* [LPT] INT16, INT32 Quantization support
* [LPT] Support build on platforms with size_t == unsigned int
* [LPT] Test and fix wrong constant
* Fix build for size_t = unsigned int
* [LPT] getDequantization: added precision limitation for dequantization constants
* [TESTS] GetDequantizationTransformation: added test case with unsupported precision
* add multimodel memleaks test runner
* Add support for is_equal_data, get_source_tensor, get_target_tensor methods in ONNX FE API (#6991)
* Change ngraph public api (#6920)
* Moved nGraph function
* Added legacy nGraph function
* Moved export API
* Moved variant
* Added old header for backward compatibility
* Introduce define
* [LPT] LP Transformations refactoring after dynamic shapes support (#6950)
* [LPT] Transformations refactoring after dynamic shapes support
* [LPT] ReshapeTransformation: 2D->2D fix
* [LPT] fixes after review
* [GPU] Fix ScatterNDUpdate unit tests (#7103)
* Deprecate ngraph file utils. Need to have common functions (#7105)
* [GPU] Get rid of memory alloc for input_layout in internal networks (#6897)
* Renamed component name (#7110)
* Propose new MaxPool-8 operation (#5359)
* MaxPool-8: pads_value attribute removal from the operator definition (#7119)
* Remove deprecated option and enable compilation without device (#6022)
* Build openvino wheel package from setup.py (#7091)
* Added ability to build wheel package by executing `setup.py bdist_wheel`
* fix linter issues
* fix formating
* remove blank line
* Support unregistered operations in MO IR Reader (#6837)
* Add support for unregistred operations in MO IR Reader
* Remove commented lines
* Add shapes equality check
* Update comments
* Update groupconv_to_conv function to support case with multiple destinations
* Add ir_data_attrs attribute to restored layers
* Update copy_shape_infer function to new graph api
* Add attribute IE to unsuppurted operations to save their attributes
* Fix wrong attribute name
* Update commentary
* Partially revert updating to new Graph API to fix regression, add appropriate comments
* Update code comments
* Rename copy_shape_infer function and add more comments
* First stage of the AUTO-MULTI merge: redirecting AUTO to the MULTI device plugin (#7037)
* Redirect -d AUTO to MULTI device plugin
Signed-off-by: Shoujiang Ma <shoujiang.ma@intel.com>
* Modify AUTO tests with MULTI config
Signed-off-by: Shoujiang Ma <shoujiang.ma@intel.com>
* Fix CI
Signed-off-by: Shoujiang Ma <shoujiang.ma@intel.com>
* Fix bug: CVS-62424
Signed-off-by: Shoujiang Ma <shoujiang.ma@intel.com>
* Add some tests for AUTO
Signed-off-by: Shoujiang Ma <shoujiang.ma@intel.com>
* Add select device logic to MULTI
Signed-off-by: Shoujiang Ma <shoujiang.ma@intel.com>
* Fix extract device name bug
Signed-off-by: Shoujiang Ma <shoujiang.ma@intel.com>
* Address reviewer's comment
Signed-off-by: Shoujiang Ma <shoujiang.ma@intel.com>
* Delete AUTO plugin source code
Signed-off-by: Shoujiang Ma <shoujiang.ma@intel.com>
* [CPU] Deform. conv. - reference enforced (#6945)
* [CPU] Bump up MKLDNN version to get fix of c26453 warning (#7089)
* OpenVINO ONNX CI Azure - update onnx/models on demand only (#7125)
* [CPU] Added improvements for StridedSlice (#6658)
* Removal of FusedOp inheritance leftovers (#7113)
* Remove FusedOp from v0::Gelu
* Update v0::Gelu NGRAPH_RTTI_DECLARATION
* Enable gelu type_prop tests
* Remove FusedOp from v0::MVN
* Remove FusedOp from HardSigmoid
* Remove FusedOp from LSTMSequence
* Remove supress deprecated
* Add missed NGRAPH_OP_SCOPE to v0 Gelu and HardSigmoid
* Move ngraph::element::Type to ov namespace (#7124)
* Moved ngraph::Type -> ov::Type
