gracehoney [Thu, 26 Apr 2018 14:33:34 +0000 (07:33 -0700)]
Merge pull request #18842 from aaroey/substr_bug_fix
Fix a bug where string::substr is used with wrong position.
Martin Wicke [Wed, 25 Apr 2018 21:30:39 +0000 (14:30 -0700)]
Merge pull request #18647 from junpenglao/ordered_bijector
Ordered bijector
Amit Patankar [Wed, 25 Apr 2018 18:49:14 +0000 (11:49 -0700)]
Merge pull request #18868 from av8ramit/fix_hp5y
Adding the h5py dependency in devel docker files.
Amit Patankar [Wed, 25 Apr 2018 17:57:06 +0000 (10:57 -0700)]
Adding install for libhdf5-serial-dev.
Amit Patankar [Wed, 25 Apr 2018 17:49:30 +0000 (10:49 -0700)]
Adding the h5py dependency in devel docker files.
Amit Patankar [Wed, 25 Apr 2018 17:04:03 +0000 (10:04 -0700)]
Merge pull request #18843 from av8ramit/fix_devel_docker
Fixing the mock import error for devel docker.
Junpeng Lao [Wed, 25 Apr 2018 17:01:38 +0000 (19:01 +0200)]
fix Non-ASCII character error
Junpeng Lao [Wed, 25 Apr 2018 08:57:46 +0000 (10:57 +0200)]
fix pylint error
Amit Patankar [Tue, 24 Apr 2018 23:34:01 +0000 (16:34 -0700)]
Fixing the mock import error for devel docker.
Billy Lamberta [Tue, 24 Apr 2018 22:48:08 +0000 (15:48 -0700)]
Merge pull request #18817 from lamberta/fix-install-linux
Docs: Clean up install_linux page
gracehoney [Tue, 24 Apr 2018 21:55:47 +0000 (14:55 -0700)]
Fix a bug where string::substr is used with wrong position.
Billy Lamberta [Tue, 24 Apr 2018 21:52:38 +0000 (14:52 -0700)]
docs: install_linux, move GPU section below install procedures.
Martin Wicke [Tue, 24 Apr 2018 19:16:18 +0000 (12:16 -0700)]
Merge pull request #18836 from tensorflow/av8ramit-patch-3
Update README.md
Billy Lamberta [Tue, 24 Apr 2018 18:09:09 +0000 (11:09 -0700)]
Review fixes to install_linux
Amit Patankar [Tue, 24 Apr 2018 17:59:10 +0000 (10:59 -0700)]
Update README.md
Billy Lamberta [Tue, 24 Apr 2018 05:08:52 +0000 (22:08 -0700)]
docs: Clean up install_linux with pip
Yifei Feng [Mon, 23 Apr 2018 22:22:57 +0000 (15:22 -0700)]
Merge pull request #18804 from tensorflow/merge_branch
R1.8 merge back to master after 1.8.0-rc1
Yifei Feng [Mon, 23 Apr 2018 21:24:11 +0000 (14:24 -0700)]
Reapply #18446.
Andy Kernahan [Mon, 23 Apr 2018 19:32:35 +0000 (20:32 +0100)]
Fix tfcompile module label. (#16582)
Yifei Feng [Mon, 23 Apr 2018 18:44:22 +0000 (11:44 -0700)]
Update resources.h
Yifei Feng [Mon, 23 Apr 2018 18:40:55 +0000 (11:40 -0700)]
Merge remote-tracking branch 'origin/r1.8' into merge_branch
Yifei Feng [Mon, 23 Apr 2018 17:53:11 +0000 (10:53 -0700)]
Merge branch 'master' into r1.8
Yifei Feng [Mon, 23 Apr 2018 17:02:49 +0000 (10:02 -0700)]
Merge pull request #18746 from yifeif/branch_193740595
Branch
193740595
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.
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
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
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
Martin Wicke [Mon, 23 Apr 2018 06:52:23 +0000 (23:52 -0700)]
Merge pull request #18728 from jerry4897/master
Typo?
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
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`
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
Martin Wicke [Mon, 23 Apr 2018 03:36:38 +0000 (20:36 -0700)]
Merge pull request #18766 from hedenface/patch-1
update $ source spacing
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
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
Yifei Feng [Mon, 23 Apr 2018 03:00:54 +0000 (20:00 -0700)]
Fix more for cuda version check.
