Edward Shogulin [Mon, 8 Jun 2020 09:55:37 +0000 (12:55 +0300)]
[LPT] [Test] DepthToSpace sporadic fail fix (#815)
Alexander Zhogov [Mon, 8 Jun 2020 09:17:50 +0000 (12:17 +0300)]
Azure CI: increase timeout for Mac to 180 min
Vladimir Paramuzov [Sun, 7 Jun 2020 19:57:20 +0000 (22:57 +0300)]
[IE CLDNN] Enable ShuffleChannels op (#787)
Edward Shogulin [Sun, 7 Jun 2020 18:12:52 +0000 (21:12 +0300)]
[LPT] [CPU] DepthToSpace transformation (#663)
* [LPT] [TEST] LayerTransformation generalization
* [LPT] DequantizationDetails extending
* [LPT] DepthToSpace transformation implementation
Alexander Zhogov [Sat, 6 Jun 2020 12:56:24 +0000 (15:56 +0300)]
Azure: Add Ninja (#803)
* Azure: Add Ninja
* Fix 'Install Ninja' on Linux
* Fix bin dir path on Windows
* Add -Wno-unused-variable on Mac
* Add -Wno-error=unused-command-line-argument on Mac
* Set CXXFLAGS for Mac
* Improvements
* Fix BIN_DIR on Linux
Vladimir Paramuzov [Fri, 5 Jun 2020 17:16:47 +0000 (20:16 +0300)]
[IE CLDNN] Enable DepthToSpace (#780)
Enabled DepthToSpace ngraph transformat
Updated implementation to support 5d and mode parameter
fsv16 direct support
Functional tests for GPU
Daria Mityagina [Fri, 5 Jun 2020 16:57:03 +0000 (19:57 +0300)]
it is duplicate of PR: #656, but without test modification (#794)
Chance Luo [Fri, 5 Jun 2020 16:47:53 +0000 (00:47 +0800)]
Disable Hw Avg Pooling for small output tensors if excludePad=true (#772)
Kami-996 [Fri, 5 Jun 2020 13:04:59 +0000 (21:04 +0800)]
add /wd4819 to disable C4819 warning, which is treated as error in win32 (#767)
Co-authored-by: jasonlee <jasonlee@qiyi.com>
emmanuelattia-philips [Fri, 5 Jun 2020 11:46:39 +0000 (13:46 +0200)]
Fix: ITT_INCLUDE_DIR was not correctly detected (#748)
Ilya Znamenskiy [Fri, 5 Jun 2020 11:28:21 +0000 (14:28 +0300)]
[IE CLDNN] GEMM int8 optimization using MMAD macro (#635)
Anton Voronov [Fri, 5 Jun 2020 11:10:55 +0000 (14:10 +0300)]
[CPU] supported ShuffleChannels and added tests (#636)
Ilya Churaev [Fri, 5 Jun 2020 10:36:35 +0000 (13:36 +0300)]
Save the name of output data if we remove previous layer (#760)
* Save the name of output data if we remove previous layer
* Added test
Pavel Rodionov [Fri, 5 Jun 2020 10:00:58 +0000 (13:00 +0300)]
[GNA] Set default GNA library to GNA2 (#771)
dmitrygo [Thu, 4 Jun 2020 12:59:14 +0000 (15:59 +0300)]
[CPU] DepthToSpace review leftovers
Maxim Vafin [Fri, 5 Jun 2020 09:34:57 +0000 (12:34 +0300)]
Add ReduceL2 decomposition (#733)
* Add ReduceL2 decomposition
* Add ReduceL2 transformation tests
* Add const propagation unit test for ReduceL2
Denis Orlov [Fri, 5 Jun 2020 08:54:03 +0000 (11:54 +0300)]
Fix hetero mode in speech sample - set config when loading network (#786)
Ilya Churaev [Fri, 5 Jun 2020 08:22:24 +0000 (11:22 +0300)]
[SAMPLES] Use defined constant instead of string (#788)
Gladilov, Gleb [Fri, 5 Jun 2020 08:16:52 +0000 (11:16 +0300)]
[IE][VPU][nGraph]: Fixes DTS transformations to properly keep outputs names (#734)
* NonZero, Broadcast
* Concat
* Gather
* [IE][VPU][nGraph]: Fixes DTS transformations to correctly keep outputs names
* [IE][VPU][nGraph]: Fixes dynamic to static shape nonzero tests
Co-authored-by: Roman Vyunov <roman.vyunov@intel.com>
Edward Shogulin [Fri, 5 Jun 2020 07:54:38 +0000 (10:54 +0300)]
[LPT] Output layers update fix (#754)
Sergey Shlyapnikov [Fri, 5 Jun 2020 07:20:58 +0000 (10:20 +0300)]
[IE CLDNN] Improve Gather performance and add fusing support (#736)
Egor Churaev [Fri, 5 Jun 2020 07:16:27 +0000 (10:16 +0300)]
[IE CLDNN] Implement NormalizeL2 int8 kernels (#720)
Ilya Churaev [Thu, 4 Jun 2020 18:21:13 +0000 (21:21 +0300)]
Avoid loading of reader if it doesn't exist (#758)
* Avoid loading of reader if it doesn't exist
* Updated error messages
Maxim Vafin [Thu, 4 Jun 2020 17:56:46 +0000 (20:56 +0300)]
Fix Proposal for the case of 2 outputs (#773)
Evgeny Lazarev [Thu, 4 Jun 2020 17:47:52 +0000 (20:47 +0300)]
Added default value for 'aligned' in the ExperimentalDetectronROIFeatureExtractor for backward compatibility (#777)
Fixed backward compatibility issue that old IRs with ExperimentalDetectronROIFeatureExtractor operation cannot be loaded with the new IE
Lukasz Debski [Thu, 4 Jun 2020 14:01:21 +0000 (16:01 +0200)]
[IE CLDNN] Permute fused ops support (#642)
Alexey Varyzgin [Thu, 4 Jun 2020 13:06:15 +0000 (16:06 +0300)]
[CPU][BF16] Default Optimisation Capability of BF16 was enabled on CPX (#647)
Maxim Vafin [Thu, 4 Jun 2020 11:48:31 +0000 (14:48 +0300)]
Add support for ONNX Pad-11 (#744)
Denis Orlov [Thu, 4 Jun 2020 11:33:09 +0000 (14:33 +0300)]
[GNA] Support 100 inputs, instead of 10 (#741)
Gorokhov Dmitriy [Thu, 4 Jun 2020 11:25:19 +0000 (14:25 +0300)]
DepthToSpace, SpaceToDepth layers optimizations (#706)
* [CPU] Updated DepthToSpace and SpaceToDepth layers to be conformant with the specification
The patch also includes n[d]hwc layout support as well as some optimizations
* [CPU][TESTS] Removed old DepthToSpace test since it doesn't corresponds to layer's specification
* [nGraph] Utilize CommonOptimizations pass with custom transformations callback
Evgenya Stepyreva [Thu, 4 Jun 2020 11:18:33 +0000 (14:18 +0300)]
[ MO ] InterpolateConcat empty sources fix (#764)
Vladimir Paramuzov [Thu, 4 Jun 2020 08:14:05 +0000 (11:14 +0300)]
[IE CLDNN] Disabled sporadic detection output tests (#740)
Vladimir Paramuzov [Thu, 4 Jun 2020 07:30:46 +0000 (10:30 +0300)]
[IE CLDNN] Remove unused fused deps for FQ (#712)
Remove unused fused FQ kernel arguments to avoid extra setArg() calls which significantly reduces host overhead
Egor Churaev [Thu, 4 Jun 2020 07:25:28 +0000 (10:25 +0300)]
[IE CLDNN] Implement EmbeddingBag operations (#623)
Implemented three operations: EmbeddingBagPackedSum,
EmbeddingBagOffsetsSum and EmbeddingSegmentsSum. These operations do
the same work but have a different format of inputs.
Anton Voronov [Thu, 4 Jun 2020 05:06:14 +0000 (08:06 +0300)]
[MKLDNN_PLUGIN] Convolution node: skip initializing of primitive descriptors for planar layout if there is already jit primitive (#672)
Ilya Lavrenov [Wed, 3 Jun 2020 19:32:55 +0000 (22:32 +0300)]
Revert "Enabled thread tests (#717)" (#756)
This reverts commit
99a2423ec0dc6d4570ca989be97da02376b3e169.
wistal [Wed, 3 Jun 2020 17:33:55 +0000 (01:33 +0800)]
MO should support LRN k param with caffe model, rather than fixed to 1 (#716)
Co-authored-by: yipengqu <yipeng.qu@intel.com>
Alexey Suhov [Wed, 3 Jun 2020 17:14:35 +0000 (20:14 +0300)]
[Docs] Fixes in readme files: (#750)
- change repo name to openvino
- update driver version
- fix path to samples data
- remove section about Movidius driver installation
- change latest release to 2020.3
- merge fixes in install_dependencies.sh from 2020 branch
Mikołaj Życzyński [Wed, 3 Jun 2020 16:44:27 +0000 (18:44 +0200)]
[IE CLDNN] Added fusing suport to all pooling kernels (#689)
adds fusing support to all available pooling kernels
tests all possible input type/output type configurations
fixes minor bug in max pooling in pooling_gpu_test.cpp
fixed minor bug with yxbf format in pooling_gpu_ref and pooling_gpu_int8_ref kernels
fixes bug with b_fs_yx_fsv32 format in pooling_gpu kernel
resolves bug with max pooling accuracy missmatch in case of non zero pad end layer parameter
resolves average pooling accuracy missmatch in case of non zero pad end layer parameter
Lukasz Debski [Wed, 3 Jun 2020 16:42:50 +0000 (18:42 +0200)]
[IE CLDNN] Fixed stack overflow in calculate_prior_boxes pass (#747)
The problem behind this error was in program_impl::init_graph() where in calculate_prior_boxes we are trying to calculate output layout of an entire network recursively which causes stack overflow. Calculating output layouts beforehand in processing order fixes this issue.
Ilya-Krylov [Wed, 3 Jun 2020 14:52:40 +0000 (17:52 +0300)]
Add 'aligned' param to ExperimentalDetectronROIFeatureExtractor for CPU plugin and MO
Tomasz Dołbniak [Wed, 3 Jun 2020 13:01:43 +0000 (15:01 +0200)]
Python API for Assign, ReadValue and ExtractImagePatches (#719)
LiweiSong [Wed, 3 Jun 2020 12:19:29 +0000 (20:19 +0800)]
mkldnn_memory_solver.hpp: include stdint.h to avoid build error (#729)
fix the following compile error:
inference-engine/src/mkldnn_plugin/mkldnn_memory_solver.hpp:60:9: error: 'int64_t' does not name a type
| 60 | int64_t size;
| | ^~~~~~~
include stdint.h to fix this.
Signed-off-by: Liwei Song <liwei.song@windriver.com>
Edward Shogulin [Wed, 3 Jun 2020 12:05:45 +0000 (15:05 +0300)]
[LPT] [TEST] Sporadic test fail fix (workaround) (#742)
Evgenya Stepyreva [Wed, 3 Jun 2020 11:14:59 +0000 (14:14 +0300)]
[ nGraph ] FP16 for evaluate (#722)
Vitaliy Urusovskij [Wed, 3 Jun 2020 10:54:38 +0000 (13:54 +0300)]
Use pre-defined DB collection names in memcheck_upload.py CLI (#651)
Use argparses `choices` for `--db_collection` option.
Also removed unnecessary redefinition of `db_collection` in memcheck_upload.py
Sergey Shlyapnikov [Wed, 3 Jun 2020 10:42:15 +0000 (13:42 +0300)]
[IE CLDNN] Improve kernel selection for b_fs_yx_fsv16 layout and optimize Convolution kernels (#730)
Anton Zaytsev [Wed, 3 Jun 2020 09:16:00 +0000 (12:16 +0300)]
[IE TESTS] Move InferRequestTests (#618)
* [IE TESTS] move Infer_request tests
* fix v0
* [ci-skip][IE TESTS] test update basic class v0
* [ci-skip][IE TESTS] test update basic class v1
* [ci-skip][IE TESTS] test update basic class
* [ci-skip][IE TESTS] test update basic class v3
* [ci-skip][IE TESTS] test update basic class final versions
* [ci-skip][IE TESTS] fix
* [ci-skip][IE TESTS] fix codestaly and comment
Co-authored-by: Irina Efode <irina.efode@intel.com>
Evgeny Talanin [Wed, 3 Jun 2020 09:15:06 +0000 (12:15 +0300)]
Skip some functional tests on VPU (#568)
Adam Osewski [Wed, 3 Jun 2020 09:01:56 +0000 (11:01 +0200)]
[ONNX] GRU and RNN operators. (#607)
* Create generic RecurrentSequenceDirection enum.
