aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Use optimized functor for conjugate transpose in MatrixInverseOp.Gravatar A. Unique TensorFlower2017-10-17
| | | | | | | Introduce convenience functions DoMatrixTranspose and DoConjugateMatrixTranspose. Misc. minor cleanup of templates in transpose_functor*. PiperOrigin-RevId: 172532252
* Go: Update generated wrapper functions for TensorFlow ops.Gravatar A. Unique TensorFlower2017-10-17
| | | | PiperOrigin-RevId: 172531834
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2017-10-17
| | | | PiperOrigin-RevId: 172530775
* Add bitwise LeftShift (aka. tf.bitwise.left_shift) and RightShift ↵Gravatar Peter Hawkins2017-10-17
| | | | | | | | (tf.bitwise_right_shift) operators to Tensorflow. Their semantics are intended to mirror numpy.left_shift and numpy.right_shift. Fix a couple of missing cases of missing uint32/uint64 support exposed when adding tests for bitshifts. PiperOrigin-RevId: 172530375
* Add environment variable to enable setting of CUDA context flags.Gravatar A. Unique TensorFlower2017-10-17
| | | | PiperOrigin-RevId: 172527804
* Fix a bug in BFCAllocator::Log2FloorNonZero() under non gnuc non windows ↵Gravatar Guangda Lai2017-10-17
| | | | | | mode, where it returns the bit index of the most significant '1' bit, starting from 1, but the expected result should start from 0 (i.e. it should return floor(log2(n)) ). PiperOrigin-RevId: 172522745
* Bugfix: add `fill_triangular`, `reduce_weighted_logsumexp`, `tridiag` toGravatar Joshua V. Dillon2017-10-17
| | | | | | `tf.contrib.distributions.__init__`. PiperOrigin-RevId: 172521590
* Improve performance of tf.space_to_depth and tf.depth_to_space for typical ↵Gravatar A. Unique TensorFlower2017-10-17
| | | | | | block sizes of NCHW data layout on GPU with a loop kernel. PiperOrigin-RevId: 172520132
* Set estimator run_config default random seed to None. This will make it ↵Gravatar Mustafa Ispir2017-10-17
| | | | | | | | aligned with other parts of the TF. Many users are not aware of impact of non-random seed. For example it may lead to train only on a small fraction of training data due to preemptions. We're changing default behavior since we consider it as a bug fix. PiperOrigin-RevId: 172519268
* A new flag `ignore_live_threads` is available on train. If set to True, it ↵Gravatar A. Unique TensorFlower2017-10-17
| | | | | | will ignore threads that remain running when tearing down infrastructure after successfully completing training, instead of throwing a RuntimeError. PiperOrigin-RevId: 172518466
* Link to Datasets doc in release notesGravatar Mark Daoust2017-10-17
| | | | PiperOrigin-RevId: 172517507
* Preserve metadata_ when using HloInstruction::CloneWithNewOperands.Gravatar A. Unique TensorFlower2017-10-17
| | | | PiperOrigin-RevId: 172517300
* Make tensor_util.constant_value_as_shape compatible with Eager modeGravatar Akshay Agrawal2017-10-17
| | | | PiperOrigin-RevId: 172514789
* Remove remaining nomac tags for TFBT as the memory corruption issue is fixed.Gravatar A. Unique TensorFlower2017-10-17
| | | | PiperOrigin-RevId: 172512636
* Internal change.Gravatar A. Unique TensorFlower2017-10-17
| | | | PiperOrigin-RevId: 172511553
* Added an additional check on the length of the values and boundaries lists.Gravatar A. Unique TensorFlower2017-10-17
| | | | PiperOrigin-RevId: 172510229
* tfdbg: add missing -f flag of the pt command to the command tableGravatar Shanqing Cai2017-10-17
| | | | PiperOrigin-RevId: 172508350
* Automated g4 rollback of changelist 172051437Gravatar A. Unique TensorFlower2017-10-17
| | | | PiperOrigin-RevId: 172493077
* Fix TF workspace issue that prevents submodules from using aws build targets ↵Gravatar Kiril Gorovoy2017-10-17
