platform/core/ml/nnfw.git
4 years ago[luci-interpreter] Add support for sqrt op (#3904)
Alexander Moiseev [Mon, 24 Aug 2020 12:42:59 +0000 (15:42 +0300)]
[luci-interpreter] Add support for sqrt op (#3904)

Add support for sqrt element-wise op in luci-interpreter.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[onert/test] Operation backend test on one_op_tests (#3939)
Hyeongseok Oh [Mon, 24 Aug 2020 11:30:05 +0000 (20:30 +0900)]
[onert/test] Operation backend test on one_op_tests (#3939)

- Introduce backend setting method in GenModelTest
- Backend test on one_op_tests

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[onert] Unify element-wise unary operations to ElementwiseUnary. (#3933)
Jang Jiseob [Mon, 24 Aug 2020 07:49:21 +0000 (16:49 +0900)]
[onert] Unify element-wise unary operations to ElementwiseUnary. (#3933)

This commit unifies element-wise activation operations to ElementwiseActivation for reducing code duplication.

Signed-off-by: ragmani <ragmani0216@gmail.com>
4 years ago[onert] Copyright fix (#3945)
Hyeongseok Oh [Mon, 24 Aug 2020 07:38:44 +0000 (16:38 +0900)]
[onert] Copyright fix (#3945)

- Add copyright to auto generated files
- Fix copyright checker to ignore NNAPI file
- Fix copyright checker to allow (year-year)
- Add copyright to ported code with small fix

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[tools] Introduce github release tool (#3900)
seongwoo chae [Mon, 24 Aug 2020 07:37:45 +0000 (16:37 +0900)]
[tools] Introduce github release tool (#3900)

This commit introduces github release tool that helps you to automate the release.

ONE-DCO-1.0-Signed-off-by: seongwoo <mhs4670go@naver.com>

4 years ago[luci] SummaryBuilder reduce LOC 2 (#3946)
SaeHie Park [Mon, 24 Aug 2020 07:32:43 +0000 (16:32 +0900)]
[luci] SummaryBuilder reduce LOC 2 (#3946)

This will relocate implementation of summary builder to reduce class LOC

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[luci] Shape inf reduce LOC S- (#3944)
SaeHie Park [Mon, 24 Aug 2020 06:36:24 +0000 (15:36 +0900)]
[luci] Shape inf reduce LOC S- (#3944)

This will relocate implementation of shape inference to reduce class LOC

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[luci] SummaryBuilder reduce LOC 1 (#3942)
SaeHie Park [Mon, 24 Aug 2020 06:11:36 +0000 (15:11 +0900)]
[luci] SummaryBuilder reduce LOC 1 (#3942)

This will relocate implementation of summary builder to reduce class LOC

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[luci] Shape inf reduce LOC M-R (#3941)
SaeHie Park [Mon, 24 Aug 2020 06:06:27 +0000 (15:06 +0900)]
[luci] Shape inf reduce LOC M-R (#3941)

This will relocate implementation of shape inference to reduce class LOC

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[onert] Fix conv2d op for dilation (#3932)
YiHyunJin [Mon, 24 Aug 2020 06:04:33 +0000 (15:04 +0900)]
[onert] Fix conv2d op for dilation (#3932)

When using MultiThread, both dilation data must be 1.
When obtaining padding data, dilation data is used.
Move additional conditions of dilation into usableMultiThreaded

Signed-off-by: YiHyunjin <hj0412.yi@samsung.com>
4 years ago[nnapi] Unify comparison ops in OperationFactory (#3875)
Hanjoung Lee [Mon, 24 Aug 2020 05:58:32 +0000 (14:58 +0900)]
[nnapi] Unify comparison ops in OperationFactory (#3875)

Unify comparison ops in OperationFactory

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[luci] Shape inf reduce for virtual outputs (#3938)
SaeHie Park [Mon, 24 Aug 2020 04:48:30 +0000 (13:48 +0900)]
[luci] Shape inf reduce for virtual outputs (#3938)

This will relocate implementation of shape inference to reduce class LOC

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[onert] Unify element-wise activation operations to ElementwiseActivation for reducin...
Jang Jiseob [Mon, 24 Aug 2020 04:37:46 +0000 (13:37 +0900)]
[onert] Unify element-wise activation operations to ElementwiseActivation for reducing code duplication (#3901)

This commit unifies element-wise activation operations to ElementwiseActivation for reducing code duplication.

Signed-off-by: ragmani <ragmani0216@gmail.com>
4 years ago[tflitefile_tool] select_operator supports Select op (#3915)
Sanggyu Lee [Mon, 24 Aug 2020 02:44:15 +0000 (11:44 +0900)]
[tflitefile_tool] select_operator supports Select op (#3915)

select_operator will support Select op.

