| da1588ccf5c62eccac8013673359ac15b43eb394 |
|
03-Jan-2018 |
A. Unique TensorFlower <gardener@tensorflow.org> |
meta_graph export: Add support to strip default valued attributes.
Following APIs now accept an additional argument (`strip_default_attrs`) to
enable/disable (default:disabled) stripping of default valued attributes in a NodeDef:
o meta_graph: export_meta_graph, create_meta_graph.
o saver: Saver.save, Saver.export_meta_graph.
o builder: SavedModelBuilder.add_meta_graph,
SavedModelBuilder.add_meta_graph_and_variables.
o estimator: Estimator.export_savedmodel.
Related changes:
o Pywrap C++ AreAttrValuesEqual to compare two AttrValue instances.
This allows for a single/canonical way of comparing AttrValues in C++/Python.
o Add a utility method to meta_graph.py to get the node def by name in a graph def.
o Update SavedModelBuilder documentation on relevance of strip_default_attrs flag.
PiperOrigin-RevId: 180619001
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 20765b3e1ae3b718699592c98aa9805cb874b6d1 |
|
29-Dec-2017 |
Patrick Nguyen <drpng@google.com> |
Merge changes from github.
PiperOrigin-RevId: 180301735
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 1938feab97e36275f18a0745804299acfe137dc8 |
|
21-Dec-2017 |
Akshay Agrawal <akshayka@google.com> |
This change adds a mechanism to the internal C API for updating an output handle's shapes and types after its source operation has been created.
Context: framework/ops.py was recently updated to use the C API when setting shapes for an op's outputs. This update broke shape inference for graph functions that captured resource handles; this, in turn, made it impossible to create graph functions from Python methods that required fully defined shapes (e.g., like MNIST's `call` method). In particular, the C API computes shapes for ops when they are created and does not update them thereafter; this is problematic because when a resource handle is captured while building a function, we need to update the captured operation's output handle in order to propagate its outputs shapes and dtypes.
PiperOrigin-RevId: 179837104
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 968da4bf2722b1303cc223e8342357d62c27dfc1 |
|
12-Dec-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Raise exception on missing unused input_map keys with C API enabled.
Without this change, the C++ ImportGraphDef API returns unused
input_map keys (which are plumbed through to the C API as
well). However, the Python import_graph_def API requires slightly
different semantics: it throws an error for unused input_map keys that
are missing from the GraphDef.
This change modifies the C and C++ APIs to limit the returned keys to
those missing from the GraphDef, and plumbs this through to the C
API-enabled import_graph_def implementation.
Note that this is a change to the existing C API. Luckily the modified
method hasn't been released yet, so it's ok to change it.
PiperOrigin-RevId: 178783957
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 85818528f3caadb9f2a8e5b350cfe210d6661a7d |
|
07-Dec-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Enable using C API with Tensor.consumers
PiperOrigin-RevId: 178189037
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 79ad4a423f3e9031eb841a164372cc7476cc112a |
|
01-Dec-2017 |
Olivia Nordquist <nolivia@google.com> |
enabling Tensor._set_shape() to work with the C API
PiperOrigin-RevId: 177543170
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 32b861d3d4f920b46954a2e02aee1fbf46a81c63 |
|
30-Nov-2017 |
Gunhan Gulsoy <gunan@google.com> |
Automated g4 rollback of changelist 177353959
PiperOrigin-RevId: 177371177
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 197850fb12015f8e01a3b9c4d3e3546fc54aaa0b |
|
29-Nov-2017 |
Olivia Nordquist <nolivia@google.com> |
enabling Tensor._set_shape() to work with the C API
PiperOrigin-RevId: 177353959
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| c2aa66115d6c2bad6752474acd3bbf9a616b487a |
|
28-Nov-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Fetch shape information from the C API when enabled.
This change makes set_shapes_for_outputs fetch the already-computed tensor shapes from the C API, rather than calling the C++ shape function. It also moves the set_shapes_for_outputs call so it works with Operations created from TF_Operations.
PiperOrigin-RevId: 177088786
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| feda969d974d02016736faab328419929428ab10 |
|
23-Nov-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Make 'name' and 'return_elements' arguments to import_graph_def work with C API.
This enables some of the tests in import_test.py to be run using the C API.
PiperOrigin-RevId: 176706698
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 8067aa0862b7cd708f3a31accc2d232bafaf9442 |
|
22-Nov-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Create Python Operations for the TF_Operations created by import_graph_def.
This change also introduces Python functionality for iterating through
every TF_Operation in the graph and every newly-added TF_Operation via
TF_GraphNextOperation.
PiperOrigin-RevId: 176694180
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| de1cd503235a32ec216533d198dd6f6318655ab2 |
|
17-Nov-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Minor tf_session.i refactoring.
Moves inline C++ helper functions to top of file, and adds
CreateWrappedTFOutput and CreateWrappedTFOperation helper functions
(this is pulling out existing functionality that will be useful moving
forward).
PiperOrigin-RevId: 176131555
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 6c95675492aa8d25619f5e4ce1674582c051a7fe |
|
08-Nov-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Allow Operation._get_attr() to work with all attr types with C API enabled
This is achieved by accessing the AttrValue directly and using the
existing Python code instead of dispatching to the specific C API attr
getter for every type. I started going down the dispatch path, but it
turns out to be a lot of code (spread across Python, C, and SWIG), and
this is likely good enough from a performance standpoint. We can
optimize in the future if necessary.
In addition, changes the colocation group logic to use _set_attr() and
get_attr(), and makes _set_attr() work with the C API disabled. This
allows the colocation tests to pass with both the C API enabled and
disabled. Without these additional changes, the "_class" attribute
would be set on the C NodeDef, and then it would try to retrieve it
from the Python NodeDef.
PiperOrigin-RevId: 175050473
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 88917888f509e3e61ffe632534476e7b09d3326a |
|
02-Nov-2017 |
Andrew Harp <andrewharp@google.com> |
Merge changes from github.
END_PUBLIC
---
Commit d77b99809 authored by Yong Tang<yong.tang.github@outlook.com>
Committed by gunan<gunan@google.com>:
Update docs for `begin_params_axis` (#13979)
This fix fixes the issue raised in 13975 where `begin_shift_axis`
is actually `begin_params_axis`.
This fix fixes 13975.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit e6a242b4e authored by Yifei Feng<fengyifei2026@gmail.com>
Committed by gunan<gunan@google.com>:
Add GCC/Compiler version to issue template. (#14113)
As suggested in #13930
---
Commit 7ece1c0b8 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Moving model_pruning library to tf.contrib
PiperOrigin-RevId: 174214419
---
Commit 693325c83 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Log the full traceback in Coordinator.request_stop if it's available
PiperOrigin-RevId: 174213375
---
Commit 6c4a769ab authored by Mark Daoust<markdaoust@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Delete duplicate label_image script.
The version in examples/label_image is more complete (with image size and normalization options), so it can be used with `mobilenets`.
Also: removed bazel from main tutorial instructions.
PiperOrigin-RevId: 174212674
---
Commit 7a5b81c29 authored by Yao Zhang<yaozhang@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Materialize shape for ShapeN.
PiperOrigin-RevId: 174211500
---
Commit 78041b1dd authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
internal change
PiperOrigin-RevId: 174211190
---
Commit 2118fcf62 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
BUILD cleanup in contrib/tensor_forest/...
PiperOrigin-RevId: 174201884
---
Commit 6849ef8f6 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
internal change.
PiperOrigin-RevId: 174197506
---
Commit 37370d98f authored by resec<resec0109@gmail.com>
Committed by gunan<gunan@google.com>:
Support more Android arch in Makefile build (#12806)
* Support more Android arch in Makefile build
* update Makefile
* fix MARCH_OPTION
* persist multiple architectures across builds
* persist multiple architectures across builds
* persist multiple architectures across builds
* persistence bug fix
* persistence bug fix
* persistence bug fix
* add -latomic to linker flags for benchmark
* Change ANDROID_OS_ARCH to ANDROID_HOST_OS_ARCH
---
Commit c40d54173 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Exposes recall_at_top_k under tf.metrics.
PiperOrigin-RevId: 174189641
---
Commit 18bf5b2d9 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Return a classifier score of the same type as the logits.
PiperOrigin-RevId: 174184871
---
Commit 9da02be11 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Make 'collections' a list, as documented and expected by downstream custom getters.
