aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* GCS Filesystem should not cache checkpoint file as we need to read the ↵Gravatar Sourabh Bajaj2018-04-11
| | | | | | updated checkpoints from the contents. PiperOrigin-RevId: 192517819
* Extend support to remove transpose/reverse on dimensions of size 1.Gravatar A. Unique TensorFlower2018-04-11
| | | | PiperOrigin-RevId: 192516190
* Adding support for batch_to_space_nd op with crops.Gravatar A. Unique TensorFlower2018-04-11
| | | | PiperOrigin-RevId: 192511036
* Adds a nodedef_fn parameter to copy_op_handler, allowing customization by ↵Gravatar A. Unique TensorFlower2018-04-11
| | | | | | | | mutating NodeDef before creating the copied operation. PiperOrigin-RevId: 192505209
* BREAKING_CHANGE: Remove event_ndims in Bijector, and require ↵Gravatar A. Unique TensorFlower2018-04-11
| | | | | | | | | | | | | | | | | | | | | | | | | `log_det_jacobian` methods to take event_ndims. The class level event_ndims parameter is being deprecated in favor of passing it in to the `log_det_jacobian` methods. Specific changes: - `log_det_jacobian` signatures are now `log_det_jacobian(input, event_ndims)` - Constructors no long have event_ndims passed in (e.g. Affine() vs. Affine(event_ndims=0)). - All bijectors must specify a subset of [forward_min_event_ndims, inverse_min_event_ndims]. This is the minimal dimensionality the bijector operates on, with it being "broadcasted" to any passed in event_ndims (e.g. Exp has forward_min_event_ndims = 0. That means it operates on scalars. However, we can use the bijector on any event_ndims > 0 (i.e. we've broadcasted the transformation to work on any amount of event_ndims > 0), and jacobian reduction will work in those cases. As a result of this change, all bijectors should "broadcast" (e.g. Sigmoid now works on any number of event_ndims). Other changes (internal and documentation): - Added clarifications on Jacobian Determinant vs. Jacobian Matrix. - Added clarifications on min_event_ndims, and what the jacobian reduction is over. - Changed caching of ildj to be keyed on event_ndims. - Several bug fixes to bugs unearthed while writing this code (e.g. transformed distribution shape computation being incorrect) PiperOrigin-RevId: 192504919
* Porting tests for the `decode_proto` and `encode_proto` to OS.Gravatar Jiri Simsa2018-04-11
| | | | PiperOrigin-RevId: 192504411
* Support for removing unfused quantized activation functions and min/max.Gravatar A. Unique TensorFlower2018-04-11
| | | | PiperOrigin-RevId: 192503204
* Replace examples/image_retraining by a pointer to TensorFlow Hub.Gravatar A. Unique TensorFlower2018-04-11
| | | | | | | https://github.com/tensorflow/hub/tree/master/examples/image_retraining has the same tool, upgraded to use TensorFlow Hub instead of raw graph defs. PiperOrigin-RevId: 192502469
* Factor out the syntactic function scope tracking into the transformer. ↵Gravatar A. Unique TensorFlower2018-04-11
| | | | | | Choosing not to do this at static analysis because it exposes the scope to any node, making it easier to use by any specialization of a transformer. PiperOrigin-RevId: 192502309
* When not necessary, avoid the creation of a `placeholder_with_default` in BN ↵Gravatar Francois Chollet2018-04-11
| | | | | | (not yet supported by TPU compilation). PiperOrigin-RevId: 192502020
* DepthwiseConv OptimizationsGravatar A. Unique TensorFlower2018-04-11
| | | | PiperOrigin-RevId: 192499401
* Internal change.Gravatar Anna R2018-04-11
| | | | PiperOrigin-RevId: 192498471
* Adding hp5y back.Gravatar Rohan Jain2018-04-11
| | | | PiperOrigin-RevId: 192491335
* Internal TF Lite test changesGravatar Austin Anderson2018-04-11
| | | | PiperOrigin-RevId: 192491201
* Increase size of ↵Gravatar Derek Murray2018-04-11
| | | | | | tensorflow/contrib/data/python/kernel_tests:batch_dataset_op_test to "medium". PiperOrigin-RevId: 192484895
* Update docs for softmax_cross_entropy_with_logits.Gravatar Adria Puigdomenech2018-04-11
| | | | PiperOrigin-RevId: 192482242
* Script to create custom_ops inside a TensorFlow graphdef.Gravatar A. Unique TensorFlower2018-04-11
| | | | PiperOrigin-RevId: 192481690
* Disable prelu tests for real now.Gravatar A. Unique TensorFlower2018-04-11
| | | | PiperOrigin-RevId: 192480452
* Fix uninitialized value.Gravatar A. Unique TensorFlower2018-04-11
| | | | PiperOrigin-RevId: 192479630
* Use tf.train.get_or_create_global_step() instead of deprecated ↵Gravatar A. Unique TensorFlower2018-04-11
| | | | | | variables.get_or_create_global_step(). PiperOrigin-RevId: 192476077
* Add gradient in cond test to match CallGradInLoop.Gravatar Jacques Pienaar2018-04-11
| | | | PiperOrigin-RevId: 192463997
* Add a clear error message for when a doc does not have a title.Gravatar Mark Daoust2018-04-11
| | | | PiperOrigin-RevId: 192463583
* Import FunctionDef as GrapplerFunctionItemGravatar A. Unique TensorFlower2018-04-11
| | | | | | | | Explicitly track function input arg expansion into Placeholders, and keep metadata to map between FunctionDef and GraphDef connectivity formats. PiperOrigin-RevId: 192462592
* Fixing dependencies.Gravatar Nupur Garg2018-04-11
| | | | PiperOrigin-RevId: 192461382
* Fix Windows GPU TensorFlow Bazel builds.Gravatar Michael Case2018-04-11
| | | | | | | The configure.py script will error out on Windows GPU builds due to NCCL attempted to be configured (and is currently Linux only). PiperOrigin-RevId: 192461362
* Temporarily remove prelu from generated_examples_zip_testGravatar A. Unique TensorFlower2018-04-11
| | | | PiperOrigin-RevId: 192453411
* Remove unused former source of tensorflow.org/tutorials/image_retraining.Gravatar A. Unique TensorFlower2018-04-11
| | | | | | | | | | | | The source of https://tensorflow.org/tutorials/image_retraining has moved from https://github.com/tensorflow/tensorflow/tree/master/tensorflow/docs_src/tutorials to https://github.com/tensorflow/hub/tree/master/docs/tutorials because of its use of TensorFlow Hub. This change replaces the now-defunct version with a pointer to the new location, in order to avoid dead code. PiperOrigin-RevId: 192451570
* Fixing propagation of minmax info on constant gather ops.Gravatar A. Unique TensorFlower2018-04-11
| | | | PiperOrigin-RevId: 192448922
* [TF:XLA] Mark oom_test as optonly, it's really slow when compiled without ↵Gravatar Benjamin Kramer2018-04-11
| | | | | | optimization. PiperOrigin-RevId: 192420481
* Create FileWriter <-> tf.contrib.summary compatibility layerGravatar Nick Felt2018-04-10
| | | | | | | | | | | | | | | This provides an implementation of FileWriter, activated by passing in a `session` parameter to the constructor, that is backed by session.run'ing graph ops that manipulate a tf.contrib.summary.create_file_writer() instance. Because tf.contrib.summary.SummaryWriters are backed by shared resources in the graph, this makes it possible to have a FileWriter and a tf.contrib.summary.SummaryWriter that both write to the same events file. This change includes some related smaller changes: - Factors out training_utils.py into a separate target to avoid a cyclic dep - Moves contrib/summary/summary_ops.py to python/ops/summary_ops_v2.py - Adds SummaryWriter.init(), .flush(), and .close() op-returning methods - Changes create_file_writer() `name` arg to default to logdir prefixed by `logdir:` so shared resources are scoped by logdir by default - Fixes a bug with tf.contrib.summary.flush() `writer` arg - Makes create_file_writer()'s max_queue arg behave as documented - Adds more testing for existing tf.contrib.summary API PiperOrigin-RevId: 192408079