* Helper class RecurrentSequenceOp.
* Add ONNX GRU & RNN operators.
* Use OutputVector.
* Update doc.
* Add UTs for GRU and skip them on IE_CPU
* Add UT for bidirectional mode and fix it.
* Normalize activation function name case.
* Add unit-tests for RNN operator.
* UT for GRU with linear_before_reset set to true.
* Fix ONNX GRU for linear_before_reset case.
* Remove unnecessary symbol export macro.
* Fix CentOS error.
* Update UTs.
- Update few tests accuracy tolerance
- Update rnn_fwd_activations with new reference values and model.
* Review comment: add check for static shape
* Add UT for RNN with constant inputs W, R.
* Skip UT with const W,R on IE_CPU
Gladilov, Gleb [Wed, 3 Jun 2020 08:43:19 +0000 (11:43 +0300)]
[IE][VPU]: Faster-RCNN fixes on myriad plugin side (#711)
* [IE][VPU]: Enables pass for propagating dynamism to network outputs
If network had dynamic output and then myriad Front-End inserted
convert stage at the end (to convert FP16 -> FP32 - output precision)
then dynamism would not be propagated - we have convert stage that
has dynamic input, but static output. As a result, we have run-time
error in Convert kernel: input and output shapes do not match.
At the moment, pass supports only Convert stage as output stage
over which we should propagate dynamism to outputs.
Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
* [IE][VPU]: Fixes parse DSR in case of output data
Replacing stage output must be done after replacing
data to shape parent, because the last one may access
original parent producer, but after replacing stage output
it'd not have one.
Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
* [IE][VPU]: Fixes MacOS build
* [IE][VPU]: Fixes shape data naming convention
Plugin part assumes that if there is dynamic data object, that's
represented as 2 different data objects (data and shape), then
shape data object has name = data object name + @shape suffix.
Pass that creates new dynamic data object should respect that
assumption.
* [IE][VPU]: Fixes dis-alignment in names of data objects representing dynamic data object
MyriadInferRequest::GetResult assumes that in case of dynamic data object
"data" data object and "shape" data object will have aligned names:
"shape" name = "data" name + "@shape" suffix.
In order to meet that expectation propagating dynamism pass must use output
data object name as prefix. Additionally, propagating pass must be applied
before converting shape notation pass in order to make output shape in IE
notation, not MDK, as MyriadInferRequest::GetResult is expecting.
Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
Mikhail Treskin [Wed, 3 Jun 2020 08:04:15 +0000 (11:04 +0300)]
Remove deprecated layer test class (#610)
* Update activation layer test
Signed-off-by: Mikhail Treskin <mikhail.treskin@intel.com>
* Get rid of LayerTestsCommonDeprecated class
Signed-off-by: Mikhail Treskin <mikhail.treskin@intel.com>
* Fix activation tests instantiations for gpu and myriad plugins
* Remove leaking inferWithInterp function
Mikołaj Życzyński [Wed, 3 Jun 2020 06:30:49 +0000 (08:30 +0200)]
[IE CLDNN] Activation with fused quantize bug fix (#613)
fixed bug connected with quantization fusing to activation
added scale and activation fusing support
added corresponding tests
Ilya Lavrenov [Wed, 3 Jun 2020 03:04:50 +0000 (06:04 +0300)]
Updated deprecated messages (#715)
Edward Shogulin [Tue, 2 Jun 2020 20:44:24 +0000 (23:44 +0300)]
[LPT] Output layers fix (#677)
Ilya Churaev [Tue, 2 Jun 2020 20:42:05 +0000 (23:42 +0300)]
Enabled thread tests (#717)
Nikolay Shchegolev [Tue, 2 Jun 2020 18:56:17 +0000 (21:56 +0300)]
[CPU] EmbeddingBagOffsetsSum, EmbeddingBagPackedSum, EmbeddingSegmentsSum operations. (#576)
* [CPU] EmbeddingBagOffsetsSum, EmbeddingBagPackedSum, EmbeddingSegmentsSum operations.