PiperOrigin-RevId: 174184867
---
Commit 16b0bb095 authored by loki der quaeler<quaeler@users.noreply.github.com>
Committed by gunan<gunan@google.com>:
Adding a feed for boolean tensors to TensorFlowInferenceInterface (#14059)
* Sublime Text index-ignore file (a copy of .gitignore)
* Adding the requested implementation to TensorFlowInferenceInterface
* Removing Sublime Text .ignore file from remote repository
* indeed there was
---
Commit fa9d8aab4 authored by Urs K?ster<ursk@users.noreply.github.com>
Committed by gunan<gunan@google.com>:
Add 'log_progress' argument for tf.estimator.Estimator's evaluate function (#13695)
* Add argument for tf.estimator.Estimator's evaluate function
* add log_progress argument to ._convert_eval_steps_to_hooks for TPU estimator
* log only every 10th step if more than 100 iterations in _StopAfterNEvalsHook
* ensure last step is logged and aim for 10 outputs total
---
Commit 07a91dac5 authored by nolan liu<nolan.liou@gmail.com>
Committed by gunan<gunan@google.com>:
make `gather` cpu kernel to be multiple threads. (#12246)
* Change the gather op to multi-thread.
* Modify the gather kernel of xla compiler in order to be compatible with multi-threads cpu kernel.
* Add prefetch logic to gather op kernel.
* Update the indention of gather op kernel code.
* Update the gather kernel code for multiple thread.
* Remove reference to ealier version of code in gather functor.
* Change the framework_lite dep of gather_functor to framework.
* Remove mutex guard in gather functor.
---
Commit a956486be authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Remove an erronous __attribute__((...)) tag.
There is no __attribute__((guarded)) or __attribute__((pt_guarded)) attribute in Clang, and if we turn on warnings for unknown attributes (which are currently turned off), this causes build failures. This means that, when the warnings are turned off, this is simply a no-op.
PiperOrigin-RevId: 174134252
---
Commit 27412f3b6 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add compiler/tf2xla/sharding_util.h with utilities for getting the core device from
a Node.
PiperOrigin-RevId: 174133602
---
Commit ab4349a26 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
BUILD cleanup in selected packages in contrib/...
PiperOrigin-RevId: 174115744
---
Commit 4aa90bfd3 authored by Justin Lebar<jlebar@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[XLA] Add HLO matchers that check parameter numbers and GTE indices.
This lets you do
EXPECT_THAT(foo, op::Parameter(42));
and
EXPECT_THAT(bar, op::GetTupleElement(baz, 8));
PiperOrigin-RevId: 174113597
---
Commit f97e7c69b authored by Olivia Nordquist<nolivia@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
partially exposing the _set_attr and _get_attr method in python
PiperOrigin-RevId: 174113043
---
Commit 8e732a312 authored by Artem Belevich<tra@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Prefer cubin over PTX when we launch CUDA kernels.
Native GPU code, if we have it, should be preferred over JIT compilation of PTX.
PiperOrigin-RevId: 174110646
---
Commit 2ccf3aba4 authored by Eugene Brevdo<ebrevdo@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Permanently remove several modules from tf.contrib.bayesflow.
These modules are very infrequently used and will not be developed moving forward.
Removing this code paves the way for remaining modules in tf.contrib.bayesflow
to move to their own repo.
PiperOrigin-RevId: 174110067
---
Commit ef7052fbd authored by Andrew Selle<aselle@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Open source build support for TensorFlow Lite Toco.
- Handle proto incompatibilities
- Mixed bazel compatibility fixes.
- Add link to absl libraries
PiperOrigin-RevId: 174103981
---
Commit d6a9cd40c authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Fix "hides overloaded virtual function" error in default/gpu_tracer.cc when compiled with -Werror,-Woverloaded-virtual.
PiperOrigin-RevId: 174101519
---
Commit b242a7988 authored by Mustafa Ispir<ispir@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Set metric variable initializers as lambda.
PiperOrigin-RevId: 174100686
---
Commit 57b1c5621 authored by Alan Yee<alyee@ucsd.edu>
Committed by drpngx<drpngx@users.noreply.github.com>:
Add deprecation notes (#12614)
* Update lookup_ops.py
Minor comment fix
* Update metrics_ops.py
Add deprecated notes
* Update tensor_util.py
Update deprecated note on remove_squeezable_dimensions
* Update metric_ops.py
Add deprecated notes
---
Commit 453dd5848 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
K-FAC: Support for tf.AUTO_REUSE when re-using registrations. Multi-tower support for FullFB, NaiveDiagonalFB. Removal of LayerCollection.generic_registrations.
PiperOrigin-RevId: 174092003
---
Commit 0a7be5a2f authored by Sanjoy Das<sanjoy@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Rename (Add|Get)ProfileResult to something more specific; NFC
PiperOrigin-RevId: 174084570
---
Commit f1916f8f6 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
- Remove slice hack to properly initialize missing entries in weight matrices
- Add real support for EmbeddingColumns / input_layer()
- Fix warmstarting for non-PartitionedVariables
PiperOrigin-RevId: 174083777
---
Commit f567ddf87 authored by Alex Sergeev<alexander.sergeev@live.com>
Committed by drpngx<drpngx@users.noreply.github.com>:
Add tf.sysconfig.get_compile_flags() & tf.sysconfig.get_link_flags() for custom operators (#13496)
* Add flags for custom op compilation
* Move ABI logic into version_info.cc
* Add #include <string> to be able to read _GLIBCXX_USE_CXX11_ABI value.
* Make flags to be lists
* Add _flag to cxx11_abi
* Address review comment.
* Move CXX import to the top level.
* Add goldens update
---
Commit 0cddb9bca authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Go: Update generated wrapper functions for TensorFlow ops.
PiperOrigin-RevId: 174074499
---
Commit ba8c38959 authored by Neal Wu<wun@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Change wide_deep.md and wide.md to reference the TensorFlow official models version rather than the tf.contrib.learn version
PiperOrigin-RevId: 174074112
---
Commit f3006422c authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Make `RunTrainOpsHook` public.
PiperOrigin-RevId: 174073925
---
Commit 21dafd6d2 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Update ops-related pbtxt files.
PiperOrigin-RevId: 174073569
---
Commit 66fc99a3b authored by Artem Belevich<tra@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[XLA:GPU] Short-circuit compilation of no-op IR -> empty PTX.
There's no point constructing/running LLVM pipeline if we know that we have no
kernels in the IR we've generated for the given HLO op. This is often the case
for ops we can optimize away at the HLO level.
PiperOrigin-RevId: 174072540
---
Commit c911d0f16 authored by Dhananjay Nakrani<dhananjayn@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Switch over python calls to RandomPoissonV2.
Part 2 of Support int32/64 in tf.random_poisson().
PiperOrigin-RevId: 174071745
---
Commit b5d5326c6 authored by Justin Lebar<jlebar@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[XLA:GPU] Fix race condition in gpu_compiler.cc.
We were racing on libdevice_dir_.
PiperOrigin-RevId: 174070334
---
Commit 35939d2d3 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[TF:XLA] Fix string to HLO opcode conversion for atan2, complex, imag and real.
Make sure that we can't forget opcodes by auto-generating the conversion
functions.
Add auto-generated functions to test HLOs for properties (like IsVariadic,
IsComparison, etc.)
This makes changing HLO more robust and easier because there are fewer places
to update when adding or removing an HLO opcode.
Also:
* Fix IsElementwiseBinary for atan2.
* Add a unit test for HLO opcode helpers.
* Express IsElementwiseBinary in terms of IsElementwise() and operand_count()
to avoid having to keep the two in sync manually.
PiperOrigin-RevId: 174069664
---
Commit 3b845c80d authored by Allen Lavoie<allenl@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Disable resnet50_graph_test under TSAN due to timeouts.
PiperOrigin-RevId: 174066937
---
Commit 8a09bbc4a authored by Igor Ganichev<iga@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add TFE_Py_TensorShapeSlice function
TFE_Py_TensorShapeSlice takes a list of EagerTensors and returns a list
of their i'th dimensions. This utility is fairly niche but it is simple
and reduces SPINN training time by over 12%.
PiperOrigin-RevId: 174065044
---
Commit 585432cc2 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Refactor ArgMin / ArgMax index ops as XlaHelpers.
PiperOrigin-RevId: 174061370
---
Commit e6faa845c authored by Michael Case<mikecase@chromium.org>
Committed by gunan<gunan@google.com>:
Merge v1.4-rc1 back into master branch. (#13960)
* Update RELEASE NOTES for TensorFlow 1.4
* Update the version strings for TF 1.4-rc0.
* Update version strings in POM files missed by update script.
* Pin TensorBoard 0.4 to TensorFlow 1.4
* Fixing the name of the disabled test. (#13592)
* Revert "Implementing ghost batch norm as defined in https://arxiv.org/pdf/1705.08741."
This reverts commit 125f7afa4a483855dc75791445d2dea64587876a.
