platform/upstream/dldt.git
3 years ago[IE CLDNN] Loosen restrictions on channels in 3d fsv16 convolution kernel (#1744)
Lukasz Debski [Tue, 1 Sep 2020 12:17:03 +0000 (14:17 +0200)]
[IE CLDNN] Loosen restrictions on channels in 3d fsv16 convolution kernel (#1744)

3 years ago[IE CLDNN] Disabled choose_impl call for generic nodes in some cases to avoid double...
Vladimir Paramuzov [Tue, 1 Sep 2020 12:00:52 +0000 (15:00 +0300)]
[IE CLDNN] Disabled choose_impl call for generic nodes in some cases to avoid double compilation (#1989)

3 years agoDo not remove convert after the topK (#1950)
Ilya Churaev [Tue, 1 Sep 2020 11:36:30 +0000 (14:36 +0300)]
Do not remove convert after the topK (#1950)

* Do not remove convert after the topK

* Added debug message

* Removed xFail

* Revert "Added debug message"

This reverts commit a01ace4ade88d73e2797b47c58db33943b0f508d.

* Added test

3 years ago[IE][VPU]: Customer model compilation error - fix(#1901)
Roman Vyunov (Intel) [Tue, 1 Sep 2020 10:03:31 +0000 (13:03 +0300)]
[IE][VPU]: Customer model compilation error - fix(#1901)

Fix of 36693 issue.

* Problem: One of the concat inputs is a constant. Adjust_data_layout pass tries to duplicate all inputs that do not meet the strides requirements, and then copy from the original input to the duplicate with strides. But duplicateData with an argument in the form of a constant also creates a constant, and then, when Copy, an error appears, the presence of a constant output, which cannot be.
* Solution: In addConvertedData create an intermediate date with the same description as the constant, and then copy the constant data into it with the required strides.

Co-authored-by: DariaMityagina <daria.mityagina@intel.com>
3 years ago[CPU] CTCLoss operation implementation. (#1482)
Nikolay Shchegolev [Tue, 1 Sep 2020 09:52:31 +0000 (12:52 +0300)]
[CPU] CTCLoss operation implementation. (#1482)

3 years ago[ONNX] Enable ScatterND operator in ONNX importer (#1987)
Katarzyna Mitrus [Tue, 1 Sep 2020 09:10:03 +0000 (11:10 +0200)]
[ONNX] Enable ScatterND operator in ONNX importer (#1987)

* Enable ONNX scatter_nd

* Tests update

3 years agoDynamic shape cell (#1683)
Szymon Durawa [Tue, 1 Sep 2020 09:01:20 +0000 (11:01 +0200)]
Dynamic shape cell (#1683)

* validate_and_infer_types() implementation

* input parameter validation for LSTM, GRU and RNN

* style-check applied

* Add LSTMSequence dynamic shape validation and test props for RNNCell, GRUCell, LSTMCell and LSTMSequence.

* recurrent_sequence.hpp moved to ngraph/core/include/ngraph/op/util/

* style check applied

* removed unused variable from LSTMSequence::validate_and_infer_types

* Add missing newline mark at the end of file.

* Add supression macro for FusedOp deprecation.

* Add element type initialization

* Apply PR review remarks

* Rewrite tests to use gtest exception assertions.

3 years agoIntepolate-4 nGraph operation (#1412)
Vladimir Gavrilov [Tue, 1 Sep 2020 03:57:34 +0000 (06:57 +0300)]
Intepolate-4 nGraph operation (#1412)

* Fixed prototype of evaluate method.

* Rewritten Interpolate-4 ctors (added argument output_shape). Corrected tests.

* Fixed typo.

* Fixed number of args of make_shared in op::v4::Interpolate::clone_with_new_inputs.

* Fixes in Interpolate-4 tests.

* Now ONNX Upsample-1 is readed as Interpolate-4 with 4 inputs.

* Code style fixes.

* Some fixes in Interpolate-4 layer test.

* Now ONNX Upsample-9 is readed as Interpolate-4 with 4 inputs.

* Small fixes.

* Some changes.

* Fixed processing of 'scales' input in evaluation of Interpolate-4: now 'scales' contains scales only from 'axes'.

* Fixes in documentation.

* Now reference implementation of Interpolate-4 is rewritten for using 3 required inputs.

* Some code style fixes.

* Small fix.

* Started to write tests for method evaluate() of Interpolate-4.

* Continued to write tests for evaluate() of Interpolate-4.

* Some fixes.

* Some additions.

* Written draft of tests for 'cubic' mode with using scales.

* Some changes.

* Started to write tests for 'nearest' mode.

* Started to write tests for 'linear_onnx' mode.

* Some changes.

* Small fixes.

* Added setup of output type.

* Small addition.

* Added debug print into Interpolate-4 evaluate.

* Now in Interpolate-4 evaluation tests 3 inputs of Interpolate are Constants.

* Small changes.

* Added some debug print.

* Added more debug print.

* Some fixes.

* Now 4th argument of runtime::interpolate has type std::vector<int64_t>.

* Added checks for result of calculations.

* Added another expected values vector for the mode 'cubic'.

* Temporarily commented result value checks  for the cubic mode.

* Written tests for 'nearest' mode.

* Some reorganization.

* Written tests for 'linear_onnx' mode.

* Fixed padding loop.

* Fixed docs for 'linear_onnx' mode.

* Written tests for 'cubic' mode.

* Deleted debug print.

* Fixed code style.

* Enabled CPU layer tests for Interpolate-4.

* Reverted changes of this file.

* Now ONNX importer reads Resize-1 as Interpolate-4 with 4 inputs.

* Now ONNX importer reads Resize-11 as Interpolate-4 with 4 inputs.

* Small fixes.

* Fixed docs.

* Added small epsilon in the shape calculation in the function op::v4::Interpolate::infer_using_scales.

* Small fix.

* Reduced size of the template function eval().

* Now the 'nearest' mode is rewritten to CoordinateTransfom instead of NDim* classes

* Now the 'cubic' mode is rewritten to use CoordinateTransformation instead of NDim classes.

* Started to rewrite 'linear' mode using CoordinateTransform.

* Started to write helper class.

* Small fix.

* Small changes.

* Some fix.

* Fixed typo.

* Now the preamble of 'linear_onnx' mode implementation is a method of helper class.

* Now the function clip_coord is the method of helper class, and the mode 'linear' uses CoordinateTransform instead of NDim classes.

* Deleted NDim classes.

* Some fixes.

* Some refactoring.

* Some refactoring: now inner calculation of 'linear' is in helper class.

* Moved reference implementation of Interpolate-4 into library with reference implementations.

* Small fix.

* Deleted commented tests.

* Code stile fixes.

* Deleted redundant type prop tests for Interpolate-4.

* Documentation fixes.

* Disabled IE_CPU tests for ONNX Resize: Interpolate-4 is not implemented in plugins.

* Temporarily disabled some ONNX tests.

* Some refactoring: deleted redundant attributes of InterpolateEval class.

* Small fix.

* Added NGRAPH_RTTI_DECLARATION and NGRAPH_RTTI_DEFINITION.

* Added debug print to 'cubic' mode calculation.

* Some deletions.

* Small fix.

* Fixed typos.

* Added another debug print.

* Now indices_shape is constructed from std::vector<std::size_t>(num_of_axes, 4) again.

* Fixed CMakeLists.txt.

* Small fix.

* Added more debug print.

* Fixed typo.

* Fixed calculation of the first argument of helper.clip_coord in the inner loop of cubic_func.

* Some code style fixes.

* Alphabetically sorted operations of opset4.

* Deleted constant cannot_define_axes.

* Used std::iota instead of loop.

* Renamed structure InfoToCallReference.

* Now void op::v4::Interpolate::validate_and_infer_types() checks an element type of an input tensor.

* Code style fix.

* Changes in reading of ONNX Resize and Upsample: we generate Interpolate-4 without 'axes' input.

* Now bodies of functions evaluate_interpolate_v4() and inline bool evaluate() are moved in the method bool op::v4::Interpolate::evaluate.

* Fixes in example of the documentation of Interpolate-4.

* Fixed typos.

* Small fix.

* Some fixes.

* Deleted some type aliases.

* Uncommented a test for 'cubic' mode.

* Small code style fix in bool op::v4::Interpolate::evaluate.

* Uncommented more test for Interpolate-4 reference implementation.

* Added more debug print.

* Some changes.

* Uncommented all tests for Interpolate-4 evaluate().

* Deleted debug print.

* Deleted 'simple' mode from the map nearest_mode_map.

* Code style fixes.

* Disabled some CPU tests.

* Some fixes.

* Small fixes.

* Some fixes.

* Fixed typo.

* Fixed typo.

* Small change.

* Fixed some typos.

* Fixed some typos.

* Fix in operator() of the class GetOriginalCoordinate.

* Disabled some CPU tests.

* Small changes.

* Deleted template function from resize.cpp.

* Code style fixes.

* Refactored the method op::v4::Interpolate::evaluate.

* Added documentation for infer_using_scales() and infer_using_shapes().

* Added documentation for classes GetNearestPixel and GetOriginalCoordinate.

* Small fixes.

* Some code style fixes.

* Small fixes.

* Some changes.

* Added NGRAPH_SUPPRESS_DEPRECATED_START and NGRAPH_SUPPRESS_DEPRECATED_END for using v0::InterpolateAttrs; and using v0::Interpolate;

* Code style fix.

* Enabled ONNX import only tests for Resize-10, Upsample-8, Upsample-9.

* Fixed element type for scales_const. Fixed targetShapes and pads in single layer tests.

* Small changes.

* Added conversion from NGRAPH to CNNLayer.

* Added documentation for the class InterpolateEval.

* Now 'nearest' mode has special tests in cpu single layer tests.

* Small changes.

* Fixes in cpu single layer tests.

* Temporarily commented Interpolate-4 in ConvertFunctionToCNNNetworkTests.

* Added some docs.

* Enabled some tests for Resize-11.

* Added test.

* Corrected expected values in the resize_downsample_scales_align_corners case.

* Added more test for the 'cubic' mode.

* Added more test for linear_onnx mode.

* Deleted debug print for linear_onnx mode.

* Deleted debug prints. Added yet another test for 'nearest' mode.

* Fixes for import of ONNX Resize-10 and Upsamples.

* Applyed Evgeny Lazarev fix for Interpolate-4 infer function.

* Code style fixes.

* Some tests were deleted from unit_test.manifest file for INTERPRETER.

* Deleted test for downscale Resize-10: results of infer are correct, but comparison is not.

* Enabled test INTERPRETER.onnx_empty_initializers_handling.

* Some fixes.

* Added the method run_with_tolerance_as_fp() to the class TestCase and the method compare_results_with_tolerance_as_fp() to the class TestCaseEngine.

* Small fix.

* Small fix.

* Added yet another type_prop test.

* Disabled CPU test IE_CPU.onnx_empty_initializers_handling.

* Code style fixes.

* Enabled some ONNX tests.

* Some changes.

* Code style fix.

* Enabled test INTERPRETER.onnx_model_round.

* Disabled tests with behavior as behavior of INTERPRETER.onnx_resize11_scales_down_linear.

* Changed tolerance for test onnx_empty_initializers_handling.

* Changed tolerance in the test onnx_resize11_sizes_linear (otherwise this test is failed in MacOS). Disabled test INTERPRETER.onnx_resize11_sizes_nearest_asymmetric_floor, because this test failed in MacOS only.

3 years agoPrint friendly name instead of name to stream (#1985)
Ilya Churaev [Tue, 1 Sep 2020 03:04:54 +0000 (06:04 +0300)]
Print friendly name instead of name to stream (#1985)

3 years agoAdded documentation about supported ONNX operations (#2004)
Ilya Churaev [Tue, 1 Sep 2020 03:04:37 +0000 (06:04 +0300)]
Added documentation about supported ONNX operations (#2004)

* Added documentation about supported ONNX operations

* Fixed comments

3 years agoRemoved NGRAPH_IE_ENABLE flag because it is always ON if unit tests are enabled ...
Ilya Churaev [Tue, 1 Sep 2020 03:03:59 +0000 (06:03 +0300)]
Removed NGRAPH_IE_ENABLE flag because it is always ON if unit tests are enabled (#2003)

3 years agoDynamic shape support for builder collapse (#1971)
Gabriele Galiero Casay [Mon, 31 Aug 2020 16:26:36 +0000 (18:26 +0200)]
Dynamic shape support for builder collapse (#1971)

3 years agoRemoved useless export from readers (#2009)
Ilya Lavrenov [Mon, 31 Aug 2020 16:00:58 +0000 (19:00 +0300)]
Removed useless export from readers (#2009)

3 years ago[MO] Fix when Crop asks for MXNet specific cmdarg `enable_ssd_gluoncv` (#1978)
Pavel Esir [Mon, 31 Aug 2020 15:06:13 +0000 (18:06 +0300)]
[MO] Fix when Crop asks for MXNet specific cmdarg `enable_ssd_gluoncv` (#1978)

3 years ago[ MO ] DepthToSpace & ShuffleChannels fusion (#2001)
Evgenya Stepyreva [Mon, 31 Aug 2020 13:20:19 +0000 (16:20 +0300)]
[ MO ] DepthToSpace & ShuffleChannels fusion (#2001)

* [ MO ] ShuffleChannel fusion

* DepthToSpace fusion

* test

* comment

3 years ago[IE CLDNN] Restrict bsv16_fsv16 convolution to int8 for batch 16 (#1954)
Alexander Chaiko [Mon, 31 Aug 2020 07:58:32 +0000 (09:58 +0200)]
[IE CLDNN] Restrict bsv16_fsv16 convolution to int8 for batch 16 (#1954)

3 years agoFixed compilation with private plugins (#1997)
Ilya Lavrenov [Mon, 31 Aug 2020 07:58:07 +0000 (10:58 +0300)]
Fixed compilation with private plugins (#1997)

3 years agoMultiple fixes related to tickets 37105 and 37615 (#1979)
Evgeny Lazarev [Mon, 31 Aug 2020 07:41:18 +0000 (10:41 +0300)]
Multiple fixes related to tickets 37105 and 37615 (#1979)

* Multiple fixes.

1. Fixes SpaceToBatch transformation to not crash if inputs are not Constant
2. Fixed eliminate_squueze, eliminate_unsqueeze to not crash when input has dynamic rank
3. Added reference implementation for the FloorMod operation

* Code style fixes

* Update floor_mod.hpp

Removed unnecessary function

3 years ago[IE CLDNN] Improve performance of fc block fp16 implementation (#1993)
Konrad Dobros [Mon, 31 Aug 2020 06:52:47 +0000 (08:52 +0200)]
[IE CLDNN] Improve performance of fc block fp16 implementation (#1993)

Main purpose of this change is to fix weird behaviour of
fully_connected_gpu_fb_io_block_fp16 implementation where it shows
severe performance drop without bias.
Additionally assembly for case with bias is improved.

3 years agoFixed compilation when NGRAPH_INTERP is not enabled (#1998)
Ilya Lavrenov [Mon, 31 Aug 2020 03:47:42 +0000 (06:47 +0300)]
Fixed compilation when NGRAPH_INTERP is not enabled (#1998)

3 years agoAdded note about QueryNetwork (#1965)
Ilya Lavrenov [Sun, 30 Aug 2020 14:07:08 +0000 (17:07 +0300)]
Added note about QueryNetwork (#1965)

3 years ago[ MO ][ RESHAPE ] Fixes in RNNSequenceNormalize transformation (#1909)
Yegor Kruglov [Sat, 29 Aug 2020 22:19:22 +0000 (01:19 +0300)]
[ MO ][ RESHAPE ] Fixes in RNNSequenceNormalize transformation  (#1909)