* Performance fix
* Perf v2
* Code style
Anna Alberska [Tue, 2 Jun 2020 18:03:27 +0000 (20:03 +0200)]
add PassManagerSettings & create more legible description for concat quantization exception and a test for it (#563)
Andrey Babushkin [Tue, 2 Jun 2020 17:22:25 +0000 (20:22 +0300)]
[Jenkinsfile] Add failFast parameter (#721)
It allows us to rebuild Jenkins build and wait until all stages are finished despite of some of them may fail
iimironov [Tue, 2 Jun 2020 16:20:29 +0000 (19:20 +0300)]
Imironov/cvs 31297 add yolov4 support (#594)
* Add transformation of softplus operation into log(1.0 + exp(x)).
Evgeny Latkin [Tue, 2 Jun 2020 14:41:19 +0000 (17:41 +0300)]
[IE][Myriad] Gather: add test case (#644)
Gleb Kazantaev [Tue, 2 Jun 2020 14:26:28 +0000 (17:26 +0300)]
Updated DeconvolutionIE to support dynamic shapes (#671)
* Updated DeconvolutionIE to support dynamic shapes
* Updated DeconvolutionIE to support output_shape input
* Updated ConvertConvolutions pass
Vladimir Paramuzov [Tue, 2 Jun 2020 14:25:41 +0000 (17:25 +0300)]
[IE CLDNN] Fixed names mapping chain in runtime graph to respect original names (#599)
Vitaliy Urusovskij [Tue, 2 Jun 2020 14:24:22 +0000 (17:24 +0300)]
Add parallel downloads to stress tests (#678)
Gleb Kazantaev [Tue, 2 Jun 2020 14:15:36 +0000 (17:15 +0300)]
specificCreator for Transpose operation (#713)
* Updated Transpose node convertor; replaced get_vector with cast_vector
* Replaced NodeCreator with specificCreator
Andrey Somsikov [Tue, 2 Jun 2020 13:46:23 +0000 (16:46 +0300)]
Enable Control Flow Guard for Windows binaries (#714)
Control Flow Guard is security option.
azhogov [Tue, 2 Jun 2020 09:52:14 +0000 (12:52 +0300)]
Revert "Add ittnotify from IntelSEAPI"
This reverts commit
0583b37a14f7f5e4514d51f4a90818e5d42b57ba.
azhogov [Tue, 2 Jun 2020 09:50:06 +0000 (12:50 +0300)]
Revert "Use ittnotify from thirdparty"
This reverts commit
3863656f446bdaa1fa7664127e315fcd97d0ef7e.
Anton Chetverikov [Tue, 2 Jun 2020 09:43:41 +0000 (12:43 +0300)]
Remove TopKnormalizer from MO IR Reader transformation_list (#590)
* Remove TopKnormalizer from transformation_list and added call of normalize_outputs to fix read/save of some models
Maksim Doronin [Tue, 2 Jun 2020 09:20:06 +0000 (12:20 +0300)]
[IE VPU] Enable s32->u8 conversion (#699)
Evgeny Lazarev [Tue, 2 Jun 2020 08:32:39 +0000 (11:32 +0300)]
Align MO requirements files (#710)
Vladimir Paramuzov [Tue, 2 Jun 2020 06:17:39 +0000 (09:17 +0300)]
[IE CLDNN] Quantized deeplabv3 optimizations (#646)
Enabled dilation for imad dw fsv16 kernel
Added argmax and mutable_data to fsv16 white list
Enabled byxf input for quantize scale_shift kernel
Somsikov, Andrey [Mon, 1 Jun 2020 07:26:34 +0000 (10:26 +0300)]
Use ittnotify from thirdparty
VTune ittnotify lack of support aarch64. Switching to use ittnotify
in sources to support any target architecture.
