Michael Case [Mon, 25 Jun 2018 17:27:33 +0000 (10:27 -0700)]
Merge pull request #20203 from alextp/cherrypicks_GUVCA
Add a single positional argument mode for shape inference in subclass…
Michael Case [Fri, 22 Jun 2018 00:01:44 +0000 (17:01 -0700)]
Merge pull request #20204 from mrry/choose_pr
1.9.0rc2 cherry-pick request: [tf.data] Properly export `choose_from_datasets()`
Derek Murray [Wed, 20 Jun 2018 17:10:55 +0000 (10:10 -0700)]
[tf.data] Properly export `tf.contrib.data.choose_from_datasets()`
PiperOrigin-RevId:
201371642
Allen Lavoie [Fri, 1 Jun 2018 02:03:21 +0000 (19:03 -0700)]
Add a single positional argument mode for shape inference in subclassed Models.
Allows fit() when call's signature looks something like call(x, training=True).
Calling conventions are "inputs", single positional, and multiple positional. Right now the distinction between "inputs" and single positional calling conventions is the text of one error message. Both support shape inference (which just hasn't been implemented for multiple positional input arguments yet).
PiperOrigin-RevId:
198815483
Michael Case [Thu, 21 Jun 2018 21:07:02 +0000 (14:07 -0700)]
Merge pull request #20195 from case540/cherry_pick_pip_external_fix
Move external/ directory in pip package.
Michael Case [Wed, 20 Jun 2018 21:08:57 +0000 (14:08 -0700)]
Move external/ directory in pip package.
Moving external/ directory in the pip packages (which is currently
installed directly into site-packages directory). Moving the
directory to tensorflow/include/external/. Also, removing all
python files from external (since it should really only contain
headers and license files.)
Michael Case [Wed, 20 Jun 2018 04:56:11 +0000 (21:56 -0700)]
Merge pull request #20130 from saeta/tfgcs3
[GCS] Typo in ConfigureGcsHook.
Brennan Saeta [Wed, 20 Jun 2018 00:47:37 +0000 (17:47 -0700)]
[GCS] Typo in ConfigureGcsHook.
This commit fixes a typo on ConfigureGcsHook that prevented its correct
operation.
Michael Case [Mon, 18 Jun 2018 18:07:48 +0000 (11:07 -0700)]
Add missing Eager relnotes for TensorFlow 1.9. (#20101)
Michael Case [Mon, 18 Jun 2018 17:13:39 +0000 (10:13 -0700)]
Merge pull request #20097 from MarkDaoust/fix_keras_guide
r1.9 cherry-pick request: Fix keras guide (docs only)
Billy Lamberta [Fri, 8 Jun 2018 22:47:19 +0000 (15:47 -0700)]
Copy edits to Keras guide, formatting, moving some things around.
Make the right TOC nav more useful.
PiperOrigin-RevId:
199863216
Michael Case [Thu, 14 Jun 2018 13:24:27 +0000 (06:24 -0700)]
Merge pull request #20020 from case540/version_string_update_19rc1
Update version strings for 1.9.0-rc1.
Michael Case [Thu, 14 Jun 2018 13:17:00 +0000 (06:17 -0700)]
Update version strings for 1.9.0-rc1.
Michael Case [Thu, 14 Jun 2018 13:11:42 +0000 (06:11 -0700)]
Merge pull request #19976 from guillaumekln/cherry-pick-best-exporter-fix
BestExporter cherry-pick request for r1.9: Only calls compare function if values were read from event file
Michael Case [Thu, 14 Jun 2018 13:07:19 +0000 (06:07 -0700)]
Merge pull request #20002 from saeta/tf19gcs2
Make GCS ops work in open source
Shanqing Cai [Thu, 14 Jun 2018 02:43:24 +0000 (22:43 -0400)]
Merge branch 'r1.9' into cherry-pick-best-exporter-fix
Brennan Saeta [Thu, 14 Jun 2018 01:05:39 +0000 (18:05 -0700)]
Update the docs and api_def.
Brennan Saeta [Wed, 13 Jun 2018 22:46:12 +0000 (15:46 -0700)]
Make GCS ops work in open source
Michael Case [Wed, 13 Jun 2018 22:11:07 +0000 (15:11 -0700)]
Merge pull request #19999 from akshaym/cherrypicks_34G4G
r1.9-rc1 cherry-pick request: bugfix/memory leak fix
Michael Case [Wed, 13 Jun 2018 20:18:44 +0000 (13:18 -0700)]
Merge pull request #19997 from MarkDaoust/keras_save_load_fix
r1.9-rc1 cherry-pick request: Keras save_weights fix
Akshay Modi [Mon, 11 Jun 2018 23:27:12 +0000 (16:27 -0700)]
Remove memory leak in read variable call, and record gradient call.
