Commit Graph

4497 Commits

Author SHA1 Message Date
Andrew Kwangwoong Park
28f708c991
[GPU] Add optimized cldnn detection output GPU implementation (#6634)
Co-authored-by: Taylor Lee <taylor.lee@intel.com>
2021-08-10 14:53:09 +03:00
Dmitrii Khurtin
4835896582
Convolution to matmul (#6803)
* [GNA] Remove transposes around MatMul

* Added tests for transformation HandleTransposesAroundMatMul

* Move IsTransposeSupported function to GNA limitations file

* Correct transpose insertion tests

* added TransposeAfterMatmul tests and moved InsertTransposeBeforeMatmul tests to handle_transposes_around_matmul.cpp

* added inifitiry loop checker and memory concat test

* fixed build errors

* changed the conditions for selecting an input of Concat for ScaleFactor calculation when entering an infinite loop

* fixed after review

Co-authored-by: Elizaveta Lobanova <elizaveta.lobanova@intel.com>
2021-08-10 14:08:17 +03:00
Evgeny Kotov
cc76d38920
remove InsertIdentityToLSTMCellPass and fix InsertIdentity (#6962)
* check is node final non-functional for grouping; remove InsertIdentityToLSTMCellPass

* code style fix
2021-08-10 13:39:16 +03:00
Maksim Kutakov
7c82ad78ee
[CPU] Add reorder if the constant memory is not aligned, and isa is SSE (#6912) 2021-08-10 11:58:57 +03:00
Ilya Churaev
e3fa6544f2
Enable IE clang format (#6938)
* Add clang-format config files for IE and include files

* Fixed IE core clang-format

* Added clang-format for plugin API

* Fixed style for plugin API

* Fixed code style

* Fixed build

* Added proposed values

* Fixed code style

* Updated config

* Fixed # for define

* Fixed comment and build

* Removed clang-config for include, fixed include for library headers

* Set column limit to 120 and space before CtorInitializedColon=true

* Fixed headers

* Added support of FOREACH_CHILD

* Fixed parameter print style

* Fixed code style

* Fixed target name

* Restore comments style for public API

* Fixed plugin API

* Applied code style

* Fixed CI
2021-08-10 11:15:08 +03:00
Vladislav Volkov
cc5dba95d4
[CPU] Memory leaks in gemm module (#6989) 2021-08-10 11:03:57 +03:00
Yury Gaydaychuk
bf6c34d7b8
[CPU] Interpolate handles inplace child layout (#6250) 2021-08-10 09:46:36 +03:00
Mikhail Nosov
c4bd0a45d3
MO: Support of discovering of suitable MOC frontend (#6888)
* MO: Support of discovering of suitable MOC frontend if --framework is not specified

* Ready for review

* Fix: don't use FrontEndManager if framework is not in list of available frontends

* Apply review comments
2021-08-10 09:23:30 +03:00
Dawid Kożykowski
2a5584791c
Revise LogicalOr operation reference implementation (#6867)
* add tests for logical or op

* remove redundant overrides

* add inputs/outputs tensor check

* create type_prop typed test for logical ops

* add new line in logical_or.cpp file

* refactor logical_and type_prop test

* update test labels

* beautify or.hpp file

* fix formatting to match clang-format

* beautifyfunctional test files

* move validate_and_infer_elementwise_logical() implementation to validate_and_infer_types()

* refactor logial or functional test to

* refactor logial aA functional test t

* update constants file

* add file to instantiate TEST_P and avoid test execution duplication

* add missing  empty lies at the end of files

* remove unused variable
2021-08-10 08:38:11 +03:00
Ilya Churaev
565627a416
Add ov::runtime::Core (#6946) 2021-08-10 07:49:37 +03:00
Polina Brzezinskaya
0c8a5d5279
[VPU] Removes MYRIAD specific tools (#6908) 2021-08-09 20:35:03 +03:00
iliya mironov
eadeae6c47
Fix ChangeOutputTypeAttributes BackReplacementPattern (#6949)
* Hot fix

* Add unit test
2021-08-09 19:22:21 +03:00
Vladimir Gavrilov
838e701e5e
ExpandRangeConstant adjustment for bidirectional Broadcast (#6739)
* Fixes in the transformation ExpandRangeConstant.

* Fixed test.

* Now we use use ShapeOf for both inputs of Broadcast.

* Now the transformation ExpandRangeConstant uses two Gather layers.

* Deletec commented code.

* Fixed tests for the transformation ExpandRangeConstant.

