* Add GatherND_8 operation
* Update shape infer function and tests
* Initial commit for nGraph GatherND_8 operation
* Add GatherNDBase class implementation
* Fix base class errors
* Add missrd header
* Update base class
* Update GatherND_8 implementation
* Fix codestyle
* Fix wrong rank
* Implement tests for gatherND_8 shape inference function
* fix codestyle
* Add limitation to doc
* Siplyfy check in shape inference
* Add more test cases
* Update shape inference function
* Add more test cases to cover all case with dynamic input shapes
* Update shape inference function
* Refactor tests
* Initial commit for gatherND_8 reference implementation
* Add visitor tests for gatherND_8 operation
* Add visitor tests for gatherND_8 operation
* Correct comment
* Add additional check is shape inference function
* Update shape inference implementation for gathernd operartion
* Fix codestyle
* Remove restriction for data is fully defined
* Update shape inference functon
* Add gatherND_8 reference to evaluate map
* Add tests for gathernd_8 reference in batch_dims > 1
* Fix codestyle
* Fix codestyle
* Fix missed check for nonetype
* Remove redundant checks for batch_dims
* Use get_shape method instead of get_input_shape in GatherND reference implementation call
* update reference implementation call
* Properly add GatherND_8 to test opset table
* Update GatherND reference implementation for support opset8 version
* Add unit tests for gatherND_8 reference implementation
* Update gatherND reference implementation call in evaluate map
* Mark failed tests for missed plugin gatherND_8 implementations
* Fix codestyle
* Fix codestyle
* Initial commit for gaherND_8 python api
* Fix codestyle
* Update tests
* Move common methods to base class
* Fix codestyle
* Revert clone_with_new_input function moving
* Add more test for reference
* partially revert API changes
* Fix test
* Fix codeslyle
* [MO, TF] Support Custom Wide and Deep CTR model by MO
It implements implicit support of EmbeddingSegmentsMean operation through decomposition.
Also, this extends the current transformation to fuse TensorFlow sub-graph (for Wide and Deep model family)
containing SparseSegmentSum and SparseSegmentMean operations into EmbeddingSegmentsSum or EmbeddingSegmentsMean.
Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Fix unit-tests after modifications of SparseToDense and EmbeddingSegmentsOperationFusing
Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Document SparseSegmentMean support
Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Add computation scheme for normalization coeffs and correct documentation
Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Initial optimization implementation
* Add tests
* Updated function and node accessors
* Avoid using static mutex in get_ordered_ops
* Apply comments
* resolve PR comments
* Move mutex up
* Simplify shared info
* Fix info copying in node copy c-tor
* Move friend declarations to public
* Fix to support node access from multiple functions in different threads
* [GNA] Improved Log activation accuracy
* [GNA] rename PWL_MAX_ITERATIONS to PWL_MAX_ITERATIONS_DEFAULT
* [GNA] Fixed invalid macros using for max segments number
* Started to write transformation that allows MKLDNN Interpolate to work with axes {1,2} or {1,2,3}.
* Written the transformation that wraps Interpolate into Transposes.
* Some fixes.
* Written transformation tests.
* Added debug prints.
* Fixed order inputs calculations for inserted Transposes.
* Fixed the transformation tests.
* Deleted commented code.
* Fixed includes.
* Small fix.
* In some places, int64_t was replaced by size_t.
* Small change.
* Small fix.
* The function transformation_is_applicable() was inlined.
* Inlined the helper function get_non_interpolated_axes().
* Inlined the function create_set_of_all_axes().
* Code style fix.
* Small fix.
* Added test.
* Some fixes.
* Fixed shapes of start, stop, step inputs of Range in the test.
* Added more tests.
* Small fix.
* Added TransposeSinking after WrapInterpolateIntoTransposes (in MKLDNN plugin).
* 1. fix for CVS-69011 2.add more test cases for slice w/ decrease_axis
* add test case simulating the last reshape2 of ocrnet which accepts slice (with decrease_axes in all dims) as its parents.
* exportModel as a helper function.