Fix #19385
PiperOrigin-RevId:
200132949
Akshay Modi [Mon, 11 Jun 2018 17:42:15 +0000 (10:42 -0700)]
Don't call back into python during insert (which will leave the set in a broken condition if the runtime decides to let another thread run).
Thank you for finding the bug. The watched_variables_ set should not really require a lock since all our functions hold the GIL (verified by looking at the generated SWIG). The reason that there was a concurrent access to the set is that the insert was calling back into python (which might release the GIL and let another thread run, which will also attempt to insert a variable and break the set).
I included the lock to be safe though, since its non-trivial to verify without looking at the generated swig wrappers that the GIL is held.
PiperOrigin-RevId:
200074843
Allen Lavoie [Mon, 11 Jun 2018 18:55:34 +0000 (11:55 -0700)]
Use the Keras session for saving/loading in TensorFlow format
Fixes issues when there's no default session
PiperOrigin-RevId:
200088574
Michael Case [Wed, 13 Jun 2018 18:28:26 +0000 (11:28 -0700)]
Merge pull request #19994 from nfelt/tensorboard-1.9
Update tensorboard dependency to 1.9.x
Michael Case [Wed, 13 Jun 2018 17:58:15 +0000 (10:58 -0700)]
Merge pull request #19963 from saeta/r1.9clusterresolver
Check to ensure the Cloud TPU is ready before resolving.
Michael Case [Wed, 13 Jun 2018 17:56:42 +0000 (10:56 -0700)]
Merge pull request #19960 from annarev/cherrypicks_NHLZ4
r1.9-rc1 cherry-pick request: Fix module docstrings
Michael Case [Wed, 13 Jun 2018 17:55:29 +0000 (10:55 -0700)]
Merge pull request #19867 from MarkDaoust/cherrypicks_613QM
1.9-rc0 cherry-pick request: Keras symbol visibility, and `Dataset` evaluation fixes.
Michael Case [Wed, 13 Jun 2018 17:53:25 +0000 (10:53 -0700)]
Merge pull request #19977 from petewarden/r1.9
r1.9-rc2 cherry-pick request: Documentation for Raspberry Pi installation
Pete Warden [Wed, 13 Jun 2018 08:06:50 +0000 (01:06 -0700)]
Documentation for Raspberry Pi installation
Nick Felt [Wed, 13 Jun 2018 07:30:09 +0000 (00:30 -0700)]
Update tensorboard dependency to 1.9.x
Brennan Saeta [Wed, 13 Jun 2018 01:09:35 +0000 (18:09 -0700)]
Check to ensure the Cloud TPU is ready before resolving.
Cherry picking this into the TF 1.9 release.
PiperOrigin-RevId:
200095692
Previous commit:
32c8013f0ab3feb139648ae759e2d0168fb5dc95
Anna R [Wed, 13 Jun 2018 00:47:58 +0000 (17:47 -0700)]
Remove unused tf_export import
Anna R [Tue, 12 Jun 2018 00:21:06 +0000 (17:21 -0700)]
Merging
Michael Case [Tue, 12 Jun 2018 23:35:00 +0000 (16:35 -0700)]
Merge branch 'r1.9' into cherrypicks_613QM
Michael Case [Tue, 12 Jun 2018 19:42:45 +0000 (12:42 -0700)]
Merge pull request #19925 from shizhiw/cherrypicks_I1PZP
release 1.9-rc1 cherry-pick request: opt out of saving tpu graph
A. Unique TensorFlower [Fri, 1 Jun 2018 23:32:20 +0000 (16:32 -0700)]
Allow user to opt out of saving metagraph for TPU with TPUEstimator.export_output().
PiperOrigin-RevId:
198944144
A. Unique TensorFlower [Tue, 29 May 2018 21:28:59 +0000 (14:28 -0700)]
In TPUEstimator.export_savedmodel(), if saving TPU metegraph fails, issue a warning instead so that user can still use the CPU metagraph.
PiperOrigin-RevId:
198458571
A. Unique TensorFlower [Sun, 27 May 2018 17:49:12 +0000 (10:49 -0700)]
TPUEstimator.export_savedmodel() saves a SavedModel with both TPU and CPU graphs.
