aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Merge changes from github.Gravatar Shanqing Cai2017-04-22
| | | | Change: 153925676
* Fix python3 build caused by recent RNNCell refactor.Gravatar Eugene Brevdo2017-04-21
| | | | Change: 153902768
* Copy function def to the optimized graph in the autoparallel optimizer.Gravatar Yuefeng Zhou2017-04-21
| | | | Change: 153896372
* RNNCell is now a subclass of tf.layers._Layer.Gravatar Eugene Brevdo2017-04-21
| | | | | | DO NOT CHERRYPICK INTO 1.1 BRANCH. This should only be released as part of TensorFlow 1.2. Change: 153891187
* Support multi-dimension weights for multi-class metrics.Gravatar A. Unique TensorFlower2017-04-21
| | | | | | Fix weighted loss. Fix some lint errors. Change: 153887559
* Add --tf_debug option flag to save_model_cliGravatar Shanqing Cai2017-04-21
| | | | Change: 153873095
* Python 3 compatibility bugfix in tensorboard/scripts/generate_testdata.py.Gravatar A. Unique TensorFlower2017-04-21
| | | | Change: 153863140
* Fix typo in 'density'.Gravatar A. Unique TensorFlower2017-04-21
| | | | Change: 153861823
* [XLA] Remove unused parameter in copy insertion pass.Gravatar A. Unique TensorFlower2017-04-21
| | | | Change: 153861680
* Fix some ClangTidy warnings in third_party/tensorflow/core/common_runtime.Gravatar Derek Murray2017-04-21
| | | | Change: 153861629
* Introduce TFDecorator, a base class for Python TensorFlow decorators. ↵Gravatar Charles Nicholson2017-04-21
| | | | | | Provides basic introspection and "unwrap" services, allowing tooling code to fully 'understand' the wrapped object. Change: 153854044
* Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2017-04-21
| | | | Change: 153851292
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2017-04-21
| | | | Change: 153850549
* [TF:XLA]Gravatar A. Unique TensorFlower2017-04-21
| | | | | This flag is unnecessary, as we can get the path to ptxas from tensorflow::CudaRoot(). Change: 153847918
* Add Mfcc op to TensorFlow for speech feature generationGravatar Pete Warden2017-04-21
| | | | Change: 153847440
* Add GrpcDebugWrapperSession to the public debug API.Gravatar A. Unique TensorFlower2017-04-21
| | | | Change: 153843609
* Branch the projector so that we have a d3v4 version with the new typings.Gravatar Nikhil Thorat2017-04-21
| | | | Change: 153837827
* Automated rollback of change 153736477Gravatar A. Unique TensorFlower2017-04-21
| | | | Change: 153825726
* Switch DirectSession to use _Arg and _Retval ops for feeding and fetching.Gravatar Derek Murray2017-04-20
| | | | | | | | | This change reduces the overhead imposed by string processing and rendezvous invocation in the DirectSession::Run() call by 1--2 microseconds per value fed or fetched. RELNOTES: Improved DirectSession::Run() overhead and error checking. Feeding a value of the wrong type will now synchronously raise an INVALID_ARGUMENT error instead of asynchronously raising an INTERNAL error. Code that depends on the (undefined) behavior when feeding a tensor of the wrong type may need to be updated. Change: 153797943
* [XLA] Fixes some div-by-zero bugs.Gravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153795265
* - Bug fix in ShapeInferenceGravatar Bjarke Hammersholt Roune2017-04-20
| | | | | | - CommaSeparatedString and VectorString added to xla_util.h - ReferenceUtil can now do more general Pad ops. Change: 153782516
* Autogenerated Change: Change TensorBoard TAG to 54Gravatar Dandelion Mané2017-04-20
| | | | Change: 153780747
* Fix serious bug which broke TensorBoard when there are many scalar runs and ↵Gravatar Dandelion Mané2017-04-20
| | | | | | few text runs. Change: 153780710
* Fix _get_arguments to work with partials.Gravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153776958
* Fix typo in "First TensorFlow program" from open source READMEGravatar Stephan Hoyer2017-04-20
| | | | | This is sort of pedantic, but the Python prompt prints strings with quotes. Change: 153772799
* Add unreduced NONE, and reduced MEAN options for losses.Gravatar A. Unique TensorFlower2017-04-20
| | | | | Remove "WEIGHTED_" prefix from other Reduction constants. Change: 153758104
* [tf contrib seq2seq] Changes to dynamic decodingGravatar Eugene Brevdo2017-04-20
| | | | | | | | | | | | 1. dynamic_decode now returns a third parameter: the sequence lengths from decoding (minibatch entries that finished earlier have a shorter sequence length) 2. beam search decoder now uses the gather_tree C++ op 3. the gather_tree c++ op now expects sequence_length to be a matrix shaped `[batch_size, beam_width]` (each beam may have its own sequence length). Change: 153756869
