aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Moves eval op from dnn_linear_combined to TargetColumn.Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127504952
* Automated rollback of change 127470901Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127493261
* Automated rollback of change 127463353Gravatar Andrew Selle2016-07-14
| | | | Change: 127492429
* Changes for C++ shape inference usability:Gravatar A. Unique TensorFlower2016-07-14
| | | | | | | - add Scalar,Vector,Matrix shorthands and use them in existing fns. - rename CreateUnknownShape to UnknownShape. - add Divide and Add functions. Change: 127487121
* Get all the gradients for a while_loop properly colocated with the ops.Gravatar Eugene Brevdo2016-07-14
| | | | | | | | | **Note**: Noam originally found this bug and wrote this solution. All credit goes to him. The change solves the problem of overloading the master device when there are lots of separate dynamic_rnns each on a different device all in the same graph. Change: 127481156
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127480049
* batch_matrix_set_diag now used in batch_matrix_diag_transform.Gravatar Ian Langmore2016-07-14
| | | | Change: 127478534
* Backported some changes to the reduction code from upstream Eigen to keep ↵Gravatar Benoit Steiner2016-07-14
| | | | | | the code in sync. Change: 127477364
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127474696
* Add C++ shape inference function to ImmutableConstGravatar Vijay Vasudevan2016-07-14
| | | | Change: 127474274
* batch_matrix_transpose added to array_ops. This transposes the last twoGravatar Ian Langmore2016-07-14
| | | | | dimensions of a Tensor. Convenience method. Change: 127474063
* Fix nightly docker build failure due to multiple wheel filesGravatar Shanqing Cai2016-07-14
| | | | | | | | | With CL/127111314, there can be multiple wheel files (different names, same content) generated in the pip_test/whl directory. But parameterized_docker_build.sh assumes a single file, which has caused nightly docker build failures such as: http://ci.tensorflow.org/view/Nightly/job/nightly-docker-cpu/58/ http://ci.tensorflow.org/view/Nightly/job/nightly-docker-gpu/55/ This CL fixes that. Change: 127473464
* Add support for DequeueUpTo in RandomShuffleQueue and uses this to add the ↵Gravatar A. Unique TensorFlower2016-07-14
| | | | | | option to allow small final batch sizes to shuffle_batch and shuffle_batch_join. Change: 127472371
* Fix a bug in _DNNLinearCombinedBaseEstimator.linear_weights_.Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127470901
* Don't skip evaluation if latest_path is None.Gravatar A. Unique TensorFlower2016-07-14
| | | | | When save_checkpoints_secs is small (note that the smaller this value is, the more frequent we checkpoint), say 1 second, there is a race condition that could lead to a None latest_path and None self._latest_path_step, which further results to an error, when the file "checkpoint" is read at the same time it is being updated. Change: 127470621
* Remove unused dependencies from ↵Gravatar A. Unique TensorFlower2016-07-14
| | | | | | third_party/tensorflow/contrib/framework/BUILD tests. Change: 127470132
* Set 'partitioner' correctly while creating var scope for wide layer weights.Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127468219
* Enable strided slice op as default slicing.Gravatar Andrew Selle2016-07-14
| | | | | | | | | | | | | | | | | This allows more rich slicing of tensors i.e. tf.assign(var, foo[-5:-10:-1, ..., 3:5, tf.newaxis]) This does not include lvalue support for assigning to a slice. It also does not include advanced indexing foo[bar] where bar is a tensor. Fixed bug in implementation where num_elements was used instead of dims for an optimization code path. Also make supervisor have an easier to read error message for dim mismatches. Change: 127463353
* Fix sampling_ops test timeout by making it a CPU-only test.Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127460861
* Allow sampling_ops to batch a list of tensors, rather just one.Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127458134
* Add InferenceContext::MakeShapeFromShapeProto(const TensorShapeProto& proto);Gravatar A. Unique TensorFlower2016-07-14
| | | | | | | | Rename CreateShapeXYZ to MakeShapeXYZ. Rename CreateDim to MakeDim. Rename CreateUnknownDim to UnknownDim. Rename CreateDimForScalarInput to MakeDimForScalarInput. Change: 127457122
* Cleanup towards requiring explicit export signatures.Gravatar David Soergel2016-07-14