| | | | | | correctly. Fixes https://github.com/tensorflow/serving/issues/615 PiperOrigin-RevId: 172489253
* Disable flaky gru_ops_test in opensource.Gravatar Gunhan Gulsoy2017-10-17
| | | | PiperOrigin-RevId: 172481341
* Create a macos py3 test script.Gravatar Gunhan Gulsoy2017-10-17
| | | | PiperOrigin-RevId: 172480793
* Reworks the imperative_grad interface.Gravatar Alexandre Passos2017-10-17
| | | | PiperOrigin-RevId: 172477878
* Disable flaky cluster tests in opensource.Gravatar Gunhan Gulsoy2017-10-17
| | | | PiperOrigin-RevId: 172477381
* [tf.contrib.seq2seq] Bugfixes to BeamSearchDecoder and GatherTree.Gravatar Eugene Brevdo2017-10-17
| | | | | | | | | | | | | | 1. Begin the gather tree at the maximum sequence length across all beams (within the batch). 2. Take a second pass starting from t=0 and mask out any beam ids past the *first* beam occurrence of end_token. 3. Update the final sequence lengths to include the first <eos> token in the beam. 4. Update dynamic_decode to allow the BeamSearchDecoder to keep track of its own "finished" states, as the shuffling in the decoder confused the tracking mechanism in dynamic_decode. This fixes a bug where beam search decoding stops early. 5. Cap sequence length used in GatherTree to min(max_time, max_seq_len(b)) to avoid accessing memory outside the dimensions of input matrices. Bugs caught by @bdaskalov on github and Pavel Sountsov. Proper solution and analysis thanks to Rui Zhao. Thanks all! Fixes #13536. PiperOrigin-RevId: 172471462
* HParams: deprecates function set_from_map in favor of new functionGravatar A. Unique TensorFlower2017-10-17
| | | | | | | | | | | override_from_dict Reasons to prefer new function name: - `set` sounds like it might return the builtin set. - There is no datatype `map` in python - it's a builtin, making the implied API a little confusing. PiperOrigin-RevId: 172471191
* Add the op->IsExpensive() argument to tracing calls.Gravatar A. Unique TensorFlower2017-10-16
| | | | PiperOrigin-RevId: 172422580
* Make Snappy header availableGravatar Justine Tunney2017-10-16
| | | | | | This is going to be useful for the tensor database I'm working on. PiperOrigin-RevId: 172412142
* Disable probable timeout flake on Ubuntu machines.Gravatar Vijay Vasudevan2017-10-16
| | | | PiperOrigin-RevId: 172408922
* math_grad: Fast path for when broadcasting is not needed.Gravatar A. Unique TensorFlower2017-10-16
| | | | PiperOrigin-RevId: 172407754
* Adding an ItemHandler that does lookups. This allows decoding of ↵Gravatar A. Unique TensorFlower2017-10-16
| | | | | | tf.Examples where IDs are not materialized (e.g. 'image/object/class/text' present but 'image/object/class/label' not). PiperOrigin-RevId: 172406978
* [XLA:GPU] Don't crash with --vmodule=gpu_compiler=2 if we can't run ptxas.Gravatar Justin Lebar2017-10-16
| | | | | | | | | | | | | | | At --vmodule=gpu_compiler=2, we run ptxas over our generated PTX, to validate it, and also to dump out stats like the number of registers used. But previously, this would fail if your GPU was anything other than sm_35 (i.e. K20/40/80), because we didn't pass down cc_major/cc_minor to ptxas. And moreover, if ptxas failed to compile your program, we'd LOG(FATAL), which is probably no what you want. This change fixes both those issues. Tested on my local GTX1080. PiperOrigin-RevId: 172403304
* Better error message for eager-specific APIsGravatar A. Unique TensorFlower2017-10-16
| | | | PiperOrigin-RevId: 172397124
* Uses head.name in name_scope. This improves the graph naming for MultiHead.Gravatar A. Unique TensorFlower2017-10-16
| | | | PiperOrigin-RevId: 172389494