ONE-DCO-1.0-Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>

4 years ago[onert] Remove ITensorBuilder::supportDynamicTensor (#3929)
Hanjoung Lee [Mon, 24 Aug 2020 02:40:52 +0000 (11:40 +0900)]
[onert] Remove ITensorBuilder::supportDynamicTensor (#3929)

Remove `ITensorBuilder::supportDynamicTensor`.
This is as adding dynamic alloc info for all has no harm, and eventually
adding dynamic alloc info will be removed.

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[onert] Move the time of choosing kernel to configure (#3924)
Jang Jiseob [Mon, 24 Aug 2020 02:27:55 +0000 (11:27 +0900)]
[onert] Move the time of choosing kernel to configure (#3924)

This commit moves the time of choosing kernel to configure in some cpu layers
  - cpu layers : BinaryArithmeticLayer, ElementwiseBinaryLayer, PoolLayer, ReduceLayer

Signed-off-by: ragmani <ragmani0216@gmail.com>
4 years ago[onert] Copy and package armcompute only in overlay (#3885)
Hyeongseok Oh [Mon, 24 Aug 2020 02:26:09 +0000 (11:26 +0900)]
[onert] Copy and package armcompute only  in overlay (#3885)

Copy to installed path and pacakge arm compute library only from overlay (ignore boost build)

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[luci] Shape inf reduce LOC D-M (#3937)
SaeHie Park [Mon, 24 Aug 2020 01:56:19 +0000 (10:56 +0900)]
[luci] Shape inf reduce LOC D-M (#3937)

This will relocate implementation of shape inference to reduce class LOC

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[onert/test] Update nnpackage test list (#3928)
Hyeongseok Oh [Mon, 24 Aug 2020 01:55:42 +0000 (10:55 +0900)]
[onert/test] Update nnpackage test list (#3928)

Update nnpackage test list file

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[nnfwapi] Improve GenTest (#3931)
Hanjoung Lee [Mon, 24 Aug 2020 01:52:02 +0000 (10:52 +0900)]
[nnfwapi] Improve GenTest (#3931)

- Aggregate variables into `_test_data` that test makers must generate
- Support multiple test cases

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[luci] Relocate shape inf common functions (#3936)
SaeHie Park [Mon, 24 Aug 2020 01:15:50 +0000 (10:15 +0900)]
[luci] Relocate shape inf common functions (#3936)

This will relocate shape inference common functions to upper side of the file

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[luci] Import validate helpers to reduce dup codes (#3920)
SaeHie Park [Mon, 24 Aug 2020 01:06:37 +0000 (10:06 +0900)]
[luci] Import validate helpers to reduce dup codes (#3920)

This will introduce ValidateHelpers to reduce duplicate codes

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[luci] Shape inf reduce LOC A-B (#3935)
SaeHie Park [Mon, 24 Aug 2020 00:35:30 +0000 (09:35 +0900)]
[luci] Shape inf reduce LOC A-B (#3935)

This will relocate implementation of shape inference to reduce class LOC

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[circledump] No opt comment for NonMaxSupressionV5 (#3796)
Alexander Moiseev [Sun, 23 Aug 2020 21:21:22 +0000 (00:21 +0300)]
[circledump] No opt comment for NonMaxSupressionV5 (#3796)

Add comment about no option in NonMaxSuppressionV5.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[luci/tests] Add Unique operation (#3919)
dr-venkman [Fri, 21 Aug 2020 13:44:41 +0000 (22:44 +0900)]
[luci/tests]  Add Unique operation (#3919)

Parent Issue : #2568
Draft PR : #2829

This commit will add unique operation to luci tests

ONE-DCO-1.0-Signed-off-by: venkat.iyer <venkat.iyer@samsung.com>

4 years ago[tflitefile_tool] support dilation in select_operator.py (#3923)
Sanggyu Lee [Fri, 21 Aug 2020 11:03:42 +0000 (20:03 +0900)]
[tflitefile_tool] support dilation in select_operator.py (#3923)

Currently select_operator ignores dilation factors.
Now, it will handle them.

ONE-DCO-1.0-Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>

4 years ago[onert/test] Split GenModelTests.cc (#3911)
Hyeongseok Oh [Fri, 21 Aug 2020 07:29:40 +0000 (16:29 +0900)]
[onert/test] Split GenModelTests.cc (#3911)

* [onert/test] Split GenModelTests.cc

- Header file for GenModelTest class
- Divide test for each operation

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Rename directory and header

OneOp test directory: one_op_tests
Header file name: GenModelTest.h

4 years ago[onert] Remove DynamicTensorMan from DynAllocInfo (#3903)
Hanjoung Lee [Fri, 21 Aug 2020 07:07:34 +0000 (16:07 +0900)]
[onert] Remove DynamicTensorMan from DynAllocInfo (#3903)

Now that there is this method `ITensor::dynamic_tensor_manager`, no need
to have `DynamicTensorManager` in `DynAllocInfo`.

% Eventually `DynAllocInfo` will be removed.

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[common-artifacts] Remove redundant space (#3927)
seongwoo chae [Fri, 21 Aug 2020 05:54:04 +0000 (14:54 +0900)]
[common-artifacts] Remove redundant space (#3927)

This commit removes redundant space in cmake file.

ONE-DCO-1.0-Signed-off-by: seongwoo <mhs4670go@naver.com>

4 years ago[luci-interpreter]Update to use Tensorflow 2.3.0 on luci-interpreter. (#3857)
struss [Fri, 21 Aug 2020 05:52:22 +0000 (14:52 +0900)]
[luci-interpreter]Update to use Tensorflow 2.3.0 on luci-interpreter. (#3857)

* [luci-interpreter]Update to use Tensorflow 2.3.0 on luci-interpreter.

This commit update to use Tensorflow 2.3.0 on luci-interpreter.

  1. Update cmake file in luci-interpreter kernel to include some file in Tensorflow(Mean kernel use this.)
  2. Update kernels for changed spec from 2.1.0 to 2.3.0.(`LeakyRelu`, `Mul`)
`LeakyRelu` : Arguments are changed.
`Mul` : kernel changed to reference to optimized.
  3. Update Kernel that accept more tensor.(handle 3~4 Tensor both).`TransposeConv`
currently, accept nullptr on 4th Tensor. it will update later.

ONE-DCO-1.0-Signed-off-by: KiDeuk Bang <rrstrous@nate.com>

* Update Testcases on TransposeConv.

ONE-DCO-1.0-Signed-off-by: KiDeuk Bang <rrstrous@nate.com>

4 years ago[locomotiv] Reduce class codes D (#3917)
SaeHie Park [Fri, 21 Aug 2020 04:54:43 +0000 (13:54 +0900)]
[locomotiv] Reduce class codes D (#3917)

This will rename methods to reduce class lines of codes of NodeExecution

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[tflitefile_tool] select_operator will show unsupported op info (#3912)
Sanggyu Lee [Fri, 21 Aug 2020 03:42:17 +0000 (12:42 +0900)]
[tflitefile_tool] select_operator will show unsupported op info (#3912)

Currently select_operator does not show any information about
what operator is not supported.
It will print BuiltinOptions index to find out easily what operator is
not supported.

ONE-DCO-1.0-Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>

4 years ago[nnpkg_run] use same format for dumping boolean (#3910)
Sanggyu Lee [Fri, 21 Aug 2020 03:42:02 +0000 (12:42 +0900)]
[nnpkg_run] use same format for dumping boolean (#3910)

nnpackage_run has stored bool as I8.
It should be same format (=U8) of python and nnpackage_run load.

ONE-DCO-1.0-Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>

4 years ago[onert] Enable L2Norm and PadV2 operation loading (#3887)
Hyeongseok Oh [Fri, 21 Aug 2020 03:36:36 +0000 (12:36 +0900)]
[onert] Enable L2Norm and PadV2 operation loading (#3887)

- Enable L2Normalization and PadV2 operation in base_loader.h
- Add model generation test

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[luci] Reduce duplicate codes in Conv2D shape-inf (#3914)
SaeHie Park [Fri, 21 Aug 2020 03:31:09 +0000 (12:31 +0900)]
[luci] Reduce duplicate codes in Conv2D shape-inf (#3914)

This will reduce duplicate codes in Conv2D type shape inference

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[locomotiv] Reduce class codes T (#3909)
SaeHie Park [Fri, 21 Aug 2020 02:33:51 +0000 (11:33 +0900)]
[locomotiv] Reduce class codes T (#3909)

This will rename methods to reduce class lines of codes of NodeExecution

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[locomotiv] Reduce class codes M-S (#3908)
SaeHie Park [Fri, 21 Aug 2020 02:15:28 +0000 (11:15 +0900)]
[locomotiv] Reduce class codes M-S (#3908)

This will rename methods to reduce class lines of codes of NodeExecution

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[locomotiv] Reduce class codes F-M (#3907)
SaeHie Park [Fri, 21 Aug 2020 02:09:03 +0000 (11:09 +0900)]
[locomotiv] Reduce class codes F-M (#3907)

This will rename methods to reduce class lines of codes of NodeExecution

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[locomotiv] Reduce class codes B-D (#3906)
SaeHie Park [Fri, 21 Aug 2020 02:08:38 +0000 (11:08 +0900)]
[locomotiv] Reduce class codes B-D (#3906)

This will rename methods to reduce class lines of codes of NodeExecution

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[circle-quantizer] verify circle file (#3894)
seongwoo chae [Fri, 21 Aug 2020 01:49:36 +0000 (10:49 +0900)]
[circle-quantizer] verify circle file (#3894)

This commit enables circle-quantizer to verify circle file.