* Revert original files
* [GNA] Fix order of SwapMatMulInput transformations (#7137)
* Moved Dimension, PartialShape, Interval, Rank to ov namespace (#7136)
* Mo implementation for If with tf extractor (#6662)
* Add tf2.x impl for If
* Fix ir_engine
* Fix opset
* Fix BOM file
* Added new test
* Fix comments
* Add subgraph_utils
* Fix comments
* Fix transform
* code refactoring
* Fix description
* rewrite support for empty tensor in if
* added onnx extractor
* delete onnx_if
* fix bug with fake_outputs
* Fix test
* Fix control_flow and fix commentaries
* create method results_mapping_and_finding_fake_outputs(output_nodes_in_subgraph,
* [GPU] Fixed 'assigned to self' error in loop_inst.h (#7126)
* [GPU] Fix build for gcc 10 (#7142)
* [GNA] Set input scale factors for imported model (#7139)
* add doc:'Paddle_Support.md' (#7122)
* add doc:'Paddle_Support.md'
* Apply suggestions from code review
Co-authored-by: Tatiana Savina <tatiana.savina@intel.com>
* Apply suggestions from code review
* Update docs/IE_DG/Paddle_Support.md
Co-authored-by: Tatiana Savina <tatiana.savina@intel.com>
Co-authored-by: Tatiana Savina <tatiana.savina@intel.com>
* Remove local configs and it's copying to bin/ for stress tests (#7131)
* Moved DEPRECATION macro and ITT domains to ov namespace (#7153)
* Moved DEPRECATION macro and ITT domains to ov namespace
* Fixed code style
* Enable NormalizeL2Fusion and LeakyReluFusion inside MOC (#7096)
* Enable NormalizeL2Fusion inside MOC
* Fix NormalizewL2 decomposition (KeepDims=True)
* Add support for ONNX Crop operator (#6956)
* Review/update spec for NotEqual operation (#6797)
* Hiding the problem, Validate() changes 'function'
* Review/update spec for NotEqual operation
* Remove unnecessary edits not related to the ticket
* Removing the extra word binary for the short description
* Re-writing detailed description
* Correcting punctuation docs/ops/comparison/NotEqual_1.md
Co-authored-by: Tatiana Savina <tatiana.savina@intel.com>
* Specifying auto_broadcast in the short description is similar to Equal spec
* The range of values for auto_brodcast is similar to Equal spec and includes the missing pdpd
Co-authored-by: Tatiana Savina <tatiana.savina@intel.com>
* Moved DiscreteTypeInfo to ov namespace (#7127)
* Moved DiscreteTypeInfo to new opset
* Revert old header
* Fixed code style
* Revise CTCLoss OP (#6953)
* Add visitor test to CTCLoss
* Add CTC Loss SSLT
* Add CTC Loss template tests
* Use ngraph rtti macros
* Code style fix
* Enable PriorBoxClustered tests (#7078)
* CumSum spec revision (#6966)
* Update detailed description
* Update exclusive attribute description
* Update Inputs/Outpu description
* Update types
* Update descriptions
* Update data input rank info
* Added common.hpp file with aliases (#7158)
* CumSum reference implementation revision (#6915)
* New CumSum implementation init
* Unified ndim approach
* Move transpose to separate function
* Move transpose to original to separate function
* Move slice_count calculation to function
* Negative axes support
* Refactor redundant copy
* Changed copy to move
* Temp more backend tests
* Add const to shape arg
* Use span for slices calculation
* Remove unused headers
* CumSum new ref tests
* Add more ref tests
* Add all cumsum modes ref tests
* new optimized cum_sum reference
* Add reverse mode
* Optimized cumsum ref
* Remove deprecated cumsum backend tests
* Add more CumSum reference tests
* Simplify CumSum shared layer tests SetUp
* Replace auto to size_t in loop
* Change static_cast to T{}
* [LPT] MarkupCanBeQuantized: handled unsupported concat (#7045)
* [LPT] MarkupCanBeQuantized: added check on unsupported concat
