aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Eliminate per-op allocation of gpu device wrapper. The PerOpGpuDevice is ↵Gravatar A. Unique TensorFlower2016-01-24
| | | | | | allocated once in the OpKernelContext::Params struct, then re-used every time a new OpKernelContext uses the Params. Thus in the executor, as long as there is more work to do the PerOpGpuDevice is not freed. Change: 112909215
* Improve error message when asking to fetch the output of a node withGravatar Josh Levenberg2016-01-24
| | | | | no outputs. Fixes #856 Change: 112903800
* Unbreak license testGravatar Geoffrey Irving2016-01-24
| | | | Change: 112903292
* Make shape inference more scalar strictGravatar Geoffrey Irving2016-01-24
| | | | | | | | We still haven't advanced to the scalar strict GraphDef version, but this change will prevent some (not all) new Python scripts from violating scalar strictness. That is, it will significantly reduce the amount of bit rot between now and when I finally get full scalar strictness submitted. Change: 112866199
* make cc ops publicGravatar Josh Levenberg2016-01-24
| | | | Change: 112846663
* Delete duplicate #include.Gravatar Josh Levenberg2016-01-24
| | | | Change: 112846648
* Use an iterator to walk down the set of free chunks, so that deletingGravatar A. Unique TensorFlower2016-01-22
| | | | | | one from the set when we find one we want is cheaper. Slight performance improvement (~0.3% on ptb_word_lm model on my desktop). Change: 112832451
* Fixes for OSS build.Gravatar Vijay Vasudevan2016-01-22
| | | | Change: 112830289
* Pin constant holding data to CPU. This makes GPU version 50-100x faster per ↵Gravatar Yaroslav Bulatov2016-01-22
| | | | | | step. Fixes #837 Change: 112829709
* Implement a promise-based RequestManager class, will be the workhorse of a ↵Gravatar A. Unique TensorFlower2016-01-22
| | | | | | new TensorFlow backend. Change: 112826468
* Disentangle the GPU code from the CPU code. This means a few things:Gravatar Josh Levenberg2016-01-22
| | | | | | | | | | | | | | | | | | * The "core_cpu_internal" build target no longer includes files from the common_runtime/gpu/ directory. * tensorflow/core internal targets instead can get access to those headers via the "gpu_runtime" target. * The class "CopyTensor" is introduced. It lives in common_runtime/ but supports registration of copy functions so the "gpu_runtime" target can add a GPU->GPU copy ability if it is linked in. This registration should make it easier to add more device types in the future. * The "core_cpu" and "core_cpu_internal" build targets no longer reference GPUUtil::CopyViaDMA; rendezvous_mgr uses CopyTensor instead. Also the "copy_tensor" build target was not needed. Change: 112821119
* Improved the performance of tensor contractions and spatial convolutions on CPU:Gravatar Benoit Steiner2016-01-22
| | | | | | * Move some checks out of inner loops * Split the mapper in 2: a base mapper, and a sub-mapper. This reduces the number of variables that are contained in the base mapper and helps reduce register spills Change: 112809881
* Added a script for rough estimation of the computation cost and parameter ↵Gravatar Pete Warden2016-01-22
| | | | | | size of a CNN. Change: 112809773
* Add symbolic gradient function for Softmax.Gravatar A. Unique TensorFlower2016-01-22
| | | | Change: 112808739
* Fixes a bug in Queue cancellation.Gravatar Derek Murray2016-01-22
| | | | | | | | | | If multiple steps are blocked on the same queue, there is a high chance that their cancellation tokens will collide, since these are dense and start at 0. This can lead to the wrong step being cancelled. This change additionally uses the `CancellationManager*` to uniquely identify an attempt for the purposes of cancellation. Change: 112801810
* Rollback of "Adding histogram_ops module containing one Op: ↵Gravatar Vijay Vasudevan2016-01-22
| | | | | | histogram_fixed_width, which updates a histogram Variable with new_values." Change: 112800368
* Change uses of TensorShape::ShortDebugString to DebugStringGravatar Geoffrey Irving2016-01-22
| | | | | | The two functions already have the same behavior, and ShortDebugString will disappear soon. Change: 112793490
* Make cuda_dnn work with R4.Gravatar Xiaoqiang Zheng2016-01-22
| | | | Change: 112755081
* Add TensorArray data_flow_ops wrapper and TensorArray gradients.Gravatar Eugene Brevdo2016-01-21
| | | | | These are undocumented features and the API can and will change. Change: 112733605
* Add TODO to fix the situation where core_cpu_internal includesGravatar Josh Levenberg2016-01-21
| | | | | all .h from common_runtime/gpu/. Change: 112732380
* Adding histogram_ops module containing one Op: histogram_fixed_width, which ↵Gravatar A. Unique TensorFlower2016-01-21
| | | | | | updates a histogram Variable with new_values. Change: 112728652
* Move #include <vector> out of port.h to users of std::vector<>.Gravatar Josh Levenberg2016-01-21
