platform/upstream/tensorflow.git
6 years agoMerge pull request #18765 from Intel-tensorflow/agramesh/String_fix
Martin Wicke [Mon, 23 Apr 2018 15:41:14 +0000 (08:41 -0700)]
Merge pull request #18765 from Intel-tensorflow/agramesh/String_fix

[INTEL MKL] Fix compilation failure.

6 years agoMerge pull request #17556 from kbsriram/strided-slice-grad
Martin Wicke [Mon, 23 Apr 2018 06:53:22 +0000 (23:53 -0700)]
Merge pull request #17556 from kbsriram/strided-slice-grad

C++ gradient for StridedSlice

6 years agoMerge pull request #15354 from facaiy/ENH/scatter_nd_supports_string
Martin Wicke [Mon, 23 Apr 2018 06:52:55 +0000 (23:52 -0700)]
Merge pull request #15354 from facaiy/ENH/scatter_nd_supports_string

CPU: Support for DT_STRING type in ScatterNd

6 years agoMerge pull request #18754 from ADiegoCAlonso/fix-numerical-warning
Martin Wicke [Mon, 23 Apr 2018 06:52:38 +0000 (23:52 -0700)]
Merge pull request #18754 from ADiegoCAlonso/fix-numerical-warning

Fix numerical warning in the Iris example

6 years agoMerge pull request #18728 from jerry4897/master
Martin Wicke [Mon, 23 Apr 2018 06:52:23 +0000 (23:52 -0700)]
Merge pull request #18728 from jerry4897/master

Typo?

6 years agoMerge pull request #18698 from imsheridan/fix_api_rendering
Martin Wicke [Mon, 23 Apr 2018 06:52:08 +0000 (23:52 -0700)]
Merge pull request #18698 from imsheridan/fix_api_rendering

Fix code block rendering in several api definitions

6 years agoMerge pull request #18715 from yongtang/18712-tf.count_nonzero
Martin Wicke [Mon, 23 Apr 2018 06:51:41 +0000 (23:51 -0700)]
Merge pull request #18715 from yongtang/18712-tf.count_nonzero

Support string tensors for `tf.count_nonzero`

6 years agoMerge pull request #18753 from ADiegoCAlonso/fix-tutorials-directories-are-not-python...
Martin Wicke [Mon, 23 Apr 2018 06:48:54 +0000 (23:48 -0700)]
Merge pull request #18753 from ADiegoCAlonso/fix-tutorials-directories-are-not-python-packages

Fix tutorials directories are not Python packages

6 years agoMerge pull request #18766 from hedenface/patch-1
Martin Wicke [Mon, 23 Apr 2018 03:36:38 +0000 (20:36 -0700)]
Merge pull request #18766 from hedenface/patch-1

update $ source spacing

6 years agoMerge pull request #18778 from yongtang/04222018-squeeze_dims
Martin Wicke [Mon, 23 Apr 2018 03:33:36 +0000 (20:33 -0700)]
Merge pull request #18778 from yongtang/04222018-squeeze_dims

Change from squeeze_dims to axis when calling tf.squeeze

6 years agoMerge pull request #18760 from yongtang/04212018-gitignore
Martin Wicke [Mon, 23 Apr 2018 03:33:00 +0000 (20:33 -0700)]
Merge pull request #18760 from yongtang/04212018-gitignore

Update .gitignore for cmake generated files

6 years agoFix warnings in tf.contrib.distributions with squeeze_dims
Yong Tang [Sun, 22 Apr 2018 17:57:31 +0000 (17:57 +0000)]
Fix warnings in tf.contrib.distributions with squeeze_dims

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoFix warnings in tf.contrib.factorization
Yong Tang [Sun, 22 Apr 2018 17:57:06 +0000 (17:57 +0000)]
Fix warnings in tf.contrib.factorization

