이상규/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 19 Sep 2018 02:18:45 +0000 (11:18 +0900)]
Remove copyright and license from externals/CMakeLists.txt (#2697)
Related Issue: #2662
We don't put license and copyright on CMakeLists.txt.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 19 Sep 2018 02:10:44 +0000 (11:10 +0900)]
[neurun] Introduce OperandPass (#2758)
`OperandPass` is a subclass of `Pass` that iterates for each
`operand::Object`. And subclasses of `OperandPass` must have
implementation of `callback` method.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
Implementation for LRN Layer (#2643)
This patch implements Local Response Normalization Layer in NNFW-PACL
Signed-off-by: shubham <shub98.gupta@samsung.com>
Enabling Logistics Op on PureACL NEON runtime (#2715)
This patch enables Logistics Op to run on NEON runtime PureACL.
Signed-off-by: b.devansh <b.devansh@samsung.com>
장지섭/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 18 Sep 2018 10:58:37 +0000 (19:58 +0900)]
[neurun] Apply memcpy for source/sink (#2754)
This commit applies memcpy for source/sink.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 18 Sep 2018 10:58:23 +0000 (19:58 +0900)]
[neurun] Introduce method `IBackendConfig::id` (#2753)
Introduce method `IBackendConfig::id()` so the backend itself holds its
string ID.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
장지섭/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 18 Sep 2018 10:58:06 +0000 (19:58 +0900)]
[neurun] Bring source/sink codes for 2D and 3D (#2752)
This commit bring source/sink codes for 2D and 3D from pureacl.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 18 Sep 2018 10:57:47 +0000 (19:57 +0900)]
[neurun] Introduce method IndexSet::replace (#2751)
Introduce operand::IndexSet::replace method which can update an element
in the set.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 18 Sep 2018 10:57:14 +0000 (19:57 +0900)]
[neurun] Introduce `operand::BackendSet` (#2750)
BackendSet is a class that is a set of Backends. It will be used from
`operand::LowerInfo` to express which backends need to allocate for a
operand.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 18 Sep 2018 10:57:04 +0000 (19:57 +0900)]
[neurun] Implement PermuteLayer kernel for cpu (#2749)
PermuteLayer basically copies data from input tensor to output tensor,
but if the tensor must be permuted it also do permute the tensor.
Currently NHWC to NCHW and NCHW to NHWC permutation is supported.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 18 Sep 2018 10:56:47 +0000 (19:56 +0900)]
[neurun] Introduce graph::Pass (#2747)
This commit introduces `graph::Pass` that can manipulate the graph
model.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 17 Sep 2018 10:58:38 +0000 (19:58 +0900)]
Install core libs into /usr/lib in tizen (#2722)
- For M2, libneuralnetworks.so should be installed on /usr/lib
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
우상정/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 17 Sep 2018 06:46:39 +0000 (15:46 +0900)]
Disable x86_64, i586 build on obs (#2716)
There is build break on obs SR. With this commit, no build is trigged
for x86_64 and i586 arch.
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
Signed-off-by: Sangjung woo <sangjung.woo@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 17 Sep 2018 06:36:14 +0000 (15:36 +0900)]
Remove scripts/command/change_gitmodule.sh (#2713)
This patch removes scripts/command/change_gitmodule.sh, since it is no
longer used, and updates its caller sites.
Same reason with #1085
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 17 Sep 2018 05:10:41 +0000 (14:10 +0900)]
Ignore acl dirty (#2712)
Ignore acl dirty to ignore build obj and binary
To ignore build directory, you should fix .gitignore in externals/acl
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 17 Sep 2018 02:02:58 +0000 (11:02 +0900)]
Fix comparison size_t with zero (#2710)
Fix comparison size_t with zero (always true)
Compare before casting from int to size_t
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 17 Sep 2018 00:35:48 +0000 (09:35 +0900)]
Download tensorflow source only (#2703)
Instead of using submodule, download source code for tensorflow source down & build
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 14 Sep 2018 10:48:16 +0000 (19:48 +0900)]
Fix NeuralNetworksLoadHelpers.h bug (#2707)
Fix NeuralNetworksLoadHelpers.h bug: same inline function name with NeuralNetworksShims.h in tensorflow lite
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 14 Sep 2018 09:20:50 +0000 (18:20 +0900)]
Remove priavte server name in framework test (#2701)
Remove priavte server name in framework test configuration
Instead we should pass by environment variable MODELFILE_SERVER
Find modelfile name for zip files
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
우상정/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 14 Sep 2018 09:16:13 +0000 (18:16 +0900)]
packaging: fix armcl package name (#2705)
armcl packages for arm and aarch64 have same package name so fix it.
