "Epsilons (Noisy Max): " + str(eps_list_nm)
-  print "Smoothed sensitivities (Noisy Max): " + str(total_ss_nm / l_list)
+  print("Epsilons (Noisy Max): " + str(eps_list_nm))
+  print("Smoothed sensitivities (Noisy Max): " + str(total_ss_nm / l_list))
 
   # If beta < eps / 2 ln (1/delta), then adding noise Lap(1) * 2 SS/eps
   # is eps,delta DP
@@ -280,12 +281,12 @@ def main(unused_argv):
   # Print the first one's scale
   ss_eps = 2.0 * beta * math.log(1/delta)
   ss_scale = 2.0 / ss_eps
-  print "To get an " + str(ss_eps) + "-DP estimate of epsilon, "
-  print "..add noise ~ " + str(ss_scale)
-  print "... times " + str(total_ss_nm / l_list)
-  print "Epsilon = " + str(min(eps_list_nm)) + "."
+  print("To get an " + str(ss_eps) + "-DP estimate of epsilon, ")
+  print("..add noise ~ " + str(ss_scale))
+  print("... times " + str(total_ss_nm / l_list))
+  print("Epsilon = " + str(min(eps_list_nm)) + ".")
   if min(eps_list_nm) == eps_list_nm[-1]:
-    print "Warning: May not have used enough values of l"
+    print("Warning: May not have used enough values of l")
 
   # Data independent bound, as mechanism is
   # 2*noise_eps DP.
@@ -294,7 +295,7 @@ def main(unused_argv):
       [logmgf_exact(1.0, 2.0 * noise_eps, l) for l in l_list])
 
   data_ind_eps_list = (data_ind_log_mgf - math.log(delta)) / l_list
-  print "Data independent bound = " + str(min(data_ind_eps_list)) + "."
+  print("Data independent bound = " + str(min(data_ind_eps_list)) + ".")
 
   return
 

research/differential_privacy/multiple_teachers/deep_cnn.py

@@ -20,6 +20,7 @@ from __future__ import print_function
 from datetime import datetime
 import math
 import numpy as np
+from six.moves import xrange
 import tensorflow as tf
 import time
 
@@ -600,5 +601,3 @@ def softmax_preds(images, ckpt_path, return_logits=False):
   tf.reset_default_graph()
 
   return preds
-
-

research/differential_privacy/multiple_teachers/input.py

@@ -24,6 +24,7 @@ import numpy as np
 import os
 from scipy.io import loadmat as loadmat
 from six.moves import urllib
+from six.moves import xrange
 import sys
 import tarfile
 

research/differential_privacy/multiple_teachers/train_student.py

@@ -19,6 +19,7 @@ from __future__ import division
 from __future__ import print_function
 
 import numpy as np
+from six.moves import xrange
 import tensorflow as tf
 
 from differential_privacy.multiple_teachers import aggregation

research/differential_privacy/privacy_accountant/python/gaussian_moments.py

@@ -40,12 +40,15 @@ To verify that the I1 >= I2 (see comments in GaussianMomentsAccountant in
 accountant.py for the context), run the same loop above with verify=True
 passed to compute_log_moment.
 """
+from __future__ import print_function
+
 import math
 import sys
 
 import numpy as np
 import scipy.integrate as integrate
 import scipy.stats
+from six.moves import xrange
 from sympy.mpmath import mp
 
 
@@ -108,10 +111,10 @@ def compute_a(sigma, q, lmbd, verbose=False):
   a_lambda_exact = ((1.0 - q) * a_lambda_first_term_exact +
                     q * a_lambda_second_term_exact)
   if verbose:
-    print "A: by binomial expansion    {} = {} + {}".format(
+    print("A: by binomial expansion    {} = {} + {}".format(
         a_lambda_exact,
         (1.0 - q) * a_lambda_first_term_exact,
-        q * a_lambda_second_term_exact)
+        q * a_lambda_second_term_exact))
   return _to_np_float64(a_lambda_exact)
 
 
@@ -125,8 +128,8 @@ def compute_b(sigma, q, lmbd, verbose=False):
   b_fn = lambda z: (np.power(mu0(z) / mu(z), lmbd) -
                     np.power(mu(-z) / mu0(z), lmbd))
   if verbose:
-    print "M =", m
-    print "f(-M) = {} f(M) = {}".format(b_fn(-m), b_fn(m))
+    print("M =", m)
+    print("f(-M) = {} f(M) = {}".format(b_fn(-m), b_fn(m)))
     assert b_fn(-m) < 0 and b_fn(m) < 0
 
   b_lambda_int1_fn = lambda z: (mu0(z) *
@@ -140,9 +143,9 @@ def compute_b(sigma, q, lmbd, verbose=False):
   b_bound = a_lambda_m1 + b_int1 - b_int2
 
   if verbose:
-    print "B: by numerical integration", b_lambda
-    print "B must be no more than     ", b_bound
-  print b_lambda, b_bound
+    print("B: by numerical integration", b_lambda)
+    print("B must be no more than     ", b_bound)
+  print(b_lambda, b_bound)
   return _to_np_float64(b_lambda)
 
