platform/core/ml/nnfw.git
4 years ago[moco-tf] Remove IR tests (#8225)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 08:23:47 +0000 (17:23 +0900)]
[moco-tf] Remove IR tests (#8225)

This will remove IR test codes that already moved to moco-lang

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[exo] Introduce TFLConcatenation converter (#8221)
남궁석/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 16 Oct 2019 08:03:58 +0000 (17:03 +0900)]
[exo] Introduce TFLConcatenation converter (#8221)

This commit will introduce `TFLConcatenation` converter in `exo`

Signed-off-by: Seok NamKoong <sk.namkoong@samsung.com>
4 years ago[exo] Refactoring: Extract code from big method (Fuse) (#8179)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 08:03:47 +0000 (17:03 +0900)]
[exo] Refactoring: Extract code from big method (Fuse) (#8179)

A piece of code is extracted as a method (`create_fused_bias()`) from big method (`Fuse`).

For diff difficulty, old code will be replace with one single call in next PR.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years ago[exo] TFLMixin for classes that have fused activation function (#8223)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 07:53:35 +0000 (16:53 +0900)]
[exo] TFLMixin for classes that have fused activation function (#8223)

This adds TFLMixin class for TFL nodes that have fused activation function

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years ago[nnfw-api] use circle loader for circle model (#8222)
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 07:51:49 +0000 (16:51 +0900)]
[nnfw-api] use circle loader for circle model (#8222)

It uses model_types from MANIFEST to determine loader.
It also added missing namespace neurun for circle_loader.

Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
4 years ago[nnpkg-test] Remove conv test (#8220)
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Wed, 16 Oct 2019 07:51:33 +0000 (16:51 +0900)]
[nnpkg-test] Remove conv test (#8220)

Remove fail conv test: neurun cannot run convolution with input weight

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[exo] TODO rigorous padding conversion (#8219)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 16 Oct 2019 07:48:34 +0000 (16:48 +0900)]
[exo] TODO rigorous padding conversion (#8219)

This commit adds TODO comments for rigorous padding conversion

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
4 years ago[exo] Introducing FuseReluPass (#8176)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 07:48:10 +0000 (16:48 +0900)]
[exo] Introducing FuseReluPass (#8176)

* [exo] Introducing FuseReluPass

This introduces FuseReluPass, which fuse relu/relu6 into 10 TensorFlow Lite ops.
This commit adds code that finds candidate subgraph.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* remove unused var (to avoid CI error)

* Update compiler/exo/src/Pass/FuseReluPass.h

Co-Authored-By: hyunsik-yoon <hyunsik.yoon@samsung.com>
4 years ago[moco-tf] Revise IR to using moco::IR (#8218)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 07:39:12 +0000 (16:39 +0900)]
[moco-tf] Revise IR to using moco::IR (#8218)

This will revise moco::tf IR and base classes to alias of moco IR and base classes

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[exo] Register FoldTransposeOfConst pass (#8202)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 16 Oct 2019 06:59:54 +0000 (15:59 +0900)]
[exo] Register FoldTransposeOfConst pass (#8202)

This commit registers FoldTransposeOfConst pass to exo TFL-level
optimization.

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
4 years ago[exo] Relu6Converter and TFL/CircleOperationExporter for TFLRelu6 (#8173)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 06:51:09 +0000 (15:51 +0900)]
[exo] Relu6Converter and TFL/CircleOperationExporter for TFLRelu6 (#8173)

This adds Relu6Converter and TFL/CircleOperationExporter for TFLRelu6.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years ago[exo] Turn ON ConvertConv2D knob (#8153)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 16 Oct 2019 06:50:32 +0000 (15:50 +0900)]
[exo] Turn ON ConvertConv2D knob (#8153)

This commit turns on ConvertConv2D knob by default

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
4 years ago[moco-tf] Revise to use moco TF dialect (#8208)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 05:20:51 +0000 (14:20 +0900)]
[moco-tf] Revise to use moco TF dialect (#8208)

This will revise moco_tf to use TF dialect from moco

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Tidy unused codes (#8204)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 05:03:39 +0000 (14:03 +0900)]
[moco-tf] Tidy unused codes (#8204)

This will remove not used anymore Concat and Padding annotations
- Remove also related codes

And remove also not used anymore ImportTarget

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Comment for Relu and Relu6 graph builder (#8201)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 05:02:46 +0000 (14:02 +0900)]
[moco-tf] Comment for Relu and Relu6 graph builder (#8201)

This will add comments fro Relu and Relu6 GraphBuilder header

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[exo-tflite] Introduce TFLConcatenate node (#7978)
남궁석/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 16 Oct 2019 04:44:05 +0000 (13:44 +0900)]
[exo-tflite] Introduce TFLConcatenate node (#7978)

* [exo-tflite] Introduce TFLConcatenate node

This commit will introduce `TFLConcatenate` node in `exo-tflite`

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

* rebase and apply comments

* fix typo

* apply comments

Co-Authored-By: sk-namkoong <sk.namkoong@samsung.com>
* add assert

* fix typo

4 years ago[res/TensorFlow] Adding Conv2D-BiasAdd-Relu6 test case (#8172)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 03:16:31 +0000 (12:16 +0900)]
[res/TensorFlow] Adding Conv2D-BiasAdd-Relu6 test case (#8172)

* [res/TensorFlow] Adding Conv2D-BiasAdd-Relu6 test case

This adds pbtxt and related files for Conv2D-BiasAdd-Relu6.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* remove a term CBR since CBR is normally used with Relu

4 years ago[moco-tf] Remove unused knobs (#8192)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 01:43:58 +0000 (10:43 +0900)]
[moco-tf] Remove unused knobs (#8192)

This will remove not used anymore ImportAs TF dialect knobs

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco] Revive old arg accessors (#8194)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 01:35:50 +0000 (10:35 +0900)]
[moco] Revive old arg accessors (#8194)

This will revive old input argument accessors for IR migration from moco-tf to moco

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Tidy includes in Op folder (#8198)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 01:35:17 +0000 (10:35 +0900)]
[moco-tf] Tidy includes in Op folder (#8198)

This will tidy includes in Op folder that is not necessary anymore

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Use headers in node graph builder (#8191)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Wed, 16 Oct 2019 00:59:42 +0000 (09:59 +0900)]
[moco-tf] Use headers in node graph builder (#8191)

This will use node header files in graph builder and cleanup unnecessary header includes

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[neurun] Apply trying const dealloc on ExecutionFactory (#8064)
김용섭/On-Device Lab(SR)/Engineer/삼성전자 [Wed, 16 Oct 2019 00:53:01 +0000 (09:53 +0900)]
[neurun] Apply trying const dealloc on ExecutionFactory (#8064)

Apply trying const dealloc on LinearExecutor and ParallelExecutor of
ExecutionFactory

Signed-off-by: Yongseop Kim <yons.kim@samsung.com>
4 years ago[exo] remove complex if statement (#8178)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 23:17:52 +0000 (08:17 +0900)]
[exo] remove complex if statement (#8178)

This removes complex if statement to reduce cyclomatic complexity.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years ago[mir-tflite-importer] Use new operations interface in importer. (#8183)
Gusev Dmitry/Engineer/AI Tools Lab /SRR/Samsung Electronics [Tue, 15 Oct 2019 21:07:09 +0000 (00:07 +0300)]
[mir-tflite-importer] Use new operations interface in importer. (#8183)

Changes due to new operations constructors signatures.

