오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Feb 2019 10:16:07 +0000 (19:16 +0900)]
Remove exeuction interface method to return plan (#4457)
Remove exeuction interface method to return plan
Make field in ANeuralNetworksExecution to reference graph model
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Feb 2019 08:37:21 +0000 (17:37 +0900)]
Introduce IExecutor interface (#4454)
* Introduce IExecutor interface
Introduce IExecutor interface
Access executor using interface on frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Executor class as final
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Feb 2019 03:43:00 +0000 (12:43 +0900)]
Use shape converter on frontend (#4452)
Use defined shape converter on frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Feb 2019 02:21:14 +0000 (11:21 +0900)]
Revert "Revert acl cmake build" (#4447)
This reverts commit
eddc0157eeba4766ce3b121fe009a24788d262ce.
Move acl build into cmake again
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
윤지영/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 21 Feb 2019 02:03:57 +0000 (11:03 +0900)]
Add gather 4D input nnapi test (#4313)
* Add gather 4D input nnapi test
For verifying 4D input case in gather operation.
Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
* Add description for this case
* Fixed wrong comments
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Feb 2019 00:50:42 +0000 (09:50 +0900)]
Revise frontend execution (#4450)
- Change return type of wrapper methods
- Change order of method declarations in wrapper class
- Add noexcept keyword
- Add log message for fail
- Exception handle on wrapper methods
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Feb 2019 00:50:27 +0000 (09:50 +0900)]
Remove useless include in frontend (#4451)
* Remove useless include in frontend
Remove deprecated or duplicated include in frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Recover including NeuralNetworks.h and NeuralNetworksEx.h
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 20 Feb 2019 09:48:15 +0000 (18:48 +0900)]
[neurun] Enable StrideSlice in neurun (#4446)
* [neurun] Enable StrideSlice in neurun
This commit enables `StrideSlice` in neurun that is from `PACL`.
* Apply OperationFactory
* add framework test
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 20 Feb 2019 07:09:13 +0000 (16:09 +0900)]
Revise frontend compilation (#4448)
- Change return type of finish wrapper method
- Move method body into .cc file
- Add noexcept keyword
- Add log message for fail
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 20 Feb 2019 02:46:56 +0000 (11:46 +0900)]
Revise frontend model input output identification (#4442)
- Introduce wrapper method to identify model input & output
- Introduce checker method to check operand usage is operation output
- Add log message for fail
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 20 Feb 2019 00:48:01 +0000 (09:48 +0900)]
Apply subgraph on compile time (#4426)
* Apply subgraph on compile time
Apply subgraph to linearization and stage generation on compile time
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
* Change the name of Element in Linear from node to subgraph
* Add todo for moving dividing graph code into graph
* Add todos for Element in Linear
* Change index of Element from const Index * to Index
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 19 Feb 2019 23:58:41 +0000 (08:58 +0900)]
[neurun] Create AddNode with OperationFactory (#4441)
Move AddNode creation from `ANeuralNetworksModel_addOperation` to
`OperationFactory`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 19 Feb 2019 01:41:50 +0000 (10:41 +0900)]
Revise frontend operation add (#4440)
- Introduce wrapper method to add operation
- Add log message for fail
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 19 Feb 2019 01:11:39 +0000 (10:11 +0900)]
[neurun] Use legacy tflite method in ConvolutionLayer for cpu (#4416)
As we've updated the tensorflow to v1.12, `ConvolutionLayer` kernel for `cpu` was updated.
However,`ConvolutionLayer` of cpu kernel for `quant8` makes wrong matching results when running genereated tests.
Legacy methods of that work well for `quant8`.
`Android NN` master barnch Conv2D also keep using legacy methods.
For consistency and being useful to support `quant8`, we can use legacy methods instead of newly added methods.
(Added TODO comment about changing to new version)
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 19 Feb 2019 01:10:51 +0000 (10:10 +0900)]
[neurun] Rename DepthwiseConv2D.cc (#4439)
Rename `DepthwiseConv2D.cc` to be `DepthwiseConv2DNode.cc` for
class-filename consistency.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 19 Feb 2019 00:49:00 +0000 (09:49 +0900)]
[neurun] Introduce OperationFactory (#4438)
This commit introduces OperationFactory:
- To extract the actual creator from `frontend/model.cc`
- To remove NN API dependency from each `operation::Node`'s constructor
To check if this works, OperationFactory generates Conv2D node.
