aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Merge changes from github.Gravatar Jonathan Hseu2016-12-22
| | | | Change: 142805270
* Introduce a shim function wrapping checks for possible SessionBundle orGravatar Sukriti Ramesh2016-12-22
| | | | | SavedModel export formats. Change: 142803996
* Optimize the case of a master communicating with an in-process worker.Gravatar Derek Murray2016-12-22
| | | | | | | | | | | | | | | | | | | This change modifies the GrpcWorkerCache so that, when a master attempts to communicate with the worker in the same process, it does so by direct method calls on a `WorkerInterface*`, without making a loopback RPC call. This change is another incremental step towards addressing issue #6256. There are further improvements possible, and we will continue to investigate them, including: * Avoiding the protobuf encoding/decoding for request/response objects where this affects performance. The zero-copy `TensorResponse` class is an example of how we could improve performance here, for `RunGraphRequest` and `RunGraphResponse` objects. * Profiling the closure creation/context switch overhead for interactions with the local worker. Change: 142793965
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-12-22
| | | | Change: 142791193
* Fix argument name in docstring.Gravatar A. Unique TensorFlower2016-12-22
| | | | Change: 142790604
* Add a small documentation blurb to __import__.py describing steps for adding ↵Gravatar A. Unique TensorFlower2016-12-22
| | | | | | more imports to this file. Change: 142784743
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-12-22
| | | | Change: 142783149
* Allows int64 indices in gather_nd gradient.Gravatar A. Unique TensorFlower2016-12-22
| | | | Change: 142781367
* tf.Operation: override default __repr__ to include node name and typeGravatar Shanqing Cai2016-12-22
| | | | | | | | | This makes print-out messages less random, more informative, and shorter in most cases. For example: Before: <third_party.tensorflow.python.framework.ops.Operation object at 0x7f6e003c6f50> After: <tf.Operation 'train_op/dnn/control_dependency' type=Identity> It also improves consistency with Tensor.__repr__(). Change: 142781125
* Allow layers to define variables in call, test that it doesn't lead to ↵Gravatar Lukasz Kaiser2016-12-22
| | | | | | duplication. Change: 142780852
* Passes ready_for_local_init_op in graph_actions.evaluate.Gravatar A. Unique TensorFlower2016-12-22
| | | | Change: 142779313
* Fix Windows cmake build by excluding newly broken test.Gravatar Jonathan Hseu2016-12-22
| | | | Change: 142778020
* Benchmark for adjust_saturation.Gravatar A. Unique TensorFlower2016-12-22
| | | | Change: 142777129
* Missed Variable to VariableV2 switch.Gravatar A. Unique TensorFlower2016-12-22
| | | | Change: 142775917
* strided_slice is replacing tf.sliceGravatar Olivia Nordquist2016-12-22
| | | | Change: 142767321
* Define GEMMLOWP_ALLOW_SLOW_SCALAR_FALLBACK to allow building withGravatar A. Unique TensorFlower2016-12-22
| | | | | | latest gemmlowp, which makes it an error by default to build without SSE4.1 on x86 or NEON on ARM. Change: 142758203
* Use concat_v2 throughout array_ops.pyGravatar A. Unique TensorFlower2016-12-22
| | | | Change: 142748584
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142732418
* Move the implementation code of rnn_cells to contrib.Gravatar Jianwei Xie2016-12-21
| | | | Change: 142730769
* Create Graph Transform Tool for rewriting model files.Gravatar Pete Warden2016-12-21
| | | | Change: 142729497
* Introduce Glorot initializers in core.Gravatar Francois Chollet2016-12-21
| | | | Change: 142728126
* Fixing typo in doc string, and improve documentation.Gravatar Makoto Uchida2016-12-21
| | | | Change: 142727588
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142726205
* Fix a typo in the split function comment.Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142724724
* Exclude the index column in pandas_input_fn - it causes errors when used ↵Gravatar A. Unique TensorFlower2016-12-21
| | | | | | with infer_real_valued_columns_from_input_fn and canned estimators. Change: 142719058
* Migrate FunctionDefHelper::Define() to NodeDef functions. For callersGravatar A. Unique TensorFlower2016-12-21
| | | | | of Define() where that doesn't work, switch to Create() instead. Change: 142718606
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142718232
* Seal tf.sets.Gravatar Patrick Nguyen2016-12-21
| | | | Change: 142717732
* Automated rollback of change 142698388Gravatar Jonathan Hseu2016-12-21
| | | | Change: 142716019
* Only register Reshape CPU kernels with type constraint of int32 onGravatar Vijay Vasudevan2016-12-21
| | | | | shape parameter. Change: 142713914
* Registers kernels for resources on slice / stack.Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142707321
* Automated rollback of change 142682138Gravatar Jonathan Hseu2016-12-21
| | | | Change: 142705711
* Populates step_container in ops_testutil.Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142700768
* Allow a tensor to have multiple aliases.Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142698388
* Android: add CHECK for if proto file exists along with log message ↵Gravatar Andrew Harp2016-12-21
| | | | | | explaining how to load from assets, and also log warning if compressed proto is larger than 64mb. Change: 142698084
* Automated rollback of change 142681535Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142694447
* Update protobuf .whl links to point to 3.1.0Gravatar Jonathan Hseu2016-12-21
| | | | Change: 142686873
* Let user define a CheckpointSaverHook instead of a default one in tf.learn.Gravatar Mustafa Ispir2016-12-21
| | | | Change: 142686791
* Add `assert_rank_in`, to handle cases where target ranks are a list, not an ↵Gravatar A. Unique TensorFlower2016-12-21
| | | | | | upper or lower bound. Change: 142684404
* Only check for early termination of split decisions every n training steps.Gravatar A. Unique TensorFlower2016-12-21
| | | | | Empirically, n ~ 75 gives the most speedup for the bootstrap method. Change: 142684023
* Seal tf.metrics.Gravatar Patrick Nguyen2016-12-21
| | | | Change: 142683937
* Update summary names in add_gradients_summaries() to conform to the new ↵Gravatar A. Unique TensorFlower2016-12-21
| | | | | | summary API convention Change: 142682974
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142682138
* Replace the Template .var_scope property with .variable_scopeGravatar Malcolm Reynolds2016-12-21
| | | | | This brings the codebase in closer alignment with the style guide: "... and do not abbreviate by deleting letters within a word." Change: 142681857
* Adding BigQuery Reader OP.Gravatar A. Unique TensorFlower2016-12-21
| | | | | This changelist adds the necessary C++ and Python code to register a new op to partition and read BigQuery tables. Change: 142681535
* Added experimental support for libxsmm sparse matrix-dense matrixGravatar A. Unique TensorFlower2016-12-21
| | | | | multiplication. Needs a new enough version of libxsmm with sparse support, and needs some patches to work reliably at all sizes. Change: 142680668
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142677515
* Rename SyncReplicasOptimizerV2 to SyncReplicasOptimizerGravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142676422
* Fix tf.contrib.training.resample_at_rate for float64 dtypeGravatar Stephan Hoyer2016-12-21
| | | | Change: 142673866
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-12-21
| | | | Change: 142669605