tensorflow · foodraider · May 4, 2017 · Apr 14, 2017 · May 8, 2017 · May 12, 2017
diff --git a/tensorflow/docs_src/performance/benchmarks.md b/tensorflow/docs_src/performance/benchmarks.md
diff --git a/tensorflow/docs_src/performance/index.md b/tensorflow/docs_src/performance/index.md
@@ -2,11 +2,19 @@
 
 Performance is often a significant issue when training a machine learning
 model.  This section explains various ways to optimize performance.  Start
-your investigation with the following guide:
+your investigation with the @{$performance_guide$Performance Guide} and then go
+deeper with techniques detailed in @{$performance_models$High-Performance Models}:
 
-  * @{$performance_guide$Performance}, which contains a collection of best
+  * @{$performance_guide$Performance Guide}, which contains a collection of best
     practices for optimizing your TensorFlow code.
 
+  * @{$performance_models$High-Performance Models}, which contains a collection
+    of advanced techniques to build highly scalable models targeting different
+    system types and network topologies.
+
+  * @{$benchmarks$Benchmarks}, which contains a collection of benchmark
+    results.
+
 XLA (Accelerated Linear Algebra) is an experimental compiler for linear
 algebra that optimizes TensorFlow computations. The following guides explore
 XLA:

diff --git a/tensorflow/docs_src/performance/leftnav_files b/tensorflow/docs_src/performance/leftnav_files
@@ -1,9 +1,12 @@
 performance_guide.md
+performance_models.md
+benchmarks.md
+quantization.md
+>>>
 xla/index.md
 xla/broadcasting.md
 xla/developing_new_backend.md
 xla/jit.md
 xla/operation_semantics.md
 xla/shapes.md
 xla/tfcompile.md
-quantization.md
diff --git a/tensorflow/docs_src/performance/performance_guide.md b/tensorflow/docs_src/performance/performance_guide.md
@@ -1,8 +1,10 @@
-# Performance
+# Performance Guide
 
 This guide contains a collection of best practices for optimizing your
 TensorFlow code. The best practices apply to both new and experienced
-Tensorflow users.
+Tensorflow users.  As a complement to the best practices in this document, the
+@{$performance_models$High-Performance Models} document links to example code
+and details for creating models that scale on a variety of hardware.
 
 ## Best Practices
 While optimizing implementations of different types of models can be different,
@@ -73,7 +75,7 @@ Unless for a special circumstance or for example code, do not feed data
 into the session from Python variables, e.g. `dictionary`.
 
 ```python
-# This will result in poor performance.
+# Using feed_dict often results in suboptimal performance when using large inputs.
 sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys})
 ```
 
@@ -141,3 +143,4 @@ bn = tf.contrib.layers.batch_norm(
 The non-fused batch norm does computations using several individual Ops. Fused
 batch norm combines the individual operations into a single kernel, which runs
 faster.
+
diff --git a/tensorflow/docs_src/performance/performance_models.md b/tensorflow/docs_src/performance/performance_models.md
diff --git a/tensorflow/examples/tutorials/estimators/abalone.py b/tensorflow/examples/tutorials/estimators/abalone.py
@@ -134,12 +134,22 @@ def main(unused_argv):
 
   # Instantiate Estimator
   nn = tf.contrib.learn.Estimator(model_fn=model_fn, params=model_params)
-
+
+  def get_train_inputs():
+    x = tf.constant(training_set.data)
+    y = tf.constant(training_set.target)
+    return x, y
+
   # Fit
-  nn.fit(x=training_set.data, y=training_set.target, steps=5000)
+  nn.fit(input_fn=get_train_inputs, steps=5000)
 
   # Score accuracy
-  ev = nn.evaluate(x=test_set.data, y=test_set.target, steps=1)
+  def get_test_inputs():
+    x = tf.constant(test_set.data)
+    y = tf.constant(test_set.target)
+    return x, y
+
+  ev = nn.evaluate(input_fn=get_test_inputs, steps=1)
   print("Loss: %s" % ev["loss"])
   print("Root Mean Squared Error: %s" % ev["rmse"])
 

diff --git a/tensorflow/python/estimator/export/export_output.py b/tensorflow/python/estimator/export/export_output.py
@@ -67,7 +67,7 @@ class ClassificationOutput(ExportOutput):
   """
 
   def __init__(self, scores=None, classes=None):
-    """Constructor for `ClassifyOutput`.
+    """Constructor for `ClassificationOutput`.
 
     Args:
       scores: A float `Tensor` giving scores (sometimes but not always

diff --git a/tensorflow/python/ops/math_ops.py b/tensorflow/python/ops/math_ops.py
@@ -209,16 +209,27 @@ def abs(x, name=None):
   Given a tensor of real numbers `x`, this operation returns a tensor
   containing the absolute value of each element in `x`. For example, if x is
   an input element and y is an output element, this operation computes
-  \\\\(y = |x|\\\\).
+  \\(y = |x|\\).
+
+  Given a tensor `x` of complex numbers, this operation returns a tensor of type
+  `float32` or `float64` that is the absolute value of each element in `x`. All
+  elements in `x` must be complex numbers of the form \\(a + bj\\). The
+  absolute value is computed as \\( \sqrt{a^2 + b^2}\\).  For example:
+  ```
+  # tensor 'x' is [[-2.25 + 4.75j], [-3.25 + 5.75j]]
+  tf.complex_abs(x) ==> [5.25594902, 6.60492229]
+  ```
 
   Args:
-    x: A `Tensor` or `SparseTensor` of type `float32`, `float64`, `int32`, or
-      `int64`.
+    x: A `Tensor` or `SparseTensor` of type `float32`, `float64`, `int32`,
+      `int64`, `complex64` or `complex128`.
     name: A name for the operation (optional).
 
   Returns:
     A `Tensor` or `SparseTensor` the same size and type as `x` with absolute
       values.
+    Note, for `complex64` or `complex128' input, the returned `Tensor` will be
+      of type `float32` or `float64`, respectively.
   """
   with ops.name_scope(name, "Abs", [x]) as name:
     if isinstance(x, sparse_tensor.SparseTensor):

diff --git a/tensorflow/tensorboard/dist/index.html b/tensorflow/tensorboard/dist/index.html
@@ -27,6 +27,6 @@
     <link rel="import" href="dist/tf-tensorboard.html">
   </head>
   <body>
-    <tf-tensorboard></tf-tensorboard>
+    <tf-tensorboard use-hash></tf-tensorboard>
   </body>
 </html>