* Disable iterator_ops_test on Windows for 1.4 release (#13609)
* Disable failing Windows tests for r1.4 release.
testRemoteIteratorUsingRemoteCallOpDirectSessionGPUCPU test is failing
with "TypeError: only integer scalar arrays can be converted to a scalar
index" on the Windows GPU Release bot. Disabling test.
* Fix typo.
* Also disalbe iterator_ops_test from contrib/.
* Add contributing authors to 1.4 Release notes.
Thanks!
* Fixes to authors.
Removed duplicate and removed googler from contributing author list.
* Fixes and additions to release notes.
Added line about Keras moving into core.
Added line about CUDA/cuDNN versions.
Added line about custom ops.
* Fixing a master regression (#13562)
* Update version strings for 1.4.0rc1
* Remaining cherry-picks for 1.4.0rc1 (#13700)
* Java: Tweak to address some Javadoc errors.
PiperOrigin-RevId: 171987329
* Fix S3 BUILD not including files explicitly.
This causes remote builds to fail since they AWS headers were missing.
PiperOrigin-RevId: 171718021
* Add missing default config setting in aws.BUILD (#13662)
* Remove setting AWS logging for S3 file system.
Was causing issues with tests. Can repro test failures on Macs by running...
bazel test --config=s3 --cache_test_results=no --test_output=streamed
//tensorflow/core/kernels:control_flow_ops_test
Possible reason for error is symbol collision with AWS logging code.
One possible solution would be to split out another shared object for
the S3 filesystem op which does not link in libtensorflow_framework.so.
This is done, for example, by libforestprotos.so in
tensorflow/contrib/tensor_forest/BUILD
PiperOrigin-RevId: 171246381
* Relanding change to add config to enable S3 file system support.
Pass --config=s3 argument to Bazel to build with S3 file system support.
Change was originally rolled back due to a failure it caused in
//tensorflow/core/kernels:control_flow_ops_test on Macs which is now fixed.
PiperOrigin-RevId: 171579378
* Update release notes about Amazon S3 file system support being default.
* Add documentation to sloppy_interleave function
PiperOrigin-RevId: 171303413
* Add `cudnn_rnn_ops` to the Windows build
Fixes #13696.
* Creating a patch for the wrong links that still point to dev. (#13753)
* tfdbg release notes in r1.4
* Fix ambiguous type comparison in s3_crypto.cc (#13758)
tensorflow/contrib/s3/s3_crypto.cc(74): error C2666:
'std::fpos<_Mbstatet>::operator ==': 3 overloads have similar conversions
could be 'bool std::fpos<_Mbstatet>::operator ==(std::streamoff) const'
or 'bool std::fpos<_Mbstatet>::operator ==(const std::fpos<_Mbstatet> &)
We were seeing this compilation error on Windows builds.
* Set estimator run_config default random seed to None. This will make it aligned with other parts of the TF. Many users are not aware of impact of non-random seed. For example it may lead to train only on a small fraction of training data due to preemptions.
We're changing default behavior since we consider it as a bug fix.
PiperOrigin-RevId: 172519268
* Move global_step_read dependency to model_fn instead of input_fn.
PiperOrigin-RevId: 172366972
* [tf.data] Fix broken implementation of `Dataset.from_generator()` on Windows.
Due to a mix-up between NumPy's default array element type for a Python `int` on Windows and Linux, a tf.py_func() in `Dataset.from_generator()` would appear to return the wrong type on Windows (np.int32 instead of np.int64).
All code using `Dataset.from_generator()` on Windows was previously broken. This change fixes both `tf.data.Dataset.from_generator()` and `tf.contrib.data.Dataset.from_generator()`. It also enables test coverage for this method on Windows, which should prevent future breakage.
PiperOrigin-RevId: 172346533
* Update RELEASE notes for change to run_config random seed.
* Disable probable timeout flake on Ubuntu machines.
PiperOrigin-RevId: 172408922
* Disabling failing contrib tests.
* Disable S3 on Windows due to build issues.
* Update serving_input_fn argument name to serving_input_receiver_fn
PiperOrigin-RevId: 172787460
* Update the C++ API guide (#13858)
- Adds the standard warning at the top that people may want the master branch
- Includes a documentation fix for 1.4 (cc_binary -> tf_cc_binary to avoid
undefined symbols).
* Add known Dataset issue to RELEASE.md. (#13870)
Adding info about issue using Unicode strings with Datasets.
* Fixes to merge.
* Fix spelling of tensorflow in install_sources.md
---
Commit 6eac524ef authored by cglewis<clewis@iqt.org>
Committed by cglewis<clewis@iqt.org>:
Use 'LABEL maintainer=' in Dockerfile
* Use 'LABEL maintainer=' in Dockerfile
This fix is a follow up of 13961 to replace `MAINTAINER`
with `LABEL maintainer=` in Dockerfile. The keyword
`MAINTAINER` has long been deprecated and is replaced by `LABEL`,
which is much more flexible and is easily searchable through `docker
inspect`.
This fix replaces remaining `MAINTAINER` with `LABEL`.
Signed-off-by: Charlie Lewis <clewis@iqt.org>
* Additional `MAITAINER` -> `LABEL`
Signed-off-by: Charlie Lewis <clewis@iqt.org>
---
Commit 469970260 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Modify quantization to support add ops that occur after Conv2D
PiperOrigin-RevId: 174058697
---
Commit 938643b56 authored by Amit Patankar<amitpatankar@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Replace the docker check with an OS check.
PiperOrigin-RevId: 174057778
---
Commit 5f1a66ccb authored by Igor Saprykin<isaprykin@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add more recovery functionality to MonitoredSession.run_step_fn.
Current implemention wouldn't recover from one of `_PREEMPTION_ERRORS` during a fetch through the raw session that is made available to the step_fn.
The changelist presents a way to map the desired functionality to the hiearchy of _MonitoredSession > (possibly!) _RecoverableSession > _CoordinatedSession > _HookedSession.
PiperOrigin-RevId: 174053865
---
Commit 9a2b0983a authored by Yifei Feng<fengyifei2026@gmail.com>
Committed by gunan<gunan@google.com>:
Add apt-key for ubuntu keyserver (#14114)
---
Commit 479ee24a0 authored by Asim Shankar<asimshankar@gmail.com>
Committed by gunan<gunan@google.com>:
eager: Update broken link in README (#14136)
---
Commit ad7bb2b9e authored by Asim Shankar<asimshankar@gmail.com>
Committed by gunan<gunan@google.com>:
eager: Update broken links in guide.md (#14135)
---
Commit c37ebf0d5 authored by Thomas Deegan<tadeegan@gmail.com>
Committed by gunan<gunan@google.com>:
Resolve //tensorflow relative to tensorflow repo so that tfcompile.bzl can be correctly loaded from another Bazel project (#14103)
---
Commit b2ff3ad96 authored by Mustafa Ispir<ispir@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Added GraphKeys.METRIC_VARIABLE collection. Added all variables under tf.metrics and tf.contrib.metrics into this collection. This will enable replication of model for evaluation. When we replicate a metric in multiple towers (let's say for each qpu we replicate same model/metric), we cannot reduce the output of metrics. On the other hand internal state (local-variables) of those metrics can reducible via sum.
PiperOrigin-RevId: 174051559
---
Commit 98dad195d authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Adds sigmoid to the list of operations that can be recomputed.
PiperOrigin-RevId: 174047825
---
Commit 123749fb1 authored by Yuan (Terry) Tang<terrytangyuan@users.noreply.github.com>
Committed by Martin Wicke<martin.wicke@gmail.com>:
Remove Scikit Flow link and description (#14036)
---
Commit 0d118e4dc authored by Benoit Steiner<bsteiner@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Implemented tensorflow::port::NominalCPUFrequency()
PiperOrigin-RevId: 174041196
---
Commit 648993e82 authored by Andrew Harp<andrew.harp@gmail.com>
Committed by Andrew Harp<andrew.harp@gmail.com>:
delete extraneous file
---
Commit c2ff8a5ab authored by Mark Daoust<markdaoust@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Delete backticks
PiperOrigin-RevId: 174030921
---
Commit 333ba224d authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Dependency information for Skylark macros
PiperOrigin-RevId: 174023371
---
Commit 9ee0cecec authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Shrink the model size for unit test.
PiperOrigin-RevId: 174001263
---
Commit c44f67a7e authored by Yifei Feng<fengyifei2026@gmail.com>
Committed by gunan<gunan@google.com>:
Disable clang_format check. (#14115)
Different clang_format version can cause different formats with the same style option. This check might be too strict. Disable for now.
