Tomasz Dołbniak [Tue, 18 Aug 2020 10:43:54 +0000 (12:43 +0200)]
v1::Pad evaluator (#1771)
* v1::Pad reference implementation
* ut fix: pad_negative_exterior_1d
* ut fix: pad_negative_exterior_1d_check_limits & pad_edge_1d
* Code formatting
* ut fix: pad_edge_1d_top_neg & pad_edge_1d_top_neg_bigger_than_tensor
* More Pad UT fixes
* Pad UT fixes: REFLECT mode
* Fix all Pad UTs
* Switch Pad evaluation in INT backend
* Non-template solution to v1::Pad::evaluate
* Always create v1::Pad with 4 inputs
* VS compilation error fix
* Python test fix
* Remove the v0::Pad constant folding pass
* Some extra checks in v1::Pad evaluator
* Code formatting
* Remove an obsolete CF test
Mateusz Bencer [Tue, 18 Aug 2020 10:10:23 +0000 (12:10 +0200)]
Add ONNX Round op (#1605)
Gladilov, Gleb [Tue, 18 Aug 2020 09:09:06 +0000 (12:09 +0300)]
[IE][VPU]: Reverts nGraph TI Unrolling for VPU (#1824)
Evgeny Latkin [Tue, 18 Aug 2020 09:05:18 +0000 (12:05 +0300)]
[IE][VPU]: Fixed MX firmware memory type patching (#1825)
Fixing Myriad X firmware loader after RTEMS 5 related changes merged into MDK
* Fixed MX firmware memory type patching
* Update firmware
* disable myriadMultipleGraphsTests_nightly
Co-authored-by: Chekhonin Andrey <andrey.chekhonin@intel.com>
Gleb Kazantaev [Tue, 18 Aug 2020 08:10:24 +0000 (11:10 +0300)]
ConvertPrecision for element::boolean (#1772)
* Added bool to u8 conversion
* Added opset1::ShapeOf handler
* Added ReduceLogicalAnd/Or support in ConvertPrecision pass
* Moved static map inside function; Updated callbacks
* Removed header
* Fixed tyle relaxed for cases when the same output consumes by multiple inputs in the same operation; added tests; fixed input types setting for already created type relaxed operations
Anna Likholat [Tue, 18 Aug 2020 05:38:48 +0000 (08:38 +0300)]
Java bindings (#506)
* java bindings added
* bulid saples alongside with IE build
* fixes
* fixes
* fixed comments
* fixes
Ilya Lavrenov [Mon, 17 Aug 2020 19:46:12 +0000 (22:46 +0300)]
Removed NGRAPH_API from reference library (#1821)
Tomasz Socha [Mon, 17 Aug 2020 19:14:39 +0000 (21:14 +0200)]
[FIX] Fix data layout for reshaped network (#1748)
* [FIX] Fix data layout for reshaped network
* [PATCH] Don't change compatible layouts
* Add UT for reshaped network
* FIX no. 1
Ilya Churaev [Mon, 17 Aug 2020 16:43:11 +0000 (19:43 +0300)]
Add static library with nGraph reference implementations (#1810)
* Removed reference implementations from public API
* Remove coordinate_transform from public API
* Introduced static library with reference implementations
Roman Kazantsev [Mon, 17 Aug 2020 16:19:59 +0000 (19:19 +0300)]
Extend MO for operation CTCLoss and partly refactor CTCGreedyDecoder (#588)
* Extend MO for operation CTCLoss
* Change sequence length format to a mask format
* Add fixes after first-round review
* Add fixes after the second-round review
* Fixing CTCLossPlusCTCGreedyDecoder transformation
Maksim Doronin [Mon, 17 Aug 2020 14:30:38 +0000 (17:30 +0300)]
[IE][VPU]: Refactor vpu configs (#1334)
* [IE][VPU]: Refactor vpu configs
* [IE][VPU][GT]: Refactor vpu configs usage in GT
* [IE][VPU]: Refactor vpu configs usage in myriad plugin
* [IE][VPU]: Refactor vpu configs usage in tests
* [IE][VPU]: Refactor vpu configs usage in deprecated_tests
* [IE][VPU]: Refactor vpu configs usage in tools
