Some OSes return filenames as Unicode, some as UTF-8.

Attempting to encode('utf-8') a non-Unicode string raises an exception.
This CL changes the behavior to encode('utf-8') iff the filename is a Unicode string.
Change: 130720340

1 files changed, 2 insertions, 1 deletions

diff --git a/tensorflow/examples/image_retraining/retrain.py b/tensorflow/examples/image_retraining/retrain.py
index b009262d6e..056664778d 100644
--- a/tensorflow/examples/image_retraining/retrain.py
+++ b/tensorflow/examples/image_retraining/retrain.py
@@ -80,6 +80,7 @@ import tensorflow as tf
 from tensorflow.python.framework import graph_util
 from tensorflow.python.framework import tensor_shape
 from tensorflow.python.platform import gfile
+from tensorflow.python.util import compat
 
 
 import struct
@@ -226,7 +227,7 @@ def create_image_lists(image_dir, testing_percentage, validation_percentage):
       # To do that, we need a stable way of deciding based on just the file name
       # itself, so we do a hash of that and then use that to generate a
       # probability value that we use to assign it.
-      hash_name_hashed = hashlib.sha1(hash_name.encode('utf-8')).hexdigest()
+      hash_name_hashed = hashlib.sha1(compat.as_bytes(hash_name)).hexdigest()
       percentage_hash = (int(hash_name_hashed, 16) % (65536)) * (100 / 65535.0)
       if percentage_hash < validation_percentage:
         validation_images.append(base_name)