---
Commit a6a618843 authored by Asim Shankar<ashankar@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
eager: Documentation and example models.
- Updated README
- A preliminary "User's Guide"
- A few example models, some with benchmarks
PiperOrigin-RevId: 173996303
---
Commit de38e5dff authored by ???<dev@goodow.com>
Committed by GitHub<noreply@github.com>:
fix broken link
---
Commit cd81bc8e0 authored by Rohan Jain<rohanj@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Adds a PrefetchWithFn op to contrib/data. Alongwith the FunctionBufferingResource, this can be used to prefetch and fill up a buffer by making repeated function calls.
Also fixes a TODO in the ProcessFLR implementation to respect alloc_attrs for Rendezvous calls.
PiperOrigin-RevId: 173990680
---
Commit 17695212c authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[TF:XLA] Don't pass HLO operands in HandleAtan2.
This makes it consistent with the rest of the Visit methods where we only
pass the HLO itself.
PiperOrigin-RevId: 173990595
---
Commit 113be5746 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
A few profiler improvements
1. Track the full allocation history of each tensor, visualized in timeline.
2. Better ProfileContext for tracing step selection.
3. Small bug fix.
PiperOrigin-RevId: 173988293
---
Commit 6d1263cdf authored by Justin Lebar<jlebar@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[XLA] Remove dead opcode kIndex.
PiperOrigin-RevId: 173987428
---
Commit a4b5356e4 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[TF:XLA] Reduce boilerplate code in HLO visitors.
Only pass the HloInstruction into visitor methods. This makes changing
instructions and visitors easier.
PiperOrigin-RevId: 173983398
---
Commit d9cee35b6 authored by LevineHuang<levinehuang@163.com>
Committed by Benoit Steiner<benoitsteiner@users.noreply.github.com>:
Typo fix in file 'fully_connected_feed.py' (#14033)
* Typo fix in file 'fully_connected_feed.py'
* Minor edits to coding style
---
Commit bb7ed1c88 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
K-FAC: Multi-tower ConvNet example.
PiperOrigin-RevId: 173982527
---
Commit 2ba529856 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Initial add of docs for Tensorflow on Mobile.
PiperOrigin-RevId: 173980290
---
Commit 187453d61 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Change momentum optimizer to allow callable learning_rate and momentum
parameters. This can be useful for implementing learninge rate decay.
PiperOrigin-RevId: 173975321
---
Commit 542b323e5 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Register quint16/qint16 for GatherOp.
PiperOrigin-RevId: 173974904
---
Commit 309e34061 authored by Allen Lavoie<allenl@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Avoid uncollectable cycles with a separate deleter object for resources.
PiperOrigin-RevId: 173972515
---
Commit 73fdaf0b5 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Summary-writing support for Evaluators.
PiperOrigin-RevId: 173971621
---
Commit 72be26dc8 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[tf.data] Iterator Save and Restore for Dataset.from_tensors(..), Dataset.from_tensor_slices(..) and dataset.concatenate(..).
PiperOrigin-RevId: 173971324
---
Commit 09f62ab38 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Speeding up the case for sparse float columns that have only 1 value.
PiperOrigin-RevId: 173971121
---
Commit c315cf1ee authored by Shanqing Cai<cais@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Internal-only changes
PiperOrigin-RevId: 173968246
---
Commit 293ba20be authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Make learning_rate_decay.piecewise_constant work in Eager mode.
PiperOrigin-RevId: 173967531
---
Commit 0e6abfcda authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
K-FAC: Example for multi-tower support for MNIST MLP.
PiperOrigin-RevId: 173967370
---
Commit b46c196e9 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
* Add graph rewrite rule that removes repeated application of scalar unary ops that are involutions (their own inverse).
* Update rewrite rule for Transpose to also handle ConjugateTranspose.
PiperOrigin-RevId: 173967184
---
Commit ff5c276ad authored by Stephan Hoyer<shoyer@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Longer README for tf.contrib.labeled_tensor
PiperOrigin-RevId: 173966577
---
Commit 558f146e1 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Go: Update generated wrapper functions for TensorFlow ops.
PiperOrigin-RevId: 173966068
---
Commit f9a673cb7 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
In the overloaded HloVerifier::CheckShape, include the failing instruction in
the error message.
PiperOrigin-RevId: 173965368
---
Commit 302ab0ff7 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Update ops-related pbtxt files.
PiperOrigin-RevId: 173965174
---
Commit 89120eb68 authored by Alexandre Passos<apassos@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
scatter_update for resource variables
PiperOrigin-RevId: 173963715
---
Commit 8f7903b4c authored by Justine Tunney<jart@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Introduce SQLite SummaryWriterInterface
This change allows tensors to be written from the graph, as they flow, directly
to the database. Many of the important details haven't been implemented yet.
This has been done with the new summary interface that's going to be used with
eager.
PiperOrigin-RevId: 173961448
---
Commit 9aaa49a4e authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Avoid using variables as booleans (similarly to tensors).
PiperOrigin-RevId: 173956625
---
Commit a60cd87c4 authored by Alexandre Passos<apassos@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
No need for unique variable names in eager.
PiperOrigin-RevId: 173954805
---
Commit f17f389d8 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add a workaround in the Grappler arithmetic optimizer for the "Add" op not being marked commutative. This will allow Grappler to dedup nodes Add(x,y) and Add(y,x).
PiperOrigin-RevId: 173950586
---
Commit e40eb810a authored by Shanqing Cai<cais@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
TFE: Add errors for classic tf.summary.* ops and FileWriter
PiperOrigin-RevId: 173949980
---
Commit 25620825b authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Dataset: Adds eager warnings to make_initializable_iterator and make_one_shot_iterator.
PiperOrigin-RevId: 173949737
---
Commit 1d6dae88e authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add check to tf.device when called with a function in eager mode.
PiperOrigin-RevId: 173947845
---
Commit 3639aa7ff authored by Alexandre Passos<apassos@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Always run iterator deleter in eager mode for safety.
PiperOrigin-RevId: 173947019
---
Commit efcbf6e34 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Supported in this CL:
* Attaching sharding descriptors to HLO ops
* Partitioning the HLO graph into per-device computations based on those sharding descriptors.
* All operator support for device placement and ops replicated on all devices.
* Elementwise op support for tiled shardings.
* 2D Convolution support for tiled shardings (no stride or dilation support).
PiperOrigin-RevId: 173946036
---
Commit 682a6ed64 authored by Jon Shlens<shlens@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Update the documentation for sample_distorted_bounding_box
PiperOrigin-RevId: 173943029
---
Commit 4f6e6ea4c authored by Sanjoy Das<sanjoy@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Fix typo in comment; NFC
PiperOrigin-RevId: 173942305
---
Commit 07584221f authored by Anna R<annarev@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Set visibility to HIDDEN for hidden Python ops in ApiDef.
PiperOrigin-RevId: 173942001
---
Commit 35cc8bb0a authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
K-FAC: Multiple minibatches support for LayerCollection.register_conv2d()
PiperOrigin-RevId: 173941279
---
Commit 32f3c3a43 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Go: Update generated wrapper functions for TensorFlow ops.
PiperOrigin-RevId: 173933228
---
Commit 8cc7b47a4 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Update ops-related pbtxt files.
PiperOrigin-RevId: 173932574
---
Commit b9337de5b authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
K-FAC: Multi-tower support for ConvKFCBasicFB
PiperOrigin-RevId: 173932013
---
Commit 1b6b7e208 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add registration for op AddV2, which is identical to Add, except that it does does not implement string concatenation. This allows us to mark AddV2 is_commutative and is_aggregate, which will allow optimizers more freedom.
PiperOrigin-RevId: 173931848
---
Commit 629e6d0c1 authored by Joshua V. Dillon<jvdillon@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Bugfix: Make `tf.contrib.distributions.Independent` tests not flaky.
PiperOrigin-RevId: 173921378
---
Commit 4b63f47d9 authored by Justin Lebar<jlebar@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[XLA:CPU] Don't crash if someone tries to do dot(X, X) or dot(X, X^T).
PiperOrigin-RevId: 173919310
---
Commit 89582677c authored by Alexandre Passos<apassos@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
EagerVariableStore, for compatibility with functional layers.
PiperOrigin-RevId: 173915730
---
Commit cef680b53 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Enable shape inference on functions in grappler.
PiperOrigin-RevId: 173914941
---
Commit e8ac0b48f authored by Akshay Agrawal<akshayka@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Report a nicer error message when differentiating a function
that returns None in eager
PiperOrigin-RevId: 173914883
---
Commit 85f8d9240 authored by Eugene Brevdo<ebrevdo@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[tensorflow training input] If SparseTensors are used in batch* ops, ensure restoration.
