platform/upstream/dldt.git
3 years agoUpdate Reduction operations specification (#1446)
Anton Chetverikov [Thu, 23 Jul 2020 12:34:53 +0000 (15:34 +0300)]
Update Reduction operations specification (#1446)

3 years agoRemove deprecated methods from node (#1369)
Ilya Churaev [Thu, 23 Jul 2020 11:44:32 +0000 (14:44 +0300)]
Remove deprecated methods from node (#1369)

3 years agoUpdate ShapeOf extender (#1406)
Anton Chetverikov [Thu, 23 Jul 2020 11:07:34 +0000 (14:07 +0300)]
Update ShapeOf extender (#1406)

3 years agoDocumentation updates (#1433)
Ilya Lavrenov [Thu, 23 Jul 2020 10:15:20 +0000 (13:15 +0300)]
Documentation updates (#1433)

3 years agoCorrected message about deprecated IR version (#1438)
Ilya Lavrenov [Thu, 23 Jul 2020 10:14:09 +0000 (13:14 +0300)]
Corrected message about deprecated IR version (#1438)

3 years agoAdd ONNX SpaceToDepth and DepthToSpace extractor (#1122)
Maxim Vafin [Thu, 23 Jul 2020 10:05:42 +0000 (13:05 +0300)]
Add ONNX SpaceToDepth and DepthToSpace extractor (#1122)

* Add ONNX SpaceToDepth extractor

* Add DepthToSpace ONNX extractor

3 years ago[IE][VPU][GT]: Change begin and end masks serialization for StridedSlice (#1417)
Andrew Bakalin [Thu, 23 Jul 2020 09:20:08 +0000 (12:20 +0300)]
[IE][VPU][GT]: Change begin and end masks serialization for StridedSlice (#1417)

* In order to support cases when begin and end size is less than input rank, the serialization was changed.
* Add tests
* Update firmware

3 years agoUpdated fluid to 4.3.0 (#1422)
Ruslan Garnov [Thu, 23 Jul 2020 09:04:31 +0000 (12:04 +0300)]
Updated fluid to 4.3.0 (#1422)

3 years ago[Stress] Redesign of MemCheckTests (#650)
Vitaliy Urusovskij [Thu, 23 Jul 2020 08:07:13 +0000 (11:07 +0300)]
[Stress] Redesign of MemCheckTests (#650)

* [Stress] Redesigned MemCheckTests: 1. Added MemCheckPipeline to incapsulate measures and logging. 2. Moved references to array

* [Stress] Added tracking of THREADS in MemCheckTests

3 years ago[IE][VPU][NGraph] Move dynamic NMS from common to vpu folder (#1409)
Andrew Bakalin [Thu, 23 Jul 2020 07:23:35 +0000 (10:23 +0300)]
[IE][VPU][NGraph] Move dynamic NMS from common to vpu folder (#1409)

* [IE][VPU]: Moves UpgradeNMS4ToNMSDynamic transformation into myriad plugin

* [IE][VPU]: Moves UpgradeNMS4ToNMSDynamic from common to vpu folder

* [IE][VPU]: Moves Dynamic NMS from common folder to vpu

* [VPU]: Makes NMS conversion unconditional

* [VPU][NGraph]: Changes dynamic NMS base class from v3 to v4

* [VPU]: Moves NMS4toDynamic transformation before common optimization

3 years ago[IE CLDNN] Shrink reshapes (#1362)
Vladimir Paramuzov [Thu, 23 Jul 2020 07:14:52 +0000 (10:14 +0300)]
[IE CLDNN] Shrink reshapes (#1362)

3 years agoFixes LFS issues (#1440)
Andrey Zaytsev [Wed, 22 Jul 2020 19:35:58 +0000 (22:35 +0300)]
Fixes LFS issues (#1440)

* Removed non-LFS images to reload them to LFS

* Reload images to LFS

* Reload images to LFS

3 years ago[IE Myriad/35383] vpu_profile is removed (#1393)
Nikita Kudriavtsev [Wed, 22 Jul 2020 14:34:16 +0000 (17:34 +0300)]
[IE Myriad/35383] vpu_profile is removed (#1393)

3 years agoTest fix import of ONNX model in serialized Protobuf binary format. (#1355)
Adam Osewski [Wed, 22 Jul 2020 10:52:53 +0000 (12:52 +0200)]
Test fix import of ONNX model in serialized Protobuf binary format. (#1355)

* Try fix parsing error.

* Small exception refinements during importing model.

* More exception refinements.

* Skip segfaulting tests.

* More clear error types and messages. Func rename.

* Fix typo.

* Check on CI whether test_onnx will work.

* Add only those file which pass tests or have failing ones skipped.

3 years agoExtend nGraph for operation CTCLoss (#1236)
Roman Kazantsev [Wed, 22 Jul 2020 10:45:42 +0000 (13:45 +0300)]
Extend nGraph for operation CTCLoss (#1236)

* Extend nGraph for operation CTCLoss

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Fixes as per comments

Co-authored-by: Nikolay Shchegolev <nikolay.shchegolev@intel.com>
3 years agoReplaced copy_with_new_args() to clone_with_new_inputs() (#1395)
Ilya Churaev [Wed, 22 Jul 2020 10:44:22 +0000 (13:44 +0300)]
Replaced copy_with_new_args() to clone_with_new_inputs() (#1395)

3 years agoExpose OpSets as part of nGraph PythonAPI (#1261)
Ewa Tusień [Wed, 22 Jul 2020 10:02:54 +0000 (12:02 +0200)]
Expose OpSets as part of nGraph PythonAPI (#1261)

3 years agoUpdate list of tests passing/failing using TestCase on IE backend (#1161)
Ewa Tusień [Wed, 22 Jul 2020 09:42:19 +0000 (11:42 +0200)]
Update list of tests passing/failing using TestCase on IE backend (#1161)

3 years ago[Stress] Small improvement of scripts (#1302)
Vitaliy Urusovskij [Wed, 22 Jul 2020 07:24:15 +0000 (10:24 +0300)]
[Stress] Small improvement of scripts (#1302)

* [Stress] Extend CLI with arguments checks for compare_memcheck_2_runs.py

* [Stress] Extend requirements.txt with PyYAML

3 years agoRemoved ie_deprecated.cpp from legacy library (#1419)
Ilya Lavrenov [Wed, 22 Jul 2020 03:44:53 +0000 (06:44 +0300)]
Removed ie_deprecated.cpp from legacy library (#1419)

3 years agoRemoved CNNLayer entries from ie_common.h (#1420)
Ilya Lavrenov [Wed, 22 Jul 2020 03:44:05 +0000 (06:44 +0300)]
Removed CNNLayer entries from ie_common.h (#1420)

* Removed useless header include

* Removed CNNLayer entries from ie_common.h

3 years agoRemoved is_parameter, is_constant and is_output (#1408)
Ilya Churaev [Wed, 22 Jul 2020 03:11:28 +0000 (06:11 +0300)]
Removed is_parameter, is_constant and is_output (#1408)

3 years agoRemoved redundant licenses (#1392)
Ilya Churaev [Tue, 21 Jul 2020 16:48:25 +0000 (19:48 +0300)]
Removed redundant licenses (#1392)

* Removed redundant licenses

* Added protobuf license

3 years agoCMAKE: added ie_libraries target (#1411)
Ilya Lavrenov [Tue, 21 Jul 2020 15:58:20 +0000 (18:58 +0300)]
CMAKE: added ie_libraries target (#1411)

3 years ago[IE CLDNN] Gemm int8 with slm optimization. Fused ops fix (#1319)
Ilya Znamenskiy [Tue, 21 Jul 2020 14:45:42 +0000 (17:45 +0300)]
[IE CLDNN] Gemm int8 with slm optimization. Fused ops fix (#1319)

3 years ago[IE TESTS] Added single layer test Mish (#1401)
Roman Lyamin [Tue, 21 Jul 2020 12:11:05 +0000 (15:11 +0300)]
[IE TESTS] Added single layer test Mish (#1401)

3 years agoPre-processing(GAPI): ARM(NEON) integration + Split, Merge, Color conversion kernels...
Anna Khakimova [Tue, 21 Jul 2020 11:19:15 +0000 (14:19 +0300)]
Pre-processing(GAPI): ARM(NEON) integration + Split, Merge, Color conversion kernels on NEON (#1315)

3 years agoSplit CNNlayer validators (#1403)
Ilya Lavrenov [Tue, 21 Jul 2020 10:57:56 +0000 (13:57 +0300)]
Split CNNlayer validators (#1403)

* Simplified CNNLayer validators: keep only parsing of arguments

* Added validators as is

* Removed parsing arguments

* Removed Backetize validator

3 years ago[IE][VPU]: Added Gelu layer (#1195)
Nikita Kudriavtsev [Tue, 21 Jul 2020 08:36:35 +0000 (11:36 +0300)]
[IE][VPU]: Added Gelu layer (#1195)

* Implement Gelu stage in GraphTransformer
* Disable GeLU decomposition in the nGraph transformations

3 years agoRemoved addLayer implementation from ngraph impl (#1400)
Ilya Lavrenov [Tue, 21 Jul 2020 03:36:18 +0000 (06:36 +0300)]
Removed addLayer implementation from ngraph impl (#1400)

3 years agoRemove redundant node methods (#1324)
Ilya Churaev [Tue, 21 Jul 2020 03:02:00 +0000 (06:02 +0300)]
Remove redundant node methods (#1324)

