diff --git a/tensorflow/tools/api/lib/BUILD b/tensorflow/tools/api/lib/BUILD
index 75cb9338d4a..ab48e6970b2 100644
--- a/tensorflow/tools/api/lib/BUILD
+++ b/tensorflow/tools/api/lib/BUILD
@@ -23,5 +23,6 @@ py_library(
         ":api_objects_proto_py",
         "//tensorflow/python:platform",
         "//tensorflow/python:util",
+        "@six_archive//:six",
     ],
 )
diff --git a/tensorflow/tools/api/lib/python_object_to_proto_visitor.py b/tensorflow/tools/api/lib/python_object_to_proto_visitor.py
index 549a11588bb..283f53882c3 100644
--- a/tensorflow/tools/api/lib/python_object_to_proto_visitor.py
+++ b/tensorflow/tools/api/lib/python_object_to_proto_visitor.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -19,8 +20,11 @@ from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
 
-import sys
 import enum
+import sys
+
+import six
+
 from google.protobuf import message
 from tensorflow.python.platform import tf_logging as logging
 from tensorflow.python.util import deprecation
@@ -191,7 +195,8 @@ class PythonObjectToProtoVisitor(object):
       if (_SkipMember(parent, member_name) or
           isinstance(member_obj, deprecation.HiddenTfApiAttribute)):
         return
-      if member_name == '__init__' or not member_name.startswith('_'):
+      if member_name == '__init__' or not six.ensure_str(
+          member_name).startswith('_'):
         if tf_inspect.isroutine(member_obj):
           new_method = proto.member_method.add()
           new_method.name = member_name
diff --git a/tensorflow/tools/common/BUILD b/tensorflow/tools/common/BUILD
index dc21877f060..5cc9fa67e6d 100644
--- a/tensorflow/tools/common/BUILD
+++ b/tensorflow/tools/common/BUILD
@@ -14,13 +14,16 @@ py_library(
     name = "public_api",
     srcs = ["public_api.py"],
     srcs_version = "PY2AND3",
-    deps = ["//tensorflow/python:util"],
+    deps = [
+        "//tensorflow/python:util",
+        "@six_archive//:six",
+    ],
 )
 
 py_test(
     name = "public_api_test",
     srcs = ["public_api_test.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":public_api",
@@ -32,13 +35,16 @@ py_library(
     name = "traverse",
     srcs = ["traverse.py"],
     srcs_version = "PY2AND3",
-    deps = ["//tensorflow/python:util"],
+    deps = [
+        "//tensorflow/python:util",
+        "@six_archive//:six",
+    ],
 )
 
 py_test(
     name = "traverse_test",
     srcs = ["traverse_test.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":test_module1",
diff --git a/tensorflow/tools/common/public_api.py b/tensorflow/tools/common/public_api.py
index f62f9bb54f1..0f788290b26 100644
--- a/tensorflow/tools/common/public_api.py
+++ b/tensorflow/tools/common/public_api.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -20,6 +21,8 @@ from __future__ import print_function
 
 import re
 
+import six
+
 from tensorflow.python.util import tf_inspect
 
 
@@ -108,9 +111,9 @@ class PublicAPIVisitor(object):
     """Return whether a name is private."""
     # TODO(wicke): Find out what names to exclude.
     del obj  # Unused.
-    return ((path in self._private_map and
-             name in self._private_map[path]) or
-            (name.startswith('_') and not re.match('__.*__$', name) or
+    return ((path in self._private_map and name in self._private_map[path]) or
+            (six.ensure_str(name).startswith('_') and
+             not re.match('__.*__$', six.ensure_str(name)) or
              name in ['__base__', '__class__']))
 
   def _do_not_descend(self, path, name):
@@ -122,7 +125,8 @@ class PublicAPIVisitor(object):
     """Visitor interface, see `traverse` for details."""
 
     # Avoid long waits in cases of pretty unambiguous failure.
-    if tf_inspect.ismodule(parent) and len(path.split('.')) > 10:
+    if tf_inspect.ismodule(parent) and len(
+        six.ensure_str(path).split('.')) > 10:
       raise RuntimeError('Modules nested too deep:\n%s.%s\n\nThis is likely a '
                          'problem with an accidental public import.' %
                          (self._root_name, path))
diff --git a/tensorflow/tools/common/traverse.py b/tensorflow/tools/common/traverse.py
index b121a87062f..5efce450dcb 100644
--- a/tensorflow/tools/common/traverse.py
+++ b/tensorflow/tools/common/traverse.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -21,6 +22,8 @@ from __future__ import print_function
 import enum
 import sys
 
+import six
+
 from tensorflow.python.util import tf_inspect
 
 __all__ = ['traverse']
@@ -59,7 +62,8 @@ def _traverse_internal(root, visit, stack, path):
     if any(child is item for item in new_stack):  # `in`, but using `is`
       continue
 
-    child_path = path + '.' + name if path else name
+    child_path = six.ensure_str(path) + '.' + six.ensure_str(
+        name) if path else name
     _traverse_internal(child, visit, new_stack, child_path)
 
 
diff --git a/tensorflow/tools/compatibility/BUILD b/tensorflow/tools/compatibility/BUILD
index 5a50d77b010..d6d882b5749 100644
--- a/tensorflow/tools/compatibility/BUILD
+++ b/tensorflow/tools/compatibility/BUILD
@@ -14,6 +14,7 @@ py_library(
     name = "ipynb",
     srcs = ["ipynb.py"],
     srcs_version = "PY2AND3",
+    deps = ["@six_archive//:six"],
 )
 
 py_library(
@@ -29,7 +30,7 @@ py_library(
 py_test(
     name = "ast_edits_test",
     srcs = ["ast_edits_test.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":ast_edits",
@@ -42,22 +43,28 @@ py_test(
 py_binary(
     name = "tf_upgrade",
     srcs = ["tf_upgrade.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
-    deps = [":tf_upgrade_lib"],
+    deps = [
+        ":tf_upgrade_lib",
+        "@six_archive//:six",
+    ],
 )
 
 py_library(
     name = "tf_upgrade_lib",
     srcs = ["tf_upgrade.py"],
     srcs_version = "PY2AND3",
-    deps = [":ast_edits"],
+    deps = [
+        ":ast_edits",
+        "@six_archive//:six",
+    ],
 )
 
 py_test(
     name = "tf_upgrade_test",
     srcs = ["tf_upgrade_test.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     tags = [
         "no_pip",
@@ -96,6 +103,7 @@ py_library(
 py_test(
     name = "all_renames_v2_test",
     srcs = ["all_renames_v2_test.py"],
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":all_renames_v2",
@@ -108,6 +116,7 @@ py_test(
 py_library(
     name = "module_deprecations_v2",
     srcs = ["module_deprecations_v2.py"],
+    srcs_version = "PY2AND3",
     deps = [":ast_edits"],
 )
 
@@ -145,13 +154,14 @@ py_binary(
         ":ipynb",
         ":tf_upgrade_v2_lib",
         ":tf_upgrade_v2_safety_lib",
+        "@six_archive//:six",
     ],
 )
 
 py_test(
     name = "tf_upgrade_v2_test",
     srcs = ["tf_upgrade_v2_test.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     tags = ["v1only"],
     deps = [
@@ -169,6 +179,7 @@ py_test(
 py_test(
     name = "tf_upgrade_v2_safety_test",
     srcs = ["tf_upgrade_v2_safety_test.py"],
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":tf_upgrade_v2_safety_lib",
@@ -208,7 +219,7 @@ py_test(
     name = "test_file_v1_0",
     size = "small",
     srcs = ["test_file_v1_0.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         "//tensorflow:tensorflow_py",
@@ -235,7 +246,7 @@ py_test(
     name = "test_file_v1_12",
     size = "small",
     srcs = ["testdata/test_file_v1_12.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     tags = ["v1only"],
     deps = [
@@ -247,7 +258,7 @@ py_test(
     name = "test_file_v2_0",
     size = "small",
     srcs = ["test_file_v2_0.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         "//tensorflow:tensorflow_py",
diff --git a/tensorflow/tools/compatibility/ast_edits.py b/tensorflow/tools/compatibility/ast_edits.py
index d6129da71dd..71fb2aee770 100644
--- a/tensorflow/tools/compatibility/ast_edits.py
+++ b/tensorflow/tools/compatibility/ast_edits.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2016 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -29,6 +30,7 @@ import traceback
 
 import pasta
 import six
+from six.moves import range
 
 # Some regular expressions we will need for parsing
 FIND_OPEN = re.compile(r"^\s*(\[).*$")
@@ -56,7 +58,7 @@ def full_name_node(name, ctx=ast.Load()):
   Returns:
     A Name or Attribute node.
   """
-  names = name.split(".")
+  names = six.ensure_str(name).split(".")
   names.reverse()
   node = ast.Name(id=names.pop(), ctx=ast.Load())
   while names:
@@ -301,7 +303,7 @@ class _PastaEditVisitor(ast.NodeVisitor):
     function_transformers = getattr(self._api_change_spec,
                                     transformer_field, {})
 
-    glob_name = "*." + name if name else None
+    glob_name = "*." + six.ensure_str(name) if name else None
     transformers = []
     if full_name in function_transformers:
       transformers.append(function_transformers[full_name])
@@ -318,7 +320,7 @@ class _PastaEditVisitor(ast.NodeVisitor):
     function_transformers = getattr(self._api_change_spec,
                                     transformer_field, {})
 
-    glob_name = "*." + name if name else None
+    glob_name = "*." + six.ensure_str(name) if name else None
     transformers = function_transformers.get("*", {}).copy()
     transformers.update(function_transformers.get(glob_name, {}))
     transformers.update(function_transformers.get(full_name, {}))
@@ -351,7 +353,7 @@ class _PastaEditVisitor(ast.NodeVisitor):
     function_warnings = self._api_change_spec.function_warnings
     if full_name in function_warnings:
       level, message = function_warnings[full_name]
-      message = message.replace("<function name>", full_name)
+      message = six.ensure_str(message).replace("<function name>", full_name)
       self.add_log(level, node.lineno, node.col_offset,
                    "%s requires manual check. %s" % (full_name, message))
       return True
@@ -363,7 +365,8 @@ class _PastaEditVisitor(ast.NodeVisitor):
     warnings = self._api_change_spec.module_deprecations
     if full_name in warnings:
       level, message = warnings[full_name]
-      message = message.replace("<function name>", whole_name)
+      message = six.ensure_str(message).replace("<function name>",
+                                                six.ensure_str(whole_name))
       self.add_log(level, node.lineno, node.col_offset,
                    "Using member %s in deprecated module %s. %s" % (whole_name,
                                                                     full_name,
@@ -394,7 +397,7 @@ class _PastaEditVisitor(ast.NodeVisitor):
     # an attribute.
     warned = False
     if isinstance(node.func, ast.Attribute):
-      warned = self._maybe_add_warning(node, "*." + name)
+      warned = self._maybe_add_warning(node, "*." + six.ensure_str(name))
 
