gracehoney [Sun, 11 Feb 2018 21:01:30 +0000 (13:01 -0800)]
Fix segment_test and run clang-format on trt_conversion.i.
gracehoney [Sun, 11 Feb 2018 06:44:01 +0000 (22:44 -0800)]
Fix python lint errors internally. One important change is to rename CreateInferenceGraph to create_inference_graph.
gracehoney [Fri, 9 Feb 2018 17:47:58 +0000 (09:47 -0800)]
Fix LINT.IfChange comment
gracehoney [Fri, 9 Feb 2018 16:46:15 +0000 (08:46 -0800)]
Fix more build dependencies and includes.
gracehoney [Fri, 9 Feb 2018 15:53:43 +0000 (07:53 -0800)]
Make build happy by adding missing OP_REQUIRES_OK and if_tensorrt
Sami Kama [Fri, 9 Feb 2018 00:44:57 +0000 (16:44 -0800)]
Merge branch 'master' into tensorrt
Jie [Thu, 8 Feb 2018 21:12:31 +0000 (13:12 -0800)]
[cleanup]
remove TRT_ShapedWeights owned_values_
fallback to converter temporary memory allocation to avoid redundant memcpy
Jonathan Hseu [Thu, 8 Feb 2018 20:32:40 +0000 (12:32 -0800)]
Merge pull request #16874 from jhseu/revert
Revert "Fix missing . (#16460)"
Jonathan Hseu [Thu, 8 Feb 2018 20:30:53 +0000 (12:30 -0800)]
Revert "Fix missing . (#16460)"
This reverts commit
5e23338fec26f0c5ad588742b8df80e5bf1a940d.
matthieudelaro [Thu, 8 Feb 2018 20:17:37 +0000 (21:17 +0100)]
Fix missing . (#16460)
Johnson145 [Thu, 8 Feb 2018 20:16:52 +0000 (21:16 +0100)]
fix typo (#16812)
Jerome [Thu, 8 Feb 2018 20:16:38 +0000 (04:16 +0800)]
Move some ndlstm functions to contrib (#16816)
* Added ctc_loss_dense_labels. This does the conversion of dense labels into sparse ones to be passed into the core ctc_loss function.
* Removed constant_op from the import.
* Matched ctc_loss_dense_labels with the other layers ops.
* Added ctc_loss_dense_labels to contrib.layers __init__.py file
* Added missing comma to list of ops.
* Reordred arguments for ctc_loss_dense_labels
Labels should be first then inputs for ctc_loss.
* Removed ctc_loss_dense_labels.
Replaced it with dense_to_sparse instead so that there'll be only one ctc_loss function.
* Replaced ctc_loss_dense_labels with dense_to_sparse
* Fixed dense_to_sparse. Some of the names of the variables did not match with that of the parameters.
* Updated documentation for dense_to_sparse since it can accept a tensor of any shape.
* Added test case for dense_to_sparse.
* Updated documentation. Dense to sparse accepts int tensors.
* Fixed testDenseFromConstantToSparse.
The sparse_to_dense order of arguments in the test are wrong and the expected constant should be of int64.
* Modified implementation of ndlstm_base_dynamic.
It now uses a BasicLSTMCell that has state_is_tuple=True to address deprecation. Right now it is still unknown why it was set to false in the first place.
* Imported lstm1d and lstm2d in ndlstm __init__.py.
Makes importing ndlstm modules easier.
* Added testGetBlocks in lstm2d_test.
* Removed testGetBlocks.py
* Modified lstm1d.ndlstm_base_unrolled to use lstm_cell with state_is_tuple = True.
* Copied some lstm2d.py functions in ndlstm module to contrib.layers.
* Update lstm1d.py
Reverted changes made.
* Update layers_test.py
Fixed failing test.
* Modified layers.py and layers_test.py.
Made them pass pylint tests.
Liang-Chi Hsieh [Thu, 8 Feb 2018 19:32:10 +0000 (04:32 +0900)]
Fix two links. (#16854)
Jie [Thu, 8 Feb 2018 19:16:08 +0000 (11:16 -0800)]
[clang-format]
Jie [Thu, 8 Feb 2018 19:09:53 +0000 (11:09 -0800)]
[update]
Avoid access of data to comply to internal build
Haichen "HC" Li [Thu, 8 Feb 2018 18:55:47 +0000 (13:55 -0500)]
remove keep_dims warning in maxout layer (#16769)
lazypanda1 [Thu, 8 Feb 2018 18:54:57 +0000 (12:54 -0600)]
Fixing assert message for beta distribution (#16786)
fo40225 [Thu, 8 Feb 2018 18:54:31 +0000 (02:54 +0800)]
python 2.7 unit test error repair on windows (#16725)
* python 2.7 unit test error repair on windows
* Fix lint error.