* initial commit

* first reshap-able variant

* right version for reshape

* comment update

* fixes for failed e2e

* set data type to ngraph TensorIterator

* Fix dynamic shapes for cells ops

* clean up

Co-authored-by: yegor.kruglov <ykruglov@nnlvdp-mkaglins.inn.intel.com>
3 years ago[IE][VPU]: firmware update (#1973)
Maksim Doronin [Sat, 29 Aug 2020 10:23:53 +0000 (13:23 +0300)]
[IE][VPU]: firmware update (#1973)

3 years ago[GNA] Added fix for eltwise layer with more 65k elements (#1943)
Andrey Dmitriev [Fri, 28 Aug 2020 15:26:31 +0000 (18:26 +0300)]
[GNA] Added fix for eltwise layer with more 65k elements (#1943)

3 years ago[IE][VPU]: ignore 2nd output of Proposal layer (#1975)
Evgeny Latkin [Fri, 28 Aug 2020 14:18:42 +0000 (17:18 +0300)]
[IE][VPU]: ignore 2nd output of Proposal layer (#1975)

* TODO: implement 2nd output, see: #-37327: Several models Failed to compile layer "proposals"

3 years agoImplement reshapeable CTCGreedyDecoderPlusSparseToDense transformation and test ...
Roman Kazantsev [Fri, 28 Aug 2020 11:28:32 +0000 (14:28 +0300)]
Implement reshapeable CTCGreedyDecoderPlusSparseToDense transformation and test (#1906)

* Implement reshapeable CTCGreedyDecoderPlusSparseToDense transformation and test

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Fix consts (after code-review #1)

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Add CTCGreedyDecoderTransformation with more generic pattern

Also it adds new middle-replacer for transforming sequence length to a mask
along with tests.

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Do fixes after review #2

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Fix after review #3

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Fix after review #4

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
3 years agorefactor import network (#1871)
Anna Alberska [Fri, 28 Aug 2020 10:48:48 +0000 (12:48 +0200)]
refactor import network (#1871)

3 years agoNormalizeL2 transformation (#1892)
Mateusz Bencer [Fri, 28 Aug 2020 03:51:46 +0000 (05:51 +0200)]
NormalizeL2 transformation (#1892)

* first version of implementation

* added unit tests

* changed multiply to pow

* doc + unit tests

* more unit tests

* code review remarks

* missing new line

* remarks

* review remarks

* Build fix - update constant check function in HSwishFusionWithClamp

Co-authored-by: mitruska <katarzyna.mitrus@intel.com>
3 years agoFix the case when run stage of verification scripts failed on Windows 10 due to diffe...
Tai-Min [Thu, 27 Aug 2020 20:58:16 +0000 (22:58 +0200)]
Fix the case when run stage of verification scripts failed on Windows 10 due to different build directory for executables (#1478)

3 years ago[IE CLDNN] Cleanup part 2 (#1865)
Vladimir Paramuzov [Thu, 27 Aug 2020 20:06:20 +0000 (23:06 +0300)]
[IE CLDNN] Cleanup part 2 (#1865)

* [IE CLDNN] Removed some unused kernels and layouts

* [IE CLDNN] Removed bsv4_fsv32 layout

* [IE CLDNN] Removed remaining BF8_XY16 usages. Removed definitions.cl

3 years agoUpdate install_openvino_dependencies.sh for Ubuntu 20.04 (#1974)
Alexey Suhov [Thu, 27 Aug 2020 19:30:30 +0000 (22:30 +0300)]
Update install_openvino_dependencies.sh for Ubuntu 20.04 (#1974)

3 years ago[IE CLDNN] Move iGPU to first position in GPU device map (#1828)
Mikhail Letavin [Thu, 27 Aug 2020 18:25:24 +0000 (21:25 +0300)]
[IE CLDNN] Move iGPU to first position in GPU device map (#1828)

3 years ago[IE][VPU]: ROIAlign : use CHWc version (#1937)
Andrey Sokolov [Thu, 27 Aug 2020 17:45:33 +0000 (20:45 +0300)]
[IE][VPU]: ROIAlign : use CHWc version (#1937)

* ROIAlign optimization (step 3): use CHWc version

3 years agoGVA/Update install_openvino_dependencies.sh for GVA on Ubuntu20.04 (#1895)
Yuliya Maskina [Thu, 27 Aug 2020 16:41:52 +0000 (19:41 +0300)]
GVA/Update install_openvino_dependencies.sh for GVA on Ubuntu20.04 (#1895)

* Update install_openvino_dependencies.sh for GVA on Ubuntu20.04

* Add dependencies

3 years agomade compatible with ubuntu 20.04 (#1798)
Nazmul Alom [Thu, 27 Aug 2020 16:40:44 +0000 (22:40 +0600)]
made compatible with ubuntu 20.04 (#1798)

* made compatible with ubuntu 20.04

* demo_benchmark_app updated

* demo_security_barrier_camera updated

* demo_squeezenet_download_convert_run updated

* Update demo_benchmark_app.sh

* Update demo_security_barrier_camera.sh

* Update demo_squeezenet_download_convert_run.sh

* unwanted file removed

3 years ago[IE CLDNN] Improved GWS for 3d fsv16 eltwise (#1957)
Vladimir Paramuzov [Thu, 27 Aug 2020 14:57:30 +0000 (17:57 +0300)]
[IE CLDNN] Improved GWS for 3d fsv16 eltwise (#1957)

3 years ago[JAVA] Use JUnit 4 features to log Java tests log (#1953)
Dmitry Kurtaev [Thu, 27 Aug 2020 14:10:24 +0000 (17:10 +0300)]
[JAVA] Use JUnit 4 features to log Java tests log (#1953)

* [JAVA] Use JUnit 4 features to log Java tests log

* [JAVA] Add device parameter for Java tests

3 years agoAdd support for ReduceL1 and ReduceL2 to ONNX Importer (#1838)
Bartosz Lesniewski [Thu, 27 Aug 2020 13:47:04 +0000 (15:47 +0200)]
Add support for ReduceL1 and ReduceL2 to ONNX Importer (#1838)

3 years ago[nGraph] RTInfo refactor (#1806)
Jan Iwaszkiewicz [Thu, 27 Aug 2020 13:41:21 +0000 (15:41 +0200)]
[nGraph] RTInfo refactor (#1806)

3 years agoUpdate README.md (#1969)
Alexey Suhov [Thu, 27 Aug 2020 13:29:11 +0000 (16:29 +0300)]
Update README.md (#1969)

3 years ago[ONNX] Update ONNX importer to use SotfPlus-4 (#1959)
Katarzyna Mitrus [Thu, 27 Aug 2020 12:55:04 +0000 (14:55 +0200)]
[ONNX] Update ONNX importer to use SotfPlus-4 (#1959)

* Use SoftPlus-4 in ONNX importer

* Tests update

3 years ago[Samples] Use get_ops for iteration over graph in samples (#1961)
Anastasia Kuporosova [Thu, 27 Aug 2020 11:14:59 +0000 (14:14 +0300)]
[Samples] Use get_ops for iteration over graph in samples (#1961)

3 years agoUse relative paths in error messages (#1963)
Ilya Churaev [Thu, 27 Aug 2020 11:04:02 +0000 (14:04 +0300)]
Use relative paths in error messages (#1963)

3 years ago[CPU] [MSVC] memcpy and memcpy_s were unified for all layers and SIMD functionality...
Alexey Varyzgin [Thu, 27 Aug 2020 10:49:31 +0000 (13:49 +0300)]
[CPU] [MSVC] memcpy and memcpy_s were unified for all layers and SIMD functionality was enabled on VS2019 (#1374)

3 years agoUpdate priorbox extenders (#1925)
Anton Chetverikov [Thu, 27 Aug 2020 10:10:44 +0000 (13:10 +0300)]
Update priorbox extenders (#1925)

3 years agoMO fusing activations (#1942)
Evgeny Lazarev [Thu, 27 Aug 2020 08:56:52 +0000 (11:56 +0300)]
MO fusing activations (#1942)

* Added HSwish operation

* Added HSwish fusing transformation

* Fixed BOM

* Added unit test for HSwish fusing transformation

* Fixed unit tests for transformations using 'build_graph_with_edge_attrs' function to build the graph

* Added fusion transformation for Swish operation

* Added fusing transformation for Softplus operation

* Added fusion transformation for Mish operation

* Added check for the node name in the unit tests

* Fixed Mish fusion pattern

* Updated Mish fusion transformation. Added unit test

* Updated HSwish fusing transformation

* Updated Swish fusion transformation and tests

* Fixed unit tests

3 years agoExperimentalDetectron* shape infer old-style (#1962)
Evgenya Stepyreva [Thu, 27 Aug 2020 08:52:04 +0000 (11:52 +0300)]
ExperimentalDetectron* shape infer old-style (#1962)

* ExperimentalDetectron* shape infer old-style

* nGraph reshape tests

3 years ago[IE CLDNN] Mixed mode support for proposal primitive (#1857)
Vladimir Paramuzov [Thu, 27 Aug 2020 08:43:24 +0000 (11:43 +0300)]
[IE CLDNN] Mixed mode support for proposal primitive (#1857)

3 years ago[MO] Relax Reshape layer hardcode under MatMul (#1921)
Evgenya Stepyreva [Thu, 27 Aug 2020 06:45:08 +0000 (09:45 +0300)]
[MO] Relax Reshape layer hardcode under MatMul (#1921)

* [MO] Relax Reshape layer hardcode under MatMul

* Memory fix

3 years agoAdded workaround for logical elementwise operations to change the constant input...
Evgeny Lazarev [Thu, 27 Aug 2020 05:46:34 +0000 (08:46 +0300)]
Added workaround for logical elementwise operations to change the constant input data type if it does not match the other one (#1955)

3 years ago[IE CLDNN] Fix fused ops in 1x1 conv fsv16 kernel (#1948)
Jedrzej Hajduczenia [Wed, 26 Aug 2020 17:58:51 +0000 (19:58 +0200)]
[IE CLDNN] Fix fused ops in 1x1 conv fsv16 kernel (#1948)

3 years ago[IE CLDNN] Fixed memory allocation for case when usm is not available (#1952)
Vladimir Paramuzov [Wed, 26 Aug 2020 17:58:21 +0000 (20:58 +0300)]
[IE CLDNN] Fixed memory allocation for case when usm is not available (#1952)

3 years agoFixed mean image support for IR v10 (#1926)
Ilya Churaev [Wed, 26 Aug 2020 17:45:54 +0000 (20:45 +0300)]
Fixed mean image support for IR v10 (#1926)

* Fixed mean image support for IR v10

* Fixed comments

3 years agoModified samples to support ONNX models (#1680)
Mikhail Ryzhov [Wed, 26 Aug 2020 15:53:24 +0000 (18:53 +0300)]
Modified samples to support ONNX models (#1680)

* Modified samples to support ONNX models

- Removed manual loading of IR .bin files (hello_classification, object_detection_sample_ssd)
- Corrected comments
- Added Note in README files

* Modified samples to support ONNX models

- Removed manual loading of IR .bin files (hello_classification, object_detection_sample_ssd)
- Corrected comments
- Added Note in README files

3 years ago[SCRIPTS] install_NEO_OCL_driver.sh: Added verifying current driver version (#1951)
Artyom Anokhov [Wed, 26 Aug 2020 14:25:33 +0000 (17:25 +0300)]
[SCRIPTS] install_NEO_OCL_driver.sh: Added verifying current driver version (#1951)

* install_NEO_OCL_driver.sh: Added verifying current driver version

* install_NEO_OCL_driver.sh: Updated removal oldest driver on Ubuntu. Updated logic on defining

* install_NEO_OCL_driver.sh: Fixed function name

3 years agoEnable eager execution while TensorFlow 2 model is loaded (#1945)
Roman Kazantsev [Wed, 26 Aug 2020 14:16:58 +0000 (17:16 +0300)]
Enable eager execution while TensorFlow 2 model is loaded (#1945)

3 years agoFixed threading configuration parameters (#1919)
Anton Pankratv [Wed, 26 Aug 2020 13:38:41 +0000 (16:38 +0300)]
Fixed threading configuration parameters (#1919)

* Fixed threading parameters

* Updated documentation

* Fix for macos

* Fix doc

3 years ago[IE CLDNN] int8 output support in fsv16 dw conv (#1886)
Vladimir Paramuzov [Wed, 26 Aug 2020 13:15:02 +0000 (16:15 +0300)]
[IE CLDNN] int8 output support in fsv16 dw conv (#1886)

3 years agoDisable CanConvertFromAOTtoSueModel test (#1949)
Kamil Magierski [Wed, 26 Aug 2020 13:09:19 +0000 (15:09 +0200)]
Disable CanConvertFromAOTtoSueModel test (#1949)

3 years agoQuery network should return all supported layers (#1845)
Anton Pankratv [Wed, 26 Aug 2020 12:33:37 +0000 (15:33 +0300)]
Query network should return all supported layers (#1845)

* Query network should return all supported layers

* Added other plugins. Fixed hetero and multi

* IRv7 test use legacy impl

3 years ago[IE TESTS] Add Reverse Seq single layer test (#1933)
Maksim Kutakov [Wed, 26 Aug 2020 12:13:32 +0000 (15:13 +0300)]
[IE TESTS] Add Reverse Seq single layer test (#1933)

* [IE TESTS] Add Reverse Seq single layer test

* Added spaces for better readability.

3 years ago[IE CLDNN] Weights reorders optimization (#1542)
Sergey Shlyapnikov [Wed, 26 Aug 2020 12:12:55 +0000 (15:12 +0300)]
[IE CLDNN] Weights reorders optimization (#1542)

* [IE CLDNN] Fix incorrect (g)_is_os_(z)yx_isv16_osv16 layout name

* [IE CLDNN] Constant blobs copying improvements

* [IE CLDNN] Weights reorders optimization

3 years ago[GNA] Aligning GNA2 and GNA Unit testing (#952)
Kamil Magierski [Wed, 26 Aug 2020 10:43:43 +0000 (12:43 +0200)]
[GNA] Aligning GNA2 and GNA Unit testing (#952)

Co-authored-by: kmagiers <kmagiers@intel.com>
3 years agoIntroduce Quantize-Dequantize to FakeQuantize transformation (#1849)
Mateusz Tabaka [Wed, 26 Aug 2020 08:51:51 +0000 (10:51 +0200)]
Introduce Quantize-Dequantize to FakeQuantize transformation (#1849)

* Introduce Quantize-Dequantize to FakeQuantize transformation

* Revert changes in DequantizeLinear

* apply code format

* Changes after review:

- description for transformation
- remove NGRAPH_CHECK and move some checks from callback to predicates in pattern
- check if out_low/high are broadcastable for FQ's first input
- fix params to copy_runtime_info

* Add type_matches and type_matches_any predicates

* Use get_single_value

* Changes after review:

- add brief description of transformation
- use get_pattern_value_map instead of get_pattern_map
- change opset1 to opset4
- fix params to copy_runtime_info

* Check result of dynamic_pointer_cast

3 years agoFixed Precision Conversion for GenericIE operation type (#1917)
Gleb Kazantaev [Wed, 26 Aug 2020 08:31:40 +0000 (11:31 +0300)]
Fixed Precision Conversion for GenericIE operation type (#1917)

* Fixed Precision Conversion for GenericIE operation type

* changed logic for setting output type in GenericIE

3 years ago[IE][VPU]: Fix NCE hang due to input dimensions exceed HW limitation (#1636)
Wenzeng Chen [Wed, 26 Aug 2020 07:38:58 +0000 (15:38 +0800)]
[IE][VPU]: Fix NCE hang due to input dimensions exceed HW limitation (#1636)

* Fixed NCE hang due to input dimensions exceed HW limitation
* Added convolution test for big input dimensions

Signed-off-by: wenzengc <wenzeng.chen@intel.com>
3 years agoImprovements for MKLDNN Input node execution (#1913)
Vladislav Volkov [Wed, 26 Aug 2020 05:38:29 +0000 (08:38 +0300)]
Improvements for MKLDNN Input node execution (#1913)

3 years agoRemoved WA with adding fake output for VariadicSplit operation (#1938)
Gleb Kazantaev [Wed, 26 Aug 2020 03:28:03 +0000 (06:28 +0300)]
Removed WA with adding fake output for VariadicSplit operation (#1938)

3 years ago[ MO ] Fix for remove_op_node_with_data_node (#1934)
Evgenya Stepyreva [Tue, 25 Aug 2020 17:30:41 +0000 (20:30 +0300)]
[ MO ] Fix for remove_op_node_with_data_node (#1934)

3 years ago[GNA] Remove profiler (#1931)
Pavel Rodionov [Tue, 25 Aug 2020 16:59:01 +0000 (19:59 +0300)]
[GNA] Remove profiler (#1931)

3 years agoSoftPlus operation implementation (#1365)
Anton Chetverikov [Tue, 25 Aug 2020 16:36:39 +0000 (19:36 +0300)]
SoftPlus operation implementation (#1365)

3 years agoCommon specific creator for ops to be converted to ie ops (#1928)
Ivan Tikhonov [Tue, 25 Aug 2020 16:23:09 +0000 (19:23 +0300)]
Common specific creator for ops to be converted to ie ops (#1928)

* Add common specific creator for ops to be converted to ie ops

* fix cpu func tests

* Add macro for specific creators

3 years agoFix output name collision for incorrect ONNX models (#1930)
Maxim Vafin [Tue, 25 Aug 2020 15:43:43 +0000 (18:43 +0300)]
Fix output name collision for incorrect ONNX models (#1930)

3 years agoAdjust tolerance values in ModelImportRunner to the ones onnxruntime … (#1894)
Mateusz Tabaka [Tue, 25 Aug 2020 14:12:55 +0000 (16:12 +0200)]
Adjust tolerance values in ModelImportRunner to the ones onnxruntime … (#1894)

3 years agoFixed order of transformation to convert the TF OD API SSD models (#1887)
Evgeny Lazarev [Tue, 25 Aug 2020 13:39:34 +0000 (16:39 +0300)]
Fixed order of transformation to convert the TF OD API SSD models (#1887)

* Fixed order of transformation to convert the TF OD API SSD models

* Refactored the sub-graph modification for the TF OD API models related to Squeeze/Reshape after SSD heads

3 years agoAdded BatchNormInference tests (#1927)
Liubov Batanina [Tue, 25 Aug 2020 13:28:05 +0000 (16:28 +0300)]
Added BatchNormInference tests (#1927)

3 years agoDeprecated migration to Core guide (#1910)
Ilya Lavrenov [Tue, 25 Aug 2020 13:00:09 +0000 (16:00 +0300)]
Deprecated migration to Core guide (#1910)

3 years ago[IE][VPU]: Update firmware (#1915)
Maksim Doronin [Tue, 25 Aug 2020 08:43:24 +0000 (11:43 +0300)]
[IE][VPU]: Update firmware (#1915)

* Firmware update contains changed conditions for removing inner dimension with length 1 in Permute kernel

3 years ago[IE][VPU]: Fix KW issues on Linux (#1916)
Maksim Doronin [Tue, 25 Aug 2020 08:42:23 +0000 (11:42 +0300)]
[IE][VPU]: Fix KW issues on Linux (#1916)

* nullptr dereference fixed in Concat and Reduce stages in Graph Transformer
* missed parameters in asserts in StridedSlice stage was added

3 years ago[LPT] BLOCKED layout ignoring (#1873)
Edward Shogulin [Tue, 25 Aug 2020 07:43:33 +0000 (10:43 +0300)]
[LPT] BLOCKED layout ignoring (#1873)

* [LPT] BLOCKED layout ignoring

* [LPT] Permute transformation tests

* [LPT] Tests rework

3 years ago[CPU] Added H-Swish activation (#1445)
Alexandra Sidorova [Tue, 25 Aug 2020 07:19:06 +0000 (10:19 +0300)]
[CPU] Added H-Swish activation (#1445)

3 years agoReference implementation for ScatterUpdate (#1678)
Adam Osewski [Tue, 25 Aug 2020 03:12:39 +0000 (05:12 +0200)]
Reference implementation for ScatterUpdate (#1678)

* Reference implementation for ScatterUpdate and use of it in evaluate.

* Review comments. Clarify comments.

* Update file directory.

* Replace scatter_update reference implementation in ngraph/core/reference/

* Remove template code from ScatterUpdate reference implementation

* Apply review requests

Co-authored-by: mitruska <katarzyna.mitrus@intel.com>
3 years ago[LPT] GemmTransformation usage fix (#1899)
Edward Shogulin [Mon, 24 Aug 2020 17:09:01 +0000 (20:09 +0300)]
[LPT] GemmTransformation usage fix (#1899)

3 years agov0::Pad operator removal (#1912)
Tomasz Dołbniak [Mon, 24 Aug 2020 16:03:01 +0000 (18:03 +0200)]
v0::Pad operator removal (#1912)

* v0::Pad operator removal

* Obsolete test removal

3 years ago[GNA] Change GNA library version to 1047.1 (#1867)
Pavel Rodionov [Mon, 24 Aug 2020 15:20:31 +0000 (18:20 +0300)]
[GNA] Change GNA library version to 1047.1 (#1867)