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoFix tf.contrib.layers warnings where squeeze_dims were used with tf.squeeze
Yong Tang [Sun, 22 Apr 2018 17:56:09 +0000 (17:56 +0000)]
Fix tf.contrib.layers warnings where squeeze_dims were used with tf.squeeze

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoFix squeeze_dims warnings in tf.contrib.learn
Yong Tang [Sun, 22 Apr 2018 17:55:35 +0000 (17:55 +0000)]
Fix squeeze_dims warnings in tf.contrib.learn

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoFix warnings in tf.contrib.tensor_forest
Yong Tang [Sun, 22 Apr 2018 17:55:06 +0000 (17:55 +0000)]
Fix warnings in tf.contrib.tensor_forest

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoFix tf.contrib.timeseries warnings related to squeeze_dims
Yong Tang [Sun, 22 Apr 2018 17:54:26 +0000 (17:54 +0000)]
Fix tf.contrib.timeseries warnings related to squeeze_dims

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoFix warnings in reduce_join_op_test.py
Yong Tang [Sun, 22 Apr 2018 17:53:58 +0000 (17:53 +0000)]
Fix warnings in reduce_join_op_test.py

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoFix warning in tf.nn ops where squeeze_dims was used with tf.squeeze
Yong Tang [Sun, 22 Apr 2018 17:53:10 +0000 (17:53 +0000)]
Fix warning in tf.nn ops where squeeze_dims was used with tf.squeeze

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoFix warning in image related ops.
Yong Tang [Sun, 22 Apr 2018 17:52:31 +0000 (17:52 +0000)]
Fix warning in image related ops.

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoChange from squeeze_dims to axis when calling tf.squeeze
Yong Tang [Sun, 22 Apr 2018 17:49:13 +0000 (17:49 +0000)]
Change from squeeze_dims to axis when calling tf.squeeze

The `squeeze_dims` in `tf.squeeze` has been deprecated in favor
of `axis` while many places still use `squeeze_dims`. That
generates lots of warnings.

This fix switches from `squeeze_dims` to `axis` to remove those warnings.

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoUpdate docs to add note and examples for tf.count_nonzero with string
Yong Tang [Sun, 22 Apr 2018 14:43:21 +0000 (14:43 +0000)]
Update docs to add note and examples for tf.count_nonzero with string

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoMerge pull request #18694 from yongtang/18692-unstack-doc
Martin Wicke [Sun, 22 Apr 2018 06:01:00 +0000 (23:01 -0700)]
Merge pull request #18694 from yongtang/18692-unstack-doc

Update docs for tf.unstack with respect to numpy compatibility

6 years agoMerge pull request #18713 from ManHyuk/fix_typo
Martin Wicke [Sun, 22 Apr 2018 05:54:45 +0000 (22:54 -0700)]
Merge pull request #18713 from ManHyuk/fix_typo

Fix typo

6 years agoMerge pull request #18757 from foo0x29a/master
Martin Wicke [Sun, 22 Apr 2018 05:18:10 +0000 (22:18 -0700)]
Merge pull request #18757 from foo0x29a/master

Fix typo

6 years agoMerge pull request #18168 from imsheridan/fix_adam_math
Martin Wicke [Sun, 22 Apr 2018 05:08:48 +0000 (22:08 -0700)]
Merge pull request #18168 from imsheridan/fix_adam_math

Fix adam optimizer related math equation rendering format

6 years agoupdate $ source spacing
Bryan Heden [Sun, 22 Apr 2018 00:45:42 +0000 (19:45 -0500)]
update $ source spacing

When viewing install_linux, the spacing was off for 'Next Steps' section.