(other nodes are generated in the as-is way)
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 19 Feb 2019 00:41:47 +0000 (09:41 +0900)]
[neurun] Introduce LowerInfoMap (#4437)
This class combines index map of `operand::LowerInfo` and
`operation::LowerInfo`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Feb 2019 23:59:56 +0000 (08:59 +0900)]
Revise setOperandValueFromMemory (#4436)
- Use wrappeer method
- Add log message for fail
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 18 Feb 2019 10:51:12 +0000 (19:51 +0900)]
[neurun] Revise FullyConncetedLayer for cpu kernel (#4413)
As we've updated the tensorflow to v1.12, `FullyConnectedLayer` kernel for `cpu` was updated, but it is not using `FullyConnected` method in `tflite` directly.
This commit revises `FullyConnectedLayer` kernel for `cpu`, for consistency and being useful to support `quant8`.
% `legacy` method in `legacy_optimized_ops.h` is used, since the method in `optimized_ops.h` does not work well now.
% It's current implementation in Android NN FUllyConncted
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Feb 2019 00:12:34 +0000 (09:12 +0900)]
Revise frontend set operand value (#4432)
- Introduce wrapper method to set operand value
- Introduce wrapper methods to check correctness
- Add log message for fail
- Move method implementation into .cc from header
- Change setOptionalOperand as private
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Feb 2019 00:12:23 +0000 (09:12 +0900)]
Remove index value casting on constructor (#4434)
Remove index value casting on constructor
Check index value and return fail when negative value on frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Sun, 17 Feb 2019 22:24:54 +0000 (07:24 +0900)]
Remove workaround for neurun framework test (#4430)
Remove workaround in Makefile
- Remove framework test list file copying
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 15 Feb 2019 02:21:54 +0000 (11:21 +0900)]
Rename utils for NNAPI to internal converters (#4429)
Rename file name from TypeConvert.* to NNAPIConvert.*
Rename util function name of NNAPI frontend struct converters
Update comments
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 15 Feb 2019 00:14:13 +0000 (09:14 +0900)]
Revise frontend operand add API (#4427)
- Add log message for fail
- Introduce wrapper method
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 15 Feb 2019 00:12:46 +0000 (09:12 +0900)]
Introduce shape converter (#4428)
Introduce shape converter from NNAPI operand type to neurun shape
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 14 Feb 2019 08:10:55 +0000 (17:10 +0900)]
[neurun] Introduce Index::valid method (#4425)
Introduce `valid` method for Index class. True if initialized otherwise
False.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 14 Feb 2019 05:30:14 +0000 (14:30 +0900)]
Apply IStage to IStageGenerator and PlanBuilder (#4414)
Apply IStage(AtomicStage) to IStageGenerator and PlanBuilder.
StageSequence will be applied for Subgraph soon.
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 14 Feb 2019 05:11:29 +0000 (14:11 +0900)]
Revise OperandUsage and related methods (#4408)
Rename OperandUsage as Usage
Remove setters for each enum value and use one setter
Rename getter and setter for usage
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
대인기/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 14 Feb 2019 04:34:11 +0000 (13:34 +0900)]
Remove tabs in libs/tflite/CMakeLists.txt (#4424)
We haven't to use tab but space instead.
This patch removes two tabs and uses spaces instead.
Signed-off-by: Inki Dae <inki.dae@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 14 Feb 2019 04:20:09 +0000 (13:20 +0900)]
[PACL] Change getARMComputePermutationVector to make it generally (#4372)
This commit changes getARMComputePermutationVector to make it generally usable to be able to support NHWC.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 14 Feb 2019 04:17:10 +0000 (13:17 +0900)]
Introduce pointer union to handle tensor data (#4421)
Introduce pointer union in cpu kernel to handle tensor data
It removes redundant reinterpet_cast
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
대인기/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 14 Feb 2019 04:07:50 +0000 (13:07 +0900)]
Use target_compile_definitions instead add_definitions (#4419)
* Use target_compile_definitions instead add_definitions
We moved TFLITE relevant flags to tflite directory so
this flag setting should affect only this target.
This patch replaces add_definitinos with target_compile_definitions
so that this setting can affect only local target.