Ben [Thu, 8 Feb 2018 18:49:23 +0000 (13:49 -0500)]
Add option to not include histograms (#16579)
* Add option to not include histograms
* Add test for model_summaries=False
Phil [Thu, 8 Feb 2018 18:48:42 +0000 (19:48 +0100)]
Adds parameter 'msg' to tf.TensorFlowTestCase. (#16667)
This commit adds a msg parameter that defaults to None to the following
functions:
- assertProtoEquals
- assertArrayNear
- assertNDArrayNear
- assertAllClose
- assertAllEqual
- assertShapeEqual
- assertDeviceEqual
Rajendra arora [Thu, 8 Feb 2018 18:48:19 +0000 (00:18 +0530)]
Fixes variable name (#16797)
* Fixes variable typo name
* fixes optional typo
* Fix typo
* fixes deallocations
* typo fix
* typo fix
* fixes variable name
* fixes param typo
* Fixes typo
* fix typo
* fixes
Brad Wannow [Thu, 8 Feb 2018 18:47:34 +0000 (12:47 -0600)]
Update CONTRIBUTING.md (#16806)
* Update CONTRIBUTING.md
Edited a few grammar issues.
* Fix
Glenn Weidner [Thu, 8 Feb 2018 18:46:12 +0000 (10:46 -0800)]
Fix error message in record_reader (#16808)
Gunhan Gulsoy [Thu, 8 Feb 2018 18:45:37 +0000 (10:45 -0800)]
Make configure script more lenient to the length of CUDA and cuDNN versions entered. (#16853)
Yong Tang [Thu, 8 Feb 2018 18:41:03 +0000 (10:41 -0800)]
Improve shape function of SampleDistortedBoundingBox and fix some test cases (#16870)
* Update test case to expose the issue of sample_distorted_bounding_box
This commit expoes the issue of sample_distorted_bounding_box,
as shape function does not check the ranking like inside Compute().
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Adjust shape to pass the test.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Improve shape function of SampleDistortedBoundingBox and fix some test cases
This fix tries to improve the shape function of SampleDistortedBoundingBox
and fix several test case issues.
As is shown in the kernel of SampleDistortedBoundingBox, the shape of
SampleDistortedBoundingBox are required to be 1-D `[height, width, channels]`
for image_size, 3-D with shape `[batch, N, 4]` for bounding_boxes.
In the test case, the uses shape is incorrect but becasue of no check in shape
function, the test case passes.
This fix adds the shape check for SampleDistortedBoundingBox, and
fixes the incorrect test cases.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
Yong Tang [Thu, 8 Feb 2018 18:35:08 +0000 (10:35 -0800)]
Fix broken link in CONTRIBUTING.md (#16869)
This fix fixes the broken link in CONTRIBUTING.md.
Without `https://`, the markdown will render the link incorrectly to
https://github.com/tensorflow/tensorflow/blob/master/www.docker.com
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
Yan Facai (颜发才) [Thu, 8 Feb 2018 16:59:08 +0000 (00:59 +0800)]
Fix static shape inference for keras.layers.LSTM (#15234)
* TST: add test case
* BUG: fix static shape inference
* TST: clean code
* BUG: support dims > 3
Scott Tseng [Thu, 8 Feb 2018 16:58:36 +0000 (00:58 +0800)]
Enable some passes for graph_transform on Windows (#16121)
* Enable some passes for graph_transform on Windows
Don't know why but the following passes are disabled on Windows:
* quantize_weights
* quantize_nodes
* round_weights
This patch re-enabled them.
* Fix BUILD file in response to format checker
Yan Facai (颜发才) [Thu, 8 Feb 2018 16:54:08 +0000 (00:54 +0800)]
py_func convert unicode string results to bytes for python2 (#16322)
* TST: add test case
* BUG: allow unicode string for python 2
* DOC: revise doc
* Fix lint error
Jie [Thu, 8 Feb 2018 05:31:25 +0000 (21:31 -0800)]
[debug]
convert const data goes out of scope.
Yong Tang [Thu, 8 Feb 2018 00:48:21 +0000 (16:48 -0800)]
Add optimized gif support for decode_gif (#16804)
* Add optimized gif support for decode_gif
While revisiting the issue of 15838, I noticed that
currently optimized gif is not supported. However,
optimized gif is actually possible to be processed
as essentially the subsequent frame just adds
the content on top of the previous frame on canvas.
This fix adds the support for optimized gif with decode_gif.
As is shown in the added test case, optimized gif (`optimized.gif`)
could be handled the same way as original gif (`scan.gif`).
This fix fixes 15838.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Format gif_io.cc with clang-format -i --style=Google
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Add test case to cover optimized gif support for decode_gif.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Add `#include <algorithm>` to fix Windows build errors
This commit add `#include <algorithm>` to fix
build errors on Windows platform.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
Yong Tang [Thu, 8 Feb 2018 00:48:00 +0000 (16:48 -0800)]
Add uint32 and uint64 kernel support for `Invert` (#15154)
* Add uint32 and uint64 kernel support for `Invert`
This fix adds uint32 and uint64 kernel support for `Invert`.
In bitwise_ops.cc, uint32 and uint64 have been registered
for `Invert` like other bitwise ops
`BitwiseAnd`/`BitwiseOr`/`BitwiseXor`/`LeftShift`/`RightShift`.
However, no uint32 and uint64 kernels available for `Invert` yet.
This fix add uint32 and uint64 kernel for `Invert`,
and adds additional test cases to cover the changes.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Add test cases for uint32 and uint64 support with `Invert`
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Add missing uint32 and uint64 in GPU for invert
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Add DEFINE_UNARY8 for invert
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
Yong Tang [Thu, 8 Feb 2018 00:46:31 +0000 (16:46 -0800)]
Support CopyFile with streaming (#12658)
* Support CopyFile with streaming
This fix tries to address the issue raised in 12641 where
it was not possible to have CopyFile with streaming. The original
implementation copies the whole content of the file to a string
buffer and write to the file. This could be an issue
if the file size is large (than memory of the host).
This fix streams the CopyFile operation.
This fix fixes 12641.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Use sendfile for CopyFile implementation in Linux
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Merge CopyFile for same fs and different fs
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* `sendfile64` -> `sendfile` to fix Android build
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Add sendfile processing for Darwin
This commit adds sendfile processing for OSX Darwin.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Not using sendfile in MacOSX
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Address review feedback
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Remove the size check and test OUT_OF_RANGE instead.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Small fixes
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
* Rename CopyFile to FileSystemCopyFile to fix Windows build errors
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
cclauss [Thu, 8 Feb 2018 00:35:29 +0000 (01:35 +0100)]
resolve undefined name array_ops (#16485)
Yusuke Yamada [Thu, 8 Feb 2018 00:35:13 +0000 (09:35 +0900)]
Fix document typo (#16489)
cclauss [Thu, 8 Feb 2018 00:28:06 +0000 (01:28 +0100)]
Fix undefined name: import as_str_any for line 35 (#16668)
flake8 testing of https://github.com/tensorflow/tensorflow on Python 2.7.14
$ __flake8 . --count --select=E901,E999,F821,F822,F823 --show-source --statistics__
```
./tensorflow/python/util/compat_internal.py:33:12: F821 undefined name 'as_str_any'
path = as_str_any(path.__fspath__())
^
```
Sergei Lebedev [Thu, 8 Feb 2018 00:26:36 +0000 (01:26 +0100)]
Fixed a typo in `group_by_window` documentation (#16711)
Marcus Ong [Thu, 8 Feb 2018 00:25:49 +0000 (18:25 -0600)]
CMake (Windows): Added support for ninja build and some fixes/changes (#16763)
* Added support for ninja build and some fixes/changes in CMake for Windows
* Fixed typo
brett koonce [Thu, 8 Feb 2018 00:21:31 +0000 (16:21 -0800)]
spelling fixes for contrib docs (#16811)
Yong Tang [Thu, 8 Feb 2018 00:21:13 +0000 (16:21 -0800)]
Remove obsolete BernoulliWithSigmoidProbs (#16846)
As was pointed out by 9485, BernoulliWithSigmoidProbs is covered
by Bernoulli and is obsolete. This fix removes BernoulliWithSigmoidProbs.
This fix closes 9485.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
hsm207 [Thu, 8 Feb 2018 00:20:15 +0000 (19:20 -0500)]
Fix typo (#16822)
Martin Wicke [Thu, 8 Feb 2018 00:14:23 +0000 (16:14 -0800)]
Merge pull request #16837 from yongtang/16534-tf.losses.softmax_cross_entropy
Remove warnings in tf.losses.softmax_cross_entropy
Jie [Wed, 7 Feb 2018 23:15:50 +0000 (15:15 -0800)]
[update]
fix naming stuff in tensorflow/contrib/tensorrt/convert/convert_nodes.cc
Sami Kama [Wed, 7 Feb 2018 22:54:49 +0000 (14:54 -0800)]
Fix swig interface problems
Francisco Guerrero [Wed, 7 Feb 2018 22:19:25 +0000 (17:19 -0500)]
Fixes issue when linking of rule '//tensorflow/contrib/lite/toco:toco… (#16838)
* Fixes issue when linking of rule '//tensorflow/contrib/lite/toco:toco' fails because LD_LIBRARY_PATH is not configured
* Check if LD_LIBRARY_PATH is in environ_cp
Andrew Harp [Wed, 7 Feb 2018 22:19:11 +0000 (17:19 -0500)]
Bump JetPack default to 3.2 in Android build script (#16842)
Rasmus Munk Larsen [Wed, 7 Feb 2018 21:47:04 +0000 (13:47 -0800)]
Merge pull request #16844 from rmlarsen/master
Fix comparison in neon_depthwise_conv_op test.
ankurtaly [Wed, 7 Feb 2018 21:17:05 +0000 (13:17 -0800)]
Merge pull request #16819 from ankurtaly/branch_184768730
Branch
184768730
Rasmus Munk Larsen [Wed, 7 Feb 2018 20:56:06 +0000 (12:56 -0800)]
Fix bug and speed up Grappler constant folding
Fix bug in and speed up ConstantFolding::CreateNodeDef():
* Fix bug trying to store more than kintmax32 values in a repeated proto field.
* Speed up populating compressed format. Example: tensorflow/python/kernel_tests/large_concat_op_test with size = 2**29+6 goes from ~30 seconds to ~15 seconds. The fraction of time spent in ConstantFolding::CreateNodeDef() goes down from about 35% to about 12%.
Rasmus Munk Larsen [Wed, 7 Feb 2018 20:45:39 +0000 (12:45 -0800)]
Use assertAllClose instead of assertArrayNear.
assertArrayNear compares absolute instead of relative error. For floating point computation you want the latter.
Rasmus Munk Larsen [Wed, 7 Feb 2018 20:37:14 +0000 (12:37 -0800)]
Merge pull request #1 from tensorflow/master
merge from master
Yifei Feng [Wed, 7 Feb 2018 19:51:55 +0000 (11:51 -0800)]
Update BUILD
gracehoney [Wed, 7 Feb 2018 19:24:09 +0000 (11:24 -0800)]
Fix #includes and namespace scope, and add missing TF_RETURN_IF_ERROR, to make internal build happy.
Ankur Taly [Wed, 7 Feb 2018 19:15:47 +0000 (11:15 -0800)]
Updated copyright from 2017 to 2018
Phil [Wed, 7 Feb 2018 18:59:59 +0000 (19:59 +0100)]
Add unsortedsegment(prod/min/max/sqrt_n/mean). (#15858)
* Add unsortedsegment(prod/min/max/sqrt_n/mean).
This commit adds CPU/GPU implementations for prod/min/max
ops and python implementations for mean/sqrt_n. Also, it adapts and unifies the
corresponding tests of all unsorted reductions.
Note: The new gradient of unsorted_segment_max fixes the crash occuring when
negative indices on CPU are used.
* update golden API
* Fix compilation of atomicAdd for cuda_arch < 600. \n This commit moves the std::complex specialization of atomicAdd below the double specialization of atomicAdd for cuda_arch 600.
* Enable bfloat16, change inline to EIGEN_STRONG_INLINE.
