Commit Graph

6170 Commits

Author SHA1 Message Date
Ilya Lavrenov
9ef48cbe6d
Moved NCC check to U20 (#9631) 2022-01-14 07:50:55 +03:00
Ilya Lavrenov
5affdc3284
Added NCC support for source files (#9630) 2022-01-14 07:50:30 +03:00
Vladislav Golubev
1f241277f4
[ngraph] Add UnsqueezeConcatReduceFusion transformation (#9183)
* [ngraph] Add UnsqueezeConcatReduceFusion transformation

* Quick refactoring & build fix

* Transformation disabled in the GNA Plugin

* Review comments applied

* codestyle fix

* test fix

Co-authored-by: Aleksandr Pertovsky <aleksandr.pertovsky@intel.com>
2022-01-14 07:36:55 +03:00
mei, yang
e6d062e395
remove option ENABLE_SHUTDOWN_PROTOBUF (#9619)
* remove option ENABLE_SHUTDOWN_PROTOBUF

* shutdown protobuf library is enabled automatically if protobuf is used in frontend
2022-01-14 05:02:26 +03:00
Mingyu Kim
11cb8ebf55
[GPU] New debug option to serialize execution of inferences (#9510) 2022-01-14 10:24:40 +09:00
Mateusz Tabaka
508af22c66
Add CompressQuantizeWeights transformation (#7538)
* Add CompressQuantizeWeights transformation

It's based on model-optimizer/extensions/back/compress_quantized_weights.py

* handle dequantization subgraph after quantization

* fix scale shift calculation

* remove TRANSFORMATIONS_API from CompressQuantizeWeights

* ZeroPointOptimizer

* add CompressQuantizeWeights pass to to ApplyMOCTransformations

* add comment

* fix code style

* cleanup

* ambiguous copy_runtime_info

* ambigous call

* fix case when zero_point is close to zero

* fix cf test case

* move files

* update tests

* add tests for evaluate_subgraph

* Address review comments

* make params static
2022-01-13 23:38:39 +01:00
Fedor Zharinov
6c69535d6c
Benchmark_app batch calculation fix (#9554)
* BenchmarkApp - batch size calculation fix

* stylefix

* -ip/op fix

* stylefix
2022-01-13 23:34:38 +03:00
Fedor Zharinov
6a126ac6bb
avg statistics calculation fix (#9612) 2022-01-13 22:57:20 +03:00
Anastasia Kuporosova
2944f39bac
[Python API] Remove extra version info from get_version (#9645) 2022-01-13 22:38:02 +03:00
Maksim Proshin
ec0dd6051a
[POT] Minor correction about OpenSource (#9625)
* Update README.md

* Update README.md
2022-01-13 20:10:34 +03:00
Mateusz Tabaka
8ba339d16f
Eliminate no-op elementwise operations (#8840)
* Eliminate no-op elementwise operations

This change adds EliminateEltwise pass to NopElimination.
EliminateEltwise removes:
- Subtract with zero
- Multiply with one
- Divide by one

* add Add support to EliminateEltwise

* fix unit test

* use get_all_data_elements_bitwise_identical instead of get_single_value

* fix are_all_data_elements_bitwise_identical for constant created from HostTensor

* fix lpt tests

* check for mul in is_dequantization_subgraph function

* optimize fetching constant value

* apply review comments
2022-01-13 20:05:47 +03:00
Liubov Talamanova
cde61f2411
[POT] Fix bug with git (#9604)
* [POT] Fix bug with git

* Add UNKNOWN_VERSION constant
2022-01-13 19:56:18 +03:00
Nikita Malinin
0166cb8f5c
[POT] New IE API integration (#9004)
* Initial commit for integration

* Apply latest updates

* Update ie_engine for multi input models

* Update reference, pylint fixed, method replacement

* Change for AC 2.0 support

* Update aac_engine with raw data processing

* Update for pylint

* Update model utils

* Disable pylint check

* Update ac_engine with the processing

* Using tensor names instead of friendly names

* Tensor names for all statistics

* BiasCorrection with tensor names

* Remove debug info usage

* Fix for sequential stats

* Fix for the scalar stats

* Update with comments

* Update tests & builder

* Update tests

(cherry picked from commit 5022ce4a4fac230ac11db68f14cd58971fe9b2a2)

* Executable_model -> compiled_model

(cherry picked from commit 0d447ff4917f764df76d9c0c0068277e5e090f3b)

* Update with PartialShapes for reshape

* Update async mode

* Update refs

* Update BC Parameter dtype

* Update simplified with inferqueue

* Disable ssd512 nightly

* BC renaming

* Revert OMZ commit

* Update references

* Update BC algo with the parameter selection

* Update AC config filter

* Pylint fixes

* Update async data

* Update pylint for ie_engine

* Update ie_engine

* Revert refs changes

(cherry picked from commit 4d3530216b3d79c33c3437a7dfcc1140693320c8)