Signed-off-by: Inki Dae <inki.dae@samsung.com>
* Do not relocate if block
also fix a typo.
Signed-off-by: Inki Dae <inki.dae@samsung.com>
* split a line
Signed-off-by: Inki Dae <inki.dae@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 14 Feb 2019 02:12:03 +0000 (11:12 +0900)]
Revise frontend model object creation and finish (#4412)
* Revise frontend model object creation and finish
- Add log message for fail
- Add noexcept keyword to wrapper method
- Move duplicate finish call check from wrapper to NNAPI implementation
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Return false if exception catched
대인기/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 14 Feb 2019 00:11:39 +0000 (09:11 +0900)]
Move TFLITE relevant flag setting to proper place (#4411)
TFLITE_MAJOR_VER and TFLITE_MINOR_VER flags are used to check
which version of TFLITE should be referred by NNFW at build time.
So these are related to only TFLITE relevant code.
This patch moves the setup script code of these two flags
from CMakeLists.txt to libs/tflite/CMakeLists.txt
Signed-off-by: Inki Dae <inki.dae@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 13 Feb 2019 10:01:20 +0000 (19:01 +0900)]
[neurun] Remove missed skip lists of cpu test for SoftMax (#4417)
This commit removes missed skip lists of cpu test for `SoftMax`.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 13 Feb 2019 09:28:51 +0000 (18:28 +0900)]
Introduce type info converter for frontend (#4418)
Introduce converter from NNAPI operand type to runtime internal type info
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 13 Feb 2019 07:44:46 +0000 (16:44 +0900)]
Support operator to compare TypeInfo (#4415)
Operator overloading to compare TypeInfo
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 13 Feb 2019 04:55:43 +0000 (13:55 +0900)]
Introduce .gitattributes for github/linguist (#4409)
This repo is shown as a Python project due to many test spec files.
This commit prevents searching in generated test directory from linguist
(repo language detection).
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 13 Feb 2019 00:58:45 +0000 (09:58 +0900)]
[neurun] Use 'GenericFullyConnectedLayer' (#4396)
* [neurun] Use 'GenericFullyConnectedLayer`
We introduced `GenericFullyConnectedLayer` to support 4d input.
This commit applies and uses `GenericFullyConnectedLayer` instead of `CLFullyConnectedLayer` directly.
and then we don't need to dealling with `4D input tensor` in `ConstantInitializer` for `FullyConnected` operation anymore, so removed it.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
* Implement calcOffset for cpu and add workaround
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 13 Feb 2019 00:45:15 +0000 (09:45 +0900)]
Remove using default backend for input and output (#4370)
Using default backend is meaningless because of input/output value setting optimization
Use frontend layout information by hardcording
Skip set input tensor value when its backend is not allocated
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 12 Feb 2019 10:12:23 +0000 (19:12 +0900)]
Modified for Tensorflow 1.12 (#4378)
* Modified for Tensorflow 1.12
Fixed error while converting pb to tflite by applying changes in Tensorflow 1.12.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* Modified to a TF 1.12 term "Tensorflow Lite Converter"
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 11 Feb 2019 10:53:09 +0000 (19:53 +0900)]
[neurun] Support 'quant8' for some cpu kernels (#4402)
This commit supports `quant8` data type for some cpu kernels except below.
- `ConvolutionLayer` : no match result
- `FullyConnectedLayer` : need to revise to use `tflite`
Signed-off-by: sjsujinkim sjsujin.kim@samsung.com
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 11 Feb 2019 05:23:48 +0000 (14:23 +0900)]
[neurun] Do not correct dims for Concat operands (#4394)
Concat operands should not change dimensions for its inputs and outputs.
For example, rank 4 (8, 8, 2048, 1) is changed to rank 3 (8, 8, 2048).
It would break Concat operation.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 11 Feb 2019 02:18:50 +0000 (11:18 +0900)]
Support NHWC to SpaceToDepthKernel (#4369)
This commit supports NHWC to SpaceToDepthKernel
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Sun, 10 Feb 2019 23:54:24 +0000 (08:54 +0900)]
Introduce IStage, AtomicStage and CompositStage (#4374)
* Introduce IStage, AtomicStage and CompositStage
Introduce IStage, AtomicStage and CompositStage
- IStage: pure interface for replacing current Stage
- AtomicStage: Stage which can be called as atomic op
- CompositStage: A Stage or Stages which can be grouped
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
* Move {Atomic,Composit}Stage to each own file
* Change struct to final class
* Change CompositStage to StageSequence
* final class to class final
대인기/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 8 Feb 2019 06:46:27 +0000 (15:46 +0900)]
Check tflite version in build time (#4371)
* Check tflite version in build time
This patch seperates what OBS_BUILD flag means into.
- OBS_BUILD flag decides where external packages should be
installed from - if OBS_BUILD == 1 then from tizen package server
otherwise from upstream.
- TFLITE_VER means what tensorflow-lite version nnfw uses - if
TFLITE_VER == 190 then nnfw uses tensorflow-lite v1.9.0 for Tizen
otherwise tensorflow-lite v1.12.0. In default TFLITE_VER is 1120.
For this, I added a TFLITE_VER flag, which will pass proper version
information - 190 for gbs build or 1120 in default - to code side.
Signed-off-by: Inki Dae <inki.dae@samsung.com>
* Seperated TFLITE_VER into TFLITE_MAJOR_VER and TFLITE_MINOR_VER
Signed-off-by: Inki Dae <inki.dae@samsung.com>
* Updated nnapi_delegate.h I missed.
Signed-off-by: Inki Dae <inki.dae@samsung.com>
* Updated and cleaned up
- Cleaned up CMakeLists.txt file
- Added TFLITE_MAJOR_VER and TFLITE_MINOR_VER flags when coverage_build == 1
Signed-off-by: Inki Dae <inki.dae@samsung.com>
* Updated description
Signed-off-by: Inki Dae <inki.dae@samsung.com>
* Fix typo
Signed-off-by: Inki Dae <inki.dae@samsung.com>
이성재/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 7 Feb 2019 08:17:30 +0000 (17:17 +0900)]
Generate compiler command information during build (#4382)
* Generate compiler command information during build
- Set CMAKE_EXPORT_COMPILE_COMMANDS option to generate
`compile_commands.json` file while cmake build.
- It will be used to generate semantic information from various tools.
- Leave a comment about unsupported feature in CMake version 3.1.0
Signed-off-by: Sung-Jae Lee <sj925.lee@samsung.com>
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 7 Feb 2019 07:46:00 +0000 (16:46 +0900)]
Introduce Subgraph as Node (#4373)
* Introduce Subgraph as Node
Introduce Subgraph as Node. This Subgraph would be handled as extrordinay
Node so that running such as group could be possible.
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
* Remove unnecessary header
* Change addOperation name and index of Element of Subgraph
이성재/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 7 Feb 2019 06:04:20 +0000 (15:04 +0900)]
Set default value of HOST_OS (#4384)
- Set default value of HOST_OS in platform identification script.
- For using cmake directly without make.
Signed-off-by: Sung-Jae Lee <sj925.lee@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 7 Feb 2019 05:46:18 +0000 (14:46 +0900)]
Add CUSTOM(Abs) op to nnapi_delegate (#4379)
This commit adds CUSTOM(Abs) op to nnapi_delegate.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Fri, 1 Feb 2019 06:14:35 +0000 (11:44 +0530)]
Arrange GTests in tizen skip lists in order (#4093)
This patch arranges the same for better organizing and tracking
Signed-off-by: prasannar <prasanna.r@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 1 Feb 2019 04:13:44 +0000 (13:13 +0900)]
[neurun] Use 'UNUSED_RELEASE' macro for unused variable (#4362)
* [neurun] Move and use 'NO_USE' macro
This commit moves and use `NO_USE` macro instead of the casting with `(void)`.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
* change macro name
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 1 Feb 2019 01:56:32 +0000 (10:56 +0900)]
Block cpu kernel build on gbs build (#4355)
Block cpu kernel build on gbs build because of build fail
Remove this if ready
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이춘석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 31 Jan 2019 22:22:45 +0000 (07:22 +0900)]
Enable DepthwiseConv2D in neurun (#4319)
* Enable DepthwiseConv2D in neurun
- Enable implicit-padding depthwise conv2d in neurun
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* enable explicit padding
- enable explicit padding
* Fix incorrect comment
* Fix incorrect year on License Statement
* fix format by checker
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 31 Jan 2019 04:16:38 +0000 (13:16 +0900)]
[neurun] Add more comment on cpu::TensorBuilder (#4365)
Add supplementary explation for the reason why `allocate` does nothing.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 31 Jan 2019 03:40:12 +0000 (12:40 +0900)]
Fix logging runtime bug (#4368)
Fix logging runtime bug: print address after console logging
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 31 Jan 2019 01:36:48 +0000 (10:36 +0900)]
Remove unnecessary kernels from ARMComputeEX (#4325)
This commit removes unnecessary kernels from ARMComputeEX
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 31 Jan 2019 00:51:54 +0000 (09:51 +0900)]
Support NHWC to DepthToSpaceKernel (#4363)
* Support NHWC to DepthToSpaceKernel
This commit supports NHWC to DepthToSpaceKernel.
* Change a preprocessor argument of kernel for DepthToSpace
This commit changes a preprocessor argument of kernel for DepthToSpace to prevent misunderstandings.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 31 Jan 2019 00:49:57 +0000 (09:49 +0900)]
[neurun] Introduce operand::IndexMap (#4364)
* [neurun] Introduce operand::IndexMap
Introduce `operand::IndexMap` which is an alias of `unordered_map` with
key as `operand::Index`.
* Update ifdef guard format
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 30 Jan 2019 08:05:54 +0000 (17:05 +0900)]
[neurun] Remove asserts for SoftMax validation to support tensors (#4361)
Removed asserts for `SoftMax` in `PACL` to support tensors.
These asserts copied from `PACL` before changing.
This commit removes asserts for `SoftMax validation` to support tensors on `neurun` too.
And removes test skip list related `SoftMax`.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 30 Jan 2019 08:04:49 +0000 (17:04 +0900)]
[neurun] Support explicit padding and activation for MaxPool2D (#4327)
We can pass all of generated `MaxPool2D` tests.
This commit supports explicit padding and activation for `MaxPool2D` to pass related tests.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 30 Jan 2019 05:31:35 +0000 (14:31 +0900)]
[neurun] Fix reshape bug (Apply #4254) (#4333)
* [neurun] Fix reshape bug (Apply #4254)
Related : #2147, #4254
In #4254, It's set `apply_dim_correction` of `input`/`output` shapes to `false in `PACL`.
This commit applies #4254 to `neurun` and remove related skip test list.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
* Revise CLTensor constructor
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 30 Jan 2019 00:51:03 +0000 (09:51 +0900)]
[neurun] Remove test skip list that can be passed (#4359)
This commit removes tests skip list that can be passed after some PR are merged.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 30 Jan 2019 00:50:05 +0000 (09:50 +0900)]
Build neurun only should be possible (#4357)
Building `nnfw` for ubuntu with `-DBUILD_NEURUN=1 -DBUILD_PURE_ARM_COMPUTE=0`
builds both of neurun and pacl. It should build neurun only as user's input parameters.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 30 Jan 2019 00:48:27 +0000 (09:48 +0900)]
Remove non-const getter in Plan (#4351)
Remove non-const getter in Plan
Use operand context and operation sequence directly for each compile phase
Generate plan after compile finished
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 29 Jan 2019 09:37:18 +0000 (18:37 +0900)]
Rename GENERATE_FILES to GENERATE_NNAPI_TESTS (#4356)
To skip generating nnapi tests, `GENERATE_FILES=0` has been used.
This patch changes the flag's name to more specific one since
`GENERATE_FILES` are too general.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 29 Jan 2019 09:31:32 +0000 (18:31 +0900)]
[neurun] Support Quant8 for Concat (#4352)
This commit supports `quant8` data type for `Concat` on `acl_cl` backend and remove skip related test list.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 29 Jan 2019 09:30:10 +0000 (18:30 +0900)]
[neurun] Support other types for ConstantInitializer (#4349)
Some tests such as `GeneratedTest.conv_quant8` have failed cause of not supporting other types for `ConstantInitializer`.
