aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* tfdbg CLI: add filtering options to the run commandGravatar Shanqing Cai2017-03-27
| | | | | | | | | | | | This is useful for limiting the amount of debug tensor data dumped and/or focusing the debugging on specific parts of the graph. For example: tfdbg> run --node_name_filter hidden tfdbg> run --op_type_filter Variable.* tfdbg> run --tensor_dtype_filter int.* RELNOTES: The "run" command of tfdbg's command-line interface now supports filtering of tensors by node name, op type and tensor dtype. Change: 151405236
* Create a new tempdir for each test to fix the race condition for sharded tests.Gravatar Jonathan Hseu2017-03-27
| | | | Change: 151403279
* Java: Update documentation with instructions for Maven and Windows.Gravatar Asim Shankar2017-03-27
| | | | | | Fixes #7877 Fixes #6926 Change: 151400818
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2017-03-27
| | | | Change: 151399581
* Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2017-03-27
| | | | Change: 151396171
* Added WAV audio file reading and writing operationsGravatar Pete Warden2017-03-27
| | | | Change: 151395519
* Avoid pruning fetch nodesGravatar Benoit Steiner2017-03-27
| | | | Change: 151394592
* Add an optional argument to set capacity of individual bucket queues in ↵Gravatar Lukasz Kaiser2017-03-27
| | | | | | bucket and bucket_by_sequence_length. Change: 151393852
* Add kernel_methods to the bigger contrib module, and indirectly to the pipGravatar Gunhan Gulsoy2017-03-27
| | | | | package. Change: 151393424
* [contrib seq2seq] Allow AttentionWrapper to store its attention history.Gravatar Eugene Brevdo2017-03-27
| | | | | | Not enabled by default, but enable it with construction flag attention_history=True. Change: 151391741
* Internal change.Gravatar A. Unique TensorFlower2017-03-27
| | | | Change: 151391358
* Turned the meta optimizer into a regular optimizerGravatar Benoit Steiner2017-03-27
| | | | Change: 151390706
* Added a new regression test to ensure control dependencies are properlyGravatar Benoit Steiner2017-03-27
| | | | | processed Change: 151383375
* Don't refer to non-existent license filesGravatar Benoit Steiner2017-03-27
| | | | Change: 151382854
* Use a initializer to initialize the RMSProp optimizer's slots.Gravatar Peter Hawkins2017-03-27
| | | | Change: 151380006
* Enable access to SecondOrStepTimer via tf.train.Gravatar Mustafa Ispir2017-03-27
| | | | Change: 151375546
* Preserve graphdef metadata when extracting a subgraph.Gravatar A. Unique TensorFlower2017-03-27
| | | | Change: 151374696
* [contrib seq2seq] Rename DynamicAttentionWrapper to AttentionWrapperGravatar Eugene Brevdo2017-03-27
| | | | Change: 151374177
* [XLA:CPU] Disable Dot strength reduction due to performance regressions. ThisGravatar Blake Hechtman2017-03-27
| | | | | issue was reported by a customer that uses XLA:CPU in production. Change: 151370263
* added more checks for experiment.Gravatar Jianwei Xie2017-03-27
| | | | Change: 151366403
* Update hexagon supported op list and fix hexagon execution sample testsGravatar A. Unique TensorFlower2017-03-27
| | | | Change: 151366249
* Output per-node-type timings to proto in benchmark_modelGravatar Andrew Harp2017-03-27
| | | | Change: 151366175
* Feed the test input with random numbers instead of constants to preventGravatar Benoit Steiner2017-03-27
| | | | | constant folding from wiping out a large part of the test network Change: 151354452
* Add support for SavedModel on Android.Gravatar A. Unique TensorFlower2017-03-27
| | | | Change: 151352935
* Android demo: Update how Gradle build dependencies are set up to resolve #8587Gravatar Andrew Harp2017-03-27
| | | | Change: 151351473
* Forgotten unref in py_func.Gravatar Alexandre Passos2017-03-27
| | | | Change: 151349771
* Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2017-03-27
| | | | Change: 151345424
* Prune unecessary Identity nodesGravatar Benoit Steiner2017-03-27
| | | | Change: 151344092
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2017-03-27
| | | | Change: 151341806
* Fix op description for BatchMatMul to state that it works for >= 2D, not > 2D.Gravatar Peter Hawkins2017-03-27
| | | | Change: 151341608
