aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Remove race in coordinator_test so it passes on tsan. Reduce sleep intervalsGravatar A. Unique TensorFlower2017-02-17
| | | | | so it runs faster (11s on tsan instead of ~30s). Change: 147893428
* Instructions on contributing to C++ gradients.Gravatar Suharsh Sivakumar2017-02-17
| | | | Change: 147892613
* Split out most of k8s_tensorflow into a library and add a way to pass anyGravatar A. Unique TensorFlower2017-02-17
| | | | | | environment variables. Add benchmark_util library that would use environemnt variable to decide on a storage location. Change: 147890534
* Swap NaN count from index 7 to 2 within DebugNumericSummary ops.Gravatar A. Unique TensorFlower2017-02-17
| | | | Change: 147888410
* Add a check if hexagon is unlocked into a script to build tf runtime with ↵Gravatar A. Unique TensorFlower2017-02-17
| | | | | | hexagon Change: 147886830
* Made tensors produced in a while loop unfeedable so we will give a more ↵Gravatar Yuan Yu2017-02-17
| | | | | | informative error message. Use queues or TensorArray to feed tensors into a while loop. Change: 147884461
* Enable tf.strided_slice GPU kernel for tf.complex64.Gravatar RJ Ryan2017-02-17
| | | | Change: 147881489
* [XLA] Support 1D and 3D convolutions in GPU backend.Gravatar A. Unique TensorFlower2017-02-17
| | | | | 1D convolutions are executed as 2D convolution with an extra singleton dimension. Change: 147877435
* Making GetNextWorkLocked a virtual function so that sub classes could ↵Gravatar Rohan Jain2017-02-17
| | | | | | override it and do something different if needed. Change: 147875810
* Add compute_time and memory_time to CostGraphDef::Node.Gravatar Yao Zhang2017-02-17
| | | | Change: 147871989
* Check if adb is installed in a sample script for hexagonGravatar A. Unique TensorFlower2017-02-17
| | | | Change: 147867469
* [XLA]Improve HLO rematerialization cost to account for duplication.Gravatar Mark Heffernan2017-02-17
| | | | | | | | A rematerialized instruction is cloned one or more times, and the original rematerialized instruction may be deleted if no uses remain. Previously this duplication (or reduction) of total computation was not accounted for in the rematerialization cost analysis. Change: 147864843
* Go: Run tests during continuous integrationGravatar Asim Shankar2017-02-17
| | | | | | | | | | | TensorFlow uses bazel to build and test. However, the TensorFlow Go API is targeted for use with the 'go' tool. This commit: - Adds a shell test so that usage with the 'go' tool can be tested with 'bazel test //tensorflow/go/...' - Installs Go in the images used in the continuous build Change: 147864583
* Add a test to build tensorflow runtime with hexagon into android_full.shGravatar A. Unique TensorFlower2017-02-17
| | | | Change: 147861747
* Java: Update READMEs to link to Javadoc and release instructions.Gravatar Asim Shankar2017-02-17
| | | | Change: 147858835
* Avoid creating a tf.Saver() when there are no variables to restore.Gravatar Sergio Guadarrama2017-02-17
| | | | Change: 147858700
* Only record cuDNN profiling results if the operation was successful.Gravatar A. Unique TensorFlower2017-02-17
| | | | Change: 147853458
* Have TensorBoard reload data more often.Gravatar Dandelion Mané2017-02-17
| | | | | Have the backend reload 5 seconds after the last reload finishes (instead of 60) and have the frontend refresh every 30 seconds. Change: 147848903
* Add "output_layer" argument to the BasicDecoder.Gravatar Eugene Brevdo2017-02-17
| | | | Change: 147845195
* Add shard_count to affine_test.Gravatar A. Unique TensorFlower2017-02-17
| | | | Change: 147844228
* tfdbg: let BaseDebugWrapperSession support graph_def and as_default()Gravatar Shanqing Cai2017-02-17
| | | | Change: 147837972
* Change gradients to be computed with respect to variable ref rather thanGravatar A. Unique TensorFlower2017-02-16
