experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[tf.data] Iterator and data/nest documentation fixes

Browse Source 
PiperOrigin-RevId: 171165796
This commit is contained in:
A. Unique TensorFlower 2017-10-05 09:50:17 -07:00 committed by TensorFlower Gardener
parent 09fa4a4e35
commit 7e7d55c0f5
2 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
tensorflow/python/data/ops/iterator_ops.py
View File

@ -138,21 +138,21 @@ class Iterator(object):
This method allows you to define a "feedable" iterator where you can choose This method allows you to define a "feedable" iterator where you can choose
between concrete iterators by feeding a value in a @{tf.Session.run} call. between concrete iterators by feeding a value in a @{tf.Session.run} call.
In that case, `string_handle` would a @{tf.placeholder}, and you would feed In that case, `string_handle` would a @{tf.placeholder}, and you would feed
it with the value of @{tf.contrib.data.Iterator.string_handle} in each step. it with the value of @{tf.data.Iterator.string_handle} in each step.
For example, if you had two iterators that marked the current position in For example, if you had two iterators that marked the current position in
a training dataset and a test dataset, you could choose which to use in a training dataset and a test dataset, you could choose which to use in
each step as follows: each step as follows:
```python ```python
train_iterator = tf.contrib.data.Dataset(...).make_one_shot_iterator() train_iterator = tf.data.Dataset(...).make_one_shot_iterator()
train_iterator_handle = sess.run(train_iterator.string_handle()) train_iterator_handle = sess.run(train_iterator.string_handle())
test_iterator = tf.contrib.data.Dataset(...).make_one_shot_iterator() test_iterator = tf.data.Dataset(...).make_one_shot_iterator()
test_iterator_handle = sess.run(test_iterator.string_handle()) test_iterator_handle = sess.run(test_iterator.string_handle())
handle = tf.placeholder(tf.string, shape=[]) handle = tf.placeholder(tf.string, shape=[])
iterator = tf.contrib.data.Iterator.from_string_handle( iterator = tf.data.Iterator.from_string_handle(
handle, train_iterator.output_types) handle, train_iterator.output_types)
next_element = iterator.get_next() next_element = iterator.get_next()

2
tensorflow/python/data/util/nest.py
View File

@ -106,7 +106,7 @@ def is_sequence(seq):
NOTE(mrry): This differs from `tensorflow.python.util.nest.is_sequence()`, NOTE(mrry): This differs from `tensorflow.python.util.nest.is_sequence()`,
which *does* treat a Python list as a sequence. For ergonomic which *does* treat a Python list as a sequence. For ergonomic
reasons, `tf.contrib.data` users would prefer to treat lists as reasons, `tf.data` users would prefer to treat lists as
implict `tf.Tensor` objects, and dicts as (nested) sequences. implict `tf.Tensor` objects, and dicts as (nested) sequences.
Args: Args: