platform/upstream/tensorflow.git
7 years agoadded audio_ops.cc to fix the Op type not registered DecodeWav error - https://github...
Jeff Tang [Thu, 1 Feb 2018 22:31:04 +0000 (14:31 -0800)]
added audio_ops.cc to fix the Op type not registered DecodeWav error - https://github.com/tensorflow/tensorflow/issues/15921 (#16537)

7 years agoFix linter error in losses_impl.py (#16675)
Gunhan Gulsoy [Thu, 1 Feb 2018 22:23:35 +0000 (14:23 -0800)]
Fix linter error in losses_impl.py (#16675)

7 years agocompile libtensorflow on windows with AVX. (#16673)
Gunhan Gulsoy [Thu, 1 Feb 2018 18:48:44 +0000 (10:48 -0800)]
compile libtensorflow on windows with AVX. (#16673)

7 years agoFix an imperfect implementation of tf.losses.mean_pairwise_squared_error (#16433)
Jin-Hwan CHO [Thu, 1 Feb 2018 18:32:36 +0000 (03:32 +0900)]
Fix an imperfect implementation of tf.losses.mean_pairwise_squared_error (#16433)

* Imperfect implementation of tf.losses.mean_pairwise_squared_error (#15968)

https://github.com/tensorflow/tensorflow/issues/15968

* To pass the test, tensorflow/python/kernel_tests/losses_test.py needs to be fixed.

7 years agoAddress sanity build issues. (#16647)
Gunhan Gulsoy [Thu, 1 Feb 2018 18:20:31 +0000 (10:20 -0800)]
Address sanity build issues. (#16647)

* Address sanity build issues.

* More fixes.

* Final pylint fixes.

7 years agoFix do_cmake_python_sanity error. (#16650)
Yifei Feng [Thu, 1 Feb 2018 18:12:34 +0000 (10:12 -0800)]
Fix do_cmake_python_sanity error. (#16650)

* Fix do_cmake_python_sanity error.

* Update python_modules.txt

7 years agoClang on Windows will define __BYTE_ORDER__ etc. for us (#16492)
Loo Rong Jie [Thu, 1 Feb 2018 18:11:25 +0000 (02:11 +0800)]
Clang on Windows will define __BYTE_ORDER__ etc. for us (#16492)

7 years agoFix for mkl_input conversion for MKL DNN. Fix also enables elemenwise operations...
AG Ramesh [Thu, 1 Feb 2018 18:10:54 +0000 (11:10 -0700)]
Fix for mkl_input conversion for MKL DNN. Fix also enables elemenwise operations. (#16557)

7 years agoPooling and AddN fixes (#16607)
Mahmoud Abuzaina [Thu, 1 Feb 2018 08:37:47 +0000 (00:37 -0800)]
Pooling and AddN fixes (#16607)

7 years agoMerge pull request #16637 from case540/branch_184052073
Amit Patankar [Thu, 1 Feb 2018 06:41:17 +0000 (22:41 -0800)]
Merge pull request #16637 from case540/branch_184052073

Branch 184052073

7 years agoSupport multiple build types in Android build.gradle with the makefile build (#16640)
Andrew Harp [Thu, 1 Feb 2018 04:43:18 +0000 (23:43 -0500)]
Support multiple build types in Android build.gradle with the makefile build (#16640)

* updating CUDA srcs for Makefile build to fix unsatisfied link error

* more makefile refactoring

* Fixing Tegra build logic for Android

* reverting ndk dir

* set ccache back to default

7 years agoRe-add missing argument specifier (#16632)
Andrew Harp [Thu, 1 Feb 2018 04:42:56 +0000 (23:42 -0500)]
Re-add missing argument specifier (#16632)

The ":" was erroneously removed in 76f70f5d62f35b5cc95121e6dfffa63a8214b626

7 years agoBy default, only download inception if it doesn't exist already (#16577)
Ben [Thu, 1 Feb 2018 03:28:25 +0000 (22:28 -0500)]
By default, only download inception if it doesn't exist already (#16577)

* By default, only download inception if it doesn't exist already

* fix filename

7 years agoUpdate docs for tf.matching_files to mention non-deterministic (#16633)
Yong Tang [Thu, 1 Feb 2018 03:28:01 +0000 (19:28 -0800)]
Update docs for tf.matching_files to mention non-deterministic (#16633)

This fix tries to close the issue of 15374 by updating the
docs of `tf.matching_files` (as well as `Dataset.list_files`
and `train.match_filenames_once`), to mention that the file names
returned could be non-deterministic.

This fix fixes 15374.

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
7 years agoChange RELEASE.md to specify CUDA 9.0 (#16486)
Anjum Sayed [Thu, 1 Feb 2018 03:22:54 +0000 (06:22 +0300)]
Change RELEASE.md to specify CUDA 9.0 (#16486)

PR for https://github.com/tensorflow/tensorflow/issues/16348 (tinyest PR ever?)

7 years agoAdd note about ptxas bug affecting XLA:GPU to relnotes. (#16636)
Justin Lebar [Thu, 1 Feb 2018 03:15:44 +0000 (19:15 -0800)]
Add note about ptxas bug affecting XLA:GPU to relnotes. (#16636)

7 years agoFix the FutureWarning: Conversion of the second argument of issubdtype from `float...
ImSheridan [Thu, 1 Feb 2018 03:03:18 +0000 (11:03 +0800)]
Fix the FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated (#16591)

7 years agoFix typos. (#16570)
hsm207 [Thu, 1 Feb 2018 03:00:51 +0000 (22:00 -0500)]
Fix typos. (#16570)

7 years agoRemove query_layer in LuongMonotonicAttention (#16602)
Colin Raffel [Thu, 1 Feb 2018 03:00:33 +0000 (19:00 -0800)]
Remove query_layer in LuongMonotonicAttention (#16602)

In the constructor for LuongMonotonicAttention, a query layer was being created but it was ultimately never used.

Fixes #16287.

7 years agoReplace 'Dan' with 'Dandelion' in the citations (#16630)
Dandelion Mané [Thu, 1 Feb 2018 02:59:31 +0000 (18:59 -0800)]
Replace 'Dan' with 'Dandelion' in the citations (#16630)

7 years agoRemove duplicated identical lines (#16573)
simsicon [Thu, 1 Feb 2018 02:59:10 +0000 (10:59 +0800)]
Remove duplicated identical lines (#16573)

7 years agoFixed typo (#16610)
Cole Gerdemann [Thu, 1 Feb 2018 02:58:56 +0000 (20:58 -0600)]
Fixed typo (#16610)

7 years agoUpdate ISSUE_TEMPLATE.md (#16635)
Mark Daoust [Thu, 1 Feb 2018 02:58:24 +0000 (18:58 -0800)]
Update ISSUE_TEMPLATE.md (#16635)

Fixes 16350

7 years agoMerge commit for internal changes
Michael Case [Thu, 1 Feb 2018 02:10:39 +0000 (18:10 -0800)]
Merge commit for internal changes

7 years agoAdd grpcio as a pip dependency of tensorflow
Shanqing Cai [Wed, 31 Jan 2018 22:55:13 +0000 (14:55 -0800)]
Add grpcio as a pip dependency of tensorflow

PiperOrigin-RevId: 184052073

7 years agoFixes minor typos.
Shivani Agrawal [Wed, 31 Jan 2018 22:35:37 +0000 (14:35 -0800)]
Fixes minor typos.

PiperOrigin-RevId: 184048812

7 years ago[XLA] Initialize linear indices used by NearComparator.
Nick Desaulniers [Wed, 31 Jan 2018 22:32:21 +0000 (14:32 -0800)]
[XLA] Initialize linear indices used by NearComparator.

In unoptimized builds, we'd see miscompares with insanely large indices when
comparing literals containing NaN.

The linear indices may never be updated/initialized if the value NaN was
compared, since:

(NaN > x) == false
(NaN < x) == false

Adds a few tests that were used when debugging, but not a perfect one for this
case. It's currently not possible to test that LiteralTestUtil::Near() fails
when given bad input since it uses the EXPECT_* family of macros, which would
cause the intentional miscompares to fail the test no matter what.

PiperOrigin-RevId: 184048275

7 years agoInternal change.
Anna R [Wed, 31 Jan 2018 22:30:22 +0000 (14:30 -0800)]
Internal change.

PiperOrigin-RevId: 184047860

7 years agoTolerate Const nodes with no data or with smaller data than is required
A. Unique TensorFlower [Wed, 31 Jan 2018 21:42:18 +0000 (13:42 -0800)]
Tolerate Const nodes with no data or with smaller data than is required
by their shape, by zero-extending the Const data to the required size.