* fix includes of cuda_device_functions; fix typo
Yong Tang [Wed, 7 Feb 2018 16:24:34 +0000 (16:24 +0000)]
Remove warnings in tf.losses.softmax_cross_entropy
This fix tries to address the issue raised in 16534 where
tf.losses.softmax_cross_entropy causes warnings due
to the calling of tf.nn.softmax_cross_entropy_with_logits.
This fix switches to tf.nn.softmax_cross_entropy_with_logits_v2
to remove the warning.
This fix fixes 16534.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
gracehoney [Wed, 7 Feb 2018 16:12:05 +0000 (08:12 -0800)]
Fix the segfault in convert_nodes.cc
cclauss [Wed, 7 Feb 2018 06:54:21 +0000 (07:54 +0100)]
Typo in variable name: BETA --> self.BETA (#16666)
__BETA__ is defined on line 118 as a class member so it can only be accessed via __self__ or via the __ElasticAverageOptimizer__.
flake8 testing of https://github.com/tensorflow/tensorflow
$ __flake8 . --count --select=E901,E999,F821,F822,F823 --show-source --statistics__
```
./tensorflow/contrib/opt/python/training/elastic_average_optimizer.py:153:27: F821 undefined name 'BETA'
self._moving_rate = BETA / communication_period / num_worker
^
```
Yong Tang [Wed, 7 Feb 2018 06:53:33 +0000 (22:53 -0800)]
Fix incorrect links in CONTRIBUTING.md (#16814)
This fix fixes two incorrect links in CONTRIBUTING.md
about license examples. The reason for broken links
is because tensorboard is in another repo.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
Ankur Taly [Wed, 7 Feb 2018 02:13:46 +0000 (18:13 -0800)]
Merge commit for internal changes
Derek Murray [Wed, 7 Feb 2018 01:37:02 +0000 (17:37 -0800)]
[tf.data] Fix a memory leak when an iterator is reinitialized many times in a session.
Previously, we would instantiate a new function handle for each
function in a dataset each time an iterator on that dataset was
initialized. These would only be deleted at session closure, which
could lead to an apparent leak of memory over the lifetime of session.
PiperOrigin-RevId:
184768730
Yuefeng Zhou [Wed, 7 Feb 2018 01:32:50 +0000 (17:32 -0800)]
Fix memory tracking in the case where temp memory is used as output memory.
Track all persistent memory allocation in allocate_persistent call except for variables and queues where persistent memory is tracked in variables ops and queue ops. Deallocation of persistent memory is ignored.
PiperOrigin-RevId:
184768231
Andrew Selle [Wed, 7 Feb 2018 01:10:31 +0000 (17:10 -0800)]
Sync the opensource and non-opensource build
PiperOrigin-RevId:
184765632
James Qin [Wed, 7 Feb 2018 00:25:59 +0000 (16:25 -0800)]
Address Adagrad/RMSProp incompatibility with CudnnRNN
CudnnRNN layers have variables of unknown shapes, which Adagrad/RMSProp didn't handle before.
This fixes 6620(#6620).
PiperOrigin-RevId:
184759579
Justin Lebar [Tue, 6 Feb 2018 23:30:40 +0000 (15:30 -0800)]
[XLA] Add and use new Literal::MakeTupleOwned overload.
Previously MakeTupleOwned was cumbersome to use, because you had to
explicitly materialize a vector<unique_ptr<Literal>>. With this new
overload, you can pass unique_ptrs directly.
PiperOrigin-RevId:
184751119
Akshay Modi [Tue, 6 Feb 2018 23:13:41 +0000 (15:13 -0800)]
Add read resource variable benchmarks.
Initial benchmarks:
entry {
name: "MicroBenchmarks.benchmark_read_variable_op_2_by_2_CPU"
iters: 30000
wall_time: 22.
9616721471
extras {
key: "examples_per_sec"
value {
double_value: 43550.8352176
}
}
}
entry {
name: "MicroBenchmarks.benchmark_read_variable_op_with_tape_2_by_2_CPU"
iters: 30000
wall_time: 27.
3616631826
extras {
key: "examples_per_sec"
value {
double_value: 36547.4859232
}
}
}
PiperOrigin-RevId:
184748548
A. Unique TensorFlower [Tue, 6 Feb 2018 23:09:51 +0000 (15:09 -0800)]
Creates tf.contrib.feature_column module.
PiperOrigin-RevId:
184747924
Alexandre Passos [Tue, 6 Feb 2018 22:43:18 +0000 (14:43 -0800)]
Handles possible infinite recursion in while loop fix.