| | | | | | | | | | | | | | | | In this CL: 1. Remove export_logistic_regression_estimator. 2. Deprecate internal graph_actions.export_for_serving(). 3. Issue warning when signature_fn is not supplied. In future CLs: 4. Delete internal graph_actions.py (together with model.py). 5. Add signature_fn to callers where needed-- maybe via a helper function that performs the current inference, issuing a warning. 6. Require signature_fn argument to export_estimator. 7. Redirect internal usages of export_estimator. Change: 127455493
* Improve error handling for scalar arguments to sparse_xentGravatar Pete Warden2016-07-14
| | | | Change: 127453531
* Modify tf-image-dashboard so that it breaks long run names at the top.Gravatar Dan Mané2016-07-14
| | | | | Fixes: https://github.com/tensorflow/tensorflow/issues/3302 Change: 127449137
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127449098
* Move broadcasting out of OperatorX. The OperatorX classes will now requireGravatar Ian Langmore2016-07-14
| | | | | | | input to follow the same (strict) shape requirements of TensorFlow batch matrix operations. Helper functions will reshape/brodcast vectors/matrices to make them compatible. Change: 127446489
* Allow batch_norm to use a is_training Tensor, which can be modified at run time.Gravatar A. Unique TensorFlower2016-07-14
| | | | | Added tf.Assert() to make sure the update_ops are not executed when is_training is False. Change: 127443556
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127443499
* Deprecate old skflow functions.Gravatar Martin Wicke2016-07-14
| | | | Change: 127441867
* Attempt to fix iris_val_based_early_stopping test.Gravatar A. Unique TensorFlower2016-07-14
| | | | Change: 127441610
* Automated rollback of change 127366296Gravatar Eugene Brevdo2016-07-14
| | | | Change: 127440874
* Reuse WorkerInterface objects. Reduces the number of allocations inGravatar A. Unique TensorFlower2016-07-14
| | | | | | | | | | | | a benchmark from ~91M for ~81M. There are future cleanups possible but not yet done: (a) Disallow direct destruction of WorkerInterface and always release back to a cache of WorkerInterface objects. (b) Replace channel cache with just a factory that knows how to create a WorkerInterface and leave caching to the higher-level cache of WorkerInterface objects. Change: 127435209
* Refactoring: Use std::unique_ptr<> to maintain pointer ownership in more ↵Gravatar A. Unique TensorFlower2016-07-14
| | | | | | | | | places in tensorflow::DirectSession and tensorflow::SimpleClientGraph. Remove unused func_defs in ExecutorsAndKeys. Fixes a few memory leaks on error paths. Change: 127425821
* Modify tensorforest training_loss to take features and labels as input to ↵Gravatar A. Unique TensorFlower2016-07-14
| | | | | | accommodate data-dependent notions of loss. Change: 127424766
* Give names to training and inference ops in tensor_forest, which helps with ↵Gravatar A. Unique TensorFlower2016-07-13
| | | | | | integration with frameworks that identify them by name. Change: 127387066
* Enable complex GPU kernels for tf.matmul and tf.batch_matmul.Gravatar RJ Ryan2016-07-13
| | | | Change: 127386123
* Made sure that philox arrays are initialized properly.Gravatar Benoit Steiner2016-07-13
| | | | Change: 127385633
* Rename io directory to learn_io because non-absolute import can confuse this ↵Gravatar Martin Wicke2016-07-13
| | | | | | with the standard io module. Change: 127385608
* Include recent C graph-construction API in release notes.Gravatar A. Unique TensorFlower2016-07-13
| | | | Change: 127385333
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-07-13
| | | | Change: 127380299
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2016-07-13
| | | | Change: 127380001
* Add TF_NodeOutputListLength() and TF_NodeInputListLength() to C APIGravatar A. Unique TensorFlower2016-07-13
| | | | | (executing a TODO). Change: 127378577
* Add BatchMatrixWithDiag op & gradientGravatar Eugene Brevdo2016-07-13
| | | | Change: 127378533
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-07-13
| | | | Change: 127376094
* Add access to mock as tf.mockGravatar A. Unique TensorFlower2016-07-13
| | | | Change: 127374393
* Fix divide-by-zero in specialized version of RequantizeMany...UsingEigen.Gravatar A. Unique TensorFlower2016-07-13
| | | | Change: 127374168
* Move SDCAOptimizer from contrib/learn to contrib/linear_optimizerGravatar A. Unique TensorFlower2016-07-13
| | | | Change: 127373128
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-07-13
| | | | Change: 127371216
* Move computation in quantized_bias_add_op.cc to quantization_utils.h.Gravatar A. Unique TensorFlower2016-07-13
| | | | Change: 127370125
* Automated rollback of change 127259950Gravatar Vijay Vasudevan2016-07-13
| | | | Change: 127366296
generated by cgit v1.2.3 (git 2.39.1) at 2024-10-18 10:57:39 +0000