Change-Id: I1104a35add8e0cfcd19625b4a79051e97e61df35
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Signed-off-by: Sangjung woo <sangjung.woo@samsung.com>
이상규/동작제어Lab(SR)/Principal Engineer/삼성전자 [Fri, 14 Sep 2018 07:24:42 +0000 (16:24 +0900)]
Change license MIT to Apache 2.0 for our ARMComputeEx sources (#2704)
Related Issue: #2663
For sources that we added under ARMComputeEx, the license is changed
to Apache 2.0, which is our company's default license for open source.
I kept ARM copyright because some code (e.g. PixelWiseDivision) uses
their code, and basically all code are based on ARM compute's skeleton.
StridedSlice and TopKV2 are implemented using TFLite kernels.
Thus the corresponding files have TensorFlow Authors copyright also.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 14 Sep 2018 06:52:54 +0000 (15:52 +0900)]
Use public vanilla arm compute library in submodule (#2696)
- change submodule url to vanilla arm compute library
- submodule commit : tag v18.05
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 14 Sep 2018 05:47:08 +0000 (14:47 +0900)]
Disable sub/broadcast model test (#2700)
`sub/broadcast` model test fails after update submodule to public acl
18.05. This commit disables sub test by default.
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 14 Sep 2018 05:06:08 +0000 (14:06 +0900)]
Environment variable to ignore md5 (#2695)
Environment variable to ignore md5 when model file is in cache directory
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이상규/동작제어Lab(SR)/Principal Engineer/삼성전자 [Fri, 14 Sep 2018 04:18:17 +0000 (13:18 +0900)]
Add MIT in tizen spec's license field (#2699)
Related Issue: #2650, #1977
MIT is specified since ARMComputeEx is introduced, which is under MIT license.
It is required to make happy SOLVe.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
이상규/동작제어Lab(SR)/Principal Engineer/삼성전자 [Fri, 14 Sep 2018 04:11:22 +0000 (13:11 +0900)]
Add MIT license in LICENSE (#2698)
Related Issue: #2650, #1977
MIT license, used in ARMComputeEx, is added in LICENSE.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 13 Sep 2018 02:43:45 +0000 (11:43 +0900)]
Environment variable setting for modelfile server (#2691)
- Support environment variable "MODELFILE_SERVER" for framework test modelfile download server
- Remove argument for mirror server
- Support zip file download
- Change variable name in config.sh to set download path: MODELFILE_SERVER_PATH
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 13 Sep 2018 02:27:46 +0000 (11:27 +0900)]
[neurun] Relocate Dumper from codegen to graph (#2690)
* [neurun] Relocate Dumper from codegen to graph
Related : #2686
This commit relocates Dumper from codegen to graph.
Dumper does not match the meaning of code generation and introducing other dumpers such as `.dot dumper` later are also same.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
* add dumper namespace
Enable Explicit Depth-wise Conv on NEON-PACL (#2680)
This patch will add support for Explicit Depthwise conv on NEON.
Signed-off-by: shubham <shub98.gupta@samsung.com>
Add pure_arm_compute layer for SquaredDifference (#2570)
This patch adds the pure_arm_compute layer for SquaredDifference.
Signed-off-by: prasannar <prasanna.r@samusng.com>
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 13 Sep 2018 02:22:51 +0000 (11:22 +0900)]
Add failed tests on vanilla acl submodule (#2694)
Some tests will fail after using vanilla acl submodule. To avoid CI
failure, add those tests into skip list.
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 13 Sep 2018 01:48:39 +0000 (10:48 +0900)]
Fix eigen package checking (#2693)
Fix eigen package checking in external
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 12 Sep 2018 10:36:17 +0000 (19:36 +0900)]
Change externals download URL (#2689)
Change externals download URL from internal mirror to github
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 12 Sep 2018 09:36:01 +0000 (18:36 +0900)]
[Contrib] OpenCL experiment code (#2685)
This code checks if devices in a GPU use same memory address.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 12 Sep 2018 09:35:45 +0000 (18:35 +0900)]
Fix gbs build fail : add python to buildrequire (#2687)
- Fix gbs build fail : add python to buildrequire
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 12 Sep 2018 09:23:18 +0000 (18:23 +0900)]
[neurun] Unify duplicate DataType definition (#2671)
This commit unifies duplicate DataType definition.