PiperOrigin-RevId:
198229550
A. Unique TensorFlower [Tue, 5 Jun 2018 22:59:21 +0000 (15:59 -0700)]
Only calls compare function if values were read from event file
PiperOrigin-RevId:
199373169
Pavithra Vijay [Thu, 7 Jun 2018 16:20:57 +0000 (09:20 -0700)]
Fix bug due to incorrect nesting of return statement in eager iterator evaluation.
PiperOrigin-RevId:
199645638
Francois Chollet [Mon, 4 Jun 2018 21:46:38 +0000 (14:46 -0700)]
Add various missing aliases for symbols in tf.keras submodules.
PiperOrigin-RevId:
199198086
Mark Daoust [Mon, 4 Jun 2018 18:11:06 +0000 (11:11 -0700)]
Fix visibility for tf.keras.__version__
PiperOrigin-RevId:
199161696
Amit Patankar [Fri, 8 Jun 2018 19:52:17 +0000 (12:52 -0700)]
Merge pull request #19866 from caisq/r1.9-patch
Update RELEASE.md (r1.9) for tfdbg and XLA
Shanqing Cai [Fri, 8 Jun 2018 18:55:26 +0000 (14:55 -0400)]
Update RELEASE.md (r1.9) for tfdbg and XLA
Amit Patankar [Thu, 7 Jun 2018 19:55:34 +0000 (12:55 -0700)]
Merge pull request #19841 from tensorflow/mrry-patch-1
Consolidate `tf.data` release notes.
Derek Murray [Thu, 7 Jun 2018 19:53:11 +0000 (12:53 -0700)]
Consolidate `tf.data` release notes.
Amit Patankar [Wed, 6 Jun 2018 22:07:00 +0000 (15:07 -0700)]
Merge pull request #19820 from av8ramit/fix_setuptools4
Adding a constraint for the setuptools version.
Amit Patankar [Wed, 6 Jun 2018 21:27:59 +0000 (14:27 -0700)]
Removing the force downgrade install.
Amit Patankar [Wed, 6 Jun 2018 21:04:40 +0000 (14:04 -0700)]
Adding a constraint for the setuptools version.
Yifei Feng [Wed, 6 Jun 2018 20:42:49 +0000 (13:42 -0700)]
Merge pull request #19819 from av8ramit/fix_setuptools3
Force downgrade setuptools for tests after tf whl is installed.
Amit Patankar [Wed, 6 Jun 2018 20:35:35 +0000 (13:35 -0700)]
Force downgrade setuptools for tests after tf whl is installed.
Amit Patankar [Wed, 6 Jun 2018 19:17:50 +0000 (12:17 -0700)]
Merge pull request #19818 from av8ramit/setuptools_fix2
Fixing the setuptools issue for pip builds.
Amit Patankar [Wed, 6 Jun 2018 18:41:23 +0000 (11:41 -0700)]
Fixing the setuptools issue for pip builds.
Amit Patankar [Wed, 6 Jun 2018 17:59:03 +0000 (10:59 -0700)]
Merge pull request #19817 from av8ramit/fix_cross_tower_ops
Iteritems is deprecated in python 3. Using items instead.
Amit Patankar [Wed, 6 Jun 2018 17:52:57 +0000 (10:52 -0700)]
Merge pull request #19815 from av8ramit/fix_operator2
Update __init__.py
Amit Patankar [Wed, 6 Jun 2018 16:51:06 +0000 (09:51 -0700)]
Iteritems is deprecated in python 3. Using items instead.
Dan Moldovan [Wed, 6 Jun 2018 12:56:29 +0000 (08:56 -0400)]
Update __init__.py
Whitelist the operators module in the main library.
Amit Patankar [Wed, 6 Jun 2018 01:08:15 +0000 (18:08 -0700)]
Merge pull request #19793 from av8ramit/add_auto
Adding the autograph operators dependency to the pip package.
Amit Patankar [Tue, 5 Jun 2018 23:29:00 +0000 (16:29 -0700)]
Adding the autograph operators dependency to the pip package.
Amit Patankar [Tue, 5 Jun 2018 23:41:19 +0000 (16:41 -0700)]
Merge pull request #19788 from av8ramit/cherrypicks2
Cherrypicks2
Amit Patankar [Tue, 5 Jun 2018 17:31:47 +0000 (10:31 -0700)]
Making setuptools the last install to ensure it's accurate.