     # All arg warnings are handled here, since only we have the args
     arg_warnings = self._get_applicable_dict("function_arg_warnings",
@@ -406,7 +409,8 @@ class _PastaEditVisitor(ast.NodeVisitor):
       present, _ = get_arg_value(node, kwarg, arg) or variadic_args
       if present:
         warned = True
-        warning_message = warning.replace("<function name>", full_name or name)
+        warning_message = six.ensure_str(warning).replace(
+            "<function name>", six.ensure_str(full_name or name))
         template = "%s called with %s argument, requires manual check: %s"
         if variadic_args:
           template = ("%s called with *args or **kwargs that may include %s, "
@@ -625,13 +629,13 @@ class _PastaEditVisitor(ast.NodeVisitor):
     # This loop processes imports in the format
     # import foo as f, bar as b
     for import_alias in node.names:
-      all_import_components = import_alias.name.split(".")
+      all_import_components = six.ensure_str(import_alias.name).split(".")
       # Look for rename, starting with longest import levels.
       found_update = False
-      for i in reversed(range(1, max_submodule_depth + 1)):
+      for i in reversed(list(range(1, max_submodule_depth + 1))):
         import_component = all_import_components[0]
         for j in range(1, min(i, len(all_import_components))):
-          import_component += "." + all_import_components[j]
+          import_component += "." + six.ensure_str(all_import_components[j])
         import_rename_spec = import_renames.get(import_component, None)
 
         if not import_rename_spec or excluded_from_module_rename(
@@ -674,7 +678,8 @@ class _PastaEditVisitor(ast.NodeVisitor):
           if old_suffix is None:
             old_suffix = os.linesep
           if os.linesep not in old_suffix:
-            pasta.base.formatting.set(node, "suffix", old_suffix + os.linesep)
+            pasta.base.formatting.set(node, "suffix",
+                                      six.ensure_str(old_suffix) + os.linesep)
 
           # Apply indentation to new node.
           pasta.base.formatting.set(new_line_node, "prefix",
@@ -720,7 +725,7 @@ class _PastaEditVisitor(ast.NodeVisitor):
 
     # Look for rename based on first component of from-import.
     # i.e. based on foo in foo.bar.
-    from_import_first_component = from_import.split(".")[0]
+    from_import_first_component = six.ensure_str(from_import).split(".")[0]
     import_renames = getattr(self._api_change_spec, "import_renames", {})
     import_rename_spec = import_renames.get(from_import_first_component, None)
     if not import_rename_spec:
@@ -918,7 +923,7 @@ class ASTCodeUpgrader(object):
   def format_log(self, log, in_filename):
     log_string = "%d:%d: %s: %s" % (log[1], log[2], log[0], log[3])
     if in_filename:
-      return in_filename + ":" + log_string
+      return six.ensure_str(in_filename) + ":" + log_string
     else:
       return log_string
 
@@ -945,12 +950,12 @@ class ASTCodeUpgrader(object):
     return 1, pasta.dump(t), logs, errors
 
   def _format_log(self, log, in_filename, out_filename):
-    text = "-" * 80 + "\n"
+    text = six.ensure_str("-" * 80) + "\n"
     text += "Processing file %r\n outputting to %r\n" % (in_filename,
                                                          out_filename)
-    text += "-" * 80 + "\n\n"
+    text += six.ensure_str("-" * 80) + "\n\n"
     text += "\n".join(log) + "\n"
-    text += "-" * 80 + "\n\n"
+    text += six.ensure_str("-" * 80) + "\n\n"
     return text
 
   def process_opened_file(self, in_filename, in_file, out_filename, out_file):
@@ -1017,8 +1022,10 @@ class ASTCodeUpgrader(object):
     files_to_process = []
     files_to_copy = []
     for dir_name, _, file_list in os.walk(root_directory):
-      py_files = [f for f in file_list if f.endswith(".py")]
-      copy_files = [f for f in file_list if not f.endswith(".py")]
+      py_files = [f for f in file_list if six.ensure_str(f).endswith(".py")]
+      copy_files = [
+          f for f in file_list if not six.ensure_str(f).endswith(".py")
+      ]
       for filename in py_files:
         fullpath = os.path.join(dir_name, filename)
         fullpath_output = os.path.join(output_root_directory,
@@ -1036,9 +1043,9 @@ class ASTCodeUpgrader(object):
     file_count = 0
     tree_errors = {}
     report = ""
-    report += ("=" * 80) + "\n"
+    report += six.ensure_str(("=" * 80)) + "\n"
     report += "Input tree: %r\n" % root_directory
-    report += ("=" * 80) + "\n"
+    report += six.ensure_str(("=" * 80)) + "\n"
 
     for input_path, output_path in files_to_process:
       output_directory = os.path.dirname(output_path)
@@ -1074,16 +1081,19 @@ class ASTCodeUpgrader(object):
     """Process a directory of python files in place."""
     files_to_process = []
     for dir_name, _, file_list in os.walk(root_directory):
-      py_files = [os.path.join(dir_name,
-                               f) for f in file_list if f.endswith(".py")]
+      py_files = [
+          os.path.join(dir_name, f)
+          for f in file_list
+          if six.ensure_str(f).endswith(".py")
+      ]
       files_to_process += py_files
 
     file_count = 0
     tree_errors = {}
     report = ""
-    report += ("=" * 80) + "\n"
+    report += six.ensure_str(("=" * 80)) + "\n"
     report += "Input tree: %r\n" % root_directory
-    report += ("=" * 80) + "\n"
+    report += six.ensure_str(("=" * 80)) + "\n"
 
     for path in files_to_process:
       if os.path.islink(path):
diff --git a/tensorflow/tools/compatibility/ipynb.py b/tensorflow/tools/compatibility/ipynb.py
index fed5f0f2bfc..77c8fedf709 100644
--- a/tensorflow/tools/compatibility/ipynb.py
+++ b/tensorflow/tools/compatibility/ipynb.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2019 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -24,6 +25,7 @@ import json
 import re
 import shutil
 import tempfile
+import six
 
 CodeLine = collections.namedtuple("CodeLine", ["cell_number", "code"])
 
@@ -31,7 +33,8 @@ def is_python(cell):
   """Checks if the cell consists of Python code."""
   return (cell["cell_type"] == "code"  # code cells only
           and cell["source"]  # non-empty cells
-          and not cell["source"][0].startswith("%%"))  # multiline eg: %%bash
+          and not six.ensure_str(cell["source"][0]).startswith("%%")
+         )  # multiline eg: %%bash
 
 
 def process_file(in_filename, out_filename, upgrader):
@@ -47,8 +50,9 @@ def process_file(in_filename, out_filename, upgrader):
         upgrader.update_string_pasta("\n".join(raw_lines), in_filename))
 
     if temp_file and processed_file:
-      new_notebook = _update_notebook(notebook, raw_code,
-                                      new_file_content.split("\n"))
+      new_notebook = _update_notebook(
+          notebook, raw_code,
+          six.ensure_str(new_file_content).split("\n"))
       json.dump(new_notebook, temp_file)
     else:
       raise SyntaxError(
@@ -78,7 +82,7 @@ def skip_magic(code_line, magic_list):
   """
 
   for magic in magic_list:
-    if code_line.startswith(magic):
+    if six.ensure_str(code_line).startswith(magic):
       return True
 
   return False
@@ -120,7 +124,7 @@ def _get_code(input_file):
         # Idea is to comment these lines, for upgrade time
         if skip_magic(code_line, ["%", "!", "?"]) or is_line_split:
           # Found a special character, need to "encode"
-          code_line = "###!!!" + code_line
+          code_line = "###!!!" + six.ensure_str(code_line)
 
           # if this cell ends with `\` -> skip the next line
           is_line_split = check_line_split(code_line)
@@ -131,14 +135,16 @@ def _get_code(input_file):
         # Sometimes, people leave \n at the end of cell
         # in order to migrate only related things, and make the diff
         # the smallest -> here is another hack
-        if (line_idx == len(cell_lines) - 1) and code_line.endswith("\n"):
-          code_line = code_line.replace("\n", "###===")
+        if (line_idx == len(cell_lines) -
+            1) and six.ensure_str(code_line).endswith("\n"):
+          code_line = six.ensure_str(code_line).replace("\n", "###===")
 
         # sometimes a line would start with `\n` and content after
         # that's the hack for this
         raw_code.append(
             CodeLine(cell_index,
-                     code_line.rstrip().replace("\n", "###===")))
+                     six.ensure_str(code_line.rstrip()).replace("\n",
+                                                                "###===")))
 
       cell_index += 1
 
diff --git a/tensorflow/tools/compatibility/tf_upgrade.py b/tensorflow/tools/compatibility/tf_upgrade.py
index 5dd548c8214..988d11b1016 100644
--- a/tensorflow/tools/compatibility/tf_upgrade.py
+++ b/tensorflow/tools/compatibility/tf_upgrade.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2016 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -20,6 +21,8 @@ from __future__ import print_function
 
 import argparse
 
+import six
+
 from tensorflow.tools.compatibility import ast_edits
 
 
@@ -245,7 +248,7 @@ Simple usage:
   else:
     parser.print_help()
   if report_text:
-    open(report_filename, "w").write(report_text)
+    open(report_filename, "w").write(six.ensure_str(report_text))
     print("TensorFlow 1.0 Upgrade Script")
     print("-----------------------------")
     print("Converted %d files\n" % files_processed)
diff --git a/tensorflow/tools/compatibility/tf_upgrade_test.py b/tensorflow/tools/compatibility/tf_upgrade_test.py
index cf05575a9dd..ca0e80564ff 100644
--- a/tensorflow/tools/compatibility/tf_upgrade_test.py
+++ b/tensorflow/tools/compatibility/tf_upgrade_test.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2016 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -46,14 +47,16 @@ class TestUpgrade(test_util.TensorFlowTestCase):
   def testParseError(self):
     _, report, unused_errors, unused_new_text = self._upgrade(
         "import tensorflow as tf\na + \n")
-    self.assertTrue(report.find("Failed to parse") != -1)
+    self.assertNotEqual(six.ensure_str(report).find("Failed to parse"), -1)
 
   def testReport(self):
     text = "tf.mul(a, b)\n"
     _, report, unused_errors, unused_new_text = self._upgrade(text)
     # This is not a complete test, but it is a sanity test that a report
     # is generating information.
-    self.assertTrue(report.find("Renamed function `tf.mul` to `tf.multiply`"))
+    self.assertTrue(
+        six.ensure_str(report).find(
+            "Renamed function `tf.mul` to `tf.multiply`"))
 
   def testRename(self):
     text = "tf.mul(a, tf.sub(b, c))\n"
diff --git a/tensorflow/tools/compatibility/tf_upgrade_v2.py b/tensorflow/tools/compatibility/tf_upgrade_v2.py
index 221353d87cd..2bd5bb984c8 100644
--- a/tensorflow/tools/compatibility/tf_upgrade_v2.py
+++ b/tensorflow/tools/compatibility/tf_upgrade_v2.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2018 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -24,6 +25,7 @@ import functools
 import sys
 
 import pasta
+import six
 
 from tensorflow.tools.compatibility import all_renames_v2
 from tensorflow.tools.compatibility import ast_edits
@@ -47,8 +49,9 @@ class VersionedTFImport(ast_edits.AnalysisResult):
 
   def __init__(self, version):
     self.log_level = ast_edits.INFO
-    self.log_message = ("Not upgrading symbols because `tensorflow." + version
-                        + "` was directly imported as `tf`.")
+    self.log_message = ("Not upgrading symbols because `tensorflow." +
+                        six.ensure_str(version) +
+                        "` was directly imported as `tf`.")
 
 
 class TFAPIImportAnalysisSpec(ast_edits.APIAnalysisSpec):
@@ -1687,7 +1690,7 @@ def _rename_if_arg_found_transformer(parent, node, full_name, name, logs,
 
   # All conditions met, insert v1 and log what we did.
   # We must have a full name, so the func is an attribute.
-  new_name = full_name.replace("tf.", "tf.compat.v1.", 1)
+  new_name = six.ensure_str(full_name).replace("tf.", "tf.compat.v1.", 1)
   node.func = ast_edits.full_name_node(new_name)
   logs.append((
       ast_edits.INFO, node.lineno, node.col_offset,
@@ -1715,8 +1718,8 @@ def _iterator_transformer(parent, node, full_name, name, logs):
   # (tf.compat.v1.data), or something which is handled in the rename
   # (tf.data). This transformer only handles the method call to function call
   # conversion.
-  if full_name and (full_name.startswith("tf.compat.v1.data") or
-                    full_name.startswith("tf.data")):
+  if full_name and (six.ensure_str(full_name).startswith("tf.compat.v1.data") or
+                    six.ensure_str(full_name).startswith("tf.data")):
     return
 
   # This should never happen, since we're only called for Attribute nodes.
@@ -2460,7 +2463,7 @@ def _name_scope_transformer(parent, node, full_name, name, logs):
 
 
 def _rename_to_compat_v1(node, full_name, logs, reason):
-  new_name = full_name.replace("tf.", "tf.compat.v1.", 1)
+  new_name = six.ensure_str(full_name).replace("tf.", "tf.compat.v1.", 1)
   return _rename_func(node, full_name, new_name, logs, reason)
 
 
diff --git a/tensorflow/tools/compatibility/tf_upgrade_v2_main.py b/tensorflow/tools/compatibility/tf_upgrade_v2_main.py
index 7288e171d72..e41577926a5 100644
--- a/tensorflow/tools/compatibility/tf_upgrade_v2_main.py
+++ b/tensorflow/tools/compatibility/tf_upgrade_v2_main.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2018 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -20,10 +21,13 @@ from __future__ import print_function
 
 import argparse
 
+import six
+
 from tensorflow.tools.compatibility import ast_edits
 from tensorflow.tools.compatibility import ipynb
 from tensorflow.tools.compatibility import tf_upgrade_v2
 from tensorflow.tools.compatibility import tf_upgrade_v2_safety
+
 # Make straightforward changes to convert to 2.0. In harder cases,
 # use compat.v1.
 _DEFAULT_MODE = "DEFAULT"
@@ -35,10 +39,10 @@ _SAFETY_MODE = "SAFETY"
 def process_file(in_filename, out_filename, upgrader):
   """Process a file of type `.py` or `.ipynb`."""
 