It can remove `cpu specific` includes for acl_cl kernel.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
Enabling Tanh on pureACL NEON runtime (#2674)
This patch enables the TANH operation on pureACL NEON runtime.
Signed-off-by: prasannar <prasanna.r@samusng.com>
Enable Squeeze on pureACL on NEON runtime (#2682)
This patch will add Squueze op for Neon runtime in pureACL
Signed-off-by: shubham <shub98.gupta@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 12 Sep 2018 07:43:57 +0000 (16:43 +0900)]
Fix externals CMake (#2683)
Fix externals CMake
- Remove unused include directories
- Remove unused packages on arm cross build
- Change eigen library name
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 12 Sep 2018 06:29:22 +0000 (15:29 +0900)]
[neurun] Introduce Permute operation (#2670)
Introduce Permute operation which is for tensor permutation. This
operation is not one of NNAPI operations but inserted internally in the
graph to convert different tensor layouts.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 12 Sep 2018 05:46:30 +0000 (14:46 +0900)]
Remove git submodule and use download (#2678)
Instead of using git submodule, use download for externals
(except tensorflow and acl)
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이상규/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 12 Sep 2018 05:43:56 +0000 (14:43 +0900)]
Fix copyright in ARMComputeEx (CLReductionMean.h, fixed_point.h) (#2665)
Related Issue: #2663
- Add our copyright since we added `CLReductionMean` kernel.
- Add our copyright since `fixed_point.h` (not existing since ACL18.08) was modified
from ACL18.x.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 12 Sep 2018 05:39:48 +0000 (14:39 +0900)]
Use CLKernellibraryEx on our own kernels (#2675)
* Use CLKernellibraryEx on our own kernels
- Change CLKernellibrary to CLKernellibraryEx
- add initilaization of CLKernellibraryEx before creating kernel
FIXME: please suggest better way
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* CLKernellibraryEx initialization on Compilation_finish
- Remove initialization on kernel's configure step
- Add initialization on Compilation_finish
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 12 Sep 2018 04:26:44 +0000 (13:26 +0900)]
Make resolve_includes.py work for Python 2 and 3 (#2681)
Previously lines with `print` work only on Python 2. Modified for Python 3.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
Implementing L2Normalization operation in NNAPI (#2642)
This commit implements L2Normalization operation in pure_arm_compute runtime.
Signed-off-by: prasannar <prasanna.r@samusng.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 11 Sep 2018 10:37:42 +0000 (19:37 +0900)]
[neurun] Fix Index casting (#2666)
* [neurun] Fix Index casting
Before Index was used `int32_t` but now it is `uint32_t`. However in
some code the casting is not changed accordingly. This commit changes
those unchanged code and fix misleading comments.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
* Return BAD_DATA instead of assertion
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Tue, 11 Sep 2018 09:24:45 +0000 (18:24 +0900)]
Use include instead of find_package for module (#2667)
Define nnfw_include to use include instead of find_package
Change ExternalProjectTools, ExternelSourceTools, and OptionTools to use nnfw_include (same with nncc)
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
Enable FLOOR on pureacl NEON runtime (#2664)
This patch enables the ```Floor``` operation on pureACL ```NEON``` runtime.
Signed-off-by: prasannar <prasanna.r@samusng.com>
Introduce HashtableLookup op in PACL runtime (#2655)
This commit introduces ANEURALNETWORKS_HASHTABLE_LOOKUP in PACL.
Related issue: #2654
Signed-off-by: Siva Sai <siva.sai@samsung.com>
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 11 Sep 2018 08:25:46 +0000 (17:25 +0900)]
[neurun] Update operand use/def info when inserting operation (#2660)
* [neurun] Update operand use/def info when inserting operation
Close : #2619
This commit updates operand use/def info when inserting operation)
- Update Use/Def info while inserting operation
- Do not allow inserting while model building
- Update TCs
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
* Seperate block for updating use/def and fix format break
Removing GeneratedTests.l2_pool_float* from the skiplist (#2648)
With #2587 and #2633 the l2_pool_float* cases passed. Removing from skip list.
