Jiri Simsa [Fri, 13 Apr 2018 00:32:36 +0000 (17:32 -0700)]
Porting tests for `rpc_op` to OS.
PiperOrigin-RevId:
192698931
A. Unique TensorFlower [Fri, 13 Apr 2018 00:07:35 +0000 (17:07 -0700)]
Move dummy AssertOp and CheckNumericsOp to //third_party/tensorflow/compiler/tf2xla/kernels.
Enable type DT_STRING for AssertOp and ConstOp, in order to make dummy Assert compile with a const string (assert message) as its input.
PiperOrigin-RevId:
192695938
Yuefeng Zhou [Fri, 13 Apr 2018 00:01:55 +0000 (17:01 -0700)]
Fix a typo in cross_tower_ops.
PiperOrigin-RevId:
192694794
A. Unique TensorFlower [Thu, 12 Apr 2018 23:59:57 +0000 (16:59 -0700)]
Misc. small optimizations in Grappler and shape inference code.
Impact on time per optimizer on inception graph:
model_pruner: 590 ms -> 550 ms (-7%)
function_optimizer: 130 ms -> 130 ms (-0%)
constant_folding: 7600 ms -> 7550 ms (-0.7%)
arithmetic_optimizer: 1860 ms -> 1550 ms (-20%)
loop_optimizer: 320 ms -> 320 ms (-0%)
dependency_optimizer: 1300 ms -> 720 ms (-45%)
layout: 1400 ms -> 1400 ms (-0%)
memory_optimizer: 4200 ms -> 3540 ms (-16%)
PiperOrigin-RevId:
192694528
Stephan Hoyer [Thu, 12 Apr 2018 23:57:40 +0000 (16:57 -0700)]
Document support for boolean values in tf.contrib.training.HParams.
PiperOrigin-RevId:
192694244
Justin Lebar [Thu, 12 Apr 2018 23:56:45 +0000 (16:56 -0700)]
[XLA:GPU] Pass all four args to custom-call convs when they're created.
A custom-call-conv should have four arguments: lhs, rhs, algorithm, and
use-tensor-cores. CudnnAlgorithmPicker did the right thing, and that
path is exercised 99% of the time. But CudnnAlgorithmPicker can fail,
and if it does, we're stuck with whatever we had before. What we had
before only had three of the four args, which is bad.
In addition to fixing it, added an e2e test that catches the bug.
PiperOrigin-RevId:
192694119
A. Unique TensorFlower [Thu, 12 Apr 2018 23:55:30 +0000 (16:55 -0700)]
Propagate sharding of the source instruction to the copies added by layout assignment.
PiperOrigin-RevId:
192693972
A. Unique TensorFlower [Thu, 12 Apr 2018 23:51:43 +0000 (16:51 -0700)]
Change assertions to use the tensor 'x' rather than 'x.op.name'. This enables eager execution in validate_args=True contexts.
PiperOrigin-RevId:
192693458
Jiri Simsa [Thu, 12 Apr 2018 23:35:47 +0000 (16:35 -0700)]
Exposing tensorflow.contrib.proto in the pip package.
PiperOrigin-RevId:
192691078
A. Unique TensorFlower [Thu, 12 Apr 2018 23:26:13 +0000 (16:26 -0700)]
Add support for RNN state array of type tf.identity.
PiperOrigin-RevId:
192689747
Benoit Steiner [Thu, 12 Apr 2018 22:41:41 +0000 (15:41 -0700)]
Added a utility to compute a topo ordering of a graph
PiperOrigin-RevId:
192683166
Derek Murray [Thu, 12 Apr 2018 22:20:34 +0000 (15:20 -0700)]
[tf.data] Clean up //tensorflow/contrib/data/python/ops/BUILD.
Create spearate targets for each submodule, so that each test can depend on
the appropriate subset of Python files.
PiperOrigin-RevId:
192679856
Derek Murray [Thu, 12 Apr 2018 22:20:18 +0000 (15:20 -0700)]
Check if the session has been deleted before releasing a callable.
In some versions of Python, the Session._session field may be cleared
(in `Session.__del__()`) before a callable that has a reference to
that Session is deleted. Add a defensive check in the
`Session._Callable.__del__()` method.
