| cb4ef362e4a18b3c42a2c90bdad8754d5ead4caf |
|
30-Nov-2017 |
Yangzihao Wang <yangzihao@google.com> |
Add native dilated support for conv2d and its gradients in cudnn v>=6.
PiperOrigin-RevId: 177382431
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 2b15badd96c651d4d191426975a1773dff4a03b8 |
|
06-Sep-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Add int8 version of fused_conv2d_bias_activation operator for the forward phase,
and support side_input and scaling parameters in float and int8 versions.
PiperOrigin-RevId: 167763219
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 10594900c5df1b84cd0336d0fb5bd0d8454bfe08 |
|
06-Sep-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Update MaxPoolV2Shape to support NCHV_VECT_C.
PiperOrigin-RevId: 167732437
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| dfeb146be1256b35453719ad5c70c6d2b6e50369 |
|
29-Aug-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Refactor Avg/MaxPoolShape(), adding MakeShapeFromFormat() and preparing for NCHW_VECT_C.
PiperOrigin-RevId: 166856720
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 91b00a110feb83b7307ea9c280142007090f3cd9 |
|
24-Aug-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Automated g4 rollback of changelist 166276461
PiperOrigin-RevId: 166305887
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 1f41602a82cb68fc7bc7e51cf9590a87ee5baf4d |
|
24-Aug-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Add int8 version of fused_conv2d_bias_activation operator for the forward phase,
and support side_input and scaling parameters in float and int8 versions.
PiperOrigin-RevId: 166276461
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 104f349e9eda70a21533294d2fb09f0bf0828834 |
|
31-Jul-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Update Conv2DShape function to handle filters that have data NCHW_VECT_C layout.
PiperOrigin-RevId: 163746769
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 9d45160c76bdb2aa7cf3cc5b318512fc7d7337a2 |
|
14-Jul-2017 |
Jingyue Wu <jingyue@google.com> |
Extend shape_inference::Conv2DShape to handle NCHW_VECT_C format.
Tested Conv2DShape with NCHW_VECT_C format.
PiperOrigin-RevId: 161879362
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| e85d3df92deb9d717befdf173966a2913ac2aea0 |
|
29-Jun-2017 |
Geoffrey Irving <geoffreyi@google.com> |
Prepare to remove a bunch of proto.h includes from tensorflow/core headers
The goal is to make kernels mostly independent of proto headers, which will let
us lock down our .so imports. This CL does not remove any actual headers, but
changes a bunch of files so that header removal is possible in a followup CL.
It also marks the headers that will be removed with
// TODO(b/62899350): Remove
RELNOTES: n/a
PiperOrigin-RevId: 160552878
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| eae7e833f104b0afc0aa7bba21a59c56612dc3df |
|
26-May-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Change shape inference so that a single resource tensor can carry
shape and type information for multiple tensors.
Apply this to QueueDequeueV2 handled by grappler.
PiperOrigin-RevId: 157163757
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| ce016c8726a9250be98337691090acb6655a0ace |
|
08-Mar-2017 |
Vijay Vasudevan <vrv@google.com> |
Add initial support for NCHW format for DepthwiseConv2D (and separable_conv2d).
This is a straightforward port of the NHWC kernels with different
input tensor indexing. This is a baseline for future optimizations,
and allows running separable nets with NCHW format throughout.
Change: 149565634
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| b177e3720721dea593f1f15ba731ab351e87d298 |
|
03-Mar-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Add the graphdef version to InferenceContext and to ShapeRefiner::AddNode.
Use this to allow loading reductions saved with older graphdefs.
Change GraphConstructor to not increase the version when importing, but instead take the min of all versions.
Change: 149152437
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| a910b52351f8341227c9814cafb0123b8c616fc6 |
|
09-Jan-2017 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Change reduce_join to treat reduction_indices in the same way as
other reduce_ ops:
- [] is no longer special, it means to do no reduction
- None is handled in python to mean "reduce all dims"
- reducing over a scalar is supported
Unlike other reduce_ ops, reduce_join still disallows duplicate
reduction indices. This is because the order of the indices matters.
