aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* [XLA] Enable HloEvaluator for constant folding, also merged a few operationsGravatar Kay Zhu2017-05-25
| | | | | | | | | | | | | | | | | | | | | | | from hlo_constant_folding to hlo_evaluator. Additionally: - In ShapeUtil::ForEachIndex: * fix a bug where visitor is called when the shape has zero elements (e.g., F32{1,0}) * added test case for ForEachIndex. - In HloEvaluator: * Instead of copying and caching a Constant instruction, return the literal directly if the instruction is constant. * Fix an issue where TUPLE and OPAQUE primitives are not keyed in the templated typed_visitor. * Use (fixed) LiteralUtil::Populate to populate resulting literal, fixes the preexisting bug in the evaluator where R0 and shape with zero size dimensions are not handled. * Refactor ElementWiseUnaryOp and HandleCompare to be templatized on the operand's type. * Refactor IsFinite to be top level since it is only applicable to floats and the return type is always boolean. * Change from std::remainder to std::fmod for kRemainder to be compliant with existing XLA behavior. * Change from std::max and std::min to std::fmax and std::fmin to handle NaNs. * Minor comments fix. - Disables constant_folding and reshape-motion for ClientLibraryTestBase so that constant folding would not affect the intended code paths to be execercised by the test. In the longer term we plan change all Constants to Parameter and re-enable constant_folding in tests. PiperOrigin-RevId: 157174708
* Disable tensorboard/plugins/images/images_plugin_test.py from windows cmake.Gravatar A. Unique TensorFlower2017-05-25
| | | | PiperOrigin-RevId: 157169680
* Automated g4 rollback of changelist 156556243Gravatar A. Unique TensorFlower2017-05-25
| | | | PiperOrigin-RevId: 157169178
* Change shape inference so that a single resource tensor can carryGravatar A. Unique TensorFlower2017-05-25
| | | | | | | | shape and type information for multiple tensors. Apply this to QueueDequeueV2 handled by grappler. PiperOrigin-RevId: 157163757
* Simplify `dense_to_sparse_tensor` and `indicators_to_sparse_ids`, and fix ↵Gravatar A. Unique TensorFlower2017-05-25
| | | | | | | | | them to work with inputs of undefined rank. Add test for `indicators_to_sparse_ids` `dtype` arg. Small update to `unstack` pydoc. PiperOrigin-RevId: 157160634
* Open-source ops to load and remap matrix (2-D) tensors. (Used for ↵Gravatar Wei Ho2017-05-25
| | | | | | loading/remapping embeddings, warm-starting weights/biases, etc.) PiperOrigin-RevId: 157148893
* Added an option to disable optimization when running grappler clusters.Gravatar Benoit Steiner2017-05-25
| | | | PiperOrigin-RevId: 157147578
* Removes unused contrib package feeder.Gravatar Alexandre Passos2017-05-25
| | | | PiperOrigin-RevId: 157146520
* Halve the number of CPU tests run in 9 files.Gravatar Vijay Vasudevan2017-05-25
| | | | | | | | | | | | | | | | | This pattern tests all tests twice (once for use_gpu=True, once for use_gpu=False). Switch to using a single test pattern that uses use_gpu=True, which tests on GPU if a GPU is available, and tests on CPU otherwise. the %test%_gpu target tests on GPU, the %test% target tests on CPU, so both are covered. The tricky situation is when you want to test locally on GPU, you have to remember to run the _gpu target, but reducing test time is more important than individual developers making mistakes at this point. PiperOrigin-RevId: 157141876
* Zero cost for _Recv op, and a few clean up in virtual_scheduler.Gravatar A. Unique TensorFlower2017-05-25
| | | | PiperOrigin-RevId: 157129614
* Make a plugin that serves data for the image dashboard.Gravatar A. Unique TensorFlower2017-05-25
| | | | | | Subsequent changes will make TensorBoard use this images plugin instead of the previous handlers for image-related data. PiperOrigin-RevId: 157123389
* Fixed the cost model to return an unknown shape instead of returning the ↵Gravatar Benoit Steiner2017-05-25
| | | | | | shape of a scalar when the shape hasn't been set. PiperOrigin-RevId: 157120482
* Deleting stale comment and clarifying error on CuDNN init failure.Gravatar A. Unique TensorFlower2017-05-25
| | | | | | The deleted comment suggests that CUDA 6.5 is enough here although we're now building binaries with CUDA 8. The updated error message makes it obvious when CuDNN fails due to an outdated driver. PiperOrigin-RevId: 157118560