ONE-DCO-1.0-Signed-off-by: seongwoo <mhs4670go@naver.com>

4 years ago[record-minmax] verify circle file (#3893)
seongwoo chae [Thu, 20 Aug 2020 23:34:34 +0000 (08:34 +0900)]
[record-minmax] verify circle file (#3893)

This commit enables record-minmax to verify circle file.

ONE-DCO-1.0-Signed-off-by: seongwoo <mhs4670go@naver.com>

4 years ago[luci/tests] Add NonMaxSuppressionV5 luci tests (#3804)
Alexander Moiseev [Thu, 20 Aug 2020 21:20:58 +0000 (00:20 +0300)]
[luci/tests] Add NonMaxSuppressionV5 luci tests (#3804)

Add testcases for NonMaxSuppressionV5 op in luci.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

Co-authored-by: Alexander Efimov <efimov.alexander@gmail.com>
4 years ago[onert] Revise OpSequence dump method (#3890)
Hanjoung Lee [Thu, 20 Aug 2020 11:13:32 +0000 (20:13 +0900)]
[onert] Revise OpSequence dump method (#3890)

- Make the method to be outside of the class
- Remove message parameter

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[onert] Use TensorRegistry directly in controlflow (#3899)
Hanjoung Lee [Thu, 20 Aug 2020 09:25:43 +0000 (18:25 +0900)]
[onert] Use TensorRegistry directly in controlflow (#3899)

Use TensorRegistry directly in controlflow KernelGenerator.

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[res] Add NonMaxSuppressionV5 recipes (#3803)
Alexander Moiseev [Thu, 20 Aug 2020 09:01:22 +0000 (12:01 +0300)]
[res] Add NonMaxSuppressionV5 recipes (#3803)

Add recipes for NonMaxSuppressionV5 op.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[luci] Enable PadV2 operator in tests (#3675)
a-zh [Thu, 20 Aug 2020 08:47:57 +0000 (11:47 +0300)]
[luci] Enable PadV2 operator in tests (#3675)

This enables tests for PadV2 operator

Parent Issue: #3362
Draft PR: #3566

ONE-DCO-1.0-Signed-off-by: Anton Zhukov <a.zhukov@samsung.com>

4 years ago[tflite2circle] verify tflite file (#3892)
seongwoo chae [Thu, 20 Aug 2020 08:40:42 +0000 (17:40 +0900)]
[tflite2circle] verify tflite file (#3892)

This commit enables tflite2circle to verify tflite file.

ONE-DCO-1.0-Signed-off-by: seongwoo <mhs4670go@naver.com>

4 years ago[onert] cpu KernelGen use _tensor_reg directly (#3895)
Hanjoung Lee [Thu, 20 Aug 2020 08:10:41 +0000 (17:10 +0900)]
[onert] cpu KernelGen use _tensor_reg directly (#3895)

cpu KernelGen use _tensor_reg directly

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[onert] Remove unused Mean IR (#3898)
Hyeongseok Oh [Thu, 20 Aug 2020 08:05:16 +0000 (17:05 +0900)]
[onert] Remove unused Mean IR (#3898)

Mean IR is merged into Reduce IR

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[doc] Write compute.md (#3879)
Hanjoung Lee [Thu, 20 Aug 2020 07:05:57 +0000 (16:05 +0900)]
[doc] Write compute.md (#3879)

This commit fills `compute.md` which is about compute directory.

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[onert/test] Fix test command warning (#3888)
Hyeongseok Oh [Thu, 20 Aug 2020 07:02:34 +0000 (16:02 +0900)]
[onert/test] Fix test command warning (#3888)

Fix test command warning: unary operator expected

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[luci/service] Correct Unique operation (#3876)
dr-venkman [Thu, 20 Aug 2020 06:04:33 +0000 (15:04 +0900)]
[luci/service] Correct Unique operation (#3876)

Parent Issue : #2568
Draft PR : #2829

This commit corrects shapeinferencing for unique operation

ONE-DCO-1.0-Signed-off-by: venkat.iyer <venkat.iyer@samsung.com>

4 years ago[onert] Remove `ITensorBuilder::tensorAt` (#3880)
Hanjoung Lee [Thu, 20 Aug 2020 05:48:15 +0000 (14:48 +0900)]
[onert] Remove `ITensorBuilder::tensorAt` (#3880)

Remove `ITensorBuilder::tensorAt` to distinguish the roles of
`ITensorBuilder` and `ITensorRegistry`. A builder is a builder and a
registry contains built tensors.