Change: 143984486
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 51ae55c4e782c7577e10e51ee54721593204f396 |
|
23-Nov-2016 |
Vijay Vasudevan <vrv@google.com> |
Relax Pool3D shape function when some dimensions are unknown.
Added test.
Fixes #5807.
Change: 140042698
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 0c2018764516add343acc9ee56c487715bc98c2a |
|
14-Nov-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Move ValidateSparseTensor to common_shape_fns.h.
Change: 139112567
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 41734d78d3facf652c25b2a2761aadd978b3f2ef |
|
31-Oct-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Automated rollback of change 137740850
Change: 137747341
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 2940b6c9ac3518b27633d90880b9092157496ee8 |
|
31-Oct-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Automated rollback of change 137731142
Change: 137740850
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 0d616c45650b8bef5da330d311bf265896328db3 |
|
31-Oct-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Adds functionality for shape inference and dtype checking of variable-based resources.
Does not implement these resources yet as this change is big enough already.
Change: 137731142
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 8cb8645322f5a738c8ce7bbfd6ebcbaac3e3ba02 |
|
26-Oct-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
In ShapeRefiner, add support for the C++ equivalent of the Python
constant_value_as_tensor functions. This follows the same lazy-evaluation as
getting constant tensors.
Add validation in InferenceContext::MakeShapeFromShapeTensor for invalid values
in the input tensor.
Change: 137231472
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 5826a9703eeba43c640aa38b09c05a3579d50802 |
|
12-Oct-2016 |
Vijay Vasudevan <vrv@google.com> |
Get rid of testing constructor for InferenceContext, fix TODOs.
Makes shapes owned by a private subclass of InferenceContext called
ShapeManager (for lack of better name), friend classed
ShapeInferenceTestutil so that conversion from string to shapes is
private. Cleaned up extra arg from constructors by changing
the two places that also do that to use a simple helper function
to convert to TensorShapeProto.
Change: 135956508
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 7d9c0c891d82fb5d35dc4669abe832708940a810 |
|
30-Sep-2016 |
Vijay Vasudevan <vrv@google.com> |
TensorFlow: for shape inference errors, provide more context in the
error message thrown by C++ (which op and what the input shapes are).
Update tests that looked for exact error messages to use .contains()
for robustness.
Change: 134796475
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 24af41cd9a7c6c6a397c1ca97b260c3ea4a04974 |
|
13-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Change C++ ReduceJoin shape fn to match the python one for the case of an empty
input dimension with the reduce_all behavior.
Change: 132952610
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 29faf36ed2341fe50a02a1ad1329b96157d96d55 |
|
09-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Switch ops in functional_ops, math_ops, state_ops, string_ops,
contrib/quantization, and contrib/layers to use C++ shape inference functions.
Implement Betainc C++ shape inference function; it's a little different from
the python one, mainly because propagating an unknown input (a_shape in this
case) when all are unknown is not correct for C++ (in case we later backwards
bind the value), since it could end up being a broadcasted scalar.
Change ReduceJoin's C++ shape inference function to match python.
Fix SymbolicGradient's shape function in C++ - there are actually more inputs
than outputs.
Changes QuantizedBiasAdd to be more precise (the C++ shape fn invokes the common bias_add to compute the shape of the first output).
Change: 132688147
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| c3a30a230f47a8ca8f5dd1dd79c63229ce1349b8 |
|
08-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Switch nn_ops shape fns to delegate to C++, for all that have a C++
implementation (fractional pool ones don't yet).
Change BiasAdd functions to require only rank 3, not 4, for NHWC. This matches
the behavior of GetBiasValueDims in bias_op.cc.
Removed unused functions common_shapes.bias_add_shape and
common_shapes.bias_add_grad_shape.
