장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 8 May 2019 05:35:03 +0000 (14:35 +0900)]
Change ConcatLayer to support NHWC (#5165)
This commit ConcatLayer to support NHWC.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 8 May 2019 05:11:21 +0000 (14:11 +0900)]
Remove pack simple kernel (#5163)
Remove pack simple kernel in libs/ARMComputeEx
Use stack kernel in acl 19.02
Fix generated test to test 2D input (permutation issue)
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 8 May 2019 05:10:54 +0000 (14:10 +0900)]
Run mixed backend for test coverage (#5164)
Use mixed backend setting for test coverage run
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
Redo PermuteLayer run as template (#5107)
* Redo PermuteLayer run as template
PermuteLayer always assumes, that its data type is float. It was causing
an error during HE scheduler implementation. PR 5073
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Revert "Not supported yet" to NYI
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Merge ctors
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 8 May 2019 01:08:08 +0000 (10:08 +0900)]
Remove nnapi dependency in neurun (#5153)
Remove nnapi dependency in neurun core and backend code
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 7 May 2019 09:29:19 +0000 (18:29 +0900)]
Move Subgraph and SubgraphContext to upper dir (#5143)
Move Subgraph and SubgraphContext from `model::operation::` to
`model::`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 7 May 2019 09:27:29 +0000 (18:27 +0900)]
Support NHWC layout for ConstantInitializer (#5139)
This commit supports NHWC layout for ConstantInitializer.
- Support NHWC layout for ConstantInitializer.
- Support calculating the offset of cpu tensor up to 4 dimensions and in NHWC.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 7 May 2019 08:35:32 +0000 (17:35 +0900)]
Remove workaround mixed test (#5156)
Remove workaround for mixed backend test because CI scripts are updated
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 7 May 2019 04:59:24 +0000 (13:59 +0900)]
Internal padding type string converter (#5151)
Use internal padding type to convert padding type to string instead of NNAPI PaddingCode
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 7 May 2019 04:59:01 +0000 (13:59 +0900)]
Move resolving transpose convolution padding (#5150)
Move resolving transpose convolution operation's padding type: backend to frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 7 May 2019 04:44:44 +0000 (13:44 +0900)]
Merge Lowered and Subgraph section in Graph (#5122)
Subgraph is also part of LOWERED phase so we should not distinguish
these members.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 7 May 2019 04:44:32 +0000 (13:44 +0900)]
Revise Job class for ParallelExecutor (#5121)
- Introduce `Backend` as a member
- Intrdouce `reset` method for next execution
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 7 May 2019 04:44:19 +0000 (13:44 +0900)]
Introduce FullyConnectedReshapingLayer into ARMComputeEX (#5140)
This commit introduces FullyConnectedReshapingLayer into ARMComputeEx.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 7 May 2019 04:43:26 +0000 (13:43 +0900)]
Remove unpack simple kernel (#5135)
* Remove unpack simple kernel
Remove unpack simple kernel in libs/ARMComputeEx
Use unstack kernel in acl 19.02
Fix generated test to test 3D input (permutation issue)
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Update comment, remove change in pack
Fix Operands::exist to be able to remove operands (#5146)
If we remove operands after building a graph, existing method doesn't
work. Moreover, complexity of std::unordered_map::find is O(1), so we
don't loose much in time
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 3 May 2019 09:41:31 +0000 (18:41 +0900)]
Support NHWC layout for calculating the offset of acl tensor (#5137)
This commit supports NHWC layout for calculating the offset of acl tensor.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 3 May 2019 07:59:42 +0000 (16:59 +0900)]
Move fullyconnected activation resolve (#5132)
Move fullyconnected operation's activation type resolve: backend to frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 3 May 2019 07:53:13 +0000 (16:53 +0900)]
Move maxpool padding & activation resolve (#5120)
Move maxpool padding & activation type resolve: backend to frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 3 May 2019 07:52:24 +0000 (16:52 +0900)]
Move OperationIndex definition to Index.h (#5117)
As all index classes other than `OperationIndex` are all in `Index.h`,
this commit moves `OperationIndex` definition to `Index.h` and removes
`OperationIndex.h`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 3 May 2019 07:34:00 +0000 (16:34 +0900)]