* [LPT] ConcatTransformation: added test-case with unsupported concat and convolution
* [LPT] added test on rtInfo check for unsupported concat
* [LPT] ConcatTransformation: added test-case with unsupported axis to plugin tests
* CVS-56144 Enable all OMZ scope (#7084)
* Install layer tests with CMake (#6892)
* add CMakeLists.txt
* add copyright docstring
* add newline after copyright
* set target name
* change TARGET to DIRECTORY
* Rename layer tests dir to avoid name conflict
* cmakelists.txt final version
* Change destination to tests\layer_tests_openvino
* Add cmake_minimum_required to CMakeLists.txt
* Update CMakeLists.txt
* ReverseSequence specification refactored (#7112)
* ReverseSequence specification refactored
* Change attribute description to avoid confusion
* Allow seq_lenghts input to be of floating-point precision
* MemCheck add INT8 models to pre-commit (#7166)
* updated desktop configs with int8 models
* updated desktop reference configs with actual values
* added commit comments
* parametrize proxy (#7174)
* Updated list of supported operations. (#6981)
* Updated list of supported layers.
* Removed Crop, softsign from Kaldi list.
* Updated limitations.
* Corrected limitations.
* Updated limitations.
* Added Einsum, corrected Where.
* Apply suggestions from code review
Co-authored-by: Anastasiya Ageeva <anastasiya.ageeva@intel.com>
Co-authored-by: Anastasiya Ageeva <anastasiya.ageeva@intel.com>
* [MO] turn on MarkSubGraphsWithCorrectLayout for TF NCHW (#7150)
* turned on MarkingSubgraphsWithCorrectLayout for TF NCHW
* restricted MarkSubgraphsWithCorrectLayout.py only to TF
* added comments why need to MarkSubgraphsWithCorrectLayout even for TF NCHW models
* [CPU][TESTS][LPT] MatMulTransformations test-cases removed from skip config (#7181)
* Fixed ngraph_onnx_importer compatibility target creation for older cmake (3.10) (#7176)
* [GPU] Fix clBuildProgram failure with ssd_mobilnet_v1_coco and batch=256 (#7121)
* Fix v0::MVN default constructor (#7175)
* [GPU] Fixes for correct MultiDevice plugin and inference request behavior (#7161)
* Fix op category section in operations spec (#7130)
* add ngraph::pass::LSTMCellDecomposition as mandatory (#7028)
* add ngraph::pass::LSTMCellDecomposition as mandatory
* move LSTMCellDecomposition just after CommonOptimizations before all convert opset transformations
* code review fixes: add flag that prevents some legacy transformations if their ngraph-based analogues were executed
* remove isNgraphPassesUsed from ModelQuantizer
* cleanups
* [IE] Convert to unsigned NMS:0 ->Gather path (#6474)
* inserted Convert to unsigned
* moved declarations from hpp into cpp, specification corrected
* added static const modifier
* updated convert specification
* minor corrections
* split into 3 passes(Init, Propogate, Update), renamed final pass to ConvertNmsGatherPathToUnsigned
* added description why transformation is needed
* added matcher for several NMS versions, removed TRANSFORMATIONS_API macros from cpp
* applied comments:
- used GraphRewrite instead of FunctionPass
- simplified some expressions
- corrected case when Converts output goes to multiple nodes
- added to MOC transformations
- other minor corrections
* removed redundant namespace prefixes
* fixed #include <ngraph/pass/graph_rewrite.hpp>
* removed matcher_scope, debug code, and redundant dynamic_cast
* [nG] [IE] use GatherBase in negative indices resolver (#7145)
* updated pattern matcher into GatherBase in negative indices resolver, so that it is triggered in all versions of operation
* copy_runtime_info fix
* added constant folding
* Nested loop (#6710)
* initial changes to support nested loop