Somsikov, Andrey [Fri, 29 May 2020 16:55:19 +0000 (19:55 +0300)]
Add ittnotify from IntelSEAPI
Adding ittnotify component of https://github.com/intel/IntelSEAPI
commit
88a56e0ecd162667c7afd2ee9969221d62a32509 (HEAD -> master, origin/master, origin/HEAD)
Merge: 6d743e1 809062a
Author: Alex <alexander.a.raud@intel.com>
Date: Wed Jul 10 15:06:46 2019 -0700
Andrew Bakalin [Mon, 1 Jun 2020 15:57:06 +0000 (18:57 +0300)]
[VPU][NGraph] Reuse NonZero evaluate in StaticShapeNonZero (#658)
* [VPU][NGraph] Reuse NonZero evaluate in StaticShapeNonZero
* [VPU][Tests] Adopt old tests to work with reverted indices
* [VPU] Update firmware
Katya [Mon, 1 Jun 2020 12:58:05 +0000 (15:58 +0300)]
[IE Python API] fix TensorDesc test file name (#701)
Maksim Doronin [Mon, 1 Jun 2020 11:51:04 +0000 (14:51 +0300)]
[IE VPU] OutShapeOfReshape per-layer tests (#631)
* [IE VPU] OutShapeOfReshape per-layer tests
* [IE VPU] Update firmware
* [IE VPU] OutShapeOfReshape: get rid of code duplication
Piotr Rozen [Mon, 1 Jun 2020 11:41:45 +0000 (13:41 +0200)]
Added speech recognition demo package for centOS (#682)
Vladislav Vinogradov [Mon, 1 Jun 2020 10:09:23 +0000 (13:09 +0300)]
[IE] Fix build error (#703)
Missing changes in transformation library due to IE API dependency removal.
Mikhail Letavin [Mon, 1 Jun 2020 09:01:28 +0000 (12:01 +0300)]
[IE CLDNN] Free up first copy of weights/biases that were transferred to USM device memory (#561)
Roman Kazantsev [Mon, 1 Jun 2020 08:48:06 +0000 (13:48 +0500)]
Fix SparseWeightedSum transform for Wide and Deep (#698)
WhereDecomposition transform is applied to Where operation in for-garbage sub-graph remained after SparseWeightedSum transform.
Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
Jedrzej Hajduczenia [Mon, 1 Jun 2020 07:38:33 +0000 (09:38 +0200)]
[IE CLDNN] Prefer bf(wz)yx format for reshape (#691)
Performance improvement for icnet-camvid model
Ilya Churaev [Mon, 1 Jun 2020 07:31:31 +0000 (10:31 +0300)]
Removed dependency on Inference Engine from transformation library (#680)
* Removed dependency on Inference Engine from transformation library
* Change transformations export macro
* Fixed comments
Ivan Tikhonov [Mon, 1 Jun 2020 06:24:16 +0000 (09:24 +0300)]
Reshape-Permute-Reshape pattern to DepthToSpace layer transformation (#601)
* implemented depth_to_space transformation
* renaming
* added functional tests, fixed mistakes in implementation of the transformation
* disable ConvertSpaceToDepth/ConvertDepthToSpace transformation for CPU plugin, enable DepthToSpaceFusion for CPU plugin only, add specific creators
* fix wrong include
* fix for functional tests: set transformation callback
* revert callback calls for CPU plugin
* move functions to .cpp file
* Apply review comments
* Apply additional review comments
* fix cast to bool type
Ewa Tusień [Mon, 1 Jun 2020 02:59:39 +0000 (04:59 +0200)]
[ONNX] Add Range op to ONNX importer (#548)
* Added Range op to ONNX importer.
