aboutsummaryrefslogtreecommitdiffhomepage
path: root/tensorflow/python/ops
Commit message (Collapse)AuthorAge
* cond_v2: raise an error if pred is a Python bool.Gravatar Skye Wanderman-Milne2018-10-10
| | | | | | This is to match the existing behavior of tf.cond. PiperOrigin-RevId: 216534084
* Silence tf.distributions deprecation messages caused by internal global ↵Gravatar Pavel Sountsov2018-10-09
| | | | | | | | | | | | function calls. E.g. register_kl calls would trigger such warnings. This spam was exacerbated by the fact that it happens before logging is initialized, so it is dumped prominently to STDERR. Worse yet it also happened no matter whether the user imported any symbols from tf.distributions or not as the relevant code is executed when you import TensorFlow. PiperOrigin-RevId: 216396036
* Removing the _SHOULD_RECORD_SUMMARIES_NAME andGravatar Rohan Jain2018-10-09
| | | | | | | _SUMMARY_WRITER_INIT_COLLECTION_NAME collections from the summaryV2 implementation. Replacing them with global variables. PiperOrigin-RevId: 216383152
* Create SDCAOptimizerV2 op to fix the "adaptative" typo.Gravatar Yuefeng Zhou2018-10-09
| | | | PiperOrigin-RevId: 216370193
* Partial support tfe.defun in tf.gradients.Gravatar Alexandre Passos2018-10-08
| | | | | | | | Doesn't attempt to deal with cases where we might have already generated the functiondef for the parent function as in that case we cannot easily modify the forward pass. PiperOrigin-RevId: 216243224
* Add tf.BenchmarkConfig that returns a session config appropriate for ↵Gravatar A. Unique TensorFlower2018-10-08
| | | | | | benchmarking. At the moment, it returns a default config with only Grappler dependency optimizer disabled. Many benchmarks wrap the subgraph they want to time in control_flow_ops.group() to avoid including the overhead of copying the output back to the Python client in the measurement. In the graph, this only adds a control dependency between the subgraph output and the fetch node, which in turn (often) causes the dependency optimizer to turn all nodes in the graph into no-ops. PiperOrigin-RevId: 216242463
* Merge pull request #21658 from lowintelligence:masterGravatar TensorFlower Gardener2018-10-08
|\ | | | | | | PiperOrigin-RevId: 216217509
* | Add support for SequenceExamples to sequence_feature_columnsGravatar Karmel Allison2018-10-08
| | | | | | | | PiperOrigin-RevId: 216210141
* | Benchmark for comparing original cond and cond_v2 performance.Gravatar Skye Wanderman-Milne2018-10-08
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This benchmark creates many intermediates values, so we can make sure there's no performance overhead (it looks like there might be currently, or it might be from some other difference). It also runs in a defun and in legacy graph mode. Results from my machine: entry { name: "CondWithManyIntermediatesBenchmark.benchmark_cond_v1_defun" iters: 500 wall_time: 1.25822591782 } entry { name: "CondWithManyIntermediatesBenchmark.benchmark_cond_v2_defun" iters: 500 wall_time: 5.99376106262 } entry { name: "CondWithManyIntermediatesBenchmark.benchmark_cond_v1_graph" iters: 500 wall_time: 2.05277585983 } entry { name: "CondWithManyIntermediatesBenchmark.benchmark_cond_v2_graph" iters: 500 wall_time: 2.84808516502 } Clearly we have some work to do! I haven't looked into the time differences at all yet. PiperOrigin-RevId: 216202325
* | Simply the logic for bubbling captured tensors when building cond_v2 grad.Gravatar Saurabh Saxena2018-10-05
| | | | | | | | | | | | | | | | | | | | The current logic tries to bubble the forward pass tensor to the outermost graph. That might not always be do-able e.g. when the cond is inside a while loop it will need to know accumulator logic for while_loop. So instead, the cond_grad now captures tensors from the forward If op's graph. When the grad If op is built these tensors will be appropriately captured by the surrounding FuncGraph. PiperOrigin-RevId: 215993009
* | Automated rollback of commit d258207f1583df4faa452265b051879af6c15dacGravatar A. Unique TensorFlower2018-10-05
| | | | | | | | PiperOrigin-RevId: 215989111
* | BEGIN_PUBLICGravatar A. Unique TensorFlower2018-10-05
| | | | | | | | | | | | | | | | Automated rollback of PR #21945 END_PUBLIC Automated rollback of commit 863f61412fcc654840c6b67473b742ea4e5e964e. Revert #21945. PiperOrigin-RevId: 215913175
* | Merge pull request #21945 from efagerho:masterGravatar TensorFlower Gardener2018-10-04
|\ \ | | | | | | | | | PiperOrigin-RevId: 215824410