PiperOrigin-RevId:
192679796
A. Unique TensorFlower [Thu, 12 Apr 2018 21:33:16 +0000 (14:33 -0700)]
K-FAC: Deprecate tf.contrib.kfac.
As LayerCollections are required to instantiate KfacOptimizer and FisherEstimator, a deprecation warning is printed upon instantiating LayerCollection.
PiperOrigin-RevId:
192671370
A. Unique TensorFlower [Thu, 12 Apr 2018 21:31:08 +0000 (14:31 -0700)]
ResolveConstantReshape transformation and fix for ResolveConstantTranspose.
PiperOrigin-RevId:
192670991
James Wexler [Thu, 12 Apr 2018 20:24:51 +0000 (13:24 -0700)]
Make new build target public.
PiperOrigin-RevId:
192659759
Alexandre Passos [Thu, 12 Apr 2018 20:24:32 +0000 (13:24 -0700)]
Fix build breakage on metagraph exporting when caching_device is set
PiperOrigin-RevId:
192659701
Priya Gupta [Thu, 12 Apr 2018 20:06:28 +0000 (13:06 -0700)]
Separate out distribute dependency out of training, as it needs to be used in summary utils (which training depends on, thus causing circular dependency).
PiperOrigin-RevId:
192656997
Justine Tunney [Thu, 12 Apr 2018 19:58:05 +0000 (12:58 -0700)]
Upgrade libjpeg-turbo
NOTE: libjpeg-turbo 1.5.90 also exists, which adds AVX2 SIMD support. However
it also comes with a CMake build rewrite and 35 micro-architecture specialized
nasm files for x86_64 alone. We do not have the cycles to update jpeg.BUILD to
incorporate those changes at this time. If anyone wants to try, please note
we'd need predicates such as the following:
config_setting(
name = "haswell_opt",
values = {
"cpu": "haswell", # First Intel chip with AVX2
"compilation_mode": "opt",
},
visibility = ["//visibility:public"],
)
config_setting(
name = "excavator_opt",
values = {
"cpu": "excavator", # First AMD chip with AVX2
"compilation_mode": "opt",
},
visibility = ["//visibility:public"],
)
PiperOrigin-RevId:
192655533
A. Unique TensorFlower [Thu, 12 Apr 2018 19:39:48 +0000 (12:39 -0700)]
Fix comment of bucket_by_sequence_length about return type of
element_length_func. Current code requires tf.int32 in order to compare with
buckets_min which is int32.
PiperOrigin-RevId:
192652917
A. Unique TensorFlower [Thu, 12 Apr 2018 19:29:48 +0000 (12:29 -0700)]
Restore dependency on estimator utils from model.
PiperOrigin-RevId:
192651583
A. Unique TensorFlower [Thu, 12 Apr 2018 19:12:16 +0000 (12:12 -0700)]
[XLA] Redesign: add XlaComputation::IsNull.
PiperOrigin-RevId:
192649052
A. Unique TensorFlower [Thu, 12 Apr 2018 19:09:43 +0000 (12:09 -0700)]
Add softsign bijector.
PiperOrigin-RevId:
192648596
Igor Ganichev [Thu, 12 Apr 2018 19:04:48 +0000 (12:04 -0700)]
Add FunctionTest.testLayerInDefun
PiperOrigin-RevId:
192647818
Allen Lavoie [Thu, 12 Apr 2018 18:59:08 +0000 (11:59 -0700)]
Start moving Checkpointable utilities toward core
Doesn't add to the public API yet, just shifts code around. Changes:
- A tiny bit of renaming (to avoid having _Checkpoint and Checkpoint in the same file)
- Removed the garbage collection decorator from a few tests due to the uuid4() garbage issue (apparently core tests get run on Python 2.7.9?)
- Renamed "Object" to "CheckpointableObject" in the proto, since core protos have Java bindings and apparently Java had something else in mind for the keyword "Object" :)
but otherwise this is a pure move.