We wanted to generate an error on that, but too many
existing graphs already rely on current lax behavior.

PiperOrigin-RevId: 184039876

7 years agoAutomated g4 rollback of changelist 184003263
A. Unique TensorFlower [Wed, 31 Jan 2018 21:35:23 +0000 (13:35 -0800)]
Automated g4 rollback of changelist 184003263

PiperOrigin-RevId: 184038801

7 years ago[XLA:CPU] Generate correct IR for integer clamp
Sanjoy Das [Wed, 31 Jan 2018 21:22:49 +0000 (13:22 -0800)]
[XLA:CPU] Generate correct IR for integer clamp

PiperOrigin-RevId: 184037078

7 years ago[TF:XLA] Implement ExtractImagePatches.
Peter Hawkins [Wed, 31 Jan 2018 20:58:05 +0000 (12:58 -0800)]
[TF:XLA] Implement ExtractImagePatches.

PiperOrigin-RevId: 184033616

7 years agoRemove contacts of ex-Googler
A. Unique TensorFlower [Wed, 31 Jan 2018 20:32:38 +0000 (12:32 -0800)]
Remove contacts of ex-Googler

PiperOrigin-RevId: 184030353

7 years ago[TF:XLA] Implement Acos, Asin, Atan in terms of Atan2 using half-angle formulae....
Peter Hawkins [Wed, 31 Jan 2018 20:28:13 +0000 (12:28 -0800)]
[TF:XLA] Implement Acos, Asin, Atan in terms of Atan2 using half-angle formulae. This may not be the most efficient implementation but it is better than no implementation.

PiperOrigin-RevId: 184029858

7 years ago[TF:XLA] Bump open source llvm revision to r323874
Sanjoy Das [Wed, 31 Jan 2018 20:27:38 +0000 (12:27 -0800)]
[TF:XLA] Bump open source llvm revision to r323874

PiperOrigin-RevId: 184029790

7 years agoIncrefing tapes when keeping them around calls to python API functions.
Alexandre Passos [Wed, 31 Jan 2018 20:22:10 +0000 (12:22 -0800)]
Increfing tapes when keeping them around calls to python API functions.

These calls can trigger GIL releases or python GC which can trigger
a change to the set of live tapes and segmentation faults.

PiperOrigin-RevId: 184029146

7 years agoDe-bazel filename_test. Part of the effort to remove all_opensource_files.
Yifei Feng [Wed, 31 Jan 2018 19:34:56 +0000 (11:34 -0800)]
De-bazel filename_test. Part of the effort to remove all_opensource_files.

PiperOrigin-RevId: 184021942

7 years agoGo: Fix Scope.WithControlDependencies array-copying behavior.
Todd Wang [Wed, 31 Jan 2018 19:32:35 +0000 (11:32 -0800)]
Go: Fix Scope.WithControlDependencies array-copying behavior.

The test fails with the old code, and passes with the new code.

PiperOrigin-RevId: 184021596

7 years agoTPUEstimator: support host_call when use_tpu=False.
Jonathan Hseu [Wed, 31 Jan 2018 19:30:36 +0000 (11:30 -0800)]
TPUEstimator: support host_call when use_tpu=False.

PiperOrigin-RevId: 184021299

7 years agoAdding tf_export decorators/calls to TensorFlow functions and constants.
Anna R [Wed, 31 Jan 2018 19:25:38 +0000 (11:25 -0800)]
Adding tf_export decorators/calls to TensorFlow functions and constants.

PiperOrigin-RevId: 184020524

7 years agoTypo Correction.
A. Unique TensorFlower [Wed, 31 Jan 2018 18:59:49 +0000 (10:59 -0800)]
Typo Correction.

PiperOrigin-RevId: 184016082

7 years agoAdd std:: to min/max in cuda code to unbreak cuda-clang build.
Eric Liu [Wed, 31 Jan 2018 18:38:57 +0000 (10:38 -0800)]
Add std:: to min/max in cuda code to unbreak cuda-clang build.

PiperOrigin-RevId: 184012479

7 years agoWhitelisting stateful op for dataset checkpointing temporarily.
Shivani Agrawal [Wed, 31 Jan 2018 18:29:29 +0000 (10:29 -0800)]
Whitelisting stateful op for dataset checkpointing temporarily.