6 years agoChanged calls to the depreacted StringPiece::contains with str_util::StrContains
AG Ramesh [Sat, 21 Apr 2018 22:25:37 +0000 (15:25 -0700)]
Changed calls to the depreacted StringPiece::contains with str_util::StrContains

6 years agoUpdate .gitignore for cmake generated files
Yong Tang [Sat, 21 Apr 2018 18:42:03 +0000 (18:42 +0000)]
Update .gitignore for cmake generated files

After running cmake on Linux with:
```
tensorflow/tools/ci_build/ci_build.sh CMAKE tensorflow/tools/ci_build/builds/cmake.sh
```

the following file is left:
```
ubuntu@ubuntu:~/tensorflow$ git status
On branch master
Your branch is up-to-date with 'origin/master'.
Untracked files:
  (use "git add <file>..." to include in what will be committed)

        api_init_files_list.txt

nothing added to commit but untracked files present (use "git add" to track)
ubuntu@ubuntu:~/tensorflow$
```

This fix updates the .gitignore file so that cmake generated files
is not added with git inadvertently.

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agofix typo
foo0x29a [Sat, 21 Apr 2018 16:30:52 +0000 (13:30 -0300)]
fix typo

6 years agoSpecify float32 as float type instead of float64
ADiegoCAlonso [Sat, 21 Apr 2018 09:56:10 +0000 (11:56 +0200)]
Specify float32 as float type instead of float64

6 years agoAdd __init__py
ADiegoCAlonso [Sat, 21 Apr 2018 09:43:51 +0000 (11:43 +0200)]
Add __init__py

6 years agoFix casting in Keras estimator (#18104)
Frédéric Branchaud-Charron [Fri, 20 Apr 2018 23:10:41 +0000 (19:10 -0400)]
Fix casting in Keras estimator (#18104)

6 years agoMKLDNN: conv2d forward DNN primitive reuse enhancement (#17943)
Guozhong Zhuang [Fri, 20 Apr 2018 20:43:06 +0000 (13:43 -0700)]
MKLDNN: conv2d forward DNN primitive reuse enhancement (#17943)

* Enable conv2d fwd primitive reuse

* coding style change based on suggestions from TF team

* minor code style fix

* refactor conv2d primitive reuse class and enhance key creation utility

* refactor by introducing ConvFwdDimensions structure

* change 'Execute' method to be a template one per PR review suggestion

* Per PR review suggestion, update DnnOp class to declared related  method as abstract ones

* refactor AddAsKey method - template for scalar value and remove Execute()which is not used yet

* rename padding_l/_r/pl/pr to padding_left or padding_right as recommended

* parameter and variable renaming - to make them more explicit

6 years ago[INTEL MKLDNN]: Upgrade mkldnn version to v13 (#18508)
jinghuangintel [Fri, 20 Apr 2018 19:20:00 +0000 (12:20 -0700)]
[INTEL MKLDNN]: Upgrade mkldnn version to v13 (#18508)

* upgrade mkldnn version to v13

* upgrade mkldnn version to v13 for all platforms

6 years agoMerge pull request #18663 from imsheridan/fix_community_render
Martin Wicke [Fri, 20 Apr 2018 18:06:46 +0000 (11:06 -0700)]
Merge pull request #18663 from imsheridan/fix_community_render

Fix incorrect format in community/documentation.md

6 years agoMerge pull request #18714 from yifeif/branch_193610264
Yifei Feng [Fri, 20 Apr 2018 17:08:40 +0000 (10:08 -0700)]
Merge pull request #18714 from yifeif/branch_193610264

Branch 193610264

6 years agoUpdate c_api_test.cc
SukHwan Kim [Fri, 20 Apr 2018 09:24:52 +0000 (18:24 +0900)]
Update c_api_test.cc

Typo

6 years agoMerge remote-tracking branch 'upstream/master' into ENH/scatter_nd_supports_string
Yan Facai (颜发才) [Fri, 20 Apr 2018 05:38:03 +0000 (13:38 +0800)]
Merge remote-tracking branch 'upstream/master' into ENH/scatter_nd_supports_string