After this CL I'll propose adding tf.train.Checkpoint to the API (currently tf.contrib.eager.Checkpoint), move the utilities that are still in contrib/eager to their own contrib directory (there will be a few more misc. utilities for inspecting checkpoints and managing dependencies), get tf.train.Saver to read object-based checkpoints for compatibility, and work on Model.save_weights/load_weights.
PiperOrigin-RevId:
192646890
Igor Ganichev [Thu, 12 Apr 2018 18:54:21 +0000 (11:54 -0700)]
Add testCompileTimeConstantsInDefun in xla
PiperOrigin-RevId:
192646199
A. Unique TensorFlower [Thu, 12 Apr 2018 18:51:34 +0000 (11:51 -0700)]
Construct Orthogonal kernels for 2d convolutions.
PiperOrigin-RevId:
192645769
A. Unique TensorFlower [Thu, 12 Apr 2018 18:46:26 +0000 (11:46 -0700)]
Go: Update generated wrapper functions for TensorFlow ops.
PiperOrigin-RevId:
192644946
Pete Warden [Thu, 12 Apr 2018 18:45:02 +0000 (11:45 -0700)]
Fixed error where no background audio samples were being used when testing no-speech clips
PiperOrigin-RevId:
192644704
A. Unique TensorFlower [Thu, 12 Apr 2018 18:35:39 +0000 (11:35 -0700)]
Fix lost dependency
PiperOrigin-RevId:
192643127
A. Unique TensorFlower [Thu, 12 Apr 2018 18:23:44 +0000 (11:23 -0700)]
Update ops-related pbtxt files.
PiperOrigin-RevId:
192640621
A. Unique TensorFlower [Thu, 12 Apr 2018 18:04:55 +0000 (11:04 -0700)]
Fix shape inference for outside_compilation clusters that include cycles.
PiperOrigin-RevId:
192637289
A. Unique TensorFlower [Thu, 12 Apr 2018 17:47:26 +0000 (10:47 -0700)]
Make default weights initializer in `base_layers.Layer` suitable for their dtype.
PiperOrigin-RevId:
192634133
Eugene Brevdo [Thu, 12 Apr 2018 17:35:41 +0000 (10:35 -0700)]
[TF] Add TensorListPushBackBatch.
Also modify code to ensure aliased forwarding happens whenever
possible with DT_VARIANT objects in ResourceVariables and in the new op.
PiperOrigin-RevId:
192632202
A. Unique TensorFlower [Thu, 12 Apr 2018 17:14:02 +0000 (10:14 -0700)]
Introduced tool to run an HLO module in replicated fashion, by infeeding random data and outfeeding the data generated at each step.
The arguments of the computation can be either read from the session module, or randomly generated.
The tool uses the raw transfer manager API to infeed and outfeed the data.
PiperOrigin-RevId:
192628605
A. Unique TensorFlower [Thu, 12 Apr 2018 16:46:34 +0000 (09:46 -0700)]
Collective Ops Part 3
BaseCollectiveExecutor and RingReducer.
This change is part of a series of changes introducing infrastructure
for collective ops and initial implementations of reduction and broadcast.
PiperOrigin-RevId:
192624521
A. Unique TensorFlower [Thu, 12 Apr 2018 16:44:16 +0000 (09:44 -0700)]
Fixing dependencies.
PiperOrigin-RevId:
192624191
A. Unique TensorFlower [Thu, 12 Apr 2018 15:49:47 +0000 (08:49 -0700)]
Updating tests containing graphs with Variables so that they Evaluate the original and optimized graphs and check if the outputs are same.
PiperOrigin-RevId:
192616402
A. Unique TensorFlower [Thu, 12 Apr 2018 13:22:30 +0000 (06:22 -0700)]
Special-case the name scoping for operator methods. TensorFlow disallows top-level name scopes to begin with underscores. Also use the transformer scope information to get to the enclosing function name.
PiperOrigin-RevId:
192600256
A. Unique TensorFlower [Thu, 12 Apr 2018 11:40:42 +0000 (04:40 -0700)]
Make DType, TensorShape, and Dimension "reducable" for pickling purposes.
PiperOrigin-RevId:
192591402
A. Unique TensorFlower [Thu, 12 Apr 2018 05:37:57 +0000 (22:37 -0700)]
Enable an r2 reduce window test case.
