Commit Graph

103 Commits

Author SHA1 Message Date
Tomasz Jankowski b1db3448e7
[ONNX] Extend ONNX Importer for operation "Einsum" (#6074) 2021-06-16 14:57:29 +02:00
Tomasz Dołbniak 9e9e33e772
Axes input handling in ReduceSum evaluator (#5614) 2021-06-10 18:13:56 +03:00
Katarzyna Mitrus bc7f61be24
PRelu reference implementation and ReshapePRelu transformation alignment (#5915)
* Apply  ReshapePRelu transformation only to const slope input

* Remove xfail from onnx backend prelu_broadcast test

* Fix and add Prelu SLT

* Update PRelu mkldnn transformation to extend broadcast support

* Fix and update PRelu reference implementation

* ONNX Prelu tests

* Add prelu backend tests

* Update ie tests manifest

* Comments clean up

* Fix STL Fill leakyslope blob

* Code refactor

* Unify layer tests slope input values generation
2021-06-08 06:51:41 +03:00
Ewa Tusień c77a13ae20
Add support for SparseTensor in Constant op (#5871) 2021-06-02 12:30:33 +02:00
Jan Iwaszkiewicz c3ca8d048e
[ONNX] ArgMin/ArgMax support for select_last_index (#5661)
* Add Reverse Op to opset

* Worksave with Reverse path

* Add last_index support

* refactor argminmax factory

* Remove old xfail, add new one

* Fix proto file for argmax

* Rewrite test for select_last_index

* Add CPU tests to Manifest

* Update manifest

* Remove Reverse from opset7

* Refactor arg_min_max factory

* Added example comment in arg_min_max

* Codestyle changes
2021-05-25 20:37:07 +03:00
Tomasz Socha 97f020481a
[ONNX][ONNX Editor] Generate an input with unknown shape and type for operators without schema. (#5531) 2021-05-24 11:04:11 +02:00
Mateusz Bencer 3dc1d7f457
[ONNX Editor] Use port indices instead of tensor names, new naming convention of inputs (#5668) 2021-05-21 12:24:51 +02:00
Michał Karzyński 87d6664669
[ONNX] Do not reshape input to DequantizeLinear if scalar (#5260)
Co-authored-by: Ewa21 <ewa.tusien@intel.com>
Co-authored-by: bsledz <bartosz.sledz@intel.com>
2021-05-19 11:26:22 +02:00
Mateusz Bencer 880b45a770
Add high level API to ONNX Editor (#4927) 2021-05-17 10:11:49 +02:00
Bartosz Sledz 81f9d78541
ONNX Subgraph - Infer types of inputs from top level Graph nodes (#5277) 2021-05-12 10:32:34 +02:00
Ewa Tusień f824f914a4
[ONNX] Add Constant op from opset 13 (#5450) 2021-05-11 11:01:27 +02:00
Tomasz Socha 64a032fa18
[ONNX] Add ONNX Compress operator. (#5400)
* [ONNX] Add support for ONNX's Compress operator

* NonZero fixes

* Add unit tests as constant_network

* Style

* Move compress tests

* xfail python tests

* Fix func test

* Review Fix I
2021-04-29 07:49:05 +03:00
Tomasz Dołbniak 36f407dd18
Proper handling of ONNX Squeeze without the axes input (#5425) 2021-04-28 21:59:32 +03:00
Tomasz Dołbniak b65003ac59
Preserve the input names when extracting subgraphs (#5207) 2021-04-21 13:16:01 +02:00
Mateusz Bencer 7ac7215924
[ONNX] Add support for ONNX ConstantFill op (#5203) 2021-04-14 16:11:21 +02:00
Bartosz Sledz 94e9b632e8
Revise ONNX Mod operator (Add support for attribute fmod == 0) (#5173) 2021-04-13 10:24:51 +02:00
Mateusz Tabaka 34385eb45d
[ONNX] use Reshape instead of Broadcast in InstanceNorm (#5148) 2021-04-12 12:03:00 +02:00
Tomasz Dołbniak 3b0c6aa408
Run IR models via TestCase (#4964) 2021-04-01 11:22:31 +02:00
Tomasz Dołbniak 170223d842
DeformableConv2D custom ONNX op (#4816) 2021-03-30 14:44:44 +02:00
Alexey Suhov 6478f1742a
Align copyright notice in python scripts (CVS-51320) (#4974)
* Align copyright notice in python scripts (CVS-51320)
2021-03-26 17:54:28 +03:00
Ilya Churaev b76e965e95
Remove obsolete cmake script and test data (#4944)
* Remove obsolete cmake script