* Disable tests for IE.
emmanuelattia-philips [Sun, 31 May 2020 20:19:37 +0000 (22:19 +0200)]
Ie capi callback with explicit calling convention (#697)
* Added explicit calling convention to CAPI callback
* Fixed typo spacing
* Renamed INFERENCE_ENGINE_CALLBACK to INFERENCE_ENGINE_C_API_CALLBAC to make the macro really specific to the C API
Gladilov, Gleb [Sun, 31 May 2020 10:17:36 +0000 (13:17 +0300)]
[IE][VPU]: Faster-RCNN fixes on myriad plugin side (#665)
* [IE][VPU]: Fixes deallocation data for cases of CMX allocator run
The final loop tries to deallocate data objects that keep shape values for
other data objects that're outputs of a model. But the case when allocator
takes only CMX data into consideration was not handled and since allocation
could not happen, it lead to fail on deallocation of a data object that has
not been allocated.
Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
* [IE][VPU]: Fixes allocator with work on data to shape edges
Since there is new relationship between data objects: some
data objects may contain shape of other data object - allocator
must properly respect that. The thing is if 2 data objects are
connected in such a way, they represent unite entity (dynamic
data object) and should have the same lifetime.
Signed-off-by: Gladilov, Gleb <gleb.gladilov@intel.com>
emmanuelattia-philips [Sat, 30 May 2020 23:25:39 +0000 (01:25 +0200)]
Added ie_core_read_network_from_memory to the C API ie_bridge. (#674)
* * Added ie_core_read_network_from_memory to the C ie_bridge.
* Added size argument for xml_content, fixed const correctness of the weight_blob, fixed unit test
* * Removed debug message
* Changed variables names from model_xxx to weights_xxx to be more consistent with the argument name of the tested function.
* Added a description for xml_content_size in ie_core_read_network_from_memory.
* * xml_content is now passed as uint8_t
* reading function factorized in the unit-test
Ilya Lavrenov [Sat, 30 May 2020 12:15:20 +0000 (15:15 +0300)]
Added dependency on ONNX reader (#693)
Denis Orlov [Fri, 29 May 2020 21:43:42 +0000 (00:43 +0300)]
[GNA] Workaround support for callbacks (#591)
Edward Shogulin [Fri, 29 May 2020 19:56:58 +0000 (22:56 +0300)]
[LPT] Convolution regression tests (#543)
* [LPT] Base test infrastructure extending & Convolution test
* [LPT] LPT test infrastructure refactoring
Anastasia Kuporosova [Fri, 29 May 2020 18:28:17 +0000 (21:28 +0300)]
[IE TOOLS] Use input_info in python benchmark app (#660)
Anastasia Kuporosova [Fri, 29 May 2020 18:20:16 +0000 (21:20 +0300)]
[Python API] Fix PreProcessInfo tests (#690)
Vladimir Gavrilov [Fri, 29 May 2020 18:01:09 +0000 (21:01 +0300)]
MO fails generating IR from XLNET model due to a bug in the transformation ConvertGroupedStridedSlice (#625)
* Small fix in the transformation ConvertGroupedStridedSlice. Now VariadicSplit is generated only in the case when node has at least 2 output nodes.
* Added unittests for the case when there is only one StridedSlice.
Ilya Churaev [Fri, 29 May 2020 17:57:32 +0000 (20:57 +0300)]
Fixed CODEOWNERS paths (#684)
Vladimir Paramuzov [Fri, 29 May 2020 17:10:30 +0000 (20:10 +0300)]
[IE CLDNN] Optimized FQ kernel in fsv16 layout (#573)
- Optimized FQ kernel in fsv16 layout. Enabled scaleshift transform for FP16 precision
- Disabled activation_opt kernel with fused ops in some cases
Evgenya Stepyreva [Fri, 29 May 2020 16:09:01 +0000 (19:09 +0300)]
[ nG ] Added F16 folding support (#686)
Ilya Churaev [Fri, 29 May 2020 14:46:40 +0000 (17:46 +0300)]
Do not use ONNX reader if ONNX importer was disabled (#683)
Anna Khakimova [Fri, 29 May 2020 12:44:12 +0000 (15:44 +0300)]
Pre-processing(GAPI): AVX2/AVX512 implementation of 3C/4C Resize via universal intrinsics. (#612)
Ilya Lavrenov [Fri, 29 May 2020 11:45:59 +0000 (14:45 +0300)]
Python: Fixed installation rules to install additional .so files generated from .pyx (#676)