* Fix ops, cleanup ToC, add latest illustrations

* PR feedback

* Support non-mkldnn fallback for Batchnorm training bprop in CPU backend

* Skip unit test on PlaidML backend

* Change method of splitting data and filters for groups

* Add support for groups included in filters shape

* Add UT for groups in weights shape

* Add helper function to check if groups are in filters shape

* Review fix I

* Remove unused variables

* Skip test on PlaidML

* Review Fix II

* Internaly store groups as dimension instead of size_t

* Review fix III

* Add updating of m_groups in pre_validate_and_infer_types()

* Provenance for builders (#3644)

* Provenance for builders

* Tests

* Tests

* Update src/ngraph/node.hpp
Co-Authored-By: Sayantan Sarkar <sayantan.sarkar@intel.com>

* style

Finish migration

* Migration to master

* Handle op constructors that construct ops

* Update provenance for topk

* Add missing input (#3691)

* Add missing input

* If builder does nothing, don't dive into the inputs

* Reverse test

* simplify, add test

* Add test

* add test

* Add RandomUniform op

* Missing files

* Make CPU compile again (RandomUniform tests still don't pass)

* Bop users of existing RNGState class and reorder some comment junk

* Try to bop RNGState in gcpu

* Add RandomUniform to CPU manifest

* Add RandomUniform to PlaidML manifest

* Add .rst for random_uniform

* Clean up junk in the .rst

* Change UniformRNGState to always use double internally

* Change weird test failure message

* RandomUniform implementation for CPU backend

* Add missing file

* Add codegen support

* Code cleanup

* Address review comments

* Fix CI failure

* Fixed CI error

* Fixed CI error

* [MLIR] Add support for parsing nGraph tensor type

Initial commit that enables nGraph parsing. It's needed for testing.

* [MLIR] Enable nGraph dialect in ngraph-opt

This PR registers nGraph dialect in ngraph-opt and prepares
nGraph lowering pass for LIT testing, fixing all the related issues.
Among other things, lowering pass has to be turned into a function pass,
dead argument in constructor was removed and `convert-ngraph-to-affine`
flag was added.

* Fix issue with function name and multiple functions

* Extend module_function.mlir lit test

* [MLIR] Add support for parsing nGraph element types

It introduces initial support for parsing nGraph signed/unsigned
integer and floating point data types.

* [MLIR] Improve gather op lowering

This PR interchanges indices and param loops in gather lowering so
that a better memory access patter is generated. Fusion of gather
with other ops is also observed with this change.

* LayerNorm (#3630)

* Constructors.

Type prop.

Decompose LayerNorm.

Add serialize.

* Add dummy test case.

* Add dummy type prop test.

* Fix some build errors.

* Remove build errors.

* Update decompose for bprop.

* Change begin_norm_axis default value to 1.

* Style.

* Reorder class members.

* Add actual type prop tests.

* Add fprop test.

* Working on bprop test.

* Bprop tests.

* Allow flattened scale and bias.

* Add support for flattened scale and bias.

* Fix incorrect type_name.

* PlaidML: Decompose fused_op LayerNorm

* Update Backprop constructors.

* PlaidML: Add missing header file.

* Remove doc about removed param.

* Fix type prop tests.

* PlaidML: Disable unit test.

* Fix stats flattening axes bug.

* Upgrade description to type_info.

Aligned with contrib/mlir directory ownership. This prevents
that Scott is added as reviewer by default.

* [MLIR] Add support for parsing nGraph tensor type

Initial commit that enables nGraph parsing. It's needed for testing.

* [MLIR] Enable nGraph dialect in ngraph-opt

This PR registers nGraph dialect in ngraph-opt and prepares
nGraph lowering pass for LIT testing, fixing all the related issues.
Among other things, lowering pass has to be turned into a function pass,
dead argument in constructor was removed and `convert-ngraph-to-affine`
flag was added.

* Fix issue with function name and multiple functions

* Extend module_function.mlir lit test

* Improve module_function.mlir test

Remove ngraph to affine dialect conversion since we just need to verify
that we can parse and print modules and functions.
Add verification for parsing the printed code.

* [MLIR] Add support for parsing nGraph element types (#3665)

* [MLIR] Add support for parsing nGraph element types

It introduces initial support for parsing nGraph signed/unsigned
integer and floating point data types.

* Improve LIT tests

Test parsing and printing of types separately from lowering to affine
since these tests will evolve differently, particularly for tensor
types.

* Missed file

I left this file behind in the previous commit

* Instantiate only valid cases for binary op constant folding kernels

* removed unused parameter

* latest PlaidML-GPU updates for 0.26

* More theme updates

* Cleaner theme to match site

* Update ngraph_theme

* Update Release Notes page for 0.26

* update theme WIP

* Update theme w/feedback

* Add top navigation menu

* more theme customization and cleanup

* Refine more details on new ngraph theme

* Refine some xsys font-kerning

* Theme updates

* Add latest release notes

* Update HE transformer doc

* Fix rendering

* Resolve doc build warnings

* Add doc versions

* test code removed

* remove test code

LIT creates "Output" folders for each directory with LIT tests.
Before this change, LIT tests were executed in source directory and
"Output" folders had to be removed manually. Git also showed them as
folders to be potentially tracked. This PR changes LIT execution to
build directory to avoid this problem. Similar approach is used in MLIR.