PiperOrigin-RevId: 184010966

7 years ago[XLA] Disable transpose folding into reduce for reduces of rank 2 or higher.
Blake Hechtman [Wed, 31 Jan 2018 18:19:17 +0000 (10:19 -0800)]
[XLA] Disable transpose folding into reduce for reduces of rank 2 or higher.

PiperOrigin-RevId: 184009219

7 years agoUpdate external protobuf codebase version
A. Unique TensorFlower [Wed, 31 Jan 2018 18:06:24 +0000 (10:06 -0800)]
Update external protobuf codebase version

PiperOrigin-RevId: 184006959

7 years agoExtending sparsify_gather to remove variables from the tensorflow summaries.
A. Unique TensorFlower [Wed, 31 Jan 2018 17:53:17 +0000 (09:53 -0800)]
Extending sparsify_gather to remove variables from the tensorflow summaries.

PiperOrigin-RevId: 184004859

7 years agoVerify contents of tensors
A. Unique TensorFlower [Wed, 31 Jan 2018 17:40:48 +0000 (09:40 -0800)]
Verify contents of tensors

PiperOrigin-RevId: 184003263

7 years agoFix docs generation for cluster_resolvers
Mark Daoust [Wed, 31 Jan 2018 16:16:59 +0000 (08:16 -0800)]
Fix docs generation for cluster_resolvers

Adds "cluster_resolver_pip" as a dependancy to opensource contrib, and applies a standard `remove_undocumented` to clear extra symbols.

Docs are build from a bazel bulid, and without this change the cluster resolvers are not directly accessible in "tf.contirb.cluster_resolver" during the docs build, so they do not get documented.

PiperOrigin-RevId: 183993115

7 years agoRemove a failing test for gpu numbers when XLA is enabled.
Eric Liu [Wed, 31 Jan 2018 09:26:29 +0000 (01:26 -0800)]
Remove a failing test for gpu numbers when XLA is enabled.

The test assumes that all devices are either CPU or GPU, which is not true when XLA is enabled.

PiperOrigin-RevId: 183956872

7 years agoGo: Support control dependencies.
Asim Shankar [Wed, 31 Jan 2018 07:15:28 +0000 (23:15 -0800)]
Go: Support control dependencies.

Fixes #16464

PiperOrigin-RevId: 183946928

7 years agoAdds input_fn-return-dataset for Per-Host input pipeline deployment.
Jianwei Xie [Wed, 31 Jan 2018 05:34:02 +0000 (21:34 -0800)]
Adds input_fn-return-dataset for Per-Host input pipeline deployment.

PiperOrigin-RevId: 183940666

7 years agoAdding tf_export decorators/calls to TensorFlow functions and constants.
Anna R [Wed, 31 Jan 2018 04:28:38 +0000 (20:28 -0800)]
Adding tf_export decorators/calls to TensorFlow functions and constants.

PiperOrigin-RevId: 183936100

7 years agoRemove Google-internal bug numbers from XLA error messages.
Bjarke Hammersholt Roune [Wed, 31 Jan 2018 04:12:22 +0000 (20:12 -0800)]
Remove Google-internal bug numbers from XLA error messages.

PiperOrigin-RevId: 183934860

7 years agointernal change
A. Unique TensorFlower [Wed, 31 Jan 2018 03:30:18 +0000 (19:30 -0800)]
internal change

PiperOrigin-RevId: 183932015

7 years agoEager support for tf.Keras.
Anjali Sridhar [Wed, 31 Jan 2018 02:50:56 +0000 (18:50 -0800)]
Eager support for tf.Keras.

PiperOrigin-RevId: 183928850

7 years agointernal change
A. Unique TensorFlower [Wed, 31 Jan 2018 02:08:52 +0000 (18:08 -0800)]
internal change

PiperOrigin-RevId: 183924520

7 years agoRemoved unused variable doReverseChannels.
A. Unique TensorFlower [Wed, 31 Jan 2018 02:03:31 +0000 (18:03 -0800)]
Removed unused variable doReverseChannels.

PiperOrigin-RevId: 183923876

7 years agoK-FAC: Wrap extract_image_patches() for compatibility with XLA.
A. Unique TensorFlower [Wed, 31 Jan 2018 01:58:28 +0000 (17:58 -0800)]
K-FAC: Wrap extract_image_patches() for compatibility with XLA.

PiperOrigin-RevId: 183923073

7 years agoAdd a CLIF wrapper for DeviceProperties.
Max Galkin [Wed, 31 Jan 2018 01:41:00 +0000 (17:41 -0800)]
Add a CLIF wrapper for DeviceProperties.