PiperOrigin-RevId:
184743192
Eugene Brevdo [Tue, 6 Feb 2018 22:37:21 +0000 (14:37 -0800)]
[TF Ops] Bugfix to Operation initializer: error message uses node_def.
self.node_def may not yet be accessible when using the C api.
PiperOrigin-RevId:
184742074
Joshua V. Dillon [Tue, 6 Feb 2018 22:17:52 +0000 (14:17 -0800)]
Automated g4 rollback of changelist
184551259
PiperOrigin-RevId:
184738583
Anna R [Tue, 6 Feb 2018 22:05:01 +0000 (14:05 -0800)]
Export CXX11_ABI_FLAG and MONOLITHIC_BUILD constants.
PiperOrigin-RevId:
184736216
gracehoney [Tue, 6 Feb 2018 22:04:35 +0000 (14:04 -0800)]
1. Use tensorflow::string instead of std::string
2. Fix swig problem in trt_conversion.i
3. Fix c++ styles
4. Fix BUILD dependencies
Alexandre Passos [Tue, 6 Feb 2018 21:54:56 +0000 (13:54 -0800)]
Another rolling back of performance regression.
PiperOrigin-RevId:
184734426
Akshay Modi [Tue, 6 Feb 2018 21:53:56 +0000 (13:53 -0800)]
Make TFE_Py_FastpathExecute work for all types of ops
MatMul benchmarks:
entry {
name: "MicroBenchmarks.benchmark_gen_math_ops_matmul_2_by_2_CPU"
iters: 30000
wall_time: 11.
580435435
extras {
key: "examples_per_sec"
value {
double_value: 86352.538781
}
}
}
entry {
name: "MicroBenchmarks.benchmark_tfe_py_fastpath_execute_matmul_2_by_2_CPU"
iters: 30000
wall_time: 7.
02576637268
extras {
key: "examples_per_sec"
value {
double_value: 142333.227004
}
}
}
PiperOrigin-RevId:
184734289
Benoit Steiner [Tue, 6 Feb 2018 21:48:55 +0000 (13:48 -0800)]
Added support for nested functions
Properly handle the case of control dependencies
PiperOrigin-RevId:
184733444
Joshua V. Dillon [Tue, 6 Feb 2018 21:03:03 +0000 (13:03 -0800)]
Add utility function which makes implicit `tf.get_variable` dependencies an
explicit argument of a callable.
PiperOrigin-RevId:
184725878
Amit Patankar [Tue, 6 Feb 2018 20:58:56 +0000 (12:58 -0800)]
Fixing the issue where an escape character was being included in the branch name changes.
PiperOrigin-RevId:
184725332
Sanjoy Das [Tue, 6 Feb 2018 20:57:07 +0000 (12:57 -0800)]
[TF:XLA] Bump open source llvm revision to r324323
PiperOrigin-RevId:
184725126
Skye Wanderman-Milne [Tue, 6 Feb 2018 20:44:12 +0000 (12:44 -0800)]
Fetch OpDefs from C API instead of using python op registry in ops.py.
PiperOrigin-RevId:
184723558
A. Unique TensorFlower [Tue, 6 Feb 2018 20:30:05 +0000 (12:30 -0800)]
Fixed typo
PiperOrigin-RevId:
184721743
Shanqing Cai [Tue, 6 Feb 2018 20:27:02 +0000 (12:27 -0800)]
tfdbg: deflake session_debug_file_test
by disabling grappler in the test.
PiperOrigin-RevId:
184721353
Alexandre Passos [Tue, 6 Feb 2018 20:04:48 +0000 (12:04 -0800)]
Second, cleaner, attempt at external control dependency handling.
PiperOrigin-RevId:
184718016
A. Unique TensorFlower [Tue, 6 Feb 2018 20:02:18 +0000 (12:02 -0800)]
Improve side_effect_guards to (1) avoid aliasing non-tensors and (2) combine aliasing with re-indenting. Move the renaming visitor into a generic utility API.
This loses potential efficiency by risking sequencing ops where there is no risk of a race. On the other hand it's still not entirely robust, and we need to raise an error where we can't guarantee that.
PiperOrigin-RevId:
184717456
Alexandre Passos [Tue, 6 Feb 2018 20:01:22 +0000 (12:01 -0800)]
Estimator should handle ResourceVariables correctly.
