IntenseWebs/openvino
3
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
3,963 Commits 38 Branches 54 Tags
ad1efb7f165dd79de3ed40a1c29ddafe94eb85ae
Commit Graph

3963 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
kblaszczak-intel
ad1efb7f16 remove font-family (#8211) 2021-10-27 12:49:15 +03:00
Elizaveta Lobanova
7eed1d910b [GNA] Limit activation output scale factor (#8106) 2021-10-26 15:38:12 +03:00
Elizaveta Lobanova
2278100258 [GNA] Fixed segments creation for multi segment pwl functions (#8044) (#8165) 
* [GNA] Fixed extreme segments creation for tanh pwl

* [GNA] Tests added

* [GNA] Added tests for different activations

* [GNA] Comments apply
 2021-10-25 11:57:40 +03:00
Nikolay Tyukaev
361dec7362 Docs to Sphinx (#8151) 
* docs to sphinx

* Update GPU.md

* Update CPU.md

* Update AUTO.md

* Update performance_int8_vs_fp32.md

* update

* update md

* updates

* disable doc ci

* disable ci

* fix index.rst

Co-authored-by: Andrey Zaytsev <andrey.zaytsev@intel.com>
 2021-10-24 17:43:00 +03:00
Alexey Lebedev
d7a917d92c [IE PYTHON] fix gil (#8145) 
* remove nogil

* Add test
 2021-10-22 16:13:52 +03:00
Roman Lyamin
326b8ad009 [GPU] Fixed matmul handling for some shapes (#6642) (#8103) 
Co-authored-by: Vladimir Paramuzov <vladimir.paramuzov@intel.com>
 2021-10-20 09:34:32 +03:00
Vladislav Golubev
cc6321f6d3 [LPT] MatmulTransformation: canBeTransformed fix (#7816) 2021-10-19 23:50:20 +03:00
Elizaveta Lobanova
5c2a39009e [GNA] Fixed insertion of delayed copy error (#7944) (#8037) 
* [GNA] Fixed error with delayed copy insertion

* [GNA] Added test
 2021-10-18 15:34:55 +03:00
Polina Brzezinskaya
a0b4394423 [VPU] Changed calling of sem_wait() to account for EINTR error code (#7815) 
Changed calling of sem_wait() to account for EINTR error code
 2021-10-15 16:18:14 +03:00
Dmitry Pigasin
85378ce176 Add --ignore-installed to the downloader requirement installation (#8025) 2021-10-15 10:42:23 +03:00
Gleb Kazantaev
e2fd335b70 Fix FrameworkNodeAttr Deserialization (#7943) 
* Fix FrameworkNodeAttr Deserialization

* Fix
 2021-10-14 13:27:50 +03:00
Elizaveta Lobanova
c5175063c8 [GNA] More precized calculation of Log pwl and pwl with sf < 1 (#7884) (#7982) 
* [GNA] More precized calculation of Log pwl and pwl with sf < 1

* [GNA] Added tests

* [GNA] Types correction

* [GNA] Added comment for absolute threshold calculation in the test
 2021-10-14 12:38:08 +03:00
Krzysztof Bruniecki
5fb5057cb6 [GNA] Fix pooling output backward compatibility (#7889) 
* [GNA] Fix pooling output backward compatibility

* Apply review
 2021-10-12 11:43:12 +03:00
Roman Lyamin
485bea73b1 [GPU] Fixed code gen with custom locale (#7548) 2021-10-08 12:54:34 +03:00
Maxim Vafin
105e67573a [MO] Fix axes in FusedBatchNorm -> MVN transformation (#7679) 
* [MO] Fix axes in FusedBatchNorm -> MVN transformation

* Improve testing

* Fix eps_mode attribute
 2021-10-08 11:21:50 +03:00
Alexander Zhogov
2282b0ee5c Change IE version to 2021.4.2 2021-10-05 13:47:30 +03:00
Anton Dudchenko
f00dc87a92 [IE][VPU] Fix execTimeMcs for VPU (#7442) (#7712) 
When serializing execGraph, milliseconds were actually written to the execTimeMcs field. It's just cherry-pick commit from master branch
 2021-09-29 21:16:33 +03:00
Mikhail Kozlov
e932a2eda5 Add virtual inheritance to shared tests (#7649) 2021-09-28 17:20:04 +03:00
Andrey Zaytsev
bd989cee67 [doc] Improved description of OpenVINO workflow (#6128) (#7703) 
* Update get_started_dl_workbench.md

POToolkit => POTool

* Update QuantizedNetworks.md

POToolkit => POTool

* Moving POT to optimizations section

Moving POT to optimizations section

* structure

* links

* step 1

* step 1 diagram

* step 2

* typo

* step 4

* step 4 diagram

* step 3

* minor corrections

* Applied comments

* Applied comments from Tatiana

* Applied comments from Alex
# Conflicts:
#	docs/index.md
#	inference-engine/thirdparty/ade

Co-authored-by: Maksim Proshin <mvproshin@gmail.com>
 2021-09-28 15:32:19 +03:00
Krzysztof Bruniecki
bdf72bcd88 [GNA] Fix KEY_EXEC_TARGET (#7671) 
* Use Gna2DeviceCreateForExport when GNA_EXEC_TARGET is != detected

* Update detected GNA device version field in GNA Device helper

   * Use EXEC instead of COMPILE TARGET to append
   CNN Legacy enforcement (GNA1)

* Apply review
 2021-09-28 11:55:28 +03:00
Egor Duplensky
6d634d09a4 [CPU] Fix mixing VEX and non-VEX instructions (#7238) 
Quote: The Skylake microarchitecture implements a different state
machine than prior generations to manage the YMM state transition
associated with mixing SSE and AVX instructions.
It no longer saves the entire upper YMM state when executing
an SSE instruction when in “Modified and Unsaved” state,
but saves the upper bits of individual register.
As a result, mixing SSE and AVX instructions will experience
a penalty associated with partial register dependency of
the destination registers being used and additional blend
operation on the upper bits of the destination registers.

Such type of penalties have a huge impact on openvino's and oneDNN's kernels.
Basically the mixing of VEX and non-VEX instructions should be avoided.
 2021-09-22 10:46:37 +03:00
Rafal Blaczkowski
05c641ff7c Update azure onnx linux 2021/4 (#7534) 
* update azure pipeline

* save
 2021-09-16 14:50:10 +03:00
Vitaliy Urusovskij
05a768e357 Bump up MKLDNN version to get fix of c26453 warning (#7090) 2021-09-16 12:57:39 +03:00
Szymon Irzabek
a0b2200408 Gna 2dconv30 releases (#7387) 
* [GNA] Port padding and 2d convolution support from master

* [GNA] Add separate fixes required after porting

* [GNA] Add support for activation without bias in padding and 2d convolution decomposition
 2021-09-15 14:43:42 +03:00
Krzysztof Bruniecki
e339afe375 IE: Fix Windows compilation on MSVC (adapt PR: #7206 from master branch) (#7486) 
`Win32` may be undefined when MSVC is used, see:
https://docs.microsoft.com/en-us/cpp/preprocessor/predefined-macros?redirectedfrom=MSDN&view=msvc-160

Signed-off-by: Karol Trzcinski <karolx.trzcinski@intel.com>
 2021-09-14 17:49:01 +03:00
Evgeny Talanin
7f47108bb3 Add apt update (#7484) 2021-09-13 13:52:21 +03:00
Elizaveta Lobanova
b13bcbcd0d [GNA] Fixed scale factors propagation for Eltwise with very different inputs ranges (#7305) (#7445) 
* [GNA] Fix scale factors propogation for Eltwise with very different inputs ranges

* [GNA] Added test

* [GNA] Added exception for scale factor <= 0

* [GNA] Disable tests with integer weights

* [GNA] Added assert for CNNLayer in getScaleFactor()

* [GNA] Added check if scale factor is inf

* [GNA] Fixed legacy tests
 2021-09-13 13:01:37 +03:00
Alexey Suhov
c2bfbf29fb [README.md] change latest release to 2021.4.1 2021.4.1 2021-09-10 00:08:38 +03:00
Artyom Tugaryov
14e67d8663 Check installed intel-opencl-icd to find version (#7427) 2021-09-08 17:22:18 +03:00
Artyom Tugaryov
d98cb7bdf8 Use dpkg-info to check installed driver (#7426) 2021-09-08 17:17:35 +03:00
Evgeny Talanin
fa18eecfb7 Revert "[IE CLDNN] Fixed code gen with custom locale (#6243)" (#7370) 
This reverts commit fadeaecb6d.
 2021-09-08 10:44:49 +03:00
Kate Generalova
3aa2f02240 doc: fix docker win links (#7402) 2021-09-07 15:17:14 +03:00
Sergey Lyubimtsev
a4fdc1c947 Fix for missed OMZ public models in openvino-dev package (#6568) (#7324) 
(cherry picked from commit b893eae9e7)
 2021-09-01 22:24:04 +03:00
Artyom Anokhov
f14f28f32b [Scripts] Detecting newest drivers on ubu20 (#7322) 
* install_NEO_OCL_driver: Added detecting current driver via intel-opencl-icd package in case of newest one on ubuntu20. Added removing intel-opencl-icd package for ubuntu.