* | | Avoid creating control edges on not-this-graph.Gravatar Alexandre Passos2018-10-04
| | | | | | | | | | | | PiperOrigin-RevId: 215811680
* | | Add "encoding" attribute to string substr op, which controls how each ↵Gravatar A. Unique TensorFlower2018-10-04
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | "character" is treated: * BYTE: Position & length refer to bytes in the string. (Default) * UTF8: The string is interpreted as UTF-8 encoded Unicode code points, and position & length are treated relative to them. RELNOTES: Add option to get substring using Unicode characters PiperOrigin-RevId: 215773373
* | | Add option in tf.gradients() to return zero tensors for unconnected gradients.Gravatar A. Unique TensorFlower2018-10-04
| | | | | | | | | | | | | | | | | | | | | | | | tf.gradients currently returns [NONE] when the gradient of unconnected variables is required. This backwards compatable change adds in the option to have zero tensors returned that match the dimensions of the input tensor. PiperOrigin-RevId: 215725488
* | | Make batch_gather work with indices of dtype int64.Gravatar Adria Puigdomenech2018-10-04
| | | | | | | | | | | | PiperOrigin-RevId: 215711383
* | | Set shape for output tensors of cond_v2.Gravatar Saurabh Saxena2018-10-02
| | | | | | | | | | | | PiperOrigin-RevId: 215492782
* | | Support shape_invariants in while_v2. Note that this arg is temporary and ↵Gravatar Saurabh Saxena2018-10-02
| | | | | | | | | | | | | | | | | | | | | | | | may be replaced by automatic shape inference in TF 2.0 (or before). Add a output_shapes attr to While op to allow output shapes to be different from the incoming loop_vars. PiperOrigin-RevId: 215446737
* | | Copy tf.distributions to tfp.distributions, and deprecate the ↵Gravatar A. Unique TensorFlower2018-10-02
| | | | | | | | | | | | | | | | | | tf.distributions API. PiperOrigin-RevId: 215441733
* | | Merge pull request #22126 from ConcurrencyPractitioner:masterGravatar TensorFlower Gardener2018-10-02
|\ \ \ | | | | | | | | | | | | PiperOrigin-RevId: 215431884
* | | | Use xlogy in a few places in TFP to avoid NaN's for certain special cases.Gravatar A. Unique TensorFlower2018-10-02
| | | | | | | | | | | | | | | | PiperOrigin-RevId: 215392621
* | | | Export endpoint for the version of the `regex_replace` function that calls ↵Gravatar A. Unique TensorFlower2018-10-02
| | | | | | | | | | | | | | | | | | | | | | | | StaticRegexReplace. PiperOrigin-RevId: 215371291
* | | | Override implementation of log survival for Exponential distribution to ↵Gravatar A. Unique TensorFlower2018-10-01
| | | | | | | | | | | | | | | | | | | | | | | | better handle small values. PiperOrigin-RevId: 215299532
* | | | Add email comment explicitly authorizing distributions/special_math.py be ↵Gravatar Joshua V. Dillon2018-10-01
| | | | | | | | | | | | | | | | | | | | | | | | released under Apache 2.0. PiperOrigin-RevId: 215296386
* | | | Fix benchmark regression.Gravatar Alexandre Passos2018-10-01
| | | | | | | | | | | | | | | | PiperOrigin-RevId: 215278033
* | | | Updating function and class tf_export decorators for endpoints according toGravatar Anna R2018-10-01
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | https://github.com/tensorflow/community/pull/16. In addition to the changes in the doc, I made the following updates (these changes make sense to me and I didn't notice them when compiling the doc): * deprecate saved_model.builder.SavedModelBuilder - replaced with saved_model.SavedModelBuilder * deprecate python_io.tf_record_iterator - replaced with io.tf_record_iterator * deprecate python_io.TFRecordWriter - replaced with io.TFRecordWriter * move reduce_join to tf.string PiperOrigin-RevId: 215253944
* | | | Add deprecation notice for BasicRNNCell, which will be replaced by ↵Gravatar Scott Zhu2018-10-01
| | | | | | | | | | | | | | | | | | | | | | | | keras.SimpleRNNCell. PiperOrigin-RevId: 215249611
* | | | Merge pull request #22473 from wangsiyu:assign_in_part_varsGravatar TensorFlower Gardener2018-10-01
|\ \ \ \ | | | | | | | | | | | | | | | PiperOrigin-RevId: 215211485
* | | | | Moves the creation of regularizer ops in get_variable out of surrounding ↵Gravatar A. Unique TensorFlower2018-10-01
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | context. This resembles the behaviour for initializer ops. PiperOrigin-RevId: 215187942
* | | | | Parallel_for: fix converters for some ops that don't support broadcasting.Gravatar A. Unique TensorFlower2018-09-30
| | | | | | | | | | | | | | | | | | | | PiperOrigin-RevId: 215133508