Move elementwise op padding & activation resolve (#5129)
Move arithmetic elementwise operation padding & activation type resolve: backend to frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 3 May 2019 07:33:40 +0000 (16:33 +0900)]
Move conv2d padding & activation resolve (#5124)
Move convolution padding & activation type resolve: backend to frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 3 May 2019 07:33:20 +0000 (16:33 +0900)]
Move averagepool padding & activation resolve (#5123)
* Move averagepool padding & activation resolve
Move average pool padding & activation type resolve: backend to frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Fix constant check assertion and fix bug
- Fix to set explicit padding type
- Remove assert because asScalar has same assertion
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 3 May 2019 07:30:37 +0000 (16:30 +0900)]
Move L2 pooling padding & activation resolve (#5126)
Move l2 pooling padding & activation type resolve: backend to frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 3 May 2019 07:26:03 +0000 (16:26 +0900)]
Move depthwise-conv padding & activation resolve (#5119)
Move depthwise convolution padding & activation type resolve: backend to frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 3 May 2019 06:51:34 +0000 (15:51 +0900)]
Runtime test with mixed backends (#5125)
* Runtime test with mixed backends
Initial version of runtime test with mixed backends (Fixed backend
assignment)
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
* Remove logging
* Fix typo comments and add TODO
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 3 May 2019 06:24:01 +0000 (15:24 +0900)]
[dotdump] Show Subgraphs (#5128)
Show subgraphs using dot's subgraph(cluster) feature.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 3 May 2019 02:04:26 +0000 (11:04 +0900)]
Change in GTests of PReLU (#5110)
* Change in GTests of PReLU
Since prelu_ex_float_1 and prelu_ex_float_2 perform similar tests
This patch adds a replacement test
Signed-off-by: prasannar <prasanna.r@samsung.com>
* Add Quant8 test cases for PReLU op
This patch adds Quant8 test cases for PReLU op
Signed-off-by: prasannar <prasanna.r@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 3 May 2019 00:56:03 +0000 (09:56 +0900)]
Move BackendSet to graph::BackendSet (#5130)
Move `BackendSet` to `graph::BackendSet`. It was in
`graph::operand::BackendSet` which is misleading.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 2 May 2019 05:47:49 +0000 (14:47 +0900)]
[neurun] Introduce CopySource and CopySink (#5116)
This commit Introduces CopySource and CopySink.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 2 May 2019 05:36:22 +0000 (14:36 +0900)]
Revise header include in OperationFactory.cc (#5101)
- Change to use brackets for core header include
- Include NNAPIConverter.h to prepare to use activation and padding type converter
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이성재/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 2 May 2019 05:35:50 +0000 (14:35 +0900)]
Renewed README.md with 2019 plan (#5102)
* Renewd README.md with 2019 plan
- Update target platform
- Trim the sentence
- Add notification about _nncc_ integration
Signed-off-by: Sung-Jae Lee <sj925.lee@samsung.com>
* Fix typo.
- remove duplicated word.
Signed-off-by: Sung-Jae Lee <sj925.lee@samsung.com>
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 2 May 2019 04:42:05 +0000 (13:42 +0900)]
[neurun] Introduce getting layout into backend (#5030)
This commit introduces getting layout into backend tensors.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 2 May 2019 04:08:40 +0000 (13:08 +0900)]
Activation builder using internal activation type (#5099)
Introduce activation builder append method using neurun internal activation type
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 2 May 2019 03:57:26 +0000 (12:57 +0900)]
Update sample app document (#5104)
- Describe runtime with tensorfliw lite is not official support
- Tensorflow -> tensorflow lite
- Update chapter "Build a model": title, etc.
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 2 May 2019 01:14:22 +0000 (10:14 +0900)]
Remove unused RandomObject.h (#5108)
Remove unused RandomObject
Random value generation for tests are defined at libs/tflite/include/tflite/Diff.h
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이성재/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 30 Apr 2019 12:26:24 +0000 (21:26 +0900)]
Software requirement specification for 2019 (#5105)
* Software requirement specification for 2019
- Initial draft of SRS
Signed-off-by: Sung-Jae Lee <sj925.lee@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 30 Apr 2019 09:05:56 +0000 (18:05 +0900)]
Introduce SubgraphIndex (#5106)
`OperationIndex` was used for Subgraph indices, which can be confusing.