* Removed obsolete models

* Removed obsolete generators

* Remove obsolete markdowns
2021-03-25 07:45:47 +03:00
Tomasz Jankowski 636986be3f
[ONNX] Add dynamism to Conv bias input (#4835) 2021-03-19 14:19:32 +01:00
Bartosz Sledz 5ce9e2ae2a
Add support for dynamic input in ONNX Instance Normalization operator (#4805) 2021-03-19 13:57:43 +01:00
Mateusz Bencer 7f3a8ce885
Move ONNX Editor to separate target (#4293)
* Subgraph extraction in ONNX models

* Windows compilation error fix + docs update

* Code cleanup after the first round of reviews

* CI compilation error fix

* Even more CI compilation error fixes

* Proper usage of ADL in generic code

* ONNX shape inference related code cleanup

* Disable the onnx test utils when pb-lite is used

* PB dependency removal from UT, strong types for input and output edges, more verbose errors in tests

* Fix for the protobuf descriptor database corruption

* testing visibility changes

* Revert the changes that didn't work

* Make tests green again?

* Make the current tests pass

* Remove the ONNX header from editor's tests

* Switch from stable_partition to remove_if because of compiler bugs

* Obsolete test removal and cmakelists cleanup in tests

* Macos failed, reverting some changes

* Handle the multiple output consumers UC

* Keep the tensor name when replacing an initializer

* Cutting a graph with multiple consumers of inputs and initializers

* Subgraph extraction with multiple initializer consumers

* Add CMakeList, move files

* fixed headers locations

* fixed calling tests

* apply code styles

* change namespace to onnx_editor

* removd not needed dependencies to onnx_editor

* Remove linking libonnx from unit-test

* Consider all flavors of protobuf libraries

* disable lto for onnx and onnx_proto

* revert set ONNX_INCLUDE_DIR change

* create new target onnx_common

* cmake dependencies clean-up

* Added onnx_common visibility

* pass ModelProto as reference + remove redundant dependencies

* expose onnx symbols by onnx_common

* remove -fno-lto, move InputEdge and OutputEdge to other header

* configure onnx linking for MSVC and APPLE

* remove unused comment

* fixed linking configuration for msvc and apple/clang

* clean cmakes

* added dependencies

* remove onnx and onnx_proto from NGRAPH_EXPORT_TARGETS_ENABLE

* make onnx and protobuf dynamic

* cmake dependency clean-up

* added onnx patch - make only onnx_proto shared

* use protobuf should depend on BUILD_SHARED_LIBS

* fixed patch

* fixed install targets

* add if to install protobuf

* fixed onnx flag

* move protobuf install

* cmake dependencies

* added protobuf header include dir + clean-up ng cmake

* added protobuf to ng targets

* fix apple build

* dependencies clean-up

* make onnx_editor static lib

* remove onnx editor visibility

* fixed protobuf export enablement

* fix apple build

* test build protobuf by fetch content

* fix apple protobuf build

* Docs update

* Producer name update in test onnx models

* More comments in the subgraph extraction code

* Merge remote-tracking branch 'upstream/master' into mbencer/MoveEditorToSeparateLib

* add dependency to onnx_proto for test utils

* dependency to ngraph_test_util

* ngraph_test_util dependency. part. 2

* conditional dependency of onnx to ngraph_test_util

* Apply suggestions from code review

Co-authored-by: Tomasz Socha <tomasz.socha@intel.com>

* review remarks

* remove onnx_proto default visibility

* camke remarks, not exporting onnx_common and onnx_editor to target set

* [test fix macos CI] Revert "remove onnx_proto default visibility"

* set onnx_proto visibility only for macos

* [PyPI] remove rpath for dylib

* Corrected cmd

* fix protobuf-lite linking

Co-authored-by: tomdol <tomasz.dolbniak@intel.com>
Co-authored-by: Tomasz Jankowski <tomasz1.jankowski@intel.com>
Co-authored-by: Mateusz Tabaka <mateusz.tabaka@intel.com>
Co-authored-by: Tomasz Socha <tomasz.socha@intel.com>
Co-authored-by: mryzhov <mikhail.ryzhov@intel.com>
2021-03-17 07:07:53 +03:00
Tomasz Jankowski 063a296842
[ONNX] Add support for Upsample opset1 (#4678) 2021-03-16 10:30:41 +01:00
Alexey Varyzgin 8cfb63cb56
[CPU] proposal_onnx tensor check was added and test was disabled (#4603) 2021-03-16 11:21:57 +03:00
Mateusz Tabaka 13066a0bc6
Propagate input rank for dynamic output shape in BatchToSpace,DepthTo… (#4588)
* Propagate input rank for dynamic output shape in BatchToSpace,DepthToSpace,Pad,SpaceToBatch,SpaceToDepth