6 years agoTST: byte string for python3
Yan Facai (颜发才) [Fri, 20 Apr 2018 05:33:05 +0000 (13:33 +0800)]
TST: byte string for python3

6 years agoUpdate doc string for tf.count_nonzero to add string type
Yong Tang [Fri, 20 Apr 2018 02:55:31 +0000 (02:55 +0000)]
Update doc string for tf.count_nonzero to add string type

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoFix unwanted typo caused protobuf load failure
imsheridan [Fri, 20 Apr 2018 03:08:34 +0000 (11:08 +0800)]
Fix unwanted typo caused protobuf load failure

6 years agoRevert part of tensorflow/core/grappler/optimizers/meta_optimizer.cc
Yifei Feng [Fri, 20 Apr 2018 02:56:09 +0000 (19:56 -0700)]
Revert part of tensorflow/core/grappler/optimizers/meta_optimizer.cc
from #18479.

6 years agoAdd n-D test cases for better coverage
Yong Tang [Fri, 20 Apr 2018 02:53:57 +0000 (02:53 +0000)]
Add n-D test cases for better coverage

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoAdd test cases with axis and keepdims for tf.count_nonzero and string
Yong Tang [Fri, 20 Apr 2018 02:51:54 +0000 (02:51 +0000)]
Add test cases with axis and keepdims for tf.count_nonzero and string

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoAdd test case for tf.string support with tf.count_nonzero
Yong Tang [Fri, 20 Apr 2018 02:48:15 +0000 (02:48 +0000)]
Add test case for tf.string support with tf.count_nonzero

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoSupport string tensors for tf.count_nonzero
Yong Tang [Fri, 20 Apr 2018 02:40:37 +0000 (02:40 +0000)]
Support string tensors for tf.count_nonzero

This fix tries to address the issue raised in 18712 where
`tf.count_nonzero` does not support string tensors.

The implementation of `tf.count_nonzero` relies on `tf.not_equal`
which actually support string tensors. The reason the string
tensor does not work is because `tf.count_nonzero` created
a numpy type `zero` which uses `input_tensor.dtype.as_numpy_dtype()`.
The numpy type `zero` is then passed to `tf.not_equal (which converts
numpy `zero` into a tensor zero). However,
`input_tensor.dtype.as_numpy_dtype()` will converts tf.string to
numpy.object thus the exception.

But that is not necessary as `zero` could be created
with `tf.zeros` directly without back and forth conversion
to numpy.

This fix fixes the issue.

This fix fixes 18712.

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
6 years agoMerge commit for internal changes
Yifei Feng [Fri, 20 Apr 2018 02:39:04 +0000 (19:39 -0700)]
Merge commit for internal changes

6 years agoSupport variable parameter structure in TPU distribution strategy.
Igor Saprykin [Fri, 20 Apr 2018 02:11:37 +0000 (19:11 -0700)]
Support variable parameter structure in TPU distribution strategy.

TPUStrategy is added to a few more tests.

There appears to be an issue with the batch norm test in minimize_loss_test where the moving averages stay at 0.  I'm trying to resolve that separately as the next CL.

PiperOrigin-RevId: 193610264

6 years agoAutomated g4 rollback of changelist 193593761
Justin Lebar [Fri, 20 Apr 2018 02:03:03 +0000 (19:03 -0700)]
Automated g4 rollback of changelist 193593761

PiperOrigin-RevId: 193609407

6 years agoInternal change.
Anna R [Fri, 20 Apr 2018 01:45:47 +0000 (18:45 -0700)]
Internal change.

PiperOrigin-RevId: 193608140

6 years ago[XLA] Fix a bug in the name_uniquer.
A. Unique TensorFlower [Fri, 20 Apr 2018 01:29:00 +0000 (18:29 -0700)]
[XLA] Fix a bug in the name_uniquer.

The problem happens because the name_uniquer stripped away the numeric suffix if it <=0. The solution is, if there was a numeric suffix, the result should also have a numeric suffix.