* [IE][VPU]: Use hddl metric vars instead of macros
* [IE][VPU]: style fixes
* [IE][VPU]: Replace MYRIAD_FORCE_RESET with MYRIAD_ENABLE_FORCE_RESET
* [IE][VPU]: Resolve confilcts with MYRIAD_TILING_CMX_LIMIT_KB
* [IE][VPU]: Remove MYRIAD_PLATFORM from tools
* [IE][VPU]: Fix DISABLE_REORDER config in StridedSlice
* [IE][VPU]: Redesign config headers hierarchy
* [IE][VPU]: Clarify comments related to vpu configs
* [IE][VPU]: Leave default values of deprecated configs
* [IE][VPU]: Revert beh tests on VPU_MYRIAD_PLATFORM
* [IE][VPU]: Revert beh tests with deprecated configs
* [IE][VPU]: Resolve conflicts
* [IE][VPU]: Revert opencv dependency
Evgeny Lazarev [Mon, 17 Aug 2020 13:32:53 +0000 (16:32 +0300)]
Enable ReduceL1 and ReduceL2 operations (#1799)
* Initial version of ReduceL1, ReduceL2 and ReduceLp enabling in the MO
* Added operations ReduceL1 and ReduceL2 to nGraph
* Removed ReduceLp. Added ReduceL1 and ReduceL2
* Separated specification of ReduceLp into ReduceL1 and ReduceL2
* Updated ReduceL1 and ReduceL2 specification
* Fixed ReduceL1 and ReduceL2 type prop tests
* Implemented nGraph transformation to decompose ReduceL1 and ReduceL2. Disabled them for CPU and GPU plugins
* Updated supported framework layers
* Added unit tests for ReduceL1 and ReduceL2 reference implementation
* Fixed ReduceXXX operations reference implementation by adding support for a new parameter 'keep_dims'
* Fixed constant folding for v0::Any
* Added ReduceL1 and ReduceL2 to Python API
* Implemented ReduceL1 and ReduceL2 decomposition tests and fixed ReduceL2 decomposition
* Added specific creator for ReduceXXX operations instead of NodeBuilders
* Fixed conversion ReduceXXX to CNNLayer
* Fixed parser for ReduceLogicalXXX operations
Mikhail Letavin [Mon, 17 Aug 2020 13:27:33 +0000 (16:27 +0300)]
[IE CLDNN] Export runtime graph as nGraph function (#1684)
Gleb Kazantaev [Mon, 17 Aug 2020 12:39:58 +0000 (15:39 +0300)]
FGTT for CF (#1792)
* Added FGTT for ConstantFolding
* Added test for CF disabling
Maxim Vafin [Mon, 17 Aug 2020 11:38:41 +0000 (14:38 +0300)]
Fix problem with renaming outputs from network (#1546)
Vladimir Paramuzov [Mon, 17 Aug 2020 11:34:49 +0000 (14:34 +0300)]
[IE CLDNN] All input layouts support in ref pooling (#1782)
Maxim Andronov [Mon, 17 Aug 2020 11:25:53 +0000 (14:25 +0300)]
[CPU] fix conv + concat case and add tests (#1747)
Ilya Churaev [Mon, 17 Aug 2020 10:37:00 +0000 (13:37 +0300)]
Removed template headers for Fused Ops (#1780)
Anastasia Kuporosova [Mon, 17 Aug 2020 09:41:54 +0000 (12:41 +0300)]
[Python API] Deprecate IENetLayer class (#1731)
* [Python API] Deprecate IENetLayer class
* update samples
* undo changes in samples
Vladimir Paramuzov [Mon, 17 Aug 2020 09:39:30 +0000 (12:39 +0300)]
[IE CLDNN] Set arguments once (#1699)
Mateusz Bencer [Mon, 17 Aug 2020 08:13:06 +0000 (10:13 +0200)]
Add reference implementation to VariadicSplit (#1698)
* implementation of variadic split ref impl
* Added tests
* added unit tests
* refactor
* review remarks
* fixed ie tests
Evgeny Latkin [Mon, 17 Aug 2020 08:02:00 +0000 (11:02 +0300)]
[IE][VPU][GT]: fix pool 3d to 2d reduction (#1760)
* Fix minor error in pooling 3D to 2D conversion
Ilya Znamenskiy [Mon, 17 Aug 2020 07:09:18 +0000 (10:09 +0300)]
[IE CLDNN] Gemm floating point unit-tests fix (#1797)