This commit introduces `SubgraphIndex` for `Subgraph`s.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
Add HowtoMakeSampleAppOnNnfw.md (#4129)
* Add HowtoMakeSampleAppOnNnfw.md
Related issue: 4128
Add tutorial about How to make a sample app on nnfw
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Fix "_nnfw_ API" as "NNAPI"
Fix "_nnfw_ API" as "NNAPI"
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
* Describe how to implement app with nnfw
Prev version was how to implement in the level of nnapi_delegate. This
version is how to create on top of nnapi_delegate
Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 30 Apr 2019 06:00:02 +0000 (15:00 +0900)]
Remove return in cpu backend kernel (#5096)
Remove unused bool type return in cpu backend kernel
Remove bool type return in helper function: always return true
Remove __wur (warning unused return) attribute in helper function
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 30 Apr 2019 05:07:21 +0000 (14:07 +0900)]
[dotdump] Remove Operation LowerInfo (#5087)
As now LowerInfo is per-subgraph(not per-operation), we need to remove
Operation LowerInfo and Backend label from `DotNodeInfo`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 30 Apr 2019 05:04:18 +0000 (14:04 +0900)]
Activation range calculator using internal type (#5093)
Introduce helper function in cpu backend for activation range calculation
Helper function use neurun internal activation type instead of NNAPI FuseCode type
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 30 Apr 2019 03:41:45 +0000 (12:41 +0900)]
Use nchw::View instead of Reader (#5084)
`Reader` is only for tensors of no paddings, while `nchw::View` can
cover padding cases.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 30 Apr 2019 03:41:23 +0000 (12:41 +0900)]
Move graph::Index to util::Index (#5094)
`graph::Index` is a class for indices in general, so it should not
belong to `graph` namespace.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 30 Apr 2019 03:41:13 +0000 (12:41 +0900)]
Remove Operation::InitParam (#5095)
This was just for creation of operations from NNAPI, so it is not used
any longer.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 30 Apr 2019 01:01:03 +0000 (10:01 +0900)]
Introduce activation and padding type converter (#5092)
Introduce nnapi to neurun internal activation and padding type converter
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 30 Apr 2019 00:50:16 +0000 (09:50 +0900)]
[neurun] Revise Executor classes to have a proper container for Subgraph (#5035)
* [neurun] Revise Executor classes to have a proper container for Subgraph
- ExecutorBase has no any container for Subgraph
- LinearExecutor has a vector container
- DataflowExecutor has a map container
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
* Extract Element as a file from Linear class
* Revise Executor classes with owned proper containers
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 30 Apr 2019 00:47:22 +0000 (09:47 +0900)]
Execute add operation in interpreter (#5088)
Implement naive interpreter execution module
Fix interpreter gtest's input/output value
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 29 Apr 2019 06:51:27 +0000 (15:51 +0900)]
Prepare interpreter add operation (#5060)
Add methods in ExecEnv to check and get tensor
Restrict ExecEnv to own tensors have buffer space
Assert for add operation
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 29 Apr 2019 04:10:00 +0000 (13:10 +0900)]
Introduce neurun activation and padding enum class (#5075)
Introduce neurun activation type enum: none, relu, relu1, relu6
Introduce neurun padding type enum: explicit, same, valid
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 26 Apr 2019 04:24:19 +0000 (13:24 +0900)]
[neurun] Enable Split op in neurun (#5068)
* [neurun] Enable Split op in neurun
This commit enables Split op in neurun for acl_cl
Signed-off-by: Seok NamKoong <sk.namkoong@samsung.com>
* Fix nnapi_delegate bug
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Fix build error by rebase
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 26 Apr 2019 00:48:56 +0000 (09:48 +0900)]
Enable custom oprations test (#5071)
Enable custom operations framework test again
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 25 Apr 2019 04:40:07 +0000 (13:40 +0900)]
[neurun] Rename classes in model::operand (#5062)
Rename all classes in `model::operand` Many changes are included in this
commit to avoid conflicts many times.