* Remove placement

* Removed validate and infer eltwise

* Remove is eltwise

* Remove support broadcast and decompose

* Removed is_op, is_parameter, is_pattern

* Fixed code style

* Added is_constant and is_output

* Removed is_communicative and is_null

* Fixed code style

* Fixed typo

* Fixed comments

* Fixed typo

* Revert is_parameter, is_output, is_result for OpenCV build

3 years agoFix for issue (#1335)
Anna Khakimova [Mon, 20 Jul 2020 17:09:23 +0000 (20:09 +0300)]
Fix for issue (#1335)

3 years agoHeader length for supported model detection is increased (#1363)
Vladislav Volkov [Mon, 20 Jul 2020 15:11:22 +0000 (18:11 +0300)]
Header length for supported model detection is increased (#1363)

3 years agoAdd mish op to ngraph (#1187)
iliya mironov [Mon, 20 Jul 2020 14:43:32 +0000 (17:43 +0300)]
Add mish op to ngraph (#1187)

* Add mish op to ngraph

* Update mish op

* Set v4 namespase for tests

* Add mish to cmake

* Add comments for mish op.

* Refactoring code style

* Update version to v1 for Mish op

* Add value propogation test for Mish op

* Refactoring mish op according to review

* Fix mish version

* Update cmake file

* Fix mish value propogation unit test

* Add unit test for mish op

Co-authored-by: Your Name <you@example.com>
3 years agoDoc Migration (master) (#1377)
Nikolay Tyukaev [Mon, 20 Jul 2020 14:36:08 +0000 (17:36 +0300)]
Doc Migration (master) (#1377)

* Doc Migration from Gitlab (#1289)

* doc migration

* fix

* Update FakeQuantize_1.md

* Update performance_benchmarks.md

* Updates graphs for FPGA

* Update performance_benchmarks.md

* Change DL Workbench structure (#1)

* Changed DL Workbench structure

* Fixed tags

* fixes

* Update ie_docs.xml

* Update performance_benchmarks_faq.md

* Fixes in DL Workbench layout

* Fixes for CVS-31290

* [DL Workbench] Minor correction

* Fix for CVS-30955

* Added nGraph deprecation notice as requested by Zoe

* fix broken links in api doxy layouts

* CVS-31131 fixes

* Additional fixes

* Fixed POT TOC

* Update PAC_Configure.md

PAC DCP 1.2.1 install guide.

* Update inference_engine_intro.md

* fix broken link

* Update opset.md

* fix

* added opset4 to layout

* added new opsets to layout, set labels for them

* Update VisionAcceleratorFPGA_Configure.md

Updated from 2020.3 to 2020.4

Co-authored-by: domi2000 <domi2000@users.noreply.github.com>
3 years agoAdded default constructor for RNNCellBase, fix conversions (#1370)
Ivan Tikhonov [Mon, 20 Jul 2020 11:15:37 +0000 (14:15 +0300)]
Added default constructor for RNNCellBase, fix conversions (#1370)

3 years agoONNX CI: add docker image cleanup (#1394)
Rafal Blaczkowski [Mon, 20 Jul 2020 10:42:52 +0000 (12:42 +0200)]
ONNX CI: add docker image cleanup (#1394)

* change stage name
 * add docker image cleanup

3 years agoAdd getting/setting friendly name for Node wrapper Py API (#1286)
Jan Iwaszkiewicz [Mon, 20 Jul 2020 08:31:49 +0000 (10:31 +0200)]
Add getting/setting friendly name for Node wrapper Py API (#1286)

3 years agoReference implementations for ReduceLogicalAnd & ReduceLogicalOr (#1333)
Tomasz Dołbniak [Mon, 20 Jul 2020 08:20:05 +0000 (10:20 +0200)]
Reference implementations for ReduceLogicalAnd & ReduceLogicalOr (#1333)

3 years agoAdd Acosh operator to ONNX importer opset4 (#1351)
Jozef Daniecki [Mon, 20 Jul 2020 08:18:03 +0000 (10:18 +0200)]
Add Acosh operator to ONNX importer opset4 (#1351)

3 years ago[Stress] Integrate compare_memcheck_2_runs call in run_memcheck.py (#1036)
Vitaliy Urusovskij [Mon, 20 Jul 2020 07:23:51 +0000 (10:23 +0300)]
[Stress] Integrate compare_memcheck_2_runs call in run_memcheck.py (#1036)

* [Stress] Define Database constant arguments in memcheck_upload.py only

* [Stress] Simplify computations using HashableDict in `compare_memcheck_2_runs`

* [Stress] Add comparison using pandas

3 years agoRemoved tests for old IR reader (#1368)
Ilya Lavrenov [Sat, 18 Jul 2020 09:42:40 +0000 (12:42 +0300)]
Removed tests for old IR reader (#1368)

3 years agoRemoved v0 fused operations (#1263)
Ilya Churaev [Fri, 17 Jul 2020 16:51:04 +0000 (19:51 +0300)]
Removed v0 fused operations (#1263)

