오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 1 Apr 2019 10:08:26 +0000 (19:08 +0900)]
Introduce cpu quant8 convolution kernel (#4910)
Introduce cpu quantized int8 convolution kernel from tflite and gemmlowp
Use kernel in neurun cpu backend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 1 Apr 2019 09:30:59 +0000 (18:30 +0900)]
Introduce float fully connected kernel (#4894)
Introduce float fully connected kernel from tflite
Use kernel in neurun cpu backend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 1 Apr 2019 07:33:38 +0000 (16:33 +0900)]
Install unzip in docker image (#4908)
Install unzip in docker image
It will be used for x86-64 test
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 1 Apr 2019 06:56:00 +0000 (15:56 +0900)]
Introduce cpu float depthwise convolution kernel and backend (#4906)
Introduce cpu float depthwise convolution kernel from tflite
Implement neurun cpu backend for depthwise convolution
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 1 Apr 2019 05:31:19 +0000 (14:31 +0900)]
Platform specific cmake option setting (#4898)
Make platform specific cmake option setting files
Platform specific setting have higher priority than common setting
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 1 Apr 2019 02:08:47 +0000 (11:08 +0900)]
Add assertion for cpu backend shape conversion (#4904)
CPU backend assume that neurun's internal shape is always same or less than 4.
It is true on current neurun implementation, but it may change.
So add assertion to check this.
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 1 Apr 2019 01:42:50 +0000 (10:42 +0900)]
Fix cmake acl module check (#4905)
Use QUIET to download source code in acl module check,
and return false when fail to get source code
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 1 Apr 2019 00:58:51 +0000 (09:58 +0900)]
[neurun] Add validation of ComparisonNode (#4901)
This commit adds validation of ComparisonNode.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 1 Apr 2019 00:57:44 +0000 (09:57 +0900)]
Fix wrong nnapi tests path in Doxyfile's exclude (#4899)
Wrong nnapi tests location is fixed in Doxyfile exclude.
It causes "input buffer overflow, can't enlarge buffer
because scanner uses REJECT" error.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 1 Apr 2019 00:40:25 +0000 (09:40 +0900)]
[neurun] Simplify Backend C Interface (#4891)
C API was 3 functions that creates each components of a Backend. This
commit revises it to have only one API function `neurun_backend_create`
which packs all the components.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 1 Apr 2019 00:37:39 +0000 (09:37 +0900)]
Move cmake buildtool flag setting (#4893)
Move cmake buildtool flag setting file into cmake/buildtool/
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 1 Apr 2019 00:37:29 +0000 (09:37 +0900)]
Introduce cpu maxpool kernel (#4897)
Introduce cpu maxpool kernel from tflite
Use kernel in neurun cpu backend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 29 Mar 2019 06:16:23 +0000 (15:16 +0900)]
Introduce float convolution kernel (#4892)
Introduce float convolution kernel from tflite
Use kernel in neurun cpu backend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 29 Mar 2019 06:16:06 +0000 (15:16 +0900)]
[neurun] Enable operations using boolean type (#4883)
This commit enable operations using boolean type.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
윤지영/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 29 Mar 2019 04:21:56 +0000 (13:21 +0900)]
Revert "[PACL] Apply DepthwiseConvolutionLayer3x3 to DepthwiseConv operation (#4354)" (#4840)
This reverts commit
34f49de36ae41f10735fd570e006981671e13a68.
The same changes were merged in acl v19.02.
