* IntelGPU backend: Use clDNN matrix operations for nGraph::Dot

* Update unit_test.manifest

* Update README.md

Fix broken link

* Update README.md

* ONNX Pad operator

* Disable Batch Norm is spatial checking

* Remove unused variable

This reverts commit c5b082c6.

* onnx: return signle nGraph function with multiple outputs
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* onnx: remove excessive code
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* Update ngraph python unit tests

* wip

* wip

* style check checking

* try cmake version of check/apply style

* update style check to print list of files in error

* cleanup

* IntelGPU backend: Use clDNN BatchNorm in operations

* IntelGPU backend: no duplications in cldnn output vector

* IntelGPU backend: PR2086. Fix clDNN batchNorm usage

* Additional info calculated & printed for all_close_f.

* Fixed format

* Tweaked test calls to all_close_f to make unambiguous.

* Tweaked test calls to all_close_f to match close_f for last unit test.

It seems wasteful to be generating debug output a character at a time
all the time, only to discard it on non-debug builds.  This change
replaces that behavior by introducing a no-op operator<< for the debug
log on non-debug builds.

* Add argmReduce support for int32 +  tests

* add new line

* add to intelGPU manifest

* Address Fenglei's comment

* address comments

* Small change to emitter logic.

* Chris's comments incorporated

* minor edits + clang

* edit

* Only synchronize stale input tensors

* Cache backend.call() compilations

* Cache invoker bindings

* Update test list

* New branch for doc changes

* Slightly out of scope doc requests to make some people happier;
this does technically change the API

* Doc versioning even with Beta release

* Don't remove markdown version of contrib guidelines after all

* Update changes from earlier branch

* Make sure ngraph-tf links to readme

* Update setup.py

* Update setup.py

* fix errors in edit

* fix ascii error

* get the right readme

* rename a variable at Scott's request

* Add license files for:
eigen: eigen-license.txt
googletest: googletest-license.txt
json: json-license.txt
llvm: llvm-license.txt
mkl-dnn: mkl-dnn-license.txt
mklml: mklml-license.txt
numpy: numpy-license.txt
tbb: tbb-license.txt

* Add tensorflow license.

* Add license files and cleanup compiler/linker options.

* Update comment on c++ standard option flag

* copy license files to install directory (#2132)

* Update to core docs

* Fix rendering of graphic

* Fix batchnorm argument order, cleanup some comments, fix backprop

* Merge error

* Clean up training function, organize inference test

* BatchNormInference tests

* Training case

* Training test

* Fix autodiff BatchNorm test

* Cleanup

* Move file to doc checkout

* Update disabled test name in igpu manifest
Fix unnused variable

* Unit tests disables

* Review comments

This re-enables `CPU.batchnorm_bprop_n4c3h2w2` was disabled. This test seems to have a checkered past:

* Back when it was called `cpu_fusion.bn_bprop_n4c3h2w2` we were observing some random crashes on macOS; this was before we officially supported macOS, so we didn't worry about it.
* More recently, the test was disabled in #2058 because it "just started failing" (no further details available).

I have tried to reproduce the failure with repeated gtest runs as follows, with no success:

* macOS Mojave with DEX
* Ubuntu 16.04 with gcc-5.4.0 and DEX
* Ubuntu 16.04 with gcc-5.4.0 and codegen (pre-built LLVM)
* Ubuntu 16.04 with clang-3.9 and DEX
* Ubuntu 16.04 with clang-3.9 and codegen (non-pre-built LLVM)

I think the best thing to do is to re-enable it, and fix it if the failure crops up again. (Who knows, maybe CI will trip over it!)

* Fix uninitialized parameter tensor data in nbench

* Remove const modifier in the loop to fix compilation error

* Add condition check before writing tensor data

* Remove the condition check when initializing pramater tensor data

- All "<>" and all \"\" are in two separate groups.

Use @loader_path for the calling library/executable
Use @rpath for dependent libraries.

* disable squeeze layout optimizations for padded layouts

* knob squeeze padded layouts