platform/core/ml/nnfw.git
5 years ago[moco/ONNX] Apply opset version to operator and frontend (#3557)
남궁석/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 22 May 2019 08:19:29 +0000 (17:19 +0900)]
[moco/ONNX] Apply opset version to operator and frontend (#3557)

* [moco/ONNX] Apply opset version to operator and frontend

This commit will apply opset version to moco ONNX frontend and
ONNX operators.

Signed-off-by: Seok NamKoong <sk.namkoong@samsung.com>
* Repositioning REGISTER_OP_BUILDER

5 years ago[moco/tf/test] test case for MaxPool (#3555)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 22 May 2019 08:08:10 +0000 (17:08 +0900)]
[moco/tf/test] test case for MaxPool (#3555)

This adds a test case for MaxPool.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
5 years ago[hermes] Introduce Sink class (#3570)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 22 May 2019 06:27:53 +0000 (15:27 +0900)]
[hermes] Introduce Sink class (#3570)

* [hermes] Introduce Sink class

This commit introduces Sink class which serves as a base class for all
log message consumers.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Update comment

5 years ago[hermes] Introduce MessageBus & MessageBuffer class (#3565)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 22 May 2019 05:47:30 +0000 (14:47 +0900)]
[hermes] Introduce MessageBus & MessageBuffer class (#3565)

* [hermes] Introduce MessageBus & MessageBuffer class

This commit introduces MessageBus & MessageBuffer classes which logging
endpoint implementation requires.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Update MessageBuffer comment (typo fix and add some details)

5 years ago[moco/tf] use make_unique instead of new for MaxPool2D (#3572)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 22 May 2019 02:13:35 +0000 (11:13 +0900)]
[moco/tf] use make_unique instead of new for MaxPool2D (#3572)

This changes new to make_unique for MaxPool2D, which was omitted in pr #3569

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
5 years ago[moco] update relu comment (#3568)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 22 May 2019 01:13:22 +0000 (10:13 +0900)]
[moco] update relu comment (#3568)

* [moco] update relu comment

This will update Relu comment about input update

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* update

5 years ago[moco] use make_unique (#3569)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 22 May 2019 01:12:46 +0000 (10:12 +0900)]
[moco] use make_unique (#3569)

This will change to use make_unique instead of direct new

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[loco] Add TenosrFlow_MaxPool_Float_000 test (#3539)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 21 May 2019 10:46:18 +0000 (19:46 +0900)]
[loco] Add TenosrFlow_MaxPool_Float_000 test (#3539)

This commit adds TensorFlow_MaxPool_Float_000 test as a loco unittest.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[moco] enable Relu_000 test (#3567)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 10:45:12 +0000 (19:45 +0900)]
[moco] enable Relu_000 test (#3567)

This will enable Relu_000 to do locomotive testing

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] Remove obsolete comment (#3566)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 10:44:52 +0000 (19:44 +0900)]
[moco] Remove obsolete comment (#3566)

This will remove obsolete comment in TF frontend

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] add and update comments tf/frontend (#3562)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 10:27:50 +0000 (19:27 +0900)]
[moco] add and update comments tf/frontend (#3562)

This will add and update comments for TF frontend

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco/tf] using static for local var inside validate() (#3564)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 09:58:15 +0000 (18:58 +0900)]
[moco/tf] using static for local var inside validate() (#3564)

Attrs names of MaxPool2D is now static.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
5 years ago[moco/ONNX] Fix wrong testcases for ONNX Identity operation (#3559)
남궁석/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 21 May 2019 09:55:45 +0000 (18:55 +0900)]
[moco/ONNX] Fix wrong testcases for ONNX Identity operation (#3559)

`moco` is using ONNX 1.4.1 and latest IR version of ONNX 1.4.1 is 4,
and opset version of IR_v4 is 9.
This commit will correct this error

Signed-off-by: Seok NamKoong <sk.namkoong@samsung.com>
5 years ago[loco] Introduce NodeMixin class (#3556)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 21 May 2019 09:30:04 +0000 (18:30 +0900)]
[loco] Introduce NodeMixin class (#3556)

* [loco] Introduce NodeMixin class

This commit extracts common shape getter/setter methods as a seperate
mix-in class.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Fix a typo