* Add tests for SpaceToDepth, DepthToSpace dynamic input
2021-03-12 07:17:24 +03:00
Bartosz Sledz c99b6feea2
Add support for input dynamic shape in ONNX LpNorm operator (#4613) 2021-03-09 13:41:55 +01:00
Bartosz Sledz fc589572a1
Fix ONNX GroupNorm and ExperimentalDetectronGroupNorm (#4579) 2021-03-09 13:40:45 +01:00
Tomasz Dołbniak a085b68d37
Subgraph extraction in ONNX model editor (#4107) 2021-03-04 10:12:37 +01:00
Tomasz Jankowski be614035d1
[ONNX] Fix unidirectional broadcast Add, Div, Mul, Sub v6 (#4462)
* Fix uni broadcast add_v6 op

* Add exception for dynamic rank

* Correct axes mapping in Div, Sub, Mul

It commonizes the approach from Add operator.

* Add unit tests to Div, Sub, Mul v6 ops

* Apply constness

* Improve add_v6 unit test readability

* Correct style

* Remove useless or redundant includes
2021-03-01 19:40:44 +03:00
Bartosz Sledz b0043bb599
ONNX Slice operator support types dynamically (#4507) 2021-02-26 16:28:51 +01:00
Katarzyna Mitrus 76fed9315e
Legacy LSTM/RNN/GRU tests clean up (#4265) 2021-02-26 14:27:18 +01:00
Bartosz Sledz 6c98171e23
Fix output names for group of Strided Slices to Variadic Split optimization (#4378)
* Add tests which reproduces issue

* Fix Output Names in StridedSlice optimization

* Fix codestyle

* Add check if at least one StridedSlice has Result operation as a consumer
2021-02-25 12:53:46 +03:00
Tomasz Jankowski 5e17926604
[ONNX Editor] Add initializers for model inputs (#4025) 2021-02-16 14:12:27 +01:00
Tomasz Socha 185aaacc07
[ONNX] Add support for Softmax v11 in opset 7 models (#4340) 2021-02-16 11:58:39 +01:00
Mateusz Bencer d821ec47ce
[ONNX] Handle optional outputs for Dropout and MaxPool (#4143) 2021-02-15 11:14:20 +01:00
Bartosz Sledz 01aaaf1138
Add support for custom ONNX operator PriorBoxClustered (#4202) 2021-02-15 10:31:30 +01:00
Katarzyna Mitrus ead6427097
ONNX RNN/GRU enable dynamic input shape (#4241) 2021-02-15 10:30:46 +01:00
Michał Karzyński 751ac1aef4
Add support for custom ExperimentalDetectron ops to ONNX Importer (#4030) 2021-02-09 10:31:32 +01:00
Tomasz Socha 4563101c5a
[ONNX Importer] Onnx importer support for brocker names (#4087)
* ONNX Importer changes

* Unit tests

* Style

* Don't set tensor names for result output tensor

* Review fx I

* fix

* Style

Co-authored-by: Vladislav Volkov <vladislav.volkov@intel.com>
2021-02-04 06:53:27 +03:00
Ewa Tusień 80781ff359
Add MVN-6 op to ONNX importer (#4012) 2021-02-01 12:39:39 +01:00
Bartosz Sledz 247606baf6
Add support for nonconstant optional NMS-5 inputs (#3640) 2021-01-26 14:41:28 +01:00
Mateusz Bencer 4e33dac7fa
Set output blobs precision for IE tests (#3905)
* Calling SetPrecission on cnn network outputs

* added tests

* get_output_name refactor

* add missing test file

* run tests on all backends (or disable if backend is not available)

* fixed tests

* fixed TestEngine
2021-01-20 10:26:41 +03:00
Tomasz Jankowski 35384a63a6
[nGraph][ONNX] Extend ONNX Importer for operation "GatherElements-6" (#3822) 2021-01-19 14:42:24 +01:00
Mateusz Bencer 4d98d7ceee
ONNX tests mismatch error (#3836) 2021-01-18 12:39:32 +01:00
Bartosz Sledz 928201bee4
Add support for ONNX Operator ReduceSum v13 and revise other Reduce operators (#3605) 2021-01-18 10:47:40 +01:00
Tomasz Dołbniak 52034a0613
ONNX model editor - replacing input shapes (#3844) 2021-01-15 17:26:44 +01:00
Mateusz Tabaka 20b9d390e1
[ONNX] don't hardcode broadcast shape in Mul, Add, Sub, Div (opset <= 6) operators (#3679) 2021-01-14 11:24:37 +01:00
Tomasz Dołbniak bd299d3daf
ONNX model editor class with input types modification support (#3714) 2021-01-13 15:02:01 +01:00