aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Pad support for quantized zero.Gravatar Suharsh Sivakumar2018-04-10
| | | | PiperOrigin-RevId: 192342172
* test previously untested eval codepaths.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192341561
* Handle duplicate features by coalescing them together into a single feature.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192341065
* Refactor layers:Gravatar Francois Chollet2018-04-10
| | | | | | | | | | | | - tf.layers layers now subclasses tf.keras.layers layers. - tf.keras.layers is now agnostic to variable scopes and global collections (future-proof). It also uses ResourceVariable everywhere by default. - As a result tf.keras.layers is in general lower-complexity, with fewer hacks and workarounds. However some of current code is temporary (variable creation should be moved to Checkpointable, arguably, and there are some dependency issues that will require later refactors). - The legacy tf.layers layers behavior is kept, with references to variable scopes and global collections injected in the subclassed tf.layers.base.Layer class (the content of tf.layers.base.Layer is the complexity differential between the old implementation and the new one). Note: this refactor does slightly change the behavior of tf.layers.base.Layer, by disabling extreme edge-case behavior that either has long been invalid, or is dangerous and should most definitely be disabled. This will not affect any users since such behaviors only existed in the base Layer unit tests. The behaviors disabled are: - Option to create reusable variables in `call` (already invalid for some time). - Option to use a variable scope to create layer variables outside of the layer while not having the layer track such variables locally. PiperOrigin-RevId: 192339798
* Update declaration order in staging ops Buffer class according to C++ style ↵Gravatar Smit Hinsu2018-04-10
| | | | | | guide PiperOrigin-RevId: 192336966
* Checkpointable: wrap restore ops in init_scopeGravatar Allen Lavoie2018-04-10
| | | | | | | | This should make restore() work with defun-wrapped code, when variables are created inside the function. Just lifts the restore code into the outer context. Adds a test for it. PiperOrigin-RevId: 192331065
* Employ array flat sizes more directly in reference_ops.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192327464
* Upgrade gRPC version and fix file duplicationGravatar Noah Eisen2018-04-10
| | | | | | This bumps the gRPC version used in OSS Tensorflow to pick up grpc/grpc#14541, which exposes gRPC serialization classes which were previously hidden in an internal namespace. Using these files eliminates files duplicated from gRPC repo PiperOrigin-RevId: 192327358
* Expand list of value-preserving ops. This will increase the number of graphs ↵Gravatar A. Unique TensorFlower2018-04-10
| | | | | | where we can apply the involution and mul->conv fusion optimizations. PiperOrigin-RevId: 192323712
* Enable loop-invariant node motion in the Grappler loop optimizer.Gravatar A. Unique TensorFlower2018-04-10
| | | | | | Thanks to the team at Alibaba, who contributed the original version of this code. PiperOrigin-RevId: 192322484
* Fix bug in TFLite Interpreter python interfaceGravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192318426
* [XLA] Fix the size of the data returned from Literal for sparse literals.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192315888
* Adds support for hoisting out common denominator in arithmetic_optimizerGravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192314177
* [TF:XLA] fix a segfault in MakeFakeArguments, and add a test case.Gravatar Nick Desaulniers2018-04-10
| | | | PiperOrigin-RevId: 192310749
* Simplify test_util.run_in_graph_and_eager_modesGravatar Asim Shankar2018-04-10
| | | | | | | - Get rid of unnecessary options - Update various resource variable tests so that they correctly exercise the cases where the variables are placed on GPU (these "with tf.device('/cpu:0')" blocks that were added for eager execution are no longer necessary) PiperOrigin-RevId: 192309109
* Place data format op on CPU:0.Gravatar Yao Zhang2018-04-10
| | | | PiperOrigin-RevId: 192302833
* Fix markdown in a couple of tf.estimator docstrings.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192299871
* Remove manifest_merger from tensorflow_demo. This is an internal-only attributeGravatar A. Unique TensorFlower2018-04-10
| | | | | | that is being removed from bazel. PiperOrigin-RevId: 192298746
* Fix capitalization (on master)Gravatar Mark Daoust2018-04-10
| | | | PiperOrigin-RevId: 192298563
* [XLA:GPU] Add infrastructure for unrolling kernels to improve bandwidth ↵Gravatar Benjamin Kramer2018-04-10
| | | | | | | | | | | | | utilization. We often have simple kernels that do very little actual work, duplicating that can increase the used bandwidth. This change introduces flags and infrastructure for unrolling kernels, it doesn't include any cost heuristics and is disabled by default. Based on code written by Bixia Zheng. PiperOrigin-RevId: 192296781
* Added minimum op, better type support in maximum.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192292693
* Update documentGravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192292160
* Suppress -Wself-assign in self-assignment tests, which triggers in newer clangGravatar A. Unique TensorFlower2018-04-10
| | | | | | revisions. PiperOrigin-RevId: 192284946
* [XLA] Redesign: implement and test custom call.Gravatar A. Unique TensorFlower2018-04-09
| | | | PiperOrigin-RevId: 192241311
* [XLA] Redesign: implement XlaBuilder::IsConstant, ↵Gravatar A. Unique TensorFlower2018-04-09