5 years ago[hermes] Introduce Message class (#3558)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 21 May 2019 09:06:29 +0000 (18:06 +0900)]
[hermes] Introduce Message class (#3558)

This commit adds hermes Message class and related tests.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[moco] add GraphUpdate test (#3560)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 08:32:09 +0000 (17:32 +0900)]
[moco] add GraphUpdate test (#3560)

* [moco] add GraphUpdate test

This will add unit test for GraphUpdate

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* use anonymous namespace

5 years ago[moco/tf] adding MaxPool2D and test (#3554)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 07:52:46 +0000 (16:52 +0900)]
[moco/tf] adding MaxPool2D and test (#3554)

This commit adds MaxPool2D and its test code.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
5 years ago[loco] Simplify test code using overloaded eq operator (#3552)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 21 May 2019 07:32:47 +0000 (16:32 +0900)]
[loco] Simplify test code using overloaded eq operator (#3552)

This commit rewrites test code which invokes value explicitly to use
overloaded eq operator (for simplicity).

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[moco/ONNX] Introduce opset_version variable (#3550)
남궁석/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 21 May 2019 05:56:30 +0000 (14:56 +0900)]
[moco/ONNX] Introduce opset_version variable (#3550)

* [moco/ONNX] Introduce opset_version variable

This commit will introduce opset_version variable which will be used to divide operations
according to proper opset version

Signed-off-by: Seok NamKoong <sk.namkoong@samsung.com>
* apply comments

5 years ago[moco] remove unused node_names (#3553)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 05:46:17 +0000 (14:46 +0900)]
[moco] remove unused node_names (#3553)

This will remove 'not-used-anymore' node_names

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[hermes] Introduce Severity class (#3551)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 21 May 2019 04:58:10 +0000 (13:58 +0900)]
[hermes] Introduce Severity class (#3551)

This commit introduces Severity class which denotes the severity of
each event.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[moco/tf] Converting functions for MaxPool2D (#3549)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 04:12:43 +0000 (13:12 +0900)]
[moco/tf] Converting functions for MaxPool2D (#3549)

This commit adds converting funtion for MaxPool2D.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
5 years ago[moco] use _table for name resolve (#3548)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 04:08:53 +0000 (13:08 +0900)]
[moco] use _table for name resolve (#3548)

This will change SymbolTable to use _table variable for name resolve

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[loco] Override eq operator over Dimension (#3545)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 21 May 2019 03:55:52 +0000 (12:55 +0900)]
[loco] Override eq operator over Dimension (#3545)

* [loco] Override eq operator over Dimension

This commit implements eq operator overriding over Dimension.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Use explicit constants for test

5 years agoAdd hermes project (#3542)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 21 May 2019 02:23:21 +0000 (11:23 +0900)]
Add hermes project (#3542)

This commit creates a directory for hermes logging framework with basic
template. Note that this commit just creates a directory and thus does
not include any actual implementation.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[moco/ONNX] Change parameter of convert_graph (#3547)
남궁석/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 21 May 2019 02:11:20 +0000 (11:11 +0900)]
[moco/ONNX] Change parameter of convert_graph (#3547)

For checking opset version of ONNX, onnx ModelProto is needed.
But until now, convert_graph has GraphProto, not ModelProto.
This commit will pass ModelProto parameter instead of GraphProto

Signed-off-by: Seok NamKoong <sk.namkoong@samsung.com>
5 years ago[loco] Introduce shape helper method (#3544)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 21 May 2019 02:11:06 +0000 (11:11 +0900)]
[loco] Introduce shape helper method (#3544)

This commit extends Push and Pull node with shape helper method.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[moco] finalize node conversion inherit classes (#3546)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 02:10:01 +0000 (11:10 +0900)]
[moco] finalize node conversion inherit classes (#3546)

This will finalize GraphBuilder and GraphUpdate inherited classes

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] use GraphUpdate to set input (#3540)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 21 May 2019 02:08:26 +0000 (11:08 +0900)]
[moco] use GraphUpdate to set input (#3540)

This will change to use GraphUpdate to set inputs of the node

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] Implement UpdateQueue::enroll() (#3538)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 20 May 2019 22:17:14 +0000 (07:17 +0900)]
[moco] Implement UpdateQueue::enroll() (#3538)

This will add UpdateQueue::enroll() implementation

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] Add UpdateQueue member (#3537)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 20 May 2019 09:29:12 +0000 (18:29 +0900)]
[moco] Add UpdateQueue member (#3537)

This will add UpdateQueue attribute in GraphBuilderContext

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] Add two GraphUpdate inherits (#3536)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 20 May 2019 08:05:42 +0000 (17:05 +0900)]
[moco] Add two GraphUpdate inherits (#3536)

* [moco] Add two GraphUpdate inherits

This will add two GraphUpdate class inherits that needs graph update(input settings), Identity and Relu.