* Added a trailing space so that ValueError reads like "..._FeatureColumn. ↵Gravatar A. Unique TensorFlower2017-05-25
| | | | | | Given" instead of "_FeatureColumn.Given". PiperOrigin-RevId: 157116424
* Refactor session.py to reference input Tensors/Operations directly instead ↵Gravatar Skye Wanderman-Milne2017-05-25
| | | | | | | | of their names This is a step towards switching to the new session API, which expects the corresponding C objects instead of names. PiperOrigin-RevId: 157114094
* Improvements for StatSummarizer when run on GPUs.Gravatar A. Unique TensorFlower2017-05-25
| | | | PiperOrigin-RevId: 157105234
* Add functionality in the ValidationMonitor to store and retrieve all ↵Gravatar A. Unique TensorFlower2017-05-24
| | | | | | evaluation metrics computed along with the best validation metric. PiperOrigin-RevId: 157081910
* Automated g4 rollback of changelist 157060591Gravatar Peter Hawkins2017-05-24
| | | | PiperOrigin-RevId: 157070223
* Enable removal of nodes with multiple inputsGravatar Pete Warden2017-05-24
| | | | PiperOrigin-RevId: 157068219
* Print out detailed node counts, shapes and costs. This is useful for a ↵Gravatar A. Unique TensorFlower2017-05-24
| | | | | | | | | | couple of reasons: 1) Shapes are useful for debugging whether the grappler operations match the expected operations. For example, checking that the batch sizes propagate to the individual matrix multiply. 2) The counts are useful for matching the number of individual grappler ops against the actual op counts from a timeline profile. This will help in debugging the cost that grappler outputs. 3) The individual per-op timings give a more detailed breakdown of how grappler is estimating the run time. PiperOrigin-RevId: 157062681
* Fix prefix of kernel for RemoteFusedGraphOpsGravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157062094
* Delete debugging code in FakeQuant* operations. They are stable now, and ↵Gravatar A. Unique TensorFlower2017-05-24
| | | | | | extra checks clutter up code. PiperOrigin-RevId: 157061530
* Issue a more user-friendly error message if a variable's initializer is from ↵Gravatar Peter Hawkins2017-05-24
| | | | | | | | inside a control-flow scope, such as tf.cond() or tf.while_loop(). Fixes #8604. PiperOrigin-RevId: 157060591
* [XLA] In ReshapeMover:Gravatar Kay Zhu2017-05-24
| | | | | | | | | | | - change OperandCanTriviallyChangeShape to only look at operand's opcode, instead of comparing the dimension of the shape between the instruction and its operand. - remove accounting for non-trivial reshapes/transposes, and instead move the requirement to FirstNonScalarReshapeOperand instead. This fixes an issue where a elementwise op's reshape0(constant0) operand gets moved (because reshape0 and constant0 are of different shapes), even when there is no benefit is doing so. Also fixes incorrect test comments. PiperOrigin-RevId: 157060319
* [XLA] Make HloModule always have a config.Gravatar Eli Bendersky2017-05-24
| | | | | | | | | | | This removes the circular dependency when creating a HloModule, then a HloModuleConfig with the help of the module's entry computation, then assigning the config back to the module. Now we have to pass a config when creating a module, or a default config gets created. This allows removing quite a bit of boilerplate code. PiperOrigin-RevId: 157059949
* Preliminary Infeed support for GPU backend.Gravatar A. Unique TensorFlower2017-05-24
| | | | | | | ** GPU transfer manager and GPU specific infeed manager/infeed buffer implementation ** Infeed thunk PiperOrigin-RevId: 157054373
* Fix bug which can cause last element of beam to have a score of infinity.Gravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157052586
* Add an error message to the bundle-shim for SessionBundle deprecation.Gravatar Sukriti Ramesh2017-05-24
| | | | PiperOrigin-RevId: 157049830
* Minor documentation fix.Gravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157047588
* Add a view that organize results by operation type:Gravatar A. Unique TensorFlower2017-05-24
| | | | | | | | | | | | | | | | | | | | | | tfprof> op -select micros,bytes,occurrence -order_by micros SoftmaxCrossEntropyWithLogits 36.58MB (100.00%, 0.05%), 1.37sec (100.00%, 23.56%), 30 MatMul 2720.57MB (99.95%, 3.66%), 988.90ms (76.44%, 17.05%), 3450 ConcatV2 741.37MB (96.29%, 1.00%), 421.44ms (59.38%, 7.27%), 6098 Mul 3957.24MB (95.29%, 5.33%), 418.90ms (52.12%, 7.22%), 9427 Add 740.05MB (89.96%, 1.00%), 335.26ms (44.89%, 5.78%), 2180 Sub 32.46MB (88.97%, 0.04%), 216.44ms The CL also: 1. Unify code view and op view implementation and rename. 2. Clean up a few unused features. 3. Unify the option semantics of all views. PiperOrigin-RevId: 157043043