Sergey Shlyapnikov [Mon, 17 Aug 2020 06:56:23 +0000 (09:56 +0300)]
[IE CLDNN] Fix for DepthToSpace fusing into convolution (#1724)
Nikolay Shchegolev [Mon, 17 Aug 2020 03:10:51 +0000 (06:10 +0300)]
[NGraph] CTCLoss reference implementation. (#1453)
Ilya Churaev [Mon, 17 Aug 2020 03:05:08 +0000 (06:05 +0300)]
Remove nGraph layout (#1763)
Pavel Esir [Sun, 16 Aug 2020 12:49:49 +0000 (15:49 +0300)]
add preliminary support of Proposal-4 in nGraph (#1448)
renamed logits -> bbox_deltas
updated ngraph unittests for Proposal
removed validate_and_infer_types Proposal-4
removed validate_and_infer_types Proposal-4
changed validate_and_infer_types in parent class of Proposal
removed get_output_size
successfully inferred Proposal on SSH and Faster-RCNN
added unittests for Proposal-4
added unittests for Proposal-4
added unittests for Proposal-4
returned back default namespace for Proposal
reduced number of outputs in v0::Proposal
correct conversion of Proposal-4 -> propodal_ie with 2 outputs
removed creator for proposal v0
removed converter for proposal v0
added Proposal-4 to MO
removed `for_deformable` attribute
added Proposal-4 to MO and nGraph Python API
removed typo in Proposal-4 specification
style corrections
style corrections and removed some redundant code
rename proposal Python api test
removed 'attrs' context from visitor
returned back AttrVisitor to check if passes OpenVINO ONNX pipeline
Should pass OpenVINO ONNX pipeline (returned back AttrVisitor just to check)
python api for Proposal-4 works ok
(style correction) python api for Proposal-4 works ok
parametrized proposal_ie some other corrections
removed 'attrs.' context from nGraph Python API tests for Proposal
minor corrections in replacer proposal->proposal_ie
corrected Python API OpenVINO-ONNX tests should pass
Improved workaround for AttributeVisitor for Proposal
Add additional check of im_info tensor shape to Proposal node in MKLDNNPlugin
😠 removed 4 extra spaces from test_dyn_attributes.py to match The Style
added new nGraph RTTI declarations, removed throwing exception in transformation
added new nGraph RTTI declarations, removed throwing exception in transformation, corrected exception in MKLDNNplugin
corrected im_info size checking in Proposal node of MKLDNNPlugin
Anton Pankratv [Sat, 15 Aug 2020 11:44:13 +0000 (14:44 +0300)]
Added split test into template plugin (#1669)
* Added split test into template plugin
* Fixed split test
* Fixed test parameters style
Alexander Zhogov [Fri, 14 Aug 2020 22:31:34 +0000 (01:31 +0300)]
Azure CI: Pipeline improvements (#1790)
* Azure CI: Pipeline improvements
* Fix checkout
* Fix path
* Fix "Install IncrediBuild"
* Add "submodules: recursive"
* Fix test run
* Fix work dir
* Fix issue
* Increase timeout
Anastasiya Ageeva [Fri, 14 Aug 2020 15:47:58 +0000 (18:47 +0300)]
Added OpenVINO Get Started Guide for Windows (#1631)
Gleb Kazantaev [Fri, 14 Aug 2020 15:35:11 +0000 (18:35 +0300)]
Removed deprecated MulAddFusion pass (#1789)
Andrew Bakalin [Fri, 14 Aug 2020 15:10:57 +0000 (18:10 +0300)]
[IE][VPU][GT]: Add pass for marking fast stages (#1579)
It was discovered that it is sometimes useful to mark fast stages (e.g. stages that process less than 100 elements) in order to be able to parse them from final performance report and estimate its contribution into performance.