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* add const

5 years ago[locomotiv] Session's graph input/output size (#3531)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 20 May 2019 07:38:23 +0000 (16:38 +0900)]
[locomotiv] Session's graph input/output size (#3531)

* [locomotiv] Session's graph input/output size

This commit introduces Session::input_size() and Session::output_size()
which are getter for graph input and output size it possesses.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
* Make them const method

* Use const variable for magic number

5 years ago[moco] add const method type (#3535)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 20 May 2019 07:32:27 +0000 (16:32 +0900)]
[moco] add const method type (#3535)

This will add const for method using const SymbolTable

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] Introduce UpdateQueue (#3533)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 20 May 2019 06:55:00 +0000 (15:55 +0900)]
[moco] Introduce UpdateQueue (#3533)

This will introduce UpdageQueue class to store graph update objects

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] Introduce GraphUpdate interface (#3525)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 20 May 2019 04:59:30 +0000 (13:59 +0900)]
[moco] Introduce GraphUpdate interface (#3525)

* [moco] Introduce GraphUpdate interface

This will introduce GraphUpdate interface

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* add default destructor

* add const

5 years agoVerify ONNXSource 1.4.1 archive checksum (#3528)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 20 May 2019 04:23:14 +0000 (13:23 +0900)]
Verify ONNXSource 1.4.1 archive checksum (#3528)

This commit updates ONNXSource 1.4.1 package to verify its archive after
download.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years agoVerify ONNXSource 1.3.0 archive checksum (#3527)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 20 May 2019 04:02:19 +0000 (13:02 +0900)]
Verify ONNXSource 1.3.0 archive checksum (#3527)

This commit updates ONNXSource 1.3.0 package to verify its archive after
downloading.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[locomotiv] How to use locomotiv (#3524)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 20 May 2019 04:02:09 +0000 (13:02 +0900)]
[locomotiv] How to use locomotiv (#3524)

* [locomotiv] How to use locomotiv

This commit introduces documentation for how to use locomotiv.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
* Review feedback updated

5 years ago[loco] Introduce Feature Permuting Codec (#3517)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 20 May 2019 03:43:35 +0000 (12:43 +0900)]
[loco] Introduce Feature Permuting Codec (#3517)

* [loco] Introduce Feature Permuting Codec

This commit implements a basic feature codec based on axis permutation.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Update comments & add a missing assert message

* Check whether permutation axes are all distinct

* Fix a typo (Encoder -> Decoder)

5 years ago[loco] Introduce IR verification infrastructure (#3526)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 20 May 2019 03:42:27 +0000 (12:42 +0900)]
[loco] Introduce IR verification infrastructure (#3526)

This commit introduces basic verifier infrastructure for loco.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years agoVerify downloaded archive (#3512)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 20 May 2019 03:20:56 +0000 (12:20 +0900)]
Verify downloaded archive (#3512)

With this commit, users are able to verify downloaded archive via
ExternalSource_Download with checksum.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[loco] Introduce all_nodes helper (#3522)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 20 May 2019 00:45:47 +0000 (09:45 +0900)]
[loco] Introduce all_nodes helper (#3522)

This commit introduces all_nodes helper which allows users to enumerate
all the nodes in a given loco graph.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[locomotiv] Support ConstGen execution (#3498)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 17 May 2019 05:08:23 +0000 (14:08 +0900)]
[locomotiv] Support ConstGen execution (#3498)

This commit starts to support ConstGen node execution.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
5 years ago[locomotiv] How to support new node execution (#3484)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Fri, 17 May 2019 05:05:14 +0000 (14:05 +0900)]
[locomotiv] How to support new node execution (#3484)

* [locomotiv] How to support new node execution

This commit introduces a guide how to support new loco node execution.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
* Review feedback updated

5 years agoRemove version-less ONNXSource package (#3518)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 17 May 2019 03:57:36 +0000 (12:57 +0900)]
Remove version-less ONNXSource package (#3518)

This commit removes version-less ONNXSource package to prevent possible
versioning issue.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years agoUse a specific ONNX version in ONNXProto (#3513)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 17 May 2019 02:22:43 +0000 (11:22 +0900)]
Use a specific ONNX version in ONNXProto (#3513)

This commit updates ONNXProto CMake package to use a specific version of
ONNX source code (1.3.0).