Yifei Feng [Mon, 23 Apr 2018 02:08:21 +0000 (19:08 -0700)]
Update tuple for cuda version with auto as it was removed in #18434.
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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
Martin Wicke [Sun, 22 Apr 2018 05:54:45 +0000 (22:54 -0700)]
Merge pull request #18713 from ManHyuk/fix_typo
Fix typo
Martin Wicke [Sun, 22 Apr 2018 05:18:10 +0000 (22:18 -0700)]
Merge pull request #18757 from foo0x29a/master
Fix typo
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
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.
AG Ramesh [Sat, 21 Apr 2018 22:25:37 +0000 (15:25 -0700)]
Changed calls to the depreacted StringPiece::contains with str_util::StrContains
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>
foo0x29a [Sat, 21 Apr 2018 16:30:52 +0000 (13:30 -0300)]
fix typo
ADiegoCAlonso [Sat, 21 Apr 2018 09:56:10 +0000 (11:56 +0200)]
Specify float32 as float type instead of float64
ADiegoCAlonso [Sat, 21 Apr 2018 09:43:51 +0000 (11:43 +0200)]
Add __init__py
Yifei Feng [Fri, 20 Apr 2018 23:43:26 +0000 (16:43 -0700)]
Merge commit for internal changes
Yunxing Dai [Fri, 20 Apr 2018 23:30:02 +0000 (16:30 -0700)]
[XLA] Remove constant cast in literal util.
It's not portable to modify an underlying char array of a c++ string object: (https://stackoverflow.com/questions/5729203/modifying-underlying-char-array-of-a-c-string-object)
RELNOTES: n/a
PiperOrigin-RevId:
193740595
Asim Shankar [Fri, 20 Apr 2018 23:18:29 +0000 (16:18 -0700)]
tf.contrib.data.scan: Support eager execution.
PiperOrigin-RevId:
193739234
Frédéric Branchaud-Charron [Fri, 20 Apr 2018 23:10:41 +0000 (19:10 -0400)]
Fix casting in Keras estimator (#18104)
A. Unique TensorFlower [Fri, 20 Apr 2018 23:05:47 +0000 (16:05 -0700)]
Remove protected data members from GraphOptimizerStage.
PiperOrigin-RevId:
193737654
Brennan Saeta [Fri, 20 Apr 2018 22:51:16 +0000 (15:51 -0700)]
Support legacy clusters
PiperOrigin-RevId:
193735742
Derek Murray [Fri, 20 Apr 2018 22:38:06 +0000 (15:38 -0700)]
Always use the local worker name in CreateWorkerSession when not doing ClusterSpec propagation.
Previously, the master would send a job name and task index in an
otherwise-empty ServerDef, and the worker would unquestioningly use
those to build its worker name. However, this would lead to errors if
the worker had a local name like "/job:worker/replica:1/task:0",
because the ServerDef doesn't support non-zero replica IDs, and so the
local worker would end up an inconsistent view of what its worker name
should be. In particular `WorkerSession::worker_name` would disagree
with the device names added during graph partitioning by the master,
which would lead to runtime failures ("InvalidArgumentError: Invalid
rendezvous key").
PiperOrigin-RevId:
193733855
A. Unique TensorFlower [Fri, 20 Apr 2018 22:32:32 +0000 (15:32 -0700)]
[XLA] Redesign: implement DumpToDirectory for the HloSession.
This is to prepare the migration of tf2xla.
PiperOrigin-RevId:
193733029
A. Unique TensorFlower [Fri, 20 Apr 2018 22:20:37 +0000 (15:20 -0700)]
[XLA] Redesign: add comparator and printer for the XlaOp.
This is to prepare the migration of tf2xla. There were some codes used ComputationDataHandle::handle() for comparison/printing. Now implement XlaOp's comparator and printer.
PiperOrigin-RevId:
193731437
A. Unique TensorFlower [Fri, 20 Apr 2018 22:19:59 +0000 (15:19 -0700)]
Do not force default layout when there is no need to.
Allow the inner computations to negotiate a root and parameter layouts different from default.