* Added pass for marking fast stages
* Introduced unit tests
Alexander Chaiko [Fri, 14 Aug 2020 13:02:37 +0000 (15:02 +0200)]
[IE CLDNN] Disable non-performant fp32 batches FC layer (#1737)
Mikhail Letavin [Fri, 14 Aug 2020 11:50:33 +0000 (14:50 +0300)]
[IE CLDNN] dp4a check that works both with old and new drivers (#1766)
Gleb Kazantaev [Fri, 14 Aug 2020 10:47:02 +0000 (13:47 +0300)]
opset4 Convolution/GroupConvolution -> Multiply fusion (#1754)
* Added new predicates for smart pattern matching
* Added ConvMul and GroupConvMul fusion passes based on opset4; Added CPU functional tests for comparing fusion accuracy
* Improved ConvMultiply fusion to support scalars; Added positive and negative tests
* Added ConvolutionBackprop/GrouConvolutionBackprop Multiply fusion; Added functional tests
Ilya Lavrenov [Fri, 14 Aug 2020 09:11:54 +0000 (12:11 +0300)]
Simplified plugin interfaces (#1745)
* Simplified plugin interface
* Allow not implemented
* Fixes
* Fixed CPU plugin tests
* Fixed tests dependencies
* Fixes
* Fixed GPU plugin compilation
* Renamed plugin
* Fixes
* Removed tests for plugin base
* Fix2
* Fix 2
* Define a macro to define plugin creation function
* Clean-up
* Fixed OSX build
* Fixed CentOS
* Fixed exception catch / throw
* Fixed clang issue
* Fixed python tests on macOsx
iliya mironov [Fri, 14 Aug 2020 07:07:58 +0000 (10:07 +0300)]
Add asinh acosh atanh to python api (#1488)
* Add asinh acosh atanh to python api
Ilya Churaev [Fri, 14 Aug 2020 03:27:58 +0000 (06:27 +0300)]
Deprecate FusedOp class (#1758)
* Deprecate FusedOps
* Try to fix windows
* Added temp headers
Alexander Zhogov [Thu, 13 Aug 2020 21:55:34 +0000 (00:55 +0300)]
Azure CI: Add IncrediBuild option: /MaxCPUS=40 (#1779)
* Azure CI: Add IncrediBuild options: /ShowTime /StopOnErrors /MaxCPUS=62
* set 48 cores
* Remove /StopOnErrors
* Set 40 cores
Ilya Lavrenov [Thu, 13 Aug 2020 17:17:30 +0000 (20:17 +0300)]
Updated mock interfaces in tests (#1762)
* Updated mock interfaces in tests
* Added mock_engine dependency
Jan Iwaszkiewicz [Thu, 13 Aug 2020 16:56:59 +0000 (18:56 +0200)]
[nGraph] Add Manager to Py API (#1533)
* Added test
* working ManagerWrapper
* Clean-up in ManagerWrapper
* worksave
* fixed building error
* Finished test of constant folding
* remove unused param
* Added get_vector function
* clean up
Alexander Peskov [Thu, 13 Aug 2020 16:06:20 +0000 (19:06 +0300)]
[CPU] Add explicit storage for MemoryNode (#895)
Sergey Lyalin [Thu, 13 Aug 2020 15:45:37 +0000 (18:45 +0300)]
TypeRelaxed implementation (#1561)
* RTTI base for ngraph::Node; cherry-pick from another branch, draft
* Added comments, moved code, switched to custom RTTI-based version of is_type
* Move rtti definitions in ngraph op class to the beginning of each class definition as a preparation for the next replacement
* Migrate part of operations to new RTTI
* Migrate GroupConvolution and Concat to new RTTI
* Apply code style for ngraph part
* Rename RTTI_DECLARATION/DEFINITION to NGRAPH_RTTI_DECLARATION/DEFINITION
* Reverted accidentally updated version of mkldnn
* TMP: rewrite RTTI back to constexprions as an attempt to fix static objects initialization order issue
* Apply ngraph code style
* Finalize move back to constexpr for RTTI
* Applied code-style
* TypeRelaxed template class implementation and necessary changes in ngraph + tests.
