aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Mark the ResourceHandleOp as inexpensive.Gravatar Derek Murray2018-09-12
| | | | | | Previously, we would schedule a closure for each ResourceHandleOp, because it is erroneously considered to be "expensive". This would cost several microseconds per op, whereas the execution cost of this kernel is as little as 100ns. This change causes these kernels to execute inline at the beginning of a step. PiperOrigin-RevId: 212712378
* Disable flaky test.Gravatar Francois Chollet2018-09-12
| | | | PiperOrigin-RevId: 212706610
* Allow bijectors to handle scalar ILDJ implementations. This also fixes the ↵Gravatar A. Unique TensorFlower2018-09-12
| | | | | | BatchNorm bijector to work with event_ndims > 1. PiperOrigin-RevId: 212705787
* Fix the colocate_with issue for Adagrad optimizerV2.Gravatar Anjali Sridhar2018-09-12
| | | | PiperOrigin-RevId: 212702577
* Use Eigen::CuboidConvolutionBackwardKernel in Conv3DBackpropFilter.Gravatar Eugene Zhulenev2018-09-12
| | | | | | | | Instead of multiple primitive Eigen ops in Conv3DBackpropFilter, call directly into Eigen function. Modest ~10-25% latency improvement and ~10-20% peak memory reduction. PiperOrigin-RevId: 212701797
* Clean-up a TODO following a previous change via b/69266521Gravatar A. Unique TensorFlower2018-09-12
| | | | PiperOrigin-RevId: 212701024
* Skip tensors with null buffers.Gravatar Suharsh Sivakumar2018-09-12
| | | | PiperOrigin-RevId: 212698623
* Improve static shape inference in grappler by propagating tensors_as_shapes ↵Gravatar Doe Hyun Yoon2018-09-12
| | | | | | | | | | | | | | | | better: Currently, static shape inference propagates shapes of tensors, but in some cases, we do need values; for this, we use input_tensors (from Const input tensor) and input_tensors_as_shapes and output_tensors_as_shapes (these are ShapeHandle format, but has values, currently only for 1D vector). This CL enhances propagation of input_tensors_as_shapes and output_tensors_as_shapes to improve static shape inference. (1) forward scalar Const as input_tensors_as_shapes (currently, only 1D vector), (2) export input_tensors_as_shapes, output const tensor, and output_tensors_as_shapes to the values of inferred input/output TensorProperties (currently, only input const tensors are exported as values), (3) use input_tensors_as_shapes as Const tensor to function input (currently, only Const tensors), (4) forward input_tensors_as_shapes to output_tensors_as_shapes for Identity op, (5) when Pack op concats scalar values to form output_tensors_as_shapes, currently it uses only input_tensors (from Const input tensors), but this CL change Pack to use input_tensors_as_shapes as well. PiperOrigin-RevId: 212696959
* Use WeakKeyDictionaries for global Keras {graph->...} mapsGravatar Igor Ganichev2018-09-12
| | | | | | | | | | | | These globals were holding onto graphs including FuncGraphs, which held onto captured tensors leaving garbage around. This change also adds a test to catch garbage like this in the future. To make the test work, I needed to manually breakup some reference cycles caused by OrderedDicts. We should probably have a custom impl of OrderedDict similar to the one in Python3 and avoid these issues. PiperOrigin-RevId: 212694290
* Preserve unique ids when serializing/deserializing HLO protos.Gravatar Mark Heffernan2018-09-12
| | | | | | Re-assigning unique IDs broke serialization of HloSchedule, and keeping IDs stable improves the fidelity of the proto serialization. This change requires that instructions in HLO module protos have valid, module-scope-unique ids so change the XLA builder to hand out module-scope-unique ids. Previously, instruction ids were only unique in the computation scope. PiperOrigin-RevId: 212692339