* | | | | Add flag for enabling while_v2.Gravatar Saurabh Saxena2018-09-28
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add a single test flag for enabling v2 control flow in tests since we do not plan to support v2 ops with legacy control flow. We have 2 test decorators now: @with_control_flow_v2: Enables all tests in a class to run with v2 control flow. @disable_control_flow_v2: Disables a test function from running in v2. I have removed the skiptests to avoid setup/teardown overheads. Enable tests in control_flow_ops_py_test that run with control_flow_v2. PiperOrigin-RevId: 214980108
* | | | | Remove @{} api_links and ban "@{}" from python and md files.Gravatar Mark Daoust2018-09-28
| | | | | | | | | | | | | | | | | | | | PiperOrigin-RevId: 214964988
| * | | | remove slashGravatar wangsiyu2018-09-28
| | | | |
* | | | | Automated rollback of commit 425e96f3ae4eb338268e3738260f9d79e4bdd893. ↵Gravatar A. Unique TensorFlower2018-09-27
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Revert #20539. PiperOrigin-RevId: 214849875
* | | | | Updating the V2 variables API.Gravatar Alexandre Passos2018-09-27
| | | | | | | | | | | | | | | | | | | | PiperOrigin-RevId: 214824023
* | | | | Update logic used in get_variable to populate custom_getter's kwargs.Gravatar A. Unique TensorFlower2018-09-27
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The new implementation ensures that the 'constraints' kwarg is propagated by customer getters whose signature includes a keyworded, variable length argument dictionary, as well as those explicitly including the 'constraints' argument. PiperOrigin-RevId: 214767296
* | | | | Add xlogy and xdivy op.Gravatar A. Unique TensorFlower2018-09-26
| | | | | | | | | | | | | | | | | | | | PiperOrigin-RevId: 214700693
* | | | | Specify a preferred_dtype=self.dtype when converting Distribution methods' ↵Gravatar Brian Patton2018-09-26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | sample-like args to Tensors. After this change, you could conceivably write tfd.Normal(0., 1.).log_prob(1) The tf core distributions can't use tfp dtype_util.common_dtype, so you can't yet write tfd.Normal(0, 1). Works around an eager bug that loses precision in the presence in tf.convert_to_tensor(0.5, preferred_dtype=tf.int32) PiperOrigin-RevId: 214666222
| | | | * Re-add compat module for leaky_relu implementation.Gravatar Cao Zongyan2018-09-26
| | | | |
| | | | * Merge remote-tracking branch 'origin'Gravatar Cao Zongyan2018-09-26
| | | | |\ | |_|_|_|/ |/| | | |
* | | | | Add "encoding" attribute to string length op, which controls how "string ↵Gravatar A. Unique TensorFlower2018-09-25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | length" is defined: * BYTE: The number of bytes in each string. (Default) * UTF8: The number of UTF-8 encoded Unicode code points in each string. RELNOTES: Add option to calculate string length in Unicode characters PiperOrigin-RevId: 214478470
* | | | | Use self.cached_session instead of self.test_session in ↵Gravatar A. Unique TensorFlower2018-09-25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | linear_operator_circulant_test. Also: * Instead of overwriting self.test_session(), overwrite self._constrain_devices_and_set_default() to remap the kernel operations (this way self.cached_session(), self.test_session() and self.session() are all correct). * Make linear_operator_test_util use self.session(graph=...) instead of self.test_session(graph=...) (semantically equivalent). PiperOrigin-RevId: 214448118
| * | | | remove warning linesGravatar wangsiyu2018-09-25
| | | | |
| * | | | fix pylintGravatar wangsiyu2018-09-25
| | | | |
| * | | | Merge branch 'master' of github.com:tensorflow/tensorflow into ↵Gravatar wangsiyu2018-09-25
| |\ \ \ \ | |/ / / / |/| | | | | | | | | assign_in_part_vars
* | | | | Unpack output of cond_v2 if it is a singleton to match behavior of cond.Gravatar Saurabh Saxena2018-09-24
| | | | | | | | | | | | | | | | | | | | PiperOrigin-RevId: 214381126
* | | | | Documentation for tf.map_fn in Eager mode.Gravatar Debidatta Dwibedi2018-09-24
| | | | | | | | | | | | | | | | | | | | PiperOrigin-RevId: 214376416
* | | | | Allow callers to specify a preferred dtype when calling convert_to_tensor.Gravatar Brian Patton2018-09-24
| | | | | | | | | | | | | | | | | | | | PiperOrigin-RevId: 214370113
| | | | * Fix build dependencies in tensorflow/cc/BUILD.Gravatar Cao Zongyan2018-09-25
| | | | |
generated by cgit v1.2.3 (git 2.39.1) at 2024-06-03 13:36:57 +0000