3 years agoReduceLp operation specification (#1205)
Anton Chetverikov [Fri, 17 Jul 2020 13:13:19 +0000 (16:13 +0300)]
ReduceLp operation specification (#1205)

* Add ReduceL2 operation specification

* Add example

* Update operation specification

* Update operation specification

* Update operation specification

* Update operation specification

* Update types in operation specification

3 years ago[nGraph][Py][Tests] Remove backup of old models (#1349)
Tomasz Socha [Fri, 17 Jul 2020 13:12:11 +0000 (15:12 +0200)]
[nGraph][Py][Tests] Remove backup of old models (#1349)

3 years agoCODEOWNERS: Add .ci & docs
Alexander Zhogov [Fri, 17 Jul 2020 12:07:58 +0000 (15:07 +0300)]
CODEOWNERS: Add .ci & docs

3 years ago[IE][VPU][GT]: Use topological order in shape allocation (#1281)
Andrew Bakalin [Fri, 17 Jul 2020 08:47:37 +0000 (11:47 +0300)]
[IE][VPU][GT]: Use topological order in shape allocation (#1281)

* Some pass creates datas duplicate with a different order from time to time (because of unordered_set usage). It leads to a different order in model->datas() list and affects the shape allocation process which relies on this order.
* Make shape allocation be relied on topological order of datas which is stable and doesn't depend on order datas creation during different passes.

3 years agoFix unsafe use of CPU_ISSET_S macro. (#1357)
Ilya Lavrenov [Fri, 17 Jul 2020 07:51:56 +0000 (10:51 +0300)]
Fix unsafe use of CPU_ISSET_S macro. (#1357)

Don't increment mapped_idx via prefix increment within the argument of the
potentially unsafe CPU_ISSET_S macro. If the macro is expanded so that the
increment expression is evaluated multiple times, it will return unexpected
results.

While the glibc implementation of CPU_ISSET_S macro seems to be safe, the musl
libc (v1.1.23) version is unsafe and will evaluate the first argument of
CPU_ISSET_S three times.

Co-authored-by: Christian Priebe <cp3213@ic.ac.uk>
3 years agoRemove get_arguments (#1323)
Ilya Churaev [Fri, 17 Jul 2020 06:50:06 +0000 (09:50 +0300)]
Remove get_arguments (#1323)

* Removed get_arguments

* Fixed code style

3 years ago[IE][VPU]: CMX limit compile option (#1268)
Nikita Kudriavtsev [Thu, 16 Jul 2020 16:54:53 +0000 (19:54 +0300)]
[IE][VPU]: CMX limit compile option (#1268)

In some networks, mvTensor would request a large CMX-DMA transfer (<512K). That starves DMA for other timing critical tasks such as SIPP. Limit CMX-DMA request size as an option in myriad_compile:
* Add compile option TILING_CMX_LIMIT_KB
* Declare compile option TILING_CMX_LIMIT_KB in IE tools (compile_tool and vpu_compile)
* Add tests for compile option TILING_CMX_LIMIT_KB. Small fix for naming behavior tests.

3 years agoAzure CI: Fix cloning git submodules (#1356)
Alexander Zhogov [Thu, 16 Jul 2020 16:29:54 +0000 (19:29 +0300)]
Azure CI: Fix cloning git submodules (#1356)

3 years agofix: inference-engine/ie_bridges/python/sample/requirements.txt to reduce vulnerabili...
Andrey Somsikov [Thu, 16 Jul 2020 15:58:26 +0000 (18:58 +0300)]
fix: inference-engine/ie_bridges/python/sample/requirements.txt to reduce vulnerabilities (#1007)

The following vulnerabilities are fixed by pinning transitive dependencies:
- https://snyk.io/vuln/SNYK-PYTHON-NUMPY-73513

Co-authored-by: snyk-bot <snyk-bot@snyk.io>
3 years ago[IE TESTS] Add PReLu single layer test (#1306)
Irina Efode [Thu, 16 Jul 2020 15:16:57 +0000 (18:16 +0300)]
[IE TESTS] Add PReLu single layer test (#1306)

* [IE TESTS] Add PReLu single layer test

* [IE TESTS] Small refactoring of test infrastructure

* [IE TESTS] Add shape for activation

3 years agoRevert "Header length for supported model detection is increased (#1340)" (#1354)
Rafal Blaczkowski [Thu, 16 Jul 2020 15:09:45 +0000 (17:09 +0200)]
Revert "Header length for supported model detection is increased (#1340)" (#1354)

This reverts commit 5b2ec7840ac3c5c1d1ec107e50e499f22c9fe7a6.