| | | | | | | | | | snapshot. Variables may create another snapshot or their ref may be exposed via public API (e.g., var.op.outputs[0] or graph.as_graph_element(var) which happens fairly often inside libraries or collection serialization). On the other hand, tf.gradients() use convert_to_tensor() which returns a snapshot, and gradients were computed with respect to this particular snapshot, which makes the gradients incorrect. Change: 147800865
* Removing note about TF 1.0 alpha, now that 1.0 has officially been released.Gravatar Sanders Kleinfeld2017-02-16
| | | | Change: 147788449
* Support softmax on hexagon and remove some custom ops from graph transfererGravatar A. Unique TensorFlower2017-02-16
| | | | Change: 147788426
* Android demo: add sample images to README.md and update model installation ↵Gravatar Andrew Harp2017-02-16
| | | | | | instructions. Change: 147787804
* Go: Add PartialRun support.Gravatar Asim Shankar2017-02-16
| | | | Change: 147783087
* Upgrade Gulp TypeScriptGravatar Justine Tunney2017-02-16
| | | | Change: 147781771
* Add std:: prefix to STL non-type names.Gravatar A. Unique TensorFlower2017-02-16
| | | | Change: 147779725
* Task 6: sealing up bayesflow.{entropy, monte_carlo, variational_inference}Gravatar Olivia Nordquist2017-02-16
| | | | Change: 147778589
* Workaround FlatMap windows build breakage.Gravatar A. Unique TensorFlower2017-02-16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Windows builds break on the following simplified example: template <typename Key, ...> class FlatSet { public: typedef Key value_type; class const_iterator { public: typedef FlatSet::value_type value_type; // Fails on windows }; }; The build succeeds by adding 'typename': typedef typename FlatSet::value_type value_type; // OK on windows <simplified log of compiler error> flatmap.h(110): warning C4346: 'difference_type': dependent name is not a type (compiling source file ...cancellation.cc) flatmap.h(110): note: prefix with 'typename' to indicate a type (compiling source file ...cancellation.cc) flatmap.h(161): note: see reference to class template instantiation 'tensorflow::gtl::FlatMap<Key,Val,Hash,Eq>::iterator' being compiled (compiling source file ...cancellation.cc) flatmap.h(376): note: see reference to class template instantiation 'tensorflow::gtl::FlatMap<Key,Val,Hash,Eq>' being compiled (compiling source file ...cancellation.cc) flatmap.h(110): error C2061: syntax error: identifier 'difference_type' (compiling source file ...cancellation.cc) </simplified log of compiler error> This is a bug in the windows compiler. It is true that FlatSet::value_type is a dependent name, but it also refers to the "current instantiation", so 'typename' shouldn't be required. For details see: http://en.cppreference.com/w/cpp/language/dependent_name#Current_instantiation But it doesn't hurt to add typename; it is simply redundant. Change: 147776071
* Deprecate `Bijector.inverse_and_inverse_log_det_jacobian`.Gravatar Joshua V. Dillon2017-02-16
| | | | Change: 147770857
* Changed the second argument of WaitForNotificationWithTimeout() to mean ↵Gravatar Yuan Yu2017-02-16
| | | | | | microseconds from milliseconds. Change: 147764063
* Upgraded Eigen to avoid a compilation warning on MacOSGravatar Benoit Steiner2017-02-16
| | | | Change: 147763615
* Add `Sigmoid` bijector.Gravatar Joshua V. Dillon2017-02-16
| | | | Change: 147761442
* Added visibility to development area.Gravatar A. Unique TensorFlower2017-02-16
| | | | Change: 147758266
* Blacklist already-failing tests in GPU PIP buildsGravatar Shanqing Cai2017-02-16
| | | | Change: 147757405
* Temporarily disable large graph tests that may be causingGravatar Vijay Vasudevan2017-02-16
| | | | | timeouts on test infra. Change: 147757378
* Add compute_output_shape method to tf.layers objects.Gravatar Eugene Brevdo2017-02-16
| | | | | | | | This is a port of the Keras get_output_shape_for layer method; and the name change was discussed with Francois first. Implemented this method for the tf.layers.Dense class. Change: 147753001
