* 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

* Temp.

* Put all the dummy files.

* Remove some compile errors.

* WIP: Add gather and gather_nd kernels.

* Temp save.

* Update comments for gather.

* Implement reference gather.

* Validate and infer shape.

* Style.

* Fix compile issues.

* Add serializer support.

* Fix interpreter compilation issues.

* WIP: Add UT

* WIP: Add UT

* gather_nd UT passing.

* Fix gather with no axis.

* Fix gather issue.

* Update unit_test.manifest for backends and add gather, gather_nd  support for generic cpu.

* Add type_prop tests.

* Add CPU builders.

* Fix codegen.

* Make some UT numbers more readable.

* Style.

* Update Copyright Year

* Update Copyright Year

* Fix Typo.

* Remove unused variable.

* fix nv gpu build error

* Fix intel gpu compilation.

* Add basic docstring.

* Allow 1D indices for gather_nd.

* Allow scalar indices for gather.

* Update unit_test manifest files.

* Style.

* Add indices element type check and add failing type_prop checks.

* Update docstring.

* Fix incorrect test names in unit_test.manifest

* [ONNX] Support for Gather op

* Remove unneeded broadcast

* Remove unused include

* Set correct default value for axis

* Add some unit tests for provenance with node replacement

* One more test

* add provenace propagation

* Adds ability to control whether provenance is on or off

* Add some unit tests for provenance with node replacement

* One more test

* add provenace propagation

* Adds ability to control whether provenance is on or off

* Removes extraneous semicolon

* Fixes include header errors

* Reimplements provenance enable flag using global variable

* Fixes style issues.

* Adds newline at the end of file.

* Fixes function return type.

* Adds RAII mechanism for localizing provenance enable state in testing.

* Improves getter/setter method name to match variable name.

* Adds s_ prefix to global variable provenance_enabled.

* [ONNX] Unify broadcast code in ngraph core and in ONNX bridge

* Use using a bit less

* Update CPU backend.

* Update INTERPRETER backend.

* Update unit tests.

* Don't use std:: prefix.

* Convert to bool for GPU backend in DEX mode.

* Handle boolean elements in Intel GPU backend.

* Disable convert to bool UT for IGPU.

* Update quoting style to fix errors raised by updated version of flake8.

* [ONNX] Erf test for integer values

* [ONNX] switch off model_erf_int32 test on GPU

* - pass tensor through reference in the lambad closure of cpu reference_erf kernnel
- add Erf unit test case to verify codepath for int32 values

* fix clang errors

* deconv optimizations for dcgan

* Added test cases

* modified some tests, not working at this point

* Removed temp code

* fixes to get unit test to pass

* Added node validation checks

* Update mkldnn emitter to memory reuse design

* Code cleanup

* Fix to enable deconv select the right kernel

* Fix file permissions

* Disabled unit test cases

* Remove unused variable

* Address PR feedback

* Removed dead code

* Style check

* removed dead code

* Temp.

* Put all the dummy files.

* Remove some compile errors.

* WIP: Add gather and gather_nd kernels.

* Temp save.

* Update comments for gather.

* Implement reference gather.

* Validate and infer shape.

* Style.

* Fix compile issues.

* Add serializer support.

* Fix interpreter compilation issues.

* WIP: Add UT

* WIP: Add UT

* gather_nd UT passing.

* Fix gather with no axis.

* Fix gather issue.

* Update unit_test.manifest for backends and add gather, gather_nd  support for generic cpu.

* Add type_prop tests.

* Add CPU builders.

* Fix codegen.

* Make some UT numbers more readable.

* Style.

* Update Copyright Year

* Update Copyright Year

* Fix Typo.

* Remove unused variable.

* fix nv gpu build error

* Fix intel gpu compilation.

* Add basic docstring.

* Allow 1D indices for gather_nd.

* Allow scalar indices for gather.

* Update unit_test manifest files.

* Style.

* Add indices element type check and add failing type_prop checks.

* Update docstring.

* Fix incorrect test names in unit_test.manifest

* Missing header

* [ONNX] Erf op support

* [ONNX] Missing CMakeLists entry for the Erf op

* [ONNX] model_erf unit test excluded from GPU

* Update quoting style to fix errors raised by updated version of flake8.

* select op add types

* add a test for double for select

* Update intelgpu_op_custom_kernels.cpp

* remove ngraph::aligned_alloc and replaced with AlignedBuffer

* allocate a minimum sized buffer is zero size is requested

* move memory alignment value to static constant

* move static constant to constexpr function

* IntelGPU backend: Custom kernels refactoring

* IntelGPU backend: remove unused header

* [ONNX] Fix default axis for legacy broadcasting

* Add new function to get default axis for legacy broadcasting

* Add std:: prefix

* Remove unnecesary default_axis function

* Style check

* Update ONNX CI

Don't send email notification when build was aborted.

* Added force parameter

* Cleanup workspace

* Cleanup docker container workspace

* - Moves some fused convolution ops to core FusedOps
- Adds support for decomposing and replacing multi-output FusedOps
- Adds query callbacks to FusedOpDecomposition to check if a FusedOp is
  supported by a backend
- Adds core fusion patterns for FusedOps
-

* style fix

* Added comments on FOP_FUSIONS

* gpu convolution 1d bug fix (#2741)

* Fix bug with dex-only compilation and addressed PR comments

* Revert "add logic in replace_node for provenance propagation (#2703)" (#2731)

This reverts commit 6c8284a3.

