platform/core/ml/nnfw.git
5 years ago[neurun] Use IObject in PermuteLayer (#4005)
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 13 Dec 2018 04:02:25 +0000 (13:02 +0900)]
[neurun] Use IObject in PermuteLayer (#4005)

This commit uses `IObject` insted of `ITensor` in `PermuteLayer`, since we don't want to use `map/unmap` functions explicitly and to get any `ACL` dependencies.

Related : #3279
Part of : #3873

Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
5 years agoFix external tensorflow lite build cmake (#4007)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 13 Dec 2018 03:21:05 +0000 (12:21 +0900)]
Fix external tensorflow lite build cmake (#4007)

- Remove BUILD_TFLITE_BENCHMARK_MODEL check for profiling source code
- Fix typo

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoRename tensorflow lite interpreter test (#4009)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 13 Dec 2018 03:20:41 +0000 (12:20 +0900)]
Rename tensorflow lite interpreter test (#4009)

* Rename tensorflow lite interpreter test

Rename tensorflow lite interpreter test by comparing with result of tensorflow
: tf_test -> tflite_test

* Fix build option description

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoMove how to run guideline into readme file (#3987)
윤지영/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 13 Dec 2018 02:14:31 +0000 (11:14 +0900)]
Move how to run guideline into readme file (#3987)

This patch moves the how to run guideline into readme file

Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
5 years agoAdd argmax quant8 to skiplist (#4004)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 13 Dec 2018 01:43:53 +0000 (10:43 +0900)]
Add argmax quant8 to skiplist (#4004)

Add argmax quant8 to skiplist: test fail

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoMove libs/misc/profiling to libs/profiling (#3997)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 13 Dec 2018 01:35:01 +0000 (10:35 +0900)]
Move libs/misc/profiling to libs/profiling (#3997)

Move libs/misc/profiling to libs/profiling for profiling library
Update cmake to link profiling library
Update header path for profiling
Update cmake for tensorflow build to support build with BUILD_TFLITE_BENCHMARK_MODEL option

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoArgMinMax CL kernel for ArgMax and ArgMin (#3867)
PRAVEEN DORESWAMY NAIDU/SNAP /SRI-Bangalore/Staff Engineer/삼성전자 [Thu, 13 Dec 2018 00:58:06 +0000 (06:28 +0530)]
ArgMinMax CL kernel for ArgMax and ArgMin (#3867)

The changes introduce the argmax and argmin CL kernel in one cl file.

Signed-off-by: Praveen D N <praveen.dn@samsung.com>
5 years agoAdd gtest to test LogicalAnd op for 2D and 3D tensors (#3938)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 13 Dec 2018 00:53:26 +0000 (06:23 +0530)]
Add gtest to test LogicalAnd op for 2D and 3D tensors (#3938)

This patch will add gtests to test logicalAnd op for 2D and 3D tensors

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoAdd gtest to test LogicalOr for 2D and 3D tensors (#3939)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 13 Dec 2018 00:53:06 +0000 (06:23 +0530)]
Add gtest to test LogicalOr for 2D and 3D tensors (#3939)

This patch will add gtests to test logicalOr for 2d and 3d tensors

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoPass explicitly local workgroup size to enqueue in acl (#3996)
윤지영/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 13 Dec 2018 00:51:13 +0000 (09:51 +0900)]
Pass explicitly local workgroup size to enqueue in acl (#3996)

This patch allows the local workgroup size(_lws_hint) value pass to enqueue function

Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
5 years ago[neurun] Introduce Config.lst (#3999)
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 13 Dec 2018 00:17:01 +0000 (09:17 +0900)]
[neurun] Introduce Config.lst (#3999)

Config.lst is a macro list with fields "Name", "Type", "Default value".

Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
5 years ago[neurun] Fix wrong size calculation when permuting (#3998)
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 12 Dec 2018 11:16:02 +0000 (20:16 +0900)]
[neurun] Fix wrong size calculation when permuting (#3998)

This commit fixes wrong size calculation when permuting.
It should be data size not pointer size.

Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
5 years agoApply dim correction to neg operator (#3986)
윤지영/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 12 Dec 2018 10:42:03 +0000 (19:42 +0900)]
Apply dim correction to neg operator (#3986)

This allows to apply dim correction to Neg operator

Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
5 years ago[neurun] Change ifdef guard in Convert.h (#3990)
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 12 Dec 2018 10:13:08 +0000 (19:13 +0900)]
[neurun] Change ifdef guard in Convert.h (#3990)

This commit changes ifdef guard in `Convert.h` to correct.

Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
5 years ago[neurun] Use host memory in acl_cl (#3974)
Дилшоджон Умронхонович Пошшоев/AI Tools Lab /SRR/Engineer/삼성전자 [Wed, 12 Dec 2018 10:10:56 +0000 (13:10 +0300)]
[neurun] Use host memory in acl_cl (#3974)

Related issue:3735
Add CLTensor::setBuffer to be able to use host memory from the acl_cl

Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
5 years agoAdd GTests to test broadcast of LogicalAND op (#3950)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 09:02:19 +0000 (14:32 +0530)]
Add GTests to test broadcast of LogicalAND op (#3950)

This patch adds GTests to test broadcast of LogicalAND op.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoADD 1D, 2D, 3D, 4D GTests for RSQRT (#3932)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 09:02:02 +0000 (14:32 +0530)]
ADD 1D, 2D, 3D, 4D GTests for RSQRT (#3932)

This patch adds 1D, 2D, 3D, 4D GTests for RSQRT

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoFix build bug by tflite library namespace (#3993)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 12 Dec 2018 08:56:23 +0000 (17:56 +0900)]
Fix build bug by tflite library namespace (#3993)

Fix build bug by tflite library namespace change

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoUpdate namespace and ifdef for tflite library (#3968)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 12 Dec 2018 08:39:07 +0000 (17:39 +0900)]
Update namespace and ifdef for tflite library (#3968)

* Update namespace and ifdef for tflite library

Make tflite/extend directory for ported codes from public tensorflow
Update namespace except extended directory
Update ifdef in header

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
* Rename tflite/extend to tflite/ext

* Fix format

* Update ifdef in ext

* Flatten namespace nnfw::tflite::interp to nnfw::tflite

* Add comment for endif

5 years agoAdd missing axises comment (#3979)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 08:38:08 +0000 (14:08 +0530)]
Add missing axises comment (#3979)

This patch will add missing axises comment.

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoFix wrong radius of LRN in NENormalization (#3977)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 08:27:54 +0000 (13:57 +0530)]
Fix wrong radius of LRN in NENormalization (#3977)

Fix wrong radius setting for LRN

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoMake TestGenerated file for each version (#3975)
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 12 Dec 2018 07:38:50 +0000 (16:38 +0900)]
Make TestGenerated file for each version (#3975)

* Make TestGenerated file for each version

On coverage build, build have failed since all_generated_tests.cpp is too
large. This patch will generate a test file for each NNAPI version.

Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* Extract common code from TestGenerated_XX.cpp

- Extract common code from TestGenerated_XX.cpp

5 years agoSupport boolean input param in tflite_run (#3958)
윤지영/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 12 Dec 2018 05:13:21 +0000 (14:13 +0900)]
Support boolean input param in tflite_run (#3958)

This patch assigns random boolean value to boolean input param.

Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
5 years agoAdd gtest to test Exp op for 3D/4D tesnors (#3797)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 04:03:42 +0000 (09:33 +0530)]
Add gtest to test Exp op for 3D/4D tesnors (#3797)

This patch will add gtest to test Exp op for tesnors having
dimensions other than 4

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoAdd CL Kernel calls for LogicalAnd op (#3970)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 02:00:35 +0000 (07:30 +0530)]
Add CL Kernel calls for LogicalAnd op (#3970)

This patch adds CL Kernel calls for LogicalAnd op

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoRemove setting of false to apply_dim_correction in compilation,cc for Broadcast suppo...
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 01:59:53 +0000 (07:29 +0530)]
Remove setting of false to apply_dim_correction in compilation,cc for Broadcast support ops (#3976)