This commit supports other types for `ConstantInitializer` and remove skip list `conv_quant8xxx`.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
윤지영/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 29 Jan 2019 07:19:04 +0000 (16:19 +0900)]
Support multiple indices upto 4D output ranks in Gather (#4339)
This commit supports below test cases:
2D input + 2D indices => 3D output
2D input + 3D indices => 4D output
3D input + 2D indices => 4D output
Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 29 Jan 2019 03:27:42 +0000 (12:27 +0900)]
Use shared pointer for operands and opertions in Plan (#4350)
Use shared pointer for operands and opertions in Plan
Define new contructor to set operands and operations
Remove setCompile method and friend class
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 28 Jan 2019 23:52:30 +0000 (08:52 +0900)]
Remove plan getter in Compiler (#4347)
Remove plan getter in Compiler
Prepare remove non-const model getter in Plan
Fix plain initialize in Compiler to return nullptr when release() is called before compile()
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이성재/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 28 Jan 2019 23:45:03 +0000 (08:45 +0900)]
update `.gitignore` (#4337)
- update `.gitignore` to support `gnu-global` tool
Signed-off-by: Sung-Jae Lee <sj925.lee@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 28 Jan 2019 10:01:06 +0000 (19:01 +0900)]
[PACL] Support ABS_EX op (#4334)
This commit supports ABS_EX operation for PACL.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 28 Jan 2019 07:56:02 +0000 (16:56 +0900)]
[neurun] Support explicit padding and activation for Conv2D (#4330)
we can pass all of generated `Conv2D` tests(except `quant8` tests).
This commit supports explicit padding and activation for `Conv2D` to pass related tests.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 28 Jan 2019 07:55:16 +0000 (16:55 +0900)]
Remove NOT_COMPILED state in compiler (#4332)
* Remove NOT_COMPILED state in compiler
Remove NOT_COMPILED in compiler's state
Make isCompiled flag to check compile result
* Rename to set compiled
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 28 Jan 2019 07:52:28 +0000 (16:52 +0900)]
Replace CLNormalizationLayerEX with CLNormalizationLayer (#4324)
This commit replaces CLNormalizationLayerEx of ARMComputeEx with CLNormalizationLayer of ARMCompute.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
이성재/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 28 Jan 2019 06:49:45 +0000 (15:49 +0900)]
fix bug (#4331)
- fix bug in case statement.
- change confusing(or maybe buggy) code to clear one.
Signed-off-by: Sung-Jae Lee <sj925.lee@samsung.com>
윤지영/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 28 Jan 2019 05:14:15 +0000 (14:14 +0900)]
Support Gather in all input ranks (#4312)
This commit supports Gather operation in all input ranks.
But it only supports 1D rank in indices.
Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 28 Jan 2019 01:10:52 +0000 (10:10 +0900)]
[neurun] Remove unnecessary header of operand (#4308)
Remove unnecessary header of `operand::Object` and added header
inclusion from the files that are depend on `operand::LowerInfo`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 28 Jan 2019 01:05:32 +0000 (10:05 +0900)]
Replace CLStridedSliceEx with CLStridedSlice (#4315)
* Replace CLStridedSliceEx with CLStridedSlice
This commit replaces CLStridedSliceEx of ARMComputeEx with CLStridedSlice of ARMCompute.
* Fix wrong initialization of std::vector
This commit fixes initialization of std::vector when compiling StridedSlice operation.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 28 Jan 2019 00:40:58 +0000 (09:40 +0900)]
[neurun] Fix `TensorInfo::total_size` (#4321)
Fix calculation of total size for a tensor to use actual element size.
Before it assumed the element size of 4.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 28 Jan 2019 00:40:21 +0000 (09:40 +0900)]
Move compute from frontend to internal (#4310)
Move compute phase from frontend to Executor class
Remove not used methods in ANeuralNetworksExecution and Executor
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 28 Jan 2019 00:39:43 +0000 (09:39 +0900)]
[neurun] Fix to set wrong length for PermuteSource/Sink (#4300)
* [neurun] Fix setting wrong length for PermuteSource/Sink
Rleated : #2147
I've tried to run the generated tests `GeneratedTests.avg_pool_quant8_1~4`,since merged all related PRs.
But the tests are failed with wrong matching.
I figured out the reason of problem is from setting wrong length for `PermuteSource/Sink` by debugging.
This commit fixes setting wrong length for `PermuteSource/Sink` and removes skip lists for that tests.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
* Fix conflict
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 28 Jan 2019 00:39:19 +0000 (09:39 +0900)]
[neurun] Change reshape kernel from `CLReshapeLayer` to `GenericReshapeLayer` (#4314)
Related : #2147
In #1353,
`GenericReshapeLayer` was introduced to remove height/width restriction, and `PACL` uses that now.