-  if in_filename.endswith(".py"):
+  if six.ensure_str(in_filename).endswith(".py"):
     files_processed, report_text, errors = \
       upgrader.process_file(in_filename, out_filename)
-  elif in_filename.endswith(".ipynb"):
+  elif six.ensure_str(in_filename).endswith(".ipynb"):
     files_processed, report_text, errors = \
       ipynb.process_file(in_filename, out_filename, upgrader)
   else:
@@ -157,24 +161,24 @@ Simple usage:
     for f in errors:
       if errors[f]:
         num_errors += len(errors[f])
-        report.append("-" * 80 + "\n")
+        report.append(six.ensure_str("-" * 80) + "\n")
         report.append("File: %s\n" % f)
-        report.append("-" * 80 + "\n")
+        report.append(six.ensure_str("-" * 80) + "\n")
         report.append("\n".join(errors[f]) + "\n")
 
     report = ("TensorFlow 2.0 Upgrade Script\n"
               "-----------------------------\n"
               "Converted %d files\n" % files_processed +
               "Detected %d issues that require attention" % num_errors + "\n" +
-              "-" * 80 + "\n") + "".join(report)
-    detailed_report_header = "=" * 80 + "\n"
+              six.ensure_str("-" * 80) + "\n") + "".join(report)
+    detailed_report_header = six.ensure_str("=" * 80) + "\n"
     detailed_report_header += "Detailed log follows:\n\n"
-    detailed_report_header += "=" * 80 + "\n"
+    detailed_report_header += six.ensure_str("=" * 80) + "\n"
 
     with open(report_filename, "w") as report_file:
       report_file.write(report)
       report_file.write(detailed_report_header)
-      report_file.write(report_text)
+      report_file.write(six.ensure_str(report_text))
 
     if args.print_all:
       print(report)
diff --git a/tensorflow/tools/compatibility/tf_upgrade_v2_test.py b/tensorflow/tools/compatibility/tf_upgrade_v2_test.py
index 1249ce7cc2f..ca915109659 100644
--- a/tensorflow/tools/compatibility/tf_upgrade_v2_test.py
+++ b/tensorflow/tools/compatibility/tf_upgrade_v2_test.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2018 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -39,7 +40,7 @@ from tensorflow.tools.compatibility import tf_upgrade_v2
 
 
 def get_symbol_for_name(root, name):
-  name_parts = name.split(".")
+  name_parts = six.ensure_str(name).split(".")
   symbol = root
   # Iterate starting with second item since 1st item is "tf.".
   for part in name_parts[1:]:
@@ -66,12 +67,13 @@ def get_func_and_args_from_str(call_str):
   Returns:
     (function_name, list of arg names) tuple.
   """
-  open_paren_index = call_str.find("(")
+  open_paren_index = six.ensure_str(call_str).find("(")
   close_paren_index = call_str.rfind(")")
 
-  function_name = call_str[:call_str.find("(")]
-  args = call_str[open_paren_index+1:close_paren_index].split(",")
-  args = [arg.split("=")[0].strip() for arg in args]
+  function_name = call_str[:six.ensure_str(call_str).find("(")]
+  args = six.ensure_str(call_str[open_paren_index +
+                                 1:close_paren_index]).split(",")
+  args = [six.ensure_str(arg).split("=")[0].strip() for arg in args]
   args = [arg for arg in args if arg]  # filter out empty strings
   return function_name, args
 
@@ -96,7 +98,7 @@ class TestUpgrade(test_util.TensorFlowTestCase, parameterized.TestCase):
           _, attr = tf_decorator.unwrap(child[1])
           api_names_v2 = tf_export.get_v2_names(attr)
           for name in api_names_v2:
-            cls.v2_symbols["tf." + name] = attr
+            cls.v2_symbols["tf." + six.ensure_str(name)] = attr
 
       visitor = public_api.PublicAPIVisitor(symbol_collector)
       visitor.private_map["tf.compat"] = ["v1"]
@@ -109,7 +111,7 @@ class TestUpgrade(test_util.TensorFlowTestCase, parameterized.TestCase):
           _, attr = tf_decorator.unwrap(child[1])
           api_names_v1 = tf_export.get_v1_names(attr)
           for name in api_names_v1:
-            cls.v1_symbols["tf." + name] = attr
+            cls.v1_symbols["tf." + six.ensure_str(name)] = attr
 
       visitor = public_api.PublicAPIVisitor(symbol_collector_v1)
       traverse.traverse(tf.compat.v1, visitor)
@@ -138,15 +140,16 @@ class TestUpgrade(test_util.TensorFlowTestCase, parameterized.TestCase):
   def testParseError(self):
     _, report, unused_errors, unused_new_text = self._upgrade(
         "import tensorflow as tf\na + \n")
-    self.assertTrue(report.find("Failed to parse") != -1)
+    self.assertNotEqual(six.ensure_str(report).find("Failed to parse"), -1)
 
   def testReport(self):
     text = "tf.angle(a)\n"
     _, report, unused_errors, unused_new_text = self._upgrade(text)
     # This is not a complete test, but it is a sanity test that a report
     # is generating information.
-    self.assertTrue(report.find("Renamed function `tf.angle` to "
-                                "`tf.math.angle`"))
+    self.assertTrue(
+        six.ensure_str(report).find("Renamed function `tf.angle` to "
+                                    "`tf.math.angle`"))
 
   def testRename(self):
     text = "tf.conj(a)\n"
@@ -169,7 +172,7 @@ class TestUpgrade(test_util.TensorFlowTestCase, parameterized.TestCase):
         _, attr = tf_decorator.unwrap(child[1])
         api_names = tf_export.get_v1_names(attr)
         for name in api_names:
-          _, _, _, text = self._upgrade("tf." + name)
+          _, _, _, text = self._upgrade("tf." + six.ensure_str(name))
           if (text and
               not text.startswith("tf.compat.v1") and
               not text.startswith("tf.compat.v2") and
@@ -198,9 +201,9 @@ class TestUpgrade(test_util.TensorFlowTestCase, parameterized.TestCase):
         api_names = tf_export.get_v1_names(attr)
         for name in api_names:
           if collect:
-            v1_symbols.add("tf." + name)
+            v1_symbols.add("tf." + six.ensure_str(name))
           else:
-            _, _, _, text = self._upgrade("tf." + name)
+            _, _, _, text = self._upgrade("tf." + six.ensure_str(name))
             if (text and
                 not text.startswith("tf.compat.v1") and
                 not text.startswith("tf.compat.v2") and
@@ -337,16 +340,16 @@ class TestUpgrade(test_util.TensorFlowTestCase, parameterized.TestCase):
 
   def testPositionsMatchArgGiven(self):
     full_dict = tf_upgrade_v2.TFAPIChangeSpec().function_arg_warnings
-    method_names = full_dict.keys()
+    method_names = list(full_dict.keys())
     for method_name in method_names:
-      args = full_dict[method_name].keys()
+      args = list(full_dict[method_name].keys())
       if "contrib" in method_name:
         # Skip descending and fetching contrib methods during test. These are
         # not available in the repo anymore.
         continue
-      elif method_name.startswith("*."):
+      elif six.ensure_str(method_name).startswith("*."):
         # special case for optimizer methods
-        method = method_name.replace("*", "tf.train.Optimizer")
+        method = six.ensure_str(method_name).replace("*", "tf.train.Optimizer")
       else:
         method = method_name
 
@@ -354,7 +357,7 @@ class TestUpgrade(test_util.TensorFlowTestCase, parameterized.TestCase):
       arg_spec = tf_inspect.getfullargspec(method)
       for (arg, pos) in args:
         # to deal with the self argument on methods on objects
-        if method_name.startswith("*."):
+        if six.ensure_str(method_name).startswith("*."):
           pos += 1
         self.assertEqual(arg_spec[0][pos], arg)
 
diff --git a/tensorflow/tools/compatibility/update/BUILD b/tensorflow/tools/compatibility/update/BUILD
index 5a74271e882..5f40406c689 100644
--- a/tensorflow/tools/compatibility/update/BUILD
+++ b/tensorflow/tools/compatibility/update/BUILD
@@ -6,7 +6,7 @@ package(
 py_binary(
     name = "generate_v2_renames_map",
     srcs = ["generate_v2_renames_map.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         "//tensorflow:tensorflow_py",
@@ -15,13 +15,14 @@ py_binary(
         "//tensorflow/tools/common:public_api",
         "//tensorflow/tools/common:traverse",
         "//tensorflow/tools/compatibility:all_renames_v2",
+        "@six_archive//:six",
     ],
 )
 
 py_binary(
     name = "generate_v2_reorders_map",
     srcs = ["generate_v2_reorders_map.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         "//tensorflow:tensorflow_py",
diff --git a/tensorflow/tools/compatibility/update/generate_v2_renames_map.py b/tensorflow/tools/compatibility/update/generate_v2_renames_map.py
index 6761fa6ae3d..6cdc4972bd5 100644
--- a/tensorflow/tools/compatibility/update/generate_v2_renames_map.py
+++ b/tensorflow/tools/compatibility/update/generate_v2_renames_map.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2018 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -23,9 +24,9 @@ To update renames_v2.py, run:
 # pylint: enable=line-too-long
 import sys
 
+import six
 import tensorflow as tf
 
-# This import is needed so that TensorFlow python modules are in sys.modules.
 from tensorflow import python as tf_python  # pylint: disable=unused-import
 from tensorflow.python.lib.io import file_io
 from tensorflow.python.platform import app
@@ -35,6 +36,7 @@ from tensorflow.tools.common import public_api
 from tensorflow.tools.common import traverse
 from tensorflow.tools.compatibility import all_renames_v2
 
+# This import is needed so that TensorFlow python modules are in sys.modules.
 