Amit Patankar [Tue, 5 Jun 2018 17:28:38 +0000 (10:28 -0700)]
Change order of installations.
Amit Patankar [Tue, 5 Jun 2018 18:31:55 +0000 (11:31 -0700)]
Fixing line too long.
Amit Patankar [Tue, 5 Jun 2018 18:11:16 +0000 (11:11 -0700)]
Fixing the adamax_test rtol to be more lenient.
Asim Shankar [Tue, 5 Jun 2018 21:28:20 +0000 (14:28 -0700)]
contrib/eagerpython/datasets: Resource naming workaround.
tensorflow/contrib/eager/python/datasets_test.py was failing on GPU
because two tests - testTensorsPlacedOnDevice() and
testTensorsExplicitPrefetchToDevice() we're creating
FunctionBufferResources with the same shared_name, leading to
unintentional interference.
This change will make the tests pass and allow the use of
tf.contrib.eager.Iterator and
tf.data.Dataset.apply(prefetching_ops.prefetch_to_device)
in the same process without interference.
However, a more appropriate fix would probably be to use
anonymous function buffering resources (similar to
AnonymousIteratorHandle) when eager execution is enabled,
doing away with sharing by name.
Amit Patankar [Tue, 5 Jun 2018 16:47:09 +0000 (09:47 -0700)]
Merge pull request #19702 from av8ramit/relnotes_19b
Updating release notes for r1.9.
Amit Patankar [Tue, 5 Jun 2018 06:06:03 +0000 (23:06 -0700)]
Merge branch 'r1.9' into relnotes_19b
Amit Patankar [Mon, 4 Jun 2018 21:26:09 +0000 (14:26 -0700)]
Fixing raspberry pi file for conflict.
Nupur Garg [Mon, 4 Jun 2018 19:08:15 +0000 (12:08 -0700)]
Fix Python API.
PiperOrigin-RevId:
199171845
Akshay Modi [Mon, 4 Jun 2018 19:05:14 +0000 (12:05 -0700)]
Fix test user ops
PiperOrigin-RevId:
199171316
Brennan Saeta [Sat, 2 Jun 2018 01:00:43 +0000 (18:00 -0700)]
Remove use of absl::make_unique
absl is not yet ready for use by open source TensorFlow. :-(
PiperOrigin-RevId:
198952953
Amit Patankar [Mon, 4 Jun 2018 20:24:31 +0000 (13:24 -0700)]
Merge branch 'relnotes_19b' of github.com:av8ramit/tensorflow into relnotes_19b
Amit Patankar [Mon, 4 Jun 2018 20:23:40 +0000 (13:23 -0700)]
I've made the updates Rajat requested. Please note the links will not work until after we have launched.
Amit Patankar [Fri, 1 Jun 2018 22:27:16 +0000 (15:27 -0700)]
Merge branch 'r1.9' into relnotes_19b
Amit Patankar [Fri, 1 Jun 2018 22:22:54 +0000 (15:22 -0700)]
Merge pull request #19703 from av8ramit/update_version_190_rc0
Updating version for 1.9.0-rc0.
Amit Patankar [Fri, 1 Jun 2018 21:26:07 +0000 (14:26 -0700)]
Updating version for 1.9.0-rc0.
Amit Patankar [Fri, 1 Jun 2018 21:19:03 +0000 (14:19 -0700)]
Updating release notes for r1.9.
Michael Case [Fri, 1 Jun 2018 19:58:16 +0000 (12:58 -0700)]
Merge pull request #19680 from case540/branch_198811639
Branch
198811639
Clayne Robison [Fri, 1 Jun 2018 19:29:39 +0000 (12:29 -0700)]
Finished incomplete support for bad usernames in the CI build scripts. ci_build.sh now passes the environment variable to the container, and the with_the_same_user script adds the --force-badname param to addgroup as well. (#19699)
Michael Case [Fri, 1 Jun 2018 05:55:46 +0000 (22:55 -0700)]
Fix import depth issue.