* Applied code-style
* Fix in fast algorithm in GraphRewrite, add new tests for this and other cases
* Make parent optional parameter for NGRAPH_RTTI_DECLARATION and remove Node::type_info; remove ability to have Node as a parent for type_info
* Try to resolve compilation error on Windows
* The next attempt to fix Windows build: re-introduce get_type_info_static
* Removed file that was removed in master and kept in this branch by mistake
* Next attempt to fix Windows build: externConstexpr
* Attempt to fix win build: extra public (suspect icc bug), remove get_type_info_static as useless.
* Next attempt to fix Windows: proxy const and constexpr
* Fixed constexpr
* Next attmpts: move get_type_info to cpp file
* Code stype fix
* Re-implemented RTTI without use of constexpr; run-time initialization is used; removed global definitions to avoid issues with order of static objects initialization
* Removed externConstexpr flag and removed TRANSFOMRATIONS_API for TypeRelaxed
* get_type_info_static initializes static local constant with type_info that is used for CLASS::type_info and CLASS::get_type_info
* Removed not needed debug output and useless comments
* Implemented better copy ctor for Node
* Fixed VisualizeTree issue for TypeRelaxed: stopped using < and > in type_info::name
* Better comments and names for methods
* Remove unused include
* Remove commented line
* Workaround for legacy conversion that uses Node::get_type_info().name as a type for the resulting CNNLayer leading to incorrect types for TypeRelaxed-based operations and then to fail in plugins
* Fixed typos, explicit ctor for TypeRelaxedBase, explanation for the need of get_overridden_output_type
* Fix typo
* Fixed issue with non-static name in type definition for TypeRelaxed and fixed WrapType to make it compatible with hierarchical relations between types
* Reverted default ctor for Output and reverted ability to reduce number of outputs for a Node; syntactically better debug message for a Node
* Cover methods of TypeRelaxedBase by tests
* Apply code-style
Andrew Bakalin [Thu, 13 Aug 2020 11:57:10 +0000 (14:57 +0300)]
[IE][VPU][GT]: Fix different blobs for the same network (#1738)
* Use vector instead of unordered_map in order to get stable blob serialization.
Evgeny Lazarev [Thu, 13 Aug 2020 10:20:29 +0000 (13:20 +0300)]
Updated a link to MO FAQ (#1750)
Alexander Zhogov [Thu, 13 Aug 2020 10:10:12 +0000 (13:10 +0300)]
Azure CI: Add Windows job with IncrediBuild (#1761)
Gorokhov Dmitriy [Thu, 13 Aug 2020 05:39:08 +0000 (08:39 +0300)]
[CPU] Disable quantize ranges validation in order to avoid regressions (#1720)
Ilya Churaev [Thu, 13 Aug 2020 04:17:24 +0000 (07:17 +0300)]
Remove some builders for old operations (#1736)
* Remove some builders
* Removed reshape v0 builder
* Fixed code style
Gleb Kazantaev [Thu, 13 Aug 2020 03:49:51 +0000 (06:49 +0300)]
nGraph passes clean up (#1742)
* Cleanup pass::Manager;Update VisualizeTree to inherit FunctionPass; Removed deprecated tranformations types
* Removed legacy code; Updated docs
Andrey Zaytsev [Wed, 12 Aug 2020 21:11:35 +0000 (00:11 +0300)]
Changed anchors (#1749)
azhogov [Wed, 12 Aug 2020 20:17:10 +0000 (23:17 +0300)]
Revert "Azure CI: Add Windows job with IncrediBuild (#1282)"
This reverts commit
41c5f2d2d67dc17c264ec54e2956fdda6c6efff9.