* Add return_nodes option to ImportGraphDefGravatar Skye Wanderman-Milne2017-10-16
| | | | | | | | | | | | The is similar to the return_tensors option. return_tensors cannot be used to fetch nodes with no outputs, so return_nodes is necessary. In addition, this change also refactors the ImportGraphDef signature to return all optional return values in a single struct. This is to keep the ImportGraphDef signature from getting too long, and also makes the call sites simpler. PiperOrigin-RevId: 172388270
* Add cc file with definition of tensorflow::gtl::nullopt.Gravatar Justin Lebar2017-10-16
| | | | | | If you ODR-use nullopt, you currently get a linker error. Oops. PiperOrigin-RevId: 172387553
* Default to procuring ResourceVariables in variable_scope.variable whenGravatar Akshay Agrawal2017-10-16
| | | | | | use_resource is not set and Eager mode is enabled. PiperOrigin-RevId: 172380659
* Enable C API for gradients_test.pyGravatar Skye Wanderman-Milne2017-10-16
| | | | PiperOrigin-RevId: 172379338
* [TF2XLA] Expand comparator and use consistently in sorting arguments.Gravatar Jacques Pienaar2017-10-16
| | | | PiperOrigin-RevId: 172376836
* Batch norm folding immediately fails if FusedBatchNorm ops are present.Gravatar A. Unique TensorFlower2017-10-16
| | | | PiperOrigin-RevId: 172374244
* Respect __array__ and __array_interface__ for string typesGravatar Mark Daoust2017-10-16
| | | | | | | | | | | | | __array__ fixes use-cases like: import tensorflow as tf import pandas as pd series = pd.Series(['a','b','c']) tf.constant(series) df = pd.DataFrame({'a':[1,2,3],'b':['a','b','c']}) tf.data.Dataset.from_tensor_slices(dict(df)) PiperOrigin-RevId: 172372593
* Close session on infeed error. This should fix most of the cases where the ↵Gravatar Russell Power2017-10-16
| | | | | | client process hangs waiting for the main training loop to exit. PiperOrigin-RevId: 172371951
* Add support for saving DT_VARIANT tensors in TensorBundle.Gravatar Saurabh Saxena2017-10-16
| | | | | | Add support for reading Varint64 to InputBuffer. PiperOrigin-RevId: 172371104
* Move global_step_read dependency to model_fn instead of input_fn.Gravatar Mustafa Ispir2017-10-16
| | | | PiperOrigin-RevId: 172366972
* Remove broken link.Gravatar A. Unique TensorFlower2017-10-16
| | | | PiperOrigin-RevId: 172366027
* Implement set_shape for EagerTensors for compatibiity with ops that call itGravatar Akshay Agrawal2017-10-16
| | | | | | | Checks if shape is not compatible with the Eager tensor's shape, raises an error if it is not. PiperOrigin-RevId: 172363347
* make_vjp in eagerGravatar Alexandre Passos2017-10-16
| | | | PiperOrigin-RevId: 172363016
* Fix divergence between core.data and contrib.data Python tests.Gravatar Jiri Simsa2017-10-16
| | | | PiperOrigin-RevId: 172353443
* [tf.contrib.seq2seq] Some light cleanup in beam search decoder code.Gravatar Eugene Brevdo2017-10-16
| | | | PiperOrigin-RevId: 172352767
* Add tf.contrib.distributions.bijectors.Gumbel.Gravatar A. Unique TensorFlower2017-10-16
| | | | PiperOrigin-RevId: 172350038
* [TF2XLA] Keep Switch and Merge nodes in own clusters.Gravatar Jacques Pienaar2017-10-16
| | | | | | | | | | | | | * Keep Switch and Merge nodes in separate clusters to avoid creating irreducible graphs; * Merge Switch nodes with common predicates; * Add support for if-then structure; * Squash trivial Switch->Merge groups; * Merge newly Merge free nodes with Switch & Merge free inputs; * Check to see if it is a Merge node before merging to common merge node; * Return an error if all Switches have not been replaced; * Add test fir tf,case; PiperOrigin-RevId: 172348729
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-06 04:51:44 +0000