PiperOrigin-RevId:
192560111
A. Unique TensorFlower [Thu, 12 Apr 2018 03:08:30 +0000 (20:08 -0700)]
Fix description of DynamicUpdateSlice.
PiperOrigin-RevId:
192550101
James Wexler [Thu, 12 Apr 2018 02:58:07 +0000 (19:58 -0700)]
Add closure_js_proto_library build for tf.example protos.
PiperOrigin-RevId:
192549109
A. Unique TensorFlower [Thu, 12 Apr 2018 02:50:10 +0000 (19:50 -0700)]
Enable a reduce window test case.
PiperOrigin-RevId:
192548652
A. Unique TensorFlower [Thu, 12 Apr 2018 02:45:42 +0000 (19:45 -0700)]
Go: Update generated wrapper functions for TensorFlow ops.
PiperOrigin-RevId:
192548367
A. Unique TensorFlower [Thu, 12 Apr 2018 02:18:20 +0000 (19:18 -0700)]
Update ops-related pbtxt files.
PiperOrigin-RevId:
192546579
A. Unique TensorFlower [Thu, 12 Apr 2018 02:14:47 +0000 (19:14 -0700)]
[XLA] Redesign: fix GetComputationGraphStats.
CreateFromProto requires that the config has proper entry_computation_layout, so give the config the program shape.
PiperOrigin-RevId:
192546316
Jonathan Hseu [Thu, 12 Apr 2018 01:57:49 +0000 (18:57 -0700)]
Switch to WaitForNotification to fix the flaky test.
See:
https://source.cloud.google.com/results/invocations/
31632a30-3728-4635-a456-
f89b9e8b9dfe/log
PiperOrigin-RevId:
192544848
A. Unique TensorFlower [Thu, 12 Apr 2018 01:38:38 +0000 (18:38 -0700)]
[XLA] Redesign: test sharding.
Also set the sharding to the instruction when created from proto.
PiperOrigin-RevId:
192543024
Derek Murray [Thu, 12 Apr 2018 01:37:47 +0000 (18:37 -0700)]
Increase size of //tensorflow/python/kernel_tests:linalg_ops_test to "medium".
PiperOrigin-RevId:
192542956
Yifei Feng [Thu, 12 Apr 2018 01:20:19 +0000 (18:20 -0700)]
In model_to_estimator, only run get_weights when there are initialized Keras variables(which assumes there exists a session). Otherwise create a session so that we can run get_config().
Actually fix #18193.
PiperOrigin-RevId:
192541442
Derek Murray [Thu, 12 Apr 2018 01:09:42 +0000 (18:09 -0700)]
Add `tf.contrib.stateless.stateless_multinomial()`.
This is a starting point for Dataset-compatible weighted sampling across a list of datasets.
PiperOrigin-RevId:
192540412
Austin Anderson [Thu, 12 Apr 2018 00:46:08 +0000 (17:46 -0700)]
Internal testing changes
PiperOrigin-RevId:
192537874
A. Unique TensorFlower [Thu, 12 Apr 2018 00:29:32 +0000 (17:29 -0700)]
Automated g4 rollback of changelist
192516190
PiperOrigin-RevId:
192536085
A. Unique TensorFlower [Thu, 12 Apr 2018 00:19:20 +0000 (17:19 -0700)]
Update ops-related pbtxt files.
PiperOrigin-RevId:
192534931
Younghee Kwon [Wed, 11 Apr 2018 23:59:45 +0000 (16:59 -0700)]
boosted_trees: make sure ensemble deserialization happens for the non-TRAIN modes too.
PiperOrigin-RevId:
192532297
Derek Murray [Wed, 11 Apr 2018 23:43:33 +0000 (16:43 -0700)]
Move callback into bound function to avoid copying.
PiperOrigin-RevId:
192530231
A. Unique TensorFlower [Wed, 11 Apr 2018 23:05:42 +0000 (16:05 -0700)]
Speed up computing mean confidence intervals by avoiding tf.while_loop.
Implement a vectorized way to compute the same thing instead.
PiperOrigin-RevId:
192524667
A. Unique TensorFlower [Wed, 11 Apr 2018 22:59:47 +0000 (15:59 -0700)]
support profiling multiple tpu through one grpc and one session.
data are saved with host prefix.