This patch removes setting of false to apply_dim_correction in compilation,cc for Broadcast support ops.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoAdd GTests to test broadcast of LogicalOR (#3949)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 01:52:17 +0000 (07:22 +0530)]
Add GTests to test broadcast of LogicalOR (#3949)

This patch adds GTests to test broadcast of LogicalOR.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years ago Redefine REDUCE_MIN in NeuralNetworkEx.h (#3925)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 01:14:51 +0000 (06:44 +0530)]
Redefine REDUCE_MIN in NeuralNetworkEx.h (#3925)

This patch redefines REDUCE_MIN in NeuralNetworkEx.h
    explaining it supports INT32 also
    Reference tf.reduce_min
    input_tensor: The tensor to reduce. Should have numeric type.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years ago Redefine REDUCE_MIN in NeuralNetworkEx.h (#3926)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 01:14:03 +0000 (06:44 +0530)]
Redefine REDUCE_MIN in NeuralNetworkEx.h (#3926)

This patch redefines REDUCE_MIN in NeuralNetworkEx.h
    explaining it supports INT32 also
    Reference tf.reduce_max
    ```
    input_tensor: The tensor to reduce. Should have numeric type.
    ```

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoAdd gtest to test neg op for int datatype (#3930)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 01:13:43 +0000 (06:43 +0530)]
Add gtest to test neg op for int datatype (#3930)

This patch will add gtest to test neg for int datatype

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoCorrection in GTests of NotEqual_Ex with truth value as 0xff (255) (#3929)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 12 Dec 2018 01:11:20 +0000 (06:41 +0530)]
Correction in GTests of NotEqual_Ex with truth value as 0xff (255) (#3929)

This patch corrects GTests of NotEqual_Ex with truth value as 0xff (255)
Related issue: #3823 #3851

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoFix Typo in internal/op/* files (#3973)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Tue, 11 Dec 2018 07:04:53 +0000 (12:34 +0530)]
Fix Typo in internal/op/* files (#3973)