3 years agoRemove implicit conversion from getInputTo, getLayerCreator (#1274)
Ilya Lavrenov [Thu, 16 Jul 2020 13:44:48 +0000 (16:44 +0300)]
Remove implicit conversion from getInputTo, getLayerCreator (#1274)

* Added ctor for CNNNetworkImpl to convert from ngraphImpl

* Re-use in all places instead of manual conversion

* Hide convertToCNNNetworkImpl usage

* Removed conversion from getCreatorLayer

* Fixes 2

* Fixes 3

* Fixes 4

* Fixed ieFuncTests

* Fixed more tests

* Fixed LPT tests

* Remove useless test

* Fixed GNA

* Fixed Gleb's comments

* Fixed Core integration tests

* Trying to fix python

* Fixed GPU tests

* Small fixes

* Fixed QueryNetwork after removing implicit conversion

* Fixed Core integration tests for QueryNetwork

* Fixed python; MULTI device QueryNetwork

* Fixed MULTI QueryNetwork

* Removed unused methods

* Enabled LPT FullyConnectedTestModel test

* Fixed typo in python

3 years ago[Stress] Specify OMZ `develop` branch in get_testdata.py (#1318)
Vitaliy Urusovskij [Thu, 16 Jul 2020 12:28:25 +0000 (15:28 +0300)]
[Stress] Specify OMZ `develop` branch in get_testdata.py (#1318)

Since in validation OMZ is used from a package (and OMZ is from develop branch), align OMZ default for get_testdata.py

3 years ago[IE TESTS] Small refactoring of test infrastructure (#1332)
Irina Efode [Thu, 16 Jul 2020 11:20:08 +0000 (14:20 +0300)]
[IE TESTS] Small refactoring of test infrastructure (#1332)

3 years agoAllow python benchmark_app load onnx model (#1283)
Nadezhda Ageeva [Thu, 16 Jul 2020 10:53:43 +0000 (13:53 +0300)]
Allow python benchmark_app load onnx model (#1283)

3 years agoSpecify operation CTCLoss-4 (#1189)
Roman Kazantsev [Thu, 16 Jul 2020 09:36:15 +0000 (14:36 +0500)]
Specify operation CTCLoss-4 (#1189)

* Specify operation CTCLoss-4

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Correct documentation for CTCLoss after #1 review

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Correct documentation for CTCLoss after #2 review

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

* Correct documentation for CTCLoss after #4 review

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Correct layout for logits and add more description for unique attribute

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Correct types for length and indices tensors

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Correct formulas and punctuation

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
3 years ago[IE CLDNN] Adjustment of layouts to choose optimal deconvolution (#781)
Alexander Chaiko [Thu, 16 Jul 2020 08:54:24 +0000 (10:54 +0200)]
[IE CLDNN] Adjustment of layouts to choose optimal deconvolution (#781)

3 years agoHeader length for supported model detection is increased (#1340)
Vladislav Volkov [Thu, 16 Jul 2020 08:12:40 +0000 (11:12 +0300)]
Header length for supported model detection is increased (#1340)

3 years agoAdded default multi threaded configuration (#1310)
Anton Pankratv [Thu, 16 Jul 2020 07:33:22 +0000 (10:33 +0300)]
Added default multi threaded configuration (#1310)

* Added default multythreaded configuration

* Fixed typo

3 years ago[IE Myriad] Remove Myriad 2 from supported devices in XLink (#1331)
Nikita Kudriavtsev [Thu, 16 Jul 2020 07:29:56 +0000 (10:29 +0300)]
[IE Myriad] Remove Myriad 2 from supported devices in XLink (#1331)

3 years agoremove nGraph deprecated methods (part 1) (#1314)
Ilya Churaev [Thu, 16 Jul 2020 03:03:59 +0000 (06:03 +0300)]
remove nGraph deprecated methods (part 1) (#1314)

* Remove remove_goe

* Remove traverse_nodes

* Removed deprecated constructors

* Removed deprecated tensor methods

* Fixed IE build

* Fixed code style

3 years ago[IE][VPU][nGraph]: Enables merging subsequent DSR operations (#1326)
Gladilov, Gleb [Wed, 15 Jul 2020 19:21:19 +0000 (22:21 +0300)]
[IE][VPU][nGraph]: Enables merging subsequent DSR operations (#1326)

Myriad plugin treats DSR operation in a way removing such operations
and connecting inputs with each other (replacing output with one of them).
Semantic of connection is one inputs contains shape of another.
Since the same data object can have exactly one shape it's prohibited
to have DSR inputs connected with another data objects
(the only allowed exception is inputs that are already connected between
each other).

As a result of nGraph -> CNN conversion some operations could be optimized
out which in turn could lead to subsequent DSR operations where each has
its own shape sub-graph. Even if shape sub-graphs are identical it's not
visible to plugin that sees incorrect inputs (inputs of DSR are already
connected, but now with each other, when second DSR is parsed).

To overcome such issue (the reason is when operations are optimized out,
their shape sub-graphs are still there), additional ngraph
transformation should be introduced to merge subsequent DSR into single
DSR operation.

