aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Adds a GuaranteeConstOp.Gravatar Vinu Rajashekhar2017-12-01
| | | | | | - Acts as indicator for the TF runtime to make possible optimizations by treating the input tensor as a constant. PiperOrigin-RevId: 177656212
* Fixing the python 3.6 build error.Gravatar Amit Patankar2017-12-01
| | | | PiperOrigin-RevId: 177655994
* Eliminate matrix multiplication with zeros.Gravatar A. Unique TensorFlower2017-12-01
| | | | PiperOrigin-RevId: 177655417
* Remove SliceProcessorConcatOffset, which is not robust as it modifies nodesGravatar Yao Zhang2017-12-01
| | | | | | | | which could be used elsewhere in the graph; SliceProcessorConcatOffset is a historical implementation anyway, and the same functionality could be provided by more recently developed SliceProcessor and SliceProcessorConst. PiperOrigin-RevId: 177653218
* closes #14981Gravatar Yifei Feng2017-12-01
| | | | PiperOrigin-RevId: 177652538
* Rename create_summary_file_writer to create_file_writerGravatar Justine Tunney2017-12-01
| | | | PiperOrigin-RevId: 177651937
* Add BF16 tests for reshape.Gravatar Yuanzhong Xu2017-12-01
| | | | PiperOrigin-RevId: 177647906
* Prefix inaccurate costs with "~" in VirtualScheduler verbose log.Gravatar Max Galkin2017-12-01
| | | | | | | | | | Fix some inaccurate estimates exposed by this approach: - propagate the inaccuracy flag when merging device stats; - estimate Const as no-op; - estimate RandomUniform, Relu and Softmax as element-wise; - consider estimates accurate for known element-wise ops in op_level_cost_estimator. PiperOrigin-RevId: 177643976
* Automated g4 rollback of changelist 177633858Gravatar Gunhan Gulsoy2017-12-01
| | | | PiperOrigin-RevId: 177640956
* Adds a ReleaseHandle method to the FunctionLibraryRuntime interface that ↵Gravatar Rohan Jain2017-12-01
| | | | | | | | allows for releasing the state associated with the handle. Also simplifies the state owned by the FunctionLibraryRuntimeImpl. Instead of having a vector of ref counted Item objects and a separate vector of function bodies, we merge it into one object that holds the entire instantiated state for the function. PiperOrigin-RevId: 177639560
* TF Eager: Add SPINN model example for dynamic/recursive NN.Gravatar Shanqing Cai2017-12-01
| | | | PiperOrigin-RevId: 177636427
* Increase tolerance in conv_ops_test to avoid flakes.Gravatar Gunhan Gulsoy2017-12-01
| | | | PiperOrigin-RevId: 177633993
* Add `tf.contrib.distributions.Autoregressive`.Gravatar Joshua V. Dillon2017-12-01
| | | | PiperOrigin-RevId: 177633858
* Remove non-exposed copy of old SummaryWriterGravatar Nick Felt2017-12-01
| | | | | RELNOTES: N/A PiperOrigin-RevId: 177631104
* Corrects documentation about what the mel processing expects as input.Gravatar A. Unique TensorFlower2017-12-01
| | | | PiperOrigin-RevId: 177629372
* Update the comment in speech_hotword_model_test.Gravatar A. Unique TensorFlower2017-12-01
| | | | PiperOrigin-RevId: 177625918
* Log the error status for failed CUDA EnablePeerAccess.Gravatar A. Unique TensorFlower2017-12-01
| | | | | | | This would help debug issues like this: #14759 PiperOrigin-RevId: 177625164
* [tpu:profiler] Add host-independent and host-dependent job info to the ↵Gravatar A. Unique TensorFlower2017-12-01
| | | | | | run environment in tf_op_stats.proto. PiperOrigin-RevId: 177624829
* Automated g4 rollback of changelist 177614538Gravatar Gunhan Gulsoy2017-12-01
| | | | PiperOrigin-RevId: 177619402
* [TF:XLA] Remove unused class XlaLocalRuntimeContext and related member ↵Gravatar A. Unique TensorFlower2017-12-01
| | | | | | | | | variables. This additional parameter to computations was only needed for the CPU backend, but it is not used anymore. The thread pool it holds is already passed via xla::ExecutableRunOptions. Also, remove unused members from XlaCompiler::CompilationResult. PiperOrigin-RevId: 177618693
* Adds -latomic as linkopts to support uses of ISO C++11 <atomic>.Gravatar A. Unique TensorFlower2017-12-01
| | | | PiperOrigin-RevId: 177616799
* Make 'producer_op_list' argument to import_graph_def work with C API.Gravatar Skye Wanderman-Milne2017-12-01
| | | | | | | | | | This changes exposes a minor functional change: unknown attrs cause an import error with the C API enabled, but only cause an error when the graph is run in the original Python code (the error will occur even if the problematic op isn't being run). It seems OK to fail faster in error cases. PiperOrigin-RevId: 177615841
* Add `dtype` keyword argument to `_EagerTensorBase`'s `__array__` method.Gravatar Akshay Agrawal2017-12-01
| | | | | | This change makes `__array__` conform to numpy's API. PiperOrigin-RevId: 177614929
* Customize request timeouts for GCS filesystem.Gravatar Brennan Saeta2017-12-01
| | | | PiperOrigin-RevId: 177614538
