See also: #6950 and #6865

Update the TensorBoard README link on r0.9.

As per comments from nvidia-docker dev @3XX0, hardcoding "1" should be
reasonably safe. The TF_CUDA_VERSION variable from the configure script
is not appropriate here (it will contain something like "7.0" or "7.5",
while the libcuda soname major version number should be "1").

For his efforts in helping to root cause and improve CPU threading performance.

R0.9: Fixes URLs and headings

Wide and Deep docs update

Change: 126101067

Adds tf.contrib.learn quickstart and tf-wide (and deep) stuff to tutorial left nav and index page (and fixes an URL in the wide overview). Also breaks the left nav and index page into sections, because the list of tutorials is getting long.
Change: 126105974

Change: 126082003

Change: 125636045

Change: 125463524

Fix the TensorBoard README link on the how to page.

Change: 125799092

Update build file to fix #2703

update docs

Use string_to_hash_bucket_fast in feature_column

R0.9 time fix

direct_session_with_tracking_alloc_test.
Change: 124970170

Change: 125205938

Version bump: 0.9.0rc0 --> 0.9.0

Updated feature_column to use string_to_hash_bucket_fast instead of deprecated string_to_hash_bucket.

Fixes to TensorArray and Functional ops

Fix genrule Python execution problem with Python3.

- TensorArray shape inference now works correctly for scalar elements
- TensorArrays each now get a unique name at runtime, per step.  This means
  that they can be used in nested functional ops (e.g. tf.scan(tf.scan(...)))

Change: 125110643

create variables. Also add missing output_size in EmbeddingWrapper (#2852).
Change: 125022470

Make RNN api public.

Change: 124202095

Change: 124305799

Cherry-picking stability and doc fixes for r0.9