Related PR : https://review.mlplatform.org/#/c/ml/ComputeLibrary/+/511/
Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 28 Mar 2019 09:44:17 +0000 (18:44 +0900)]
[neurun] DataflowExecutor indentation fix (#4890)
* [neurun] DataflowExecutor indentation fix
Remove the extra block in DataflowExecutor implementation.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
* Remove another block
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이춘석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 28 Mar 2019 09:43:37 +0000 (18:43 +0900)]
[neurun] Add NEON backend (#4522)
* [neurun] Add CMakeLists.txt for NEON
Add CMakeLists.txt for NEON backend
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* [neurun] Replicate the acl_cl backend code
- This is just a replication of acl_cl backend code
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* [neurun] Add acl_neon sub dir
Add acl_neon sub dir for NEON backend
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* [neurun] Fix build failure
[neurun] Fix build failure in NEON
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* Add acl_neon backend into Config file
Add acl_neon backend into Config file
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* Use NEReshapeLayer for runinng mobilenet model
With this patch, neurun can run mobilenet
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* Fix build error after rebase
Fix build error after rebase
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* Fix copyright year and macros
Fix copyright year and macros
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* Remove unsupported op from StageGenerator.cc
Remove unsupported op from StageGenerator.cc
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* Use NEFullyConnectedLayer instead of GenericFCLayer
Use NEFullyConnectedLayer instead of GenericFCLayer
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* [neurun] Fix build error
[neurun] Fix build error #2
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* [neurun] Reuse cl version of convert and Swizzle
Also fixed INETensor class: it didn't implemented one abstract member of ITensor
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* [neurun] Rebase onto master
[neurun] Rebase changes onto master
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* [neurun] Phase 2: Reuse other backends codes where needed
Reuse other backends codes where needed
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* [neurun] Fix linking errors of reused codes
Fix linking errors of reused codes
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Separate acl common files into single dir
Pull all files, that are used in both acl_cl and acl_neon into separate
dir
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Cleanup virtual function declarations in neurun backends
Cleanup virtual function declarations in neurun backends.
This is borrowed from PR 4791
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Rename make_acl_function as asAclFunction
Fix coding style of make_acl_function
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Move cpu::Object into neurun_core::backend
Move cpu::Object into neurun_core since it is used in NEON also
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Rebase onto master: conflict because of TensorInfo
Rebase onto master: conflict because of TensorInfo
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Fix copyright year of acl_common/TemplTensorBuilder.h
Fix copyright year of acl_common/TemplTensorBuilder.h
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Merge onto master: fixStageGenerator
Merge onto master: fixStageGenerator
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
윤지영/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 28 Mar 2019 07:57:28 +0000 (16:57 +0900)]
[pureacl] Use fused_act parameter in Conv and DepthwiseConv acl layer (#4851)
* [pureacl] Use fused_act parameter in Conv and DepthwiseConv acl layer
acl v19.03 support fused_act parameter in Conv and DepthwiseConv layer.
Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
* Revert DepthwiseConvolution changes
Activation layer information only works for QASYMM8 in DepthwiseConvolutionLayer.
* Add comments for configure of ConvolutionLayer
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 28 Mar 2019 07:56:37 +0000 (16:56 +0900)]
[Interpreter] Allocate and assign constant tensor (#4888)
Allocate tensor for constant operand
Assign constant tensor to exec-env
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 28 Mar 2019 07:21:29 +0000 (16:21 +0900)]
[cker] Introduce average pool kernel (#4889)
Introduce average pool kernel from tflite
Use kernel in neurun cpu backend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 28 Mar 2019 06:30:23 +0000 (15:30 +0900)]