* Add unit test for model_to_estimator where inpu_fnGravatar Zhenyu Tan2018-09-12
| | | | | | returns features and labels as a list instead of dict. PiperOrigin-RevId: 212685344
* Internal change.Gravatar Anna R2018-09-12
| | | | PiperOrigin-RevId: 212684548
* Added ability to bucket without padding, as sparse tensors to ↵Gravatar A. Unique TensorFlower2018-09-12
| | | | | | `bucket_by_sequence_length`. PiperOrigin-RevId: 212684420
* Support coverage penalty for beam search decoder (according to ↵Gravatar A. Unique TensorFlower2018-09-12
| | | | | | https://arxiv.org/pdf/1609.08144.pdf). PiperOrigin-RevId: 212683753
* Create experimental DNN Estimators with support for Path-Integrated ↵Gravatar A. Unique TensorFlower2018-09-12
| | | | | | Gradients annotations. PiperOrigin-RevId: 212682657
* [XLA] A queue interface to allow fusion in different orders.Gravatar Yuanzhong Xu2018-09-12
| | | | PiperOrigin-RevId: 212674212
* Add a preconf/ directory for preconfigured targets / toolchains.Gravatar A. Unique TensorFlower2018-09-12
| | | | PiperOrigin-RevId: 212668653
* Correct argument name in declaration of StronglyConnectedComponentsGravatar James Keeling2018-09-12
| | | | | | | | This now matches the definition. I fixed it here rather than in the definition as it seems every call to this function names the variable "num_components". I also tidied up the comment a little. PiperOrigin-RevId: 212668416
* Disable third_party/tensorflow/contrib/lite/testing:zip_test* in {a,m,t}san ↵Gravatar Yanan Cao2018-09-12
| | | | | | tests PiperOrigin-RevId: 212668288
* [XLA] Teach Literal to round trip S8 through protosGravatar Benjamin Kramer2018-09-12
| | | | | | Just for completeness, this isn't being used currently. PiperOrigin-RevId: 212667528
* Merge pull request #21490 from tensorflow:rasmi-patch-1Gravatar TensorFlower Gardener2018-09-12
|\ | | | | | | PiperOrigin-RevId: 212667200
* | Point distributions code to tf.linalg rather than tf.contrib.linalgGravatar Ian Langmore2018-09-12
| | | | | | | | PiperOrigin-RevId: 212658047
* | Roll forward change "Move control flow functionalization as a graph ↵Gravatar Tong Shen2018-09-12
| | | | | | | | | | | | optimization pass, instead of a step in XlaCompiler.". PiperOrigin-RevId: 212657932
* | Add linear_operator_addition to tensorflow/python/. A subsequent CLGravatar Ian Langmore2018-09-12
| | | | | | | | | | | | | | will remove this from contrib. linear_operator_addition is hidden from the public API. PiperOrigin-RevId: 212655087
* | Change HandleFromInput() to return a `const ResourceHandle&` and avoid ↵Gravatar Derek Murray2018-09-12
| | | | | | | | | | | | | | | | copying that type. This avoids unnecessary string copies and deallocations in the ReadVariableOp, and similar ops. PiperOrigin-RevId: 212652588
* | Add basic type propagation for unsupported ops in TFLite conversionGravatar Jared Duke2018-09-12
| | | | | | | | PiperOrigin-RevId: 212651704
* | Fix typo in Tensorflow control_flow_ops_py_test.Gravatar A. Unique TensorFlower2018-09-12
| | | | | | | | | | | | The test would fall back to GPU:0 when unable to find a GPU. This should be CPU. PiperOrigin-RevId: 212649435
* | Fix convolution bug when input and filter dimensions matchGravatar Jared Duke2018-09-12
| | | | | | | | | | | | | | | | | | | | | | | | | | TFLite has an optimized matmul path for cases where the input and filter tensors have matching width+height. However, this case doesn't properly account for multiple *batches*. Account for this and add an appropriate test. Credit to zgxnet for the bug and proposed fix. Fixes #21817 PiperOrigin-RevId: 212645329