This patch fixes typo issues in internal/op/*.h files
changes:
```defualt -> default```

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoFix typo on SimpleArgMinMax.cc (#3972)
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Tue, 11 Dec 2018 06:15:18 +0000 (15:15 +0900)]
Fix typo on SimpleArgMinMax.cc (#3972)

fix typo : defualy -> default

Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
5 years agoUse TFLiteConverter instead of deprecated API in tensorflow_model_freezer (#3917)
윤지영/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 11 Dec 2018 05:15:19 +0000 (14:15 +0900)]
Use TFLiteConverter instead of deprecated API in tensorflow_model_freezer (#3917)

Add prerequires section in readme
Use tflite v1.12 tool to support the extended operators
Fix some tc errors

Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
5 years agoAdd sample py example for some operator (#3919)
윤지영/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 11 Dec 2018 04:35:50 +0000 (13:35 +0900)]
Add sample py example for some operator (#3919)

Added operators:
ARGMAX, ARGMIN, LOGICAL_AND, LOGICAL_NOT, LOGICAL_OR, STACK, UNSTACK

Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
5 years agoUpdate tflite library name in cmake (#3967)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Tue, 11 Dec 2018 02:41:31 +0000 (11:41 +0900)]
Update tflite library name in cmake (#3967)

Update tflite library name in cmake to nnfw_lib_tflite

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years ago[neurun] Remove default values in View.h (#3966)
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 11 Dec 2018 02:08:18 +0000 (11:08 +0900)]
[neurun] Remove default values in View.h (#3966)

This commit removes default values of `feature_index_to_byte_offset`'s parameter in `View.h`.

Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
5 years agoUpdate namespace of nnapi library (#3965)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Tue, 11 Dec 2018 00:39:06 +0000 (09:39 +0900)]
Update namespace of nnapi library (#3965)

Update namespace of nnapi library: nnfw::support::nnapi to nnfw::nnapi
Update header ifdef

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years ago[neurun] Fix wrong parameter in nchw View.h (#3963)
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Mon, 10 Dec 2018 11:26:29 +0000 (20:26 +0900)]
[neurun] Fix wrong parameter in nchw View.h (#3963)

This commit fixes wrong parameter for 3 rank in nchw `View.h`.

Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
5 years agoRename libs/util to libs/misc and update cmake (#3962)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 10 Dec 2018 11:23:19 +0000 (20:23 +0900)]
Rename libs/util to libs/misc and update cmake (#3962)

Rename libs/util to libs/misc
Use static build only

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoResolve inconsistencies in the boolean data values (#3957)
윤지영/동작제어Lab(SR)/Engineer/삼성전자 [Mon, 10 Dec 2018 08:26:57 +0000 (17:26 +0900)]
Resolve inconsistencies in the boolean data values (#3957)

* Resolve inconsistencies in the boolean data values

Since the data is stored in the index value of the input/output tensors,
it is modified to use the index value when it accesses the data.

Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
* Use meaningful variable names

Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
5 years agoMove libs/support/tflite to libs/tflite (#3920)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 10 Dec 2018 06:40:20 +0000 (15:40 +0900)]
Move libs/support/tflite to libs/tflite (#3920)

Move libs/support/tflite to libs/tflite
Update header include path

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoChange nnapi library name and update cmake (#3956)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 10 Dec 2018 06:32:43 +0000 (15:32 +0900)]
Change nnapi library name and update cmake (#3956)

Chane nnapi library name: nnfw_support_nnapi -> nnfw_lib_nnapi
Specify library type as static
Remove unused nnapi library link

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoRemove redundant .cl files regarding reduce operations (#3952)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 06:26:02 +0000 (11:56 +0530)]
Remove redundant .cl files regarding reduce operations (#3952)

This patch removes redundant .cl files regarding reduce operations
In issue: #3448
Since Unification of reduce kernel is merged,
      removal of existing implementation can be removed
Hence removing the same.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoUpdate skiplist for tizen (#3953)
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 10 Dec 2018 06:02:11 +0000 (15:02 +0900)]
Update skiplist for tizen (#3953)

Now, the skiplist of tizen is the same as linux's

Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
5 years agoAdd PReLU CL Kernel calls from PACL (#3916)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 05:13:42 +0000 (10:43 +0530)]