Introduce concat kernel (#4887)
Introduce concat kernel from tflite
Use concat kernel in cpu backend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 28 Mar 2019 06:24:18 +0000 (15:24 +0900)]
[neurun] Rename Executor as LinearExecutor (#4874)
Before this was the only Executor so its name was "Executor", now that
we have more Executors so rename it to prevent from confusion.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 28 Mar 2019 04:06:26 +0000 (13:06 +0900)]
Introduce cpu kernel in libs (#4886)
* Introduce cpu kernel in libs
Introduce cpu kernel in libs for cpu backend and interpreter
Introduce shape implementation from tflite
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Fix header define and license format
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 28 Mar 2019 04:05:58 +0000 (13:05 +0900)]
Move cmake compile flag setting (#4880)
* Move cmake compile flag setting
Move target independent/dependent copmile flag setting into ApplyCompileFlags.cmake
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Remove setting duplication
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 28 Mar 2019 04:02:08 +0000 (13:02 +0900)]
Introduce ExecEnv for intepreter (#4882)
Define ExecEnv to gather interpreter environment - status, memory, etc
Assign input and output tensor into ExecEnv
Fix unittest
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 27 Mar 2019 10:56:38 +0000 (19:56 +0900)]
[neurun] Enable setting the data type of mode operand::Object (#4881)
This commit Enable setting the data type of made operand::Object.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 27 Mar 2019 09:53:14 +0000 (18:53 +0900)]
Fix cl codes' format (#4820)
Fix cl codes' format using c++ format
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 27 Mar 2019 08:08:07 +0000 (17:08 +0900)]
Move cmake platform identifier (#4878)
* Move cmake platform identifier
Move cmake file for platform identify
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Move IdentifyPlatform.cmake and use nnfw_include
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 27 Mar 2019 07:21:14 +0000 (16:21 +0900)]
GBS neurun build (#4876)
Build neurun as default on gbs build
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 27 Mar 2019 05:29:03 +0000 (14:29 +0900)]
Fix build error by conflict (#4873)
Fix build error by operationFactory parameter change
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 27 Mar 2019 05:06:34 +0000 (14:06 +0900)]
Move backend unittest (#4870)
Move backend unittest into each backend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김수진/Quality Tool Lab(SR)/Engineer/삼성전자 [Wed, 27 Mar 2019 04:55:51 +0000 (13:55 +0900)]
[neurun] Enable LocalResponseNorm, DepthToSpace, ReduceMin ops (#4714)
Related : #3708
This commit enables `LocalResponseNormalization`, `DepthToSpace`, `ReduceMin` ops for `acl_cl`.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 27 Mar 2019 04:52:16 +0000 (13:52 +0900)]
[neurun] Reuse code in Executor::execute (#4869)
Reuse setting inputs and outputs part with template method pattern.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 27 Mar 2019 04:27:18 +0000 (13:27 +0900)]
[neurun] Add passing a param of operand::Set when creating the Node (#4867)
This commit adds passing a param of operand::Set when creating the Node in order to adjust operand information for each operation.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 27 Mar 2019 03:36:47 +0000 (12:36 +0900)]
Coverage test with log (#4868)
Enable logging and dot-dump on coverage test
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 27 Mar 2019 01:32:50 +0000 (10:32 +0900)]
[neurun] Bugfix - running with multiple backends (#4866)
Due to not defining DefBackend for model inputs, there was a problem
when running with multiple backends.
Fix #4828
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 26 Mar 2019 23:57:50 +0000 (08:57 +0900)]
[neurun] Promote OperandContext to ExecutorBase (#4864)
As all the executors have OperandContext in common, we can promote it to
its base class `ExecutorBase`.
Part of #4823
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 26 Mar 2019 11:20:15 +0000 (20:20 +0900)]
Enable external IO memory set validation test (#4860)
Implement NNAPI to set input and output using external memory
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 26 Mar 2019 08:25:55 +0000 (17:25 +0900)]
Enable supported validation test (#4862)
Enable validation test
- SetInput
- SestOutput
- EventWait
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 26 Mar 2019 06:34:51 +0000 (15:34 +0900)]
Make sure standalone scripts are not source'd (#4853)
Scripts under standalone are supposed to run, not sourced.
On error, it would exit, it may cause the shell terminated.