Signed-off-by: Dmitry Gusev <d.gusev@partner.samsung.com>
4 years ago[mir] Add methods for accessing Operation's inputs (#8188)
Sergei Barannikov/Engineer/AI Tools Lab /SRR/Samsung Electronics [Tue, 15 Oct 2019 21:05:45 +0000 (00:05 +0300)]
[mir] Add methods for accessing Operation's inputs (#8188)

The `getInput` method is always used in pair with `getProducer`.
This PR adds `getInputProducer` method for convenience.

Signed-off-by: Sergei Barannikov <s.barannikov@samsung.com>
4 years ago[nnc] Fix unittest for CombineTransposes optimization (#8187)
Sergei Barannikov/Engineer/AI Tools Lab /SRR/Samsung Electronics [Tue, 15 Oct 2019 21:02:41 +0000 (00:02 +0300)]
[nnc] Fix unittest for CombineTransposes optimization (#8187)

Get the correct IDs of Input ops.

Signed-off-by: Sergei Barannikov <s.barannikov@samsung.com>
4 years ago[mir2loco] Use new operations interface in mir2loco.test (#8185)
Gusev Dmitry/Engineer/AI Tools Lab /SRR/Samsung Electronics [Tue, 15 Oct 2019 19:20:34 +0000 (22:20 +0300)]
[mir2loco] Use new operations interface in mir2loco.test (#8185)

Changes related to operations constructors renewal.

Signed-off-by: Dmitry Gusev <d.gusev@partner.samsung.com>
4 years ago[nnc] Use new operations interface. (#8184)
Gusev Dmitry/Engineer/AI Tools Lab /SRR/Samsung Electronics [Tue, 15 Oct 2019 19:15:02 +0000 (22:15 +0300)]
[nnc] Use new operations interface. (#8184)

Changes due to renewed operations constructors.

Signed-off-by: Dmitry Gusev <d.gusev@partner.samsung.com>
4 years ago[mir-caffe2-importer] Use new operations interface in importer. (#8181)
Gusev Dmitry/Engineer/AI Tools Lab /SRR/Samsung Electronics [Tue, 15 Oct 2019 17:19:34 +0000 (20:19 +0300)]
[mir-caffe2-importer] Use new operations interface in importer. (#8181)

Changes due to new operations constructors signatures.

Signed-off-by: Dmitry Gusev <d.gusev@partner.samsung.com>
4 years ago[mir-caffe-importer] Use new operations interface in importer. (#8180)
Gusev Dmitry/Engineer/AI Tools Lab /SRR/Samsung Electronics [Tue, 15 Oct 2019 17:07:26 +0000 (20:07 +0300)]
[mir-caffe-importer] Use new operations interface in importer. (#8180)

Changes due to new operations constructors signatures.

Signed-off-by: Dmitry Gusev <d.gusev@partner.samsung.com>
4 years ago[mir] Rename AffineQuantization methods (#8177)
Pavel Iliutchenko/AI Tools Lab /SRR/Engineer/Samsung Electronics [Tue, 15 Oct 2019 15:09:04 +0000 (18:09 +0300)]
[mir] Rename AffineQuantization methods (#8177)

* Remove plural variant of methods

Signed-off-by: Pavel Iliutchenko <p.iliutchenk@samsung.com>
4 years ago[tool] Validation model file check in tflite_benchmark (#8149)
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 15 Oct 2019 11:23:42 +0000 (20:23 +0900)]
[tool] Validation model file check in tflite_benchmark (#8149)

Pass verifier and use VerifyAndBuildFromFile in tflite_benchmark

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years agoUpdate HowToUseDockerImage.md (Fix broken link) (#8174)
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 15 Oct 2019 11:21:26 +0000 (20:21 +0900)]
Update HowToUseDockerImage.md (Fix broken link) (#8174)

This commit fixes broken link to relative address

Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
4 years ago[mir] Creation InputOp from TensorType (#8137)
Pavel Iliutchenko/AI Tools Lab /SRR/Engineer/Samsung Electronics [Tue, 15 Oct 2019 10:31:37 +0000 (13:31 +0300)]
[mir] Creation InputOp from TensorType (#8137)