PiperOrigin-RevId: 193606838

6 years agoFix the error looking for libhdfs.so, Mac OS using libhdfs.dylib (#18486)
Jiajia Li [Fri, 20 Apr 2018 01:22:26 +0000 (09:22 +0800)]
Fix the error looking for libhdfs.so, Mac OS using libhdfs.dylib (#18486)

6 years agoFix incorrect math equation renderings broken by backtick (#18386)
ImSheridan [Fri, 20 Apr 2018 01:20:38 +0000 (09:20 +0800)]
Fix incorrect math equation renderings broken by backtick (#18386)

* Fix incorrect `` typo format

* Remove breaking ``` for math equations

* fix one more typo

* fix more math equation broken ` typos in py

6 years agoOptimize Graph function library.
A. Unique TensorFlower [Fri, 20 Apr 2018 01:17:02 +0000 (18:17 -0700)]
Optimize Graph function library.

PiperOrigin-RevId: 193605910

6 years agoRespect any device filters in {Create,Delete}WorkerSessions().
Derek Murray [Fri, 20 Apr 2018 01:12:57 +0000 (18:12 -0700)]
Respect any device filters in {Create,Delete}WorkerSessions().

This is another step towards enabling us to turn on explicit worker
sessions for all master sessions.

PiperOrigin-RevId: 193605565

6 years agointernal
Lukasz Kaiser [Fri, 20 Apr 2018 00:39:09 +0000 (17:39 -0700)]
internal
END_PUBLIC

BEGIN_PUBLIC
Automated g4 rollback of changelist 193571934

PiperOrigin-RevId: 193602050

6 years agoMerge pull request #18707 from joel-shor/master
Joel Shor [Fri, 20 Apr 2018 00:40:59 +0000 (03:40 +0300)]
Merge pull request #18707 from joel-shor/master

[tf.data] Allow `sample_from_datasets` to accept a tf.Dataset object …

6 years agoInternal testing changes
Austin Anderson [Fri, 20 Apr 2018 00:30:49 +0000 (17:30 -0700)]
Internal testing changes

PiperOrigin-RevId: 193601134

6 years agoUpdate DecodeProtoOp so that it returns explicitly specified default values for missi...
A. Unique TensorFlower [Fri, 20 Apr 2018 00:27:04 +0000 (17:27 -0700)]
Update DecodeProtoOp so that it returns explicitly specified default values for missing fields.

PiperOrigin-RevId: 193600735

6 years agoAdd support for non-Tensor args in recompute_grad
A. Unique TensorFlower [Fri, 20 Apr 2018 00:26:41 +0000 (17:26 -0700)]
Add support for non-Tensor args in recompute_grad

Previously, the function decorated by recompute_grad had to have a signature that contained only positional arguments, and all those arguments had to be Tensors. Most "layers" users define however have non-Tensor arguments (for example, various hyperparameters) and often have keyword arguments as well. This change allows a user to use whatever function signature they wish while being explicit about which arguments are Tensors.

PiperOrigin-RevId: 193600682

6 years agofix typo
manhyuk [Fri, 20 Apr 2018 00:28:45 +0000 (09:28 +0900)]
fix typo

6 years agofix typo
manhyuk [Fri, 20 Apr 2018 00:28:37 +0000 (09:28 +0900)]
fix typo

6 years agoMinor code refactoring.
Jiri Simsa [Fri, 20 Apr 2018 00:21:50 +0000 (17:21 -0700)]
Minor code refactoring.

PiperOrigin-RevId: 193600173

6 years ago[XLA] De-unique_ptr-ify ShapedBuffer and ScopedShapedBuffer.
Justin Lebar [Fri, 20 Apr 2018 00:18:10 +0000 (17:18 -0700)]
[XLA] De-unique_ptr-ify ShapedBuffer and ScopedShapedBuffer.