Signed-off-by: Praveen D N <praveen.dn@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Tue, 11 Sep 2018 07:39:27 +0000 (16:39 +0900)]
Change gtest download cmake script (#2659)
Change gtest download cmake script simliar with nncc
Prepare supporting public server and mirror server download
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김용섭/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 11 Sep 2018 06:41:39 +0000 (15:41 +0900)]
[Generated Tests] Update tests for embedding lookup (#2657)
* [Generated Tests] Update tests for embedding lookup
- Remove embedding_lookup in skip lists for pure-acl
- Add embedding_lookup_2d_nnfw for testing 2d tensor
- Add embedding_lookup_4d_nnfw for 4d tensor
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
* Append embed_xd_nnfw tests to neuron list
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Tue, 11 Sep 2018 06:35:04 +0000 (15:35 +0900)]
arm_compute extended library for nnfw (#2653)
* arm_compute extented library for nnfw
1. Add libarm_compute_ex.so into nnfw
- Add CLKernelLibraryEx class : For newly added ops(e.g. cast,
gather, topk2 ..), CLKernelLibraryEx will be used on next patch.
2. Generate clembed files
- now, generate embeded_kernel by default
- add python script
- add generated files into gitignore list
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
* Add samsung license / update gitignore
- Add samsung license into CLKernellibraryEx.h
- Update gitignore for embeded files and comments
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 11 Sep 2018 06:23:48 +0000 (15:23 +0900)]
[neurun] Introduce set operators for LayoutSet (#2658)
Introduce Union, Intersect and Minus operators to operand::LayoutSet.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 11 Sep 2018 06:23:10 +0000 (15:23 +0900)]
[neurun] Add assert condition to insertOperation (#2639)
The node to be added in `Graph::insertOperation` must not have any
inputs.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
김용섭/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 11 Sep 2018 04:28:31 +0000 (13:28 +0900)]
Implement EMBEDDING_LOOKUP op in pure_acl on cpu (#2647)
* Implement EMBEDDING_LOOKUP op in pure_acl on cpu
Implement EMBEDDING_LOOKUP op in pure_acl on cpu
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
* Apply comments
- For 1d tensor, asTensorShape(shape, true)
- Use output_info for offset
* Change output to values
이상규/동작제어Lab(SR)/Principal Engineer/삼성전자 [Tue, 11 Sep 2018 04:14:06 +0000 (13:14 +0900)]
Add run_benchmark_tflite_model.sh (#2640)
Related Issue: #2544
`run_benchmark_tflite_model.sh` will run `tflite_benchmark_model` tool
with `run_benchmark_tflite_model.in` as input.
`run_benchmark_tflite_model.in` can have several lines.
Each line is a pair of (model name, its parameters for `tflite_benchmark_model`).
Its delimeter is `\t`.
For example, see `run_benchmark_tflite_mode.in`.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
이상규/동작제어Lab(SR)/Principal Engineer/삼성전자 [Tue, 11 Sep 2018 01:15:36 +0000 (10:15 +0900)]
Add our license under runtimes/pure_arm_compute (#2649)
Related Issue: #2644
Add our licenses for all *.[cpp,cc,h] under runtimse/pure_arm_compute
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Tue, 11 Sep 2018 01:13:09 +0000 (10:13 +0900)]
[neurun] Add license to source files (#2651)
This commit prepends license comment to the source files.
Resovle #2645
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Tue, 11 Sep 2018 01:02:14 +0000 (10:02 +0900)]
Parameter to set mirror sever (#2652)
Add new parameter in run_test.sh to set model file's mirror server
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 10 Sep 2018 04:21:19 +0000 (13:21 +0900)]
Set umask for temp file generation (#2637)
Set umask and recover after generate temporary file to avoid security issue
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
Adding explicit padding support for L2Pool2D (#2633)
This commit enables L2Pool2D operation with explicit padding.
Related issue: #2562
Signed-off-by: Siva Sai <siva.sai@samsung.com>
이춘석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Mon, 10 Sep 2018 02:16:49 +0000 (11:16 +0900)]
Remove tensor.h deps of RSQRT, SQDiff (#2638)
To fix gbs build, remove tensor.h deps of RSQRT and SQDiff
Signed-off-by: Chunseok Lee <chunseok.lee@samsung.com>
이상규/동작제어Lab(SR)/Principal Engineer/삼성전자 [Mon, 10 Sep 2018 02:06:40 +0000 (11:06 +0900)]
BUILD_TYPE=Release is specified in docker_env for tflite_benchmark_model (#2641)
Previously, build_type was specified in CI's groovy script.
But I would like to set all environment in one place - docker env file.
Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Fri, 7 Sep 2018 04:01:00 +0000 (13:01 +0900)]
Revise framework test list (#2636)
- Change directory structure: test/{tflite op name}/{testname}
- Add new framework tests to support more operators
- Update test list for neurun and benchmark_op
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
김용섭/동작제어Lab(SR)/Engineer/삼성전자 [Fri, 7 Sep 2018 01:07:09 +0000 (10:07 +0900)]
Introduce EMBEDDING_LOOKUP op to pure_acl (#2631)
Introduce EMBEDDING_LOOKUP op to pure_acl
Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Fri, 7 Sep 2018 01:00:00 +0000 (10:00 +0900)]
[neurun] Implement Lowering operands (#2630)
This commit implements lowering for operands in `Graph::lower`.
operands lowering does the following:
- Normalize to 4D shape (rank greater than 4 does not supported)
- Collect info of operand layouts for def/use
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Sep 2018 10:24:06 +0000 (19:24 +0900)]
[neurun] Introduce asShape4D (#2626)
This commit introduces asShape4D function to convert from
`operand::Shape` to `operand::LowerInfo::Shape4D`. This is
normalization to 4D shape.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Sep 2018 10:23:56 +0000 (19:23 +0900)]
[neurun] Add LowerInfo to operand::Object (#2628)
Add LowerInfo to operand::Object as a member.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 6 Sep 2018 10:23:45 +0000 (19:23 +0900)]
Diff: initialize output tensor (#2627)
Initialize output tensor as zero
Need to RNN hidden state initialize
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Sep 2018 09:48:48 +0000 (18:48 +0900)]
[neurun] Introduce operand::LowerInfo (#2622)
This commit introduces operand::LowerInfo which contains normalized
4D shape and tensor layout.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 6 Sep 2018 09:23:52 +0000 (18:23 +0900)]
Add Transpose into nnapi_delegate (#2624)
This commit enables TF Lite to call Transpose operation of NNAPI.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Sep 2018 08:24:01 +0000 (17:24 +0900)]
[neurun] Add basic methods to LayoutSet (#2623)
Introduce `contains` and `size` method for user queries.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
Enabling L2Pool2D Operation in NNAPI (#2587)
* Enabling PAD operation in NNAPI
This commit enables PAD operation in pure_arm_compute runtime
Signed-off-by: Siva Sai <siva.sai@samsung.com>
* Enabling L2Pool2D Operation in NNAPI
This commit enables L2Pool2D (Implicit Padding) operation in PACL runtime.
Related issue: #2562
Signed-off-by: Siva Sai <siva.sai@samsung.com>
장지섭/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Sep 2018 05:34:56 +0000 (14:34 +0900)]
Remove unnecessary CAST_CL from pure_arm_compute (#2617)
This commit removes unnecessary CAST_CL from pure_arm_compute.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
장지섭/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Sep 2018 05:33:29 +0000 (14:33 +0900)]
Implement SPACE_TO_DEPTH operation for CPU (#2614)
This commit implements SPACE_TO_DEPTH operation in pure_arm_compute for CPU.
- Add a layer that supports SPACE_TO_DEPTH for CPU.
- Add a function to get ::arm_compute::Coordinates by the rank.
- Implement SPACE_TO_DEPTH operation.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 6 Sep 2018 04:17:09 +0000 (13:17 +0900)]
Use ExternalData for zero length value (#2618)
Use ExternalData for object with no value (nullptr buffer from setOperandValue)
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Thu, 6 Sep 2018 01:33:55 +0000 (10:33 +0900)]
TFlite file generator for FLOOR, PAD, SQUEEZE, TRANSPOSE (#2609)
This python script generates four tflite files for FLOOR, PAD, SQUEEZE and TRANSPOSE respectively.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Sep 2018 01:33:22 +0000 (10:33 +0900)]
[neurun] Add IBackendConfig getter to Backend (#2611)
Add `IBackendConfig` getter `config()` method to `Backend`. We need to
keep it otherwise we do not have any way to call `getOperandLayout()`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Sep 2018 01:33:08 +0000 (10:33 +0900)]
[neurun] Introduce operand::LayoutSet (#2613)
LayoutSet is a set data structure for operand::Layout.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Thu, 6 Sep 2018 01:32:26 +0000 (10:32 +0900)]
[neurun] Move Source and Sink to `exec` dir (#2612)
Move `Source` and `Sink` to exec directory and change namespace from
`internal` to `exec`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Thu, 6 Sep 2018 01:02:22 +0000 (10:02 +0900)]
[Model Parse] Parsing exceptional case: tensor index (#2593)
Parsing exceptional case when tensor index is less than 0
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
장지섭/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 10:43:34 +0000 (19:43 +0900)]
Introduce SPACE_TO_DEPTH operation into pure_arm_compute (#2601)
This commit introduce SPACE_TO_DEPTH operation into pure_arm_compute.
Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 09:50:26 +0000 (18:50 +0900)]
[neurun] Change namespace of BackendXX to backend (#2603)
* [neurun] Change namespace of BackendXX to backend
Tshi commit changes namespace of BackendXX to backend.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
* Revert BackendResolver namespace
* Remove neurun::
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 5 Sep 2018 09:50:11 +0000 (18:50 +0900)]
Revert makfile to build pacl only (#2607)
Revert makfile to build pacl & test, not to build neurun
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 09:05:57 +0000 (18:05 +0900)]
[neurun] Support iteration for operand::Object (#2608)
Support const/non-const iteration for operand::Object.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
윤현식/동작제어Lab(SR)/Principal Engineer/삼성전자 [Wed, 5 Sep 2018 08:36:15 +0000 (17:36 +0900)]
[PureACL] Enable NEON to run Mobilenet (#2600)
This commit enables NEON to run Mobilenet by providing `depthwise_conv` and `reshape` for NEON.
Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
윤지영/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 08:32:28 +0000 (17:32 +0900)]
Add fullyconnected and mean tests using tflite (#2606)
This commit add fullyconnected and mean directory for testing these tflite files.
It will automatically download the files from npuarchive.mooo server and test it.
Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 08:24:12 +0000 (17:24 +0900)]
[neurun] Enable compile option Wall/Wextra/Werror for backend modules (#2604)
This commit enables compile option Wall/Wextra/Werror for backend modules.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 08:03:56 +0000 (17:03 +0900)]
[neurun] Use unordered_map instead of map (#2605)
Update `operand::Set` and `operation::Set` to use `std::unordered_map`
rather than `std::map`. Also update use key as Index instead of `int`.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
김수진/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 05:42:08 +0000 (14:42 +0900)]
[neurun] Remove unused includes for BackendResolver (#2596)
Related #2510
This commit removes unused includes for BackendResolver.
Signed-off-by: sjsujinkim <sjsujin.kim@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 5 Sep 2018 05:41:56 +0000 (14:41 +0900)]
Define constructor for profile context (#2597)
Define constructor for profile context to initialize field
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 5 Sep 2018 05:35:51 +0000 (14:35 +0900)]
Initialize feature shape (#2598)
Initialize feature shape by constructor to remove uninitialized field
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 05:06:58 +0000 (14:06 +0900)]
[neurun] Use Index as key for unordered_map (#2594)
This commit revises `unordered_map`s to use `graph::Index` as key
which used `int` as key because hash function was not exist.
Related PR : #2590
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 04:18:02 +0000 (13:18 +0900)]
[neurun] Introduce getOperandLayout (#2591)
Introduce `IBackendConfig::getOperandLayout` which returns
`operand::Layout` for each backend. Note that this assumes a backend
only support one type of operand layout.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
SquaredDifference nnapi delegate and Custom ops addition (#2569)
This patch adds changes for SquaredDifference in nnapi delegate
and also adds related CustomOps changes.
SquaredDifference kernel has been added.
Signed-off-by: prasannar <prasanna.r@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 04:15:24 +0000 (13:15 +0900)]
[neurun] Use Index as key for unordered_set (#2590)
This commit revises `unordered_set`s to use `graph::Index` as key
which used `int` as key because hash function was not exist.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
이한종/동작제어Lab(SR)/Engineer/삼성전자 [Wed, 5 Sep 2018 04:14:12 +0000 (13:14 +0900)]
[neurun] Do not use addTensorBuilder (#2573)
PR #2563 has introduced `IPlanBuilder::addTensorBuilder`. However
adding a call to `addTensorBuilder` for all nodes from Planner is an
overhead for development. So this commit revises to remove the calls.
However with this commit, `Linear::markTensor` method returns
`TensorBuilderSet` and passes it to `PlanBuilder::finalize`. Still
need to enhance this structure, so TODO comments are left.
Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
오형석/동작제어Lab(SR)/Staff Engineer/삼성전자 [Wed, 5 Sep 2018 03:15:55 +0000 (12:15 +0900)]
[Model Parser] Update tflite file API (#2580)
Update tflite file API: schema in tensorflow r1.10
Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>