aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
...
* [TF:XLA] Improvements to resource variables:Gravatar Peter Hawkins2017-02-27
| | | | | | | | | | | | | | * enable compilation of VarIsInitializedOp. * fix deprecated variable initializer in variable_ops_test.py * simplify variable logic in XlaContext, move intelligence into XlaOpKernelContext. * add resource variable support in the contrib layers library. Cleanups and refactorings: * merge XlaCompiler::CompileSubComputation with XlaCompiler::CompileFunction. * pass XlaCompiler arguments consistently via XlaCompiler::Options. * split the two roles of XlaCompiler::CompilationResult::input_shapes into input_mapping and xla_input_shapes. * initialize the numpy and Python seeds to a constant for XLA test cases. Change: 148683645
* Refactor the buffer forwarding code:Gravatar A. Unique TensorFlower2017-02-27
| | | | | | | | * Moves the core forwarding logic to a new function forward_input that is not restricted to forwarding to an output slot, but also allows, e.g., reusing input buffers as temporaries or variables. * Gets rid of forward_input_to_output_with_same_shape that is now unused. * Adds convenience methods input_memory_type(), output_memory_type, and (private) input_is_ref() to OpKernelContext. * Misc. small cleanups. Change: 148683484
* We've moved from using the docs in g3doc/ to docs_src/ files, whichGravatar A. Unique TensorFlower2017-02-27
| | | | | get additional processing before being published. Change: 148682342
* Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2017-02-27
| | | | Change: 148682151
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2017-02-27
| | | | Change: 148681977
* Fix assert error in evaluation_test on windows.Gravatar Jianwei Xie2017-02-27
| | | | Change: 148681573
* Bugfix: Ensure tf.random_gamma can never return 0. Fixes #7474.Gravatar Joshua V. Dillon2017-02-27
| | | | Change: 148678769
* Move cluster initialization to happen during training using training hooksGravatar A. Unique TensorFlower2017-02-27
| | | | | | | | | instead of depending on variable initialization logic. This avoids deadlocks that can happen if QueueRunners are used to generate data. Deadlocks would happen since initialization would wait for QueueRunner to generate data, but QueueRunners were not started till variable initialization was complete. Also some refactoring of the code to move Hooks and model function out of the Estimator class. Change: 148678573
* Fix spelling errors.Gravatar Patrick Nguyen2017-02-27
| | | | Change: 148678164
* Fix a bug where queue_parsed_examples fails for single tensors.Gravatar A. Unique TensorFlower2017-02-27
| | | | Change: 148676738
* [TF:XLA] Refactor build targets in //tensorflow/compiler/jit, split ops into ↵Gravatar Peter Hawkins2017-02-27
| | | | | | | more traditional ops/ and kernels/ directories. No functional changes. Change: 148673534
* - Upgraded libxsmm to 1.7.1.Gravatar A. Unique TensorFlower2017-02-27
| | | | | | | | | | | | | - Applied LLVM optimization patch to libxsmm (https://github.com/hfp/libxsmm/commit/0e412d5d2769a8754cace64e56e26e14093f887d.patch). - Limited outstanding libxsmm sparse matrix multiply handle counts to limit memory usage for temporary space. - Added extra logging to libxsmm handle management in TensorFlow. - Added support for running multiple sparse matrix multiplies simultaneously in performance benchmark to match some practical use cases. - Added more size combinations to sparse matrix multiply benchmark. - Fixed dependencies for xsmm_conv2d_test. Change: 148672973
* [XLA] Set layout of fused instructions in `ReshapeMover`Gravatar A. Unique TensorFlower2017-02-27
| | | | Change: 148671742
* [TF:XLA] Silence a number of compiler warnings, in particular warnings ↵Gravatar Peter Hawkins2017-02-27
| | | | | | | repeatedly issued by code in headers. Fixes #7919 Change: 148670735
* Measure feed/fetch overhead on CPU only, to focus on Session-related overhead.Gravatar Derek Murray2017-02-27
| | | | Change: 148669949
* Add the job's DeviceSet to the options for optimization passes that run in ↵Gravatar A. Unique TensorFlower2017-02-27
| | | | | | the session master. Change: 148669552
* [XLA] Add HLO call graph.Gravatar Mark Heffernan2017-02-27
| | | | | | | | Add an HLO call graph class. The object includes a node for each computation and forward and backwards links between them. Node include the calling context: either parallel (eg, kMap) or sequential (eg, kCall). The class is not used anywhere yet, but there are numerous potential uses throughout the HLO level. Change: 148669295
* Migrate tf-audio-dashboard to webfiles.Gravatar Dandelion Mané2017-02-27
| | | | Change: 148668422
* Internal-only changesGravatar Shanqing Cai2017-02-27
| | | | Change: 148668420
* Remove the libtensorflow_c.so target.Gravatar Asim Shankar2017-02-27