* [XLA] Fix the size of data buffer for sparse literals.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192404543
* [XLA] Redesign: implement ComputeHost.Gravatar A. Unique TensorFlower2018-04-10
| | | | | | Also support convert from/to proto for ComputeHost. PiperOrigin-RevId: 192403660
* [XLA] Redesgin: implement and test Gather, Conditional.Gravatar A. Unique TensorFlower2018-04-10
| | | | | | Also support convert from/to proto for Gather. PiperOrigin-RevId: 192400659
* [XLA] Redesign: implment and test CrossReplicaSum.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192397189
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192392702
* [StreamExecutor] Merge StreamExecutor's and XLA's StatusOr classes.Gravatar Justin Lebar2018-04-10
| | | | | | | | | StatusOr is a...complicated class to write. It's really not good to have two copies of it. They've diverged (the XLA one is more sophisticated), and this may be causing upstream build problems with gcc6. PiperOrigin-RevId: 192392111
* [XLA] Redesign: implement and test BatchNormXXX.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192391748
* experimental C API: Fix compilation failure in Windows.Gravatar Asim Shankar2018-04-10
| | | | | | | | | | | | | The functions added in https://github.com/tensorflow/tensorflow/commit/be917027e37c5e8f21f6ba07f24bdbf072cf6dfd are temporary, and their existence breaks compilation in MSVC because of https://docs.microsoft.com/en-us/cpp/c-language/maximum-string-length and https://docs.microsoft.com/en-us/cpp/error-messages/compiler-errors-1/compiler-error-c2026 So just disabling it in Windows for now. PiperOrigin-RevId: 192391164
* Checkpointable: remove colocation constraints from restore opsGravatar Allen Lavoie2018-04-10
| | | | | | Mystery solved thanks to log_device_placement. PiperOrigin-RevId: 192389574
* Remove BN workaround for resource variable gradients bug that was recently ↵Gravatar Francois Chollet2018-04-10
| | | | | | fixed. PiperOrigin-RevId: 192388867
* Merge changes from github.Gravatar Michael Case2018-04-10
| | | | PiperOrigin-RevId: 192388250
* Fix for users who were passing `Dimension` type as `units` arg in `Dense`.Gravatar Francois Chollet2018-04-10
| | | | PiperOrigin-RevId: 192387984
* TFTS: De-flake the LSTM testGravatar Allen Lavoie2018-04-10
| | | | | | Disabling the value-based check for now. Hopefully the shapes are deterministic. PiperOrigin-RevId: 192383553
* Forward the status from LookupResource to GetInputTensorFromVariable rather ↵Gravatar Allen Lavoie2018-04-10
| | | | | | than returning a generic error status PiperOrigin-RevId: 192382499
* [XLA] Redesign: implement Conj.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192381481
* [XLA] Redesign: implement SliceInDim.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192381080
* [XLA] Redesign: implement ReduceAll.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192380688
* [XLA] Redesign: implement and test SelectAndScatter.Gravatar A. Unique TensorFlower2018-04-10
| | | | PiperOrigin-RevId: 192380121
* Update LogToSTDErr for TF Lite usageGravatar Austin Anderson2018-04-10
| | | | PiperOrigin-RevId: 192379483
* Add types to error message in case of mismatch. NFC.Gravatar Jacques Pienaar2018-04-10
| | | | PiperOrigin-RevId: 192370979
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-10 02:23:33 +0000