Commit Graph
93 Commits
Author SHA1 Message Date
Vladislav Golubev 6630ae82d2 [LPT] Avoid AddTransformation for the cases which introduce Infinite constants after math calculations (#15496) 2023-02-08 13:06:24 +00:00
Ilya Lavrenov 1f3e469c5e Added -Wall for Clang and GCC (#15513)
* Added -Wall for Clang and GCC

* Fixes

* Don't use /J

* Fixed warnings

* Fixed warnings

* More fixes

* Fixed for MSVC

* Fixed more warnings on Windows

* Suppressed some warnings in template plugin

* Update src/tests/functional/plugin/shared/include/behavior/plugin/caching_tests.hpp

* Added suppression for PT FE

* Suppressed warnings in TF FE

* Suppressed warnings on Core unit tests

* Suppress warnings in python

* Suppressed Windows warning for 3rd party modules

* Suppresed one more warning
2023-02-08 15:01:00 +04:00
Edward Shogulin c1782d9e3d [LPT] clang merge warning fix (#15502) 2023-02-03 21:21:23 +00:00
Mang Guo 9e83b081f4 Add gather lpt transformation (#14597)
* Add gather lpt transformation

* Add per-channel gather lpt dequantization support

* Fix review comments

* Add GPU test case

* Fix clang-format error gpu case  build error

* Fix comments

* Fix clang-format check fail

* Update docs

* Fix comments

* Add Gather opset1 quantization support
2023-02-02 16:13:52 +01:00
Vladislav Golubev d1397b7b48 [LPT] Rank limitations removed (#14785)
* [LPT] LayerTransformation: removed legacy rank checks

* [LPT] Added test cases with 1D and 6D ranks & existing tests corrected
2023-01-31 00:26:59 +00:00
Mateusz Tabaka 3ad92084d6 Fix MarkDequantizationSubgraph when Mul has Convert on both inputs (#14997)
Ticket: 100042
2023-01-26 08:55:23 +00:00
Tomasz Jankowski 988847f559 Transformation component transition to OV namespace (phase 4) (#13558)
* Use openvino pass graph_rewrite

* Replace use of ngraph::pass to ov::pass

* Remove ngraph::pass aliases

* Update after rebase

* Fix post merge conflicts

* Apply code style

* Apply code style

* Remove bypassing inclusion

* Apply code style
2023-01-25 01:02:35 +03:00
Ilya Churaev 0c9abf43a9 Updated copyright headers (#15124)
* Updated copyright headers

* Revert "Fixed linker warnings in docs snippets on Windows (#15119)"

This reverts commit 372699ec49.
2023-01-16 11:02:17 +04:00
Oleg Pipikin b808fb7f4f Enable warnings for transformation tests (#14699)
* Remove warning suppression from tests

* Fix warnings MSVC ov_transformations_tests

* Fix warnings MSVC ov_lp_transformations_tests

* Fix code style
2022-12-17 01:54:24 +03:00
Oleg Pipikin 8e3c3f5c04 Enable clang-format for tests (#14667) 2022-12-17 01:54:07 +03:00
Oleg Pipikin 161a0a6319 Disable clang format for lp transformation test for fix CI (#14629) 2022-12-15 11:12:35 +01:00
Oleg Pipikin cc386d13ec Fix code style for transformation and lp transformations tests (#14073)
* Fix code style for transformation tests

* Fix code style for lp transformation tests
2022-12-13 15:19:09 +04:00
Vladislav Golubev b11adcdde4 [LPT] FQ Decomposition improvement (#14203)
* [LPT] FQ Decomposition modified to create FQ without constants with several consumers

* [LPT] Added subgraph test

* CPUTestsBase: Added the ability to check only fusing results

* [CPU] Added subgraph test

* LPT review comments applied

* CPUTestsUtils: added special string 'anytype' for selectedType check ignore
2022-11-30 01:04:10 +00:00
Vladislav Golubev f06c44115f [LPT] Check on isQuantized fixed (#14151) 2022-11-28 13:58:42 +04:00
Edward Shogulin cdb711a069 LPT: element-wise ops fuse extension to support updatePrecisions (#14141) 2022-11-25 13:49:03 +04:00
Mateusz Tabaka e105b3ea37 [LPT] fix weights output channel index in ConvolutionBackpropDataTransformation (#14124)
ConvolutionBackpropData weights have following shape {input channels, output_channels, ...},
so unlike forward convolution, the output channels are on the second dimension.
2022-11-23 09:01:59 +01:00
Oleg PipikinandIlya Churaev 6b4b01aecf Move transformation and lp transformations tests from IeFuncTests and IeUnitTests (#14096)
* Move transformation and lp transformations tests from IeFuncTests and IeUnitTests