| | | | | | | | | | | | | | | | | | This was identical to libtensorflow.so for all but a few days in the life of libtensorflow_c.so's existence. Remove libtensorflow_c.so to remove any confusion about which target to use for the TensorFlow C library. (https://github.com/tensorflow/tensorflow/commit/783c52e and #3830 were the changes that resulted in two targets doing the same thing) I have made attempts to remove use of libtensorflow_c.so in other projects such as: https://github.com/LaurentMazare/tensorflow-ocaml/pull/19 https://github.com/tensorflow/rust/pull/56 https://github.com/malmaud/TensorFlow.jl/pull/157 Change: 148665838
* Move estimator.evaluate into core.Gravatar A. Unique TensorFlower2017-02-27
| | | | Change: 148664638
* C API: Keep the TF_LoadSessionFromSavedModel symbol in Android builds.Gravatar Asim Shankar2017-02-27
| | | | | | | Though, the function will return an error. Required for #7688 Change: 148664204
* Added a benchmark for low-level `Session::Run()` overheads.Gravatar Derek Murray2017-02-27
| | | | Change: 148659423
* Set the clock rate for the host accuratelyGravatar David Majnemer2017-02-27
| | | | | | We used CLOCKS_PER_SEC which is defined only in terms of the clock system call. What we needed to use was the hardware clock rate. Change: 148655561
* Stop the gradients on the sampled_values of the compute_sampled_logits.Gravatar A. Unique TensorFlower2017-02-27
| | | | Change: 148653993
* Internal only changeGravatar A. Unique TensorFlower2017-02-27
| | | | Change: 148647686
* Disable map inliner on CPU due to bug and add two tests that demonstrate the ↵Gravatar Bjarke Hammersholt Roune2017-02-26
| | | | | | bug. Change: 148607148
* Use std::unordered_set instead of std::set in DifferentialSet.Gravatar Jeffrey A. Dean2017-02-25
| | | | Change: 148563356
* Move numpy_input_fn and pandas_input_fn from contrib to core.Gravatar Jianwei Xie2017-02-25
| | | | Change: 148560715
* Automated rollback of change 148492669Gravatar Manjunath Kudlur2017-02-25
| | | | Change: 148552367
* Refine the UI of health pills.Gravatar A. Unique TensorFlower2017-02-24
| | | | | Changed colors used to render health pills. Added a drop shadow with an SVG gaussian filter (CSS does not suffice) to make the light gray (for the positive values category) stand out. Rewrote rendering logic to use a discrete gradient for rendering health pill colors to allow for the filter. Change: 148522653
* Temporarily disable buffer forwarding for GPU memory since we don't see a ↵Gravatar A. Unique TensorFlower2017-02-24
| | | | | | large gain in performance and potentially introduce bugs in Cuda kernels that rely on reads from read-only (texture cache) memory. Change: 148520565
* Make doc's generate.py into a library.Gravatar A. Unique TensorFlower2017-02-24
| | | | Change: 148513339
* Move estimator.fit into core.Gravatar Mustafa Ispir2017-02-24
| | | | Change: 148507023
* [XLA] Test uniformity of uniform discrete RNGGravatar A. Unique TensorFlower2017-02-24
| | | | Change: 148501408
* [XLA] Make reshape mover appropriately modify shapes of fused parametersGravatar A. Unique TensorFlower2017-02-24
| | | | Change: 148500449
* Add a registry for remote fused graph executor to remove dependency from ↵Gravatar A. Unique TensorFlower2017-02-24
| | | | | | RemoteFusedGraphExecuteOp to GraphTransferer Change: 148496466
* Rollback the change to depend on header only protobuf library.Gravatar Manjunath Kudlur2017-02-24
| | | | Change: 148492669
* Typo in reader v1 backwards compatibility code.Gravatar Alexandre Passos2017-02-24
| | | | Change: 148490238
* Add remote_fused_graph_execute_info to abstract hexagon specific parametersGravatar A. Unique TensorFlower2017-02-24
| | | | Change: 148489487
* Fix typo in tf.contrib.layers.batch_norm, which wasn't adding gamma to ↵Gravatar A. Unique TensorFlower2017-02-24
| | | | | | custom variables collections (when supplied). Change: 148488709
* Automated rollback of change 148480357Gravatar A. Unique TensorFlower2017-02-24
| | | | Change: 148488129
* [XLA] Add an IsFinite operation that tests elementwise whether values are ↵Gravatar Peter Hawkins2017-02-24
| | | | | | finite (i.e., not NaN or Inf). Change: 148485205
* Add sort_by_execution_order to transforms used by eight-bit-calculationsGravatar A. Unique TensorFlower2017-02-24
| | | | Change: 148481141
* Moved common code out of RNN Estimators.Gravatar A. Unique TensorFlower2017-02-24
| | | | Change: 148480357
* Remove the data download option from TensorBoard.Gravatar Dandelion Mané2017-02-24
| | | | | The data download links are buggy, and complicated to use. Change: 148478960
* Move InputFnOps to core as ServingInputReceiverGravatar David Soergel2017-02-24
| | | | Change: 148477564
* Update module docstrings to (consistently) link to the guide in theGravatar A. Unique TensorFlower2017-02-24
| | | | | | body instead of the title (consistently). Also fix some malformed @{$...} references and titles starting with "##". Change: 148476930
* Open source HParams object and proto.Gravatar A. Unique TensorFlower2017-02-24
| | | | Change: 148474914
* Fix the open source test.Gravatar Jianwei Xie2017-02-24
| | | | Change: 148470369
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-01 15:51:54 +0000