PiperOrigin-RevId:
192523668
A. Unique TensorFlower [Wed, 11 Apr 2018 22:44:55 +0000 (15:44 -0700)]
Adding support of core feature columns and losses to gradient boosted trees estimators.
PiperOrigin-RevId:
192521398
A. Unique TensorFlower [Wed, 11 Apr 2018 22:37:49 +0000 (15:37 -0700)]
Revealing the range of node ids in the latest layer via resource' state
PiperOrigin-RevId:
192520351
Derek Murray [Wed, 11 Apr 2018 22:33:06 +0000 (15:33 -0700)]
Increase size of //tensorflow/python/kernel_tests:sets_test to "medium".
PiperOrigin-RevId:
192519639
Alexandre Passos [Wed, 11 Apr 2018 22:23:17 +0000 (15:23 -0700)]
Fixes issue where name scope collisions could lead to an invalid variable in the metagraph.
PiperOrigin-RevId:
192518307
Sourabh Bajaj [Wed, 11 Apr 2018 22:20:11 +0000 (15:20 -0700)]
GCS Filesystem should not cache checkpoint file as we need to read the updated checkpoints from the contents.
PiperOrigin-RevId:
192517819
A. Unique TensorFlower [Wed, 11 Apr 2018 22:09:07 +0000 (15:09 -0700)]
Extend support to remove transpose/reverse on dimensions of size 1.
PiperOrigin-RevId:
192516190
A. Unique TensorFlower [Wed, 11 Apr 2018 21:39:11 +0000 (14:39 -0700)]
Adding support for batch_to_space_nd op with crops.
PiperOrigin-RevId:
192511036
A. Unique TensorFlower [Wed, 11 Apr 2018 21:04:09 +0000 (14:04 -0700)]
Adds a nodedef_fn parameter to copy_op_handler, allowing customization by mutating
NodeDef before creating the copied operation.
PiperOrigin-RevId:
192505209
A. Unique TensorFlower [Wed, 11 Apr 2018 21:02:49 +0000 (14:02 -0700)]
BREAKING_CHANGE: Remove event_ndims in Bijector, and require `log_det_jacobian` methods to take event_ndims.
The class level event_ndims parameter is being deprecated in favor of passing it in
to the `log_det_jacobian` methods.
Specific changes:
- `log_det_jacobian` signatures are now `log_det_jacobian(input, event_ndims)`
- Constructors no long have event_ndims passed in (e.g. Affine() vs. Affine(event_ndims=0)).
- All bijectors must specify a subset of [forward_min_event_ndims, inverse_min_event_ndims]. This is the minimal dimensionality the bijector operates on, with it being "broadcasted" to any passed in event_ndims (e.g. Exp has forward_min_event_ndims = 0. That means it operates on scalars. However, we can use the bijector on any event_ndims > 0 (i.e. we've broadcasted the transformation to work on any amount of event_ndims > 0), and jacobian reduction will work in those cases.
As a result of this change, all bijectors should "broadcast" (e.g. Sigmoid now works on any number of event_ndims).
Other changes (internal and documentation):
- Added clarifications on Jacobian Determinant vs. Jacobian Matrix.
- Added clarifications on min_event_ndims, and what the jacobian reduction is over.
- Changed caching of ildj to be keyed on event_ndims.
- Several bug fixes to bugs unearthed while writing this code (e.g. transformed distribution shape computation being incorrect)
PiperOrigin-RevId:
192504919
Jiri Simsa [Wed, 11 Apr 2018 20:59:58 +0000 (13:59 -0700)]
Porting tests for the `decode_proto` and `encode_proto` to OS.
PiperOrigin-RevId:
192504411
A. Unique TensorFlower [Wed, 11 Apr 2018 20:53:01 +0000 (13:53 -0700)]
Support for removing unfused quantized activation functions and min/max.
PiperOrigin-RevId:
192503204
A. Unique TensorFlower [Wed, 11 Apr 2018 20:48:43 +0000 (13:48 -0700)]
Replace examples/image_retraining by a pointer to TensorFlow Hub.
https://github.com/tensorflow/hub/tree/master/examples/image_retraining
has the same tool, upgraded to use TensorFlow Hub instead of raw graph defs.
