aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
...
* Turn the following ops into Identity.Gravatar A. Unique TensorFlower2018-03-10
| | | | | | | | | * Slice when the Size input matches the size of the input tensor * Tile when the multiples input is a tensor of '1' * Pad/PadV2 when the paddings input is a tensor of 0 * Squeeze when the squeeze dimensions are known to be > 1 PiperOrigin-RevId: 188609800
* Increment node_ids when merging CostGraphDef.Gravatar Rui Zhao2018-03-10
| | | | PiperOrigin-RevId: 188586552
* [XLA] Speed up colocated buffer merging.Gravatar Michael Kuperstein2018-03-09
| | | | PiperOrigin-RevId: 188581202
* Fix DepthToSpace and SpaceToDepth to silently return instead of failing when ↵Gravatar A. Unique TensorFlower2018-03-09
| | | | | | the input tensor is empty. PiperOrigin-RevId: 188580972
* Fix docstring for `embedding_lookup_sparse`.Gravatar A. Unique TensorFlower2018-03-09
| | | | | | Example with weighted mean combiner implies that single-key embeddings not normalized (the weighted sum answer). However, the code and test shows normalization regardless of number of keys. PiperOrigin-RevId: 188575982
* Unified test util PlaceHolderFloat() into PlaceHolder(), and extended the latterGravatar Mingsheng Hong2018-03-09
| | | | | | to take a TF_DataType param. PiperOrigin-RevId: 188570493
* Propagate min/max for StridedSliceGravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188569611
* Move optimizations to arithmetic optimizer stagesGravatar A. Unique TensorFlower2018-03-09
| | | | | | | | 1) Redundant Bitcast 2) Redundant Cast 3) Remove inverse transpose PiperOrigin-RevId: 188569367
* Add experimental Session::MakeCallable() API and implement it for DirectSession.Gravatar Derek Murray2018-03-09
| | | | | | | | | | | | | | | | | | | The intent of this new API matches the Python `tf.Session.make_callable()` method: it splits the two roles of the `Session::Run()` method into separate methods: 1. `Session::MakeCallable()` takes information about a subgraph (such as the names of nodes to feed and fetch), and prunes and optimizes that graph, returning a simple handle. 2. `Session::RunCallable()` takes that handle, plus any values to be fed, and executes the graph, returning whatever outputs are produced. This split moves string processing off the critical path of running a step. We also add a new method `Session::ReleaseCallable()` that makes it possible to free the resources associated with a cached subgraph, and could be useful for seldom-executed graphs such as initializers. PiperOrigin-RevId: 188566635
* Fix flakiness in common_runtime/function_test.cc.Gravatar Derek Murray2018-03-09
| | | | | | | | The flakiness was due to nondeterministic names being chosen for folded constants; the fix was to split out the source of the nondetermism into a separate test. PiperOrigin-RevId: 188565362
* Allowing ReorderActivationFunctions to reorder output arrays and adding ↵Gravatar A. Unique TensorFlower2018-03-09
| | | | | | support for propagating activation functions into Gather ops and as constant ops. PiperOrigin-RevId: 188556574
* [XLA] Add a whole graph execution interface.Gravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188554206
* ProcFLR: Include the remote function target in the function_keyGravatar Brennan Saeta2018-03-09
| | | | PiperOrigin-RevId: 188548941
* Move a utility function into the eager C internal APIGravatar Akshay Modi2018-03-09
| | | | PiperOrigin-RevId: 188548393
* Copy `replicate_model_fn` to core.Gravatar Igor Saprykin2018-03-09
| | | | PiperOrigin-RevId: 188547527
* Migrate tf.contrib.bayesflow.optimizers to tfp.optimziers.Gravatar Joshua V. Dillon2018-03-09
| | | | PiperOrigin-RevId: 188547477
* Eager: Fix a Dimension PyObject leak, test for it.Gravatar Allen Lavoie2018-03-09
| | | | PiperOrigin-RevId: 188540944
* Internal Change.Gravatar Michael Case2018-03-09
| | | | PiperOrigin-RevId: 188540659
* Part of the update of tf.keras to the Keras 2.1.5 API.Gravatar Francois Chollet2018-03-09
| | | | PiperOrigin-RevId: 188540513
* Automated g4 rollback of changelist 188501394Gravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188536863
* [TF:XLA] Bump open source llvm revision to r326989Gravatar Sanjoy Das2018-03-09
| | | | PiperOrigin-RevId: 188536576
* Add smart_case that calls smart_cond.Gravatar Skye Wanderman-Milne2018-03-09
| | | | PiperOrigin-RevId: 188534066
* Remove the nondeterminism from a test for initializing variables from ↵Gravatar A. Unique TensorFlower2018-03-09
| | | | | | checkpoints. PiperOrigin-RevId: 188533156
* [XLA:Doc] Fix an typo "Alternately" -> "Alternatively".Gravatar Kay Zhu2018-03-09
| | | | PiperOrigin-RevId: 188532135
* LSTM support: Add non-uint8 quantized operators.Gravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188529107