These are already notionally equivalent to T* and unique_ptr<T>, so
having a unique_ptr of a {Scoped,}ShapedBuffer is pretty redundant.

Also clean up the ScopedShapedBuffer API a bit.

PiperOrigin-RevId: 193599773

6 years agoRemoved deprecated methods from tensorflow::StringPiece.
A. Unique TensorFlower [Fri, 20 Apr 2018 00:17:05 +0000 (17:17 -0700)]
Removed deprecated methods from tensorflow::StringPiece.

This will allow tensorflow::StringPiece to be more easily replaced with absl::string_view as absl::string_view does not contain those methods.

PiperOrigin-RevId: 193599651

6 years agoMerge pull request #18582 from ntenenz/master
Martin Wicke [Fri, 20 Apr 2018 00:11:18 +0000 (17:11 -0700)]
Merge pull request #18582 from ntenenz/master

Remove conditional scope logic now that "current_arg_scope" exists in contrib

6 years agoMerge pull request #18686 from ManHyuk/fix_typo
Martin Wicke [Fri, 20 Apr 2018 00:08:39 +0000 (17:08 -0700)]
Merge pull request #18686 from ManHyuk/fix_typo

fix typo

6 years agoMerge pull request #18646 from joeyearsley/patch-4
Martin Wicke [Fri, 20 Apr 2018 00:06:20 +0000 (17:06 -0700)]
Merge pull request #18646 from joeyearsley/patch-4

Update debugger.md

6 years agoMerge pull request #18616 from dcirne/batch_size
Martin Wicke [Fri, 20 Apr 2018 00:01:46 +0000 (17:01 -0700)]
Merge pull request #18616 from dcirne/batch_size

QueueOptions var data types to size_t

6 years agoRun EvaluateNodes for ModelPruner test except for NoPruning.
A. Unique TensorFlower [Thu, 19 Apr 2018 23:53:14 +0000 (16:53 -0700)]
Run EvaluateNodes for ModelPruner test except for NoPruning.

PiperOrigin-RevId: 193596812

6 years agoMerge pull request #18605 from imsheridan/improve_deprecated_argument
Martin Wicke [Thu, 19 Apr 2018 23:51:49 +0000 (16:51 -0700)]
Merge pull request #18605 from imsheridan/improve_deprecated_argument

Improve deprecation assignment with deprecated_argument_lookup

6 years ago[tf.data] Correct / clarify docstring for `weights` as a dataset.
joel-shor [Thu, 19 Apr 2018 23:49:58 +0000 (02:49 +0300)]
[tf.data] Correct / clarify docstring for `weights` as a dataset.

This is a noop.

6 years agoImprove error messages for LiteralTestUtil::Near.
Mark Heffernan [Thu, 19 Apr 2018 23:35:40 +0000 (16:35 -0700)]
Improve error messages for LiteralTestUtil::Near.
Previously error messages for mismatches were difficult to read with much of the space taken by useless stack traces. This CL cleans up the message considerably and adds additional information including statistics about the values and mismatches.

PiperOrigin-RevId: 193594593

6 years agoCleaned up the handling of merge nodes
Benoit Steiner [Thu, 19 Apr 2018 23:30:26 +0000 (16:30 -0700)]
Cleaned up the handling of merge nodes

PiperOrigin-RevId: 193593810

6 years agoMake xla/service:cpu_plugin depend on the StreamExecutor host platform.
Justin Lebar [Thu, 19 Apr 2018 23:30:02 +0000 (16:30 -0700)]
Make xla/service:cpu_plugin depend on the StreamExecutor host platform.

PiperOrigin-RevId: 193593761

6 years agoAutomated g4 rollback of changelist 193564222
A. Unique TensorFlower [Thu, 19 Apr 2018 22:56:17 +0000 (15:56 -0700)]
Automated g4 rollback of changelist 193564222

PiperOrigin-RevId: 193588935

6 years agoInternal-only change.
Justin Lebar [Thu, 19 Apr 2018 22:55:53 +0000 (15:55 -0700)]
Internal-only change.