Note that KernelGenerator and ConstantInitializer now use TensorRegistry
to find tensor objects which takes most part of the changes.

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[onert] Simplify CF KernelGenerator::getTensor (#3869)
Hanjoung Lee [Thu, 20 Aug 2020 03:41:12 +0000 (12:41 +0900)]
[onert] Simplify CF KernelGenerator::getTensor (#3869)

Simplify `backend::controlflow::KernelGenerator::getTensor`.
`TensorBuilders` has the exact function already.

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[infra] Update eigen download url (#3883)
Hyeongseok Oh [Thu, 20 Aug 2020 02:59:01 +0000 (11:59 +0900)]
[infra] Update eigen download url (#3883)

bitbucket no more support source code download service

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[onert] Support dilation for Conv2d (#3845)
YiHyunJin [Wed, 19 Aug 2020 09:26:14 +0000 (18:26 +0900)]
[onert] Support dilation for Conv2d (#3845)

It applied dilation data in conv2d And Added test case
for conv2d with dilation data

Signed-off-by: YiHyunjin <hj0412.yi@samsung.com>
4 years ago[circle2circle] Fix segfault for invalid input (#3874)
SaeHie Park [Wed, 19 Aug 2020 09:09:16 +0000 (18:09 +0900)]
[circle2circle] Fix segfault for invalid input (#3874)

This will fix to handle segment fault error for invalid input file

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[tool/model2nnpkg] Fix string comparison bug (#3871)
Hyeongseok Oh [Wed, 19 Aug 2020 09:07:47 +0000 (18:07 +0900)]
[tool/model2nnpkg] Fix string comparison bug (#3871)

Fix string comparison test bug

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[res] TensorFlowLiteRecipes for PadV2 (#3676)
a-zh [Wed, 19 Aug 2020 08:55:53 +0000 (11:55 +0300)]
[res] TensorFlowLiteRecipes for PadV2 (#3676)

This adds PadV2 model recipe

Parent Issue: #3362
Draft PR: #3566

ONE-DCO-1.0-Signed-off-by: Anton Zhukov <a.zhukov@samsung.com>

4 years ago[luci/export] Support NonMaxSuppressionV5 op (#3802)
Alexander Moiseev [Wed, 19 Aug 2020 08:49:35 +0000 (11:49 +0300)]
[luci/export] Support NonMaxSuppressionV5 op (#3802)

Add support for NonMaxSuppressionV5 op in luci exporter.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[luci/import] Add NonMaxSuppressionV5 op (#3801)
Alexander Moiseev [Wed, 19 Aug 2020 08:46:41 +0000 (11:46 +0300)]
[luci/import] Add NonMaxSuppressionV5 op (#3801)

Support NonMaxSuppressionV5 in luci importer.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[luci/service] Support NonMaxSuppressionV5 op (#3800)
Alexander Moiseev [Wed, 19 Aug 2020 08:46:29 +0000 (11:46 +0300)]
[luci/service] Support NonMaxSuppressionV5 op (#3800)

Add shape and type inference for NonMaxSuppressionV5 in luci service.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[luci/logex] Support NonMaxSuppressionV5 op (#3799)
Alexander Moiseev [Wed, 19 Aug 2020 08:46:17 +0000 (11:46 +0300)]
[luci/logex] Support NonMaxSuppressionV5 op (#3799)

Add support for NonMaxSuppressionV5 op in luci/logex.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[onert] Add ifdef guard to controlflow TensorReg (#3868)
Hanjoung Lee [Wed, 19 Aug 2020 08:34:06 +0000 (17:34 +0900)]
[onert] Add ifdef guard to controlflow TensorReg (#3868)

Add ifdef guard to controlflow TensorRegistry.