PiperOrigin-RevId:
193731341
Akshay Modi [Fri, 20 Apr 2018 21:58:56 +0000 (14:58 -0700)]
Move the guts of TFE_Execute into EagerExecute
PiperOrigin-RevId:
193728072
A. Unique TensorFlower [Fri, 20 Apr 2018 21:32:07 +0000 (14:32 -0700)]
internal
END_PUBLIC
BEGIN_PUBLIC
Automated g4 rollback of changelist
193600682
PiperOrigin-RevId:
193723856
A. Unique TensorFlower [Fri, 20 Apr 2018 21:30:48 +0000 (14:30 -0700)]
Updating Generate Random Tensor to generate tensors whose values are small and do not cause overflow for arithmetic operations.
PiperOrigin-RevId:
193723661
A. Unique TensorFlower [Fri, 20 Apr 2018 21:25:57 +0000 (14:25 -0700)]
Pin rbe-debian8-tf container tp a newer base image
- Also improve how numpy is installed (not compiling from source) for containers based on other distros than Ubuntu14.04
PiperOrigin-RevId:
193722848
Derek Murray [Fri, 20 Apr 2018 20:58:51 +0000 (13:58 -0700)]
Automated g4 rollback of changelist
193694958
PiperOrigin-RevId:
193718607
A. Unique TensorFlower [Fri, 20 Apr 2018 20:56:55 +0000 (13:56 -0700)]
Track dependencies between outside_compilation clusters so that control edges can be correctly added to sequence compiled computations.
PiperOrigin-RevId:
193718295
Brian Patton [Fri, 20 Apr 2018 20:54:00 +0000 (13:54 -0700)]
Fix for: Suggest braces around initialization of subobject.
PiperOrigin-RevId:
193717872
A. Unique TensorFlower [Fri, 20 Apr 2018 20:48:37 +0000 (13:48 -0700)]
Merged commit includes the following changes:
193717076 by yifeif:
Automated g4 rollback of changelist
193713153.
--
193716750 by fchollet:
Refactor `tf.keras.layers.Embedding` layer to use `embedding_lookup` instead of `gather`. This makes the layer TPU-compatible.
--
193716664 by A. Unique TensorFlower:
Go: Update generated wrapper functions for TensorFlow ops.
--
193713153 by power:
Experimental Keras TPU compatibility layer.
--
PiperOrigin-RevId:
193717076
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
A. Unique TensorFlower [Fri, 20 Apr 2018 20:18:02 +0000 (13:18 -0700)]
Update ops-related pbtxt files.
PiperOrigin-RevId:
193712839
Allen Lavoie [Fri, 20 Apr 2018 19:40:57 +0000 (12:40 -0700)]
Add a utility to visualize object-based checkpoints
Useful for generating a warm fuzzy feeling that everything you think should be saved was saved, and for explaining what object-based checkpointing is. (Also useful on the former front will be a planned "assert that all of this Graph's trainable variables are accessible from object X" function.)
Somewhat hacky since it generates strings rather than using the pydot bindings (and so works without a pydot dependency).
PiperOrigin-RevId:
193708003
A. Unique TensorFlower [Fri, 20 Apr 2018 19:37:39 +0000 (12:37 -0700)]
Changed tf_to_tflite build rule.
PiperOrigin-RevId:
193707628
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
Yangzihao Wang [Fri, 20 Apr 2018 19:05:11 +0000 (12:05 -0700)]
Add native dilated support for conv3d and its gradients in cudnn v>=6.
PiperOrigin-RevId:
193703316
Dimitris Vardoulakis [Fri, 20 Apr 2018 19:02:56 +0000 (12:02 -0700)]
[TF:XLA] Now that the compiler no longer introduces implicit broadcasts, forbid them in the HLO verifier.
PiperOrigin-RevId:
193702874
Anna R [Fri, 20 Apr 2018 19:01:21 +0000 (12:01 -0700)]
Prefix compat import with underscore in meta_graph_transform.py so that it
doesn't get exported as part of API: https://www.tensorflow.org/versions/r1.8/api_docs/python/tf/contrib/meta_graph_transform/meta_graph_transform
PiperOrigin-RevId:
193702570
A. Unique TensorFlower [Fri, 20 Apr 2018 18:43:48 +0000 (11:43 -0700)]
[XLA] Fix a bug in ToProto: don't add gather attributes twice.
PiperOrigin-RevId:
193699745
Akshay Modi [Fri, 20 Apr 2018 18:34:55 +0000 (11:34 -0700)]
Move the guts of TFE_Op into EagerOperation
PiperOrigin-RevId:
193698320
Akshay Modi [Fri, 20 Apr 2018 18:24:53 +0000 (11:24 -0700)]
Allow creating tensors from numpy arrays, and other various constants - try #2
Allow type-inference from a different input tensor, similar to args_to_matching_eager.