PiperOrigin-RevId: 193588868

6 years ago[tf.data] Allow `sample_from_datasets` to accept a tf.Dataset object for `weights`.
joel-shor [Thu, 19 Apr 2018 22:41:28 +0000 (01:41 +0300)]
[tf.data] Allow `sample_from_datasets` to accept a tf.Dataset object for `weights`.

Tested:
bazel test :interleave_dataset_op_test

6 years agoSimplifying cols_to_vars update
Rohan Jain [Thu, 19 Apr 2018 22:32:37 +0000 (15:32 -0700)]
Simplifying cols_to_vars update

PiperOrigin-RevId: 193585237

6 years agoPrevent a bool field from being accessed when uninitialized.
Derek Murray [Thu, 19 Apr 2018 22:29:21 +0000 (15:29 -0700)]
Prevent a bool field from being accessed when uninitialized.

PiperOrigin-RevId: 193584746

6 years agoConvert a local variable and mutex to a struct so GUARDED_BY annotation works correctly.
A. Unique TensorFlower [Thu, 19 Apr 2018 22:27:19 +0000 (15:27 -0700)]
Convert a local variable and mutex to a struct so GUARDED_BY annotation works correctly.

PiperOrigin-RevId: 193584438

6 years agoOnly generate floating points that are fractions like n / 256, since they are RGB...
A. Unique TensorFlower [Thu, 19 Apr 2018 22:09:58 +0000 (15:09 -0700)]
Only generate floating points that are fractions like n / 256, since they are RGB pixels. This fixes RGBToHSVTest.testBatch on low-precision dtypes like bfloat16.

PiperOrigin-RevId: 193581652

6 years agodisabling opensource testing for failing xla test
Olivia Nordquist [Thu, 19 Apr 2018 21:59:25 +0000 (14:59 -0700)]
disabling opensource testing for failing xla test

PiperOrigin-RevId: 193579805

6 years agoReplaced calls to deprecated tensorflow::StringPiece methods with their
A. Unique TensorFlower [Thu, 19 Apr 2018 21:35:27 +0000 (14:35 -0700)]
Replaced calls to deprecated tensorflow::StringPiece methods with their
tensorflow::str_util equivalents.

This will allow the deprecated methods to be removed.

PiperOrigin-RevId: 193575992

6 years agoThe HLO element type converter must remove side effecting instructions like Rng
Sanjoy Das [Thu, 19 Apr 2018 21:30:27 +0000 (14:30 -0700)]
The HLO element type converter must remove side effecting instructions like Rng

The CPU backend does not know how to lower bf16 typed RNG nodes so even unused
instances of these can't remain in the HLO IR.
HloComputation::ReplaceInstruction keeps these Rng nodes around since it doesn't
remove side effecting nodes.
PiperOrigin-RevId: 193575183

6 years agoRecordReader: Simplify interface contract and implementation.
Asim Shankar [Thu, 19 Apr 2018 21:10:01 +0000 (14:10 -0700)]
RecordReader: Simplify interface contract and implementation.

Prior to this change, RecordReader had the following contract:
- Records can be read in any order, EXCEPT if compression or buffering was
  enabled.
- If the underlying file is being concurrently written to
  then calls to ReadRecord() may fail (because of an incomplete
  record near the end of a file), but a retry may succeed (once the
  record is written), EXCEPT if compression or buffering is enabled
  (in which case the failure will be terminal).

  This "retry-may-succeed" behavior is relied upon by tensorboard
  (https://github.com/tensorflow/tensorboard/blob/1.7/tensorboard/backend/event_processing/event_file_loader.py#L55)
  where one process (typically the model training process) is writing
  tf.summary events to an event file and another process (tensorboard)
  is concurrently reading it.

