aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2017-02-06
| | | | Change: 146741901
* Java: Add the UINT8 data type.Gravatar Asim Shankar2017-02-06
| | | | Change: 146741882
* Add state_size, output_size, and static shape checking to ResidualWrapper.Gravatar Eugene Brevdo2017-02-06
| | | | Change: 146740995
* Remove dirty tag when generating __version__Gravatar Gunhan Gulsoy2017-02-06
| | | | Change: 146740764
* Change dependencies for the _android_only binaries.Gravatar A. Unique TensorFlower2017-02-06
| | | | Change: 146739972
* tfdbg CLI: Make CursesUI command history persist across sessionsGravatar Shanqing Cai2017-02-06
| | | | | | | by using a .tfdbg_history file under user's home directory. Previously, each instance of CommandHistory held ephemeral command history, which goes away as soon as, e.g., a run-start UI or run-end UI exists. Change: 146739723
* contrib/android: Add a RunStats interface.Gravatar Asim Shankar2017-02-06
| | | | | | | | | | | Add a RunStats Java class that is a limited version of what StatSummarizer in C++ does. Between this and the rest of the Java API, I think we have the required feature coverage to switch the implementation of TensorFlowInferenceInterface from custom JNI to being pure-Java, implemented on top of the full Java API. Change: 146738405
* Add nccl_py to the contrib build file.Gravatar A. Unique TensorFlower2017-02-06
| | | | Change: 146738266
* Split input related estimator tests into a new file to fix timeout issue.Gravatar Mustafa Ispir2017-02-06
| | | | Change: 146735638
* Check size before passing the size to TensorArray.Gravatar Yuefeng Zhou2017-02-06
| | | | Change: 146731752
* scalar_summary changed to summary.scalarGravatar Olivia Nordquist2017-02-06
| | | | | | summarywriter link updated -- it's deprecated but this points to the correct one tensorboard tutorial renamed Change: 146728429
* [Docs] Distributed Execution: added missing links to classes.Gravatar Derek Murray2017-02-06
| | | | Change: 146725170
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2017-02-06
| | | | Change: 146722046
* Bump the website and the dependency to use protobuf 3.2.0.Gravatar Jonathan Hseu2017-02-06
| | | | | This version removes the 64 MB default limit for serialized proto sizes. Change: 146721509
* [Docs] Fix a few incorrect names and missing links in queues HOWTO.Gravatar Derek Murray2017-02-06
| | | | Change: 146718366
* Better deprecation warnings for the losses move.Gravatar Alexandre Passos2017-02-06
| | | | Change: 146717512
* Add a way to build a fused graph def to run the graph in a remote processorGravatar A. Unique TensorFlower2017-02-06
| | | | Change: 146717131
* [XLA] Get correct path to FileCheck when TEST_SRCDIR is not passed.Gravatar Justin Lebar2017-02-06
| | | | Change: 146715468
* making it exactly like the code snippetGravatar Olivia Nordquist2017-02-06
| | | | Change: 146713025
* Use MonitoredTrainingSession in distributed how-to.Gravatar Mustafa Ispir2017-02-06
| | | | Change: 146712044
* [XLA] On GPUs, lower pow(A, .5) as sqrt(A) and pow(A, -.5) as 1/sqrt(A).Gravatar Justin Lebar2017-02-06
| | | | Change: 146699987
* Fixes a bug where we were trying to access weight tensor for inference. This ↵Gravatar Zakaria Haque2017-02-06
| | | | | | | broke export/inference for any classification model using weights. + Added a test. Change: 146697049
* [XLA] Add CodegenTestBase::set_fast_math_disabled.Gravatar Justin Lebar2017-02-06
| | | | | | This lets us control whether fast-math is enabled during tests, now that fast-math is not controlled via flag. Change: 146696974
* [XLA] Fix parameter type of IsLiteralWithValue in algebraic_simplifier.cc.Gravatar Justin Lebar2017-02-06
| | | | | It should match the param type of LiteralUtil::IsAll. Change: 146692488
* [TF:XLA] Remove duplicate main() from randomized_tests.Gravatar Peter Hawkins2017-02-06
| | | | Change: 146690299
* [XLA] Add LiteralUtil::IsAllFloat.Gravatar Justin Lebar2017-02-06
| | | | | | This is like LiteralUtil::IsAll, but for floating-point values. Used in a later patch. Change: 146687678
* Fixes a couple of mis-formatted error messages.Gravatar Adam Roberts2017-02-06
| | | | Change: 146684854
* Streamlined representation of PendingCounts data structure.Gravatar Jeffrey A. Dean2017-02-06
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | (1) We now allocate a contiguous array of memory to hold count information for each node in a frame, allocating either 8 bits to hold a PendingCounts::PackedCounts structure (if the dead and pending count fields can fit in 8 bits), or an aligned region of 64 bits to hold a PendingCounts::LargeCounts structure (if the counts are too large to fit in the 8-bit representation). The byte offset and representation of the data for a given node are now held in a 32-bit handle. This makes all the count data contiguous for a given frame (better cache locality), and avoids the need for the FlatMap that we were using to hold a mapping from id to LargeCounts for those nodes that needed the overflow space. (2) Sped up the fast path of adjustments needed when normal non-merge nodes complete by adding a specialized "adjust_for_activation" routine that does all the necessary bookkeeping in one routine, rather than having to manipulate the count data structure several times. (3) Several tensorflow/core/common_runtime/executor.cc changes related to this: . Stored the new PendingCounts::Handle for a given node in the ItemState data structure, rather than in a separate frame_local_ids_ array. This means that we touch one less cache line per edge in ActivateNodes. . Changed initialization code to use the new PendingCounts::AllocateHandle routine for allocating space to hold the counts for a node. This allowed simplifying the ControlFlowInfo that is computed at initialization (we just need the unique set of frame names now, rather than a mapping from frame name to number of nodes in that frame). (4) Updated tensorflow/core/common_runtime/pending_counts_test.cc to reflect the new interface and to test the new compound adjust_for_activation routine. Speeds up an image processing benchmark on my desktop from 3370 images/sec to 3540 images/sec (+5.0%). Change: 146684811
* tfdbg stepper: display last updated variablesGravatar Shanqing Cai2017-02-06
| | | | | | * Add public last_updated() method to NodeStepper class. * The Stepper CLI uses this new method to inform the user of what variables (if any) are updated in the last cont/step action. Change: 146683089
* Comment out some of the more expensive microbenchmark cases for tf.multinomial.Gravatar Zongheng Yang2017-02-06
| | | | Change: 146677928
* Explicitly colocate the read with resource handle.Gravatar Alexandre Passos2017-02-06
| | | | | This ensures slots end up colocated on the right device. Change: 146660303
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2017-02-06
| | | | Change: 146642486
* Fix typo in documentation of KMeansClustering.Gravatar A. Unique TensorFlower2017-02-06
| | | | Change: 146639240
* Make legacy_seq2seq seq2seq models perform a **deep** copy of the incoming cell.Gravatar Eugene Brevdo2017-02-05
| | | | | Shallow copies are not enough when using a MultiRNNCell. Change: 146611566
* Allow lookup_table_ops to have bools as values.Gravatar A. Unique TensorFlower2017-02-05
| | | | Change: 146608995
* Set up directory structure to open source distributed tensorflow ↵Gravatar Brennan Saeta2017-02-04
| | | | | | benchmarking scripts Change: 146578260
* tfdbg: stepper: use dumped intermediate tensors during cont() callsGravatar Shanqing Cai2017-02-04
| | | | | | | Each cont() call will use TFDBG's own RunOptions.debug_options to generate intermediate tensor dumps and load the dumps and cached the relevant DebugTensorDatum objects. If future cont() calls require the intermediate tensor values, they will be obtained from the cached DebugTensorDatum objects, saving unnecessary recomputation. The use of such cached intermediate tensor dumps can be disabled in individual cont() calls by using "use_dumped_intermediates=False". Change: 146572672
* Go: Check in generated code for ops.Gravatar Asim Shankar2017-02-04
| | | | Change: 146572093
* Update SavedModel half_plus_two data and generation code.Gravatar Sukriti Ramesh2017-02-04
| | | | Change: 146565394
* Small change to new doc generator.Gravatar A. Unique TensorFlower2017-02-04
| | | | Change: 146564734
* Disable broken tensorboard test in windows cmake build.Gravatar Gunhan Gulsoy2017-02-04
| | | | Change: 146552777
* A few changes to get tf_cc_logged_benchmark's running under tensorflow/. ↵Gravatar A. Unique TensorFlower2017-02-04
| | | | | | Pass in benchmark argument to tf_cc_logged_benchmark. Also, add _gpu suffix to benchmarks. Change: 146552100
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2017-02-03
| | | | Change: 146548294
* Unit tests for logistic, OneHotCategorical, RelaxedBernoulli, ↵Gravatar A. Unique TensorFlower2017-02-03
| | | | | | RelaxedOneHotCategorical distributions. Fixed some bugs in the implementations that caused tests not to pass. Using numpy.finfo(np_dtype).tiny to prevent underflow issues. Change: 146548214
* Do not log test results using the text format, the JSON format is sufficient.Gravatar A. Unique TensorFlower2017-02-03
| | | | Change: 146539043
* Replace calculations in scheduled sampling code with gather/scatter_nd.Gravatar Eugene Brevdo2017-02-03
| | | | | | gather has a GPU kernel, and scatter_nd will soon have a GPU kernel. this should allow more of the calculations to stay on the GPU. Change: 146538474
* This implemented Jeff's idea to avoid most of the refcount related atomic ↵Gravatar Yuan Yu2017-02-03
| | | | | | operations when propagating tensors to the input slots of its consumers. Change: 146536099
* TensorBoard: Migrate vz_{sorting,line_chart} to webfilesGravatar Justine Tunney2017-02-03
| | | | Change: 146535501
* Add TODOs on several REGISTER_OP calls that should get deprecated.Gravatar A. Unique TensorFlower2017-02-03
| | | | | | Remove HistogramAccumulatorSummary from hidden_ops.txt and logging_ops.py; this op no longer exists. Change: 146534447
* Go: Enable custom headers in files generated by genop.Gravatar Asim Shankar2017-02-03
| | | | Change: 146532572
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-03 22:45:01 +0000