* install_NEO_OCL_driver: Fixed pattern for parsing driver version for newest drivers
 2021-09-01 15:00:24 +03:00
Artyom Anokhov
e3baff25a6 install_NEO_OCL_driver: Added ocl-icd-libopencl1 for ubuntu18 case (#7314) 2021-08-31 19:46:51 +03:00
Anton Voronov
708825b439 [CPU] fixed convolution outputShape in ConvDWConv fusing (#7290) 2021-08-31 15:21:15 +03:00
Mikhail Letavin
91d85a88a1 [GPU] Remove unused copy constructor (#7221) 2021-08-31 14:14:51 +03:00
Denis Orlov
f77c3d7fdc [GNA] Include documentation for GNA3 QoS (#7046) 
* [GNA] Include documentation for GNA3 QoS

* Fix according to review

* Fix the driver version
 2021-08-31 12:54:05 +03:00
Artyom Anokhov
36f2e63c9c Update install neo driver script with 21.29.20389 (#7244) 
* install_NEO_OCL_driver: Added installing 21.29.20389

* install_NEO_OCL_driver: Updated CMDs for installing 21.29.20389 on Ubuntu

* Update scripts/install_dependencies/install_NEO_OCL_driver.sh

Co-authored-by: Kate Generalova <kate.generalova@intel.com>

* Update scripts/install_dependencies/install_NEO_OCL_driver.sh

Co-authored-by: Kate Generalova <kate.generalova@intel.com>

* install_NEO_OCL_driver: Updated getting values from dictionaries

* install_NEO_OCL_driver: Updated descriptions

* install_NEO_OCL_driver: Added more packages to install

* install_NEO_OCL_driver: Added ocl-icd-2.2.12-1.el8.x86_64 for RHEL

* install_NEO_OCL_driver: Updated error message. Replaced `dnf update --refresh` with --refresh option for dnf install. Added software-properties-common in case if user doesn't have apt-add-repository.

* install_NEO_OCL_driver.sh: Fixed syntax

Co-authored-by: Kate Generalova <kate.generalova@intel.com>
 2021-08-31 11:45:02 +03:00
Mikhail Ryzhov
d474617d12 [GNA] Fixed import of model with several inputs (#7277) (#7278) 
* [GNA] Fixed import of model with several intputs

* Fixed copyright year
 2021-08-31 11:09:57 +03:00
Elizaveta Lobanova
64a896c22e [GNA] Fixed accuracy degradation caused by the input quantization restriction (#7260) 2021-08-27 11:48:57 +03:00
Elizaveta Lobanova
2dcd09055f [GNA] Fixed calculation of input scale factor and search of the next layer for FQ (#7246) 
* [GNA] Fixed search of the next layer for FQ

* [GNA] Fixed calculation of input scale factor for POT-quantized model in the case if the first layer after input is activation
 2021-08-26 15:02:23 +03:00
Andrey Sapozhnikov
1a656f4e44 [GNA] Plugin transition to the library v3.0 (#7241) 
* GNA Plugin transition to the library v3.0

* Remove Gna2RequestConfigEnableHardwareConsistency calls

Co-authored-by: Krzysztof Bruniecki <krzysztof.bruniecki@intel.com>
 2021-08-26 01:00:07 +03:00
Mikhail Ryzhov
0702b44174 [Samples] Enabled oname for imported models (#7228) 2021-08-25 12:57:27 +03:00
Krzysztof Bruniecki
ce21344585 [GNA] Fixes for GNA 3.0 library (#7180) 
* Pass compileTarget to am_intel_dnn

* Enable tests for GNA lib version prefix 3.0

* Fix conv split transform for 2d cnn tests

* Apply review
 2021-08-25 10:06:10 +03:00
Ilya Sharikov
3a28ffaf57 Updated path to gflags (#7200) 2021-08-24 13:57:40 +03:00
Anton Dudchenko
f63be649eb Update FW (#7209) 2021-08-24 12:02:23 +03:00
Polina Brzezinskaya
32a9e98437 [VPU] Added ConvertGather7ToGather1 pass to frontend (#7184) 2021-08-24 11:03:23 +03:00
Victor Kuznetsov
b76c903745 add pytest html (#7169) 2021-08-20 16:58:39 +03:00
Mikhail Ryzhov
4714b8edb8 [GNA] Set input scale factors for imported model (#7139) (#7140) 
* [GNA] Set custom scale factor for imported model
 2021-08-19 14:48:37 +03:00