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
3 years ago[IE][VPU][nGraph]: Fixes Reshape's shape infer method (#1327)
Gladilov, Gleb [Wed, 15 Jul 2020 19:17:36 +0000 (22:17 +0300)]
[IE][VPU][nGraph]: Fixes Reshape's shape infer method (#1327)

Previously, if Reshape had input pattern with values [0, -1] - it
propagated dynamic shape through a function. At the same time,
taking "0" and "-1" interpretation into consideration, it turns out
in such cases we could just propagate the same input shape in case of
2D input.

For Faster-RCNN this fix makes static dimensions on dynamic paths static.

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
3 years ago[IE][VPU][nGraph]: Fixes StridedSlice DTS (#1328)
Gladilov, Gleb [Wed, 15 Jul 2020 16:43:31 +0000 (19:43 +0300)]
[IE][VPU][nGraph]: Fixes StridedSlice DTS (#1328)

* In case of Begin/End/Stride inputs of StridedSlice have rank less
than input data rank - remaining dimensions must be kept unchanged.
* Previous, implementation had UB in such cases - out of bound
vector element access

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
3 years agoAdd Input and Output class to Py API (#1284)
Jan Iwaszkiewicz [Wed, 15 Jul 2020 13:32:24 +0000 (15:32 +0200)]
Add Input and Output class to Py API (#1284)

3 years ago[ONNX] Exception handling refinements. (#1266)
Adam Osewski [Wed, 15 Jul 2020 12:02:18 +0000 (14:02 +0200)]
[ONNX] Exception handling refinements. (#1266)

3 years ago[IE Common][Tests] saturated_cast: refactoring & tests (#1304)
Andrew Bakalin [Wed, 15 Jul 2020 10:48:57 +0000 (13:48 +0300)]
[IE Common][Tests] saturated_cast: refactoring & tests (#1304)

* [IE Common] Refactor saturated_cast

* [IE Common][Tests] Add tests for saturated casts

* [IE Common] Review fixes

* [IE Common] Make enable_if check a template parameter

3 years agoReverse Sequence code clean up (#1303)
Jan Iwaszkiewicz [Wed, 15 Jul 2020 10:48:53 +0000 (12:48 +0200)]
Reverse Sequence code clean up (#1303)

3 years agoAdded mish layer doc opset (#1149)
iliya mironov [Wed, 15 Jul 2020 07:30:33 +0000 (10:30 +0300)]
Added mish layer doc opset (#1149)

* Added mish layer doc opset

* Refactoring mish spec

* Update mish spec

* Change output description of Mish layer

* Fix Mish according to review

* Refactoring Mish and GELU spec according to code review

* Update formula for ops in spec

* Refactoring spec text

* Update Mish opset

* Change Mish version from 1 to 4

* Sort opset4

Co-authored-by: Your Name <you@example.com>
3 years agoSpecify, review and approve operation Interpolate-4 (#1035)
Vladimir Gavrilov [Wed, 15 Jul 2020 07:27:56 +0000 (10:27 +0300)]
Specify, review and approve operation Interpolate-4 (#1035)

* Added documentation for Interpolate-3.

* Some fixes.

* Fixed some typos.

* Now Interpolate-3 is Interpolate-4.

* Fixed typo.

* DEleted unused 'mode' 'area'.

* Fixed some typos.

* Now 'axes' attribute is an input of Interpolate.

* Added description of variants of nearest_mode.

* Added descriptions of coordinate transformation modes.

* Now 'axes' is an optional input.

* Fixed typo.

3 years ago[IE TESTS] Add ShapeOf SingleLayerTest (#1285)
Anton Zaytsev [Tue, 14 Jul 2020 20:55:32 +0000 (23:55 +0300)]
[IE TESTS] Add ShapeOf SingleLayerTest (#1285)

* [IE TESTS] add single layer test ShapeOf

* [IE TESTS] update for master

* [IE TESTS] add subgraph test

* [IE TESTS] update todo in skip_tests_config

* [IE TESTS] update skip_tests_config

* [IE TESTS] update skip_tests_config

* [IE TESTS] update opset3

3 years agoupdate system requirements (#1321) (#1322)
Alexey Suhov [Tue, 14 Jul 2020 19:06:55 +0000 (22:06 +0300)]
update system requirements (#1321) (#1322)

* update system requirements

* update release version in readme

3 years agoFix that brings back the MULTI's ability to add/remove devices (to the priorities...
Maxim Shevtsov [Tue, 14 Jul 2020 16:15:52 +0000 (19:15 +0300)]
Fix that brings back the MULTI's ability to add/remove devices (to the priorities list) on the fly. Presumably was lost during refactoring. (#1309)

the point is that we should check the ORIGINALLY (largest) list of the devices (actually ExecutableNetworks for them) to see if the device is just added back

3 years ago[LPT] FuseFakeQuantizeAndScaleShift transformation for last layer fix (#1291)
Edward Shogulin [Tue, 14 Jul 2020 15:55:06 +0000 (18:55 +0300)]
[LPT] FuseFakeQuantizeAndScaleShift transformation for last layer fix (#1291)

