* IntelGPU backend: helper function improvement

* PR1541. Change string to stringsteam

* Validate/infer types as a virtual function
BroadcastLike, ScalarConstantLike
Set output shape separately from output types
Remove some obsolete tests
Replace *Like ops with fixed type versions

* Review comments

* Remove TensorViewType

* Remove PrimaryTensorView

* Reviewer comments

* Style fix.

* Another try at GPU

* GPU

* Simplify class hierarchy

* More of previous.

* Better errors

* Remove unneeded classes

* Merge

* Update Python bindings

* clang-format

* Remove redundant files

* Remove commented out lines

* clang-format

* Group includes

* Missing tests

* Merge error

* Expose onnx_import C++ interface in Python API.

* Pretty printing.

* Add Computation interface using Ngraph Function objects plus UT.

* Apply code format.

* Remove unnecessary stream open mode.

- Code formatting.

* Fix onnx_import submodule visibility.

- Folder restructurization.

* Fix some small errors.

- Wrong function type annotations.
- Class doc.
- Code formatting.
- Class inheritance from object.

* Use modified Runtime class interface.

* Add model for test_onnx_import.

* Revert back to old API.

- Use of Function object in Computation class.

* Use of previous verions API.

* Small refactoring

* Code cleanup

* onnxifi: upgrade libonnx
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* onnx: upgrade libonnx to version 1.3
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* [ONNX] Concat operator

* Style fix

* only print details if details enabled

* refactor print routines to main file

* refactor for multi dir support

* dir support prints nice results

* use line comments instead of multiline comments for license header

* update more

* update new files

* more header updates

* style

* normalize doxygen keywords

* more normalization

* normalize more onnx files

* yet more normalization

* fix doxygen error

* [ONNX] Sum op

* [ONNX] Generic variadic op template

* Add support for Min op

* clang-format

* Add support for Max op

* Add support for Mean op

* Docs, code cleanup

* Docs, code cleanup

* IntelGPU backend: SigmoidBackprop operation

* Requested changes done

* Delete namespaces in function names

* Apply code format check

disabled RNN test to workaround RNN unit test failure on MAC due to bug in MKLDNN scratchpad creation (#1502)

* IntelGPU backend: AvgPoolBackprop operation

* Requested changes done

* Make gen_window_loop() static & initiate win_elems_size through shape_size

* [ONNXIFI] Fix namespace

* Revert "[ONNXIFI] Fix namespace"

This reverts commit acbf95afafd451c0b23adc8cd83563feeb1c2816.

* Include cstddef

* Update onnxifi.cpp

* Update onnxifi.cpp

* address klocwork warning

* fix PR comments

* onnx: upgrade to version 1.3 pre-release
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* onnx: fix to specific commit from master branch
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* fix style

* [ONNX] Softmax operator

* Review fix pt. 1

* Review fix pt. 2

* Add softmax test

* Update onnx_import.cpp

* cmake: use GNUInstallDirs to autodetect installation directories (Linux)
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* define LINUX variable to detect Linux
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* Add Travis testing on CentOS
* Disable building CPU Backend
* Remove Python dependencies from CentOS for now

* normalize comments

* address review comments

* [ONNX] MatMul operator

* Add NL on EOF

* Review fix pt. 1

* onnxifi: add 'onnx' as external project
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* onnxifi: public interface function stubs
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* onnxifi: fix the typo in function name
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* onnx: fix compilation Ubuntu 16.04
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* onnxifi: update to dummy backend
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* onnxifi: style cleanup
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* Bug fixes to account for padded layouts and correct handling for reshape that transposes and changes shape

* Use nullptr instead of dummy buffer in memory primitive creation

* add cuda reduce

* clang format

* fix bugs

* fix bug

* add 1d reduce

* clang format

* fix bugs

* unroll loop

* remove debug info

* revert tests

* unroll 1D reduce op

* add comments

* using cudnn for nd to scalar reduction

* remove cuda 1d reduction since cudnn version is faster

* remove 1D kernel

* fix variable name

* resolve Chris's comments

* non_reduce_in_strides to non_reduce_strides