Michael Case [Fri, 1 Jun 2018 04:55:11 +0000 (21:55 -0700)]
Potential fix to layout_optimizer_test.py
Jason Zaman [Fri, 1 Jun 2018 03:52:17 +0000 (11:52 +0800)]
BUILD: dont force stripping (#19599)
* BUILD: dont force stripping
Build systems must not strip binaries, it makes it impossible for
distros to ship debugging symbols for packages.
bazel build has a --strip option to allow the user to generate stripped
binaries in a configurable way, that should be used instead.
https://fedoraproject.org/wiki/Packaging:Debuginfo
https://wiki.gentoo.org/wiki/Project:Quality_Assurance/Backtraces#Stripping
Signed-off-by: Jason Zaman <jason@perfinion.com>
* configure: add --strip=always to bazelrc
Michael Case [Fri, 1 Jun 2018 03:44:41 +0000 (20:44 -0700)]
Add new line to make buildifier happy
Michael Case [Fri, 1 Jun 2018 03:27:17 +0000 (20:27 -0700)]
Merge branch 'master' into branch_198811639
Michael Case [Fri, 1 Jun 2018 02:46:48 +0000 (19:46 -0700)]
Fix sanity issues.
Felix Abecassis [Fri, 1 Jun 2018 02:45:15 +0000 (04:45 +0200)]
docker: update cuDNN to 7.1.4.18 (#19636)
Signed-off-by: Felix Abecassis <fabecassis@nvidia.com>
Pete Warden [Fri, 1 Jun 2018 02:44:05 +0000 (19:44 -0700)]
Manual roll back of PR #19443, because it causes the Raspberry Pi build to fail (#19678)
Michael Case [Fri, 1 Jun 2018 01:31:23 +0000 (18:31 -0700)]
Fix lite.py Python TypeError.
Michael Case [Fri, 1 Jun 2018 01:29:32 +0000 (18:29 -0700)]
Merge commit for internal changes
Allen Lavoie [Fri, 1 Jun 2018 01:17:48 +0000 (18:17 -0700)]
Make the TFOptimizer wrapper checkpointable.
TensorFlow Optimizers compiled with a Model will now have their state saved and restored with save_weights/load_weights.
PiperOrigin-RevId:
198811639
Peter Hawkins [Fri, 1 Jun 2018 01:09:50 +0000 (18:09 -0700)]
Automated g4 rollback of changelist
198803131
PiperOrigin-RevId:
198810875
Saurabh Saxena [Fri, 1 Jun 2018 00:20:31 +0000 (17:20 -0700)]
Throw a more informative error message when checkpointing an input pipeline
containing a ShuffleDataset with reshuffle_each_iteration=True.
This is a temporary fix till we figure out how to handle this use-case.
PiperOrigin-RevId:
198805344
Justin Lebar [Fri, 1 Jun 2018 00:19:25 +0000 (17:19 -0700)]
[XLA] Fix handling of CustomCall's window and dnums.
CustomCall can have a window and convolution-dimension-numbers, so
HloInstruction needs to handle this in Clone() and Identical().
PiperOrigin-RevId:
198805211
Derek Murray [Fri, 1 Jun 2018 00:18:54 +0000 (17:18 -0700)]
[tf.data] Update `DatasetBase::DebugString()` to be const in the docs.
PiperOrigin-RevId:
198805143
Peter Hawkins [Fri, 1 Jun 2018 00:17:13 +0000 (17:17 -0700)]
Mark tensorflow/contrib/learn:estimator_test as optonly because it is flaky due to timeouts without optimization.
PiperOrigin-RevId:
198804880
Peter Hawkins [Fri, 1 Jun 2018 00:03:07 +0000 (17:03 -0700)]
[TF:XLA] Preliminary support for tpu.replicate() inside of TF control flow (such as tf.while_loop()).
Register the remaining control-flow operators on XLA devices.
PiperOrigin-RevId:
198803131
Blake Hechtman [Thu, 31 May 2018 23:58:05 +0000 (16:58 -0700)]
[XLA] Fix batchnorm rewriter to not use implicit broadcasts. Algebraic
simplifier reshape change is now covered by ReshapeMover.
PiperOrigin-RevId:
198802494
Bixia Zheng [Thu, 31 May 2018 23:25:24 +0000 (16:25 -0700)]
Handle FilterLayout::kOutputYXInput in FilterDescriptor::ToShortString.
This fixes an error when running resnet50_batch128_fp16 with --v=2.
PiperOrigin-RevId:
198798196
A. Unique TensorFlower [Thu, 31 May 2018 23:21:24 +0000 (16:21 -0700)]
Edited the landing page for the Performance section. Reorganized content and removed references to content that is being deleted.
PiperOrigin-RevId:
198797662
Jiri Simsa [Thu, 31 May 2018 23:18:15 +0000 (16:18 -0700)]
[tf.data] Reflect `MakeIterator` signature change in documentation.
PiperOrigin-RevId:
198797254