This forces the ST restore op to be called if any tensors are accessed at the output
of the batch, thus fixing a memory leak.
Solution suggested by Derek Murray.
Fixes #13999.
PiperOrigin-RevId: 173904309
---
Commit 7fd261602 authored by Skye Wanderman-Milne<skyewm@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add TF_GraphVersions() to C API and use in Graph.graph_def_versions()
PiperOrigin-RevId: 173902666
---
Commit 4723f8f6e authored by RJ Ryan<rjryan@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Support SymbolicGradient for functions with non-trainable arguments.
The non-trainable arguments end up with None as their incoming out_grad, which is not a valid input to SymbolicGradient (inputs have to be convertible to Tensor, and None isn't).
PiperOrigin-RevId: 173901727
---
Commit 494672475 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Added "NOTE: You may only install TensorFlow on 64-bit machines" to all the
TensorFlow Install guides.
PiperOrigin-RevId: 173899394
---
Commit b73743e3a authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Remove accidental disablation of (already manual) tests.
PiperOrigin-RevId: 173898910
---
Commit ce0238198 authored by Skye Wanderman-Milne<skyewm@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add ability to fetch return nodes and unused input mappings from C API GraphDef import
This change introduces yet another ImportGraphDef function to the C
API (TF_GraphImportGraphDefWithResults), but this one has extensible
return values so we shouldn't have to add more in the future.
This change also modifies the ImportGraphDef C interface to manage all
string data for the user.
PiperOrigin-RevId: 173894710
---
Commit ef4490f63 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
BUILD cleanup in contrib/...
PiperOrigin-RevId: 173889798
---
Commit 2e54fd6de authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Adds eager execution compatibility note in Readers, Queues, and QueueRunner.
Raises a RuntimeError in base classes for QueueBase, ReaderBase, and QueueRunner.
PiperOrigin-RevId: 173888425
---
Commit 32ab30cb0 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Fixes typo in compatibility.
PiperOrigin-RevId: 173887031
---
Commit 325c8e5ef authored by Justine Tunney<jart@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Improve C++ SQLite veneer
- Use shared_ptr for Sqlite
- Don't need unique_ptr on SqliteStatement
- Don't need db namespace
- Include SQL in error statuses
PiperOrigin-RevId: 173802267
---
Commit 0eba15fe6 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Adds eager compatability message for PartitionedVariable.
PiperOrigin-RevId: 173772851
---
Commit e7645b629 authored by Justin Lebar<jlebar@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[XLA] DOT dumper: Handle fusion nodes nested inside other nodes (e.g. map).
PiperOrigin-RevId: 173752314
---
Commit 8ec7540e0 authored by Shanqing Cai<cais@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
TFE: Fix pip test for tf.contrib.summary
Fixes test failure in tensorflow/contrib/summary:summary_ops_test, e.g.,
http://ci.tensorflow.org/job/tensorflow-cl-cpu-python3-pip/10933/console
PiperOrigin-RevId: 173749502
---
Commit c16797ec3 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Adds eager execution compatibility note in Estimators.
Raises a RuntimeError in Estimator base class.
PiperOrigin-RevId: 173744765
---
Commit e8a62a30b authored by ???<dev@goodow.com>
Committed by GitHub<noreply@github.com>:
Fix minor typo
---
Commit 36696ad58 authored by ???<dev@goodow.com>
Committed by Larry Tin<dev@goodow.com>:
tf.zeros doesn't accept a tensor argument
ValueError: Shape must be rank 1 but is rank 0 for 'zeros_2' (op: 'Fill') with input shapes: [], [].
---
Commit 9f4b12bb5 authored by Justin Lebar<jlebar@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[XLA] DOT dumper: Print constant shape when we elide the constant's value.
For example, instead of "operand 1 = %constant.42", we now print
"operand 1 = %constant.42 (f32[100])".
PiperOrigin-RevId: 173741373
---
Commit 45c5118f0 authored by Mark Heffernan<meheff@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
When creating an HloModule from an HloProto construct the HloModuleConfig
with a correct ProgramShape which matches the shapes of the entry computation.
Previously the module config had a bogus or default constructed ProgramShape.
PiperOrigin-RevId: 173741104
---
Commit 09a89ae57 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add `tf.contrib.distributions.bijectors.Reshape`.
PiperOrigin-RevId: 173740491
---
Commit 729db035e authored by Mark Daoust<markdaoust@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Allow compatibility notes in class, property and module doc-strings
PiperOrigin-RevId: 173739674
---
Commit ca56fa49a authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Go: Update generated wrapper functions for TensorFlow ops.
PiperOrigin-RevId: 173739110
---
Commit 48df7c972 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Update ops-related pbtxt files.
PiperOrigin-RevId: 173738765
---
Commit fb2c84cb2 authored by Jeremy Lau<lauj@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Internal change
PiperOrigin-RevId: 173738655
---
Commit 245a5c171 authored by Akshay Agrawal<akshayka@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Make functional_ops compatible with eager exeuction by ignoring
caching devices when in eager mode
PiperOrigin-RevId: 173737949
---
Commit d1c59bd37 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add tf.quantize op, which is the same as tf.quantize_v2.
PiperOrigin-RevId: 173735986
---
Commit 3ff9c8d2a authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Fix typos in Linear Model Tutorial samples
1. test_file_name is undefined (should be test_file.name)
2. train_file_name is undefined (should be train_file.name)
PiperOrigin-RevId: 173733442
---
Commit abbab2430 authored by Michael Case<mikecase@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add bazel mirror links for newly added workspace dependencies.
PiperOrigin-RevId: 173732606
---
Commit 46a577feb authored by Derek Murray<mrry@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[CMake] Generate audio_ops wrappers in the CMake build.
Fixes #14004.
PiperOrigin-RevId: 173732397
---
Commit 7cb7f88c5 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add count metric, a helper function that computes the total number or total weight of examples.
PiperOrigin-RevId: 173731046
---
Commit e1d7615eb authored by Alexandre Passos<apassos@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Fix issue with gradients of functions which return multiple values.
PiperOrigin-RevId: 173730922
---
Commit 80374a7b4 authored by Joshua V. Dillon<jvdillon@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Breaking change: Rename `tf.contrib.distributions.Independent` parameter from
`reduce_batch_ndims` to `reinterpreted_batch_ndims`. Also change default;
`reinterpreted_batch_ndims` default has semantics of `tf.layers.flatten`, i.e.,
all batch dimensions except the first (batch axis 0) are interpretted as being
part of the event.
PiperOrigin-RevId: 173729585
---
Commit 5426a3c93 authored by Allen Lavoie<allenl@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add tfe.get_optimizer_variables for fetching a list of variables which an
optimizer has created. Useful for saving them if executing eagerly.
PiperOrigin-RevId: 173726859
---
Commit 02f55400f authored by Alexandre Passos<apassos@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
custom_gradient functions should be able to return their inputs
PiperOrigin-RevId: 173723462
---
Commit 78bac7290 authored by Shanqing Cai<cais@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
TFE: Add compatbility doc string to add_to_collection() and friends
PiperOrigin-RevId: 173716912
---
Commit 9bf00c371 authored by Alexandre Passos<apassos@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Shorter import for tfe.
PiperOrigin-RevId: 173716375
---
Commit 0bc432a44 authored by Shanqing Cai<cais@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
TFE: Add compatibility errors and doc strings to queues, input pipelines and Supervisor
PiperOrigin-RevId: 173712330
---
Commit e9af1af4f authored by Amit Patankar<amitpatankar@google.com>
Committed by Amit Patankar<amitpatankar@google.com>:
Fixing the sources docs in master.
---
Commit b31b08bb0 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Adds randomized tests for newly introduced complex and related ops.
PiperOrigin-RevId: 173709206
---
Commit 466b9ecf8 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Report total number of bytes to be transferred when the curl request makes no progress.
PiperOrigin-RevId: 173707608
---
Commit 7c4e98eb4 authored by Igor Ganichev<iga@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add Tensor._rank() getter
It appears to speed up SPINN model by about 1%, which is not much, but
this method is very simple and easier to use than len(tensor._shape_tuple())
PiperOrigin-RevId: 173703259
---
Commit d7cffe9c0 authored by Allen Lavoie<allenl@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Adds save and restore methods to tfe.Network
Save just saves the variables to a checkpoint. Restore either restores immediately or defers the restoration to variable creation time with a custom getter.
PiperOrigin-RevId: 173703075
---
Commit 9158f974a authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Use tf.app.run in gcs_smoke, so that the flags are explicitly parsed, instead of parsed when first accessed.
