aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
...
* Expand all saveable operations to generate a single C++ restore call.Gravatar Russell Power2018-01-04
| | | | | | | | This allows us to avoid repeated index lookups and perform a sequential scan of the index in the common case where we are doing a full restore, or a restore from a sub-model. It also dramatically reduces excessive restore parallelism. Testing with a checkpoint with 1000 100x100 tensors, restoring from CNS drops from ~1m to ~5 seconds. PiperOrigin-RevId: 180827583
* Automated g4 rollback of changelist 179552496Gravatar HyoukJoong Lee2018-01-04
| | | | PiperOrigin-RevId: 180825516
* adding tests to make sure that the public methods that could either call ↵Gravatar Olivia Nordquist2018-01-04
| | | | | | C-API or python API are returning the same types--mostly the string and unicode types. PiperOrigin-RevId: 180821956
* Clarify confusing documentation regarding FinalBeamSearchDecoderOutput.Gravatar Adam Roberts2018-01-04
| | | | PiperOrigin-RevId: 180820115
* Correct mention of channel ordering (we support NHWC only)Gravatar Andrew Selle2018-01-04
| | | | PiperOrigin-RevId: 180818654
* Java: Package GPU native libraries in Maven.Gravatar Asim Shankar2018-01-04
| | | | | | | | | | | | | | | | | | | | | And update version to 1.5.0-rc0. Starting with 1.5.0-rc0, it will be possible to use GPUs from TensorFlow Java programs by adding the following to the application's pom.xml: <dependency> <groupId>org.tensorflow</groupId> <artifactId>libtensorflow</artifactId> <version>1.5.0-rc0</version> </dependency> <dependency> <groupId>org.tensorflow</groupId> <artifactId>libtensorflow_jni_gpu</artifactId> <version>1.5.0-rc0</version> </dependency> Updates #12909 PiperOrigin-RevId: 180816468
* Expand error-message when a valid export is not found at the specified ↵Gravatar Sukriti Ramesh2018-01-04
| | | | | | export location in bundle-shim. PiperOrigin-RevId: 180814251
* Adds the lookup_config property to the _WeightedSparseColumn class.Gravatar A. Unique TensorFlower2018-01-04
| | | | PiperOrigin-RevId: 180811282
* Ensure decode_raw on an empty string returns an empty tensor.Gravatar A. Unique TensorFlower2018-01-04
| | | | PiperOrigin-RevId: 180809736
* Support scalar and vector condition for select.Gravatar Yao Zhang2018-01-04
| | | | PiperOrigin-RevId: 180809175
* Allowing override of common_env.sh python directory.Gravatar Amit Patankar2018-01-04
| | | | PiperOrigin-RevId: 180806246
* Currently LayoutAssignment only deals with a module at a time andGravatar A. Unique TensorFlower2018-01-04
| | | | | | | | | | | | | | | knows nothing about other modules. Therefore, sends and recvs are given a default layout to ensure that the layout is the same across all communicating modules. This CL adds a ChannelLayoutConstraints object which can be passed into LayoutAssignment. When a send or recv is layed out, its chosen layout is saved into the ChannelLayoutConstraints. When, later, another instruction is seen in a different module that uses the same communicating channel ID, the layout is constrained to be the same as in the previous module. PiperOrigin-RevId: 180771799
* Internal changeGravatar A. Unique TensorFlower2018-01-04
| | | | PiperOrigin-RevId: 180762948
* Support reduction and histogram summary ops.Gravatar Yao Zhang2018-01-03
| | | | PiperOrigin-RevId: 180755908
* Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2018-01-03
| | | | PiperOrigin-RevId: 180754374
* [XLA:CPU] Count cycles in non-entry computationsGravatar Sanjoy Das2018-01-03
| | | | | | | | | | | | This change teaches XLA to maintain cycle counters specifically for non-entry computations, like computations representing the body of a While. Without this change, instructions in such non-entry computations are noted as taking 0.00% of their execution time which isn't ideal. Implementation-wise, this just falls out of uniformly using a std::unordered_map<T, int64> for both the HloInstruction->ProfileIndex and the HloComputation->ProfileIndex mappings. PiperOrigin-RevId: 180750463
* Made the shape of the elements stored in a TensorArrayV3 available to shape ↵Gravatar Benoit Steiner2018-01-03
| | | | | | inference. PiperOrigin-RevId: 180750305
* [XLA:python] Plumb Python exceptions through API instead of OrDie'ing.Gravatar Chris Leary2018-01-03
| | | | PiperOrigin-RevId: 180749642
* tfdbg: add option to disable sending of traceback and source codeGravatar Shanqing Cai2018-01-03
| | | | | | | from TF runtime to debug server, using TensorBoardDebugWrapperSession and TensorBoardDebugHook. PiperOrigin-RevId: 180748754
* Automated g4 rollback of changelist 180738639Gravatar Eugene Brevdo2018-01-03
| | | | PiperOrigin-RevId: 180747698
* [XLA] Fix xla_data.proto for protocGravatar David Majnemer2018-01-03
| | | | | | The "reserved" keyword doesn't work in enumerations, replace it with a comment. PiperOrigin-RevId: 180747499
* Merge changes from github.Gravatar Raghuraman Krishnamoorthi2018-01-03
| | | | PiperOrigin-RevId: 180746153
* Support begin and end mask for strided slice.Gravatar Yao Zhang2018-01-03
| | | | PiperOrigin-RevId: 180744498
* [XLA:CPU] Cleanups to VectorSupportLibrary, TargetMachineFeatures and ↵Gravatar Sanjoy Das2018-01-03
| | | | | | | | | | | DotOpEmitter - Move VectorSupportLibrary to under service/cpu since it is specific to the CPU backend. - Use TargetMachineFeatures to infer the vector width in DotOpEmitter - Move the kAvxVectorSize magic constant into TargetMachineFeatures PiperOrigin-RevId: 180740693