* fixed issues
* fixed nested loop extraction
* added comments
* removed unneeded comments
* review fix
* added tests
* turned off loop tests on GPU
* set xfail for TF tests
* removed TF test to move it in another repo
* fix typo in comment
* move duplicated code to separate functions;
added asserts
* add function for onnx constant creation;
add function to create body of loop
add comments to test
* move main change for nested loop to separate function
* install necessary dirs for tests (#7044)
* install necessary dirs to tests
* rem RUNTIME from install step
* fix paths
* fix install paths
* fix install paths: add destination dirs
* add pandas
* fix requirements conflict - change pytest version to ~5
* remove comment from requirements.txt
* upd numpy version
* Added openvino infer request API (#7151)
* [CPU] Removed eltwise overhead on execution stage (#6760)
* [GNA] For similar records, the pattern length was increased to 4 in the algorithm for determining infinite cycles. (#7165)
* for similar records, the pattern length was increased to 4
* Added comments
* [CPU] Enable direct copy implementation for u8->u8 reorder. (#7043)
* [CPU] Fix not expected No-Preprocess Exception with RGB to BGR conversion (#6954)
* [CPU] Avoid inserting additional transpose + reorder after RNN node. (#5921)
* [MO] Replacing StridedSlice with Squeeze/Unsqueeze (#6693)
* added reinterp_shape parameter to tf ss extractor
* removed reinterp_shape
* added transformation to replace ss
* updated bom
* fix for e2e tests
* updated a case when shrink_axis_mask and new_axis_mask are both initialized
* unittests
* added comments
* updated graph_condition
* comments resolving
* updated the case, when shrink_axis_mask and new_axis_mask are both initialized
* added layer tests for squeeze/unsqueeze cases
* remove case when shrink and new axis masks are both set
* [VPU] Added ConvertGather7ToGather1 pass to frontend (#7183)
This pr adds ConvertGather7ToGather1 pass to frontend before MergeGatherGatherElements pass, to make it so that when MergeGatherGatherElements is ran, any v7::Gather will be replaced with v1::Gather
* [MO] Add transformation for single CTCGreedyDecoder operation (#7023)
* Add transformation for single CTCGreedyDecoder operation
* Fix style in op specification
* Update transformation logic
* refactor old tests and add tests for new transformation
* Move tf specific front transformations to tf folder
* Update transformation logic and comments
* Add run_after function and update comments
* Add output_sparse_format attribute to extractor
* Update transformation conditions and tests
* Fix incorrect comment
* Move sparse_to_dense_replacer to front/tf folder to fix problems with class registration
* Update import
* Update output ports handling in transformation
* Update test
* Fix BOM file
* Update pattern for ctcloss transformation
* Fix and refactor tests for ctcloss transform
* Update transformation conditions
* Add support opset11 for gemm normalizer (#6733)
* Add support opset11 for gemm normolizer
* Add layer test for gemm opset 11
* Fix layer test
* Fix layer test
* Refactoring according to code review
* Fix
* Update biases norm
* Refactoring matmul norm
* Fix accoding to review
* Fix alpha parameter
* Fix variable naming
* Refactoring according to code review
* Add support for ONNX RandomUniform and RandomUniformLike ops (#7190)
* remove adaptive pool2d shape check in ngraph paddle frontend (#7074)
* remove adaptive pool2d shape check in ngraph paddle frontend
* add ngraph paddle frontend dynamic pool2d test
* Revise ReverseSequence reference implementation (#7117)