 _OUTPUT_FILE_PATH = 'third_party/tensorflow/tools/compatibility/renames_v2.py'
 _FILE_HEADER = """# Copyright 2018 The TensorFlow Authors. All Rights Reserved.
@@ -178,7 +180,8 @@ def update_renames_v2(output_file_path):
   rename_lines = [
       get_rename_line(name, canonical_name)
       for name, canonical_name in all_renames
-      if 'tf.' + name not in manual_renames]
+      if 'tf.' + six.ensure_str(name) not in manual_renames
+  ]
   renames_file_text = '%srenames = {\n%s\n}\n' % (
       _FILE_HEADER, ',\n'.join(sorted(rename_lines)))
   file_io.write_string_to_file(output_file_path, renames_file_text)
diff --git a/tensorflow/tools/docs/BUILD b/tensorflow/tools/docs/BUILD
index e4806027a91..68f04f20dc3 100644
--- a/tensorflow/tools/docs/BUILD
+++ b/tensorflow/tools/docs/BUILD
@@ -31,6 +31,7 @@ py_library(
         "doc_generator_visitor.py",
     ],
     srcs_version = "PY2AND3",
+    deps = ["@six_archive//:six"],
 )
 
 py_test(
@@ -39,12 +40,13 @@ py_test(
     srcs = [
         "doc_generator_visitor_test.py",
     ],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":doc_generator_visitor",
         ":generate_lib",
         "//tensorflow/python:platform_test",
+        "@six_archive//:six",
     ],
 )
 
@@ -59,7 +61,7 @@ py_test(
     name = "doc_controls_test",
     size = "small",
     srcs = ["doc_controls_test.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":doc_controls",
@@ -77,6 +79,7 @@ py_library(
         "//tensorflow/python:platform",
         "//tensorflow/python:util",
         "@astor_archive//:astor",
+        "@six_archive//:six",
     ],
 )
 
@@ -84,11 +87,12 @@ py_test(
     name = "parser_test",
     size = "small",
     srcs = ["parser_test.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":parser",
         "//tensorflow/python:platform_test",
+        "@six_archive//:six",
     ],
 )
 
@@ -96,6 +100,7 @@ py_library(
     name = "pretty_docs",
     srcs = ["pretty_docs.py"],
     srcs_version = "PY2AND3",
+    deps = ["@six_archive//:six"],
 )
 
 py_library(
@@ -120,7 +125,7 @@ py_test(
     name = "generate_lib_test",
     size = "small",
     srcs = ["generate_lib_test.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":generate_lib",
@@ -132,7 +137,7 @@ py_test(
 py_binary(
     name = "generate",
     srcs = ["generate.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":generate_lib",
@@ -165,9 +170,12 @@ py_test(
 py_binary(
     name = "generate2",
     srcs = ["generate2.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
-    deps = [":generate2_lib"],
+    deps = [
+        ":generate2_lib",
+        "@six_archive//:six",
+    ],
 )
 
 py_library(
@@ -184,16 +192,18 @@ py_library(
     name = "py_guide_parser",
     srcs = ["py_guide_parser.py"],
     srcs_version = "PY2AND3",
+    deps = ["@six_archive//:six"],
 )
 
 py_test(
     name = "py_guide_parser_test",
     size = "small",
     srcs = ["py_guide_parser_test.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":py_guide_parser",
         "//tensorflow/python:client_testlib",
+        "@six_archive//:six",
     ],
 )
diff --git a/tensorflow/tools/docs/doc_generator_visitor.py b/tensorflow/tools/docs/doc_generator_visitor.py
index 6157eb1b7fc..ec2102a5935 100644
--- a/tensorflow/tools/docs/doc_generator_visitor.py
+++ b/tensorflow/tools/docs/doc_generator_visitor.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -48,7 +49,7 @@ class DocGeneratorVisitor(object):
   def set_root_name(self, root_name):
     """Sets the root name for subsequent __call__s."""
     self._root_name = root_name or ''
-    self._prefix = (root_name + '.') if root_name else ''
+    self._prefix = (six.ensure_str(root_name) + '.') if root_name else ''
 
   @property
   def index(self):
@@ -178,7 +179,7 @@ class DocGeneratorVisitor(object):
       A tuple of scores. When sorted the preferred name will have the lowest
       value.
     """
-    parts = name.split('.')
+    parts = six.ensure_str(name).split('.')
     short_name = parts[-1]
 
     container = self._index['.'.join(parts[:-1])]
diff --git a/tensorflow/tools/docs/doc_generator_visitor_test.py b/tensorflow/tools/docs/doc_generator_visitor_test.py
index 1c2635d4a8c..29ec1f8437d 100644
--- a/tensorflow/tools/docs/doc_generator_visitor_test.py
+++ b/tensorflow/tools/docs/doc_generator_visitor_test.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -20,6 +21,8 @@ from __future__ import print_function
 
 import types
 
+import six
+
 from tensorflow.python.platform import googletest
 from tensorflow.tools.docs import doc_generator_visitor
 from tensorflow.tools.docs import generate_lib
@@ -29,9 +32,9 @@ class NoDunderVisitor(doc_generator_visitor.DocGeneratorVisitor):
 
   def __call__(self, parent_name, parent, children):
     """Drop all the dunder methods to make testing easier."""
-    children = [
-        (name, obj) for (name, obj) in children if not name.startswith('_')
-    ]
+    children = [(name, obj)
+                for (name, obj) in children
+                if not six.ensure_str(name).startswith('_')]
     super(NoDunderVisitor, self).__call__(parent_name, parent, children)
 
 
diff --git a/tensorflow/tools/docs/generate_lib.py b/tensorflow/tools/docs/generate_lib.py
index 18d3a8349e8..77a685062ae 100644
--- a/tensorflow/tools/docs/generate_lib.py
+++ b/tensorflow/tools/docs/generate_lib.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -147,8 +148,10 @@ def write_docs(output_dir,
     duplicates = [item for item in duplicates if item != full_name]
 
     for dup in duplicates:
-      from_path = os.path.join(site_api_path, dup.replace('.', '/'))
-      to_path = os.path.join(site_api_path, full_name.replace('.', '/'))
+      from_path = os.path.join(site_api_path,
+                               six.ensure_str(dup).replace('.', '/'))
+      to_path = os.path.join(site_api_path,
+                             six.ensure_str(full_name).replace('.', '/'))
       redirects.append((
           os.path.join('/', from_path),
           os.path.join('/', to_path)))
@@ -167,7 +170,7 @@ def write_docs(output_dir,
     # Generate table of contents
 
     # Put modules in alphabetical order, case-insensitive
-    modules = sorted(module_children.keys(), key=lambda a: a.upper())
+    modules = sorted(list(module_children.keys()), key=lambda a: a.upper())
 
     leftnav_path = os.path.join(output_dir, '_toc.yaml')
     with open(leftnav_path, 'w') as f:
@@ -183,16 +186,15 @@ def write_docs(output_dir,
         if indent_num > 1:
           # tf.contrib.baysflow.entropy will be under
           #   tf.contrib->baysflow->entropy
-          title = module.split('.')[-1]
+          title = six.ensure_str(module).split('.')[-1]
         else:
           title = module
 
         header = [
-            '- title: ' + title,
-            '  section:',
-            '  - title: Overview',
-            '    path: ' + os.path.join('/', site_api_path,
-                                        symbol_to_file[module])]
+            '- title: ' + six.ensure_str(title), '  section:',
+            '  - title: Overview', '    path: ' +
+            os.path.join('/', site_api_path, symbol_to_file[module])
+        ]
         header = ''.join([indent+line+'\n' for line in header])
         f.write(header)
 
@@ -211,8 +213,9 @@ def write_docs(output_dir,
   # Write a global index containing all full names with links.
   with open(os.path.join(output_dir, 'index.md'), 'w') as f:
     f.write(
-        parser.generate_global_index(root_title, parser_config.index,
-                                     parser_config.reference_resolver))
+        six.ensure_str(
+            parser.generate_global_index(root_title, parser_config.index,
+                                         parser_config.reference_resolver)))
 
 
 def add_dict_to_dict(add_from, add_to):
@@ -345,7 +348,7 @@ def build_doc_index(src_dir):
   for dirpath, _, filenames in os.walk(src_dir):
     suffix = os.path.relpath(path=dirpath, start=src_dir)
     for base_name in filenames:
-      if not base_name.endswith('.md'):
+      if not six.ensure_str(base_name).endswith('.md'):
         continue
       title_parser = _GetMarkdownTitle()
       title_parser.process(os.path.join(dirpath, base_name))
@@ -353,7 +356,8 @@ def build_doc_index(src_dir):
         msg = ('`{}` has no markdown title (# title)'.format(
             os.path.join(dirpath, base_name)))
         raise ValueError(msg)
-      key_parts = os.path.join(suffix, base_name[:-3]).split('/')
+      key_parts = six.ensure_str(os.path.join(suffix,
+                                              base_name[:-3])).split('/')
       if key_parts[-1] == 'index':
         key_parts = key_parts[:-1]
       doc_info = _DocInfo(os.path.join(suffix, base_name), title_parser.title)
@@ -367,8 +371,8 @@ def build_doc_index(src_dir):
 class _GuideRef(object):
 
   def __init__(self, base_name, title, section_title, section_tag):
-    self.url = 'api_guides/python/' + (('%s#%s' % (base_name, section_tag))
-                                       if section_tag else base_name)
+    self.url = 'api_guides/python/' + six.ensure_str(
+        (('%s#%s' % (base_name, section_tag)) if section_tag else base_name))
     self.link_text = (('%s > %s' % (title, section_title))
                       if section_title else title)
 
@@ -447,7 +451,7 @@ def update_id_tags_inplace(src_dir):
       # modified file contents
       content = tag_updater.process(full_path)
       with open(full_path, 'w') as f:
-        f.write(content)
+        f.write(six.ensure_str(content))
 
 
 EXCLUDED = set(['__init__.py', 'OWNERS', 'README.txt'])
@@ -512,7 +516,7 @@ def replace_refs(src_dir,
       content = reference_resolver.replace_references(content,
                                                       relative_path_to_root)
       with open(full_out_path, 'wb') as f:
-        f.write(content.encode('utf-8'))
+        f.write(six.ensure_binary(content, 'utf-8'))
 
 
 class DocGenerator(object):
diff --git a/tensorflow/tools/docs/parser.py b/tensorflow/tools/docs/parser.py
index d87f9585f20..d6426cfd6de 100644
--- a/tensorflow/tools/docs/parser.py
+++ b/tensorflow/tools/docs/parser.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -28,6 +29,7 @@ import re
 
 import astor
 import six
+from six.moves import zip
 
 from google.protobuf.message import Message as ProtoMessage
 from tensorflow.python.platform import tf_logging as logging
@@ -50,7 +52,7 @@ def is_free_function(py_object, full_name, index):
   if not tf_inspect.isfunction(py_object):
     return False
 
-  parent_name = full_name.rsplit('.', 1)[0]
+  parent_name = six.ensure_str(full_name).rsplit('.', 1)[0]
   if tf_inspect.isclass(index[parent_name]):
     return False
 
@@ -112,14 +114,14 @@ def documentation_path(full_name, is_fragment=False):
   Returns:
     The file path to which to write the documentation for `full_name`.
   """
-  parts = full_name.split('.')
+  parts = six.ensure_str(full_name).split('.')
   if is_fragment:
     parts, fragment = parts[:-1], parts[-1]
 
-  result = os.path.join(*parts) + '.md'
+  result = six.ensure_str(os.path.join(*parts)) + '.md'
 
   if is_fragment:
-    result = result + '#' + fragment
+    result = six.ensure_str(result) + '#' + six.ensure_str(fragment)
 
   return result
 
@@ -288,7 +290,7 @@ class ReferenceResolver(object):
         self.add_error(e.message)
         return 'BAD_LINK'
 
