* Function for retrieving number of node outputs.

* Handle optional trimmed outputs.

* Fix compilation err on clang.

* Fix error for number of outputs.

- Iterate over the minimum of number of outputs we return and the number
  of outputs of respective node in the graph. Some outputs may be
  optional and trimmed, as well as for some op implementations we may
  return not all outputs (ie. Dropout - where we do not return additional
  optional output).

* Update graph.cpp

* Add dropout ONNX op.

* Revert to iterate over node outputs in graph.

* Use more apropriate word in comment.

* Add QuantizedConcat

* Remove unused variables and add check for size of mins and maxes vector

* Resolve conflicts

* Merged with master and addressed some PR feedback

* Avoid float comparison

* make min/max vector, add dequant/quanti

* fix dequant/quant scales

* fix CI build issue

* use calls for new backend API in unit tests

* fix compile error

* fix compile error

* IntelGPU backend: Comvolution support for double and code minor clean up

* PR2479. custom kernel selection fix

* [ONNX] Overriding custom ops

* Add UT

* Style Check

* Review & style fix

Changing clDNN to latest commit.

* IntelGPU backend: Quantize operations

* Update intelgpu_op_custom_kernels.cpp

* User friendly assert message

* User friendly assert message

* Update UT

* Utility function for reading binary file content.

* Style apply.

* Review. Add sanity check on file size.

* Style-apply.:

* Auto detect OSX_SYSROOT for macos >= 10.14

* Fix potential regex issue.

* Use git shallow clone if CMake >= 3.6. Remove BUILD_BYPRODUCTS.

* Fix ONNX, protobuf integration.

* [ONNX] Enable OneHot operation

* Add UT

* Style check

* Change converts

* Change assert messages

* Update comments

* Update onehot.cpp

* rename tensor view to tensor

* one more

* revert change

* handle channelwise scale mask

* add Quantize op to cpu_layout

* enable Quantize to use input layout for output

* add comments for channelwise mask

* add Dequantize to cpu_layout

* resolve style-check

* fix unsupported layout issue

* fix s8s8 reorder issue

* Use uniform naming for external project prefixes and use INSTALL_DIR instead of SOURCE_DIR for referencing prebuilt external libraries.

* TBB: Use single variable to share path across files.

* pass HostTensor all the way down

* fix distributed build error

* Cleaner API doc reference for compile call

* Add a useful table for nGraph namespaces

* Remove layout namespace

* Show exploding kernel problem on illustration like IEEE preso

* Inplace convert for bitcasting ops

* Dont passthrough blocked layouts

* More checks for types not handled by mkldnn

* IntelGPU backend: Separate backend and executable classes (#2447)

* Add notice to main ToC page about distributed training

* adding quantized convolution builders for ONNX

* fix build error

* fix CMakeLists.txt spacing