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[nnc] Use versioned ONNXSource package (#3515)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 17 May 2019 02:22:31 +0000 (11:22 +0900)]
[nnc] Use versioned ONNXSource package (#3515)

This commit updates nnc build script to find a specific version of
ONNXSource package.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[loco] Introduce asserts inside postorder_traversal (#3506)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 17 May 2019 02:19:02 +0000 (11:19 +0900)]
[loco] Introduce asserts inside postorder_traversal (#3506)

This commit introduces several asserts inside postorder_traversal in
order to faciliate debugging.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[moco/tf] Show executed commands (#3510)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 17 May 2019 02:14:49 +0000 (11:14 +0900)]
[moco/tf] Show executed commands (#3510)

With this commit, runall test script records executed commands into the log file.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[locomotiv] Relocate codes that may be duplicates (#3503)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Fri, 17 May 2019 02:14:38 +0000 (11:14 +0900)]
[locomotiv] Relocate codes that may be duplicates (#3503)

This will relocate relu codes that may be duplicates in the future.

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[loco] Introduce Domain enum class (#3500)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 16 May 2019 23:51:20 +0000 (08:51 +0900)]
[loco] Introduce Domain enum class (#3500)

* [loco] Introduce Domain enum class

This commit introduces Domain enum class which denotes the kind of
a loco value.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Update comment

5 years ago[locomotiv] Remove data update duplication in Forward execution (#3507)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 16 May 2019 07:50:27 +0000 (16:50 +0900)]
[locomotiv] Remove data update duplication in Forward execution (#3507)

This commit removes NodeData update duplication in switch statement in
Forward execution.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
5 years ago[locomotiv] Remove data update duplication in Push execution (#3496)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 16 May 2019 06:57:41 +0000 (15:57 +0900)]
[locomotiv] Remove data update duplication in Push execution (#3496)

* [locomotiv] Remove data update duplication in Push execution

This commit removes NodeData update duplication in switch statement in
Push execution.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
* Add assertion

5 years ago[loco] Introduce Feature Encode/Decode node (#3495)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 16 May 2019 06:26:04 +0000 (15:26 +0900)]
[loco] Introduce Feature Encode/Decode node (#3495)

This commit extends loco IR with FeatureEncode/FeatureDecode nodes.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[locomotiv] Introduce ReLU (#3485)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 16 May 2019 04:54:25 +0000 (13:54 +0900)]
[locomotiv] Introduce ReLU (#3485)

* [locomotive] Introduce ReLU

This will enable ReLU FLOAT32 execution

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* remove unused using

* use index enumerator

* remove unused code

5 years ago[moco] import TensorFlow Const for some float values (#3494)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 16 May 2019 04:29:23 +0000 (13:29 +0900)]
[moco] import TensorFlow Const for some float values (#3494)

This will enable TensorFlow Const conversion where value is given with float value but with partial elements.

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] import TensorFlow Const for binary encoded (#3492)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 16 May 2019 03:55:44 +0000 (12:55 +0900)]
[moco] import TensorFlow Const for binary encoded (#3492)

This will enable TensorFlow Const conversion where value is given as tensor_content.

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] enable TensoFlow Relu node import (#3488)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 16 May 2019 03:49:28 +0000 (12:49 +0900)]
[moco] enable TensoFlow Relu node import (#3488)

This will enable TensorFlow Relu node import in frontend.

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[loco] Introduce Feature Encoder & Decoder (#3479)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Thu, 16 May 2019 03:47:07 +0000 (12:47 +0900)]
[loco] Introduce Feature Encoder & Decoder (#3479)

* [loco] Introduce Feature Encoder & Decoder

This commit introduces Feature Encoder & Decoder interface.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Update comments per feedback

* Another comment update

* Remove unnecessary space

5 years ago[moco] test material for TensorFlow Relu (#3490)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 16 May 2019 03:43:29 +0000 (12:43 +0900)]
[moco] test material for TensorFlow Relu (#3490)

This will add fist unit test material for importing TensorFlow Relu node.