PiperOrigin-RevId: 173702828
---
Commit 3d39b32b9 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Fix a tfprof bug. Throws an error when the flops cannot be calculated.
PiperOrigin-RevId: 173702740
---
Commit 73155f56a authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[TF:XLA] Small code cleanup. Re-alphabetized.
PiperOrigin-RevId: 173702336
---
Commit 32bcf46f1 authored by Mustafa Ispir<ispir@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
internal
PiperOrigin-RevId: 173697389
---
Commit 97484a4d9 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Update ops-related pbtxt files.
PiperOrigin-RevId: 173690751
---
Commit 873ef2ca3 authored by Oleg Zabluda<ozabluda@gmail.com>
Committed by GitHub<noreply@github.com>:
Fix documentation error in tf.size() - output type
---
Commit 16538dab7 authored by Alexandre Passos<apassos@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Saves summaries in the mnist example.
PiperOrigin-RevId: 173690505
---
Commit 6b05b36cd authored by Jiri Simsa<jsimsa@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Generalizing sloppy_interleave, making sloppiness an option.
PiperOrigin-RevId: 173687797
---
Commit 7775a6604 authored by Michael Case<mikecase@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Internal Change
PiperOrigin-RevId: 173685895
---
Commit 5120e75cf authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Yong Tang<yong.tang.github@outlook.com>:
Move `@compatibility(eager)` from class docstring to __init__ docstring
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit 7d7b2ec58 authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Yong Tang<yong.tang.github@outlook.com>:
Also fixes `@end_compatiblity` -> `@end_compatibility`
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit 96dc501cd authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Yong Tang<yong.tang.github@outlook.com>:
Fix incorrect annotation tag in tf.Variable
In tf.Variable the annotation tag of `@compatiblity` should be `@compatibility`
---
Commit c22973867 authored by Mark Daoust<markdaoust@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Delete bad links (md links not supported in html blocks).
PiperOrigin-RevId: 173680417
---
Commit 4198e27be authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
[XLA:CPU] [XLA:GPU] Adds compiler support for C64 primitive type, including relevant elementwise unary and binary op lowering for CPU and GPU.
We use a named LLVM struct "complex64", laid out the same as std::complex<float>. This named struct is accessed via the llvm::Module, which required changes to accessors of PrimitiveTypeToIrType & friends.
Ops that require atan2 (in particular, angle and log) are only supported on GPU at this point. LLVM lacks a CPU intrinsic for atan or atan2, whereas libdevice provides this for GPU.
PiperOrigin-RevId: 173676849
---
Commit 4ae245a7d authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
n/a (internal change only)
PiperOrigin-RevId: 173674697
---
Commit 0ccf5cf60 authored by A. Unique TensorFlower<gardener@tensorflow.org>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Limit the amount of logspam a use of GraphKeys.VARIABLES causes.
Multiple copies of this warning next to each other often make logs unreadable.
PiperOrigin-RevId: 173672701
---
Commit a7b872527 authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Yong Tang<yong.tang.github@outlook.com>:
Fix an ouput typo in `ci_sanity.sh`
In the last PR #13924 (clang sanity check) the output message should be changed:
`due to the absence of Python code changes`
->
`due to the absence of .h or .cc code changes`
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit 58d2c5f50 authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Shanqing Cai<cais@google.com>:
Add `SANITY_STEPS_DESC` for do_clang_format_check (#14030)
* Add `SANITY_STEPS_DESC` for do_clang_format_check
This fix is a follow up to PR #13924 to add the corresponding
description in `SANITY_STEPS_DESC`.
See comment #13924#discussion_r147314599
for details.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Update description for Clang Format Check
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit 62a9ab28c authored by ???<dev@goodow.com>
Committed by GitHub<noreply@github.com>:
fix broken link
---
Commit c6292a3f9 authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Yong Tang<yong.tang.github@outlook.com>:
Sanitize decode_csv_op.cc with `clang-format -i`
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit 285ea3910 authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Yong Tang<yong.tang.github@outlook.com>:
Add test cases for `double` support of `tf.decode_csv`
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit 73aaed655 authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Yong Tang<yong.tang.github@outlook.com>:
Update docs for `double` support on `tf.decode_csv`
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit 3595d1613 authored by Yong Tang<yong.tang.github@outlook.com>
Committed by Yong Tang<yong.tang.github@outlook.com>:
Add `double` support for `tf.decode_csv`
In the current tensorflow `tf.decode_csv` accepts
`float`, `int32`, `int64`, `string` but not `double`.
It seems adding `double` support makes sense as `StringToNumber`
already support `double` type.
This fix adds `double` support for `tf.decode_csv`
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
---
Commit 37d483fda authored by Sergii Khomenko<sergii.khomenko@stylight.com>
Committed by Sergii Khomenko<sergii.khomenko@stylight.com>:
Fix a typo
---
Commit 9c8a520b0 authored by Justine Tunney<jart@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
Add WriteEvent method to SummaryWriterInterface
Another change will follow that adds an op for this method. It will be useful
for loading event logs into other types of summary writer implementations, like
a database.
This change might also make the new summary file writer go faster, due to less
memory copying.
PiperOrigin-RevId: 173640116
---
Commit a49455812 authored by Eugene Brevdo<ebrevdo@google.com>
Committed by TensorFlower Gardener<gardener@tensorflow.org>:
BEGIN_PUBLIC
Automated g4 rollback of changelist 172654120
PiperOrigin-RevId: 174388998
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| f97e7c69b84dac8c3c8c78204d48816036b9bead |
|
01-Nov-2017 |
Olivia Nordquist <nolivia@google.com> |
partially exposing the _set_attr and _get_attr method in python
PiperOrigin-RevId: 174113043
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| d719036e9f43cb878abaa1bf6f9bf651522f1394 |
|
28-Sep-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Fetch Operation.inputs from the C API
This is tested by a number of existing C API-enabled tests,
e.g. framework/ops_test.py. I also added some checks to
testUpdateInput() which would fail without thie change (since
Operation._update_input() does not update the Python input if the C
API is enabled).
PiperOrigin-RevId: 170284504
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 4b9e50686928b858a0045e1ddcc0bab4acc7ff37 |
|
08-Sep-2017 |
Allen Lavoie <allenl@google.com> |
Fix C API Python wrapping for TF_DeviceListMemoryBytes and TF_Dim
It was returning Swig wrapped pointers, since Swig thinks int64_t is a complex
type.
PiperOrigin-RevId: 168005582
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 9624d165f1f2c717eda96464fee8bf7229cc14f5 |
|
31-Aug-2017 |
Igor Ganichev <iga@google.com> |
Add function support to Tensorflow C API
This change adds minimal functionality. Support for FunctionOptions,
attributes, output name rewriting, function name generation, etc is
comming next.
PiperOrigin-RevId: 167091238
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 13eb3b90e9ed8778ffd2b1bf6401677938b1ec39 |
|
10-Aug-2017 |
Alexandre Passos <apassos@google.com> |
Experimental C and Python APIs to invoke TensorFlow kernels on concrete values.
PiperOrigin-RevId: 164902588
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 09b9ce84fa7c7741eeb3b9203a802fd27492ad34 |
|
02-Aug-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Use SWIG_exception_fail instead of SWIG_fail
SWIG_fail results in a mysterious
"SystemError: error return without exception set"
error message without any stack trace or other helpful info.
SWIG_exception_fail provides much more context shortening
debug time.
PiperOrigin-RevId: 164036157
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| ce952bde5ea34708160f3e97d039617a0c11bd5f |
|
17-Jul-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Create ScopedTFStatus to manage TF_Status lifetime.
In addition, this change:
- Adds a SWIG typemap to automatically unwrap ScopedTFStatus's for TF_Status* inputs
- Creates a new c_api_util.py file
- Switches a few TF_Status users to use raise_exception_on_not_ok_status() instead
PiperOrigin-RevId: 162243205
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 45a58d378ea081907d1aa7d8963c40e7fb3dcfe8 |
|
12-Jul-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Introduce Python-only extensions to the C API
Implements an incomplete version of Operation._add_control_input()
using a new extension to make sure the plumbing works.
This also adds header guards to c_api_internal.h, which were missing. For some reason the missing guards caused problems in the cmake build even though there doesn't appear to be any #include cycles.
PiperOrigin-RevId: 161705859
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 0bfccd3994d7ee9cfee4dd1e337047ab6abbe25c |
|
11-Jul-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Use C API to implement Operation.type
The code is excercised by many tests in ops_test.py, math_ops_test*.py,
session_*test.py.