* Introduced tf.estimator.classifier_parse_example_spec a utility function to ↵Gravatar Mustafa Ispir2017-05-24
| | | | | | help parsing for classifiers. PiperOrigin-RevId: 157041528
* Tutorial for tf.contrib.kernel_methods module.Gravatar Petros Mol2017-05-24
| | | | PiperOrigin-RevId: 157040972
* Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157039326
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157038372
* [XLA] Remove stream priming from BackendGravatar Eli Bendersky2017-05-24
| | | | | | | | This was originally added a long time ago to help with GPU benchmarks; I think it adds needless complexity to the Backend class, which we intend to simplify. If this is still an issue for benchmarks, it should be done on a higher level. PiperOrigin-RevId: 157037991
* Implement quantized addition op, with NEON-acceleration for ARM devicesGravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157037658
* C API/Windows: Export API functions that were accidentally left out.Gravatar Asim Shankar2017-05-24
| | | | PiperOrigin-RevId: 157037270
* Allow uses of over-parameterized separable_conv.Gravatar Vijay Vasudevan2017-05-24
| | | | | | | Fixes #4330. RELNOTES: Allow uses of over-parameterized separable convolution. PiperOrigin-RevId: 157035904
* Arrange TensorForest tree_utils sources properly to avoid double linking.Gravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157035658
* Add various XLA flags to tfcompile.Gravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157034590
* tfdg: minor doc fixGravatar Shanqing Cai2017-05-24
| | | | | | | debug_utils is an internal target that shouldn't be used by users. Replacing it with the correct debug / tf_debug import. PiperOrigin-RevId: 157032842
* Added a new IsConstant() function.Gravatar Benoit Steiner2017-05-24
| | | | PiperOrigin-RevId: 157032759
* Extract the scalars dashboard to a proper pluginGravatar A. Unique TensorFlower2017-05-24
| | | | | | | | | | | | | | | | | | | | | | | | | This CL achieves the following: - The output of the /data/runs endpoint no longer contains any references to "scalars." - The router no longer has any special handling for scalars. - The application (application.py) and its tests no longer have any special handling for scalars. As part of the process of pulling out the scalars dashboard, this leaves behind some assorted infrastructure to make it easier to migrate other plugins (primarily, in the router). Tests done so far: - All tests descended from the "tensorboard" directory - Manual interaction of the following forms: expanding a scalar plot; enlarging a scalar plot; shrinking a scalar plot; using the run selector's check boxes and radio buttons; using the run selector's regex filter; creating a tag group; viewing and scrolling through images, along with a casual check of the other dashboards - Manual inspection of the /data/runs output to make sure that it contains no scalar data PiperOrigin-RevId: 157028690
* Change CreateThread() back to passing by value.Gravatar Peter Hawkins2017-05-24
| | | | PiperOrigin-RevId: 157020704
* [tf contrib layers] Add options to layer_norm to configure layer axes.Gravatar Eugene Brevdo2017-05-24
| | | | | | | | This is a fully backwards compatible change that allows the user to decide: * Which axes to normalize * The shape of the scale/center variables (these are broadcasted). PiperOrigin-RevId: 157019098
* Fix bug in TensorForest for mutli-regression statistics.Gravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157018996
* Script builds a framework file that can be used in CocoaPods distributionGravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157016549
* Remove the 'aliases' section for things with only one name, and make it ↵Gravatar Mark Daoust2017-05-24
| | | | | | | | clear that it is a list of aliases. Add parent class to class pages. PiperOrigin-RevId: 157011127
* changing dropout utils interfaceGravatar A. Unique TensorFlower2017-05-24
| | | | PiperOrigin-RevId: 157011106
* Track statistics for looping graphs in benchmark, and add missing kernelsGravatar Pete Warden2017-05-24
| | | | PiperOrigin-RevId: 157010139
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-06 00:55:50 +0000