 
@@ -188,10 +191,10 @@ def compute_a_mp(sigma, q, lmbd, verbose=False):
   a_lambda_second_term = integral_inf_mp(a_lambda_second_term_fn)
 
   if verbose:
-    print "A: by numerical integration {} = {} + {}".format(
+    print("A: by numerical integration {} = {} + {}".format(
         a_lambda,
         (1 - q) * a_lambda_first_term,
-        q * a_lambda_second_term)
+        q * a_lambda_second_term))
 
   return _to_np_float64(a_lambda)
 
@@ -210,8 +213,8 @@ def compute_b_mp(sigma, q, lmbd, verbose=False):
   b_fn = lambda z: ((mu0(z) / mu(z)) ** lmbd_int -
                     (mu(-z) / mu0(z)) ** lmbd_int)
   if verbose:
-    print "M =", m
-    print "f(-M) = {} f(M) = {}".format(b_fn(-m), b_fn(m))
+    print("M =", m)
+    print("f(-M) = {} f(M) = {}".format(b_fn(-m), b_fn(m)))
     assert b_fn(-m) < 0 and b_fn(m) < 0
 
   b_lambda_int1_fn = lambda z: mu0(z) * (mu0(z) / mu(z)) ** lmbd_int
@@ -223,8 +226,8 @@ def compute_b_mp(sigma, q, lmbd, verbose=False):
   b_bound = a_lambda_m1 + b_int1 - b_int2
 
   if verbose:
-    print "B by numerical integration", b_lambda
-    print "B must be no more than    ", b_bound
+    print("B by numerical integration", b_lambda)
+    print("B must be no more than    ", b_bound)
   assert b_lambda < b_bound + 1e-5
   return _to_np_float64(b_lambda)
 

research/domain_adaptation/domain_separation/dsn_eval.py

@@ -19,6 +19,7 @@
 import math
 
 import numpy as np
+from six.moves import xrange
 import tensorflow as tf
 
 from domain_adaptation.datasets import dataset_factory

research/gan/cifar/util.py

@@ -18,6 +18,7 @@ from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
 
+from six.moves import xrange
 import tensorflow as tf
 tfgan = tf.contrib.gan
 

research/gan/image_compression/networks_test.py

@@ -19,6 +19,7 @@ from __future__ import division
 from __future__ import print_function
 
 import tensorflow as tf
+from six.moves import xrange
 import networks
 
 

research/gan/mnist/util.py

@@ -24,6 +24,7 @@ from __future__ import print_function
 
 
 import numpy as np
+from six.moves import xrange
 import tensorflow as tf
 
 ds = tf.contrib.distributions

research/gan/mnist_estimator/train.py

@@ -22,6 +22,7 @@ import os
 
 import numpy as np
 import scipy.misc
+from six.moves import xrange
 import tensorflow as tf
 
 from mnist import data_provider

research/im2txt/README.md

@@ -119,8 +119,8 @@ First ensure that you have installed the following required packages:
 * **NumPy** ([instructions](http://www.scipy.org/install.html))
 * **Natural Language Toolkit (NLTK)**:
     * First install NLTK ([instructions](http://www.nltk.org/install.html))
-    * Then install the NLTK data ([instructions](http://www.nltk.org/data.html))
-
+    * Then install the NLTK data package "punkt" ([instructions](http://www.nltk.org/data.html))
+* **Unzip**
 ### Prepare the Training Data
 
 To train the model you will need to provide training data in native TFRecord
@@ -145,7 +145,7 @@ available space for storing the downloaded and processed data.
 MSCOCO_DIR="${HOME}/im2txt/data/mscoco"
 
 # Build the preprocessing script.
-cd tensorflow-models/im2txt
+cd research/im2txt
 bazel build //im2txt:download_and_preprocess_mscoco
 
 # Run the preprocessing script.
@@ -212,7 +212,7 @@ INCEPTION_CHECKPOINT="${HOME}/im2txt/data/inception_v3.ckpt"
 MODEL_DIR="${HOME}/im2txt/model"
 