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[luci] Support for requantization from int8 to uint8 (#3830)
Hyukjin Jeong [Wed, 19 Aug 2020 08:28:10 +0000 (17:28 +0900)]
[luci] Support for requantization from int8 to uint8 (#3830)

This supports requantization from int8 to uin8

ONE-DCO-1.0-Signed-off-by: Hyukjin Jeong <hj1.jeong@samsung.com>

4 years ago[res/TensorFlowLiteRecipes] Correct Unique recipe (#3862)
dr-venkman [Wed, 19 Aug 2020 08:21:43 +0000 (17:21 +0900)]
[res/TensorFlowLiteRecipes] Correct Unique recipe (#3862)

Parent Issue : #1606
Draft PR : #2829

This commit adds a correction for unique recipes that were introduced in #3410

ONE-DCO-1.0-Signed-off-by: venkat.iyer <venkat.iyer@samsung.com>

4 years ago[onert] build nnpkg_run by default for android (#3860)
Sanggyu Lee [Wed, 19 Aug 2020 08:13:54 +0000 (17:13 +0900)]
[onert] build nnpkg_run by default for android (#3860)

nnpackage_run will be built by default for android.

Note: nnpackage_run will be built without feature of loading and dumping
hdf5 if you don't specify pre-built hdf5 libraries.

ONE-DCO-1.0-Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>

4 years ago[infra/cmake] Boost config and Install path (#3867)
Hyeongseok Oh [Wed, 19 Aug 2020 07:50:32 +0000 (16:50 +0900)]
[infra/cmake] Boost config and Install path (#3867)

- Introduce boost source download mirror environment: BOOST_URL
- Changebuild path to intermediate workspace (obj/externals/boost)
- Change user-config.jam generation path to intermediate workspace (obj/externals/boost)
- Change built boost install path to overlay

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[onert] Add TensorRegistry for acl backends (#3865)
Hanjoung Lee [Wed, 19 Aug 2020 06:59:55 +0000 (15:59 +0900)]
[onert] Add TensorRegistry for acl backends (#3865)

- Make `AclTensorBuilder:tensorRegistry` return a tensor registry
- Simplify getting tensor of `controlflow::getTensorBuilder`

This is to make all the backends have `TensorRegistry`.

Part of #3809

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[tool/nnpkg] Set nnpacakge name (#3861)
Hyeongseok Oh [Wed, 19 Aug 2020 06:49:36 +0000 (15:49 +0900)]
[tool/nnpkg] Set nnpacakge name (#3861)

Allow to set nnpackage name by option -p

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[onert] Unify pooling operations to Pool2D (#3853)
Jang Jiseob [Wed, 19 Aug 2020 06:30:21 +0000 (15:30 +0900)]
[onert] Unify pooling operations to Pool2D (#3853)

This commit unifies pooling operations to Pool2D for reducing code duplication.

Signed-off-by: ragmani <ragmani0216@gmail.com>
4 years ago[onert/util] Introduce MDTable write type of ChromeTracingObserver (#3858)
Yongseop Kim [Wed, 19 Aug 2020 06:10:46 +0000 (15:10 +0900)]
[onert/util] Introduce MDTable write type of ChromeTracingObserver (#3858)

* [onert/util] Introduce MDTable write type of ChromeTracingObserver

Introduce new write type of ChromeTracingObserver - MD Table

ONE-DCO-1.0-Signed-off-by: Yongseop Kim <yons.kim@samsung.com>

* Add virtual destructor

ONE-DCO-1.0-Signed-off-by: Yongseop Kim <yons.kim@samsung.com>

* Revert change of WriteFormat

ONE-DCO-1.0-Signed-off-by: Yongseop Kim <yons.kim@samsung.com>

4 years ago[nnpackage_tool] update tflite2circle (#3852)
Sanggyu Lee [Wed, 19 Aug 2020 03:16:22 +0000 (12:16 +0900)]
[nnpackage_tool] update tflite2circle (#3852)

- Update tensorflow schema version v1.12.0 -> v1.13.1
- Remove unnecessary default value writings in json

ONE-DCO-1.0-Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>

4 years ago[onert] Fix HEScheduler and running script (#3855)
Hanjoung Lee [Wed, 19 Aug 2020 02:18:31 +0000 (11:18 +0900)]
[onert] Fix HEScheduler and running script (#3855)

- Fix HEScheduler to make it work
- Update testing script to make it work

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[common-artifacts] Add NonMaxSuppressionV5 tests (#3798)
Alexander Moiseev [Tue, 18 Aug 2020 21:13:41 +0000 (00:13 +0300)]
[common-artifacts] Add NonMaxSuppressionV5 tests (#3798)

Exclude tests for NonMaxSuppressionV5 op in common artifacts.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[luci] Enable PadV2 operator in services (#3677)
a-zh [Tue, 18 Aug 2020 17:18:51 +0000 (20:18 +0300)]
[luci] Enable PadV2 operator in services (#3677)

This enables PadV2 support in services

Parent Issue: #3362
Draft PR: #3566

ONE-DCO-1.0-Signed-off-by: Anton Zhukov <a.zhukov@samsung.com>

4 years ago[tflite2circle] Enable NonMaxSuppressionV5 op (#3794)
Alexander Moiseev [Tue, 18 Aug 2020 16:12:01 +0000 (19:12 +0300)]
[tflite2circle] Enable NonMaxSuppressionV5 op (#3794)