* Enable tests in Azure CI

Co-authored-by: Ilya Churaev <ilya.churaev@intel.com>
2022-11-22 07:46:19 +04:00
2a7dc5404b [Core/CC]Verify add_matcher scope for Conditinal Compilation. (#13148)
* My verification:
libopenvino.so original size: 10799560
After adding this patch, new size: 10684744

Reduce: (10799560 - 10684744)/1024.0 = 112.125K.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* try to fix clang issue

* Merge xuejun/cc_test

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Test

Signed-off-by: xuejun <Xuejun.Zhai@intel.com>

* TEST1

Signed-off-by: xuejun <Xuejun.Zhai@intel.com>

* TEST2

Signed-off-by: xuejun <Xuejun.Zhai@intel.com>

* Complete Add_matcher wraper.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* register_pass rebase

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* fix clang issue.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* fix some replace error.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* 1: Rename ADD_MATCHER_SCOPE_WITH_OBJ -> ADD_MATCHER_SCOPE;
2: Remove debug info;
3: Fix bug: REGISTER_PASS_WITH_FALSE_

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Add ADD_MATCHER_SCOPE_WITHOUT_NSPACE macro

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* fix macro define issue.

* Add register_pass for cnn_network_ngraph_impl.cpp

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Try to fix vpux plugin fail issue.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Fix clang issue

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Merge macro "REGISTER_PASS_MODEL_SCOPE" and "REGISTER_PASS_FUNCTION_SCOPE" to "REGISTER_PASS_SCOPE"

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Rename "REGISTER_PASS_SCOPE_WITH_FALSE" to "REGISTER_DISABLED_PASS_SCOPE_WITH"

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Rename ADD_MATCHER_SCOPE_WITHOUT_OBJ to ADD_MATCHER_SCOPE_FOR_THIS
ADD_MATCHER_SCOPE_WITHOUT_NSPACE to ADD_MATCHER_SCOPE_FOR_THIS_WITHOUT_NSPACE

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* In order to implement CC, I have to move "StridesOptimization()" constructor to cpp file.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Remove "SCOPE" in macro.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Recover changes to fix clang issue.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Remove "REGISTER_PASS_MODEL_IF"

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* 1: Remove "ADD_MATCHER_FOR_THIS_WITHOUT_NSPACE";
2: Remove param "nspace" in Macro, replace with "using namespace " in local.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* fix missing update.

* For MACRO REGISTER_PASS:
1: Only keep a external macro define;
2: Judge 3 possibilities, if one of them is true, it will be disable;

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Optimize ADD_MATCHER_FOR_THIS.
-#    define ADD_MATCHER_FOR_THIS(region, ...)        add_matcher<region>(__VA_ARGS__);
+#    define ADD_MATCHER_FOR_THIS(region, ...)        ADD_MATCHER(this, region, __VA_ARGS__)

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Update src/common/conditional_compilation/include/openvino/cc/pass/itt.hpp

Co-authored-by: Ilya Churaev <ilyachur@gmail.com>

* Update src/common/conditional_compilation/include/openvino/cc/pass/itt.hpp

Co-authored-by: Ilya Churaev <ilyachur@gmail.com>

* Update src/common/conditional_compilation/include/openvino/cc/pass/itt.hpp

Co-authored-by: Ilya Churaev <ilyachur@gmail.com>

* Update src/common/conditional_compilation/include/openvino/cc/pass/itt.hpp

Co-authored-by: Ilya Churaev <ilyachur@gmail.com>

* clang format

* Update OR defiation in MACRO, increasing code scalability.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

* Fix window compiles error.

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>

Signed-off-by: Yan, Xiping <xiping.yan@intel.com>
Signed-off-by: xuejun <Xuejun.Zhai@intel.com>
Co-authored-by: xuejun <Xuejun.Zhai@intel.com>
Co-authored-by: Ilya Churaev <ilyachur@gmail.com>
2022-11-20 19:02:42 +03:00
Tingqian Li 80971ae8e8 [CPU] Optimize fused FQ/Eltwise nodes to use less binary postOps (#13502) 2022-11-16 00:26:02 +03:00
Tomasz Jankowski c5eb25d3ea Transformation component transition to OV namespace (phase 2) (#13557) 2022-11-15 04:40:48 +03:00
Xiake Sun 7ab81d0950 Add missing opset5 header for LSTMSequence in LPT (#13953) 2022-11-11 09:43:31 +03:00
5eef0298d9 Transformation component transition to OV namespace (phase 1) (#13469)
* Move GeluFusion into ov namespace

It required to move utils as well.

