* - add safety checks for mkldnn-assignement for quantized inner product

* - add asserts for unsupported data types in builder & emitter code of Quantized Dot

* Add fused version of Elu op

* Refactor ONNX importer prelu function to use fused op

* Style check

* Add docstrings

* Move make_constant_node to op/util

* Use make_constant_node helper

* Remove unneeded std:: prefixes

* Remove make_constant_node function, use builder::make_constant

* Remove redundant includes

* Add Elu to serializer

* Add Elu tests

* Add Elu tests to type prop

* Add Elu to list of ops unsupported on iGPU

* Add Elu to list of ops unsupported on iGPU

* Disable tests in iGPU manifest

* rework PerformanceCounter class

* Update intel gpu to new PerformanceCounter and enable performance on interpreter

* Add new HybridTensor class. Remove Tensor parent.

* upate more backends

* fix GPU tensor

* update intel gpu backend

Added NGRAPH_JSON_ENABLE to conditionally enable code/features relying on external JSON libraries (#2821)

* Update LLVM to 8.0.0

* Required codegen changes to use LLVM 8.0.0

* First rev of NGTF build from nGraph

* Copyright year update

* WIP

* Update test_ngtf_build.py

Updated with the latest repo URL

* Add NgraphTestCase test runner class

* Code improvements

* clang-format

* Add support for multiple output types

* Add reading inputs/outputs from files

* Refactor model_quant_conv_linear_3d

* Add shape checking to NgraphTestCase

* Add expected output shape to model_quant_conv_linear_3d

* Code review

* Remove small data files and move values to test code for legibility

* clang-format

* Refactor model_quant_conv_linear_2d

* Refactor model_dequantize_linear_1d_zero_scale_uint8_negative_axis

* Refactor model_dequantize_linear_1d_zero_scale_int8_4d

* Refactor model_dequantize_linear_1d_zero_scale_int8

* Refactor model_dequantize_linear_1d_zero_scale_uint8

* Refactor model_dequantize_linear_scalar_zero_scale_int8

* Refactor model_dequantize_linear_scalar_zero_scale_uint8

* Refactor model_quantize_linear_axis_negative

* Refactor model_quantize_linear_axis_zero

* Refactor model_quantize_linear_zero_point

* Add shape checking to NgraphTestCase::run

* Refactor NgraphTestCase::run

Workaround for gcc 4.8.2: Suppress mkl-dnn warning manually since pragma GCC diagnostic ignored is not working on gcc 4.8.2 (#2812)

* IntelGPU backend: Set of fusion operations

* avoid test failed after pr merge

* Add numeric_limits for bfloat16 type

* fix constexpr

* Fix broken test in shape specialization pass

* Roll back unnecessary changes to Add

* [ONNX] Extended types support for NgraphTestCase

* [ONNX] Move the value comparators to the NgraphTestCase class

* Add ConstantToBroadcast pass

* address review comment

* fix merge error

* Don't create a separate .so for python onnx importer.

* Pass outer make's -j values to sub make used in python_wheel.

* don't emit empty objects

* Add method to set the option to add output shapes to json

* more cleanup

* constant/io sizes

* add sizes and counts for all data io in a json

* Add details to model sizing

* change filemode

* change filemode

* fix filemode

* Adding auto pad to convolution

* Added auto pad to pooling ops and moved auto pad computation to utility method

* Added serializer support for autopadding. workaround for clang macro warning

* Style fix

* Addressed PR feedback

* Fix docstrings for same_upper and same_lower

* CMAKE_CFG_INTDIR does not work at install time. Use CMAKE_INSTALL_CONFIG_NAME on Windows.

* Rename NGRAPH_*_INSTALLSRC_DIRECTORY to NGRAPH_*_INSTALL_SRC_DIRECTORY

* wip

* Make a copy of the FunctionCall's internal function prior to compilation

* Make sure unit test is only included with interpreter and cpu backends enabled

* include rc version in NGRAPH_VERSION_SHORT

* Create a separate version string for python wheel.

// ... some content ... \
On Clang or AppleClang
PaddlePaddle sets -Werror=comment

* - add method to determine if the function contains nodes with dynamic shape
- add virtual method to node to determine if the node contains any partial shape.

* - add support to reshape op to determine if the reshape node as dynamic attributes

* Address PR comments

* - check all the inputs to determine if the node has dynamic_shape
- addressed Adam comments

* - add support to abort graph match if the nodes are dynamic

* Addressed PR comments

* - remove strict check on dynmic shape in the matcher

* Added EnumMask.

* added unit tests for EnumMask.

* fix.

* forgot to save file.

* add pass property to matcher.

* Clean up and made interface more restrictive, easier to use.

* addressed PR feedbacks.

* Move ONNX tests to the test/onnx subdirectory

* Change non-conforming test names

* Refactor quant and reshape tests to separate files

* Refactor tests into separate files

* Rename test in GPU excludes