3 years agoDisable eager execution affecting TensorFlow 1 concept model handling (#1914)
Roman Kazantsev [Mon, 24 Aug 2020 15:01:45 +0000 (18:01 +0300)]
Disable eager execution affecting TensorFlow 1 concept model handling (#1914)

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
3 years ago[IE CLDNN] Fix concat in place with conv interactions (#1714)
Konrad Dobros [Mon, 24 Aug 2020 14:37:46 +0000 (16:37 +0200)]
[IE CLDNN] Fix concat in place with conv interactions (#1714)

This change fixes concatenation in place optimization where it may
interact with convolution that uses physical padding.
One of such cases is where input to optimized concatenation is also
input to convolution, so it should have padding to enable optimized
implementation.
Previously for all concatenation inputs padding was overriden with only
concatenation axis being padded.
This change fixes this issue by propagating padding across inputs and
output.

3 years agoAdd reshaping of dynamic network in python tests (#1850)
Tomasz Socha [Mon, 24 Aug 2020 11:58:38 +0000 (13:58 +0200)]
Add reshaping of dynamic network in python tests (#1850)

3 years ago[Python API] Deprecation of DataPtr.creator_layer and DataPtr.input_to properties...
Anastasia Kuporosova [Mon, 24 Aug 2020 11:57:55 +0000 (14:57 +0300)]
[Python API] Deprecation of DataPtr.creator_layer and DataPtr.input_to properties (#1905)

3 years agoPartial revert a commit with reference implementations of PriorBox(Clustered) ops...
Ivan Tikhonov [Mon, 24 Aug 2020 11:31:32 +0000 (14:31 +0300)]
Partial revert a commit with reference implementations of PriorBox(Clustered) ops (#1878)

* partial revert a commit with reference implementations of PriorBox(Clustered), disable references for this ops

* ngraph codestyle

* disable const folding unit tests for PriorBox(Clustered)

* fix arm build

* fix unit test

* Revert "fix unit test"

This reverts commit 1fe59e55d620c430e81c52edd3fc46f2b5b1a9b9.

3 years ago[ MO ] Reinfer shape sub-graphs once (#1908)
Evgenya Stepyreva [Mon, 24 Aug 2020 11:30:41 +0000 (14:30 +0300)]
[ MO ] Reinfer shape sub-graphs once (#1908)

* [ MO ] Reinfer shape sub-graphs once

* feedback

* feedback

3 years agoAdded test for Pad (#1902)
Liubov Batanina [Mon, 24 Aug 2020 11:16:29 +0000 (14:16 +0300)]
Added test for Pad (#1902)

* Addded test for Pad

* Add builder

3 years agoFix Assign, Matmul ops for correct work with keep_shape_ops feature for Kaldi models...
Ivan Tikhonov [Mon, 24 Aug 2020 11:11:01 +0000 (14:11 +0300)]
Fix Assign, Matmul ops for correct work with keep_shape_ops feature for Kaldi models (#1885)

* fix assign and matmul ops

* fix unit test

* static shapes by default for kaldi

3 years ago[IE CLDNN] Enabled tensor pitches in topk kernel for 0 input (#1898)
Vladimir Paramuzov [Mon, 24 Aug 2020 10:03:39 +0000 (13:03 +0300)]
[IE CLDNN] Enabled tensor pitches in topk kernel for 0 input (#1898)

3 years ago[IE CLDNN] Forcing bfzyx format in MVN layer (#1903)
Lukasz Debski [Mon, 24 Aug 2020 10:03:21 +0000 (12:03 +0200)]
[IE CLDNN] Forcing bfzyx format in MVN layer (#1903)

3 years agoReduce number of ops needed to create InstanceNorm (#1896)
Mateusz Tabaka [Mon, 24 Aug 2020 09:26:15 +0000 (11:26 +0200)]
Reduce number of ops needed to create InstanceNorm (#1896)

* Reduce number of ops needed to create InstanceNorm

InstanceNorm in onnx importer creates the same subgraph for Mean twice - once for Variance and once for actual Mean.
This change makes InstanceNorm to use single Mean which is shared by numerator and Variance.

Also enables IE_CPU.onnx_model_instance_normalization test case

* Revert changes to .gitignore

3 years ago[IE CLDNN] FP16 support for some quantized kernels (#1855)
Vladimir Paramuzov [Mon, 24 Aug 2020 05:47:45 +0000 (08:47 +0300)]
[IE CLDNN] FP16 support for some quantized kernels (#1855)

3 years agoLinux clang compilation error fix (#1891)
Tomasz Dołbniak [Fri, 21 Aug 2020 14:22:54 +0000 (16:22 +0200)]
Linux clang compilation error fix (#1891)

3 years ago[ MO ] Change layout of Shape sub-graphs once (#1875)
Evgenya Stepyreva [Fri, 21 Aug 2020 13:27:02 +0000 (16:27 +0300)]
[ MO ] Change layout of Shape sub-graphs once (#1875)

* [ MO ] Change layout of Shape sub-graphs once

3 years agoAccept empty initializers from ONNX models (#1776)
Tomasz Dołbniak [Fri, 21 Aug 2020 12:15:27 +0000 (14:15 +0200)]
Accept empty initializers from ONNX models (#1776)

* Accept empty initializers from ONNX models

* Create dummy constants instead of null nodes

* Code formatting

* Test the empty initializers handling in onnx importer