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2018-01-03
| | | | PiperOrigin-RevId: 180739249
* Automated g4 rollback of changelist 180733901Gravatar A. Unique TensorFlower2018-01-03
| | | | PiperOrigin-RevId: 180738639
* Add critical section resource and op that allows execution within the criticalGravatar Eugene Brevdo2018-01-03
| | | | | | | | | section. This op (not in the public API yet) allows exclusive access to certain resources by bottlenecking subgraph access to run within a serialized critical section. PiperOrigin-RevId: 180733901
* Make init_scope install a new graph when graph stack is empty in graph mode.Gravatar Akshay Agrawal2018-01-03
| | | | | | | | Prior to this change, entering an init_scope while in graph mode with an empty graph stack would enable eager execution, if eager execution had previously been enabled; this wasn't the desired behavior. PiperOrigin-RevId: 180730994
* Fix bug with imported while loops with C API enabled.Gravatar Skye Wanderman-Milne2018-01-03
| | | | | | | | | | | | Specifically, make control_flow_util.GetOutputContext robust to imported exit nodes (which don't have control flow contexts). This was a bug prior to the C API being enabled in that imported exit nodes would not have contexts, but it happened to not be exposed. Note that importing a metagraph will add the contexts back after doing the initial import, but there's still a window where no contexts are assigned. PiperOrigin-RevId: 180730785
* Update RNN/LSTM performance docsGravatar A. Unique TensorFlower2018-01-03
| | | | PiperOrigin-RevId: 180730491
* Run C++ control flow validation on FunctionDefs before running.Gravatar Skye Wanderman-Milne2018-01-03
| | | | | | | | | | Clients should ideally prevent such functions from being created in the first place, but we still want the runtime to be robust to malformed functions. Trying to run functions with invalid control flow constructs can result in crashes or hangs, so we want to catch it before running. PiperOrigin-RevId: 180727589
* [XLA] Remove RNG_BERNOULLIGravatar David Majnemer2018-01-03
| | | | | | | RNG_BERNOULLI is easy to compose out of other operations and appears to provide no real benefit. Let's remove it. PiperOrigin-RevId: 180726889
* Fix tf.nn.fractional_max_pool output have same batch size when feed with ↵Gravatar A. Unique TensorFlower2018-01-03
| | | | | | different input batch size. Fixes #14985. PiperOrigin-RevId: 180724096
* Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2018-01-03
| | | | PiperOrigin-RevId: 180721390
* [XLA] Remove a false invariant used in generating test inputs.Gravatar Nick Desaulniers2018-01-03
| | | | | | | | | This invariant is incorrect; the parameter will never be the shape of the use's first operand; in fact the param should match operand(1)'s shape for DynamicSlice and operand(2)'s shape for DynamicUpdateSlice (since start_indices is the argument that we are trying to constrain the bounds for). This is checked for in CreateLiteralForConstrainedUses(). PiperOrigin-RevId: 180717508
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2018-01-03
| | | | PiperOrigin-RevId: 180705266
* Boilerplate: port a subset of the anno and quoting packages from Tangent. ↵Gravatar A. Unique TensorFlower2018-01-03
| | | | | | Also includes a pretty printer API. PiperOrigin-RevId: 180698491
* Automated g4 rollback of changelist 180670333Gravatar Anna R2018-01-03
| | | | PiperOrigin-RevId: 180691955
* Set application_id and user_version on TensorBoard DBGravatar Justine Tunney2018-01-03
| | | | | | | Formatting was also cleaned up, without making any schema or documentation changes. PiperOrigin-RevId: 180688094
* Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2018-01-03
| | | | PiperOrigin-RevId: 180679688
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2018-01-03
| | | | PiperOrigin-RevId: 180677538
* Add Snappy support to SQLiteGravatar Justine Tunney2018-01-03
| | | | | | | | | This change also puts the necessary build infrastructure in place so we can continue to build many proper lightweight SQLite extensions in the future that are easy for users to copy over into their own environments, so they can readily access their data. PiperOrigin-RevId: 180673039
* K-FAC: Add "OnehotCategoricalLogitsNegativeLogProbLoss" to ↵Gravatar A. Unique TensorFlower2018-01-03
| | | | | | kfac.loss_functions_lib._allowed_symbols. PiperOrigin-RevId: 180672608
* [xla] Add RNG ops to local Python client.Gravatar A. Unique TensorFlower2018-01-03
| | | | PiperOrigin-RevId: 180671590
* Removing doc strings from REGISTER_OP calls in core/ops.Gravatar Anna R2018-01-03
| | | | PiperOrigin-RevId: 180670333
* Fix "the the" typoGravatar Mark Daoust2018-01-03
| | | | PiperOrigin-RevId: 180669193
* Add low pass filtering to the classification results.Gravatar Mark Daoust2018-01-03
| | | | | | | | This makes the output much more consistent. Also round the probabilities to 2 decimal places. PiperOrigin-RevId: 180666095
* Fix the scope to prepend to names when importing metagraphs.Gravatar A. Unique TensorFlower2018-01-03
| | | | | | See #15326 PiperOrigin-RevId: 180645977
* Fixed a typo that resulted in the graph being processed in reverse topologicalGravatar Benoit Steiner2018-01-02
| | | | | | order. This made shape inference far less efficient than it should be in some cases. PiperOrigin-RevId: 180629882
* Automated g4 rollback of changelist 180622078Gravatar Sanjoy Das2018-01-02
| | | | PiperOrigin-RevId: 180628481
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-09 20:47:26 +0000