* Added setOutputType in Operation
* Added InputOp constructor for TensorType
* Marked other constructor as DEPRECATED

Signed-off-by: Pavel Iliutchenko <p.iliutchenk@samsung.com>
4 years ago[moco-integration-test] adding NET_0034 (#8169)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 10:18:34 +0000 (19:18 +0900)]
[moco-integration-test] adding NET_0034 (#8169)

NET_0034 (conv2d-sub) was added into test.lst.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years ago[tf2tflite] Adding NET_0034 into test.lst (#8170)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 10:17:15 +0000 (19:17 +0900)]
[tf2tflite] Adding NET_0034 into test.lst (#8170)

This adds NET_0034 (conv2d-sub) into test.lst.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years ago[exo] Initializing TFLAdd's fused activation func in existing test code (#8168)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 10:13:43 +0000 (19:13 +0900)]
[exo] Initializing TFLAdd's fused activation func in existing test code (#8168)

Some test code does not initializes TFLAdd's fused activation func. This commit adds that.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years ago[exo] type and shape inference for Relu6 (#8171)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 10:13:32 +0000 (19:13 +0900)]
[exo] type and shape inference for Relu6 (#8171)

This adds type and shape inference for Relu6.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years ago[exo] Pass to fold Transpose + Const (#8155)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 15 Oct 2019 08:02:57 +0000 (17:02 +0900)]
[exo] Pass to fold Transpose + Const (#8155)

This commit introduces pass to fold Transpose of Const node

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
4 years ago[moco-tf] Node GraphBuilder header (#8158)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 08:02:27 +0000 (17:02 +0900)]
[moco-tf] Node GraphBuilder header (#8158)

* [moco-tf] Node GraphBuilder header

This will introduce header for GraphBuilder for nodes that doesn't have one.

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

4 years ago[exo-tflite] test cases for FuseConv2DAddSubPass (#8101)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 07:49:27 +0000 (16:49 +0900)]
[exo-tflite] test cases for FuseConv2DAddSubPass (#8101)

* [exo-tflite] test cases for FuseConv2DAddSubPass

This adds four test cases for FuseConv2DAddSubPass.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
* make const value 1.5, 3

4 years ago[res] Remove deprecated test case (#8165)
채성우/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 15 Oct 2019 07:46:28 +0000 (16:46 +0900)]
[res] Remove deprecated test case (#8165)

This commit removes deprecated test case.

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
4 years ago[exo] Initialize TFLConv2D member vars (#8166)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 07:45:57 +0000 (16:45 +0900)]
[exo] Initialize TFLConv2D member vars (#8166)

This adds an initialization of TFLConv2D member vars.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years ago[exo] Add fusedActivationFunction to TFLAdd and TFLMul (#8161)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 07:42:59 +0000 (16:42 +0900)]
[exo] Add fusedActivationFunction to TFLAdd and TFLMul (#8161)

* [exo] Add fusedActivationFunction to TFLAdd and TFLMul

TFLAdd and TFLMul now has an attribute for fusedActivationFunction.

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

* remove unused param

* defined and used init_fused_act_func

* remove constructor. init member vars directly

4 years ago[nncc] Remove ENABLE_CONTRIB_BUILD option (#8156)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 15 Oct 2019 06:07:46 +0000 (15:07 +0900)]
[nncc] Remove ENABLE_CONTRIB_BUILD option (#8156)

This option was introduce when nncc has "contrib" directory, but
"contrib" directory no longer exists.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
4 years ago[CI] Revise test report directory (#7994)
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 15 Oct 2019 06:01:08 +0000 (15:01 +0900)]
[CI] Revise test report directory (#7994)

Divide test report directory for each scheduler
Use mixed backend test for graph generation test

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years agoNNAS create-package command (#8086)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 15 Oct 2019 05:28:03 +0000 (14:28 +0900)]
NNAS create-package command (#8086)

* NNAS create-package command

This commit introduces create-package command.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
* Add TODO and some comments

4 years ago[loco] Add comment to TensorConstantPad (#8022)
채성우/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 15 Oct 2019 03:38:26 +0000 (12:38 +0900)]
[loco] Add comment to TensorConstantPad (#8022)

* [loco] Add comment to TensorConstantPad

This commit adds comment to TensorConstantPad IR in loco.