* Migrate doc changes to r0.18 (#2738)

* Migrate doc changes

* Add TensorFlow version change

* Use only convolution direct algo for non float convs (#2752)

* Build branch from scratch if building upon cached master fails

* Revert "Build branch from scratch if building upon cached master fails"

This reverts commit 0141d4151d0fbfd76733708833180e36adbd51b2.

* Revert "Revert "Build branch from scratch if building upon cached master fails""

This reverts commit 64cf85c101664544a1db8736bc6ae35cb075ef77.

* Update ONNX CI workflow

- Periodically build base nGraph image in a separate job
- Use that base to build branches in CI
- If that fails - build from scratch

* Moved base builder script

* Fixed ngraph repo path

* Fix branch name

* Fix CI root path

* Docker login must be done explicitly

* Fix credentials variable

* Fix pullImage method name

* Removed Docker registry addresses

* Set default branch to master

* Clone branch specified as parameter

* Fix try catch error

* Fix build script execution

* Added docker_registry parameter to start container script

* Calling setConfigurationStatus directly

* Revert "Calling setConfigurationStatus directly"

This reverts commit ea16a54ed1919f08bf0e5e1c4e8284146bcbaf6d.

* Fix missing $

* Fix password redirection to stdin

* Removed internal address

* Fix closing bracket

* Replace assert and NGRAPH_ASSERT with NGRAPH_CHECK; deprecate old classes

* Rename node_validation_assertion_string to something slightly more descriptive

* Fix syntax error in GPU BE

* Fix some messed-up includes

* More comments, more tests

* Fix reference to properly include check.hpp (should unbreak codegen)

* Add check.hpp include to gpu_util.cpp

* Comment updates

* Comment update

* Oops

* PR feedback: better formatting when context info or explanation are missing

* Add NodeInput and NodeOutput classes

* Deprecate Node::get_inputs, Node::get_outputs, Node::get_output_inputs. Remove Node::get_input_from and Node::get_output_from

* Privatize most fields of Node

* Make deprecation of descriptor-munching classes optional

* Review comments

* Adapt ReshapeSinking to use raw pointers for NodeInput

* Fix ZDTE (thought I had already done in this branch, weird); style

* wip

* Change get_node_outputs() and get_node_inputs() to return vectors

* Updates after merge

* Whoops, forgot to define these functions

* {NodeInput,NodeOutput} -> {Input,Output}

* Kill shared_ptr in Output

* Move Input and Output into node.hpp

* Templatize the underlying node (sub)type in Input and Output

* Eliminate some get_input_* and get_output_* functions

* Change get_outputs and get_inputs back to their original names; rename NGRAPH_DEPRECATE_IO_DESCRIPTORS to NGRAPH_DEPRECATE_OLD_NODE_APIS

* Miscellaneous cleanup

* More cleanup

* Unbreak CPU build

* Simplify unit tests

* Make Node less friendly

* Deprecate more get_output_* and get_input_* functions

* A couple of PR comments

* Make the deprecation stuff more generally available

* Better comment

* Be more consistent about [] vs. at

* batch dot WIP.

* cpu backend refactor and unit tests pass.

* WIP.

* batch dot interpreter impelementation.

* minor clean up.

* more clean up.

* patching the gpu backends.

* added more tests, fixes, etc.

* fixed compile error.

* renamed batch dot to batch matmul.

* refactor WIP.

* fixes some tests and formating.

* more fixes.

* add top_k operation

* modify topk definition

* Update intelgpu_backend.cpp

* Update intelgpu_op_custom_kernels.cpp

* add topk to graph visualization

* enable index_element_type in graph visualization

* minor changes

* Serializer supports new pad api

* Addressed reviewer comment

* Fixed compiler warning

* Enable previously CI-failing test

* Disable max_3d_to_scalar_int32, still failing on CI

* Remove tests

* [Standalone] Decouple MKLDNN primitive build from code generation

This patch introduces a new pass, MKLDNNPrimitiveBuildPass, which
iterates over all the ops assigned to MKLDNN and builds their
corresponding primitives. Primitive indexes are stored in MKLDNNEmitter
and can easily be retrieved with the get_primitive_index(node)
interface. This decouples the creation of primitives from codegen and
fixes the problem of MKLDNN primitives being created twice
(CommonFunctionCollection pass and codegen).

Current assertions only allow the creation of a single primitive per
node but it should be simple to remove this when needed. Using a pass
might not be the best approach here but I found it convenient for the
current needs and it should be straightforward to convert into a utility,
if needed.

These changes caused a conflict with recently introduced
'build_quantized_inner_product*'. These new build methods will be ported
in a follow up patch to new build approach.

* Remove unrelated comment

* Remove TensorView code

* Set m_node_primitive_map from MKLDNNPrimitiveBuildPass

* Move node->primitive map from mkldnn pass to external function

* Fix struct/class inconsistency en fw declaration

* Add Quantized MatMul OP

* Rename qmatmul -> quantized_matmul

* Use auto everywhere