-    string = re.sub(SYMBOL_REFERENCE_RE, strict_one_ref, string)
+    string = re.sub(SYMBOL_REFERENCE_RE, strict_one_ref, six.ensure_str(string))
 
     def sloppy_one_ref(match):
       try:
@@ -333,7 +335,7 @@ class ReferenceResolver(object):
   @staticmethod
   def _link_text_to_html(link_text):
     code_re = '`(.*?)`'
-    return re.sub(code_re, r'<code>\1</code>', link_text)
+    return re.sub(code_re, r'<code>\1</code>', six.ensure_str(link_text))
 
   def py_master_name(self, full_name):
     """Return the master name for a Python symbol name."""
@@ -389,11 +391,11 @@ class ReferenceResolver(object):
       manual_link_text = False
 
     # Handle different types of references.
-    if string.startswith('$'):  # Doc reference
+    if six.ensure_str(string).startswith('$'):  # Doc reference
       return self._doc_link(string, link_text, manual_link_text,
                             relative_path_to_root)
 
-    elif string.startswith('tensorflow::'):
+    elif six.ensure_str(string).startswith('tensorflow::'):
       # C++ symbol
       return self._cc_link(string, link_text, manual_link_text,
                            relative_path_to_root)
@@ -401,7 +403,8 @@ class ReferenceResolver(object):
     else:
       is_python = False
       for py_module_name in self._py_module_names:
-        if string == py_module_name or string.startswith(py_module_name + '.'):
+        if string == py_module_name or string.startswith(
+            six.ensure_str(py_module_name) + '.'):
           is_python = True
           break
       if is_python:  # Python symbol
@@ -421,7 +424,7 @@ class ReferenceResolver(object):
     string = string[1:]  # remove leading $
 
     # If string has a #, split that part into `hash_tag`
-    hash_pos = string.find('#')
+    hash_pos = six.ensure_str(string).find('#')
     if hash_pos > -1:
       hash_tag = string[hash_pos:]
       string = string[:hash_pos]
@@ -520,10 +523,10 @@ class _FunctionDetail(
 
   def __str__(self):
     """Return the original string that represents the function detail."""
-    parts = [self.keyword + ':\n']
+    parts = [six.ensure_str(self.keyword) + ':\n']
     parts.append(self.header)
     for key, value in self.items:
-      parts.append('  ' + key + ': ')
+      parts.append('  ' + six.ensure_str(key) + ': ')
       parts.append(value)
 
     return ''.join(parts)
@@ -587,7 +590,7 @@ def _parse_function_details(docstring):
   item_re = re.compile(r'^   ? ?(\*?\*?\w[\w.]*?\s*):\s', re.MULTILINE)
 
   for keyword, content in pairs:
-    content = item_re.split(content)
+    content = item_re.split(six.ensure_str(content))
     header = content[0]
     items = list(_gen_pairs(content[1:]))
 
@@ -634,7 +637,8 @@ def _parse_md_docstring(py_object, relative_path_to_root, reference_resolver):
 
   atat_re = re.compile(r' *@@[a-zA-Z_.0-9]+ *$')
   raw_docstring = '\n'.join(
-      line for line in raw_docstring.split('\n') if not atat_re.match(line))
+      line for line in six.ensure_str(raw_docstring).split('\n')
+      if not atat_re.match(six.ensure_str(line)))
 
   docstring, compatibility = _handle_compatibility(raw_docstring)
   docstring, function_details = _parse_function_details(docstring)
@@ -698,8 +702,9 @@ def _get_arg_spec(func):
 
 
 def _remove_first_line_indent(string):
-  indent = len(re.match(r'^\s*', string).group(0))
-  return '\n'.join([line[indent:] for line in string.split('\n')])
+  indent = len(re.match(r'^\s*', six.ensure_str(string)).group(0))
+  return '\n'.join(
+      [line[indent:] for line in six.ensure_str(string).split('\n')])
 
 
 PAREN_NUMBER_RE = re.compile(r'^\(([0-9.e-]+)\)')
@@ -761,9 +766,9 @@ def _generate_signature(func, reverse_index):
         default_text = reverse_index[id(default)]
       elif ast_default is not None:
         default_text = (
-            astor.to_source(ast_default).rstrip('\n').replace('\t', '\\t')
-            .replace('\n', '\\n').replace('"""', "'"))
-        default_text = PAREN_NUMBER_RE.sub('\\1', default_text)
+            six.ensure_str(astor.to_source(ast_default)).rstrip('\n').replace(
+                '\t', '\\t').replace('\n', '\\n').replace('"""', "'"))
+        default_text = PAREN_NUMBER_RE.sub('\\1', six.ensure_str(default_text))
 
         if default_text != repr(default):
           # This may be an internal name. If so, handle the ones we know about.
@@ -797,9 +802,9 @@ def _generate_signature(func, reverse_index):
 
   # Add *args and *kwargs.
   if argspec.varargs:
-    args_list.append('*' + argspec.varargs)
+    args_list.append('*' + six.ensure_str(argspec.varargs))
   if argspec.varkw:
-    args_list.append('**' + argspec.varkw)
+    args_list.append('**' + six.ensure_str(argspec.varkw))
 
   return args_list
 
@@ -879,7 +884,7 @@ class _FunctionPageInfo(object):
 
   @property
   def short_name(self):
-    return self._full_name.split('.')[-1]
+    return six.ensure_str(self._full_name).split('.')[-1]
 
   @property
   def defined_in(self):
@@ -998,7 +1003,7 @@ class _ClassPageInfo(object):
   @property
   def short_name(self):
     """Returns the documented object's short name."""
-    return self._full_name.split('.')[-1]
+    return six.ensure_str(self._full_name).split('.')[-1]
 
   @property
   def defined_in(self):
@@ -1091,9 +1096,12 @@ class _ClassPageInfo(object):
       base_url = parser_config.reference_resolver.reference_to_url(
           base_full_name, relative_path)
 
-      link_info = _LinkInfo(short_name=base_full_name.split('.')[-1],
-                            full_name=base_full_name, obj=base,
-                            doc=base_doc, url=base_url)
+      link_info = _LinkInfo(
+          short_name=six.ensure_str(base_full_name).split('.')[-1],
+          full_name=base_full_name,
+          obj=base,
+          doc=base_doc,
+          url=base_url)
       bases.append(link_info)
 
     self._bases = bases
@@ -1121,7 +1129,7 @@ class _ClassPageInfo(object):
       doc: The property's parsed docstring, a `_DocstringInfo`.
     """
     # Hide useless namedtuple docs-trings
-    if re.match('Alias for field number [0-9]+', doc.docstring):
+    if re.match('Alias for field number [0-9]+', six.ensure_str(doc.docstring)):
       doc = doc._replace(docstring='', brief='')
     property_info = _PropertyInfo(short_name, full_name, obj, doc)
     self._properties.append(property_info)
@@ -1255,8 +1263,8 @@ class _ClassPageInfo(object):
 
         # Omit methods defined by namedtuple.
         original_method = defining_class.__dict__[short_name]
-        if (hasattr(original_method, '__module__') and
-            (original_method.__module__ or '').startswith('namedtuple')):
+        if (hasattr(original_method, '__module__') and six.ensure_str(
+            (original_method.__module__ or '')).startswith('namedtuple')):
           continue
 
         # Some methods are often overridden without documentation. Because it's
@@ -1294,7 +1302,7 @@ class _ClassPageInfo(object):
       else:
         # Exclude members defined by protobuf that are useless
         if issubclass(py_class, ProtoMessage):
-          if (short_name.endswith('_FIELD_NUMBER') or
+          if (six.ensure_str(short_name).endswith('_FIELD_NUMBER') or
               short_name in ['__slots__', 'DESCRIPTOR']):
             continue
 
@@ -1332,7 +1340,7 @@ class _ModulePageInfo(object):
 
   @property
   def short_name(self):
-    return self._full_name.split('.')[-1]
+    return six.ensure_str(self._full_name).split('.')[-1]
 
   @property
   def defined_in(self):
@@ -1425,7 +1433,8 @@ class _ModulePageInfo(object):
                   '__cached__', '__loader__', '__spec__']:
         continue
 
-      member_full_name = self.full_name + '.' + name if self.full_name else name
+      member_full_name = six.ensure_str(self.full_name) + '.' + six.ensure_str(
+          name) if self.full_name else name
       member = parser_config.py_name_to_object(member_full_name)
 
       member_doc = _parse_md_docstring(member, relative_path,
@@ -1680,20 +1689,21 @@ def _get_defined_in(py_object, parser_config):
   # TODO(wicke): And make their source file predictable from the file name.
 
   # In case this is compiled, point to the original
-  if path.endswith('.pyc'):
+  if six.ensure_str(path).endswith('.pyc'):
     path = path[:-1]
 
   # Never include links outside this code base.
-  if path.startswith('..') or re.search(r'\b_api\b', path):
+  if six.ensure_str(path).startswith('..') or re.search(r'\b_api\b',
+                                                        six.ensure_str(path)):
     return None
 
-  if re.match(r'.*/gen_[^/]*\.py$', path):
+  if re.match(r'.*/gen_[^/]*\.py$', six.ensure_str(path)):
     return _GeneratedFile(path, parser_config)
   if 'genfiles' in path or 'tools/api/generator' in path:
     return _GeneratedFile(path, parser_config)
-  elif re.match(r'.*_pb2\.py$', path):
+  elif re.match(r'.*_pb2\.py$', six.ensure_str(path)):
     # The _pb2.py files all appear right next to their defining .proto file.
-    return _ProtoFile(path[:-7] + '.proto', parser_config)
+    return _ProtoFile(six.ensure_str(path[:-7]) + '.proto', parser_config)
   else:
     return _PythonFile(path, parser_config)
 
diff --git a/tensorflow/tools/docs/parser_test.py b/tensorflow/tools/docs/parser_test.py
index 6ecdf521cdc..15d4cad89cc 100644
--- a/tensorflow/tools/docs/parser_test.py
+++ b/tensorflow/tools/docs/parser_test.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -23,6 +24,8 @@ import functools
 import os
 import sys
 
+import six
+
 from tensorflow.python.platform import googletest
 from tensorflow.python.util import tf_inspect
 from tensorflow.tools.docs import doc_controls
@@ -180,7 +183,8 @@ class ParserTest(googletest.TestCase):
 
     # Make sure the brief docstring is present
     self.assertEqual(
-        tf_inspect.getdoc(TestClass).split('\n')[0], page_info.doc.brief)
+        six.ensure_str(tf_inspect.getdoc(TestClass)).split('\n')[0],
+        page_info.doc.brief)
 
     # Make sure the method is present
     self.assertEqual(TestClass.a_method, page_info.methods[0].obj)
@@ -236,7 +240,7 @@ class ParserTest(googletest.TestCase):
     #   'Alias for field number ##'. These props are returned sorted.
 
     def sort_key(prop_info):
-      return int(prop_info.obj.__doc__.split(' ')[-1])
+      return int(six.ensure_str(prop_info.obj.__doc__).split(' ')[-1])
 
     self.assertSequenceEqual(page_info.properties,
                              sorted(page_info.properties, key=sort_key))
@@ -378,7 +382,8 @@ class ParserTest(googletest.TestCase):
 
     # Make sure the brief docstring is present
     self.assertEqual(
-        tf_inspect.getdoc(test_module).split('\n')[0], page_info.doc.brief)
+        six.ensure_str(tf_inspect.getdoc(test_module)).split('\n')[0],
+        page_info.doc.brief)
 
     # Make sure that the members are there
     funcs = {f_info.obj for f_info in page_info.functions}
@@ -422,7 +427,8 @@ class ParserTest(googletest.TestCase):
 
     # Make sure the brief docstring is present
     self.assertEqual(
-        tf_inspect.getdoc(test_function).split('\n')[0], page_info.doc.brief)
+        six.ensure_str(tf_inspect.getdoc(test_function)).split('\n')[0],
+        page_info.doc.brief)
 
     # Make sure the extracted signature is good.
     self.assertEqual(['unused_arg', "unused_kwarg='default'"],
@@ -461,7 +467,8 @@ class ParserTest(googletest.TestCase):
 
