Fix lint issues introduced by my pull from GitHub.

Change: 149985352

tensorflow/contrib/layers/python/layers/layers_test.py

@@ -1709,7 +1709,7 @@ class BatchNormTest(test.TestCase):
     with self.test_session():
       reg = lambda x: 0.1 * math_ops.reduce_sum(x)
       images = np.random.uniform(size=(5, height, width, 3)).astype('f')
-      output = _layers.batch_norm(images, param_regularizers={'beta': reg})
+      _layers.batch_norm(images, param_regularizers={'beta': reg})
       self.assertEqual(
           len(ops.get_collection(ops.GraphKeys.REGULARIZATION_LOSSES)), 1)
       beta_decay = ops.get_collection(ops.GraphKeys.REGULARIZATION_LOSSES)[0]
@@ -1720,7 +1720,7 @@ class BatchNormTest(test.TestCase):
     with self.test_session():
       reg = lambda x: 0.1 * math_ops.reduce_sum(x)
       images = np.random.uniform(size=(5, height, width, 3)).astype('f')
-      output = _layers.batch_norm(
+      _layers.batch_norm(
           images, param_regularizers={'gamma': reg}, scale=True)
       self.assertEqual(
           len(ops.get_collection(ops.GraphKeys.REGULARIZATION_LOSSES)), 1)

tensorflow/contrib/slim/python/slim/learning.py

@@ -18,9 +18,9 @@ This script contains various functions for training models. These include
 manipulating gradients, creating a `train_op` (an operation that computes the
 loss and applies the gradients) and a training loop function. The training loop
 allows the user to pass in the `train_op` and runs the optimization according
-to user-specified arguments. Note that the training loop uses the tf.train.Supervisor
-and its managed_session in its implementation to ensure the ability of worker
-processes to recover from failures.
+to user-specified arguments. Note that the training loop uses the
+tf.train.Supervisor and its managed_session in its implementation to ensure the
+ability of worker processes to recover from failures.
 
 ************************************
 * A simple working training script *

tensorflow/examples/learn/iris.py

@@ -16,8 +16,11 @@
 from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
-from sklearn import model_selection
+
+
 from sklearn import metrics
+from sklearn import model_selection
+
 import tensorflow as tf
 
 

tensorflow/python/framework/test_util.py

@@ -34,6 +34,7 @@ from google.protobuf import text_format
 from tensorflow.core.framework import graph_pb2
 from tensorflow.core.protobuf import config_pb2
 from tensorflow.python import pywrap_tensorflow
+from tensorflow.python.client import device_lib
 from tensorflow.python.client import session
 from tensorflow.python.framework import device as pydev
 from tensorflow.python.framework import errors
@@ -44,7 +45,6 @@ from tensorflow.python.platform import googletest
 from tensorflow.python.platform import tf_logging as logging
 from tensorflow.python.util import compat
 from tensorflow.python.util.protobuf import compare
-from tensorflow.python.client import device_lib
 
 
 def gpu_device_name():
@@ -54,6 +54,7 @@ def gpu_device_name():
       return x.name
   return ""
 
+
 def assert_ops_in_graph(expected_ops, graph):
   """Assert all expected operations are found.
 

tensorflow/python/kernel_tests/matmul_op_test.py

@@ -18,8 +18,8 @@ from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
 
-import numpy as np
 import operator
+import numpy as np
 
 from tensorflow.python.framework import constant_op
 from tensorflow.python.framework import ops
@@ -157,9 +157,10 @@ try:
   # @ operator supported since python 3.5.
   infix_matmul = operator.matmul
 except AttributeError:
+
   # For earlier versions of python, emulate regular behavior.
   # Useful to build and test for 3.5+ on earlier versions.
-  def infix_matmul(x, y):
+  def infix_matmul(x, y):  # pylint: disable=invalid-name
     try:
       r = type(x).__matmul__(x, y)
     except AttributeError:

tensorflow/python/kernel_tests/variable_scope_test.py

@@ -723,11 +723,11 @@ class VariableScopeTest(test.TestCase):
 
   def testGetCollection(self):
     with self.test_session():
-      a = variable_scope.get_variable("a", [])
-      b = variable_scope.get_variable("b", [], trainable=False)
+      _ = variable_scope.get_variable("a", [])
+      _ = variable_scope.get_variable("b", [], trainable=False)
       with variable_scope.variable_scope("foo_") as scope1:
-        a = variable_scope.get_variable("a", [])
-        b = variable_scope.get_variable("b", [], trainable=False)
+        _ = variable_scope.get_variable("a", [])
+        _ = variable_scope.get_variable("b", [], trainable=False)
         self.assertEqual([
             v.name
             for v in scope1.get_collection(ops.GraphKeys.TRAINABLE_VARIABLES)
@@ -737,8 +737,8 @@ class VariableScopeTest(test.TestCase):
             for v in scope1.get_collection(ops.GraphKeys.GLOBAL_VARIABLES)
         ], ["foo_/a:0", "foo_/b:0"])
       with variable_scope.variable_scope("foo") as scope2:
-        a = variable_scope.get_variable("a", [])
-        b = variable_scope.get_variable("b", [], trainable=False)
+        _ = variable_scope.get_variable("a", [])
+        _ = variable_scope.get_variable("b", [], trainable=False)
         self.assertEqual([
             v.name
             for v in scope2.get_collection(ops.GraphKeys.TRAINABLE_VARIABLES)
@@ -758,21 +758,22 @@ class VariableScopeTest(test.TestCase):
 
   def testGetTrainableVariables(self):
     with self.test_session():
-      a = variable_scope.get_variable("a", [])
+      _ = variable_scope.get_variable("a", [])
       with variable_scope.variable_scope("foo") as scope:
-        b = variable_scope.get_variable("b", [])
-        c = variable_scope.get_variable("c", [], trainable=False)
+        _ = variable_scope.get_variable("b", [])
+        _ = variable_scope.get_variable("c", [], trainable=False)
         self.assertEqual([v.name
                           for v in scope.trainable_variables()], ["foo/b:0"])
 
   def testGetGlobalVariables(self):
     with self.test_session():
-      a = variable_scope.get_variable("a", [])
+      _ = variable_scope.get_variable("a", [])
       with variable_scope.variable_scope("foo") as scope:
-        b = variable_scope.get_variable("b", [])
+        _ = variable_scope.get_variable("b", [])
         self.assertEqual([v.name
                           for v in scope.global_variables()], ["foo/b:0"])
 
+
 def axis0_into1_partitioner(shape=None, **unused_kwargs):
   part = [1] * len(shape)
   return part

tensorflow/python/training/monitored_session.py

@@ -67,8 +67,8 @@ class Scaffold(object):
   The following pieces are directly accessible as attributes of the `Scaffold`
   object:
 
-  * `saver`: A `tf.train.Saver` object taking care of saving the variables.  Picked
-    from and stored into the `SAVERS` collection in the graph by default.
+  * `saver`: A `tf.train.Saver` object taking care of saving the variables.
+    Picked from and stored into the `SAVERS` collection in the graph by default.
   * `init_op`: An op to run to initialize the variables.  Picked from and
     stored into the `INIT_OP` collection in the graph by default.
   * `ready_op`: An op to verify that the variables are initialized.  Picked
@@ -124,7 +124,8 @@ class Scaffold(object):
       local_init_op: Optional op to initialize local variables.
       summary_op: Optional op to gather all summaries.  Must return a scalar
         string tensor containing a serialized `Summary` proto.
-      saver: Optional `tf.train.Saver` object to use to save and restore variables.
+      saver: Optional `tf.train.Saver` object to use to save and restore
+        variables.
     """
 
     # NOTE(touts): modifying the init function to be passed the scaffold is a