PiperOrigin-RevId:
192502469
A. Unique TensorFlower [Wed, 11 Apr 2018 20:47:46 +0000 (13:47 -0700)]
Factor out the syntactic function scope tracking into the transformer. Choosing not to do this at static analysis because it exposes the scope to any node, making it easier to use by any specialization of a transformer.
PiperOrigin-RevId:
192502309
Francois Chollet [Wed, 11 Apr 2018 20:46:03 +0000 (13:46 -0700)]
When not necessary, avoid the creation of a `placeholder_with_default` in BN (not yet supported by TPU compilation).
PiperOrigin-RevId:
192502020
A. Unique TensorFlower [Wed, 11 Apr 2018 20:29:12 +0000 (13:29 -0700)]
DepthwiseConv Optimizations
PiperOrigin-RevId:
192499401
Anna R [Wed, 11 Apr 2018 20:22:53 +0000 (13:22 -0700)]
Internal change.
PiperOrigin-RevId:
192498471
Rohan Jain [Wed, 11 Apr 2018 19:33:04 +0000 (12:33 -0700)]
Adding hp5y back.
PiperOrigin-RevId:
192491335
Austin Anderson [Wed, 11 Apr 2018 19:32:08 +0000 (12:32 -0700)]
Internal TF Lite test changes
PiperOrigin-RevId:
192491201
Derek Murray [Wed, 11 Apr 2018 18:52:52 +0000 (11:52 -0700)]
Increase size of tensorflow/contrib/data/python/kernel_tests:batch_dataset_op_test to "medium".
PiperOrigin-RevId:
192484895
Adria Puigdomenech [Wed, 11 Apr 2018 18:36:56 +0000 (11:36 -0700)]
Update docs for softmax_cross_entropy_with_logits.
PiperOrigin-RevId:
192482242
A. Unique TensorFlower [Wed, 11 Apr 2018 18:33:38 +0000 (11:33 -0700)]
Script to create custom_ops inside a TensorFlow graphdef.
PiperOrigin-RevId:
192481690
A. Unique TensorFlower [Wed, 11 Apr 2018 18:26:25 +0000 (11:26 -0700)]
Disable prelu tests for real now.
PiperOrigin-RevId:
192480452
A. Unique TensorFlower [Wed, 11 Apr 2018 18:21:48 +0000 (11:21 -0700)]
Fix uninitialized value.
PiperOrigin-RevId:
192479630
A. Unique TensorFlower [Wed, 11 Apr 2018 18:02:48 +0000 (11:02 -0700)]
Use tf.train.get_or_create_global_step() instead of deprecated variables.get_or_create_global_step().
PiperOrigin-RevId:
192476077
Jacques Pienaar [Wed, 11 Apr 2018 16:53:21 +0000 (09:53 -0700)]
Add gradient in cond test to match CallGradInLoop.
PiperOrigin-RevId:
192463997
Mark Daoust [Wed, 11 Apr 2018 16:50:40 +0000 (09:50 -0700)]
Add a clear error message for when a doc does not have a title.
PiperOrigin-RevId:
192463583
A. Unique TensorFlower [Wed, 11 Apr 2018 16:43:32 +0000 (09:43 -0700)]
Import FunctionDef as GrapplerFunctionItem
Explicitly track function input arg expansion into Placeholders,
and keep metadata to map between FunctionDef and GraphDef connectivity
formats.
PiperOrigin-RevId:
192462592
Nupur Garg [Wed, 11 Apr 2018 16:34:51 +0000 (09:34 -0700)]
Fixing dependencies.
PiperOrigin-RevId:
192461382
Michael Case [Wed, 11 Apr 2018 16:34:44 +0000 (09:34 -0700)]
Fix Windows GPU TensorFlow Bazel builds.
The configure.py script will error out on Windows GPU builds due
to NCCL attempted to be configured (and is currently Linux only).