* Set a more descriptive error when a Run() call is cancelled due to session ↵Gravatar Derek Murray2017-02-16
| | | | | | | | closure. Also fixes a potential memory leak, where the worker-side state of a failed Run() call would not be cleaned up. Change: 147752067
* Integrate SyncReplicasOptimizer with Estimators.Gravatar Mustafa Ispir2017-02-16
| | | | | There is an hidden dependency between when 'apply_gradient' and get_chief_queue_runner() are called. This cl postpones creation of the queue to the initialization of the Session. In Estimator, Session is created after forming the graph/training-op. That means it is after the apply_gradient. Change: 147746938
* [TF:XLA] Refactor XlaOpRegistry, moving metadata about how to compile ↵Gravatar Peter Hawkins2017-02-16
| | | | | | | operators on a device into a struct. No functional changes. Change: 147741833
* Add default values to max_message_size of grpc APIs. They have changed at head.Gravatar A. Unique TensorFlower2017-02-16
| | | | Change: 147738820
* C API: Tests and cleanup for partial run.Gravatar Asim Shankar2017-02-16
| | | | | | | | | - Add a unittest for TF_SessionPRun - Add TF_DeletePRunHandle. This provides a way for callers to safely delete the handle (without knowing the details of the allocator used by TF_SessionPRunSetup) and also provides a placeholder for cancellation of PRun state in the future. Change: 147738191
* Android demo: Add example images for inclusion in README.mdGravatar Andrew Harp2017-02-16
| | | | Change: 147737818
* Temporarily disable feeder_test because it's flaky.Gravatar Alexandre Passos2017-02-16
| | | | Change: 147734454
* Make FlatSet and FlatMap behave more like the std equivalents.Gravatar A. Unique TensorFlower2017-02-16
| | | | | | | | | | | | | | | | | | | | | | | | | | This fixes a collection of annoyances that prevented Flat{Set,Map} from being a pure drop-in replacement for std::unordered_{set,map}. I haven't exhaustively ensured compatibility; I started with stuff I've actually run into: * Construction via std::initializer_list FlatSet<int> set({1, 2, 3}); FlatMap<int, int> map({{1, 10}, {2, 20}, {3, 30}}); * Define an iterator_category for compatibility with std algorithms. E.g. the code below would yield a compile error with the useful bit truncated, and you'd need to find and set --cxxopt=-fshow-overloads=all to figure it out. FlatSet<int> set({1, 2, 3}); vector<int> vec(set.begin(), set.end()); // used to fail, now works * Defining the iterators with forward_iterator_tag requires postfix ++ Admittedly I haven't actually run into this, but it's easy to add. * Previously FlatSet::iterator allowed mutation of the set keys, which would corrupt the internal representation. I've taken the standard approach of defining iterator as an alias of const_iterator, so that we'll now get a compile error if you mistakenly do this. I haven't actually run into this either, but it seems like a worthwhile change. Change: 147731813
* Expanding the optimization that avoids making an extra copy when restoring ↵Gravatar A. Unique TensorFlower2017-02-16
| | | | | | slices to include the case when the source and destination have the same shape and are full slices with respect to their shape. Change: 147726529
* [XLA] Guard rematerialization warning with proper conditionGravatar Mark Heffernan2017-02-16
| | | | | | | | Previously rematerialization issued a warning unconditionally that rematerialization failed to reduce memory below the specified limit. With this change, the warning is emitted only if rematerialization actually failed. Change: 147724519
* Remove debug_ops from kernels:arrayGravatar Shanqing Cai2017-02-15
| | | | Change: 147678642
* Removes exports_files(["LICENSE"]) which exports a non-existent LICENSE file.Gravatar A. Unique TensorFlower2017-02-15
| | | | Change: 147676615
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-01 02:46:57 +0000