To prevent unintended use style, each script is checked whether
it is sourced or not.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 26 Mar 2019 06:10:53 +0000 (15:10 +0900)]
Remove gbs build check for cpu backend build (#4858)
To build neurun cpu backend, check tensorflow lite internal source code instead of gbs build flag
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 26 Mar 2019 04:56:41 +0000 (13:56 +0900)]
[neurun] Move out of the graph of neurun to collect the model's information (#4838)
This commit moves out of the graph of neurun to collect the model's infomation.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 26 Mar 2019 04:19:32 +0000 (13:19 +0900)]
Remove architecture check for backend build (#4854)
To build neurun acl_cl backend, check acl support instead of target architecture name
Fix arm architecture flag bug
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 26 Mar 2019 02:24:43 +0000 (11:24 +0900)]
Check invalid NNAPI input and output index (#4849)
* Check invalid NNAPI input and output index
Check too large index by return invalid index value in wrppaer function
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Fix validation check log
김수진/Quality Tool Lab(SR)/Engineer/삼성전자 [Tue, 26 Mar 2019 02:24:17 +0000 (11:24 +0900)]
[neurun] Enable ArgMax, Dequantize, Mean ops (#4751)
* [neurun] Enable ArgMax, Dequantize, Mean ops
This commit enables `ArgMax`, `Dequantize`, `Mean` ops for `acl_cl`.
* fix build break
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 26 Mar 2019 00:53:12 +0000 (09:53 +0900)]
Check input and output tensor length (#4847)
Check correctness of length for input and output tensor setting based on NNAPI spec
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 25 Mar 2019 11:09:20 +0000 (20:09 +0900)]
Fix test script bug (#4837)
* Fix test script bug
- Check copy and script failure
- Fix test driver parameter
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Add set -e for all standalone scripts
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 25 Mar 2019 10:19:26 +0000 (19:19 +0900)]
Fix rootfs default setting bug (#4846)
Fix rootfs default setting bug by changing path
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 25 Mar 2019 10:19:15 +0000 (19:19 +0900)]
[neurun] Extract OperandContext from Plan (#4834)
Extract OperandContext from Plan so OperandContext can be merged in
ExecutorBase. (Both Executor and DataflowExecutor contains
OperandContext so we can promote it to ExecutorBase, later.)
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 25 Mar 2019 10:02:38 +0000 (19:02 +0900)]
Move util for frontend only (#4843)
* Move util for frontend only
Move util implementation for NNAPI frontend only
* Define static function for convert
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 25 Mar 2019 10:01:58 +0000 (19:01 +0900)]
Change compiler::TensorInfo to model::operand::Info (#4830)
Change compiler::TensorInfo to model::operand::Info
All data fields in TensorInfo is model operand's info
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
Redo TensorBuilderSet as unordered_set (#4836)
Redo TensorBuilderSet as unordered_set since we don't care if it is
sorted or not
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 25 Mar 2019 09:15:43 +0000 (18:15 +0900)]
Rename toolchain setting files (#4839)
- Rename toolchain setting directory and files for cross build to cmake/buildtool/cross/toolchain_$(TARGET_ARCH_LC)-$(TARGET_OS).cmake
- Remove needless x86-64 cross build toolchain setting
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 25 Mar 2019 05:07:42 +0000 (14:07 +0900)]
Move cmake options (#4841)
Move cmake options for build and download into cmake/CfgOptionFlgs.cmake
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 25 Mar 2019 03:21:16 +0000 (12:21 +0900)]
Introduce tensor interface and class for interpreter (#4818)
Introduce tensor interface for interpreter
Introduce ROTensor (read-only tensor) and Tensor (writable tensor) class
Generate interpreter tensor when input/output is set
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 22 Mar 2019 08:41:35 +0000 (17:41 +0900)]
[neurun] Abstract compilation into functions (#4803)
* [neurun] Abstract compilation into functions
Abstract compilation procedure as a function for each Executors.
They are introduced as static, private and has a parameter and a return
value. It is intended to explicitly show which is input and output for
compilation.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
* Rename functions prefix - compile to create
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Fri, 22 Mar 2019 07:31:05 +0000 (16:31 +0900)]
[neurun] Replace uint32_t with Index in Set (#4816)
operand::Set and operation::Set have hardcoded uint32_t and assertion,
which is based on wrong assumption that Index is always uint32_t.