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[locomotiv] Docstring for NodeData and make_data() (#3483)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Thu, 16 May 2019 03:37:59 +0000 (12:37 +0900)]
[locomotiv] Docstring for NodeData and make_data() (#3483)

* [locomotiv] Docstring for NodeData and make_data()

This commit adds docstring for NodeData class and make_data() function.
NodeData is read-only, and not modifiable after creation.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
* Review feedback updated

5 years ago[moco] import TensorFlow Const for one value data (#3491)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 16 May 2019 03:36:13 +0000 (12:36 +0900)]
[moco] import TensorFlow Const for one value data (#3491)

This will enable TensorFlow Const conversion where there is only one float value.
And add a test code for this case.

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] import Const attributes and value data (#3487)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 16 May 2019 01:54:15 +0000 (10:54 +0900)]
[moco] import Const attributes and value data (#3487)

This will enable import for Const attributes and data where there are same number of const value data.
Add test code for this case.

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco] introduce test helper header (#3476)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Thu, 16 May 2019 00:51:47 +0000 (09:51 +0900)]
[moco] introduce test helper header (#3476)

* [moco] introduce test helper header

This will add some helpers for node unit testing

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* move outside namespace

* use istream

5 years agoAdd ONNXSource 1.3.0 package (#3481)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 15 May 2019 10:37:59 +0000 (19:37 +0900)]
Add ONNXSource 1.3.0 package (#3481)

This commit introduces ONNXSource v1.3.0 package.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[loco] Eliminate CRTP from Named class and rename it to NameInfo (#3464)
Роман Михайлович Русяев/AI Tools Lab /SRR/Staff Engineer/삼성전자 [Wed, 15 May 2019 10:23:33 +0000 (13:23 +0300)]
[loco] Eliminate CRTP from Named class and rename it to NameInfo (#3464)

* rename Named class to NameInfo
* remove virtual destructor because this class is not designed for public inheritance

Signed-off-by: Roman Rusyaev <r.rusyaev@samsung.com>
5 years ago[nnkit/moco/tf] Build moco_test_tf only when TensorFlow is found (#3465)
Ivan Vagin/AI Tools Lab /SRR/Engineer/삼성전자 [Wed, 15 May 2019 09:14:32 +0000 (12:14 +0300)]
[nnkit/moco/tf] Build moco_test_tf only when TensorFlow is found (#3465)

Build moco_test_tf only when TensorFlow is found

Signed-off-by: Ivan Vagin <ivan.vagin@samsung.com>
5 years ago[moco] introduce empty TF Const converter (#3477)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 15 May 2019 08:28:29 +0000 (17:28 +0900)]
[moco] introduce empty TF Const converter (#3477)

This will introduce empty Const node converter

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[moco/tf] Introduce get_tensor_attr (#3468)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 15 May 2019 06:42:42 +0000 (15:42 +0900)]
[moco/tf] Introduce get_tensor_attr (#3468)

This will introduce get_tensor_attr method to get tensor attribure from node

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[loco] Introduce ReLU node (#3472)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 15 May 2019 06:42:26 +0000 (15:42 +0900)]
[loco] Introduce ReLU node (#3472)

This will extend loco IR with ReLU node.

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
5 years ago[loco] Introduce MaxPool2D node (#3473)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 15 May 2019 06:42:15 +0000 (15:42 +0900)]
[loco] Introduce MaxPool2D node (#3473)

This commit introduces MaxPool2D node.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years agoDistinguish unversioned ONNXSource package (#3469)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 15 May 2019 05:23:13 +0000 (14:23 +0900)]
Distinguish unversioned ONNXSource package (#3469)

This commit introduces ONNXSourceConfigVersion.cmake for unversioned
package in order to eliminate ambiguity in package searching.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[loco] Introduce Window class (#3459)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 14 May 2019 10:59:59 +0000 (19:59 +0900)]
[loco] Introduce Window class (#3459)

This commit declares (generic) Window class, and provides its
specialization over 2D Window.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[loco] Introduce Stride class (#3461)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 14 May 2019 10:59:49 +0000 (19:59 +0900)]
[loco] Introduce Stride class (#3461)