* Update OMZ submodule
2022-01-13 19:21:37 +03:00
Chenhu Wang
b33dcd02c6
fix windows debug test failure (#9635) 2022-01-13 18:53:20 +03:00
Vladimir Paramuzov
20d9d551eb
[GPU] Layout API improvement to replace direct named dimension usages from tensor (#9056) 2022-01-13 17:02:13 +03:00
Dawid Kożykowski
5059184789
remove unnecessary warning (#9243) 2022-01-13 14:50:56 +01:00
Vasily Shamporov
b7e8ef910d
Fix a paragraph in LPT docs w.r.t mixed precision (#9570) 2022-01-13 15:58:14 +03:00
Ilya Lavrenov
fdee2d5728
Disabled dynamic Windows build in Azure (#9568) 2022-01-13 15:44:04 +03:00
Alexander Zhogov
04c9a55b45
Revert "Deprecate GNA_LIB_N_THREADS and SINGLE_THREAD parameters (#9486)" (#9632)
This reverts commit c06d4e7fb7.
2022-01-13 15:41:15 +03:00
Roman Kazantsev
a53a4da3a2
[MO, Frontends] Implement ChangePlaceholderType as Offline Transformation (#8245)
* [Common FrontEnd] Implement ChangePlaceholderType transformation with OldApiMap

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Add ChangePlaceholderType pass into frontend normalize

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Add a comment about the transformation

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Fix build issue

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Apply code review comments: remove ngraph namespace rountine, etc.

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Remove use of extra Node object in set_old_api_map call

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Create separate pass for legacy nGraph transformations for MO with new frontend

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Fix build for ieFuncTests

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Fix build issue with namespace conflicts

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Revert "Fix build issue with namespace conflicts"

This reverts commit 50a189f4e5.

* Move legacy transformations to ngraph namespace

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Fix build issue with unresolved MOCLegacyTransformation

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Adjust code to new OldApiMap API

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Resolve merge conflicts

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Fix build issue

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Do not set legacy type for parameter in case user defined type

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
2022-01-13 15:29:51 +03:00
Ilya Churaev
f74b9caabb
Removed old dynamism API from infer request (#9557)
* Removed old dynamism API from infer request

* Fixed segfault

* Fixed some tests

* Added method to set inputs, outputs

* Fix python tests

* Try to fix window build
2022-01-13 14:54:14 +03:00
Ilya Lavrenov
9fe5165ffa
Support of clang-11 (#9620) 2022-01-13 14:31:25 +03:00
Nadezhda Ageeva
c06d4e7fb7
Deprecate GNA_LIB_N_THREADS and SINGLE_THREAD parameters (#9486) 2022-01-13 14:30:41 +03:00
Maxim Andronov
2b996c32ad
[CPU] FullyConnected dynamism support (#9380) 2022-01-13 13:18:02 +03:00
Gorokhov Dmitriy
c36afa495e
Temporarily disable onnx_if_inside_if unit test on CPU (#9624) 2022-01-13 13:14:37 +03:00
Mikhail Nosov
fc6eff048f
Python bindings: expose InputTensorInfo::set_memory_type (#9608)
* Exposing set_memory_type method

* Update quotes
2022-01-13 12:48:26 +03:00
Steve Yoo
a23a398dc0
Add SLT to Template Plugin: NonMaxSuppression-4,3,1 (#9511)
* Add SLT to Template Plugin: NonMaxSuppression-4

* Add SLT to Template Plugin: NonMaxSuppression-4,3,1
2022-01-13 12:42:56 +03:00
Jade Cho
35bcb77182
[GPU] Merge reorder into next onednn conv (#9601)
+ When the format of onednn convolution is bs_fs_yx_bsv32_fsv32, the reorder for dependency primitive of fused eltwise sum is removed.
2022-01-13 18:32:50 +09:00
Sergey Shlyapnikov
bd12648ec0
[GPU] Fix users invalidation and output layout recalculation in graph dump mode (#9439) 2022-01-13 12:29:43 +03:00
mei, yang
6ae0ba481e
fix build issue on android on shutdown protobuf (#9615)
* fix build issue on android on shutdown protobuf

* add ie_dependent_option for shutdown protobuf
2022-01-13 11:47:54 +03:00
Mateusz Bencer
080477aa9c
MO fallback to old path if some features are not supported yet (#9034) 2022-01-13 09:39:14 +01:00
Vladimir Dudnik
28fb55dffe
[IE Samples][OV2.0] final clean up of old API headers (#9494)
* final clean up of old API headers, compile_tool separated from samples