Add PReLU CL Kernel calls from PACL (#3916)

This patch adds PReLU CL Kernel calls from PACL.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoAdd 3D 4D Gtests for SquaredDifferenceEx op (#3933)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 05:07:17 +0000 (10:37 +0530)]
Add 3D 4D Gtests for SquaredDifferenceEx op (#3933)

This patch adds 3D 4D Gtests for SquaredDifferenceEx op.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoAdd GTests for SQRT_EX 3D 4D case (#3934)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:58:05 +0000 (07:28 +0530)]
Add GTests for SQRT_EX 3D 4D case (#3934)

This patch GTests for SQRT_EX 3D 4D case

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoAdd validations for Neg op in CLKernel.cpp (#3931)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:56:28 +0000 (07:26 +0530)]
Add validations for Neg op in CLKernel.cpp (#3931)

This patch will add validations req to be tested
before executing neg op

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoRemove unwanted apache license from RSQRT Gtest (#3937)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:55:18 +0000 (07:25 +0530)]
Remove unwanted apache license from RSQRT Gtest (#3937)

This patch removes unwanted apache license from RSQRT Gtest ```rsqrt_ex_float1.mod.py```

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoImplement LogicalOr op in PACL (#3944)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:53:41 +0000 (07:23 +0530)]
Implement LogicalOr op in PACL (#3944)

This patch will add implementation of LogicalOr op
in PACL

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoAdd gtest to test quant8 datatype for NotEqual Op (#3757)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:44:09 +0000 (07:14 +0530)]
Add gtest to test quant8 datatype for NotEqual Op (#3757)

This patch will add gtest to test quant8 datatype for NotEqual Op

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoAdd NNAPI delegation for LOGICAL_OR (#3904)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:42:50 +0000 (07:12 +0530)]
Add NNAPI delegation for LOGICAL_OR (#3904)

This patch adds NNAPI delegation for LOGICAL_OR.
Related issue: #3459

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoGeneric EmbeddingLookup kernel to handle tensors upto rank=4 (#3794)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:28:33 +0000 (06:58 +0530)]
Generic EmbeddingLookup kernel to handle tensors upto rank=4 (#3794)

This patch unifies the 3 different embedding lookup kernel into
one kernel so as to make a generic cl to support upto 4d tensors.

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoDefining default constructor in Pack and Unpack layer (#3907)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:28:15 +0000 (06:58 +0530)]
Defining default constructor in Pack and Unpack layer (#3907)

This patch will define default constructor in Pack and
Unpack Layers

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoGeneric HashTableLookup kernel to handle tensors upto rank=4 (#3798)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:27:55 +0000 (06:57 +0530)]
Generic HashTableLookup kernel to handle tensors upto rank=4 (#3798)

This patch unifies the 3 different hashtable lookup kernel into
one kernel so as to make a generic cl to support upto 4d tensors.

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoAdd gtest to test Argmax op for tensors of QASYMM8 dataype (#3802)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:26:14 +0000 (06:56 +0530)]
Add gtest to test Argmax op for tensors of QASYMM8 dataype (#3802)

This patch will add gtest for Argmax to test for tensors of
QASYMM8 dataype

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoRemove passing gtests from skiplist (#3908)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:23:18 +0000 (06:53 +0530)]
Remove passing gtests from skiplist (#3908)

This patch will remove all passing gtests from skiplist
- local_response_norm_float_4
- GeneratedTests.unpack*

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoAdd license comments in SimpleUnpackLayer* (#3905)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:22:13 +0000 (06:52 +0530)]
Add license comments in SimpleUnpackLayer* (#3905)

This patch will add license commnets in SimpleUnpackLayer.cc
andSimpleUnpackLayer.h

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoAdd gtest for split with negative axis. (#3915)
Devansh Bansal/SNAP /SRI-Bangalore/Engineer/삼성전자 [Mon, 10 Dec 2018 01:21:57 +0000 (06:51 +0530)]
Add gtest for split with negative axis. (#3915)

This patch adds a test case for split with negative axis.

Signed-off-by: b.devansh <b.devansh@samsung.com>
5 years agoRemove generated test cpp (#3921)
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Mon, 10 Dec 2018 00:59:59 +0000 (09:59 +0900)]
Remove generated test cpp (#3921)