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
* make comment more clear.

* apply comment.

* modify comment.

* modify comment.

4 years ago[android] Remove unncessary CMakeLists comments (#8151)
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 15 Oct 2019 03:37:00 +0000 (12:37 +0900)]
[android] Remove unncessary CMakeLists comments (#8151)

Remove comments that are no longer helpful.

Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
4 years ago[exo] Using C++ 11 style static_assert (#8152)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 03:29:43 +0000 (12:29 +0900)]
[exo] Using C++ 11 style static_assert (#8152)

This modifies static_assert from c++ 17 style to c++ c++ 11 style.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years ago[tflite_loader] Model file verification (#8132)
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 15 Oct 2019 03:00:25 +0000 (12:00 +0900)]
[tflite_loader] Model file verification (#8132)

Verification check before tflite model loading

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[nnapi] Fix transpose conv NNAPI test (#8127)
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Tue, 15 Oct 2019 01:38:09 +0000 (10:38 +0900)]
[nnapi] Fix transpose conv NNAPI test (#8127)

Set kernel to constant, not model input

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years agoApply optimized cpu kernel for SoftMaxFloat32 (#7943)
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Tue, 15 Oct 2019 01:36:00 +0000 (10:36 +0900)]
Apply optimized cpu kernel for SoftMaxFloat32 (#7943)

This commit applies optimized cpu kernel for SoftMaxFloat32.
  - Introduce optimized cpu kernel of SoftMax for float
  - Apply that kernel for cpu backend of neurun

Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
4 years ago[nnpkg-run] allow --load to accept little endian (#8116)
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 01:17:18 +0000 (10:17 +0900)]
[nnpkg-run] allow --load to accept little endian (#8116)

It will allow little endianness in `nnpackage-run`.
`hdf5` internally convert endianness if necessay.
All we need is to allow little endian from if-condition.

Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
4 years ago[tf2circle] Initial copy from tf2tflite (#8145)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 00:59:15 +0000 (09:59 +0900)]
[tf2circle] Initial copy from tf2tflite (#8145)

* [tf2circle] Initial copy from tf2tflite

This will add working tf2circle from copy of tf2tflite
- CustomOp support is not added cause of duplicate code for now

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

* remove test material conversion related

* fix build error, remove unused, fix comment

4 years ago[moco] Add naming rule as a comment (#8147)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 00:49:11 +0000 (09:49 +0900)]
[moco] Add naming rule as a comment (#8147)

This will add some comment for naming rule for TF dialect OPCODE and input accessor

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Add naming rule as a comment (#8144)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 00:27:23 +0000 (09:27 +0900)]
[moco-tf] Add naming rule as a comment (#8144)

* [moco-tf] Add naming rule as a comment

This will add some comment for naming rule for TF dialect OPCODE and input/output accessor

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

4 years ago[moco-tf] Graph build Relu Relu6 as TF dialect (#8141)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 00:21:13 +0000 (09:21 +0900)]
[moco-tf] Graph build Relu Relu6 as TF dialect (#8141)

This will update graph builder to produce only TF dialect for Relu and Relu6 node

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Graph build MaxPool as TF dialect (#8140)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 00:20:56 +0000 (09:20 +0900)]
[moco-tf] Graph build MaxPool as TF dialect (#8140)