* | Support providing default gradient for variant tensors in tf.gradients call.Gravatar Saurabh Saxena2018-09-12
| | | | | | | | PiperOrigin-RevId: 212645190
* | Parameterize test matrix_band_part_testGravatar Yanan Cao2018-09-12
| | | | | | | | PiperOrigin-RevId: 212643986
* | Remove compat.forward_compatible horizon checks for StaticRegexReplace.Gravatar Jeremiah Harmsen2018-09-12
| | | | | | | | PiperOrigin-RevId: 212642629
* | compat: Update forward compatibility horizon to 2018-09-12Gravatar A. Unique TensorFlower2018-09-12
| | | | | | | | PiperOrigin-RevId: 212601523
* | Automated rollback of commit 4c936f1b220676d0d427f5f38b4111cfb9011b5aGravatar A. Unique TensorFlower2018-09-12
| | | | | | | | PiperOrigin-RevId: 212600364
* | Automated rollback of commit c5267a54a63a08234a0314888f6cfe842647a73bGravatar A. Unique TensorFlower2018-09-12
| | | | | | | | PiperOrigin-RevId: 212595533
* | [tf.data] Thread parent cancellation manager through to MapDefunOp ↵Gravatar Rachel Lim2018-09-11
| | | | | | | | | | | | correctly. Refactor MapDefunOp for correctness + better C++ hygiene PiperOrigin-RevId: 212589285
* | Update documentation for SavedModel tags.Gravatar Christina Sorokin2018-09-11
| | | | | | | | PiperOrigin-RevId: 212582822
* | Implementation of square.Gravatar A. Unique TensorFlower2018-09-11
| | | | | | | | PiperOrigin-RevId: 212577288
* | Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2018-09-11
| | | | | | | | PiperOrigin-RevId: 212569958
* | Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2018-09-11
| | | | | | | | PiperOrigin-RevId: 212568029
* | Add square to schema.Gravatar A. Unique TensorFlower2018-09-11
| | | | | | | | PiperOrigin-RevId: 212565231
* | [tf.data] Mechanism for collecting processing time information and modeling ↵Gravatar Jiri Simsa2018-09-11
| | | | | | | | | | | | performance. PiperOrigin-RevId: 212557406
* | Make the visualization tool handle null shape gracefully.Gravatar Yunlu Li2018-09-11
| | | | | | | | PiperOrigin-RevId: 212556651
* | Rename "_periods" private property in ARModel with "_periodicities" to make ↵Gravatar A. Unique TensorFlower2018-09-11
| | | | | | | | | | | | it more accurate. PiperOrigin-RevId: 212555968
* | Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2018-09-11
| | | | | | | | PiperOrigin-RevId: 212555263
* | Automated rollback of commit d31f360e1574553ed23b8d483512a2065ac426ebGravatar A. Unique TensorFlower2018-09-11
| | | | | | | | PiperOrigin-RevId: 212551965
* | Merge pull request #21860 from Intel-tensorflow:add-mkl-cc-testsGravatar TensorFlower Gardener2018-09-11
|\ \ | | | | | | | | | PiperOrigin-RevId: 212551803
* | | Removes confusing comment in floordiv docstring.Gravatar A. Unique TensorFlower2018-09-11
| | | | | | | | | | | | PiperOrigin-RevId: 212548263
* | | Internal change.Gravatar Nupur Garg2018-09-11
| | | | | | | | | | | | PiperOrigin-RevId: 212545735
* | | Move AutoGraph to core. This CL moves the entirety of the code base, keeping ↵Gravatar Dan Moldovan2018-09-11
| | | | | | | | | | | | | | | | | | the frontend autograph module in contrib for backward compatibility. Certain files, like notebooks and the readme file may be referenced from the outside, so a copy of those is kept as well. In addition, the notebooks subdirectory of examples is also kept in contrib because the extension the build file relies on is not available in the PIP package. PiperOrigin-RevId: 212543067
* | | Merge pull request #18293 from rongjiecomputer:llvmGravatar TensorFlower Gardener2018-09-11
|\ \ \ | | | | | | | | | | | | PiperOrigin-RevId: 212543062
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-01 14:23:05 +0000