Change: 132597521
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| b74a74393b0d4ce118cc051d91f7b5996bf3dd98 |
|
07-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Delegate to C++ shape inference function for some conv and pooling functions.
Change several C++ shape inference functions to not return an error if an input
dimension is unknown; this more closely matches the python functions.
Change: 132459740
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 7e538db04d9ecf2432f8f0c7bf2d6d5c1f994e1a |
|
07-Sep-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
When calling C++ shape functions from python shape functions,
allow passing the result of python's constant_value for some input tensors.
Use this on some existing math ops to delegate to python.
This updates Pack to merge all inputs before checking ndims is None, which
allows more merging possibilities.
Changes to Reduction C++ shape function to match python version.
Change: 132434990
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| f6bc8cabbd3ac1fb3acc36d3edbdce672cae7d12 |
|
24-Aug-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Add shape_inference::ShapeHandle and shape_inference::DimensionHandle to
replace uses of const Shape* and const Dimension*. This change only adds a
typedef and updates references. A later change will make DimensionHandle and
ShapeHandle real types instead of typedefs (to further hide the pointer
access).
Change: 131118981
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 8b409dfa8de95fb58e511592ccefda25e97d816f |
|
10-Aug-2016 |
Vijay Vasudevan <vrv@google.com> |
TensorFlow: shape function in C++ for Reduction operations.
Change ReduceJoin to have the same shape function as other reduction
operations, including the same rank requirements.
Change: 129888814
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 19ad04564a70ae0134c044666460f47714e287f1 |
|
01-Aug-2016 |
Vijay Vasudevan <vrv@google.com> |
TensorFlow: Add Conv3D/MaxPool3D/AvgPool3D C++ shape inference functions .
Change: 129011665
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 757f4d9f7b459af252766cdd258d534d9d5813e5 |
|
29-Jul-2016 |
Vijay Vasudevan <vrv@google.com> |
TensorFlow: implement MaxPoolShape common function + tests, add to
ops that need it.
Moved one of the implementation-specific checks in the shape function
to the kernel instead, where it belongs.
Change: 128834923
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| eaf5ad2ce467cd8730cbbab0cd96dee0d92399c1 |
|
27-Jul-2016 |
Vijay Vasudevan <vrv@google.com> |
TensorFlow: Annotate UnknownShape common shape function to all
ops that currently do the same in python.
Change: 128627877
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 8d1f819db3f27a544071eaf81d60e0c37d015fc7 |
|
23-Jul-2016 |
Vijay Vasudevan <vrv@google.com> |
Add C++ Shape function for AvgPool
Change: 128219991
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 2641ef0c440161c8f75683d29d4ec53c52de51ed |
|
22-Jul-2016 |
Vijay Vasudevan <vrv@google.com> |
Add ShapeInference functions for DepthwiseConv2DNative, and Multiply
support to ShapeInference in support of DepthwiseConv.
Change: 128114077
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 0d81d9c2c762ddf23bece3afa09456ee2534131a |
|
21-Jul-2016 |
Vijay Vasudevan <vrv@google.com> |
Added C++ Shape inference function for Conv2D.
Moved some of the utilities we use for computing 2d output shapes into
common_shape_fns.h to share some of this code.
Change: 128008947
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| 976e7ff02122a1c313408ca14abb2a6b24392c12 |
|
16-Jul-2016 |
A. Unique TensorFlower <gardener@tensorflow.org> |
Infer the number of outputs properly in shape_inference::InferenceContext,
based on the nodedef and opdef.
Change: 127602218
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|
| ff46ed44c13068af63bc601129666e731683d1b9 |
|
16-Jul-2016 |
Vijay Vasudevan <vrv@google.com> |
Implement more common shape functions in C++ with tests:
- MatMulShape
- BiasAddShape
- NoOutputs
- ScalarShape
Change: 127592611
/external/tensorflow/tensorflow/core/framework/common_shape_fns_test.cc
|