* Make TopologicalSort return an error status if the sorting fails.Gravatar A. Unique TensorFlower2017-12-01
| | | | PiperOrigin-RevId: 177612830
* Internal ChangeGravatar A. Unique TensorFlower2017-12-01
| | | | PiperOrigin-RevId: 177611775
* Remove stale comment.Gravatar Nick Desaulniers2017-12-01
| | | | PiperOrigin-RevId: 177611737
* Moved compute_pi from utils.py to fisher_blocks.py and added global constant ↵Gravatar A. Unique TensorFlower2017-12-01
| | | | | | that controls the type of pi-based damping adjustment to use. Current options are "tracenorm", which is what we were already doing and is the default, and "off". PiperOrigin-RevId: 177610677
* Seed the time series LSTM example unit test.Gravatar Allen Lavoie2017-12-01
| | | | PiperOrigin-RevId: 177606245
* Add `cross_entropy` and `kl_divergence` to `tf.distributions.Distribution`.Gravatar Joshua V. Dillon2017-12-01
| | | | PiperOrigin-RevId: 177605983
* Store and log request URI when requests hang.Gravatar Brennan Saeta2017-12-01
| | | | PiperOrigin-RevId: 177596564
* Split out a "graph" library from "core_cpu" to allow finer-grained dependencies.Gravatar Derek Murray2017-12-01
| | | | | | | This is a step towards supporting custom Dataset ops built as external libraries. PiperOrigin-RevId: 177595688
* 1) Make `tensor_pool` support a list of input_values, so we can store ↵Gravatar A. Unique TensorFlower2017-12-01
| | | | | | | | (generated_input, generated_data) tuples in the pool. 2) Rename `tensor_pool` function to `random_tensor_pool` so the function name does not collide with the pkg name. PiperOrigin-RevId: 177594443
* TFE: Remove contrib/eager/python/sumary_writer.pyGravatar Shanqing Cai2017-12-01
| | | | | | Use tf.contrib.summary.create_file_writer, instead. PiperOrigin-RevId: 177588097
* Build demo app for SmartReplyGravatar A. Unique TensorFlower2017-12-01
| | | | PiperOrigin-RevId: 177559103
* Adds a fisher block for fully connected recurrent layers.Gravatar A. Unique TensorFlower2017-11-30
| | | | | | | | `FullyConnectedSeriesFB` uses an approximation to the Fisher information matrix designed for RNNs. This CL also adds support for dtypes other than `float32` to `fisher_factors`. PiperOrigin-RevId: 177558080
* Add additional linkopts argument to tf_custom_op_library.Gravatar A. Unique TensorFlower2017-11-30
| | | | PiperOrigin-RevId: 177555877
* [XLA] Hlo parser: support rng and reduce-precision. Also simplify the lexer ↵Gravatar A. Unique TensorFlower2017-11-30
| | | | | | by regarding several things as identifier. PiperOrigin-RevId: 177548483
* Use latest nsync in tensorflow. Latest nsync builds with bazel on FreeBSD.Gravatar A. Unique TensorFlower2017-11-30
| | | | PiperOrigin-RevId: 177545934
* Disable tuning for now. Re-enable when measurement-based estimator is ready.Gravatar Yao Zhang2017-11-30
| | | | PiperOrigin-RevId: 177545499
* enabling Tensor._set_shape() to work with the C APIGravatar Olivia Nordquist2017-11-30
| | | | PiperOrigin-RevId: 177543170
* Automated g4 rollback of changelist 177505909Gravatar Yao Zhang2017-11-30
| | | | PiperOrigin-RevId: 177540002
* Register more ops with bfloat16 types.Gravatar Peter Hawkins2017-11-30
| | | | PiperOrigin-RevId: 177537667
* Output unknown dimension root nodes with --vmodule=graph_properties=2Gravatar Max Galkin2017-11-30
| | | | PiperOrigin-RevId: 177535370
* Support compressed TensorProto format in constant folding for types iny16, ↵Gravatar A. Unique TensorFlower2017-11-30
| | | | | | | | int8, uint8, and bool, in addition to float ,double, int32, and int64, which were already supported. Add unit test for all types. PiperOrigin-RevId: 177533200
* Merge changes from github.Gravatar Sourabh Bajaj2017-11-30
| | | | PiperOrigin-RevId: 177526301
* [TF:XLA] Use output spatial dimensions instead of a transpose for convGravatar David Majnemer2017-11-30
| | | | | | backwards filter PiperOrigin-RevId: 177522710
* Add int64 support to XLA Shape op.Gravatar A. Unique TensorFlower2017-11-30
| | | | PiperOrigin-RevId: 177519992
* [TF:XLA] Make tf_cnn_benchmarks run on CPU with XLA.Gravatar A. Unique TensorFlower2017-11-30
| | | | | | Adds _cpu_jit to tf_cnn_benchmarks_xla BUILD rule and fixes an issue in XLA bridge triggered by XLA CPU compilation of whole graphs. In particular, modifies mark_for_compilation_pass.cc to skip _Retval nodes when looking for compilation candidates in the top level function. _Retval nodes are introduced in the input subgraph as a replacement for fetches. Including _Retval nodes into XLA clusters confuses encapsulate subgraph pass that expects a graph with no pre-existing _Retval nodes. PiperOrigin-RevId: 177518178
* [XLA] Sanitize hlo names to match regexp "[a-zA-Z_][a-zA-Z0-9_.-]*".Gravatar A. Unique TensorFlower2017-11-30
| | | | PiperOrigin-RevId: 177518046
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-28 02:07:54 +0000