* Internal ChangeGravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188528771
* Automated g4 rollback of changelist 188492233Gravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188525453
* Automated g4 rollback of changelist 188433328Gravatar Gunhan Gulsoy2018-03-09
| | | | PiperOrigin-RevId: 188525171
* Move warm_starting_util from third_party/tensorflow/python/estimator to ↵Gravatar A. Unique TensorFlower2018-03-09
| | | | | | third_party/tensorflow/python/training (move WarmStartSettings definition to third_party/tensorflow/python/estimator/estimator.py), and make _warm_start() public under tf.train.warm_start(). WarmStartSettings and VocabInfo are both available under tf.estimator, and VocabInfo is also available under tf.train. PiperOrigin-RevId: 188522820
* Allowing for FunctionLibraryRuntime::Run calls to not be provided with a ↵Gravatar Rohan Jain2018-03-09
| | | | | | | | runner to execute kernels with. In that case, it defaults to using the threadpool provided by the device. Also makes sure each device has a default threadpool to fall back on. PiperOrigin-RevId: 188520648
* Add more debugging output, filtering by int_type in XLA test, more tests.Gravatar Jacques Pienaar2018-03-09
| | | | PiperOrigin-RevId: 188513895
* Add bool type for tflite.Gravatar Mingxing Tan2018-03-09
| | | | PiperOrigin-RevId: 188512706
* [XLA:GPU] Don't fuse get-tuple-element.Gravatar Justin Lebar2018-03-09
| | | | | | | | Fusing GTE works, but it's slower than not fusing. (In some sense, GTE is *always* fused; it's just that our "implicit fusion" implementation is faster than our explicit fusion implementation.) PiperOrigin-RevId: 188509801
* Make SetNumThreads apply to the eigen threads. (This creates a dependency on ↵Gravatar A. Unique TensorFlower2018-03-09
| | | | | | eigen!) PiperOrigin-RevId: 188504172
* Automated g4 rollback of changelist 188397087Gravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188503184
* [XLA:GPU] Convert FusionMergerTest to use module strings.Gravatar Justin Lebar2018-03-09
| | | | | | | | This is a nice cleanup, but it also makes this a proper unit test -- the module strings we use are post fusion. (Without module strings, fusion computations are a real pain to create.) PiperOrigin-RevId: 188502642
* Implement partial constant folding for Concat.Gravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188501394
* Ensure that the rank of the input to tf.Unique is 1 as shape inference time ↵Gravatar Benoit Steiner2018-03-09
| | | | | | instead of letting the kernel catch invalid inputs. PiperOrigin-RevId: 188496351
* Use the multithreaded conv only when threads are available.Gravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188495357
* Enable the Grappler arithmetic optimizer for all python tests.Gravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188492233
* Desugar IfExp nodesGravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188491604
* Adding support for constant Gather ops.Gravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188478173
* Convert Squeeze into Reshape and support squeezes on all dimensions.Gravatar A. Unique TensorFlower2018-03-09
| | | | PiperOrigin-RevId: 188477922
* [StreamExecutor] Remove ThenDoHostCallbackForTest -- it's identical to ↵Gravatar Justin Lebar2018-03-09
| | | | | | | | | | | ThenDoHostCallback. The reason this came about is: ThenDoHostCallback was once private, and ThenDoHostCallbackForTest was public. Then at some point ThenDoHostCallback became public, but the *ForTest one was never removed. PiperOrigin-RevId: 188459741
* Fix misleading comment.Gravatar Sanjoy Das2018-03-09
| | | | PiperOrigin-RevId: 188450336
* Remove a layer of templatizationGravatar Sanjoy Das2018-03-08
| | | | | | | | | | | | | With this change - HloTestBase always calls HloRunner with an array of non-owning Literal pointers as arguments - HloRunner no longer has a general LiteralPtr, but just provides explicit overloads for std::unique_ptr<Literal> and Literal* This was prompted by a dependent change that needs to call HloTestBase::RunAndCompare with Literal* arguments. PiperOrigin-RevId: 188446331
* Collapse adjacent dimensions that have no paddings.Gravatar Jingyue Wu2018-03-08
| | | | | | | | | | For example, tf.pad(<4D tensor>, [[0, 0], [0, 0], [0, 0], [0, 1]]) is equivalent to a 2D pad, which is faster. PiperOrigin-RevId: 188440916
* Removing certain attributes from pom-android.xml.Gravatar Amit Patankar2018-03-08
| | | | PiperOrigin-RevId: 188433792
* Adding AudioSpectrogram and MFCC operator support to tfliteGravatar Andrew Harp2018-03-08
| | | | PiperOrigin-RevId: 188433328
* Fix typo in error messageGravatar Sanjoy Das2018-03-08
| | | | PiperOrigin-RevId: 188425637
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-07 10:58:17 +0000