Alexander Zhogov [Wed, 12 Aug 2020 17:22:00 +0000 (20:22 +0300)]
Azure CI: Add Windows job with IncrediBuild (#1282)
* Azure CI: Add Windows job with IncrediBuild
* Update IB version to 9.4.6
* Fix "Clone submodules"
* Update IB version to 9.5
* Update install link
* Add debug out
* Update debug out
* Remove debug out
* Disable initiator machine from acting as helpers
Anna Alberska [Wed, 12 Aug 2020 14:37:00 +0000 (16:37 +0200)]
[GNA] fix scale factor issue in remove permutation test (#1740)
Mateusz Tabaka [Wed, 12 Aug 2020 14:06:09 +0000 (16:06 +0200)]
Improve SpaceToDepth tests (#1661)
Ilya Churaev [Wed, 12 Aug 2020 12:39:51 +0000 (15:39 +0300)]
Removed ngraph assertion (#1719)
Ilya Churaev [Wed, 12 Aug 2020 12:39:23 +0000 (15:39 +0300)]
Removed cpio (#1735)
Alexey Moskalev [Wed, 12 Aug 2020 10:17:34 +0000 (13:17 +0300)]
Update issue templates
Removing wrong template
Alexey Moskalev [Wed, 12 Aug 2020 10:16:22 +0000 (13:16 +0300)]
Update issue templates
removing wrong templates
Alexey Moskalev [Wed, 12 Aug 2020 10:13:41 +0000 (13:13 +0300)]
Update issue templates
First version
Ilya Lavrenov [Wed, 12 Aug 2020 09:00:30 +0000 (12:00 +0300)]
Define a macro to define plugin creation function (#1727)
Konrad Dobros [Wed, 12 Aug 2020 06:01:19 +0000 (08:01 +0200)]
[IE CLDNN] Add asymmetric dw convolution improvements (#1251)
This change adds full support for asymmetric quantization to optimized
depthwise convolution, adds slm optimization and other minor
improvements.
Issue: CVS-25122
Roman Kazantsev [Wed, 12 Aug 2020 05:55:56 +0000 (08:55 +0300)]
Add requirements_tf2.txt to package_BOM.txt (#1728)
Ilya Churaev [Wed, 12 Aug 2020 03:57:38 +0000 (06:57 +0300)]
Removed constant folding pass for reverse (#1716)
Dmitry Kurtaev [Wed, 12 Aug 2020 03:56:19 +0000 (06:56 +0300)]
size_t on 32bit OS (#1721)
Andrey Zaytsev [Tue, 11 Aug 2020 16:10:56 +0000 (19:10 +0300)]
Merging Documentation updates for 2020.4 (#1672) (#1726)
Evgenya Stepyreva [Tue, 11 Aug 2020 15:47:27 +0000 (18:47 +0300)]
[ MO ] Turning GNMT KSO OFF (#1718)
Ivan Tikhonov [Tue, 11 Aug 2020 15:46:57 +0000 (18:46 +0300)]
Unroll transformation for TensorIterator (#1259)
* unroll ti transformation, lstm sequence ie, rnn sequence ie
* Update unroll ti transformation, added GRUSequenceIE op, fixed several ti e2e tests
* apply ngraph codestyle
* fix naming after unroll transformation
* Added default constructor for RNNCellBase, fix conversions
* copy runtime info
* added UnrollTI unit tests
* clean up, move sequence ops in a separate PR
* clean up, ngraph code style
* temporary disable ngraph reader unit tests for ti
* fix unit tests on windows
* naming: use name of tensor after unroll tensor iteration transformation
* apply transformations to tensor iterator body, separate pass for ti transformations, fix naming issue
* fix build
* remove TensorIterationTransformations pass
* fix includes
* resolve conflicts
* fix build: incorrect includes
* remove split/concat for single iteration of TI, update to opset4, unit tests
* use matcher pass instead of graph rewrite
* try to enable UnrollTI transformation for all plugins
* disable unrollTI transformation for cpu plugin
* resolve review comments, enable unit tests
* update transformation description
* fix unit tests
* update transformation pipeline
* clean up
* clean up
* resolve review comments