* [LPT] FuseFakeQuantizeAndScaleShift transformation for last layer fix

* [LPT] refactoring

* [LPT] FuseFakeQuantizeAndScaleShift test: last layer name validation was added

3 years ago[IE CLDNN] WA to use bfyx format if applicable for bf(w)zyx Gather input (#1056)
Jedrzej Hajduczenia [Tue, 14 Jul 2020 15:00:51 +0000 (17:00 +0200)]
[IE CLDNN] WA to use bfyx format if applicable for bf(w)zyx Gather input (#1056)

3 years ago[IE CLDNN] Disable inserting reorders if num_dims mismatch (#1023)
Jedrzej Hajduczenia [Tue, 14 Jul 2020 14:59:20 +0000 (16:59 +0200)]
[IE CLDNN] Disable inserting reorders if num_dims mismatch (#1023)

3 years agoAdd OpenVINO-ONNX CI check (#688)
Rafal Blaczkowski [Tue, 14 Jul 2020 12:57:27 +0000 (14:57 +0200)]
Add OpenVINO-ONNX CI check (#688)

3 years ago[IE][VPU][Tests]: Fix M2 on deprecated tests for StridedSlice (#1300)
Andrew Bakalin [Tue, 14 Jul 2020 10:13:56 +0000 (13:13 +0300)]
[IE][VPU][Tests]: Fix M2 on deprecated tests for StridedSlice (#1300)

* Disable reorder in Myriad2 cases in StridedSlice deprecated_tests

3 years ago[ONNX] Quantize linear using FakeQuantize (#1169)
Adam Osewski [Tue, 14 Jul 2020 08:55:07 +0000 (10:55 +0200)]
[ONNX] Quantize linear using FakeQuantize (#1169)

3 years agoEnable importing of TF_NASNet_Mobile (#1252)
Adam Osewski [Tue, 14 Jul 2020 08:54:39 +0000 (10:54 +0200)]
Enable importing of TF_NASNet_Mobile (#1252)

3 years agoRemoved max pool v0 (#1277)
Ilya Churaev [Tue, 14 Jul 2020 07:27:51 +0000 (10:27 +0300)]
Removed max pool v0 (#1277)

* Removed MaxPool v0

* Removed atan2

* Removed and operation

3 years agoAdded U32 precision (#1297)
Ilya Churaev [Tue, 14 Jul 2020 07:27:10 +0000 (10:27 +0300)]
Added U32 precision (#1297)

3 years ago[IE CLDNN] Optimize kernel cache memory usage in GPU plugin (#1233)
Mikhail Letavin [Mon, 13 Jul 2020 15:33:32 +0000 (18:33 +0300)]
[IE CLDNN] Optimize kernel cache memory usage in GPU plugin (#1233)

3 years ago[IE][VPU][nGraph]: Enables dynamic Reshape with non-const pattern support in myriad...
Gladilov, Gleb [Mon, 13 Jul 2020 15:19:05 +0000 (18:19 +0300)]
[IE][VPU][nGraph]: Enables dynamic Reshape with non-const pattern support in myriad plugin (#1159)

* [IE][nGraph]: Introduces PartialShape ctor from values vector

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
* [IE][VPU][nGraph]: Moves evaluateTargetShape to common utilities

The same functionality - get upper-bound shape estimation for dynamic
input - is needed in dynamic Reshape along with dynamic Broadcast.
Return value type has been changed from PartialShape to vector<int64_t>.
The reason is Reshape encodes special values (0, -1) into input values
that define output shape. Representing those values (which upper-bound
provides evaluateTargetShape) as PartialShape leads to incorrect
representation vector with -1 as dynamic shape - which is not expected.

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
* [IE][VPU][nGraph]: Introduces StaticShapeReshape

In comparison with original Reshape StaticShapeReshape propagates
upper-bound shape through a function in case of dynamic input. To do so,
shape inference method gets upper-bound shape from evaluateTargetShape,
decodes special values (0, -1) in it and then propagate the result.

Output shape processing happens only once, because if shape inference
were called after ShapeOf operations have been optimized out on dynamic
path, then evaluateTargetShape will require evaluate method for all
operations that appear in function before current Reshape. Since
evaluate method is implemented not for all operations it lead to
Faster-RCNN compilation error.

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
* [IE][VPU][nGraph]: Updates Reshape DTS on StaticShapeReshape

In case of non-const Reshape input that defines output shape DTS uses
StaticShapeReshape which propagates upper-bound shape evaluated from
this input through a function.

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
* [IE][VPU][nGraph][Tests]: Refactoring DTS Reshape tests

The only changes are:

* header files include reordering
* indentation/wrapping fixing

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
* [IE][VPU][nGraph]: Moves ShapeOf transformation out of DTS scope

In comparison with DTS ShapeOf transformation needs to work on whole
function. Separating these 2 transformations makes testing easier since
now it's possible to call specific DTS without ShapeOf transformation
and vice versa.