PiperOrigin-RevId: 161460668
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| cab048ecde3c3567b188a93f62d51af3b6b5b078 |
|
06-Jul-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Use C API to implement Operation.control_inputs
Operation.control_inputs must return Python Operation instances,
but C API returns TF_Operation*s. After retrieving TF_Operation*s,
we get the Python Operation object from the graph using TF_Operation's
name.
There is currently just one test for this functionality, namely
ControlDependenciesTest.testBasic. We can't enable other tests
because C API does not accept unregistered operations.
In following changes we will deal with this problem thereby enabling
more tests.
PiperOrigin-RevId: 161106091
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 980d3f2be66e62f9240d9154add9f14697957714 |
|
21-Jun-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Use C API to implement Operation.name property
This name property is used in many existing tests including those that
already run with C API enabled (math_ops_test, framework_ops_test,
session_test, session_partial_run_test, math_ops_test_gpu, etc).
PiperOrigin-RevId: 159645767
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 348812b4a56ca460d2006abf1032f9ad9c86a084 |
|
17-Jun-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Use new C API for Session.partial_run and its setup
PiperOrigin-RevId: 159298135
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 813be3fb0e42cdf1efa5ea37aa4aafbf520c0b98 |
|
27-May-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Optionally use new session API in Python API
The new session API is enabled if ops._USE_C_API is True, otherwise the original behavior is maintained (the default). Most of the tests in session_test.py are now run with the new API. The new session API runs the TF_Graph that is built alongside the Python graph when ops._USE_C_API is True.
PiperOrigin-RevId: 157272139
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 52c1afbb5490e13f2a674a79b81b947ade8ea437 |
|
18-May-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Light refactoring of tf_session_helper.h/cc
A few changes to make these helper functions easier to understand and in preparation for adding TF_SessionRun() functionality:
- move all typedefs to .h
- get rid of simple std::vector<...> typedefs
- rename TF_Tensor_to_PyObject() to TFTensorToPyArray()
- pull out TF_Run_wrapper_helper() functionality into new PyArrayToTFTensor() function
PiperOrigin-RevId: 156480714
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| b9cdcbe3a019f09eee7e7c9c9039e4533931e1e7 |
|
10-May-2017 |
Skye Wanderman-Milne <skyewm@google.com> |
Create a TF_Graph alongside the Python graph.
This is a first step towards porting the Python API to use the C API. As the Python Graph and Operations are constructed, an analogous TF_Graph and TF_Operations are created via SWIG. Currently nothing is done with the TF_Graph; a next step will be switching to the new TF_Session API which runs a TF_Graph directly (instead of a GraphDef).
This new functionality is disabled by default and can be manually enabled by setting the _USE_C_API global in ops.py. For this patch I only enabled it for a single test file. I tried enabling it for all TF Python tests and manually disabling it for unsupported tests, but there were too many failing tests (although most tests passed). See ops.py for a TODO list of unsupported functionality.
I benchmarked building an Inception model, and building the TF_Graph incurs a 20% overhead to the total graph construction time. Note that this patch does not remove any existing Python functionality; another next step will be recovering this time by removing redundant Python code. There is no measurable overhead with the new functionality disabled.
PiperOrigin-RevId: 155655064
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 624f0d402383ba4476e6131cbe1d421e3710c1bd |
|
13-Apr-2017 |
Shanqing Cai <cais@google.com> |
Improve efficiency of get_session_handle
Change get_session_handle() back to use v1 of GetSessionHandleOp, instead of v2. The v1 op returns a string, instead of an encoded ResourceHandle, which gets rid of the need to call ParseFromString() in Python.
Change: 153101229
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| c73c2eef3d565e4a84c6df7038113491bc0ce616 |
|
24-Mar-2017 |
Asim Shankar <ashankar@google.com> |
Delete C library handle when done to avoid leaking memory.
Change: 151074297
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| fdd94f24e9ebfe58bd967a8c2c7b8985c03ff80b |
|
08-Nov-2016 |
Asim Shankar <ashankar@google.com> |
C API: Rename TF_Session to TF_DeprecatedSession.
This is one step towards having a stable C API by the time we get to TensorFlow
1.0. A follow-up step will involve renaming TF_SessionWithGraph to TF_Session.
We want to encourage all client languages to use TF_SessionRun,
TF_GraphImportGraphDef etc. (instead of TF_Run and TF_ExtendGraph), hence
the choice of names. Ideally, all client languages will use these
functions, but it is likely that Python will continue to use
TF_DeprecatedSession for a while.
Change: 138454433
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| f56f563a92b4804f5ffcd502cd787b3a2011f3cc |
|
05-Oct-2016 |
Zongheng Yang <zongheng@google.com> |
Move the SWIG typemap for protobufs to platform/base.i.
Change: 135234473
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 9814f5efc297692ef04309d94eb582d19bf2aef0 |
|
18-Sep-2016 |
Derek Murray <mrry@google.com> |
SWIG simplification: Move the FeedVector conversion into C++.
This change moves the %typemap(in) FeedVector into C++ code in
tf_session_helper.cc, which now implements the conversion from a
string-to-array dictionary into the arguments to `TF_Run()`.
Change: 133494142
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 10cb80be1d91227f794cb8c7bf4400b05f2d7326 |
|
17-Sep-2016 |
Derek Murray <mrry@google.com> |
Remove unused dependency on numpy.i.
Change: 133439235
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 09045e49d1e2d3322e8ca8cc40f0e5db10f13c58 |
|
06-Sep-2016 |
Andrew Selle <aselle@google.com> |
Include sha hash in python variable __version__
It is necessary to symlink in files from .git/ in order to make
bazel aware of changes to the current head. As it is this is not
completely reliable when git repositories are in a dirty index
state. First class support for bazel git a reported bug but
not a high priority.
./configure sets up the symlinks by calling the gen_git_source.py
a bazel genrule calls gen_git_source.py to generate version_info.cc
Also changed cmake and make to build this properly.
Change: 132328009
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 788f359b7218ad46696c15459c89688ffe70955e |
|
30-Jul-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Move C API files (and related files used by SWIG wrappers)
to new tensorflow/c directory.
Change: 128855990
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 6c14cd8e1bfbf1484ef22c82ac3badc53ed73f7d |
|
27-Jun-2016 |
Derek Murray <mrry@google.com> |
Validate the types of arguments to `tf.Session.__init__()`.
Fixes #3035.
Change: 125992640
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 7083e2af69f5692d74fc6b1148c0ecea0c562274 |
|
24-Jun-2016 |
Sherry Moore <sherrym@google.com> |
Added tf.container(container_name) context manager.
Added Python interfaces to reset resource containers.
pywrap_tensorflow.TF_Reset(target, [containers]) will release resources
in all listed containers, while pywrap_tensorflow.TF_Reset(target) will
release resources in all the containers. In particular, Variables cached
on the devices will be cleared by this call.
Change: 125790643
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 0cf9ed3a719c0782695154d5a0bca260001cec15 |
|
02-Jun-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Update copyright for 3p/tf/python.
Change: 123900456
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 94d27b6852b3e331fd9d64a0533f0fc27af05bfd |
|
06-May-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Next pass at converting all return statuses to errors.OpError exceptions
in Python.
- Add a tf_status util file to convert between Status and TF_Status.
- Use TF_Status for the swigged APIs in session, checkpoint_reader, and server_lib.
- Converts all these tf_status to exceptions in Python with the new context handler.
- Remove the old StatusNotOK exception.
Change: 121644982
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| b80a4a8732dfcbbeb348b9f8470aa420bab3931b |
|
04-May-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Raise proper tensorflow exceptions in a few more places:
- Add a context manager to deal with C apis that want a status object passed
as an extra argument.
- Use the context manager in session.py and session.i
- Update tests.
Change: 121430994
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| ce9bddfc5ab1420a21140f259d37a4b4bff6a7b5 |
|
14-Apr-2016 |
Derek Murray <mrry@google.com> |
Fix ASAN breakage due to delete/delete[] confusiong in PRun SWIG wrapper.
Change: 119861147
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 37606a4c63364c56a0834d281023b62d2bda6cd8 |
|
23-Mar-2016 |
Vijay Vasudevan <vrv@google.com> |
Merge changes from github, some fixes to adhere somewhat
to our requirements for skflow.
Change: 117901053
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| b27022dd3ccb0b415cef3505ee0ee65f25b2ffcb |
|
15-Mar-2016 |
Manjunath Kudlur <keveman@gmail.com> |
Fix GetOpList and GetPythonWrappers SWIG wrappers for Python 3.
- Return the result from GetOpList as uninterpreted bytes object.
- Write a input typemap for GetPythonWrappers to receive python 'bytes'
object and convert to const char* pointer and length.