This patch replaces uint32_t with Index.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 22 Mar 2019 06:30:41 +0000 (15:30 +0900)]
Fix tizen cross build bug (#4831)
Fix to use nnfw command in tizen cross build script
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
Андрей Шедько/AI Tools Lab /SRR/Engineer/삼성전자 [Fri, 22 Mar 2019 03:45:27 +0000 (06:45 +0300)]
Cleanup virtual function declarations in neurun backends (#4791)
Cleanup virtual function declarations in neurun cpu and acl_cl backends
Signed-off-by: Andrei Shedko <a.shedko@samsung.com>
[neurun] Remove unneeded dependency of neurun-core on tflite (#4817)
Looks like it is a remnant of times when neurun was being built as a single library (now it is divided into several parts, such as core, front end and back end).
Signed-off-by: Sergei Barannikov <s.barannikov@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Mar 2019 23:42:44 +0000 (08:42 +0900)]
Remove skiplist for pacl x86-64 (#4821)
Remove skiplist for pacl x86-64 because pacl is not available for x86-64
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Mar 2019 23:35:44 +0000 (08:35 +0900)]
Enable NNAPI compilation validation test (#4808)
* Enable NNAPI compilation validation test
Enable ValidationTestCompilation.Finish test
Define compile state STARTED
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Rename NONE to CREATED
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Mar 2019 10:49:29 +0000 (19:49 +0900)]
Remove simple operation in ARMComputeEx (#4814)
Remve simple operation in ARMComputeEx except pack/unpack
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 21 Mar 2019 10:49:09 +0000 (19:49 +0900)]
make distclean should remove nnapi generated tests (#4812)
`make distclean` should remove all derived files after git cloned.
nnapi generated tests are among them.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Mar 2019 10:48:51 +0000 (19:48 +0900)]
Add reduce max and custom operator framework test (#4811)
Add custom op for reduce max and reduce sum test (for tizen build)
Add custom op squared difference and abs test
Add reduce_max op test
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Mar 2019 10:48:06 +0000 (19:48 +0900)]
Set input and output tensor info for interpreter (#4798)
- Set input and output tensor info for interpreter
- Fix TensorInfo bug
- Remove sink/source implementation
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Mar 2019 10:07:36 +0000 (19:07 +0900)]
Remove NENormalizationLayer (#4813)
Remove deprecated NENormalizationLayer and NENormalizationLayerExKernel implementation
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 21 Mar 2019 05:20:40 +0000 (14:20 +0900)]
[neurun] Enable Gather, Neg, Abs ops (#4740)
This commit enables `Gather`, `Neg`, `Abs` ops for `acl_cl`.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Mar 2019 05:06:46 +0000 (14:06 +0900)]
Enable NNAPI I/O identify validation check (#4810)
Enable ValidationTestIdentify.DuplicateInputs and ValidationTestIdentify.DuplicateOutputs for acl_cl kernel
Enable all ValidationTestIdentify test for cpu kernel
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 21 Mar 2019 03:46:03 +0000 (12:46 +0900)]
[neurun] Add additional skip lists related to unsupported U8 (#4809)
As We've discussed in #4797 and already merged skip list PR #4805, but some tests related to unsupported U8 are also failed.
This commit adds additional skip lists related to unsupported U8.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 21 Mar 2019 02:18:15 +0000 (11:18 +0900)]
[Tools] tool to convert pb file (binary) to pbtxt file (text) (#4788)
* [Tools] Convert pb file to pbtxt file
This tool converts frozen pb file in tensorflow to pbtxt text file.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* remove common.py
* remove underbar from parameter
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Mar 2019 01:40:39 +0000 (10:40 +0900)]
Introduce paramter checker (#4804)
Introduce parameter checker. As a first commit,
- Check add node only
- Check activation paramter is constant only
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 21 Mar 2019 01:25:00 +0000 (10:25 +0900)]
Rename reduce max operation (#4806)
Rename NNAPI Ex operation for reduce max
TENSORFLOW_MAX_EX -> REDUCE_MAX_EX
Update NeuralNetworksEx.h comment
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 21 Mar 2019 01:06:45 +0000 (10:06 +0900)]
Remove const keyword from converting functions in NNAPIConvert (#4802)
This commit removes const keyword from converting functions in NNAPIConvert.