With this change, the intent is to remove the EXCEPTions and have the
same behavior irrespective of compression/buffering.

Additionally, fix a memory leak when ZlibInputStream::Reset() is invoked.

PiperOrigin-RevId: 193571934

6 years agoMake flat_transforms_to_matrices and matrices_to_flat_transforms public available.
A. Unique TensorFlower [Thu, 19 Apr 2018 21:05:06 +0000 (14:05 -0700)]
Make flat_transforms_to_matrices and matrices_to_flat_transforms public available.

PiperOrigin-RevId: 193571089

6 years agotools/lib_package: Fix typo in README
Asim Shankar [Thu, 19 Apr 2018 20:38:43 +0000 (13:38 -0700)]
tools/lib_package: Fix typo in README

PiperOrigin-RevId: 193566850

6 years agoMerge pull request #18701 from tensorflow/yifeif-patch-3
Amit Patankar [Thu, 19 Apr 2018 20:38:07 +0000 (13:38 -0700)]
Merge pull request #18701 from tensorflow/yifeif-patch-3

Fix doc gen error

6 years agoAllow output has a different shape from input in the image.transform (#17011).
A. Unique TensorFlower [Thu, 19 Apr 2018 20:21:25 +0000 (13:21 -0700)]
Allow output has a different shape from input in the image.transform (#17011).

PiperOrigin-RevId: 193564222

6 years agoSupport various shapes in TPU DistributionStrategy.
Igor Saprykin [Thu, 19 Apr 2018 20:19:27 +0000 (13:19 -0700)]
Support various shapes in TPU DistributionStrategy.

PiperOrigin-RevId: 193563912

6 years agoCollective Ops Part 4
A. Unique TensorFlower [Thu, 19 Apr 2018 20:09:07 +0000 (13:09 -0700)]
Collective Ops Part 4

Add Broadcaster.
A few minor adjustments to CollectiveParams and RMA.

This change is part of a series of changes introducing infrastructure
for collective ops and initial implementations of reduction and broadcast.

PiperOrigin-RevId: 193562391

6 years agoUpdate adam.py
Mark Daoust [Thu, 19 Apr 2018 20:11:04 +0000 (13:11 -0700)]
Update adam.py

6 years agoUpdate api_def_ResourceApplyAdam.pbtxt
Mark Daoust [Thu, 19 Apr 2018 20:09:59 +0000 (13:09 -0700)]
Update api_def_ResourceApplyAdam.pbtxt

6 years agoUpdate api_def_ApplyAdam.pbtxt
Mark Daoust [Thu, 19 Apr 2018 20:09:24 +0000 (13:09 -0700)]
Update api_def_ApplyAdam.pbtxt

6 years agoUpdate adam.py
Mark Daoust [Thu, 19 Apr 2018 20:08:37 +0000 (13:08 -0700)]
Update adam.py

6 years agoMake PmfToQuantizedCdf op to make adjustments if the sum of quantized pmf is
Sung Jin Hwang [Thu, 19 Apr 2018 19:44:21 +0000 (12:44 -0700)]
Make PmfToQuantizedCdf op to make adjustments if the sum of quantized pmf is
less than 2**precision.

Prior to the change, the op did nothing when the sum of quantized pmf was less
than 2**precision. While the produced CDF was valid for range coders,
adjustments to CDF could be made to achieve better compression rate.

PiperOrigin-RevId: 193558740

6 years agoFix bug in ring_reducer.cc abort handling.
A. Unique TensorFlower [Thu, 19 Apr 2018 19:32:52 +0000 (12:32 -0700)]
Fix bug in ring_reducer.cc abort handling.

PiperOrigin-RevId: 193557334

6 years agoMerge pull request #18655 from langner/patch-1
Martin Wicke [Thu, 19 Apr 2018 19:29:14 +0000 (12:29 -0700)]
Merge pull request #18655 from langner/patch-1

Remove over-indentation