Remove generated test cpp as these files are unused since #3874.

Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
5 years agoMove Padding module from `internal` to `util`. (#3882)
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Fri, 7 Dec 2018 07:48:57 +0000 (16:48 +0900)]
Move Padding module from `internal` to `util`. (#3882)

Move Padding module from `internal` to `util`.

Part of #3764

Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
5 years agoAdd NNAPI delegation for LOGICAL_AND (#3903)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Fri, 7 Dec 2018 07:48:06 +0000 (13:18 +0530)]
Add NNAPI delegation for LOGICAL_AND (#3903)

This patch adds NNAPI delegation for LOGICAL_AND.
Related issue: #3459

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoValidation check for addOperation (#3911)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 7 Dec 2018 07:46:20 +0000 (16:46 +0900)]
Validation check for addOperation (#3911)

- Check operation type is validate
- Add try-catch exception and return ANEURALNETWORKS_BAD_STATE when exception is raised
- Set return value when addOperationEx is not failed

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoAdd ExtendCMakeFunction module and add_subdirectories function (#3912)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 7 Dec 2018 07:15:23 +0000 (16:15 +0900)]
Add ExtendCMakeFunction module and add_subdirectories function (#3912)

- Add ExtendCMakeFunction module to simplify cmake
- Introduce add_subdirectories function to support adding all subdirectories at once.  (It's comes from nncc repo)
- Use add_subdirectories function in libs

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoUnify REDUCE_SUM and REDUCE_MAX operation in CL kernel (#3885)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Fri, 7 Dec 2018 07:06:03 +0000 (12:36 +0530)]
Unify REDUCE_SUM and REDUCE_MAX operation in CL kernel (#3885)

This patch unifies REDUCE_SUM and REDUCE_MAX operation in CL kernel
This is done to reduce code redundancy.
Related issue: #3771

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoAdd comments to define alpha input in internal/op/PReLU.cc (#3900)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Fri, 7 Dec 2018 06:44:13 +0000 (12:14 +0530)]
Add comments to define alpha input in internal/op/PReLU.cc (#3900)

This patch adds comments to define alpha input in internal/op/PReLU.cc.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoChange in kernel name for LogicalBinaryOp (#3913)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Fri, 7 Dec 2018 06:44:04 +0000 (12:14 +0530)]
Change in kernel name for LogicalBinaryOp (#3913)

This patch corrects the change in kernel name for LogicalBinaryOps (AND & OR)

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoAdd PReLU layer in PACL (#3901)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Fri, 7 Dec 2018 06:43:21 +0000 (12:13 +0530)]
Add PReLU layer in PACL (#3901)

This patch adds PReLU layer in PACL.
Issue: #3459 for Epic4

PReLU implemented as,
      f(x) = Relu(x) + (negative_alpha * Neg(x, activation=Relu))

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoMove libs/support/nnapi to libs/nnapi (#3895)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 7 Dec 2018 03:24:37 +0000 (12:24 +0900)]
Move libs/support/nnapi to libs/nnapi (#3895)

Move libs/support/nnapi to libs/nnapi
Update header include path

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoUpdate namespace and header define for cpp14 library (#3896)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 7 Dec 2018 02:13:20 +0000 (11:13 +0900)]
Update namespace and header define for cpp14 library (#3896)

Update namespace to nnfw::cpp14
Update define to __NNFW_CPP14_xxx__
Update usage of cpp14 library to updated namespace

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoRemove unused headers from Unpack op in internal/layers (#3906)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Fri, 7 Dec 2018 01:20:10 +0000 (06:50 +0530)]
Remove unused headers from Unpack op in internal/layers (#3906)

This patch will remove unused headers included in Unpack
related files in internal/layers

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoAdd GTests for PReLU op (#3899)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Fri, 7 Dec 2018 01:14:24 +0000 (06:44 +0530)]
Add GTests for PReLU op (#3899)

This patch adds GTest for PReLU op
Issue: #3459.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoAdding assertions in TransposeConv op. (#3887)
Devansh Bansal/SNAP /SRI-Bangalore/Engineer/삼성전자 [Fri, 7 Dec 2018 00:51:47 +0000 (06:21 +0530)]
Adding assertions in TransposeConv op. (#3887)

This patch adds necessary assertions in TransposeConv.

Signed-off-by: b.devansh <b.devansh@samsung.com>
5 years agoReturn error for not finished model compile (#3897)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 7 Dec 2018 00:50:43 +0000 (09:50 +0900)]