* [XLA] Fix the parameter instruction printing issueGravatar A. Unique TensorFlower2017-04-20
| | | | | | | | Append the parameter number to the fusion parameter name, and use the parameter name rather the instruction name in creating the new parameter. Show the paramameter number when printing out parameter instructions. Change: 153752424
* Add a subgraph recomputation helper function to Grappler's memory optimzerGravatar A. Unique TensorFlower2017-04-20
| | | | | No optimizations make use of it yet. Change: 153751713
* Fix missing assertion methods in core_rnn_cell_test.pyGravatar Shanqing Cai2017-04-20
| | | | | The assertEmpty and assertNontEmpty methods do not appear to be universally available across the environments that TF is tested on. Change: 153751634
* Fix const-ness of master_env.Gravatar Brennan Saeta2017-04-20
| | | | | | | | A recent change in grpc_server_lib removed const from the master_session_factory function. This change fixes master_env.h to accurately reflect the new type signature. Change: 153751309
* Add more tests for broadcast_shape.Gravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153748425
* Renamed kernel for AudioSpectrogram op to match op nameGravatar Pete Warden2017-04-20
| | | | Change: 153747523
* Added missing support for uint16 type in the queues.Gravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153741673
* Android: Create TensorFlow AAR as part of nightly build, so that it will be ↵Gravatar Andrew Harp2017-04-20
| | | | | | | | | | | | | | | | | | | | downloadable at https://ci.tensorflow.org/view/Nightly/job/nightly-android/ This contains TensorFlowInferenceInterface and the Java API, as well as all of the native prebuilt libraries. This means that TF can be integrated into an Android Studio app simply by downloading the AAR file to e.g. aar/, and then adding the following into a gradle build file: allprojects { repositories { jcenter() flatDir { dirs 'aar' } } } dependencies { compile(name:'tensorflow', ext:'aar') } Change: 153741338
* Expose tf.contrib.staging.StagingArea.Gravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153739304
* [XLA] Fix incorrect commentsGravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153737501
* Automated rollback of change 153709951Gravatar Shanqing Cai2017-04-20
| | | | Change: 153736477
* Fix graph_transform documentation:Gravatar A. Unique TensorFlower2017-04-20
| | | | | | | | - remove \ from within strings - remove :0 from inputs and outputs, so fold_constants works - make sure fold_(old_)batch_norms runs before quantize_weigths and round_weights. Change: 153728959
* Mention RDMA support in the release notes.Gravatar Jonathan Hseu2017-04-20
| | | | Change: 153726368
* Update nccl to a newer version, removed patching of nccl on download.Gravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153722905
* Automated rollback of change 153358570Gravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153718592
* Create a per session inter op thread pool instead of relying on a global threadGravatar Benoit Steiner2017-04-20
| | | | | pool: this ensures that the thread pool is reset whenever the session is reset. Change: 153718067
* Internal C API headers.Gravatar Alexandre Passos2017-04-20
| | | | Change: 153717143
* Adding python_configure.bzl to (partially) replace python_config.shGravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153709951
* [XLA:HLO] Also clone metadata when cloning instructions e.g. in fusion.Gravatar A. Unique TensorFlower2017-04-20
| | | | | Without the metadata, it's hard to correlate HLO instructions to TF ops after fusion. Change: 153709552
* [TF-XLA] Include TF metedata in HLO GraphDef node's attributes.Gravatar A. Unique TensorFlower2017-04-20
| | | | | This includes TF op names and op types. Change: 153708854
* removing unused variable LLVM_HOST_COMPILER_PATHGravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153707385
* Properly aggregate memcpyH2DAsync in the "per node type" summaries in ↵Gravatar A. Unique TensorFlower2017-04-20
| | | | | | StatSummarizer. Change: 153705191
* Clarify docstring: tf.Variable(variable_def) does not add to the graph.Gravatar A. Unique TensorFlower2017-04-20
| | | | Change: 153680516
generated by cgit v1.2.3 (git 2.39.1) at 2024-06-29 02:20:10 +0000