PiperOrigin-RevId: 183921389

7 years agoGive tf.contrib.summary a default suffix of .v2. This is so that it doesn't conflict...
Jonathan Hseu [Wed, 31 Jan 2018 01:28:42 +0000 (17:28 -0800)]
Give tf.contrib.summary a default suffix of .v2. This is so that it doesn't conflict with v1 summaries that are being created at the same time.

By default, MonitoredTrainingSession and Estimator will create v1 summaries, so using both currently creates a race condition, and the last one to create the event file writer wins.

PiperOrigin-RevId: 183919988

7 years agoUpdate graph_matcher to include OneofPattern.
Suharsh Sivakumar [Wed, 31 Jan 2018 01:03:49 +0000 (17:03 -0800)]
Update graph_matcher to include OneofPattern.

PiperOrigin-RevId: 183916547

7 years agooptimize_for_inference_lib.fold_batch_norms() preserves data_format (#16075)
yegord [Wed, 31 Jan 2018 23:02:25 +0000 (00:02 +0100)]
optimize_for_inference_lib.fold_batch_norms() preserves data_format (#16075)

Fixes https://github.com/tensorflow/tensorflow/issues/15034

7 years agoMerge pull request #16612 from joel-shor/master
Yifei Feng [Wed, 31 Jan 2018 21:08:13 +0000 (13:08 -0800)]
Merge pull request #16612 from joel-shor/master

Simplify loader_impl.py logic around main Op Tensor.

7 years agoMerge pull request #16446 from freedomtan/label_image_tflite_bilinear_pr
Rasmus Munk Larsen [Wed, 31 Jan 2018 19:22:18 +0000 (11:22 -0800)]
Merge pull request #16446 from freedomtan/label_image_tflite_bilinear_pr

use tflite bilinear op to resize input of label_image

7 years agoFixed iOS build script for all architectures, fixes #12904 (#16559)
Pete Warden [Wed, 31 Jan 2018 19:11:58 +0000 (11:11 -0800)]
Fixed iOS build script for all architectures, fixes #12904 (#16559)

7 years agoadd brackets and new line
Koan-Sin Tan [Wed, 31 Jan 2018 05:22:54 +0000 (13:22 +0800)]
add brackets and new line

Add curly brackets and new line to address the format problem
mentioned in review.

7 years agoSimplify loader_impl.py logic around main Op Tensor.
joel-shor [Wed, 31 Jan 2018 03:06:48 +0000 (19:06 -0800)]
Simplify loader_impl.py logic around main Op Tensor.

7 years ago[tf.data] Native support for `tf.SparseTensor` and serialization tests for the `tf...
Shivani Agrawal [Wed, 31 Jan 2018 00:14:17 +0000 (16:14 -0800)]
[tf.data] Native support for `tf.SparseTensor` and serialization tests for the `tf.data` methods that can produce `tf.SparseTensor` elements.

PiperOrigin-RevId: 183909820

7 years agoFix _standardize_input_data function to allow users to use list of
Anjali Sridhar [Wed, 31 Jan 2018 00:12:32 +0000 (16:12 -0800)]
Fix _standardize_input_data function to allow users to use list of
ints/floats and list of lists as training and validation data input.

PiperOrigin-RevId: 183909545

7 years ago[XLA] Add Clamp and Round to the local Python XLA client.
A. Unique TensorFlower [Tue, 30 Jan 2018 23:53:55 +0000 (15:53 -0800)]
[XLA] Add Clamp and Round to the local Python XLA client.

PiperOrigin-RevId: 183906722

7 years agoInternal change.
Anna R [Tue, 30 Jan 2018 23:36:08 +0000 (15:36 -0800)]
Internal change.

PiperOrigin-RevId: 183904042

7 years agoMerge pull request #16604 from rmlarsen/branch_183881907
Rasmus Munk Larsen [Tue, 30 Jan 2018 23:33:13 +0000 (15:33 -0800)]
Merge pull request #16604 from rmlarsen/branch_183881907

Branch 183881907

7 years agoInternal change.
Nupur Garg [Tue, 30 Jan 2018 23:11:33 +0000 (15:11 -0800)]
Internal change.

PiperOrigin-RevId: 183900332

7 years agoSupporting paths for build_server script.
Amit Patankar [Tue, 30 Jan 2018 23:07:03 +0000 (15:07 -0800)]
Supporting paths for build_server script.