Return error for not finished model compile (#3897)

When NNAPI call try to compile model not finished, return BAD_STATE error

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoRemove template runtime (#3845)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 6 Dec 2018 09:08:34 +0000 (18:08 +0900)]
Remove template runtime (#3845)

Remove experimental template runtime in runtimes/template

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoImplementation of CUSTOM Unpack op (#3669)
Vishal Keshav/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 6 Dec 2018 08:27:47 +0000 (13:57 +0530)]
Implementation of CUSTOM Unpack op (#3669)

ACL implementation
Used CLSubtensor and CLReshape
CLPermute to overcome backward swizzle

Signed-off-by: Vishal Keshav <vishal.k1@samsung.com>
5 years agoDefine GENERATE_FILES for test generation (#3890)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 6 Dec 2018 08:11:38 +0000 (17:11 +0900)]
Define GENERATE_FILES for test generation (#3890)

Introduce GENERATE_FILES environment variable to skip running generate_test.sh
If you set GENERATE_FILES 0, makefile skip running generate_test.sh. Default is 1 (run generate_test.sh default)

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoMove include/nnfw/std (#3892)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 6 Dec 2018 08:09:41 +0000 (17:09 +0900)]
Move include/nnfw/std (#3892)

Move include/nnfw/std to libs/cpp14/
Update include path
Update cmake to include cpp14 library using interface

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years ago[neurun] Move Swizzle.h to acl backend dir (#3893)
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Dec 2018 08:08:54 +0000 (17:08 +0900)]
[neurun] Move Swizzle.h to acl backend dir (#3893)

Move Swizzle.h to acl backend dir since it is only used there.

Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
5 years ago[neurun] Remove backend/cpu dependancy from PermutateSource/Sink (#3889)
Дилшоджон Умронхонович Пошшоев/AI Tools Lab /SRR/Engineer/삼성전자 [Thu, 6 Dec 2018 07:25:33 +0000 (10:25 +0300)]
[neurun] Remove backend/cpu dependancy from PermutateSource/Sink (#3889)

Related issue: #3888
Remove backend/cpu dependancy from PermutateSource/Sink

Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
5 years agoFix wrong radius of LRN cl kernel (#3861)
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 6 Dec 2018 06:39:36 +0000 (15:39 +0900)]
Fix wrong radius of LRN cl kernel (#3861)

* Fix wrong radius setting of LRN cl kernel

Fix wrong radius setting for LRN

Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* Remove unused variables

5 years agoChange the order of Unpack Input parameter in NeuralNetworksEx.h (#3865)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 6 Dec 2018 06:07:22 +0000 (11:37 +0530)]
Change the order of Unpack Input parameter in NeuralNetworksEx.h (#3865)

This patch changes the order of Unpack Input parameter in NeuralNetworksEx.h
The num is seconds and axis is third param.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoRedefine PACK_EX in NeuralNetworksEx.h (#3864)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 6 Dec 2018 06:06:56 +0000 (11:36 +0530)]
Redefine PACK_EX in NeuralNetworksEx.h (#3864)

As per @hseok82-oh's words

>I found out runtime is updated to use additional parameter num and this PR will use this change. Please update comment PACK_EX in NeuralNetworksEx.h using another PR.

This patch redefines PACK_EX in NeuralNetworksEx.h.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years ago[neurun] Do backends as shared and update BackendManager to support plugin (#3793)
Дилшоджон Умронхонович Пошшоев/AI Tools Lab /SRR/Engineer/삼성전자 [Thu, 6 Dec 2018 06:06:23 +0000 (09:06 +0300)]
[neurun] Do backends as shared and update BackendManager to support plugin (#3793)

Related issue: #3703
This is the initial step of converting backends to plugins
Update BackendManager to creage a map using plugins

Signed-off-by: Poshshoev Dilshodzhon <d.poshshoev@samsung.com>
5 years agoRemove #failed comment from skiplist (#3878)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 6 Dec 2018 06:04:19 +0000 (11:34 +0530)]
Remove #failed comment from skiplist (#3878)

This pacth will remove #failed comment from skiplist

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoIntroduce enum for EQUALITY op. (#3853)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 6 Dec 2018 05:13:31 +0000 (10:43 +0530)]
Introduce enum for EQUALITY op. (#3853)

This patch introduces an enum for EQUALITY operations
like EQUAL, NOT_EQUAL.
This is related to #3851.
This is reduce code redundancy to a good extent in future.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoDisable nnapi_test of Custom/TensorflowMax (#3850)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 6 Dec 2018 04:37:23 +0000 (10:07 +0530)]