* Move AddFakeQuantizeFusion into ov namespace

* Move AlignEltwiseInputRanks into ov namespace

* Move BatchToSpaceFusion into ov namespace

* Move BinarizeWeights into ov namespace

* Move BroadcastElementwiseFusion into ov namespace

* Move ClampFusion into ov namespace

* Move Concats and Eliminations into ov namespace

* Move many passes and such  into ov namespace

* Move many passes and such into ov namespace

ngraph_ops
op_conversions

* Move control_flow into ov namespace

* Apply style

* Move smart_reshape into ov namespace

* Move opset_conversions into ov namespace

* Move low_precision into ov namespace

* Apply style

* Move fused_names into ov namespace

* Add missing aliases

* Move few passes into ov namespace

ConvertPrecision
FixRtInfo
InitNodeInfo

* Move CommonOptimizations into ov namespace

* Fix convert_precision conflicting namespaces

* Apply code style

* ov-nize Type Relaxed definitions

* Remove useless aliases

bulkly added by script

* Revert opsets

* Address review comments

* Remove NGRAPH_RTTI

* Clean up remains of yet another implicit inclusion

Co-authored-by: Andrei Kochin <andrei.kochin@intel.com>
Co-authored-by: Ivan Tikhonov <ivan.tikhonov@intel.com>
2022-11-11 02:52:19 +03:00
Egor Duplenskii c1733b5542 [CPU][LPT] Support quantized RNN (LSTM and GRU algorithms) (#12981) 2022-11-10 23:48:51 +03:00
Mateusz Tabaka 154850e8ca Eliminate dequantization shift when zero point == 0 (#13353)
Ticket: 91111
2022-10-27 01:37:13 +02:00
Ilya Churaev 2a344b66cb Update to the latest cmake for Windows (#13604)
* Update to the latest cmake for Windows

* Skip new warnings

* Added compiler flags for old cmake

* Disabled some warnings

* Suppress new warnings

* Fixed some errors

* Disable some warning for lpt

* Fix some warnings

* Disable new warnings

* Fix some warnings

* Disabled Was error for arm

* Fixed some errors

* Fixed some new warnings

* Fixed auto plugin

* Fixed Windows warning

* Fixed Windows

* Disable some warnings

* Some changes

* Small test fixes

* Fixed python build

* Added skip warnings for legacy python API
2022-10-25 18:11:46 +04:00
Vladislav Golubev a25c2ba665 [LPT] Precision restrictions extending (#13172)
* [LPT] Precision restrictions extending

* Compilation error fixed

* Markup for SubgraphOps fixed

* cleanup

* Review fixes: removed vector<size_t> wrapper in PrecisionRestriction class

* Review fixes: added comments

* docs compilation fix
2022-10-19 12:29:32 +02:00
Vladislav Golubev 372fe475c9 [LPT] Removed legacy limitations on dequantization propagation for several transformations (#13048) 2022-10-10 10:58:23 +04:00
Ilya Churaev 8a9c19e3eb Warning as error for Windows (#13291)
* parent 6e7016ccda
author Ilya Churaev <ilya.churaev@intel.com> 1664281499 +0400
committer Ilya Churaev <ilya.churaev@intel.com> 1664510018 +0400

Fixed warnings on local machine

* Added CMAKE_COMPILE_WARNING_AS_ERROR usage

* Fixed style

* Fixed merge conflicts

* Fixed typo

* Fixed myriad build for macOS

* Fixed warning

* Fixed tests

* Disabled incorrect test

* Try to fix linux tests

* Revert "Try to fix linux tests"

This reverts commit 29224c93ff.

* Fixed tests

* Revert logic with incorrect cast

* Fixed log softmax

* Disable warning as error for cuda

* Try to fix inference_engine_s

* Fixed cmake

* Revert "Fixed cmake"

This reverts commit 87e9e4e674.

* Revert "Try to fix inference_engine_s"

This reverts commit a1adca8b05.

* WA for static symbols in inference_engine_s test library

* Fixed code style

* Fixed static definition for master

* Revert "Fixed static definition for master"

This reverts commit 20d00d215a.

* Revert "Fixed code style"

This reverts commit 0eb2362543.

* Revert "WA for static symbols in inference_engine_s test library"

This reverts commit 75ef86a79d.

* Fixed linker issue for Windows

* Disable WaE by default

* Disable warning as error in the developer package

* Try to fix dev package

* Try to fix Windows Jenkins

* Revert old behavior for tread_warn_as_err variable
2022-10-06 13:44:21 +04:00
Vladislav Golubev 88e4ac5e53 [LPT] Fixes for the cases with convert before subtraction constant (#12835) 2022-09-06 20:41:29 +03:00
Egor Duplenskii a89c4cfc3f [LPT] Correct a check for whether model is quantized (#12364)
Look inside subgraph operations, such as TensorIterator, Loop, If, etc
2022-08-03 20:36:01 +04:00
River Li 0ef4ccb86f [CC]Fix CC issue for transformation (#12292)
* Revert "Fixed 3 naming issue"

This reverts commit a92d3cfff5.