Also DynamicToStaticShapeOf has been renamed into
EliminateShapeOfAfterDSR since transformation doesn't introduce new DSR
operations.

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
* [VPU][Tests]: Introduces DTS Reshape tests with non-const pattern

New StaticShapeReshape constructor has been added as well, since test
fixture should create it from reshape parameters, not reshape itself.

Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
3 years agofix strided slice neg out of bounds ends (#1177)
Maxim Andronov [Mon, 13 Jul 2020 14:40:24 +0000 (17:40 +0300)]
fix strided slice neg out of bounds ends (#1177)

3 years ago[IE Samples] Add api arg to classification sample (#943)
Anastasia Kuporosova [Mon, 13 Jul 2020 11:48:40 +0000 (14:48 +0300)]
[IE Samples] Add api arg to classification sample (#943)

3 years ago[IE CLDNN] LRN int8 fsv16 optimizations (#814)
Egor Churaev [Mon, 13 Jul 2020 10:25:15 +0000 (03:25 -0700)]
[IE CLDNN] LRN int8 fsv16 optimizations (#814)

JIRA: 32367

3 years ago[CPU] Add check quantize ranges (#850)
Maxim Andronov [Mon, 13 Jul 2020 09:48:00 +0000 (12:48 +0300)]
[CPU] Add check quantize ranges (#850)

3 years ago[IE][VPU]: adjust batch - support dynamic number of iterations (#1114)
Andrey Sokolov [Mon, 13 Jul 2020 08:49:20 +0000 (11:49 +0300)]
[IE][VPU]: adjust batch - support dynamic number of iterations (#1114)

* support dynamic number of iterations in "AdjustBatch" pass
* add unit tests for this case

3 years agoThrow special exception if IR v7 is passed, but no IR v7 reader (#1293)
Ilya Lavrenov [Mon, 13 Jul 2020 03:13:59 +0000 (06:13 +0300)]
Throw special exception if IR v7 is passed, but no IR v7 reader (#1293)

3 years agoOptimisations for binary operations broadcast. Phase 2. (#1295)
Vladislav Volkov [Mon, 13 Jul 2020 03:11:02 +0000 (06:11 +0300)]
Optimisations for binary operations broadcast. Phase 2. (#1295)

3 years agoRemove old Scatter operations (#1265)
Ilya Churaev [Mon, 13 Jul 2020 03:02:20 +0000 (06:02 +0300)]
Remove old Scatter operations (#1265)

4 years ago[IE][VPU][GT]: Process StridedSlice stage on device as one kernel (#1244)
Andrew Bakalin [Fri, 10 Jul 2020 11:32:49 +0000 (14:32 +0300)]
[IE][VPU][GT]: Process StridedSlice stage on device as one kernel (#1244)

* Remove replacement of StridedSlice with other stages and execute it on device as one kernel.
* Refactor strided slice tests to be able to parametrize it by precision.
* Update firmware.

4 years ago[TESTS] Added Comparison and Logical single layer tests (#1242)
Gorokhov Dmitriy [Fri, 10 Jul 2020 10:56:22 +0000 (13:56 +0300)]
[TESTS] Added Comparison and Logical single layer tests (#1242)

4 years agoRemoved adjoints (#1269)
Ilya Churaev [Fri, 10 Jul 2020 10:49:43 +0000 (13:49 +0300)]
Removed adjoints (#1269)

4 years ago[GNA] Support in GNA plugin for power layer with non-1 exponents (#997)
Bartosz Sochacki [Fri, 10 Jul 2020 10:39:29 +0000 (12:39 +0200)]
[GNA] Support in GNA plugin for power layer with non-1 exponents (#997)

* added support for power layer with non-1 exponents to GNA plugin

* reverted a change caused by merge issue

* fixes for review comments (typo fix - lrelu instead of leru, unnamed structure instead of of named one in union with arguments of activation function, name fix - input instead of inputs),

scale-shift implementation based on affine layer instead of PWL,

* fixed code style

* fixes for coding style in scale_factor_calc.hpp

* added domain for power function

* fixed review comment - power function specific methods

* added check if dynamic casting was successful

* removed I16 as it is not supported by ngraph

* fixed initialization per review comment

4 years ago[GNA] MemoryStateImpl (#839)
Kamil Magierski [Fri, 10 Jul 2020 10:37:12 +0000 (12:37 +0200)]
[GNA] MemoryStateImpl (#839)

* [GNA] fix query state for GNAMemoryState

* [GNA] MemoryState implementation:

Fix shared tests
Fix smoke tests
fix SetState
Implement LastState
Fix Reset

* Move getPrecision() to GNAMemoryState

Change Smoke Reset() test to check resetting one state

* [GNA] add dequantize to getLastState()

* code refactor

Co-authored-by: Anna Alberska <anna.alberska@intel.com>
Co-authored-by: kmagiers <kmagiers@intel.com>