- `operand::Object` -> `Operand`
- `operand::Info` -> `OperandInfo`
- `operand::Data` -> `Data`
- `operand::DataType` -> `DataType`
- `operand::TypeInfo` -> `TypeInfo`
- `operand::Shape` -> `Shape`
- `operand::Index` -> `OperandIndex`
- `operand::IO::Index` -> `IOIndex`
- `operand::IndexMap` -> `OperandIndexMap`
- `operand::IndexSequence` -> `OperandIndexSequence`
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 25 Apr 2019 02:18:24 +0000 (11:18 +0900)]
Remove usage field in operand (#5066)
Remove usage field and related methods in internal operand
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 24 Apr 2019 08:08:11 +0000 (17:08 +0900)]
[neurun] Rename classes in model::operation (#5018)
* [neurun] Rename classes in model::operation
Rename the rest of classes in `model::operation` other than Operation
derivatives and subgraph modules. Many changes are included in this
commit to avoid conflicts many times.
- `operation::NodeVisitor` -> `OpVisitor`
- `operation::OperandConstraint` -> `OperandConstraint`
- `operation::IndexList` -> `OperationIndexList`
- `operation::Set` -> `OperationManager`
- `operation/Op.lst` -> `Op.lst`
- `operation/Node.Include.h` -> `Operation.Include.h`
Part of #4985
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
* Fix nits
* Rename OpXXX to OperationXXX
- OpVisitor.h => OperationVisitor.h
- Op.lst => Operations.lst
* Rebase fix
* Rename OperationManager to Operations
* Fix ifdef guards - OperandConstraint and Operation
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 24 Apr 2019 08:00:12 +0000 (17:00 +0900)]
Implement COPY kernel for PermuteLayer (#5052)
* Implement COPY kernel for PermuteLayer
PermuteLayer was supported NHWC to NCHW or vice versa. So it was not
able to handle just COPY case.
For #5044
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
* Handle paddings
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 24 Apr 2019 07:59:51 +0000 (16:59 +0900)]
Remove operand usage check in neurun internal (#5057)
Remove operand usage check assertion
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 24 Apr 2019 05:13:15 +0000 (14:13 +0900)]
Check operand usage in frontend (#5056)
Check operand usage by using operand usage vector in nnapi frontend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 24 Apr 2019 03:54:42 +0000 (12:54 +0900)]
Remove make tflite (#5055)
Remove make tflite command: not used anymore for svace and coverity check
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 24 Apr 2019 00:48:31 +0000 (09:48 +0900)]
Update howto document (#5047)
Update howto document's tflite model link
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 23 Apr 2019 10:37:34 +0000 (19:37 +0900)]
Rename hasData to isConstant (#5046)
Rename hasData to isConstant
It is used to check operand is constant
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 23 Apr 2019 10:32:22 +0000 (19:32 +0900)]
Fix LD_LIBRARY_PATH setting for model benchmark (#5051)
Fix LD_LIBRARY_PATH setting for model benchmark to use pure acl
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 23 Apr 2019 10:32:00 +0000 (19:32 +0900)]
Fix validation test (#5034)
Set operand's value for activation to ANEURALNETWORKS_FUSED_NONE
Test model's input operand numbers and size are also changed: [0, 1, 2] -> [0, 1]
By this, we can enable all validation test on acl-cl backend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 23 Apr 2019 08:31:17 +0000 (17:31 +0900)]
Manage operand stack and operation stack for interpretation (#5020)
- Introduce operand ready checker
- Introduce operation stack to save prepared operation
- Manage operand stack and operation stack and show log
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 23 Apr 2019 07:59:17 +0000 (16:59 +0900)]
Introduce operand usage vectorin frontend (#5039)
Collect operand usage info in frontend
Clear when model building is finished
Prepare removing nnapi-dependent usage info in neurun model operand
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 23 Apr 2019 07:58:13 +0000 (16:58 +0900)]
Introduce build option for tflite accuracy tool (#5048)
Introduce build option for tflite accuracy tool and set off as default
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 23 Apr 2019 07:09:14 +0000 (16:09 +0900)]
Change default runtime to neurun (#5042)
* Change default runtime to neurun
Change default runtime: pacl -> neurun
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Enable loggin runtime default
[batch_run] A batch execution tool used to run experiments (#4507)
* Related: #4405
The batch execution tool can be used to run experiments. It reads a neural network model from a file and a series of input images from a directory, runs each image through the network, and collect
statistics, such as execution time and accuracy.