     # Make sure the brief docstring is present
     self.assertEqual(
-        tf_inspect.getdoc(test_function_with_args_kwargs).split('\n')[0],
+        six.ensure_str(
+            tf_inspect.getdoc(test_function_with_args_kwargs)).split('\n')[0],
         page_info.doc.brief)
 
     # Make sure the extracted signature is good.
@@ -751,7 +758,8 @@ class TestParseFunctionDetails(googletest.TestCase):
 
     self.assertEqual(
         RELU_DOC,
-        docstring + ''.join(str(detail) for detail in function_details))
+        six.ensure_str(docstring) +
+        ''.join(str(detail) for detail in function_details))
 
 
 class TestGenerateSignature(googletest.TestCase):
diff --git a/tensorflow/tools/docs/pretty_docs.py b/tensorflow/tools/docs/pretty_docs.py
index 1a3e79621f8..d7237b1a39a 100644
--- a/tensorflow/tools/docs/pretty_docs.py
+++ b/tensorflow/tools/docs/pretty_docs.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2015 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -28,6 +29,7 @@ from __future__ import division
 from __future__ import print_function
 
 import textwrap
+import six
 
 
 def build_md_page(page_info):
@@ -83,7 +85,8 @@ def _build_class_page(page_info):
   """Given a ClassPageInfo object Return the page as an md string."""
   parts = ['# {page_info.full_name}\n\n'.format(page_info=page_info)]
 
-  parts.append('## Class `%s`\n\n' % page_info.full_name.split('.')[-1])
+  parts.append('## Class `%s`\n\n' %
+               six.ensure_str(page_info.full_name).split('.')[-1])
   if page_info.bases:
     parts.append('Inherits From: ')
 
@@ -222,7 +225,7 @@ def _build_module_page(page_info):
       parts.append(template.format(**item._asdict()))
 
       if item.doc.brief:
-        parts.append(': ' + item.doc.brief)
+        parts.append(': ' + six.ensure_str(item.doc.brief))
 
       parts.append('\n\n')
 
@@ -234,7 +237,7 @@ def _build_module_page(page_info):
       parts.append(template.format(**item._asdict()))
 
       if item.doc.brief:
-        parts.append(': ' + item.doc.brief)
+        parts.append(': ' + six.ensure_str(item.doc.brief))
 
       parts.append('\n\n')
 
@@ -246,7 +249,7 @@ def _build_module_page(page_info):
       parts.append(template.format(**item._asdict()))
 
       if item.doc.brief:
-        parts.append(': ' + item.doc.brief)
+        parts.append(': ' + six.ensure_str(item.doc.brief))
 
       parts.append('\n\n')
 
@@ -273,7 +276,7 @@ def _build_signature(obj_info, use_full_name=True):
         '```\n\n')
 
   parts = ['``` python']
-  parts.extend(['@' + dec for dec in obj_info.decorators])
+  parts.extend(['@' + six.ensure_str(dec) for dec in obj_info.decorators])
   signature_template = '{name}({sig})'
 
   if not obj_info.signature:
@@ -313,7 +316,7 @@ def _build_function_details(function_details):
   parts = []
   for detail in function_details:
     sub = []
-    sub.append('#### ' + detail.keyword + ':\n\n')
+    sub.append('#### ' + six.ensure_str(detail.keyword) + ':\n\n')
     sub.append(textwrap.dedent(detail.header))
     for key, value in detail.items:
       sub.append('* <b>`%s`</b>: %s' % (key, value))
diff --git a/tensorflow/tools/docs/py_guide_parser.py b/tensorflow/tools/docs/py_guide_parser.py
index b00694dc403..70149c4dd9e 100644
--- a/tensorflow/tools/docs/py_guide_parser.py
+++ b/tensorflow/tools/docs/py_guide_parser.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2017 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -21,14 +22,16 @@ from __future__ import print_function
 
 import os
 import re
+import six
 
 
 def md_files_in_dir(py_guide_src_dir):
   """Returns a list of filename (full_path, base) pairs for guide files."""
   all_in_dir = [(os.path.join(py_guide_src_dir, f), f)
                 for f in os.listdir(py_guide_src_dir)]
-  return [(full, f) for full, f in all_in_dir
-          if os.path.isfile(full) and f.endswith('.md')]
+  return [(full, f)
+          for full, f in all_in_dir
+          if os.path.isfile(full) and six.ensure_str(f).endswith('.md')]
 
 
 class PyGuideParser(object):
diff --git a/tensorflow/tools/docs/py_guide_parser_test.py b/tensorflow/tools/docs/py_guide_parser_test.py
index 168b0535a94..2975a1a6575 100644
--- a/tensorflow/tools/docs/py_guide_parser_test.py
+++ b/tensorflow/tools/docs/py_guide_parser_test.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2017 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -20,6 +21,8 @@ from __future__ import print_function
 
 import os
 
+import six
+
 from tensorflow.python.platform import test
 from tensorflow.tools.docs import py_guide_parser
 
@@ -38,7 +41,7 @@ class TestPyGuideParser(py_guide_parser.PyGuideParser):
 
   def process_in_blockquote(self, line_number, line):
     self.calls.append((line_number, 'b', line))
-    self.replace_line(line_number, line + ' BQ')
+    self.replace_line(line_number, six.ensure_str(line) + ' BQ')
 
   def process_line(self, line_number, line):
     self.calls.append((line_number, 'l', line))
diff --git a/tensorflow/tools/git/BUILD b/tensorflow/tools/git/BUILD
index fb6a07133e7..8a47f4c4c2d 100644
--- a/tensorflow/tools/git/BUILD
+++ b/tensorflow/tools/git/BUILD
@@ -11,6 +11,7 @@ package(
 py_binary(
     name = "gen_git_source",
     srcs = ["gen_git_source.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
+    deps = ["@six_archive//:six"],
 )
diff --git a/tensorflow/tools/git/gen_git_source.py b/tensorflow/tools/git/gen_git_source.py
index d29b535ae30..011406e2288 100755
--- a/tensorflow/tools/git/gen_git_source.py
+++ b/tensorflow/tools/git/gen_git_source.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2016 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -26,13 +27,16 @@ NOTE: this script is only used in opensource.
 from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
-from builtins import bytes  # pylint: disable=redefined-builtin
+
 import argparse
+from builtins import bytes  # pylint: disable=redefined-builtin
 import json
 import os
 import shutil
 import subprocess
 
+import six
+
 
 def parse_branch_ref(filename):
   """Given a filename of a .git/HEAD file return ref path.
@@ -161,10 +165,13 @@ def get_git_version(git_base_path, git_tag_override):
   unknown_label = b"unknown"
   try:
     # Force to bytes so this works on python 2 and python 3
-    val = bytes(subprocess.check_output([
-        "git", str("--git-dir=%s/.git" % git_base_path),
-        str("--work-tree=" + git_base_path), "describe", "--long", "--tags"
-    ]).strip())
+    val = bytes(
+        subprocess.check_output([
+            "git",
+            str("--git-dir=%s/.git" % git_base_path),
+            str("--work-tree=" + six.ensure_str(git_base_path)), "describe",
+            "--long", "--tags"
+        ]).strip())
     version_separator = b"-"
     if git_tag_override and val:
       split_val = val.split(version_separator)
diff --git a/tensorflow/tools/graph_transforms/BUILD b/tensorflow/tools/graph_transforms/BUILD
index adafe2aca12..0f5c298b48b 100644
--- a/tensorflow/tools/graph_transforms/BUILD
+++ b/tensorflow/tools/graph_transforms/BUILD
@@ -3,7 +3,6 @@
 
 load(
     "//tensorflow:tensorflow.bzl",
-    "if_not_v2",
     "if_not_windows",
     "tf_cc_binary",
     "tf_cc_test",
diff --git a/tensorflow/tools/lib_package/BUILD b/tensorflow/tools/lib_package/BUILD
index 174f21610dc..1162ad5de47 100644
--- a/tensorflow/tools/lib_package/BUILD
+++ b/tensorflow/tools/lib_package/BUILD
@@ -163,6 +163,7 @@ genrule(
         "@com_google_protobuf//:LICENSE",
         "@snappy//:COPYING",
         "@zlib_archive//:zlib.h",
+        "@six_archive//:LICENSE",
     ] + select({
         "//tensorflow:android": [],
         "//tensorflow:ios": [],
@@ -235,6 +236,7 @@ genrule(
         "@zlib_archive//:zlib.h",
         "@grpc//:LICENSE",
         "@grpc//third_party/address_sorting:LICENSE",
+        "@six_archive//:LICENSE",
     ] + select({
         "//tensorflow:android": [],
         "//tensorflow:ios": [],
diff --git a/tensorflow/tools/tensorflow_builder/compat_checker/BUILD b/tensorflow/tools/tensorflow_builder/compat_checker/BUILD
index b60a7df1b76..d2119dd1e63 100644
--- a/tensorflow/tools/tensorflow_builder/compat_checker/BUILD
+++ b/tensorflow/tools/tensorflow_builder/compat_checker/BUILD
@@ -13,9 +13,12 @@ package(
     ],
 )
 
+licenses(["notice"])  # Apache 2.0
+
 py_library(
     name = "compat_checker",
     srcs = ["compat_checker.py"],
+    srcs_version = "PY2AND3",
     deps = [
         "//tensorflow/python:platform",
         "//tensorflow/python:util",
@@ -29,6 +32,7 @@ py_test(
     data = [
         "//tensorflow/tools/tensorflow_builder/compat_checker:test_config",
     ],
+    python_version = "PY3",
     tags = ["no_pip"],
     deps = [
         ":compat_checker",
diff --git a/tensorflow/tools/tensorflow_builder/compat_checker/compat_checker.py b/tensorflow/tools/tensorflow_builder/compat_checker/compat_checker.py
index d65d7727ffa..ec8a0ba6f96 100644
--- a/tensorflow/tools/tensorflow_builder/compat_checker/compat_checker.py
+++ b/tensorflow/tools/tensorflow_builder/compat_checker/compat_checker.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2019 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -23,16 +24,11 @@ import re
 import sys
 
 import six
+from six.moves import range
+import six.moves.configparser
 from tensorflow.python.platform import tf_logging as logging
 from tensorflow.python.util import tf_inspect
 
-# pylint: disable=g-import-not-at-top
-if six.PY2:
-  import ConfigParser
-else:
-  import configparser as ConfigParser
-# pylint: enable=g-import-not-at-top
-
 PATH_TO_DIR = "tensorflow/tools/tensorflow_builder/compat_checker"
 
 
@@ -56,8 +52,8 @@ def _compare_versions(v1, v2):
     raise RuntimeError("Cannot compare `inf` to `inf`.")
 
   rtn_dict = {"smaller": None, "larger": None}
-  v1_list = v1.split(".")
-  v2_list = v2.split(".")
+  v1_list = six.ensure_str(v1).split(".")
+  v2_list = six.ensure_str(v2).split(".")
   # Take care of cases with infinity (arg=`inf`).
   if v1_list[0] == "inf":
     v1_list[0] = str(int(v2_list[0]) + 1)
@@ -380,7 +376,7 @@ class ConfigCompatChecker(object):
     curr_status = True
 