PiperOrigin-RevId: 183899675

7 years agoAlways evaluate() on TPU in TPUEstimator unless use_tpu is set to False.
A. Unique TensorFlower [Tue, 30 Jan 2018 23:05:49 +0000 (15:05 -0800)]
Always evaluate() on TPU in TPUEstimator unless use_tpu is set to False.

PiperOrigin-RevId: 183899481

7 years agoOptimize memory allocation of TFLite StridedSlice Op for constant tensors.
A. Unique TensorFlower [Tue, 30 Jan 2018 23:04:14 +0000 (15:04 -0800)]
Optimize memory allocation of TFLite StridedSlice Op for constant tensors.

PiperOrigin-RevId: 183899156

7 years agoAssign*VariableOp: Documentation fix.
Asim Shankar [Tue, 30 Jan 2018 23:00:10 +0000 (15:00 -0800)]
Assign*VariableOp: Documentation fix.

These ops have no outputs.
The way to ensure ordering is by adding dependencies via control inputs.

Helps with #16464

PiperOrigin-RevId: 183898409

7 years agoExpose the decorator in the main API. Move the top-level implementation files into...
A. Unique TensorFlower [Tue, 30 Jan 2018 22:47:11 +0000 (14:47 -0800)]
Expose the decorator in the main API. Move the top-level implementation files into a submodule, to avoid problems around the remove_undocumented call.

PiperOrigin-RevId: 183896273

7 years ago[XLA] Add a test for scalar kRoundNearestAfz
David Majnemer [Tue, 30 Jan 2018 22:28:06 +0000 (14:28 -0800)]
[XLA] Add a test for scalar kRoundNearestAfz

PiperOrigin-RevId: 183892483

7 years agoAdd `tf.contrib.distributions.Kumaraswamy`.
A. Unique TensorFlower [Tue, 30 Jan 2018 22:07:26 +0000 (14:07 -0800)]
Add `tf.contrib.distributions.Kumaraswamy`.

PiperOrigin-RevId: 183888726

7 years agoRemove unnecessary dependency from xla:types
Todd Wang [Tue, 30 Jan 2018 22:06:07 +0000 (14:06 -0800)]
Remove unnecessary dependency from xla:types

PiperOrigin-RevId: 183888495

7 years agoMerge commit for internal changes
Rasmus Larsen [Tue, 30 Jan 2018 22:11:58 +0000 (14:11 -0800)]
Merge commit for internal changes

7 years agoAdd and Mul support broadcasting.
A. Unique TensorFlower [Tue, 30 Jan 2018 21:58:05 +0000 (13:58 -0800)]
Add and Mul support broadcasting.

PiperOrigin-RevId: 183886920

7 years ago[tf.data] Add `IteratorContext::allocator()`.
Derek Murray [Tue, 30 Jan 2018 21:26:51 +0000 (13:26 -0800)]
[tf.data] Add `IteratorContext::allocator()`.

This enables the various iterator implementations to use the actual allocator for the device on which they are running, rather than defaulting to `cpu_allocator()` (which is typically a plain malloc). In future, this will enable allocating iterator outputs in CUDA-pinned memory (and GPU memory).

PERFORMANCE NOTE: In sessions where `ConfigProto.force_gpu_compatible == True`, this change has the effect of allocating all input pipeline tensors in CUDA-pinned memory. Previous if this flag was set, only the tensors allocated during function execution would be allocated in this space, and other tensors (e.g. the result of a `Dataset.batch()` would be allocated using `cpu_allocator()` (i.e. `malloc()`). This change should lead to more efficient communication between a host-side input pipeline and GPUs, but it may also create more pressure on the CUDA host allocator (whose default maximum size is 64GB). The "TF_CUDA_HOST_MEM_LIMIT_IN_MB" environment variable can be used to override this value.

This change is a starting point for working on issue #13610.

PiperOrigin-RevId: 183881907

7 years agoEager: Update documentation to reflect that you can use TensorFlow 1.5
Asim Shankar [Tue, 30 Jan 2018 21:20:13 +0000 (13:20 -0800)]
Eager: Update documentation to reflect that you can use TensorFlow 1.5

PiperOrigin-RevId: 183880991

7 years agoAdd missing dependency.
A. Unique TensorFlower [Tue, 30 Jan 2018 21:10:54 +0000 (13:10 -0800)]
Add missing dependency.