* ReverseSequence ngraph op shell revision with type_prop tests
* Add attribute count check in visitor test
* Refactor backend tests to template plugin test with reference values
* Rename cpu SLT instances
* Add op to list of trusted operations
* Rewrite validation check for input type due to backward compatibility
* Reference implementation speed up by replacing index function call of CoordinateTransform by precalculated strides
* Moved attribute_adapter, attribute_visitor files to ov namespave (#7179)
* Fixed nGraph build
* Fixed nGraph unit tests
* Fixed func tests
* Fix some operators
* Fixed build
* Try to fix specialization in different namespace
* Try to fix build
* Fixed element_type
* memleaks multimodel supporting
* revert rebase mistake
* add newline at the end of config file
* fix log messages
* refine memelaks test case class
* remove temporary decision designed to save in memory pipeline functions parameters
* code consistency
* rework example of new memleak tests config format
* oop in testcases
* mistype
* set num of iterations in example test config to previous value
* add multiproc stress unit tests
* Add more cases
* remove unique_ptr test objects saving logic
* switch memleak test configs to new format
* swith weekly memleak test config to new format
* Clarify new get_testdata script arg
* clang-format
* wrong changes
* Add docstring to generateTestsParamsMemLeaks()
* add explanation what update_item_for_name() doing
* Autodetect stress framework while parsing models
* adjust the wording
* Shorten test cases names
* fix get_testdata for memcheck tests
Co-authored-by: Mateusz Bencer <mateusz.bencer@intel.com>
Co-authored-by: Ilya Churaev <ilya.churaev@intel.com>
Co-authored-by: Vladislav Golubev <vladislav.golubev@intel.com>
Co-authored-by: Sergey Shlyapnikov <sergey.shlyapnikov@intel.com>
Co-authored-by: Vladimir Paramuzov <vladimir.paramuzov@intel.com>
Co-authored-by: Ilya Sharikov <ilya.sharikov@intel.com>
Co-authored-by: Michał Karzyński <michal.karzynski@intel.com>
Co-authored-by: Tomasz Dołbniak <tomasz.dolbniak@intel.com>
Co-authored-by: Daria Mityagina <daria.mityagina@intel.com>
Co-authored-by: Sergey Lyubimtsev <sergey.lyubimtsev@intel.com>
Co-authored-by: Anton Chetverikov <Anton.Chetverikov@intel.com>
Co-authored-by: Shoujiang Ma <shoujiang.ma@intel.com>
Co-authored-by: Yury Gaydaychuk <yury.gaydaychuk@intel.com>
Co-authored-by: Vitaliy Urusovskij <vitaliy.urusovskij@intel.com>
Co-authored-by: Rafal Blaczkowski <rafal.blaczkowski@intel.com>
Co-authored-by: Alexandra Sidorova <alexandra.sidorova@intel.com>
Co-authored-by: Katarzyna Mitrus <katarzyna.mitrus@intel.com>
Co-authored-by: Elizaveta Lobanova <elizaveta.lobanova@intel.com>
Co-authored-by: Eugeny Volosenkov <eugeny.volosenkov@intel.com>
Co-authored-by: Paul Youngsoo Ahn <paul.y.ahn@intel.com>
Co-authored-by: Mikhail Ryzhov <mikhail.ryzhov@intel.com>
Co-authored-by: Liu Bo <bo4.liu@intel.com>
Co-authored-by: Tatiana Savina <tatiana.savina@intel.com>
Co-authored-by: Olesya Martinyuk <olesya.martinyuk@intel.com>
Co-authored-by: Gleb Kazantaev <gleb.nnstu@gmail.com>
Co-authored-by: Nikita Semaev <nikita.semaev@intel.com>
Co-authored-by: Bartosz Lesniewski <bartosz.lesniewski@intel.com>
Co-authored-by: Anton Pankratv <anton.pankratov@intel.com>
Co-authored-by: Anastasiia Urlapova <anastasiia.urlapova@intel.com>
Co-authored-by: Gabriele Galiero Casay <gabriele.galiero.casay@intel.com>
Co-authored-by: Anastasia Popova <anastasia.popova@intel.com>
Co-authored-by: Anastasiya Ageeva <anastasiya.ageeva@intel.com>
Co-authored-by: Pavel Esir <pavel.esir@intel.com>
Co-authored-by: Maksim Shabunin <maksim.shabunin@gmail.com>