* Revert "Fix CC issues for transformation and snippets"

This reverts commit d08a3f5aac.

* Fix NGRAPH_PASS_CALLBACK issue to make it can work

* Fix matcher name missing issue
2022-07-27 16:17:53 +04:00
a571539107 Optimize FullyConnected FakeQuantize post-ops (#11819)
* Optimize FullyConnected FakeQuantize post-ops

* matmul bias fuse

* Add simplifyToScale for FakeQuantize and use it in FC and Conv.

* Add fakequantize documentation

* Update doc and fix accuracy issue

* Update doc

* Fix accuracy regression

* Generalize the judgment Criteria about fake quantization with scale

* Update document

Co-authored-by: Zhang Yi3 <yi3.zhang@intel.com>
Co-authored-by: xuchen-intel <chen.xu@intel.com>
2022-07-05 09:39:42 +08:00
River Li 0571124fd3 Fix CC issues for transformation and snippets (#11798)
* Fix CC issues for transformation and snippets

Matcher should be enabled if it was hit during analyze stage.

* Fixed 3 naming issue
2022-06-13 13:36:35 +08:00
Edward Shogulin 542a374c40 [LPT] Introduce new quantization mode attribute (#11380) 2022-04-04 13:27:03 +03:00
Edward Shogulin 5d821453ae [LPT] Introduce new granularity attribute instead of OperationPerTensorQuantizationRestriction (#11330) 2022-04-03 19:35:04 +03:00
Andrey Zaytsev cad355a03e Docs labels adjustment (#11227) (#11294)
* Adjusted documentation labels

* Renamed images

* fix doc tests

Co-authored-by: CCR\ntyukaev <nikolay.tyukaev@intel.com>
# Conflicts:
#	docs/IE_PLUGIN_DG/ExecutableNetwork.md
2022-04-01 15:06:55 +03:00
Egor Shulman 23476c8eee CC for LPT transformation call in CPU plugin (#11341)
* Use CC in LPT

* Applied comment
2022-03-31 12:35:27 +03:00
Ilya Lavrenov 19d0e5ba52 CMAKE: IE_VERSION => OpenVINO_VERSION (#11242)
* IE_VERSION => OpenVINO_VERSION

* Reverted installation of python unconditionally
2022-03-28 19:32:21 +03:00
Vladimir Zinoviev 4a2d0f39dd [LPT] Turn back checks in reshape transformation when subtract is absent (#10939) 2022-03-15 11:34:12 +03:00
Ilya Churaev ad1c4a24c3 Deprecate version inside DiscreteTypeInfo (#10781)
* Deprecate version inside DiscreteTypeInfo

* Fixed code style

* Fixed openvino for macOS

* Fixed build for macOS

* Fixed errors for Windows build
2022-03-14 21:18:00 +03:00
Ilya Churaev 0bc6196d96 Migrate to new RTTI for all transformations and graph structures (#10703)
* Migrate to new RTTI for all transformations and graph structures

* Fixed code style
2022-03-14 06:57:21 +03:00
Vladimir Zinoviev eebe8c70f9 [LPT] Fix out of bounds access in reshape (#10791) 2022-03-11 18:04:14 +03:00
Edward Shogulin 5be402750a [LPT] FuseConvert transformation extension (#10558)
* [LPT] FuseConvert transformation extension

* [LPT] Tests

* [LPT] Cleanup & tests refactoring
2022-02-22 02:02:11 +03:00
Nikita Demashov f7a85c59fe [LPT] Disable Move Fake Quantize on shuffle channels pattern (#10389)
* added shuffle channels check

* refactoring
2022-02-21 10:01:37 +03:00
Edward Shogulin 17311c46b3 [LPT] checkElementwise extending for 1D tensor (#10498) 2022-02-18 09:41:17 +03:00
Vladislav Golubev ff4e97ab09 [LPT] Security fixes (#10465) 2022-02-17 13:47:27 +03:00
Vladislav Golubev fa4246d531 [LPT] Security fixes (#10381) 2022-02-16 10:31:17 +03:00
Ilya Znamenskiy 523adff17a [GPU] Fully connected int8 optimizations, some fixes, better fused ops support (#10035) 2022-02-15 12:33:16 +03:00
Mingyu Kim 13c024b7a3 Remove unnecessary cout message (#10346) 2022-02-15 16:14:56 +09:00
Maksim Derbasov bb0d82f724 Fix warnings (#10278) 2022-02-14 07:48:41 +03:00