Support NonMaxSuppressionV5 op in tflite2circle.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[luci/lang] Support NonMaxSuppressionV5 op (#3797)
Alexander Moiseev [Tue, 18 Aug 2020 16:02:59 +0000 (19:02 +0300)]
[luci/lang] Support NonMaxSuppressionV5 op (#3797)

Add support for NonMaxSuppressionV5 op in luci IR.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[tfldump] No opt comment for NonMaxSuppressionV5 (#3795)
Alexander Moiseev [Tue, 18 Aug 2020 15:55:57 +0000 (18:55 +0300)]
[tfldump] No opt comment for NonMaxSuppressionV5 (#3795)

Comment about no option in NonMaxSuppressionV5.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[tflchef] Support NonMaxSuppressionV5 op (#3793)
Alexander Moiseev [Tue, 18 Aug 2020 15:55:24 +0000 (18:55 +0300)]
[tflchef] Support NonMaxSuppressionV5 op (#3793)

Support NonMaxSuppressionV5 op in tflchef.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[benchmark] Measure memory usage at the end of each phase (#3843)
JiHwan Yeo [Tue, 18 Aug 2020 09:42:31 +0000 (18:42 +0900)]
[benchmark] Measure memory usage at the end of each phase (#3843)

- Typically, the memory usage is largest at the end of the run
  - However, memory usage is not measured at the end of each phase
  - This PR measures memory usage at the end of each phase

ONE-DCO-1.0-Signed-off-by: JiHwan Yeo <jihwan.yeo@samsung.com>

4 years ago[onert] Remove method `TensorBuilder::iterate` (#3842)
Hanjoung Lee [Tue, 18 Aug 2020 07:45:13 +0000 (16:45 +0900)]
[onert] Remove method `TensorBuilder::iterate` (#3842)

Remove method `TensorBuilder::iterate` which is never used.

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[nncc/infra]Add GEMMLowp and Ruy source for Tensorflow 2.3.0 (#3849)
struss [Tue, 18 Aug 2020 06:50:37 +0000 (15:50 +0900)]
[nncc/infra]Add GEMMLowp and Ruy source for Tensorflow 2.3.0 (#3849)

This commit add GEMMLowp and Ruy source for Tensorflow 2.3.0

ONE-DCO-1.0-Signed-off-by: KiDeuk Bang <rrstrous@nate.com>

4 years ago[onert] Throw when no backends loaded (#3829)
Hanjoung Lee [Tue, 18 Aug 2020 03:24:01 +0000 (12:24 +0900)]
[onert] Throw when no backends loaded (#3829)

Throw when no backends were loaded rather than assertion.

Add a regression test.

Resolve #3826

ONE-DCO-1.0-Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>

4 years ago[luci] Apply disable_validation to optimized model (#3847)
SaeHie Park [Tue, 18 Aug 2020 01:42:36 +0000 (10:42 +0900)]
[luci] Apply disable_validation to optimized model (#3847)

This will apply disable_validation option to optimized model validation
- for real model with dynamic shape in Concat

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[luci] Fix typo in command help (#3848)
SaeHie Park [Tue, 18 Aug 2020 01:41:34 +0000 (10:41 +0900)]
[luci] Fix typo in command help (#3848)

This will fix typo in command help of disable_validation

ONE-DCO-1.0-Signed-off-by: SaeHie Park <saehie.park@gmail.com>

4 years ago[pota-value-test] Add test data for channel-wise quant (#3841)
Hyukjin Jeong [Mon, 17 Aug 2020 21:33:22 +0000 (06:33 +0900)]
[pota-value-test] Add test data for channel-wise quant (#3841)

This adds test data for CWQ of following Ops

- Conv2D, DepthwiseConv2D, FullyConnected, TransposeConv

ONE-DCO-1.0-Signed-off-by: Hyukjin Jeong <hj1.jeong@samsung.com>

4 years ago[tflite2circle] Add shape signature during conversion (#3834)
Seok Namkoong [Fri, 14 Aug 2020 08:10:01 +0000 (17:10 +0900)]
[tflite2circle] Add shape signature during conversion (#3834)

Until now, `tflite2circle` didn't add shape_signature of TFLite model.
This commit will add shape_signature to circle model.