* Rewritten the transformation ExpandRangeConstant using Select.
2021-08-09 16:49:07 +03:00
Ilya Lavrenov
6a63cb9122
Enabled naming style check for FEM, ONNX FE (#6970)
* Enabled naming style check for FEM, ONNX FE

* Used custom version of ncc

* Added self-check

* Removed submdoule

* Moved to public ncc
2021-08-09 15:38:08 +03:00
Dawid Kożykowski
461584ffb1
Update spec for LogicalOr operation (#6852)
* update markdown file for logical or operator

* update information on broadcasting in logical operators spec

* change *T* to *T_BOOL*

* add information about input shape

* remove divergence between logical_or and logical_and specs
2021-08-09 13:50:48 +03:00
Min, Byungil
9899264d61
Resolve python build issue on Windows (#6517)
Signed-off-by: Min, Byungil <byungil.min@intel.com>
2021-08-09 12:53:29 +03:00
Anastasiya Koryachikhina
ad9078d9eb
add itt_collector build (#6859)
* add itt_collector build

* changed format for itt_collector build

* replace build_dir with sea_itt_lib_path

* change type of build_target

* replace os.join with Path

* change built_target arg type
2021-08-09 12:40:56 +03:00
Gabriele Galiero Casay
aa3645cb53
Avoid template test from executing multiple times (#6948) 2021-08-09 11:42:05 +03:00
Mateusz Bencer
3826a0d08d
Remove leaky relu alpha check (#6910)
* remove leaky relu check

* make alpha_node scalar
2021-08-09 07:26:37 +03:00
Gleb Kazantaev
b11a2220b0
Enable ConstantFolding inside MO Backend (#6916) 2021-08-07 12:40:46 +03:00
Ilya Lavrenov
5d781afe73
Naming convention tool integration (#6967)
* Fixed cycle dependency

* Added NCC infra

* Add submodule

* Added substitution for clang lib

* target depends on target_ncc_check

* Try find_package(Clang QUIET)

* Added github actions

* Try U18 + clang9

* Fixed nnc

* Updated style

* Works for clang-9

* Works for clang-7

* Fixes
2021-08-07 09:55:45 +03:00
Ilya Lavrenov
141f401575
Fixed cycle dependency (#6959) 2021-08-06 21:15:18 +03:00
Aleksandr Voron
74ac48a297
Update CMakeLists.txt (#6965) 2021-08-06 21:14:44 +03:00
Szymon Irzabek
862176d1d4
[GNA] Add support for strides with dilations and POT fake quantize layers (#6900) 2021-08-06 14:45:11 +03:00
Gabriele Galiero Casay
371fc7aee4
Enable low precision types in ConvertLike operation reference implementation (#6890)
* Enable low precision types for ConvertLike operation

* Migrate backend unit test suite to template plugin reference tests

* Fix typo in naming convention

* Avoid duplication of template plugin tests execution

* Add file to instantiate TEST_P and avoid test execution duplication
2021-08-06 11:19:16 +03:00
Edward Shogulin
43f18da413
[LPT] Legacy compliance restrictions removal all: Reshape (#6870)
* [LPT] Reshape: legacy compliance restrictions removal

* [LPT] comment fixes
2021-08-06 10:27:34 +03:00
Ilya Churaev
fc39303677
Moved IE public headers to IE library (#6941)
* Moved IE public headers to IE library

* Fixed documentation

* Fixed fast build

* Fixed doc
2021-08-06 10:05:51 +03:00
Sergey Shlyapnikov
f231fbd551
[GPU] Fix invalid memory access in replace_all_usage fucntion (#6906) 2021-08-06 08:40:34 +03:00
Mikhail Ryzhov
89fba123bb
[GNA] Added support of FQ layers for outputs (#6905)
* [GNA] Fixed FQ pass for several outputs

* Added tests

* Used the latest opeset8
2021-08-05 22:12:52 +03:00
Edward Shogulin
c08bc75766
[LPT] Optimize quantization scheme for SPR/ATS (#6808)
* [LPT] Optimize quantization scheme for SPR/ATS

* [LPT] updatePrecisions = false support
2021-08-05 21:03:03 +03:00
Mateusz Bencer
e4dfff387b
Apply moc transformations on FE API path (#6871)
* apply moc transformations

* changed type of net in apply_moc_transformations

* review remarks

* args->argv typo
2021-08-05 18:08:40 +03:00
Sergey Shlyapnikov
79e9190838
[GPU] Fix incorrect fusions indexes for eltwise ref kernel (#6936) 2021-08-05 13:40:52 +03:00
Michał Karzyński
01a5db7cc8
Add support for Affine operator (#6884)
* Add support for Affine operator

* Add support for Affine operator
2021-08-05 12:45:06 +03:00
Egor Duplensky
f584cf9727
[CPU] Update batch limit right after setting it (#6513)
Seem to be more convenient than checking and updating batch limit
in scope of Infer request call.
2021-08-05 08:21:16 +03:00
Patryk Elszkowski
1b4d456f8e
remove legacy fake quantize reference implementation API (#6810) 2021-08-05 07:50:47 +03:00
Yuan Hu
a913950039
revise tan op (#6567)
* revise tan op

Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>

* update doc

add examples in desciption
add the unit of measure
clear input type

Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>

* add template plugin test case for int type

Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>

* add template plugin test case for uint and float

remove the float test in backend

Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>

* modify document

change type to any supported numeric type

Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>

* fix compile error in openvino-lin

Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
2021-08-05 07:48:58 +03:00
Ilya Churaev
031381dc07
Small api changes (#6923)
* Deprecate specialize_function

* Changed RTMap

* Fixed python build
2021-08-04 21:34:11 +03:00
Mikhail Nosov
12e2b3e4a9
Pyngraph tests: Install & run on Azure-Linux CI (#6927)
* Pyngraph tests: Install & run on Azure-Linux CI

* Typo correction

* more ci corrections
2021-08-04 20:29:49 +03:00
hyunback kim
f00dabc184
Add input image scale option in benchmark App (#5848)
* Add input image scale flag in benchmark app.