Roman Kazantsev [Tue, 11 Aug 2020 15:02:05 +0000 (18:02 +0300)]
Separate MO configuration for TensorFlow 2 model conversion (#1685)
* Separate MO configuration for TensorFlow 2 model conversion
Also, it updates documentation including steps to convert
TF2 model with a custom layer in Keras H5 format into SavedModel
* Do fixes based on the first-round code review
Mateusz Tabaka [Tue, 11 Aug 2020 14:28:24 +0000 (16:28 +0200)]
Improve DepthToSpace tests (#1659)
Anastasia Kuporosova [Tue, 11 Aug 2020 14:27:49 +0000 (17:27 +0300)]
[Tools] Install compile tool to tool directory (#1649)
Tomasz Dołbniak [Tue, 11 Aug 2020 13:59:15 +0000 (15:59 +0200)]
Make Clip work for dynamic input (#1666)
Jan Iwaszkiewicz [Tue, 11 Aug 2020 13:57:31 +0000 (15:57 +0200)]
[nGraph] Create Python API support for rt_info (#1696)
Vladimir Paramuzov [Tue, 11 Aug 2020 13:40:52 +0000 (16:40 +0300)]
[IE CLDNN] Enabled fsv16 asymmetric first conv (#1372)
Ilya Churaev [Tue, 11 Aug 2020 12:28:14 +0000 (15:28 +0300)]
Remove GetOutputElement op (#1604)
Anna Alberska [Tue, 11 Aug 2020 12:20:19 +0000 (14:20 +0200)]
[GNA] Support of NHWC conv2d with N=1 H=1 and 1xk Kernel (#1209)
* [GNA] Support of NHWC conv2d with N=1 H=1 and 1xk Kernel
* [GNA] add test for comparing optimization outputs & cpplint fixes
* fix getInputTo() & fix cpplint
* fix tests
* revert kernel padding
* add AddConvolutionKernelPadPass & refactor
* cpplint fix
* fix CI issues & add layout sensitive dimensions
* move kernel padding issue to another branch
* add more legible error descriptions
* fix legacy tests & disable 3d input convolution tests
* change comment messages
* fix additional convolution kernel padding for PWL case
Egor Churaev [Tue, 11 Aug 2020 11:52:04 +0000 (14:52 +0300)]
[IE CLDNN] Fix reshape for yxfb layout (#1632)
In one of the network it was the following pipeline:
```
FullyConnected -> Reshape -> FullyConnected
```
And the output of Reshape wasn't in the same order as input for this
layer. I found that the problem was connected with format of the layers.
During optimization passes this pipeline was transformed to the
following:
```
FullyConnected -> Reorder -> Reshape -> Reorder -> FullyConnected
```
Both `FullyConnected` layers works with `yxfb` format. This is why
Reorder layer after the Reshape has output layout with format `yxfb` and
`reshape_in_layout.format` returns `yxfb` format. But in this case we
have to convert Reshape to `bfyx` format because in this case we won't
change the order of elements.
I replaced `reshape_in_layout.format` (which returns `yxfb`) and
explicitly set `bfyx` format.
JIRA: 35288
Pavel Rodionov [Tue, 11 Aug 2020 11:37:40 +0000 (14:37 +0300)]
[GNA] Bump GNA2 version to 1047 (#1629)
Pavel Rodionov [Tue, 11 Aug 2020 11:35:39 +0000 (14:35 +0300)]
[GNA] Remove empty PWL (#1224)
Anna Alberska [Tue, 11 Aug 2020 09:46:27 +0000 (11:46 +0200)]
[GNA] Add Basic_LSTM_S test (#805)
* add Basic_LSTM_S test
* add comparing with model with unrolled TI
* move computing reference output to overridden CalculateRefs()
Gleb Kazantaev [Tue, 11 Aug 2020 09:14:14 +0000 (12:14 +0300)]
Removed CNNNetwork BlobTransformer (#1709)
* Removed CNNNetwork BlobTransformer
* Removed inference_engine_lp_transformations dependency for GNA and VPU plugins
Denis Orlov [Tue, 11 Aug 2020 09:13:06 +0000 (12:13 +0300)]
[GNA] Fixes in checks, asserts, etc. (#903)