In general, const is used on a return value of RVO because const affect move semantics.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 21 Mar 2019 01:05:45 +0000 (10:05 +0900)]
Append nnapi tests of comparison into skiplist of neurun (#4805)
This commit appends nnapi tests of comparison into skiplist of neurun.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
Vladimir Plazun/AI Tools Lab /SRR/Engineer/삼성전자 [Thu, 21 Mar 2019 00:28:13 +0000 (03:28 +0300)]
Cleanup build warnings (#4794)
Remove redundant `virtual` declarations and use `override` instead( as per Core Guidelines C.128 )
Signed-off-by: Vladimir Plazun <v.plazun@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 20 Mar 2019 08:37:04 +0000 (17:37 +0900)]
[neurun] Merge Equal and NotEqual to Comparison op (#4797)
In https://github.sec.samsung.net/STAR/nnfw/pull/4738#discussion_r156180, we've discussed to group all operations in `::arm_compute::ComparisonOperation`.
This commit introduces `ComparisonNode` that includes operations in `::::arm_compute::ComparisonOperation` and intergrate `Equal` and `NotEqual` to `ComparisonNode`
`ComparisonNode` can be support below.
```
enum class ComparisonType
{
Equal,
NotEqual,
Greater,
GreaterEqual,
Less,
LessEqual
};
```
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 20 Mar 2019 07:36:22 +0000 (16:36 +0900)]
Fix cmake regular expression bug (#4790)
Use variable setting instead of regular expression
because option setting by cmake configure is not working
It is temporary solution until changing option setting for each architecture
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 20 Mar 2019 04:32:40 +0000 (13:32 +0900)]
[neurun] Enable Equal, SquaredDifference, TopKV2 ops (#4738)
* [neurun] Enable Equal, SquaredDifference, TopKV2 ops
This commit enables `Equal`, `SquaredDifference`, `TopKV2` ops for `acl_cl`.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
* Fix build break
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 20 Mar 2019 02:11:50 +0000 (11:11 +0900)]
[neurun] Remove unnecessary target include dirs (#4787)
Remove `target_include_directories` for NEURUN_INCLUDE_DIR.
However variable NEURUN_INCLUDE_DIR cannot be removed for now since
cpu/acl_cl kernel module improperly depend on this dir.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
Vladimir Plazun/AI Tools Lab /SRR/Engineer/삼성전자 [Wed, 20 Mar 2019 01:12:22 +0000 (04:12 +0300)]
Fix missing symbol in logging backend (#4793)
* Fix missing symbol in logging backend
Add missing ANeuralNetworksCompilation_free in logging backend
Signed-off-by: Vladimir Plazun <v.plazun@samsung.com>
* Format fix
Signed-off-by: Vladimir Plazun <v.plazun@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 19 Mar 2019 08:09:46 +0000 (17:09 +0900)]
[neurun] Extract core module CMakeLists.txt (#4789)
Extract CMakeLists.txt file for neurun/core.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 19 Mar 2019 08:07:32 +0000 (17:07 +0900)]
Save backend dlopen handle (#4784)
Save backend dlopen handle in map member variable
It uses to avoid warning by handle lost without dlclose(), but it will be used to other purpose later
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 19 Mar 2019 08:06:49 +0000 (17:06 +0900)]
[neurun] Enable LogicalOr, LogicalNot ops (#4734)
* [neurun] Enable LogicalOr, LogicalNot ops
Related : #3708
This commit enables `LogicalOr`, `LogicalNot` ops for `acl_cl`.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
* Change LogicalOr kernel to CLBitwiseOr
* Fix build break
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 19 Mar 2019 06:45:52 +0000 (15:45 +0900)]
Introduce TensorInfo for interpreter (#4780)
Introduce TensorInfo for interpreter
Current implementation is copy of compiler::TensorInfo
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 19 Mar 2019 06:45:18 +0000 (15:45 +0900)]
Interpreter unittest (#4783)
* Interpreter unittest
Introduce interpreter unittest: unittest for ExecManager
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Add comment and namespace
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 19 Mar 2019 03:58:45 +0000 (12:58 +0900)]
[neurun] Re-organize neurun core directories (#4781)
Introduce core directory and separate headers into two types - private
and public. Public header may be used by frontend and backend
implemenatations.