Change: 117258253
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| af4e6fc72586239d73402564042d5ad4aa640d22 |
|
15-Mar-2016 |
Derek Murray <mrry@google.com> |
Fix //tensorflow/python:server_lib_test in Python 3.
Adds a typemap for the `const char* target` argument to
`TF_SetTarget()` that accepts any object that may be converted to a
null-terminated string.
Change: 117241091
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| b54ad5788b7cec515d81e9f7e3c9811fd8181235 |
|
14-Mar-2016 |
Zongheng Yang <zongheng.y@gmail.com> |
Expose RunOptions and RunOutputs to Python's Run().
The Run() call that takes these protos previously only worked in C++.
Example usage in Python (see also new unit test in session_test.py):
sess.run(constant_op.constant(1.0),
options=run_options,
run_outputs=run_outputs)
A thin TF_Buffer struct is introduced to the C API, which takes care of
handling in/out bytes. For instance, the "in protobuf" RunOptions is
serialized to bytes, which then get typemap'd to TF_Buffer. The "out protobuf"
RunOutputs is handled analogously.
Change: 117152952
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 1442f04dfbd715ead1d32b6395f5346e5256578f |
|
04-Mar-2016 |
Sherry Moore <sherrym@google.com> |
Moved config.proto from core/framework to core/protobuf.
Change: 116396958
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 5ff6d34a05b2eb49f7a79a4d0b78ada9a6842b6c |
|
02-Feb-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Fix a Python 3 incompatibility in the SWIG bindings.
Change: 113579944
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 8a59748c087a2fee535c0d5067dbabb01920e812 |
|
29-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Use cc_binary rather than cc_library to reduce size of native library in APK from 5.5mb to 3.2mb (compressed).
Change: 113369407
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| a18d48f80f118b28d351001a9b4c47613e228f34 |
|
26-Jan-2016 |
Geoffrey Irving <geoffreyi@google.com> |
Change GraphDef versions to use version, min_consumer, min_producer
Since all GraphDef versions to date are forwards compatible, we make no attempt
to merge old-style and new-style GraphDef version information. The old int32
version field is ignored (and deprecated) and the new VersionDef versions field
defaults to producer 0, min_consumer 0 if left out.
As a benefit, once this CL is in we can immediately bump the scalar strictness
GraphDef version, since consumers will either ignore the new versions (before
this CL) or be forward compatible (after this CL).
Later, I'll use the same mechanism and protobuf to add versioning to checkpoints.
Change: 113091281
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 2ab8e07fb717c0d1a6dd20b41d0e84986e961335 |
|
25-Jan-2016 |
Geoffrey Irving <geoffreyi@google.com> |
Make EqualGraphDef ignore versions and expose to Python
The intention is for EqualGraphDef to be the only way downstream tests compare
GraphDefs in golden file tests.
Change: 112967255
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 9b70316263eb74476ab96b7c0f300c4d90223425 |
|
25-Jan-2016 |
Vijay Vasudevan <vrv@google.com> |
Running our linter on a lot of files.
Change: 112920860
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| db7478e8998f7703c57a75a950c905ec0cb59d7b |
|
20-Jan-2016 |
Josh Levenberg <josh11b@tensorflow.org> |
First step of moving files out of tensorflow/core/public/. Here
we copy the original files to their new location and make the public/
versions #include the new location. Once all references are updated
to point to the new location, we can delete the originals in public/.
Change: 112622561
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 1c579361cd1e088dd5e05a394b1561a73e3667ba |
|
05-Jan-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Added 'logging' import to control_flow_ops which is used in the file but not imported.
Change: 110842260
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 54a644f33f34a32fa2cb5e7a489c64540b16e166 |
|
04-Dec-2015 |
Vijay Vasudevan <vrv@google.com> |
TensorFlow: upstream changes to git
Change 109366961
TensorFlow BUILD: now that we have an ops library,
set linkstatic to 1. This fixes a breakage in the would-be
opensource build, and it *might* mean we can get rid of
all of the RequireDefaultOps() calls in our code.
The ops library is much smaller than the kernels library that was
previously linked together. We set linkstatic=0 presumably since we
didn't want to package a static copy of the kernels (very large)
everywhere. But the op definitions are small, so this seems like a
safe change to make. Time to build the various tests was not
any longer after this change, and inspecting the example_trainer
binary showed no large increase.
Change 109363613
TensorFlow: new graph_def_builder_test needs to RequireDefaultOps.
Change 109362569
Split ":ops" out of ":kernels" target in tensorflow/core.
Change 109360666
Catch dtype and some shape errors sooner in `QueueBase`.
Some avoidable errors were not being caught (e.g. the dtypes of the
enqueue components were not checked against the queue's dtypes in
Python), leading to cryptic messages at runtime. After this CL, they
will be caught earlier.
Change 109359569
TensorFlow: Expect g_ != nullptr in test
Change 109350735
Add a version number to GraphDef
We would like to be able to deprecate behavior in newly generated graphs
without invalidating tensorflow's ability to read and evaluate old graphs.
For this purpose, GraphDef now has a version field which can be checked inside
op kernels to determine how backwards compatible to be. version.h defines
TF_GRAPHDEF_VERSION_MIN and TF_GRAPHDEF_VERSION_MAX specifying the range of
supported GraphDef versions in the current version of tensorflow.
Also expose tf.__version__ and tf.__graph_def_version{,_min,_max}__ for Python
interrogation purposes.
Whenever we want to deprecate or change some GraphDef semantics, we will
proceed as follows:
1. Bump TF_GRAPHDEF_VERSION_MAX, leaving TF_GRAPHDEF_VERSION_MIN unchanged.
Describe the change in graph.proto, include the date introduced.
2. In each relevant kernel, implement the new behavior if the GraphDef version
is new, but preserve the old behavior for previous GraphDef versions.
3. Wait six months or so (we need to formalize this somewhere).
4. Bump TF_GRAPHDEF_VERSION_MIN and remove the backwards compatibility.
The GraphDef version is distinct from the open source version, but at least
(4) and possibly (1) correspond to major version number bumps.
The first GraphDef version bump is the upcoming scalar strictness change,
which affects Google users only since open source is already scalar strict.
This commit does not yet plumb the version number into OpKernelConstruction
so that ops can access it. That will follow.
Change 109350260
Made TensorShapeProto implicitly convertible to TensorShape.
Base CL: 109366982
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 854f49bd43588c062b046384f239f64a3d819702 |
|
25-Nov-2015 |
Manjunath Kudlur <keveman@gmail.com> |
TensorFlow: Upstream changes to git
Changes:
- Updates to docs
- Several changes for Python 3 compatibility
- Added license headers
Base CL: 108710566
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 4213ac97be449d0e40631a314d2b7bd3901d4967 |
|
17-Nov-2015 |
Vijay Vasudevan <vrv@google.com> |
TensorFlow: conv improvements, label_image example, and
a few other changes.
Changes:
- Some improvements to convolution by using 32-bit indices by
@benoitsteiner. Not all calls converted yet. Also some
improvements to pooling as well by @benoitsteiner.
- Improvements to sparse matmul CPU implementation by Ashish
- Some fixes to warnings by @vrv
- Doc fixes to padding by @Yangqing
- Some improvements to Tensor wrappers by Eider
- Speed up of matrix inverse on CPU by Rasmus
- Add an example of doing image inference from a pre-trained model
by @petewarden.
- fixed formula in mnist example by nodir
- Updates to event accumulator by Cassandra
- Slight changes to tensor c api by @mrry
- Handling of strings in listdiff by Phil
- Fix negative fraction-of-queue-full stats by Frank
- Type-checking improvement to importer by Yaroslav
- logdir recursive search for Tensorboard by @danmane
- Session.run() checks for empty graph by Manoj
Base CL: 108013706
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| 011e9baccd343eb943d25014c4e8aec53eac396b |
|
13-Nov-2015 |
Vijay Vasudevan <vrv@google.com> |
TensorFlow: a few small updates.
Changes:
- Fix softmax formula in word2vec to remove an extra exp()
by @gouwsmeister
- Python3 fixes to remove basestring / support for unicode by @mrry
- Remove some comments by Josh
- Specify exact versions of bower dependencies for TensorBoard by
@danmane.
Base CL: 107742361
/external/tensorflow/tensorflow/python/client/tf_session.i
|
| f41959ccb2d9d4c722fe8fc3351401d53bcf4900 |
|
07-Nov-2015 |
Manjunath Kudlur <keveman@gmail.com> |
TensorFlow: Initial commit of TensorFlow library.
TensorFlow is an open source software library for numerical computation
using data flow graphs.
Base CL: 107276108
/external/tensorflow/tensorflow/python/client/tf_session.i
|