Ilya Churaev [Tue, 11 Aug 2020 08:11:33 +0000 (11:11 +0300)]
Changed structure of nGraph core library (#1658)
Rafal Blaczkowski [Tue, 11 Aug 2020 07:28:55 +0000 (09:28 +0200)]
Enable Model Zoo in OpenVINO-ONNX CI (#1660)
Katarzyna Mitrus [Tue, 11 Aug 2020 06:54:08 +0000 (08:54 +0200)]
HSwish operation specification (#1708)
* HSwish specification init
* Update docs/ops/activation/HSwish_4.md
Co-authored-by: Michał Karzyński <4430709+postrational@users.noreply.github.com>
* Update docs/ops/opset4.md
Co-authored-by: Michał Karzyński <4430709+postrational@users.noreply.github.com>
Ilya Znamenskiy [Tue, 11 Aug 2020 06:54:00 +0000 (09:54 +0300)]
[IE CLDNN] Gemm fp16/fp32 optimized kernel (#1646)
Evgenya Stepyreva [Tue, 11 Aug 2020 06:34:04 +0000 (09:34 +0300)]
[ MO ] Fixed layout interpretation for 4/5D tensors calculated from ShapeOfs (#1634)
Jan Iwaszkiewicz [Tue, 11 Aug 2020 04:16:11 +0000 (06:16 +0200)]
Fixed access to the data of FP16 IRs with nGraph Python API (#1707)
Ilya Lavrenov [Tue, 11 Aug 2020 04:06:24 +0000 (07:06 +0300)]
Moved plugin-specific utils from blob_factory.hpp (#1710)
Dmitry Kurtaev [Mon, 10 Aug 2020 20:10:46 +0000 (23:10 +0300)]
Add python executable for RPI compilation Docker (#1530)
Maxim Vafin [Mon, 10 Aug 2020 18:10:45 +0000 (21:10 +0300)]
[MO] Support ONNX QuantizeLinear (#1451)
* [MO] Support ONNX QuantizeLinear
* Update docs
* Fix cast type
* Fix error messages
Ilya Lavrenov [Mon, 10 Aug 2020 15:33:25 +0000 (18:33 +0300)]
Moved legacy API to legacy/ subfolder for include (#1677)
Ilya Churaev [Mon, 10 Aug 2020 15:25:44 +0000 (18:25 +0300)]
Removed onnx_import folder from src (#1706)
Ilya Lavrenov [Mon, 10 Aug 2020 15:24:54 +0000 (18:24 +0300)]
Changed ICNNNetwork to CNNNetwork in QueryNetwork (#1704)
Ilya Churaev [Mon, 10 Aug 2020 15:24:25 +0000 (18:24 +0300)]
Fixed input/output shape initialization (#1695)
* Fixed input/output shape initialization
* Use template_extension library in tests
Gleb Kazantaev [Mon, 10 Aug 2020 13:32:03 +0000 (16:32 +0300)]
Removed transformations _tbl.hpp files (#1700)
Evgeny Lazarev [Mon, 10 Aug 2020 12:51:21 +0000 (15:51 +0300)]
Enable swish (#1682)
* Draft version of the Swish nGraph operation and fusing transformations for different approaches to express the operation
* Swish fusing transformation refactoring
* Added Swish operation and extractor for TF. Removed unfolding transformation for the operation.
* Added SwishIE. Implemented transformation to convert Swish to SwishIE.
* Code style fixes
* Updated Swish reference implementation. Added tests for shape and value inference
* Fixed code style for Python API
* Fixed unit test
* Apply review comments
* Use matcher_pass_callback
* Make m_alpha attribute protected in the SwishIE operation
* Fixed Swish op PythonAPI test
Ilya Lavrenov [Mon, 10 Aug 2020 12:43:25 +0000 (15:43 +0300)]
Fixed CPU performance (#1702)
Evgenya Stepyreva [Mon, 10 Aug 2020 12:36:05 +0000 (15:36 +0300)]
[ MO ] Extended Const->Result replacer (#1688)
* [ MO ] Extended Const->Result replacer
Kamil Magierski [Mon, 10 Aug 2020 12:23:25 +0000 (14:23 +0200)]
[GNA] Fix cases when Gna2ModelGetLastError() returns unknown error (#1255)
Co-authored-by: kmagiers <kmagiers@intel.com>
Mateusz Bencer [Mon, 10 Aug 2020 11:48:18 +0000 (13:48 +0200)]
Extend dynamic shape support for ops which use auto padding mode (#1432)