New directory structure is as follows:
```
- neurun
- frontend (as-is)
- backend (as-is)
- core
- include
- src
```
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 19 Mar 2019 00:35:20 +0000 (09:35 +0900)]
Remove "externals/CMakeLists.txt" (#4777)
This commit moves all the scripts related android support into contrib/
directory, and removes "externals/CMakeLists.txt" from the repo.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Mar 2019 12:30:54 +0000 (21:30 +0900)]
Use pre-built TensorFlow Lite (#4779)
This commit updates TensorFlow Lite CMake package to lookup pre-built
binaries from the system if BUILD_TENSORFLOW_LITE is set as OFF.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Mar 2019 09:34:07 +0000 (18:34 +0900)]
Check format all files as default (#4775)
Check format all c and c++ files as default
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Mar 2019 09:33:52 +0000 (18:33 +0900)]
Require TensorFlow Lite when building nnfw_lib_tflite (#4774)
nnfw_lib_tflite depends on TensorFlow Lite, but there is no explicit
find_package statement.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 18 Mar 2019 08:24:31 +0000 (17:24 +0900)]
[neurun] Organize directories of backend impls (#4773)
Extract backend implementations outside of `src` directory.
Now the directory structure is:
```
- neurun
- frontend
- nnapi
- backend
- acl_cl
- cpu
- src (will be `core`)
```
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Mar 2019 08:24:03 +0000 (17:24 +0900)]
Introduce interpreter ExecManager (#4747)
To introduce interpreter ExecManager, rename Interpereter class
ExecManager will handle execution environment and invoke intepreter
Execution environment and interpreter will be introduced as separated module later
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Mar 2019 08:14:40 +0000 (17:14 +0900)]
Move scripts related with tflite_benchmark_model (#4770)
This commit extracts scripts related with tflite_benchmark_model
from "externals/CMakeLists.txt", and move these scripts into
tflite_benchmark_model project.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Mar 2019 07:25:44 +0000 (16:25 +0900)]
Allow multiple TensorFlowLite package request (#4771)
The current implementation does not work in the presence of multiple
nnfw_find_package(TensorFlowLite ...) invocations.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
김수진/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 18 Mar 2019 06:42:37 +0000 (15:42 +0900)]
[neurun] Enable PReLU, TransposeConv, SQRT ops (#4733)
This commit enables `PReLU`, `TransposeConv`, `SQRT` ops for `acl_cl`.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Mar 2019 05:08:11 +0000 (14:08 +0900)]
Fix format in contrib (#4762)
Fix format in contrib except benchmark_acl
(bencmark_acl is ported codes from arm compute)
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 18 Mar 2019 04:59:50 +0000 (13:59 +0900)]
[neurun] Fix NNAPI frontend directory structure (#4769)
Reduce redundant frontend directory.
From `neurun/frontend/nnapi/frontend` to `neurun/frontend/nnapi`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Mar 2019 04:55:22 +0000 (13:55 +0900)]
Introduce TensorFlow Lite package (#4766)
This commit extracts TensorFlow Lite build script from
"external/CMakeLists.txt", and makes is as a separate TensorFlow
Lite package.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 18 Mar 2019 04:46:46 +0000 (13:46 +0900)]
Use regular expression to check arm architecture (#4765)
Use regular arm* and aarch* to check arm architecture
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>