* make cpplint happy
2022-01-13 11:12:20 +03:00
Jade Cho
113014161d
Add a gpu func test to compare with ngraph runtime (#9591)
+ Compare conv result of gpu plugin with ngraph result.
2022-01-13 11:03:52 +03:00
Vladimir Paramuzov
ef3df6de03
[GPU] Fixed uninit axis field in pooling prim (#9598) 2022-01-13 10:53:20 +03:00
Vladimir Paramuzov
f68dd90413
[GPU] Get rid of PreAllocator usages to allow setShape (#9576) 2022-01-13 10:52:59 +03:00
Chen Xu
e329baef04
[CPU] Cache for Reduce node (#9519) 2022-01-13 10:42:24 +03:00
Vladislav Golubev
456347e1b6
[Transformations] ConvertDivide disabled transformation for divide ops from precision sensitive subgraphs (#9561)
* [Transformations] Skip ConvertDivide if Divide is on fp16 shapeOf subgraph

* lambda definition moved out of loop

* review fixes
2022-01-13 10:32:44 +03:00
song, bell
282f37cdae
fix sporadic crash (#9616)
Signed-off-by: fishbell <bell.song@intel.com>
2022-01-13 10:25:53 +03:00
Zhang Yi
6d3f99c7db
[CPU] Pooling & FakeQuantize cache (#9503)
* [CPUCache]Impl FakeQuantize

* [CPUCache]Impl Pooling

* [CPUCache]Fix Pooling ==

* [CpuCache]apply review comments

* [CpuCache]fix pooling ==
2022-01-13 10:25:39 +03:00
Mikhail Znamenskiy
5b40f381cb
Fix for benchmark_app: set model stream flags (#9609) 2022-01-13 10:15:33 +03:00
Chenhu Wang
ac45196ce2
[CPU] Interpolate runtime params cache (#9524) 2022-01-13 10:01:02 +03:00
Mang Guo
f562e5572f
[CPU] Enable cache for LRN (#9589) 2022-01-13 09:54:00 +03:00
Ilya Lavrenov
c3d2af7501
Fixed CVS-75003 (#9552)
* Fixed CVS-75003

* Allow dynamic shapes
2022-01-13 06:46:52 +03:00
Mingyu Kim
3ae3c8dfb5
[GPU] Enable a network with onednn deconv (#9597) 2022-01-13 12:03:08 +09:00
Indira Salyahova
c597bc8928
[POT] requirements install telemetry library (#9605)
* Update telemetry.py

* Update setup.py

* Update test_telemetry.py

* Update telemetry.py
2022-01-13 03:30:58 +03:00
Ilya Lavrenov
f222814f0c
Updated documentation for OV Inference Runtime (#9573)
* Updated documentation for OV Inference Runtime

* Updated documentation

* Fixed comments
2022-01-13 01:05:17 +03:00
Eugeny Volosenkov
00da13b058
Support TF ScatterND and CenterNet (#9257)
* add ScatterND for tf

* Fix ApplyPermutation

* fix description of TFScatterNDDecomposition

* fix permutation

* fix package_POM

* fix pom file

* fix BOM file

* fix bom file

* Added layer tests for TF ScatterND

* fix comments

* fix bom file

* Add additional tests

* Add additional tests

* Added ConvertLike to ScatterND decomposition
2022-01-12 23:15:45 +03:00
Mikhail Nosov
896532ace2
[OV2.0] PrePostProcessor dump to output stream for debugging purposes (#9580)
* Calculate model layout based on 'tensor' layout and convert steps

Previously, 'model layout' is set to '...' by default,
thus no shape conversion happened when tensor layout is set to 'NHWC', then there was explicit convert_layout "NCHW"

Now "model layout" is calculated based on tensor layout and conversion steps:
Examples:
 1) Tensor: NHWC, Convert: NCHW. Result: NCHW
 2) Tensor: NHWC, Convert: 0312. Result: NCHW

* Initial move of tensor data calculation

* Moved 'impls' to new file

* Postprocessing + unit tests

* clang-format fix

* Added more details to preprocessing nodes
- Mean/Scale - will print mean/scale values
- Convert type - will print type
- Convert layout - will print destination layout
- Convert color - will print destination color

It is needed to troubleshoot the problems. If error occurs, message will not display last op's target shape/layout/type

* Add python bindings

* update tests

* Added memory type to dump if set

* Code style fix

* unity build fix

* Dump tensor if only memory type is set

* Added debug print

* Fix Param->Result case

Previously, layout was set by preprocessing set to old parameter as well
This is incorrect because in case of exception layout info will not be reverted

In this case old Result pointed to old Parameter and was able to preserve runtime info

After fixing of this, case Param->Result was broken if revalidation is not triggerred

Fix is to detect 'Result' as a consumer of some parameter and force revalidation in this case

* Revert occasionally committed line

* And one more line
2022-01-12 22:00:32 +03:00
Ilya Churaev
e78ada1178
Introduce dev_api for core component (#9571)
* Introduce dev_api for core component

* Fixed CPU plugin for static build

* Fixed comments
2022-01-12 20:23:18 +03:00
Vladislav Golubev
01689ee408
[CommonOptimizations] SimplifySecondInputOfReshape fix (#9210)
* [CommonOptimizations] SimplifySecondInputOfReshape fix

* GroupedGatherElimination: added Gather v8 support
2022-01-12 17:50:49 +03:00