| | | | | After this we can replace port.h with types.h. Change: 112727463
* Export ToGraphDef() from function.ccGravatar A. Unique TensorFlower2016-01-21
| | | | | | The standard Graph::ToGraphDef function does not correctly handle the graphs produced by TensorFlow function expansion, e.g. inlining often produces duplicate node names. This CL also adds a parameter which can be used to produce more legible node names which help when viewing the GraphDef in the TensorFlow graph visualizer. Change: 112719928
* Need to check if the op is in a loop context.Gravatar A. Unique TensorFlower2016-01-21
| | | | Change: 112709520
* Fix a test for Python 3Gravatar Geoffrey Irving2016-01-21
| | | | Change: 112703190
* Drop unused gemmlowp dependency.Gravatar Josh Levenberg2016-01-21
| | | | Change: 112699191
* Fix and re-enable tests for matrix_triangular_solve op.Gravatar A. Unique TensorFlower2016-01-21
| | | | Change: 112688071
* Unspecified message.Gravatar A. Unique TensorFlower2016-01-21
|
* Unspecified message.Gravatar A. Unique TensorFlower2016-01-21
|
* Unspecified message.Gravatar A. Unique TensorFlower2016-01-21
|
* Unspecified message.Gravatar A. Unique TensorFlower2016-01-21
|
* Unspecified message.Gravatar A. Unique TensorFlower2016-01-21
|
* Change: 112644520Gravatar Manjunath Kudlur2016-01-20
|
* Rollback of "Updated TensorBoard logic for discarding events after crashes ↵Gravatar Vijay Vasudevan2016-01-20
| | | | | | to discard out of order events only within a particular tag. This was changed because race conditions in supervisor was causing many events to be unintentionally discarded." Change: 112644077
* Change: 112641949Gravatar Manjunath Kudlur2016-01-20
|
* Change: 112640197Gravatar Manjunath Kudlur2016-01-20
|
* Unspecified message.Gravatar A. Unique TensorFlower2016-01-20
|
* Change: 112636955Gravatar Manjunath Kudlur2016-01-20
|
* Prepare to hide tf.tensor_utilGravatar Geoffrey Irving2016-01-20
| | | | | | | | | | | | | | | | | 1. There is a new tf.unsupported module to hold things which some people use but which we don't yet support. 2. tf.tensor_util.ConstantValue is now tf.unsupported.constant_value. Most users use this, but tf.tensor_util.ConstantValue is still available; it will be removed in a following commit. 3. tensor_util.MakeTensorShapeProto is now make_tensor_shape_proto. It looks like all users of this access the tensor_util module directly (not through tf), so for now it is not in unsupported. This commit does not remove tensor_util from tf.__all__; a few more downstream users must be changed before that can happen. Change: 112626961
* First step of moving files out of tensorflow/core/public/. HereGravatar Josh Levenberg2016-01-20
| | | | | | | we copy the original files to their new location and make the public/ versions #include the new location. Once all references are updated to point to the new location, we can delete the originals in public/. Change: 112622561
* Improve exception message when `TensorShape.with_rank()` failsGravatar Derek Murray2016-01-20
| | | | | | | | | | | Previously this would pass on the exception message from `TensorShape.merge_with()`, which is cryptic for users who don't (and shouldn't need to) understand how shape inference works. This would arise, for example, in the error message for `tf.matmul()` when passed a non-matrix, as noted here: http://stackoverflow.com/questions/34908033/tensorflow-exception-with-matmul Change: 112621185
* Adds `tf.decode_json_example()` op that parses JSON into binary Example protos.Gravatar Derek Murray2016-01-20
| | | | Change: 112615357
* Removes some unintended change.Gravatar A. Unique TensorFlower2016-01-20
| | | | Change: 112611994
* More function support.Gravatar A. Unique TensorFlower2016-01-20
| | | | Change: 112611228
* Rearrange BUILD into sections, and in particular put the targets inGravatar Josh Levenberg2016-01-20
| | | | | | | | | | the public section in a sensible order (smallest to largest, matching the documentation comment). Plus: * Update documentation to reflect that test_main is not public. * Remove "friends" package_group now that it is unused. Change: 112605117
* Change: 112595580Gravatar Manjunath Kudlur2016-01-20
|
* update docsGravatar Martin Wicke2016-01-20
| | | | Change: 112591828
* Change DebugString behavior to ShortDebugStringGravatar Geoffrey Irving2016-01-20
| | | | | | | | | The old behavior of DebugString is needlessly verbose and is quite confusing for scalar shapes (it produced the empty string). Now DebugString is the same as ShortDebugString. A future commit will remove ShortDebugString. Change: 112590646
* Many tensorflow/core build clean ups.Gravatar Josh Levenberg2016-01-20
| | | | Change: 112523833
* Add support for Refs in control flow ops (Merge, Switch, Enter, Exit).Gravatar Eugene Brevdo2016-01-20
| | | | Change: 112505342
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-05 04:01:10 +0000