PiperOrigin-RevId:
184717305
Alexandre Passos [Tue, 6 Feb 2018 19:58:36 +0000 (11:58 -0800)]
Support resource variables in moving averages.
PiperOrigin-RevId:
184716895
Alexandre Passos [Tue, 6 Feb 2018 19:51:47 +0000 (11:51 -0800)]
Contrib estimator should handle ResourceVariables properly.
PiperOrigin-RevId:
184715896
Mingsheng Hong [Tue, 6 Feb 2018 19:51:18 +0000 (11:51 -0800)]
Internal change
PiperOrigin-RevId:
184715822
Anna R [Tue, 6 Feb 2018 19:07:04 +0000 (11:07 -0800)]
Adding tf_export decorators/calls to TensorFlow functions and constants.
PiperOrigin-RevId:
184708277
Yao Zhang [Tue, 6 Feb 2018 19:00:58 +0000 (11:00 -0800)]
Enable layout optimizer by default
PiperOrigin-RevId:
184707084
Amit Patankar [Tue, 6 Feb 2018 19:04:17 +0000 (11:04 -0800)]
Merge pull request #16785 from av8ramit/merge_branch
Merging the 1.6 branch back into master.
A. Unique TensorFlower [Tue, 6 Feb 2018 18:48:28 +0000 (10:48 -0800)]
Support training with model parallelism in tpu_estimator.
PiperOrigin-RevId:
184704902
Benoit Steiner [Tue, 6 Feb 2018 18:43:45 +0000 (10:43 -0800)]
Use the function name as the grappler item id
PiperOrigin-RevId:
184704131
ted chang [Tue, 6 Feb 2018 18:44:48 +0000 (10:44 -0800)]
remove write_version=saver_pb2.SaverDef.V1 (#15966)
Amit Patankar [Tue, 6 Feb 2018 18:40:54 +0000 (10:40 -0800)]
Merge pull request #16789 from ManHyuk/fix_typo
Fix typo
Michael Case [Tue, 6 Feb 2018 18:28:19 +0000 (10:28 -0800)]
Merge pull request #16792 from case540/branch_184622482
Branch
184622482
A. Unique TensorFlower [Tue, 6 Feb 2018 18:13:04 +0000 (10:13 -0800)]
Add local interconnect data to DeviceLocality.
This information can be used within a distributed implementation
when deciding how to route data transfers that might involve more
than one hop. By default the new fields are populated according to
StreamExecutor::CanEnablePeerAccessTo(), however a platform-specific
implementation can augment them with more detailed values.
Do some refactoring of gpu_device and gpu_device_factory, making
GetDeviceLocalities() and GetInterconnectMaps() into virtual
functions.
PiperOrigin-RevId:
184698821
Benoit Steiner [Tue, 6 Feb 2018 18:11:26 +0000 (10:11 -0800)]
Explicitly specify the value of all the attributes for the newly created Assign
nodes since we can't always rely on TF calling AddDefaultAttrsToGraphDef.
PiperOrigin-RevId:
184698463
A. Unique TensorFlower [Tue, 6 Feb 2018 17:40:53 +0000 (09:40 -0800)]
Fix bug in and speed up ConstantFolding::CreateNodeDef():
* Fix bug trying to store more than kintmax32 values in a repeated proto field.
* Speed up populating compressed format. Example: tensorflow/python/kernel_tests/large_concat_op_test with size = 2**29+6 goes from ~30 seconds to ~15 seconds. The fraction of time spent in ConstantFolding::CreateNodeDef() goes down from about 35% to about 12%.
PiperOrigin-RevId:
184693749
gracehoney [Tue, 6 Feb 2018 17:34:17 +0000 (09:34 -0800)]
Revert "Remove LICENSE file. It will be added internally instead."
This reverts commit
d0c2b84c2d6cd3c235671e67f22e3639b4b2b54c.
Andrew Selle [Tue, 6 Feb 2018 17:28:24 +0000 (09:28 -0800)]
Add a utility function to partition TensorFlow Lite graphs into subgraphs.
This will be used for the forthcoming delegate interface. The delegate
interface will allow parts of the TensorFlow lite graph to be sent to
other accelerators. NNAPI will be implemented to move to this interface.
That will allow partial delegation to NNAPI.
PiperOrigin-RevId:
184692126