Signed-off-by: Saulo A. Moraes <s.moraes@samsung.com>
* Remove boost_ext reference
Build script cleanup, remove unnecessary reference to boost_ext.
Signed-off-by: Saulo A. Moraes <s.moraes@samsung.com>
* Related: #4405
Code review fixes and tool rename from batch_run to tflite_accuracy.
Signed-off-by: Saulo A. Moraes <s.moraes@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 23 Apr 2019 04:20:20 +0000 (13:20 +0900)]
Remove operand usage getter methods (#5038)
- Remove unused isModelInput() method
- Remove usage() method. Use hasData() and model's input/output
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 22 Apr 2019 03:50:16 +0000 (12:50 +0900)]
Ignore try-run result on cross build (#5031)
* Ignore try-run result
Ignore try-run result on cross build when finding cmake threads package
Threads package is checked when finding pre-installed tflite package
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Fix comment
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 22 Apr 2019 02:25:55 +0000 (11:25 +0900)]
Change ARGMAX definition (#5021)
Chagne ARGMAX definition same with current NNAPI
- Allow one axis reduce
- Output rank = input rank - 1
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 19 Apr 2019 08:38:44 +0000 (17:38 +0900)]
Add printing info for filled memory in model_parser (#5028)
* Add printing info for filled memory in model_parser
Add printing info for filled memory such as weight in model_parser
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
* fix failed format-check
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 19 Apr 2019 05:40:57 +0000 (14:40 +0900)]
Use POSITION_INDEPENDENT_CODE property for tflite build (#5026)
Use cmake POSITION_INDEPENDENT_CODE property instead of setting gcc -fPIC compile option directly
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 18 Apr 2019 07:01:17 +0000 (16:01 +0900)]
[neurun] Apply code for merging node into subgraph to Graph (#5019)
Apply code for merging node into subgraph to Graph instead of 1 subrgaph
1 node code
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 18 Apr 2019 05:36:19 +0000 (14:36 +0900)]
Remove OBS_BUILD cmake define in Makefile (#5017)
Remove OBS_BUILD cmake define in cmake configure because OBS_BUILD is not used in cmake any more
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 18 Apr 2019 04:39:15 +0000 (13:39 +0900)]
Prepare operand stack in interpreter (#5015)
Prepare operand stack in interpreter to save prepared operands to use
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 18 Apr 2019 02:02:42 +0000 (11:02 +0900)]
[neurun] Apply SubgraphContext(1 subgraph 1 node) (#5014)
Apply SubgraphContext(1 subgraph 1 node) to Graph, Linear, Compiler and
{Dataflow|Linear}Executor
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 18 Apr 2019 00:57:48 +0000 (09:57 +0900)]
Fix CMake to use STATUS for message (#5016)
This will fix message() to use STATUS instead if "--" string
Signed-off-by: SaeHie Park <saehie.park@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 17 Apr 2019 07:15:33 +0000 (16:15 +0900)]
Interpreter ITensor Inherit backend ITensor (#5005)
Interpreter ITensor Inherit backend ITensor for more compatibility with other backends
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 17 Apr 2019 06:16:19 +0000 (15:16 +0900)]
[neurun] Rename operation::Node as Operation (#5013)
This is for reducing namespace operation.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
Fix a couple of noticeable clang static analyzer warnings (#4981)
* Fix a couple of noticeable clang static analyzer warnings
Fixed passing vector by value, lack of default switch case, dead code
Signed-off-by: Alexey Grebenkin <a.grebenkin@samsung.com>
* Correcting for PR comments
* Correcting for more PR comments
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 17 Apr 2019 01:12:48 +0000 (10:12 +0900)]
Fix define in cker header file (#5007)
Fix define __NNFW_CKER_CONCATENATION_H_ => __NNFW_CKER_CONCATENATION_H__
Fix define __NNFW_CKER_FXIED_POINT_H__ => __NNFW_CKER_GEMMLOWP_FIXED_POINT_H__
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 17 Apr 2019 01:10:19 +0000 (10:10 +0900)]
Fix OperationIndex build break from Subgraph (#5012)
It fixes build-break caused by Subgraph's use of model::operation::Index,
which is renamed to model::OperationIndex.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 17 Apr 2019 00:14:19 +0000 (09:14 +0900)]
Introduce interpreter class (#5006)
* Introduce interpreter class
Introduce interpreter class: run on assigned ExecEnv
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Fix define for header
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 16 Apr 2019 08:28:08 +0000 (17:28 +0900)]
[neurun] Introduce SubgraphContext class (#4995)
This class would be used as a context class which has Subgraphs instead
of SubgraphSet in Graph
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 16 Apr 2019 06:52:28 +0000 (15:52 +0900)]
Modify cpu average pool kernel (#5004)
Move loop invariant codes
Calculate filter count at once and compare
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 16 Apr 2019 02:28:08 +0000 (11:28 +0900)]
[neurun] Rename operation::Index as OperationIndex (#5002)
This is for reducing namespace operation.