PiperOrigin-RevId:
192461362
A. Unique TensorFlower [Wed, 11 Apr 2018 15:30:18 +0000 (08:30 -0700)]
Temporarily remove prelu from generated_examples_zip_test
PiperOrigin-RevId:
192453411
A. Unique TensorFlower [Wed, 11 Apr 2018 15:11:50 +0000 (08:11 -0700)]
Remove unused former source of tensorflow.org/tutorials/image_retraining.
The source of https://tensorflow.org/tutorials/image_retraining has moved from
https://github.com/tensorflow/tensorflow/tree/master/tensorflow/docs_src/tutorials
to https://github.com/tensorflow/hub/tree/master/docs/tutorials
because of its use of TensorFlow Hub.
This change replaces the now-defunct version with a pointer to the new
location, in order to avoid dead code.
PiperOrigin-RevId:
192451570
A. Unique TensorFlower [Wed, 11 Apr 2018 14:47:26 +0000 (07:47 -0700)]
Fixing propagation of minmax info on constant gather ops.
PiperOrigin-RevId:
192448922
Benjamin Kramer [Wed, 11 Apr 2018 09:16:25 +0000 (02:16 -0700)]
[TF:XLA] Mark oom_test as optonly, it's really slow when compiled without optimization.
PiperOrigin-RevId:
192420481
Nick Felt [Wed, 11 Apr 2018 06:44:12 +0000 (23:44 -0700)]
Create FileWriter <-> tf.contrib.summary compatibility layer
This provides an implementation of FileWriter, activated by passing in a `session` parameter to the constructor, that is backed by session.run'ing graph ops that manipulate a tf.contrib.summary.create_file_writer() instance. Because tf.contrib.summary.SummaryWriters are backed by shared resources in the graph, this makes it possible to have a FileWriter and a tf.contrib.summary.SummaryWriter that both write to the same events file.
This change includes some related smaller changes:
- Factors out training_utils.py into a separate target to avoid a cyclic dep
- Moves contrib/summary/summary_ops.py to python/ops/summary_ops_v2.py
- Adds SummaryWriter.init(), .flush(), and .close() op-returning methods
- Changes create_file_writer() `name` arg to default to logdir prefixed by `logdir:` so shared resources are scoped by logdir by default
- Fixes a bug with tf.contrib.summary.flush() `writer` arg
- Makes create_file_writer()'s max_queue arg behave as documented
- Adds more testing for existing tf.contrib.summary API
PiperOrigin-RevId:
192408079
A. Unique TensorFlower [Wed, 11 Apr 2018 05:44:36 +0000 (22:44 -0700)]
[XLA] Fix the size of data buffer for sparse literals.
PiperOrigin-RevId:
192404543
A. Unique TensorFlower [Wed, 11 Apr 2018 05:29:13 +0000 (22:29 -0700)]
[XLA] Redesign: implement ComputeHost.
Also support convert from/to proto for ComputeHost.
PiperOrigin-RevId:
192403660
A. Unique TensorFlower [Wed, 11 Apr 2018 04:42:14 +0000 (21:42 -0700)]
[XLA] Redesgin: implement and test Gather, Conditional.
Also support convert from/to proto for Gather.
PiperOrigin-RevId:
192400659
A. Unique TensorFlower [Wed, 11 Apr 2018 03:48:57 +0000 (20:48 -0700)]
[XLA] Redesign: implment and test CrossReplicaSum.
PiperOrigin-RevId:
192397189
A. Unique TensorFlower [Wed, 11 Apr 2018 02:44:00 +0000 (19:44 -0700)]
Update ops-related pbtxt files.
PiperOrigin-RevId:
192392702
Justin Lebar [Wed, 11 Apr 2018 02:34:54 +0000 (19:34 -0700)]
[StreamExecutor] Merge StreamExecutor's and XLA's StatusOr classes.
StatusOr is a...complicated class to write. It's really not good to
have two copies of it. They've diverged (the XLA one is more
sophisticated), and this may be causing upstream build problems with
gcc6.
PiperOrigin-RevId:
192392111
A. Unique TensorFlower [Wed, 11 Apr 2018 02:31:05 +0000 (19:31 -0700)]
[XLA] Redesign: implement and test BatchNormXXX.
PiperOrigin-RevId:
192391748