| | | | | | | | | XlaBuidler::BuildConstantSubGraph, and Client::ComputeConstant(XlaComputation...). - Since the builder no longer holds a client, we moved the ComputeConstant to the client side so that it can communicate with the service side. Now we add XlaBuilder::BuildConstantSubGraph, which is only responsible for building a subgraph that is compile-time constant. - Before this change, every XlaBuilder has a unique id. Now since it also builds constant subgraph, we give every XlaComputation being built a global unique id, and uniquify instruction names when actually building the XlaComputation. PiperOrigin-RevId: 192236997
* Implementation of ArgMaxGravatar A. Unique TensorFlower2018-04-09
| | | | PiperOrigin-RevId: 192236845
* Updating three tests in constant_folding_test.cc with PlaceHolders Nodes to ↵Gravatar A. Unique TensorFlower2018-04-09
| | | | | | check EvaluateNodes returns the same output for the original and optimized graph PiperOrigin-RevId: 192235310
* Fix the GCS Kokoro build badge links.Gravatar Michael Case2018-04-09
| | | | | | | | The original badges I added were redirected/authenticated links (I think). This resulted in broken images on the README. These new links should not be redirected and should just link to the badge images. PiperOrigin-RevId: 192230689
* Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2018-04-09
| | | | PiperOrigin-RevId: 192227995
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2018-04-09
| | | | PiperOrigin-RevId: 192226063
* [XLA] Attach a reference client to the client_library_test_base, and ↵Gravatar A. Unique TensorFlower2018-04-09
| | | | | | | | | implement ComputeAndCompare for XlaBuilder. - In ComputeAndCompare(XlaBuilder..), compute the reference result by executing on the interpreter backend. - Also migrate some tests to use the new ComputeAndCompare(XlaBuilder..). PiperOrigin-RevId: 192225152
* Include standard headers when libc++ requires them.Gravatar A. Unique TensorFlower2018-04-09
| | | | PiperOrigin-RevId: 192224104
* Add the total device time and total host time to tf_op_stats.Gravatar A. Unique TensorFlower2018-04-09
| | | | PiperOrigin-RevId: 192223560
* [XLA] Redesign: implement fft.Gravatar A. Unique TensorFlower2018-04-09
| | | | PiperOrigin-RevId: 192220070
* [tf.data] Add parameter for select_cols in tf.decode_csv for option to parse ↵Gravatar Rachel Lim2018-04-09
| | | | | | only a subset of CSV columns; added same parameter to make_csv_dataset function in tf.contrib.data. PiperOrigin-RevId: 192219509
* boosted_trees: early stop hooks are fixed to stop at the right momentGravatar Younghee Kwon2018-04-09
| | | | | by reading tensor values in a separate session after train_op run. PiperOrigin-RevId: 192217338
* Remove extra #define EIGEN_USE_THREADS.Gravatar Skye Wanderman-Milne2018-04-09
| | | | | | | It breaks some builds with error: "EIGEN_USE_THREADS" redefined [-Werror] PiperOrigin-RevId: 192216796
* [slim] Allow passing timeout_fn to 'evaluation_loop' so that caller can have ↵Gravatar A. Unique TensorFlower2018-04-09
| | | | | | more control over what to do if checkpoints_iterator times out. PiperOrigin-RevId: 192216302
* Don't fail when running shape inference on a graph that contains functions.Gravatar Benoit Steiner2018-04-09
| | | | PiperOrigin-RevId: 192215493
* Export recurrent and its RNN implementation in tf.contrib.Gravatar Patrick Nguyen2018-04-09
| | | | PiperOrigin-RevId: 192210794
* Internal ChangeGravatar A. Unique TensorFlower2018-04-09
| | | | PiperOrigin-RevId: 192209093
* [XLA] Redesign: implement infeed and outfeed.Gravatar A. Unique TensorFlower2018-04-09
| | | | | | | - XlaBuilder::Infeed is basically ComputationBuilder::Infeed + UserComputation::AddInfeedInstruction + ComputationLowerer::Visit + HloInstruction::CreateInfeed. - Similar for Outfeed. PiperOrigin-RevId: 192206502
* Hide slide_dataset from the new API.Gravatar Anna R2018-04-09
| | | | PiperOrigin-RevId: 192204209
* Add bitcast-convert support to the evaluator and as a method on Literal.Gravatar Mark Heffernan2018-04-09
| | | | PiperOrigin-RevId: 192197163
* [XLA] Use ThreadPool in a safer way.Gravatar Michael Kuperstein2018-04-09
| | | | | | ThreadPool joins its threads when it is destroyed, and there's no way to explicitly join. This means passing a ThreadPool and then scheduling in the callee is risky. PiperOrigin-RevId: 192193752
* Don't run test flaky under TSAN.Gravatar A. Unique TensorFlower2018-04-09
| | | | PiperOrigin-RevId: 192193350
* Acknowledges tape in graph mode doesn't like unknown shapesGravatar Alexandre Passos2018-04-09
| | | | PiperOrigin-RevId: 192192757
* Tweak the context stack so init_scope works with eager GraphsGravatar Allen Lavoie2018-04-09
| | | | | | | Previously breaking out into Graphs created with eager execution enabled would enter the graph but not re-enable eager execution. PiperOrigin-RevId: 192192109
* Don't initialize global threadpool in GraphRunner.Gravatar Skye Wanderman-Milne2018-04-09
| | | | | | | | | | | | | | | TF_Graph creates a ShapeRefiner, which in turn creates a GraphRunner, which prior to this change would eventually create a LocalDevice that initialized the global eigen threadpool. This prevents users from specifying a custom number of threads for the pool via a ConfigProto. This change introduces a new device class, SingleThreadedCpuDevice, that can be used for light-weight computations without initializing the threadpool. Addresses #18300. PiperOrigin-RevId: 192188031
* Fix GPUDebugAllocator and GPUNanResetAllocator:Gravatar Guangda Lai2018-04-09
| | | | | | | | 1. Eliminate the checks when the pointer is nullptr. 2. Fix nan array size. Some application can produce a zero size which would cause out-of-range access error. PiperOrigin-RevId: 192186224
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-05 00:37:00 +0000