 # Build the model.
-cd tensorflow-models/im2txt
+cd research/im2txt
 bazel build -c opt //im2txt/...
 
 # Run the training script.
@@ -306,7 +306,7 @@ VOCAB_FILE="${HOME}/im2txt/data/mscoco/word_counts.txt"
 IMAGE_FILE="${HOME}/im2txt/data/mscoco/raw-data/val2014/COCO_val2014_000000224477.jpg"
 
 # Build the inference binary.
-cd tensorflow-models/im2txt
+cd research/im2txt
 bazel build -c opt //im2txt:run_inference
 
 # Ignore GPU devices (only necessary if your GPU is currently memory

research/im2txt/im2txt/data/build_mscoco_data.py

@@ -97,6 +97,7 @@ import threading
 
 import nltk.tokenize
 import numpy as np
+from six.moves import xrange
 import tensorflow as tf
 
 tf.flags.DEFINE_string("train_image_dir", "/tmp/train2014/",

research/im2txt/im2txt/evaluate.py

@@ -76,7 +76,7 @@ def evaluate_model(sess, model, global_step, summary_writer, summary_op):
   start_time = time.time()
   sum_losses = 0.
   sum_weights = 0.
-  for i in xrange(num_eval_batches):
+  for i in range(num_eval_batches):
     cross_entropy_losses, weights = sess.run([
         model.target_cross_entropy_losses,
         model.target_cross_entropy_loss_weights
@@ -143,7 +143,7 @@ def run_once(model, saver, summary_writer, summary_op):
           global_step=global_step,
           summary_writer=summary_writer,
           summary_op=summary_op)
-    except Exception, e:  # pylint: disable=broad-except
+    except Exception as e:  # pylint: disable=broad-except
       tf.logging.error("Evaluation failed.")
       coord.request_stop(e)
 

research/im2txt/im2txt/run_inference.py

@@ -70,7 +70,7 @@ def main(_):
     generator = caption_generator.CaptionGenerator(model, vocab)
 
     for filename in filenames:
-      with tf.gfile.GFile(filename, "r") as f:
+      with tf.gfile.GFile(filename, "rb") as f:
         image = f.read()
       captions = generator.beam_search(sess, image)
       print("Captions for image %s:" % os.path.basename(filename))

research/learned_optimizer/metaopt.py

@@ -21,6 +21,7 @@ import sys
 import time
 
 import numpy as np
+from six.moves import xrange
 import tensorflow as tf
 
 from learned_optimizer.optimizer import trainable_optimizer

research/learning_to_remember_rare_events/data_utils.py

@@ -29,6 +29,7 @@ import numpy as np
 from scipy.misc import imresize
 from scipy.misc import imrotate
 from scipy.ndimage import imread
+from six.moves import xrange
 import tensorflow as tf
 
 

research/learning_to_remember_rare_events/memory.py

@@ -23,6 +23,7 @@ published as a conference paper at ICLR 2017.
 """
 
 import numpy as np
+from six.moves import xrange
 import tensorflow as tf
 
 

research/learning_to_remember_rare_events/train.py

@@ -26,6 +26,7 @@ import os
 import random
 
 import numpy as np
+from six.moves import xrange
 import tensorflow as tf
 
 import data_utils

research/lfads/synth_data/generate_itb_data.py

@@ -18,6 +18,7 @@ from __future__ import print_function
 import h5py
 import numpy as np
 import os
+from six.moves import xrange
 import tensorflow as tf
 
 from utils import write_datasets
@@ -47,12 +48,12 @@ flags.DEFINE_float("max_firing_rate", 30.0,
 flags.DEFINE_float("u_std", 0.25,
                    "Std dev of input to integration to bound model")
 flags.DEFINE_string("checkpoint_path", "SAMPLE_CHECKPOINT",
-                    """Path to directory with checkpoints of model 
+                    """Path to directory with checkpoints of model
                     trained on integration to bound task. Currently this
                     is a placeholder which tells the code to grab the
-                    checkpoint that is provided with the code 
-                    (in /trained_itb/..). If you have your own checkpoint 
-                    you would like to restore, you would point it to 
+                    checkpoint that is provided with the code
+                    (in /trained_itb/..). If you have your own checkpoint
+                    you would like to restore, you would point it to
                     that path.""")
 FLAGS = flags.FLAGS
 

research/lfads/synth_data/generate_labeled_rnn_data.py

@@ -18,6 +18,7 @@ from __future__ import print_function
 import os
 import h5py
 import numpy as np
+from six.moves import xrange
 
 from synthetic_data_utils import generate_data, generate_rnn
 from synthetic_data_utils import get_train_n_valid_inds