PiperOrigin-RevId: 183879566

7 years agoEnable bulk restoration by default.
Russell Power [Tue, 30 Jan 2018 21:02:25 +0000 (13:02 -0800)]
Enable bulk restoration by default.

This enables loading multiple tensors in single call, allowing for better buffering and reduced load on distributed filesystems.

PiperOrigin-RevId: 183878169

7 years agoSupport outfeed host calls in TPUEstimator. That implicitly allows us to support...
Jonathan Hseu [Tue, 30 Jan 2018 20:49:18 +0000 (12:49 -0800)]
Support outfeed host calls in TPUEstimator. That implicitly allows us to support tf.contrib.summary (see the unit test).

We may change the recommended way to use summaries later.

PiperOrigin-RevId: 183876356

7 years agoAutomated g4 rollback of changelist 183846994
A. Unique TensorFlower [Tue, 30 Jan 2018 20:36:51 +0000 (12:36 -0800)]
Automated g4 rollback of changelist 183846994

PiperOrigin-RevId: 183874527

7 years agoFix bad logging call in warmstarting_utils.py.
Martin Wicke [Tue, 30 Jan 2018 20:32:30 +0000 (12:32 -0800)]
Fix bad logging call in warmstarting_utils.py.

PiperOrigin-RevId: 183873925

7 years agoAdd check macro for not-equal.
A. Unique TensorFlower [Tue, 30 Jan 2018 20:13:58 +0000 (12:13 -0800)]
Add check macro for not-equal.

PiperOrigin-RevId: 183871336

7 years agoEliminate crash on a 'no error' return from DecodeGif when parsing an invalid
David G. Andersen [Tue, 30 Jan 2018 20:06:40 +0000 (12:06 -0800)]
Eliminate crash on a 'no error' return from DecodeGif when parsing an invalid
gif.  (Previous code tried to strcat a null).

PiperOrigin-RevId: 183870288

7 years agoRetryingFileSystem::FlushCaches() calls the base FileSystem's FlushCaches().
A. Unique TensorFlower [Tue, 30 Jan 2018 20:01:02 +0000 (12:01 -0800)]
RetryingFileSystem::FlushCaches() calls the base FileSystem's FlushCaches().

PiperOrigin-RevId: 183869325

7 years agoAdd static_sample flag to Mixture, permitting calls to `sample` to not rely on dynami...
A. Unique TensorFlower [Tue, 30 Jan 2018 20:00:04 +0000 (12:00 -0800)]
Add static_sample flag to Mixture, permitting calls to `sample` to not rely on dynamic tensor indexing. This allows for some static graph compilation optimizations, but at the expense of sampling all underlying distributions in the mixture.

PiperOrigin-RevId: 183869189

7 years ago[TF:XLA] Bump open source llvm revision to r323761
Sanjoy Das [Tue, 30 Jan 2018 19:53:26 +0000 (11:53 -0800)]
[TF:XLA] Bump open source llvm revision to r323761

PiperOrigin-RevId: 183868087

7 years agoK-FAC: expose set_global_constants() for tf.contrib.kfac.utils
A. Unique TensorFlower [Tue, 30 Jan 2018 19:47:43 +0000 (11:47 -0800)]
K-FAC: expose set_global_constants() for tf.contrib.kfac.utils
PiperOrigin-RevId: 183867014

7 years agoDelete dead code in Layer.
Akshay Agrawal [Tue, 30 Jan 2018 19:42:31 +0000 (11:42 -0800)]
Delete dead code in Layer.

PiperOrigin-RevId: 183866106

7 years ago[XLA] Fix tools broken by cl/183837856
Chris Leary [Tue, 30 Jan 2018 19:22:12 +0000 (11:22 -0800)]
[XLA] Fix tools broken by cl/183837856

PiperOrigin-RevId: 183862522

7 years agoMake loss_ops_test.py work with C API enabled.
Skye Wanderman-Milne [Tue, 30 Jan 2018 19:18:03 +0000 (11:18 -0800)]
Make loss_ops_test.py work with C API enabled.

PiperOrigin-RevId: 183861779

7 years agoAdd py2tf to contrib_py.
A. Unique TensorFlower [Tue, 30 Jan 2018 19:09:06 +0000 (11:09 -0800)]
Add py2tf to contrib_py.

PiperOrigin-RevId: 183860192