ONE-DCO-1.0-Signed-off-by: Seok NamKoong <seok9311@naver.com>

4 years ago[onert] Unify element-wise binary operations to ElementwiseBinary (#3791)
Jang Jiseob [Fri, 14 Aug 2020 07:41:26 +0000 (16:41 +0900)]
[onert] Unify element-wise binary operations to ElementwiseBinary (#3791)

This commit unifies element-wise binary operations to ElementwiseBinary for reducing code duplication.
  - ElementwiseBinary ops: LogicalAnd, LogicalOr, Maximum, Minimum

Signed-off-by: ragmani <ragmani0216@gmail.com>
4 years agoDo not publish onert-minimal-app on x86 tizen (#3838)
chunseoklee [Fri, 14 Aug 2020 06:58:08 +0000 (15:58 +0900)]
Do not publish onert-minimal-app on x86 tizen (#3838)

- Do not publish onert-minimal-app on x86 tizen

ONE-DCO-1.0-Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>

4 years ago[tool/tflitefile] Add requirements.txt file (#3815)
Hyeongseok Oh [Fri, 14 Aug 2020 05:43:57 +0000 (14:43 +0900)]
[tool/tflitefile] Add requirements.txt file (#3815)

- Add requirements.txt
- Remove deprecated python path setting in code

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[onert] Change param ReduceType to template in base_loader (#3812)
Jang Jiseob [Fri, 14 Aug 2020 04:15:51 +0000 (13:15 +0900)]
[onert] Change param ReduceType to template in base_loader (#3812)

This commit changes param ReduceType to template in base_loader.

Signed-off-by: ragmani <ragmani0216@gmail.com>
4 years ago[onert] div will use MatchElementsSize instead of MatchFlatSize (#3819)
Sanggyu Lee [Fri, 14 Aug 2020 03:54:25 +0000 (12:54 +0900)]
[onert] div will use MatchElementsSize instead of MatchFlatSize (#3819)

MatchFlatSize checks whether tww inputs are same dimension.
It will cause assertion failure. Thus it will use MatchElementsSize.

ONE-DCO-1.0-Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>

4 years ago[tool/update-version] change name and interface of update-version (#3824)
Sanggyu Lee [Fri, 14 Aug 2020 02:47:54 +0000 (11:47 +0900)]
[tool/update-version] change name and interface of update-version (#3824)

- It uses similar syntax for onert version
- ./onert-version.sh will show current version.
- ./onert-version.sh -s 1.10.0 will set version information.
- name is changed since it does more than update.

ONE-DCO-1.0-Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>

4 years ago[tool/update-version] Show current version (#3821)
Hyeongseok Oh [Fri, 14 Aug 2020 01:06:13 +0000 (10:06 +0900)]
[tool/update-version] Show current version (#3821)

Add -v option to show current version
Add execute permission
Rename script

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[onert] Fix weight deallocation to work on android (#3807)
JiHwan Yeo [Fri, 14 Aug 2020 01:04:58 +0000 (10:04 +0900)]
[onert] Fix weight deallocation to work on android (#3807)

- Weight deallocation is enabled when `__ARM_NEON__` flag is exist
  - This flag is legacy and not supported on android
  - It is replaced by `__ARM_NEON`
- Use both `__ARM_NEON` and `__ARM_NEON__` flags to enable weight deallocation on android

ONE-DCO-1.0-Signed-off-by: JiHwan Yeo <jihwan.yeo@samsung.com>

4 years ago[onert] Generate test package using ls (#3822)
Hyeongseok Oh [Fri, 14 Aug 2020 00:41:25 +0000 (09:41 +0900)]
[onert] Generate test package using ls (#3822)

Find directories to package by using shell's "ls" command

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[res] Add examples for NonMaxSuppressionV5 op (#3792)
Alexander Moiseev [Thu, 13 Aug 2020 22:49:55 +0000 (01:49 +0300)]
[res] Add examples for NonMaxSuppressionV5 op (#3792)

Add examples for NonMaxSuppressionV5 op.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>

4 years ago[luci] Enable PadV2 operator in export (#3679)
a-zh [Thu, 13 Aug 2020 15:24:05 +0000 (18:24 +0300)]
[luci] Enable PadV2 operator in export (#3679)

This enables PadV2 operator in luci export

Parent Issue: #3362
Draft PR: #3566

ONE-DCO-1.0-Signed-off-by: Anton Zhukov <a.zhukov@samsung.com>

4 years ago[luci/tests] Add NonMaxSuppressionV4 luci tests (#3705)
Alexander Moiseev [Thu, 13 Aug 2020 13:30:35 +0000 (16:30 +0300)]
[luci/tests] Add NonMaxSuppressionV4 luci tests (#3705)

Add testcases for NonMaxSuppressionV4 op in luci.

ONE-DCO-1.0-Signed-off-by: Alexander Moiseev <a.moiseev@samsung.com>