* [loco] Introduce Stride class

This commit declares (generic) Stride class, and provides its
specialization over 2D.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Mark as a final class

5 years ago[loco] Introduce FeatureIndex class (#3458)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 14 May 2019 09:50:38 +0000 (18:50 +0900)]
[loco] Introduce FeatureIndex class (#3458)

This commit introduces FeatureIndex class which is the offset of an
element in a feature map.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[nnkit/moco/tf] `runall` and build script (#3449)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 14 May 2019 03:55:39 +0000 (12:55 +0900)]
[nnkit/moco/tf] `runall` and build script (#3449)

This commit adds `runall` and build script, working for `Placeholder_000` test case.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
5 years ago[loco] Introduce FeatureShape class (#3448)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 14 May 2019 03:50:38 +0000 (12:50 +0900)]
[loco] Introduce FeatureShape class (#3448)

This commit introduces FeautreShape class which describes the shape of
Feature Map.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[nnkit/moco/tf] Adds getConstFloatTensor for InputTensorContext (#3447)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 14 May 2019 00:50:06 +0000 (09:50 +0900)]
[nnkit/moco/tf] Adds getConstFloatTensor for InputTensorContext (#3447)

This adds getConstFloatTensor() method body for InputTensorContext, which is called by HDF5::Export.cpp

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
5 years ago[loco] Introduce TensorIndex class (#3445)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 13 May 2019 23:35:46 +0000 (08:35 +0900)]
[loco] Introduce TensorIndex class (#3445)

* [loco] Introduce TensorIndex class

This commit introduces TensorIndex class to loco IR.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Fix a typo

5 years ago[MIR] Added Readme (#3271)
Андрей Шедько/AI Tools Lab /SRR/Engineer/삼성전자 [Mon, 13 May 2019 16:23:55 +0000 (19:23 +0300)]
[MIR] Added Readme (#3271)

Added a readme for Model IR

Signed-off-by: Andrei Shedko <a.shedko@samsung.com>
5 years ago[nnkit/moco/tf] Fix bug while calling ModelSignature (#3446)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 13 May 2019 10:12:11 +0000 (19:12 +0900)]
[nnkit/moco/tf] Fix bug while calling ModelSignature (#3446)

With this commit, node name (without, e.g., ":0") is passed to ModelSignature, not a tensor name with, e.g., ":0".

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
5 years ago[nnkit/moco/tf] Restructuring 'using' (#3438)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 13 May 2019 09:46:33 +0000 (18:46 +0900)]
[nnkit/moco/tf] Restructuring 'using' (#3438)

`using` is restructured.
ParsedTensors are common so move `using` with ParsedTensors to top level. Others were moved to method level.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
5 years ago[nnkit/moco/tf] backend of nnkit (#3444)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 13 May 2019 09:13:37 +0000 (18:13 +0900)]
[nnkit/moco/tf] backend of nnkit (#3444)

* [nnkit/moco/tf] backend of nnkit

This commit adds backend of nnkit for nnkit/moco/tf.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* indentation

5 years ago[CMake] Check TensorFlow version (#3373)
채성우/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 13 May 2019 09:13:14 +0000 (18:13 +0900)]
[CMake] Check TensorFlow version (#3373)

* [CMake] Check TensorFlow version

This commit introduce check of tensorflow version.

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
* This commit introduce .FORMATCHECKED

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
* This commit apply format.patch

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
* This commit make code safe.

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
* This commit remove modification of 'format'

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
5 years ago[locomotiv] Session frees node annotations on destruction (#3441)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 13 May 2019 07:46:07 +0000 (16:46 +0900)]
[locomotiv] Session frees node annotations on destruction (#3441)

* [locomotiv] Session frees node annotations on destruction

This commit introduces destructor of Session. Previously node
annotations set by Session, like NodeDataImpl, remain even after
Session closed. They were not memory leak by definition, as they can
be accessed through another Session. However node annotations maybe
preserved until program ends, so it is like memory leak. This commit
would prevent that situation.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
* Delete unused using Index

* Session is final class

5 years ago[loco] Introduce TensorShape class (#3443)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 13 May 2019 07:42:03 +0000 (16:42 +0900)]
[loco] Introduce TensorShape class (#3443)