* fix: rank get_length type
* Enable LTO for oneDNN GPU
* Use importer target for onednn_gpu
* Fixed install of OpenCL, ittapi
* Try to enable samples build
* Fixed ie_wheel build
* Disabled samples
* Fixed incorrect install of pugixml in dynamic build
* Fixed clDNN_unit_tests64 link
* Minimized a number of hardcoded paths
* Switched off ITT
* Disable TBBBind for static Windows
* Fixed link on Windows
* Fixes
* Static build for Window
* Disable unit-tests
* Disabled TF
* Fixed add dependency for PDPD unit tests
* Disabled GNA
* Enabled ngraph test utils
* Fixes
* Fixes
* Disable clDNN for static build on Windows
* Dynamic MCVS runtime for OpenCL
* Disable ITT traces for static libs
* Ability to switch TBBBind_2_5 OFF
* Disabled GPU on Windows build
* Fixed libGNA link
* Disabled some GNA tests
* Fixed multiple definitions issue
* Hide FrontEnd::load in DLL
* Install libGNAStubs
* Trying to fix Windows
* Trying to fix frontends
* Passed compilation on Windows
* Updated install rules for ngraph backends
* Try CPU, TEMPLATE tests
* Enabled unit-tests
* Enabled ngraph tests
* Try to fix unit-test
* Fixed unit-test on Linux
* Fixed compilation with ENABLE_IR_V7_READER=OFF
* Fixed tests on Linux
* Enable cpuFuncTests only in static build
* Try to fix testMockPluginFrontEnd test
* Try to enable TF
* Fixed python tests
* Partial revert
* Removed OV_FRONTEND_PATH
- PrePostProcessor takes 'function' argument in constructor
- PrePostProcessor::build() doesn't take any function anymore
- PrePostProcessor::input() method to get reference to input
- PrePostProcessor::output() method to get reference to output
- InputInfo - add getters of tensor, preprocess, network
- OutputInfo - add getters of tensor, preprocess, network
Samples:
ClassificationSampleAsync - use new getters
Inference engine:
- Use new getters in ie_network_reader.cpp
TODO: Consider removal of builder-like API in PrePostProcessor, InputInfo, OutputInfo
* Added new Tensor dispatch and improvements, replace injections with inheritance of pybind objects, remove Blob from public python API.
* Clean-up tests and API from unused classes
* Remove unused import
* cpp codestyle
* Update AsyncInferQueue with python wrapper
* Codestyle cpp
* Applying comments
* Common tensor setting for requests
* add base ngraph::SoftSign operation wrapper; add substite softsign pass draft code
* add SoftSign convertor from ngraph to CNNNetwork
* add unit tests; write DoTransformation
* cpplint fixes
* add pass description
* cleanups; always use add layer
* remove bf16 since it is CPU specific
* add softsign evaluate unit test; code review fixes
* add unit tests on checking const values
* inherit softsign from :UnaryElementwiseArithmetic; code review fixes
* remove unneeded visit_attributes
* remove NGRAPH_TYPE_CASE macro
* use ngraph::op::util::get_single_value in pattern checking
* use legacy SubstituteSoftSignPass if there are FQ layers; build fix
* remvoe deprecated softsign tests
* fix deprecated error
* add draft unit tests on softsign operation with new api
* use new test API for softsign
* use another evaluate virtual method
* Add GatherND_8 operation
* Update shape infer function and tests
* Initial commit for nGraph GatherND_8 operation
* Add GatherNDBase class implementation
* Fix base class errors
* Add missrd header
* Update base class
* Update GatherND_8 implementation
* Fix codestyle
* Fix wrong rank
* Implement tests for gatherND_8 shape inference function
* fix codestyle
* Add limitation to doc
* Siplyfy check in shape inference
* Add more test cases
* Update shape inference function
* Add more test cases to cover all case with dynamic input shapes
* Update shape inference function
* Refactor tests
* Initial commit for gatherND_8 reference implementation
* Add visitor tests for gatherND_8 operation
* Add visitor tests for gatherND_8 operation
* Correct comment
* Add additional check is shape inference function
* Update shape inference implementation for gathernd operartion
* Fix codestyle
* Remove restriction for data is fully defined
* Update shape inference functon
* Add gatherND_8 reference to evaluate map
* Add tests for gathernd_8 reference in batch_dims > 1
* Fix codestyle
* Fix codestyle
* Fix missed check for nonetype
* Remove redundant checks for batch_dims
* Use get_shape method instead of get_input_shape in GatherND reference implementation call
* update reference implementation call
* Properly add GatherND_8 to test opset table
* Update GatherND reference implementation for support opset8 version
* Add unit tests for gatherND_8 reference implementation
* Update gatherND reference implementation call in evaluate map
* Mark failed tests for missed plugin gatherND_8 implementations
* Fix codestyle
* Fix codestyle
* Fix codestyle
* Update tests
* Move common methods to base class
* Fix codestyle
* Revert clone_with_new_input function moving
* Add more test for reference
* partially revert API changes
* revise LSTM spec
Signed-off-by: fishbell <bell.song@intel.com>
* add param check and related test case
Signed-off-by: fishbell <bell.song@intel.com>
* fix clang-format
Signed-off-by: fishbell <bell.song@intel.com>
* use static_cast to replace c style force conversion
Signed-off-by: fishbell <bell.song@intel.com>
* Migrate Movement2 (BatchToSpace-2, DepthToSpace-1, Roll-7, ShuffleChannels-1, SpaceToBatch-2, SpaceToDepth-1)
* Fix typos while copying
* Update to opset namespace and add i4, u4, bf16 tests when available
* Remove i4, u4 tests from BatchToSpace and Roll
* [GPU] fix constant shape if its user has scalar inputs
Currently for nodes like:
```
Multiply(input{}, mul_const{2}) -> mul{2}
```
cldnn sets 'mul_const' shape to (1, 2, 1, 1), which may not be correct if it's followed by
```
Add(mul{2}, add_const{2}) -> add{2}
```
because 'mul' shape is (1, 2, 1, 1), 'add_const' shape is (2, 1, 1, 1)
and the result becomes (2, 2, 1, 1) instead of expected (2, 1, 1, 1)
* Add tests for the issue