- user set input image scale with -iscale.
  input is divided by scale.

Signed-off-by: hyunback <hyunback.kim@intel.com>

* Apply image scale, mean parameter in benchmark APP

Means and sacles values per channel

Signed-off-by: hyunback <hyunback.kim@intel.com>

* Fix clang-format

Signed-off-by: hyunback <hyunback.kim@intel.com>

* fix clang-format issue2.

Signed-off-by: hyunback <hyunback.kim@intel.com>

* Update benchmark tool to align the format of mean and sacle values with MO arguments.

Signed-off-by: hyunback <hyunback.kim@intel.com>

* Remove debug print.

Signed-off-by: hyunback <hyunback.kim@intel.com>
2021-08-04 19:51:46 +03:00
Anastasia Kuporosova
585f211e0a
[IE Python API] fix test query state write buffer (#6928) 2021-08-04 19:43:41 +03:00
Aleksandr Voron
cad700444e
Update CMakeLists.txt (#6840) 2021-08-04 17:58:20 +03:00
Liu Bo
45982fc30f
supported Paddle Op list (#6687)
* supported Paddle Op list

* fix 'interp' Ops limitation description

* Op limitation update for 'range' and 'arg_max'

remove 'range' Op limitation and add output limitation for 'arg_max'  Op.

* Update Supported_Frameworks_Layers.md

update limitations of 'fill_constant_batch_size_like', 'slice', 'unsqueeze2'

* Update Supported_Frameworks_Layers.md

update limitation way

* Update docs/MO_DG/prepare_model/Supported_Frameworks_Layers.md

Co-authored-by: Anastasiya Ageeva <anastasiya.ageeva@intel.com>

Co-authored-by: Anastasiya Ageeva <anastasiya.ageeva@intel.com>
2021-08-04 17:21:51 +03:00
Vladimir Paramuzov
29b527353a
[GPU] Replace internal TBlobs with mappable ClBlobs in all InferRequests to ensure correct work on discrete GPUs (#6922)
Co-authored-by: Mikhail Letavin <mikhail.letavin@intel.com>
2021-08-04 16:53:56 +03:00
Gleb Kazantaev
73e8e91ffd
Enable NopEliminatoin inside MO Backend (#6874)
* Refactor NopElimination; Execute NopEliminatoin inside MOC Backend

* Add missing header

* Refactor tests; use weak_ptr to avoid excess consumers for node

* Add test to check number of shared_ptr usages for graph nodes during Manager execution

* Remove AlgebraicSimplification forewer
2021-08-04 13:53:06 +03:00
Anton Romanov
0acfc8b5ae
Update weights size (#6909) 2021-08-04 13:37:24 +03:00
Anastasia Kuporosova
dd742de105
[Python API] removal of deprecated python api (#6282)
* [Python API] removal of deprecated python api

* [Python API] removal of deprecated python api
2021-08-04 11:46:39 +03:00
Anton Pankratv
4d8825e4f8
Stream executor config returns strings (#6830) 2021-08-04 11:09:30 +03:00
Zhang Yi
0aa6b07628
[FrontEnd]enable pdpd ops conversion part3 (#6636)
* [FrontEnd]enable pdpd ops conversion part3

* Add adaptive pool2d op conversion (#1)

* param support tensor (#2)

* add missing sync_batch_norm

* Update pow.cpp

* deal empty axis (#5)

* deal empty axis

* apply review comments

* fix code style

* fix code style

* change shape to i32

* fix code in shape

* fix code style

* fix paddle code style

* remove redandent ops

* fix maxAdativePool

* fix expand_v2

* remove redandent code

Co-authored-by: Mang Guo <mang.guo@intel.com>
Co-authored-by: Luo Cheng <cheng.luo@intel.com>
2021-08-04 11:04:19 +03:00
Alexey Lebedev
6f23458534
disable debug postfix on mac (#6827)
Co-authored-by: Your Name <you@example.com>
2021-08-04 10:45:41 +03:00
Mateusz Tabaka
8c4cee1b58
[GPU] Handle negative axis in concat op (#6839)
* [clDNN] Handle negative axis in concat op

That enables following models for onnx importer path:
yolact-resnet50-fpn-pytorch
yolact-resnet101-fpn-pytorch
yolact-darknet53-fpn-pytorch
swin-tiny-patch4-window7-224
action-recognition-mkinetics-res34-mhsa
driver-action-recognition-adas-0002-decoder
horizontal-text-detection-0001

* Add tests for negative axis in Concat op
2021-08-04 10:44:22 +03:00