This commit introduces TensorShape class to loco IR.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years agoExtract nnkit-tflite (#3442)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 13 May 2019 07:29:28 +0000 (16:29 +0900)]
Extract nnkit-tflite (#3442)

This commit extracts nnkit modules related with TensorFlow Lite as
nnkit-tflite.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[nnkit/tf] fix wrong name of method (#3440)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 13 May 2019 05:29:26 +0000 (14:29 +0900)]
[nnkit/tf] fix wrong name of method (#3440)

This commit renames isInt32Tensor to isS32Tensor.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
5 years ago[tfldump] Add readme (#3297)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 13 May 2019 04:48:32 +0000 (13:48 +0900)]
[tfldump] Add readme (#3297)

* [tfldump] Add readme

This will add readme for tfldump project

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
* fix typo and some comments

5 years ago[locomotiv] Support Forward node (#3436)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 13 May 2019 03:46:54 +0000 (12:46 +0900)]
[locomotiv] Support Forward node (#3436)

This commit starts to support loco::Forward node. In terms of node
execution, Forward calculation is same with Push's.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
5 years ago[locomotiv] Use break rather than return (#3437)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 13 May 2019 03:44:58 +0000 (12:44 +0900)]
[locomotiv] Use break rather than return (#3437)

For Push node execution, let's use break rather than return inside
switch statement, as it feels more natural.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
5 years ago[nnkit] int32 support for nnkit::TensorContext (#3413)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 13 May 2019 03:44:38 +0000 (12:44 +0900)]
[nnkit] int32 support for nnkit::TensorContext (#3413)

* [nnkit] int32 support for nnkit::TensorContext

This commit adds methods to support int32 for nnkit::TensorContext

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* Int32->S32, added header

* cluster S32 function tegether

5 years agoExtract nnkit-caffe (#3435)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 13 May 2019 02:12:39 +0000 (11:12 +0900)]
Extract nnkit-caffe (#3435)

This commit extracts nnkit-caffe from nnkit, and simplifes the depedency
check in CMakeLists.txt.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[locomotiv] Use macro to iterate supported Nodes (#3429)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 13 May 2019 02:12:01 +0000 (11:12 +0900)]
[locomotiv] Use macro to iterate supported Nodes (#3429)

* [locomotiv] Use macro to iterate supported Nodes

This commit introduces list of supported Nodes for execution, and use
them with macro to iterate and do same thing through Nodes.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
* clang-format off for comment indentation

5 years ago[loco] Introduce ConstGen node (#3425)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Sun, 12 May 2019 23:47:43 +0000 (08:47 +0900)]
[loco] Introduce ConstGen node (#3425)

* [loco] Introduce ConstGen node

This commit extends loco IR with ConstGen node.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Add assert

* Fix a bug and add a corresponding test

5 years ago[nnkit/moco/tf] backend implementation (#3424)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Sun, 12 May 2019 23:45:58 +0000 (08:45 +0900)]
[nnkit/moco/tf] backend implementation (#3424)

* [nnkit/moco/tf] backend implementation

This commit adds backend implementation for nnkit/moco/tf.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* move using into class

* fixed for locomotiv set_input

5 years agoExtract "nnkit-intf" from "nnkit" (#3433)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 10 May 2019 09:49:30 +0000 (18:49 +0900)]
Extract "nnkit-intf" from "nnkit" (#3433)

This commit extracts "nnkit-intf" from "nnkit" as the first step toward
nnkit directory restructuring.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
5 years ago[nnkit] Run onnx model using onnx runtime (#3432)
윤지영/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 10 May 2019 08:11:10 +0000 (17:11 +0900)]
[nnkit] Run onnx model using onnx runtime (#3432)

This patch has TensorContext class to run pre and post actions.

Signed-off-by: Jiyoung Yun <jy910.yun@samsung.com>
5 years ago[loco] Introduce input_nodes/output_nodes helper (#3428)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Fri, 10 May 2019 06:41:54 +0000 (15:41 +0900)]
[loco] Introduce input_nodes/output_nodes helper (#3428)

* [loco] Introduce input_nodes/output_nodes helper

This commit introduces input_nodes/output_nodes helper functions which
allow users to easily enumerate graph-level input/output nodes.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Update test name