* Strided slice

* Strided slice

* default value for strides

* Added new strided slice test, enabled old tests, refactoring

* Refactoring

* Autogenerated file: dyn_replace_slice tests

* Renaming

* Fix codestyle

* Fix build on MacOS

* Fix codestyle

* Add several tests in unit_test.manifest to skip it on PlaidML

* Disable all dyn_replace_slice tests on PlaidML

* Add new enum: RoundingType for pooling operations

* Add v1::AvgPool op

* Add v1::MaxPool op

* Fix comments format

* Fix problem with forward declaration

* new UT & fix some bugs

* add pdpd autob spec and ut

* add pdpd autob to pass check

* add dummy dynamic ut

* add dynamic test case

* use int64_t

* pass dynamic ut

* add pdpd style kernel

* handle pdpd validation

* add dummy pdpd style bcast

* implement pdpd style bcast

* add validation

* use output arg

* use ngraph_check
Co-Authored-By: Adam Procter <adam.m.procter@intel.com>

* enable error print

* accuracy mismatch in plaidml

* fix ut on windows

* make separated tests

* fix trailing one case

* fix warning

* [MLIR] Encapsulate internal nodes of CompiledKernel Op to prevent later passes from accessing them.

* Fix style and namespace.

* Address PR feedback.

* Move encapsulate_nodes from pass to CompiledKernel Op.

* Address PR feedback.

* Follow MLIR naming convention.

* Allow TBB to be excluded from CPU build.

* Change some CMake scope.

* m_use_tbb can be removed if TBB is not enabled.

* Revert some CMake linkage scope as it breaks unit test.

* Migrate changes from #3652.

* [MLIR] Add support for parsing nGraph tensor type

Initial commit that enables nGraph parsing. It's needed for testing.

* Clang format

* Fix return in parser

* [MLIR] Add `ngraph` prefix to MLIR flags

Some flags collision with some MLIR flags.

* [MLIR] Add support for nGraph tensor type in parser

Initial commit that enables nGraph parsing. It's needed for testing.

* Rename ngraph print flag

* Rename ngraph dump mlir flags

* Clang format

* Revert "[MLIR] Add support for nGraph tensor type in parser"

This reverts commit ae371d6a5c8ea590322d5d3b9ba110159d4bf5fa.

* Update MKLDNN to v0.21

* Changed for MKLML asset for WIN32

* test bfloat16 with CPU maxpool

* Move test to cpu_test

* convert result from bfloat16 to float

* Modify result op to support bfloat16

* Correct comment

* Add a simple convolution test case

* initialize floats with float literals

* Guard bf16 tests with MKLDNN_VERSION_MAJOR >= 1

* [ONNX] Added function for converting negative axes.

* [ONNX] Bugfix

* [ONNX] Code formatting

* [ONNX] Changed name of function.

* Added validation for axis/axes.

* Changed abs to llabs for int64 argument.

* Changed validation function.

* [ONNX] Added function for converting negative axes.

* [ONNX] Bugfix

* [ONNX] Code formatting

* [ONNX] Changed name of function.

* Added validation for axis/axes.

* Changed abs to llabs for int64 argument.

* Changed validation function.

* Changed validation function.

* Changed validation function

* Changed rank size in onehot

* Apply suggestions from code review
Co-Authored-By: Tomasz Socha <tomasz.socha@intel.com>

* Code refactoring.

* Update flatten.cpp

* Convert remaining dynamic_pointer_casts for ops
Make is_type a function instead of a method
Direct implementation of `is_parameter`

* Add exports

* Missing includes

* headers

* missing export

* Refactor cpu/builder/dot.cpp

* Support only float and i64

* Refactor softmax and reduction to support only float and i64

* Add a new file for customized kernels

* Add compile time flag NGRAPH_CPU_LARGE_BINARY

* Add compilation flag checks in reduction.hpp

* change dot to only support float and i64

* Revert "Refactor cpu/builder/dot.cpp"

This reverts commit 0d53f27fde64872aff096f12ee9b79e5db7a7fee.

* style

* Consolidate macros

* Refactor slice ops, reshape & pad

* cleanup

* Gather op

* Concat, Convert and broadcast op

* Tile and reverse_sequence op

* Scatter_add op

* ET other then fp and i64 go through ref

* tests passing

* Consolidate macros

* Address feedback

* Fall back to reference for pad and reshape

* add scatter_add reference fallback

* Undo concat change

* Undo tile and update slice as they dont have reference implementation

* Remove update slice condition

* Gather op type check

* change routine name

* Build-time configurability for datatypes that requires optimized kernels on CPU backend (#3592)

* Inline function

* Change condition

* VS compiler workaround

* Add comment for VS workaround

* More fixes for VS

* More wrapping of nested macros for VS

* More wrapper

* variable name change and more wrapping of macros

* test

* Style and refactor

* Wrap macros

* Add a seperate macro for SELECT_KERNEL

* Change SELECT_KERNEL_3ARGS

* Unwrap couple of macros

* Syntax

* Add a new macro for fixed number of args for VS compiler

* Comment/Fake ectx

* Add detailed comment for the workaround for VS compiler

* Comment all unused ectx variables

* Templated calls to reference kernels

* const args

* Change softmax ref definition to take double

* Hardcode softmax kernel to take double ..testing

* Fix softmax

* Latest illustrations of nG components and features

* Update ABOUT.md

* Update full stack diagram

* fix typo

* Updating as per discussion

* Gather:v1 was introduced

* Added support for negative axis

* Removed unsused serialization

* Code review remarks introduced

* Change returned type of get_axis method

* Code review remarks introduced

* Chnaged axis_node to scalar during transformation

* Clang style applied

* Fixed clang errors

* style