- Update TFE_Py_TensorShapeSlice to take tuples.
- Update int values to allow int/long in py2
END_PUBLIC
BEGIN_PUBLIC
Automated g4 rollback of changelist
192184809
PiperOrigin-RevId:
193696790
James Qin [Fri, 20 Apr 2018 18:23:29 +0000 (11:23 -0700)]
Temporarily set cudnn Rnn math precision to fp32.
Problem:
When calling cudnnGetRNNLinLayerMatrixParams(), return error CUDNN_STATUS_BAD_PARAM if:
* RNN descriptor set math precision = CUDNN_DATA_FLOAT
* input descriptor dataType = CUDNN_DATA_HALF
* weight descriptor dataType= CUDNN_DATA_HALF
If updating Rnn descriptor math precision to CUDNN_DATA_HALF, then no error.
cudnn 7.1.4 will fix the problem.
PiperOrigin-RevId:
193696566
akindyakov [Fri, 20 Apr 2018 18:23:15 +0000 (11:23 -0700)]
Speed up safe_strtod and safe_strtof functions by using double-conversion library
Closes #12102.
PiperOrigin-RevId:
193696537
Derek Murray [Fri, 20 Apr 2018 18:13:16 +0000 (11:13 -0700)]
Use CreateWorkerSession and DeleteWorkerSession for all distributed sessions.
This change adds a phase to the session creation protocol: the master now contacts all workers to register a session handle and create a "WorkerSession" on each worker before it first registers or runs a graph on any worker. Subsequent requests to a worker ensure that the worker has the session handle registered before performing the request, and an AbortedError is raised if the worker has not (e.g. because it restarted after a failure).
As a result, more failure cases are covered by the high-level APIs (tf.estimator, Slim, etc.) that recreate the session on receiving an AbortedError. Previously, there was a possible race condition in which a PS task could restart between variable initialization and the first step, leading to a FailedPreconditionError ("Attempting to use uninitialized value") that would not be handled by the high-level APIs.
PiperOrigin-RevId:
193694958
Sanjoy Das [Fri, 20 Apr 2018 18:10:56 +0000 (11:10 -0700)]
[XLA:CPU] Use Eigen for F64 dot operations
PiperOrigin-RevId:
193694613
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
Junpeng Lao [Fri, 20 Apr 2018 18:06:02 +0000 (20:06 +0200)]
minor code styling
A. Unique TensorFlower [Fri, 20 Apr 2018 17:34:49 +0000 (10:34 -0700)]
Disable constant folding and arithmetic optimizations for functions.
PiperOrigin-RevId:
193688466
Jacques Pienaar [Fri, 20 Apr 2018 17:32:24 +0000 (10:32 -0700)]
Adding Shape inference functions to outfeed enqueue ops.
PiperOrigin-RevId:
193688099
A. Unique TensorFlower [Fri, 20 Apr 2018 17:16:03 +0000 (10:16 -0700)]
Change the TF record reader to use 16MB buffering by default in order to improve performance.
PiperOrigin-RevId:
193685521
Yifei Feng [Fri, 20 Apr 2018 17:08:40 +0000 (10:08 -0700)]
Merge pull request #18714 from yifeif/branch_193610264
Branch
193610264
James Keeling [Fri, 20 Apr 2018 16:54:50 +0000 (09:54 -0700)]
Correct error in "Adding An Op" docs.
The macro `REGISTER_KERNEL_BUILDER` always declared a functor specialized on floats, instead of the type actually passed into the macro.
PiperOrigin-RevId:
193682519
Mark Daoust [Fri, 20 Apr 2018 16:31:08 +0000 (09:31 -0700)]
Don't delete inbound_nodes and outbound_nodes, these no longer exist.
PiperOrigin-RevId:
193679512
Saurabh Saxena [Fri, 20 Apr 2018 16:23:52 +0000 (09:23 -0700)]
Fix msan error in MapAndBatchDataset.
While checkpointing tensors in BatchResult.output save only the initialized slice.
If the final batch is short, the entire batch tensor may not be initialized.
PiperOrigin-RevId:
193678679
Sanjoy Das [Fri, 20 Apr 2018 16:20:36 +0000 (09:20 -0700)]
[TF:XLA] Bump open source llvm revision to r330313
PiperOrigin-RevId:
193678317