     # Initialize config parser for parsing version requirements file.
-    parser = ConfigParser.ConfigParser()
+    parser = six.moves.configparser.ConfigParser()
     parser.read(self.req_file)
 
     if not parser.sections():
@@ -643,7 +639,7 @@ class ConfigCompatChecker(object):
     if filtered[-1] == "]":
       filtered = filtered[:-1]
     elif "]" in filtered[-1]:
-      filtered[-1] = filtered[-1].replace("]", "")
+      filtered[-1] = six.ensure_str(filtered[-1]).replace("]", "")
     # If `]` is missing, then it could be a formatting issue with
     # config file (.ini.). Add to warning.
     else:
@@ -792,7 +788,7 @@ class ConfigCompatChecker(object):
       Boolean that is a status of the compatibility check result.
     """
     # Check if all `Required` configs are found in user configs.
-    usr_keys = self.usr_config.keys()
+    usr_keys = list(self.usr_config.keys())
 
     for k in six.iterkeys(self.usr_config):
       if k not in usr_keys:
@@ -809,10 +805,10 @@ class ConfigCompatChecker(object):
     for config_name, spec in six.iteritems(self.usr_config):
       temp_status = True
       # Check under which section the user config is defined.
-      in_required = config_name in self.required.keys()
-      in_optional = config_name in self.optional.keys()
-      in_unsupported = config_name in self.unsupported.keys()
-      in_dependency = config_name in self.dependency.keys()
+      in_required = config_name in list(self.required.keys())
+      in_optional = config_name in list(self.optional.keys())
+      in_unsupported = config_name in list(self.unsupported.keys())
+      in_dependency = config_name in list(self.dependency.keys())
 
       # Add to warning if user config is not specified in the config file.
       if not (in_required or in_optional or in_unsupported or in_dependency):
diff --git a/tensorflow/tools/tensorflow_builder/compat_checker/compat_checker_test.py b/tensorflow/tools/tensorflow_builder/compat_checker/compat_checker_test.py
index bd0d50a9f99..b12815b555d 100644
--- a/tensorflow/tools/tensorflow_builder/compat_checker/compat_checker_test.py
+++ b/tensorflow/tools/tensorflow_builder/compat_checker/compat_checker_test.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2019 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -88,7 +89,7 @@ class CompatCheckerTest(unittest.TestCase):
     # Make sure no warning or error messages are recorded.
     self.assertFalse(len(self.compat_checker.error_msg))
     # Make sure total # of successes match total # of configs.
-    cnt = len(USER_CONFIG_IN_RANGE.keys())
+    cnt = len(list(USER_CONFIG_IN_RANGE.keys()))
     self.assertEqual(len(self.compat_checker.successes), cnt)
 
   def testWithUserConfigNotInRange(self):
@@ -106,7 +107,7 @@ class CompatCheckerTest(unittest.TestCase):
     err_msg_list = self.compat_checker.failures
     self.assertTrue(len(err_msg_list))
     # Make sure total # of failures match total # of configs.
-    cnt = len(USER_CONFIG_NOT_IN_RANGE.keys())
+    cnt = len(list(USER_CONFIG_NOT_IN_RANGE.keys()))
     self.assertEqual(len(err_msg_list), cnt)
 
   def testWithUserConfigMissing(self):
diff --git a/tensorflow/tools/tensorflow_builder/config_detector/BUILD b/tensorflow/tools/tensorflow_builder/config_detector/BUILD
index 6227366fec1..ab52eb33fdc 100644
--- a/tensorflow/tools/tensorflow_builder/config_detector/BUILD
+++ b/tensorflow/tools/tensorflow_builder/config_detector/BUILD
@@ -14,22 +14,24 @@ py_binary(
     data = [
         "//tensorflow/tools/tensorflow_builder/config_detector/data/golden:cuda_cc_golden",
     ],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":cuda_compute_capability",
         "@absl_py//absl:app",
         "@absl_py//absl/flags",
+        "@six_archive//:six",
     ],
 )
 
 py_binary(
     name = "cuda_compute_capability",
     srcs = ["data/cuda_compute_capability.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         "@absl_py//absl:app",
         "@absl_py//absl/flags",
+        "@six_archive//:six",
     ],
 )
diff --git a/tensorflow/tools/tensorflow_builder/config_detector/config_detector.py b/tensorflow/tools/tensorflow_builder/config_detector/config_detector.py
index 680247d19b8..2c24780bcfd 100755
--- a/tensorflow/tools/tensorflow_builder/config_detector/config_detector.py
+++ b/tensorflow/tools/tensorflow_builder/config_detector/config_detector.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2019 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -70,6 +71,7 @@ import subprocess
 import sys
 from absl import app
 from absl import flags
+import six
 
 from tensorflow.tools.tensorflow_builder.config_detector.data import cuda_compute_capability
 
@@ -182,7 +184,7 @@ def get_cpu_type():
   """
   key = "cpu_type"
   out, err = run_shell_cmd(cmds_all[PLATFORM][key])
-  cpu_detected = out.split(":")[1].strip()
+  cpu_detected = out.split(b":")[1].strip()
   if err and FLAGS.debug:
     print("Error in detecting CPU type:\n %s" % str(err))
 
@@ -201,7 +203,7 @@ def get_cpu_arch():
   if err and FLAGS.debug:
     print("Error in detecting CPU arch:\n %s" % str(err))
 
-  return out.strip("\n")
+  return out.strip(b"\n")
 
 
 def get_distrib():
@@ -216,7 +218,7 @@ def get_distrib():
   if err and FLAGS.debug:
     print("Error in detecting distribution:\n %s" % str(err))
 
-  return out.strip("\n")
+  return out.strip(b"\n")
 
 
 def get_distrib_version():
@@ -233,7 +235,7 @@ def get_distrib_version():
         "Error in detecting distribution version:\n %s" % str(err)
     )
 
-  return out.strip("\n")
+  return out.strip(b"\n")
 
 
 def get_gpu_type():
@@ -251,7 +253,7 @@ def get_gpu_type():
   key = "gpu_type_no_sudo"
   gpu_dict = cuda_compute_capability.retrieve_from_golden()
   out, err = run_shell_cmd(cmds_all[PLATFORM][key])
-  ret_val = out.split(" ")
+  ret_val = out.split(b" ")
   gpu_id = ret_val[0]
   if err and FLAGS.debug:
     print("Error in detecting GPU type:\n %s" % str(err))
@@ -261,10 +263,10 @@ def get_gpu_type():
     return gpu_id, GPU_TYPE
   else:
     if "[" or "]" in ret_val[1]:
-      gpu_release = ret_val[1].replace("[", "") + " "
-      gpu_release += ret_val[2].replace("]", "").strip("\n")
+      gpu_release = ret_val[1].replace(b"[", b"") + b" "
+      gpu_release += ret_val[2].replace(b"]", b"").strip(b"\n")
     else:
-      gpu_release = ret_val[1].replace("\n", " ")
+      gpu_release = six.ensure_str(ret_val[1]).replace("\n", " ")
 
     if gpu_release not in gpu_dict:
       GPU_TYPE = "unknown"
@@ -285,7 +287,7 @@ def get_gpu_count():
   if err and FLAGS.debug:
     print("Error in detecting GPU count:\n %s" % str(err))
 
-  return out.strip("\n")
+  return out.strip(b"\n")
 
 
 def get_cuda_version_all():
@@ -303,7 +305,7 @@ def get_cuda_version_all():
   """
   key = "cuda_ver_all"
   out, err = run_shell_cmd(cmds_all[PLATFORM.lower()][key])
-  ret_val = out.split("\n")
+  ret_val = out.split(b"\n")
   filtered = []
   for item in ret_val:
     if item not in ["\n", ""]:
@@ -311,9 +313,9 @@ def get_cuda_version_all():
 
   all_vers = []
   for item in filtered:
-    ver_re = re.search(r".*/cuda(\-[\d]+\.[\d]+)?", item)
+    ver_re = re.search(r".*/cuda(\-[\d]+\.[\d]+)?", item.decode("utf-8"))
     if ver_re.group(1):
-      all_vers.append(ver_re.group(1).strip("-"))
+      all_vers.append(six.ensure_str(ver_re.group(1)).strip("-"))
 
   if err and FLAGS.debug:
     print("Error in detecting CUDA version:\n %s" % str(err))
@@ -409,13 +411,13 @@ def get_cudnn_version():
   if err and FLAGS.debug:
     print("Error in finding `cudnn.h`:\n %s" % str(err))
 
-  if len(out.split(" ")) > 1:
+  if len(out.split(b" ")) > 1:
     cmd = cmds[0] + " | " + cmds[1]
     out_re, err_re = run_shell_cmd(cmd)
     if err_re and FLAGS.debug:
       print("Error in detecting cuDNN version:\n %s" % str(err_re))
 
-    return out_re.strip("\n")
+    return out_re.strip(b"\n")
   else:
     return
 
@@ -432,7 +434,7 @@ def get_gcc_version():
   if err and FLAGS.debug:
     print("Error in detecting GCC version:\n %s" % str(err))
 
-  return out.strip("\n")
+  return out.strip(b"\n")
 
 
 def get_glibc_version():
@@ -447,7 +449,7 @@ def get_glibc_version():
   if err and FLAGS.debug:
     print("Error in detecting GCC version:\n %s" % str(err))
 
-  return out.strip("\n")
+  return out.strip(b"\n")
 
 
 def get_libstdcpp_version():
@@ -462,7 +464,7 @@ def get_libstdcpp_version():
   if err and FLAGS.debug:
     print("Error in detecting libstdc++ version:\n %s" % str(err))
 
-  ver = out.split("_")[-1].replace("\n", "")
+  ver = out.split(b"_")[-1].replace(b"\n", b"")
   return ver
 
 
@@ -485,7 +487,7 @@ def get_cpu_isa_version():
   found = []
   missing = []
   for isa in required_isa:
-    for sys_isa in ret_val.split(" "):
+    for sys_isa in ret_val.split(b" "):
       if isa == sys_isa:
         if isa not in found:
           found.append(isa)
@@ -539,7 +541,7 @@ def get_all_configs():
   json_data = {}
   missing = []
   warning = []
-  for config, call_func in all_functions.iteritems():
+  for config, call_func in six.iteritems(all_functions):
     ret_val = call_func
     if not ret_val:
       configs_found.append([config, "\033[91m\033[1mMissing\033[0m"])
@@ -557,10 +559,10 @@ def get_all_configs():
           configs_found.append([config, ret_val[0]])
           json_data[config] = ret_val[0]
         else:
-          configs_found.append(
-              [config,
-               "\033[91m\033[1mMissing " + str(ret_val[1])[1:-1] + "\033[0m"]
-          )
+          configs_found.append([
+              config, "\033[91m\033[1mMissing " +
+              six.ensure_str(str(ret_val[1])[1:-1]) + "\033[0m"
+          ])
           missing.append(
               [config,
                "\n\t=> Found %s but missing %s"
@@ -587,7 +589,7 @@ def print_all_configs(configs, missing, warning):
   llen = 65  # line length
   for i, row in enumerate(configs):
     if i != 0:
-      print_text += "-"*llen + "\n"
+      print_text += six.ensure_str("-" * llen) + "\n"
 
     if isinstance(row[1], list):
       val = ", ".join(row[1])
@@ -629,7 +631,7 @@ def save_to_file(json_data, filename):
     print("filename: %s" % filename)
     filename += ".json"
 
-  with open(PATH_TO_DIR + "/" + filename, "w") as f:
+  with open(PATH_TO_DIR + "/" + six.ensure_str(filename), "w") as f:
     json.dump(json_data, f, sort_keys=True, indent=4)
 
   print(" Successfully wrote configs to file `%s`.\n" % (filename))
diff --git a/tensorflow/tools/tensorflow_builder/config_detector/data/cuda_compute_capability.py b/tensorflow/tools/tensorflow_builder/config_detector/data/cuda_compute_capability.py
index 4dba961b037..9d17dbc9178 100644
--- a/tensorflow/tools/tensorflow_builder/config_detector/data/cuda_compute_capability.py
+++ b/tensorflow/tools/tensorflow_builder/config_detector/data/cuda_compute_capability.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2019 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -42,6 +43,7 @@ import re
 from absl import app
 from absl import flags
 