Disable nnapi_test of Custom/TensorflowMax (#3850)

This patch disables nnapi_test of Custom/TensorflowMax.
Since TensorFlowMax is builtin op as REDUCE_MAX from V1.12,
      we are removing the trace of TensorFlowMax as a custom op.

This is required for PRs #3849 #3846

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoRemove UPDATE_MODULE in makefile (#3891)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 6 Dec 2018 01:49:02 +0000 (10:49 +0900)]
Remove UPDATE_MODULE in makefile (#3891)

Remove UPDATE_MODULE environment variable to ignore submodule update: deprecated feature because we don't use submodule

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years ago[tfltool] Enable select operator for Pack op (#3817)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 6 Dec 2018 00:57:36 +0000 (06:27 +0530)]
[tfltool] Enable select operator for Pack op (#3817)

This patch will enable select operator tflitetool for pack op

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years ago[tfltool] Enable select operator for transpose op (#3811)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 6 Dec 2018 00:57:23 +0000 (06:27 +0530)]
[tfltool] Enable select operator for transpose op (#3811)

This patch will enable transpose op in tflitetool in select_operator.py

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years agoAdd unified CL Kernels for Binary LOGICAL OPS (AND & OR) (#3752)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Thu, 6 Dec 2018 00:56:50 +0000 (06:26 +0530)]
Add unified CL Kernels for Binary LOGICAL OPS (AND & OR) (#3752)

This issue tracks the implementation of unified CL Kernels for Binary LOGICAL OPS (AND & OR).

Unified CL kernel can be implemented.

This patch is an implementation of unified CL Kernel for Binary LOGICAL OPS (AND & OR).

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years ago[neurun] Fix wrong ifdef guard in Sink (#3883)
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Dec 2018 00:21:33 +0000 (09:21 +0900)]
[neurun] Fix wrong ifdef guard in Sink (#3883)

This commit fixes wrong ifdef guard in `Sink`.

Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
5 years ago Generate generated test cpp files on configuration (#3874)
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Dec 2018 00:21:21 +0000 (09:21 +0900)]
 Generate generated test cpp files on configuration (#3874)

* Rename generated dir to generated_remove

`generated` dir will be used for generated cpp files by
`generate_test.sh`. And `generated_remove` will be removed later.

Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
* Generate generated tests on demand

Generate generated tests on demand

Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
5 years agoFix include dir setting in cmake (#3881)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 5 Dec 2018 07:29:41 +0000 (16:29 +0900)]
Fix include dir setting in cmake (#3881)

Remove libs/util/include in NNFW_INCLUDE_DIR list
Fix cmake include for libs/util
Fix cmake link for neurun cpu kernel

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoFix header files format in libs (#3879)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 5 Dec 2018 07:29:18 +0000 (16:29 +0900)]
Fix header files format in libs (#3879)

Fix header files format in libs except comes from public tensorflow

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
5 years agoUnify REDUCE_MIN and REDUCE_MAX kernels (#3859)
Prasanna R/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 5 Dec 2018 07:21:00 +0000 (12:51 +0530)]
Unify REDUCE_MIN and REDUCE_MAX kernels (#3859)

This patch unifies reduce_min and reduce_max cl kernels
Related issue: #3771

Done to reduce code redundancy.

Signed-off-by: prasannar <prasanna.r@samsung.com>
5 years agoUse asPadstrideInfo utility to calc pad values in TransposeConv op (#3705)
Shubham Gupta/SNAP /SRI-Bangalore/Engineer/삼성전자 [Wed, 5 Dec 2018 07:14:35 +0000 (12:44 +0530)]
Use asPadstrideInfo utility to calc pad values in TransposeConv op (#3705)

This patch removes the internal calc of pad values during execution and uses asPadstrideInfo
utility to calc the same.

Signed-off-by: shubham <shub98.gupta@samsung.com>
5 years ago[neurun] Remove internal/nnapi (#3880)
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Dec 2018 07:13:44 +0000 (16:13 +0900)]
[neurun] Remove internal/nnapi (#3880)

Now that we removed the use of classes in `internal::nnapi::kernel`, we
can remove this directory.

Related PR : #3862

Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
5 years agoMove include/support (#3876)
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 5 Dec 2018 06:24:07 +0000 (15:24 +0900)]
Move include/support (#3876)

Move include/support to libs/support/include/support
Update cmake for include
Add .FORMATDENY to avoid format checking

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>