Part of #4985
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 15 Apr 2019 23:44:31 +0000 (08:44 +0900)]
Remove axis conversion in libs/ARMComputeEx (#4992)
Remove axis conversion in libs/ARMComputeEx because conversion should done before configuration
Fix axis conversion in pack/unpack acl-cl extend kernel
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 15 Apr 2019 23:44:19 +0000 (08:44 +0900)]
Move gemmlowp code in cker (#4998)
Move gemmlowp code in cker into cker/gemmlowp
Add namespace for gemmlowp
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 15 Apr 2019 23:44:08 +0000 (08:44 +0900)]
[neurun] Change from SubgraphSet to SubgraphSequence (#4996)
Change the name SubgraphSet. Actually SubgraphSet has elements in order
so that SubgraphSequence is proper name.
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 15 Apr 2019 07:47:47 +0000 (16:47 +0900)]
Remove OpenCLEx codes (#4993)
- CLKernelLibraryEx includes OpenCL.h, not OpenCLEx.h
- There is no problem now
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 15 Apr 2019 03:55:03 +0000 (12:55 +0900)]
Interpreter use info structure in model (#4988)
Instead of using interpreter's own structure, use tensor info structure in model
Move comment and copy constructor
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 15 Apr 2019 03:49:13 +0000 (12:49 +0900)]
Remove script call in acl test (#4991)
Remove neurun acl-neon test script call in acl-cl test script
It will be called directly on CI infra.
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 11 Apr 2019 08:02:19 +0000 (17:02 +0900)]
Remove install backend kernel (#4979)
Remove install static kernel library because they are linked at each backend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 11 Apr 2019 07:36:05 +0000 (16:36 +0900)]
Move backend cmake setting into each backend (#4976)
Move backend cmake variable setting into each backend
Remove target check in backend
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 11 Apr 2019 05:38:16 +0000 (14:38 +0900)]
Add more coverage test (#4977)
* Add more coverage test
Call scripts for each backend test
Add acl_neon backend kernel test
Add dataflow executor test
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Fix script name typo
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 11 Apr 2019 04:12:05 +0000 (13:12 +0900)]
[neurun] add ExecManager::{setInput,setOutput} unit tests (#4972)
Add unittest for specifying unspecified dimension with setInput and setOutput.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 11 Apr 2019 03:22:56 +0000 (12:22 +0900)]
Remove cpu backend build dependency with tflite (#4975)
Remove header path and linking dependency with tflite in cpu backend build cmake
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 11 Apr 2019 01:23:37 +0000 (10:23 +0900)]
[ubun] Introduce Softmax microbenchmark (#4974)
With this commit, ubun now includes a Softmax microbenchmark based on
cker library.
Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 10 Apr 2019 08:04:14 +0000 (17:04 +0900)]
[neurun] Remove type and shape argument from Executor's setInput and setOutput (#4957)
IExecutor, ExecutorBase, ExecManager will have setInput and setOutput that don't require
`type` and `shape` parameters. In this case, `type` and `shape` from model will be used.
ExecManager test is also modified to use newly introduced setInput and setOutput.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 10 Apr 2019 07:21:19 +0000 (16:21 +0900)]
Remove unused util function in cpu backend (#4971)
Remove unused util function in cpu backend: type conversion to tflite internal type
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>