+import six
 import six.moves.urllib.request as urllib
 
 FLAGS = flags.FLAGS
@@ -61,21 +63,18 @@ def retrieve_from_web(generate_csv=False):
     NVIDIA page. Order goes from top to bottom of the webpage content (.html).
   """
   url = "https://developer.nvidia.com/cuda-gpus"
-  source = urllib.urlopen(url)
+  source = urllib.request.urlopen(url)
   matches = []
   while True:
     line = source.readline()
     if "</html>" in line:
       break
     else:
-      gpu = re.search(
-          r"<a href=.*>([\w\S\s\d\[\]\,]+[^*])</a>(<a href=.*)?.*",
-          line
-      )
+      gpu = re.search(r"<a href=.*>([\w\S\s\d\[\]\,]+[^*])</a>(<a href=.*)?.*",
+                      six.ensure_str(line))
       capability = re.search(
           r"([\d]+).([\d]+)(/)?([\d]+)?(.)?([\d]+)?.*</td>.*",
-          line
-      )
+          six.ensure_str(line))
       if gpu:
         matches.append(gpu.group(1))
       elif capability:
@@ -155,15 +154,15 @@ def create_gpu_capa_map(match_list,
 
         gpu = ""
         cnt += 1
-        if len(gpu_capa.keys()) < cnt:
+        if len(list(gpu_capa.keys())) < cnt:
           mismatch_cnt += 1
-          cnt = len(gpu_capa.keys())
+          cnt = len(list(gpu_capa.keys()))
 
       else:
         gpu = match
 
   if generate_csv:
-    f_name = filename + ".csv"
+    f_name = six.ensure_str(filename) + ".csv"
     write_csv_from_dict(f_name, gpu_capa)
 
   return gpu_capa
@@ -179,8 +178,8 @@ def write_csv_from_dict(filename, input_dict):
     filename: String that is the output file name.
     input_dict: Dictionary that is to be written out to a `.csv` file.
   """
-  f = open(PATH_TO_DIR + "/data/" + filename, "w")
-  for k, v in input_dict.iteritems():
+  f = open(PATH_TO_DIR + "/data/" + six.ensure_str(filename), "w")
+  for k, v in six.iteritems(input_dict):
     line = k
     for item in v:
       line += "," + item
@@ -203,7 +202,7 @@ def check_with_golden(filename):
   Args:
     filename: String that is the name of the newly created file.
   """
-  path_to_file = PATH_TO_DIR + "/data/" + filename
+  path_to_file = PATH_TO_DIR + "/data/" + six.ensure_str(filename)
   if os.path.isfile(path_to_file) and os.path.isfile(CUDA_CC_GOLDEN_DIR):
     with open(path_to_file, "r") as f_new:
       with open(CUDA_CC_GOLDEN_DIR, "r") as f_golden:
diff --git a/tensorflow/tools/test/BUILD b/tensorflow/tools/test/BUILD
index 5607fe73361..b09acad5036 100644
--- a/tensorflow/tools/test/BUILD
+++ b/tensorflow/tools/test/BUILD
@@ -22,17 +22,19 @@ py_library(
     ],
     srcs_version = "PY2AND3",
     deps = [
+        "//tensorflow:tensorflow_py",
         "//tensorflow/core:protos_all_py",
         "//tensorflow/python:client",
         "//tensorflow/python:errors",
         "//tensorflow/python:platform",
+        "@six_archive//:six",
     ],
 )
 
 py_binary(
     name = "system_info",
     srcs = ["system_info.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     deps = [
         ":system_info_lib",
@@ -50,16 +52,20 @@ py_library(
         ":system_info_lib",
         "//tensorflow/core:protos_all_py",
         "//tensorflow/python:platform",
+        "@six_archive//:six",
     ],
 )
 
 py_binary(
     name = "run_and_gather_logs",
     srcs = ["run_and_gather_logs.py"],
-    python_version = "PY2",
+    python_version = "PY3",
     srcs_version = "PY2AND3",
     visibility = ["//visibility:public"],
-    deps = [":run_and_gather_logs_main_lib"],
+    deps = [
+        ":run_and_gather_logs_main_lib",
+        "@six_archive//:six",
+    ],
 )
 
 py_library(
@@ -72,6 +78,7 @@ py_library(
         "//tensorflow/core:protos_all_py",
         "//tensorflow/python:client_testlib",
         "//tensorflow/python:platform",
+        "@six_archive//:six",
     ],
 )
 
diff --git a/tensorflow/tools/test/gpu_info_lib.py b/tensorflow/tools/test/gpu_info_lib.py
index 3a4ff4fdff4..e9fb227baf1 100644
--- a/tensorflow/tools/test/gpu_info_lib.py
+++ b/tensorflow/tools/test/gpu_info_lib.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2016 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -21,6 +22,9 @@ from __future__ import print_function
 import ctypes as ct
 import platform
 
+import six
+from six.moves import range
+
 from tensorflow.core.util import test_log_pb2
 from tensorflow.python.framework import errors
 from tensorflow.python.platform import gfile
@@ -30,10 +34,11 @@ def _gather_gpu_devices_proc():
   """Try to gather NVidia GPU device information via /proc/driver."""
   dev_info = []
   for f in gfile.Glob("/proc/driver/nvidia/gpus/*/information"):
-    bus_id = f.split("/")[5]
-    key_values = dict(line.rstrip().replace("\t", "").split(":", 1)
-                      for line in gfile.GFile(f, "r"))
-    key_values = dict((k.lower(), v.strip(" ").rstrip(" "))
+    bus_id = six.ensure_str(f).split("/")[5]
+    key_values = dict(
+        six.ensure_str(line.rstrip()).replace("\t", "").split(":", 1)
+        for line in gfile.GFile(f, "r"))
+    key_values = dict((k.lower(), six.ensure_str(v).strip(" ").rstrip(" "))
                       for (k, v) in key_values.items())
     info = test_log_pb2.GPUInfo()
     info.model = key_values.get("model", "Unknown")
diff --git a/tensorflow/tools/test/run_and_gather_logs.py b/tensorflow/tools/test/run_and_gather_logs.py
index f6b25bbfccb..a1486826615 100644
--- a/tensorflow/tools/test/run_and_gather_logs.py
+++ b/tensorflow/tools/test/run_and_gather_logs.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2016 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -25,9 +26,10 @@ from string import maketrans
 import sys
 import time
 
+import six
+
 from google.protobuf import json_format
 from google.protobuf import text_format
-
 from tensorflow.core.util import test_log_pb2
 from tensorflow.python.platform import app
 from tensorflow.python.platform import gfile
@@ -83,8 +85,9 @@ def main(unused_args):
   if FLAGS.test_log_output_filename:
     file_name = FLAGS.test_log_output_filename
   else:
-    file_name = (name.strip("/").translate(maketrans("/:", "__")) +
-                 time.strftime("%Y%m%d%H%M%S", time.gmtime()))
+    file_name = (
+        six.ensure_str(name).strip("/").translate(maketrans("/:", "__")) +
+        time.strftime("%Y%m%d%H%M%S", time.gmtime()))
   if FLAGS.test_log_output_use_tmpdir:
     tmpdir = test.get_temp_dir()
     output_path = os.path.join(tmpdir, FLAGS.test_log_output_dir, file_name)
@@ -92,7 +95,8 @@ def main(unused_args):
     output_path = os.path.join(
         os.path.abspath(FLAGS.test_log_output_dir), file_name)
   json_test_results = json_format.MessageToJson(test_results)
-  gfile.GFile(output_path + ".json", "w").write(json_test_results)
+  gfile.GFile(six.ensure_str(output_path) + ".json",
+              "w").write(json_test_results)
   tf_logging.info("Test results written to: %s" % output_path)
 
 
diff --git a/tensorflow/tools/test/run_and_gather_logs_lib.py b/tensorflow/tools/test/run_and_gather_logs_lib.py
index 3b4921bb983..f629e3a10b6 100644
--- a/tensorflow/tools/test/run_and_gather_logs_lib.py
+++ b/tensorflow/tools/test/run_and_gather_logs_lib.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2016 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -25,6 +26,8 @@ import subprocess
 import tempfile
 import time
 
+import six
+
 from tensorflow.core.util import test_log_pb2
 from tensorflow.python.platform import gfile
 from tensorflow.tools.test import gpu_info_lib
@@ -118,12 +121,15 @@ def run_and_gather_logs(name, test_name, test_args,
     IOError: If there are problems gathering test log output from the test.
     MissingLogsError: If we couldn't find benchmark logs.
   """
-  if not (test_name and test_name.startswith("//") and ".." not in test_name and
-          not test_name.endswith(":") and not test_name.endswith(":all") and
-          not test_name.endswith("...") and len(test_name.split(":")) == 2):
+  if not (test_name and six.ensure_str(test_name).startswith("//") and
+          ".." not in test_name and not six.ensure_str(test_name).endswith(":")
+          and not six.ensure_str(test_name).endswith(":all") and
+          not six.ensure_str(test_name).endswith("...") and
+          len(six.ensure_str(test_name).split(":")) == 2):
     raise ValueError("Expected test_name parameter with a unique test, e.g.: "
                      "--test_name=//path/to:test")
-  test_executable = test_name.rstrip().strip("/").replace(":", "/")
+  test_executable = six.ensure_str(test_name.rstrip()).strip("/").replace(
+      ":", "/")
 
   if gfile.Exists(os.path.join("bazel-bin", test_executable)):
     # Running in standalone mode from core of the repository
@@ -136,14 +142,17 @@ def run_and_gather_logs(name, test_name, test_args,
   gpu_config = gpu_info_lib.gather_gpu_devices()
   if gpu_config:
     gpu_name = gpu_config[0].model
-    gpu_short_name_match = re.search(r"Tesla (K40|K80|P100|V100)", gpu_name)
+    gpu_short_name_match = re.search(r"Tesla (K40|K80|P100|V100)",
+                                     six.ensure_str(gpu_name))
     if gpu_short_name_match:
       gpu_short_name = gpu_short_name_match.group(0)
-      test_adjusted_name = name + "|" + gpu_short_name.replace(" ", "_")
+      test_adjusted_name = six.ensure_str(name) + "|" + gpu_short_name.replace(
+          " ", "_")
 
   temp_directory = tempfile.mkdtemp(prefix="run_and_gather_logs")
-  mangled_test_name = (test_adjusted_name.strip("/")
-                       .replace("|", "_").replace("/", "_").replace(":", "_"))
+  mangled_test_name = (
+      six.ensure_str(test_adjusted_name).strip("/").replace("|", "_").replace(
+          "/", "_").replace(":", "_"))
   test_file_prefix = os.path.join(temp_directory, mangled_test_name)
   test_file_prefix = "%s." % test_file_prefix
 
diff --git a/tensorflow/tools/test/system_info_lib.py b/tensorflow/tools/test/system_info_lib.py
index 59a30f9a368..ccdd8f3bf2a 100644
--- a/tensorflow/tools/test/system_info_lib.py
+++ b/tensorflow/tools/test/system_info_lib.py
@@ -1,3 +1,4 @@
+# Lint as: python2, python3
 # Copyright 2016 The TensorFlow Authors. All Rights Reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -29,6 +30,8 @@ import socket
 # OSS tree.  They are installable via pip.
 import cpuinfo
 import psutil
+
+import six
 # pylint: enable=g-bad-import-order
 
 from tensorflow.core.util import test_log_pb2
@@ -81,7 +84,8 @@ def gather_cpu_info():
   # Gather num_cores_allowed
   try:
     with gfile.GFile('/proc/self/status', 'rb') as fh:
-      nc = re.search(r'(?m)^Cpus_allowed:\s*(.*)$', fh.read().decode('utf-8'))
+      nc = re.search(r'(?m)^Cpus_allowed:\s*(.*)$',
+                     six.ensure_text(fh.read(), 'utf-8'))
     if nc:  # e.g. 'ff' => 8, 'fff' => 12
       cpu_info.num_cores_allowed = (
           bin(int(nc.group(1).replace(',', ''), 16)).count('1'))