* Added UGRNN and Intersection RNN cells from the paper: Capacity and ↵Gravatar A. Unique TensorFlower2017-03-27
| | | | | | Trainability in Recurrent Neural Networks Change: 151339833
* Update DynamicAttentionWrapper's output_size to return _cell's output_size ifGravatar A. Unique TensorFlower2017-03-27
| | | | | output_attention is False. Change: 151337979
* Automated rollback of change 151331705Gravatar A. Unique TensorFlower2017-03-27
| | | | Change: 151335029
* py_func no longer copies outbut buffer (attempt #2)Gravatar Alexandre Passos2017-03-27
| | | | | Missed an incref the first time. Change: 151332384
* Comment change s/contrib.layer/contrib.layers/Gravatar A. Unique TensorFlower2017-03-27
| | | | Change: 151332329
* Remove debugging log lines.Gravatar Brennan Saeta2017-03-27
| | | | Change: 151331999
* Prune unecessary Identity nodesGravatar Benoit Steiner2017-03-27
| | | | Change: 151331705
* sample_stats module added to distributions/. Starting by adding percentile, ↵Gravatar Ian Langmore2017-03-27
| | | | | | roughly equivalent to np.percentile Change: 151317484
* Remove unnecessary instance variable.Gravatar Brennan Saeta2017-03-26
| | | | | requested_port_ is only used within a single function, and thus would be better as a local variable. Change: 151285269
* Add a test documenting custom_getter behaviour when reuse=True.Gravatar Lukasz Kaiser2017-03-26
| | | | Change: 151282782
* Add int32 version of Mul kernel to the slim kernels.Gravatar Manjunath Kudlur2017-03-26
| | | | Change: 151267767
* This CL adds:Gravatar Petros Mol2017-03-25
| | | | | | - an approximate kernel mapper for the RBF kernel (based on Random Fourier Features) - a (canned) tf.learn kernel-based classifier. Change: 151237967
* Consolidate worker state behind a session-centric abstraction.Gravatar Brennan Saeta2017-03-25
| | | | | | | | | | | | | | | | | | | | | | State in workers is currently splayed across graph_mgr, rendezvous_mgr, and additional components. This has resulted in it being difficult to ensure proper cleanup and shut down of the worker components. In addition to paving the way for a more reliable shut down, this CL also sets up the beginnings of ClusterSpec propagation. ClusterSpec propagation is a capability upgrade for TensorFlow that should make it much easier to (1) build distributed TensorFlow clusters, and (2) handle node failures. After the ClusterSpec propagation capability is fully implemented, the TensorFlow workers can be booted independently of each other, and with no knowledge about others. A client can then query a central cluster scheduler or other API to find all of the workers, and then send the ClusterDef (ClusterSpec) to the TF master, which then propagates that along to all of the workers. This change is only the first of a sequence to fully implement ClusterSpec propagation in TensorFlow. Change: 151229111
* Automated rollback of change 151188187Gravatar A. Unique TensorFlower2017-03-24
| | | | Change: 151201476
* Added a new OpPerformanceList messageGravatar Benoit Steiner2017-03-24
| | | | Change: 151191936
* py_func return values are no longer copied.Gravatar Alexandre Passos2017-03-24
| | | | Change: 151188187
* Expose models in contrib.cudnn_rnnGravatar Skye Wanderman-Milne2017-03-24
| | | | Change: 151185292
* Move the core API & implementation of the TF-Serving batching library to ↵Gravatar Christopher Olston2017-03-24
| | | | | | tensorflow/contrib/. (Leave shims in the old location for now.) Change: 151184823
* [XLA] Rematerialize subcomputations.Gravatar Mark Heffernan2017-03-24
| | | | | | | Extend HLO rematerialization to rematerialize subcomputations in addition to the entry computations. Outer nesting levels of computations are rematerialized before inner nesting levels because inner subcomputations may be while bodies where rematerialization is more expensive. Also Also fix latent bug in call_graph dealing with fusion instructions, and extend HloInstruction::Clone to accept a string suffix (eg, "remat") for the clone name. Change: 151179956
* Adds an AllocatorWrapper class.Gravatar Vinu Rajashekhar2017-03-24
| | | | | | - Useful for allocators which want to override only some functionality of another allocator. Change: 151179248
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-05 21:13:08 +0000