This commit changes reshape kernel from `CLReshapeLayer` to `GenericReshapeLayer`.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 25 Jan 2019 06:46:58 +0000 (15:46 +0900)]
Avoid redefinition of macro functions (#4320)
Avoid redefinition of CAST_CL and CAST_NE that generates warning.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 25 Jan 2019 06:24:33 +0000 (15:24 +0900)]
Fix concat for not rank 4 (#4317)
Now concat operation rank is not 4 can run using cpu implementation
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 25 Jan 2019 00:40:39 +0000 (09:40 +0900)]
[neurun] Move Model class to `model` (#4303)
Move Model class to `model` from `graph` since `Model` represents pure
neural network primitive elements.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 24 Jan 2019 03:33:56 +0000 (12:33 +0900)]
Update the description of Gather (#4277)
This commit updates the description of the types that CLGather and CLGatherKerenl support.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 24 Jan 2019 00:00:56 +0000 (09:00 +0900)]
Introduce BUILD_UBEN option (#4309)
* Introduce BUILD_UBEN option
This commit introduces BUILD_UBEN option to fix Tizen build break.
* Set default value
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 23 Jan 2019 23:44:24 +0000 (08:44 +0900)]
[PACL] Replace CLReduceOperation with CLReduceMean. (#4304)
This commit replaces CLReduceOperation with CLReduceMean in case of mean operation.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 23 Jan 2019 23:44:12 +0000 (08:44 +0900)]
[PACL] Replace CLLogicalNot with CLBitwiseNot (#4299)
This commit replaces CLLogicalNot of ARMComputeEx with CLBitwiseNot of ARMCompute
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 23 Jan 2019 23:39:33 +0000 (08:39 +0900)]
Move input and output setting into Executor class (#4293)
* Move input and output setting into Executor class
Move source (input setting) and sink (output setting) into Executor class
* Fix source list name to same as directory name
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 23 Jan 2019 23:39:13 +0000 (08:39 +0900)]
[neurun] Extract sizeOfDataType as a function (#4307)
Extract sizeOfDataType as a function which were implemented in class
`Object`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 23 Jan 2019 10:24:14 +0000 (19:24 +0900)]
Introduce uben (#4301)
* Introduce uben
This commit introduces uben (Micro BENchmark) suite based on nonius
benchmark framework.
Initial version includes convolution 3x3 (with unit stride) benchmark
over ARM Compute CL Direct/GEMM/Winograd kernels.
* Fix a bug in ker_info method
* Use CLWinogradConvolutionLayer correctly
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 23 Jan 2019 07:35:21 +0000 (16:35 +0900)]
[PACL] Replace CLPixelWiseDivision with CLArithmeticDivision (#4302)
This commit replaces CLPixelWiseDivision of ARMComputeEx with CLArithmeticDivision of ARMCompute.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 23 Jan 2019 05:04:05 +0000 (14:04 +0900)]
[neurun] Support activation for AvgPool2D (#4297)
Related : #2147
This commit supports unimplemented activation for AvgPool2D.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 23 Jan 2019 04:59:15 +0000 (13:59 +0900)]
[PACL] Add the declaration of asTensorInfo (#4298)
This commit adds the declaration of asTensorInfo.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 22 Jan 2019 09:59:13 +0000 (18:59 +0900)]
[neurun] Support explicit padding for AvgPool2D (#4278)
* [neurun] Support explicit padding for AvgPool2D
Related : #2147
This commit supports explicit padding for `AvgPool2D`.
- Merge `Implicit` and `Explicit` `Node` compared to `PACL`
- Remove related test skip list
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
* move explicit_padding to param, add TODO comment
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 22 Jan 2019 06:28:22 +0000 (15:28 +0900)]
Move source and sink in wrapper (#4290)
* Move source and sink in wrapper
Move source and sink setting and call in wrapper into Executor class
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Remove header include
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 22 Jan 2019 04:14:12 +0000 (13:14 +0900)]
[PACL] Replace CLBatchToSpaceND with CLBatchToSpaceLayer (#4283)
This commit replaces CLBatchToSpaceND of ARMComputeEx with CLBatchToSpaceLayer of ARMCompute
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>