Co-authored-by: Andrew Kwangwoong Park <andrew.kwangwoong.park@intel.com>
Co-authored-by: Mikhail Letavin <mikhail.letavin@intel.com>
Co-authored-by: Evgeny Kotov <evgeny.kotov@intel.com>
Co-authored-by: Svetlana Dolinina <svetlana.a.dolinina@intel.com>
Co-authored-by: Victor Kuznetsov <victor.kuznetsov@intel.com>
Co-authored-by: Dmitrii Khurtin <dmitrii.khurtin@intel.com>
Co-authored-by: Ivan Novoselov <ivan.novoselov@intel.com>
Co-authored-by: Aleksandr Pertovsky <aleksandr.pertovsky@intel.com>
Co-authored-by: Nikolay Shchegolev <nikolay.shchegolev@intel.com>
Co-authored-by: Yegor Kruglov <yegor.kruglov@intel.com>
Co-authored-by: Polina Brzezinskaya <polina.brzezinskaya@intel.com>
Co-authored-by: iliya mironov <iliya.mironov@intel.com>
Co-authored-by: mei, yang <yang.mei@intel.com>
* Change sort functions to fix incorrect order
* Add separate sorts for different cases
* Update sort function to correct control flow edges handling
* Refactor all sorts for dictionary ordering
* Update build_graph function for correct control_flow edges creation
* Add tests for nodes and ports sort equality
* Fix wrong port value
* Add direct conversion from bool to int
* Use .replace instead of .strip
* Add test with both control_flow and ordinary edges, refactored edges lists filling
* Refactored nodes dict filling
* Delete unused test code
* Allocate internal buffer to usm_device when one of the input tensor is from usm_device.
Allocate output tensors if there is no user which is cpu impl.
* Move intermediate buffer allocation to primitive_inst
* Allocate to usm_host when the internal buffer is allocated close to limitation of device memory
* Remove internal_buffer_info and replace it with vector of layout.
Updated conditions to use alloc_type w.r.t the availability.
* Allocate internal buffer within primitive_inst construction
* Fixed device_mem allocation condition aligned with driver team
- Single allocation should be less than CL_DEVICE_MAX_MEM_ALLOC_SIZE
- Total allocation for a kernel should be less than CL_DEVICE_GLOBAL_MEM_SIZE
* Apply review comment
* Add set_argument/s methods to nGraph Python API
* Apply formatting
* Clear inputs in Node::set_arguments
* Run all tests in one container
* Fix formatting
* Add unit test
* Specify own CXX configuration for time && stress tests
* Included os_flags for build gflags
* Updated README
* Updated commands for build tests in README
* Use new cmake config
* [GNA] Add support for DWSC, other fixes and code refactoring.
* [GNA] Change supported layout to NHWC
* [GNA] Detect bias const only on second position, move verification of dwsc to matcher
* Added RandomUniformFusion transformation.
* Extended transformations for case with Convert, extended transformations for general min and max value case.
* Set to const unchanged variables.
* Apply suggestions from code review
Co-authored-by: Gleb Kazantaev <gleb.nnstu@gmail.com>
* Reformat code, small corrections.
* Added const shape checks.
* Fixed transformation for case of different const ranks.
* Added type checks.
* Apply suggestions from code review
Co-authored-by: Gleb Kazantaev <gleb.nnstu@gmail.com>
* United RandomUniformMulFusion and RandomUniformAddFusion to single transformation.
* Added negative tests.
* Used get_constant_from_source().
* Moved transformation to common fusions.
* Added const refs.
* Update inference-engine/src/transformations/src/transformations/common_optimizations/random_uniform_fusion.cpp
Co-authored-by: Gleb Kazantaev <gleb.nnstu@gmail.com>
* Changed to single class.
* Corrected IRs checks in layer tests.
* Small corrections.
Co-authored-by: Gleb Kazantaev <gleb.nnstu@gmail.com>
