| 348bf0c436e4f571022bc08d0699e3b257125467 |
|
08-Feb-2018 |
Michael Case <mikecase@google.com> |
Merge commit for internal changes
|
| d90054e7c0f41f4bab81df0548577a73b939a87a |
|
07-Feb-2018 |
Michael Case <mikecase@google.com> |
Merge changes from github.
PiperOrigin-RevId: 184897758
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|
| 9d84147b0a895ef970f60dde6b78a5bfcc8ba4d7 |
|
24-Jan-2018 |
Sami Kama <samikama@users.noreply.github.com> |
Workaround 'too perfect forwarding' issue in variant_op_registry (#16309)
* Workaround 'too perfect forwarding' issue in variant_op_registry
* Fix comment issue created by clang-format
* Fix friend declaration for clang
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|
| cf57817f554fa7bfe7c134453ef9cf4374aef23d |
|
21-Nov-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Automated g4 rollback of changelist 175593063
PiperOrigin-RevId: 176445215
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|
| 7d197b3ab4b5389c19ab7742cf5d473d5fa91b08 |
|
14-Nov-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Automated g4 rollback of changelist 175563786
PiperOrigin-RevId: 175593063
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|
| 0a79d4ab6ab961d7f36cbb3a14cfaff2152e70fd |
|
13-Nov-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Moved tensorflow::StringPiece::Hasher out of tensorflow::StringPiece and renamed it tensorflow::StringPieceHasher. This allows tensorflow::StringPiece to be more easily replaced with absl::string_view (which does not contain a Hasher struct).
PiperOrigin-RevId: 175563786
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|
| 0a11eaffc985ad6abd3a0e792061e1880766674a |
|
03-Oct-2017 |
Eugene Brevdo <ebrevdo@google.com> |
Internal Variant API allowing registering Variants to be copied from/to GPU.
Adds a test in the variant_op_copy_test.
Modifies the base GPUDevice to use this registry if it sees a singleton variant.
Modifies the rendezvous manager to do the same.
PiperOrigin-RevId: 170908757
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|
| 6e2f629aafd71ab9231ee1830f96a43488d9d48a |
|
20-Sep-2017 |
Eugene Brevdo <ebrevdo@google.com> |
Add tf.conj support for Variants, using a new variant op registration.
PiperOrigin-RevId: 169430180
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|
| 477a221a2ffee7261220ad6c0f4f8c76a5eb7931 |
|
08-Sep-2017 |
Eugene Brevdo <ebrevdo@google.com> |
Modify variant registry to have UnaryOp and BinaryOp registrations. Speed up registry lookup.
* Op type is described as an enum (separate enums for unary and binary ops).
* Modified ZerosLike registrations to unary registrations with ZEROS_LIKE enum.
* Added Add(a,b) registrations as binary registrations with ADD enum.
* AddN op uses ADD BinaryOp registrations and ZerosLike op modified to use
ZEROS_LIKE UnaryOp registrations.
* Modified the registry tables' keys from string type to StringPiece type.
The reduced copying should speed up registry lookups by ops. Required creating
a backing store for device and type_name strings passed in at registration.
PiperOrigin-RevId: 168020449
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|
| 0acf5bb38a8f208c6d9f048579a076d5bc6ff0be |
|
01-Sep-2017 |
Eugene Brevdo <ebrevdo@google.com> |
Added registry for variants for op ZerosLike.
* Updated tf.zeros_like python wrapper to avoid calling tf.zeros on Variants.
* tf.zeros_like for variants calls the appropriate callback for the given device
and type_name.
PiperOrigin-RevId: 167317274
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|
| 105dd2aa22c7aaac401d2d8ed3131b5d81fdc227 |
|
24-Aug-2017 |
Eugene Brevdo <ebrevdo@google.com> |
Implement Variant changes suggested by Manjunath Kudlur; simplify API.
1. Remove the self-mutating MaybeDecodeAndGet<>() in favor of the simple,
original get<>().
2. Remove mutexes and locks from everything.
To allow tensorflow RPC to continue working transparently:
3. Add a registration mechanism for Variant decoding -- used by
variant_coding within TensorFlow.
4. Add a MaybeDecodeAndCopy<>() const; for use by callers of Session::Run
that get Tensors back which may contain encoded Variants.
5. Small bugfixes to Variant Encode impls and tensor FromProtoField and
variant_coding.
PiperOrigin-RevId: 166383211
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|
| 2f6f329aac6e4f23ef20f6e37165a235e232e772 |
|
22-Aug-2017 |
Eugene Brevdo <ebrevdo@google.com> |
Add variant dtype op registry, first registration mechanism: shape ops.
A followup PR will have the actual modifications to the shape ops that
make use of this registry.
PiperOrigin-RevId: 166077425
/external/tensorflow/tensorflow/core/framework/variant_op_registry.h
|