This will update graph builder to produce only TF dialect for MaxPool node

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Graph build Identity as TF dialect (#8139)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Tue, 15 Oct 2019 00:15:30 +0000 (09:15 +0900)]
[moco-tf] Graph build Identity as TF dialect (#8139)

This will update graph builder not to produce only TF dialect for Identity node
- related Importer test is also modified to work with this change

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Graph build Concat as TF dialect (#8131)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 22:20:22 +0000 (07:20 +0900)]
[moco-tf] Graph build Concat as TF dialect (#8131)

This will remove Graph builder fro Concat as Canonical dialect and build only TF dialect

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Graph build Const as TF dialect (#8133)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 22:19:58 +0000 (07:19 +0900)]
[moco-tf] Graph build Const as TF dialect (#8133)

This will remove Graph builder for Const as Canonical dialect and build only TF dialect

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Graph build Conv2D as TF dialect (#8135)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 22:19:37 +0000 (07:19 +0900)]
[moco-tf] Graph build Conv2D as TF dialect (#8135)

This will remove Graph builder for Conv2D as Canonical dialect and build only for TF dialect

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Graph build BiasAdd as TF dialect (#8130)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 22:19:18 +0000 (07:19 +0900)]
[moco-tf] Graph build BiasAdd as TF dialect (#8130)

This will remove Graph builder for BiasAdd as Canonical dialect and build only for TF dialect

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Graph build for AvgPool as TF dialect (#8128)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 22:10:00 +0000 (07:10 +0900)]
[moco-tf] Graph build for AvgPool as TF dialect (#8128)

This will remove Graph builder for AvgPool as Canonical dialect and build only for TF dialect

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years agoRefine [mir] operations constructors in order to simplify interface. (#8044)
Gusev Dmitry/Engineer/AI Tools Lab /SRR/Samsung Electronics [Mon, 14 Oct 2019 18:07:38 +0000 (21:07 +0300)]
Refine [mir] operations constructors in order to simplify interface. (#8044)

Operations attributes are united into structures.

Signed-off-by: Dmitry Gusev <d.gusev@partner.samsung.com>
4 years ago[mir] Introduce AffineQuantization in MIR (#8077)
Pavel Iliutchenko/AI Tools Lab /SRR/Engineer/Samsung Electronics [Mon, 14 Oct 2019 16:01:22 +0000 (19:01 +0300)]
[mir] Introduce AffineQuantization in MIR (#8077)

* Implemented AffineQuantization
* Added helpers for setting Quantization to TensorVariant and TensorType

Signed-off-by: Pavel Iliutchenko <p.iliutchenk@samsung.com>
4 years ago[exo-tflite] Overload to_shape_description for MatrixShape (#8009)
Sergei Barannikov/Engineer/AI Tools Lab /SRR/Samsung Electronics [Mon, 14 Oct 2019 12:39:10 +0000 (15:39 +0300)]
[exo-tflite] Overload to_shape_description for MatrixShape (#8009)

Add overload of `to_shape_description` for `MatrixShape`.

Signed-off-by: Sergei Barannikov <s.barannikov@samsung.com>
4 years ago[docs] Introduce CrossBuildForAndroid.md for nnfw (#7976)
이한종/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 14 Oct 2019 12:33:24 +0000 (21:33 +0900)]
[docs] Introduce CrossBuildForAndroid.md for nnfw (#7976)

* [docs] Introduce CrossBuildForAndroid.md for nnfw

This commit introduces `CrossBuildForAndroid.md` for Android Cross
Build Instruction.

Resolve #7892

Signed-off-by: Hanjoung Lee <hanjoung.lee@samsung.com>
* Symlink Makefile rather than copying

* Complete the incomplete sentence

4 years agoMake to support BatchToSpaceND op for acl neon (#7986)
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 14 Oct 2019 11:24:19 +0000 (20:24 +0900)]
Make to support BatchToSpaceND op for acl neon (#7986)

This commit makes to support BatchToSpaceND op for acl neon.

Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
4 years ago[mir] Introduce new operations: Quantize and Dequantize (#8050)
Pavel Iliutchenko/AI Tools Lab /SRR/Engineer/Samsung Electronics [Mon, 14 Oct 2019 11:04:15 +0000 (14:04 +0300)]
[mir] Introduce new operations: Quantize and Dequantize (#8050)

* Append QuantizeOp and DequantizeOp on mir

Signed-off-by: Pavel Iliutchenko <p.iliutchenk@samsung.com>
4 years agoMake to support ResizeBilinear op on tflite_loader (#8008)
장지섭/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 14 Oct 2019 09:59:15 +0000 (18:59 +0900)]
Make to support ResizeBilinear op on tflite_loader (#8008)

This commit makes to support ResizeBilinear op on tflite_loader.

Signed-off-by: jiseob.jang <jiseob.jang@samsung.com>
4 years ago[loco] TensorConstantPad shape and type inference (#8043)
채성우/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 14 Oct 2019 08:04:34 +0000 (17:04 +0900)]
[loco] TensorConstantPad shape and type inference (#8043)

* [loco] TensorConstantPad shape and type inference

This commit support shape and type inference of TensorConstantPad.

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

4 years ago[exo] Set DataFormat of subgraph for Circle (#8113)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 08:01:17 +0000 (17:01 +0900)]
[exo] Set DataFormat of subgraph for Circle (#8113)

* [exo] Set DataFormat of subgraph for Circle

This will set DataFormat of the sub graph for Circle format as default

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

4 years agoInstall exo and exo-tflite (#8107)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 14 Oct 2019 07:46:43 +0000 (16:46 +0900)]
Install exo and exo-tflite (#8107)

From now on, install command will put exo and exo-tflite librarry into
the specified output directory.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
4 years agoAdd "NET_0010" manifest (#8059)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 14 Oct 2019 07:32:48 +0000 (16:32 +0900)]
Add "NET_0010" manifest (#8059)

This commit adds NET_0010 test manifest.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
4 years ago[moco-tf] Remove unused WindowData header (#8121)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 07:25:34 +0000 (16:25 +0900)]
[moco-tf] Remove unused WindowData header (#8121)

This will remove not used anymore WindowData.h file

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Remove unused PadData and StrideData (#8115)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 06:55:34 +0000 (15:55 +0900)]
[moco-tf] Remove unused PadData and StrideData (#8115)

This will remove not used anymore PadData and StrideData annotations

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] Turn ON knobs for resolving Reshape (#8117)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 14 Oct 2019 06:52:36 +0000 (15:52 +0900)]
[moco-tf] Turn ON knobs for resolving Reshape (#8117)

This commit turns on following knobs:
- ResolveDuplicateReshape
- ResolveRedundantReshape

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
4 years ago[moco-tf] Fix Placeholder import (#8114)
박천교/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 14 Oct 2019 06:49:58 +0000 (15:49 +0900)]
[moco-tf] Fix Placeholder import (#8114)

Previously, Placeholder with negative value was not caught by assertion,
and converted to Pull with overflowed value. This commit fixes this
behaviour

Signed-off-by: Cheongyo Bahk <ch.bahk@samsung.com>
4 years ago[tf2tflite] use exo (#8109)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 06:49:37 +0000 (15:49 +0900)]
[tf2tflite] use exo (#8109)

This will update tf2tflite to use exo instead of exo_tflite

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[exo] Update includes (#8105)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 06:49:14 +0000 (15:49 +0900)]
[exo] Update includes (#8105)

This will remove not needed anymore header includes for ExportUtils

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[exo] use circle for comments and methods (#8111)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 06:48:38 +0000 (15:48 +0900)]
[exo] use circle for comments and methods (#8111)

This will update to use Circle for comments and local methods

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[nnfw-api] Introduce resize input API (#8034)
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 14 Oct 2019 06:36:33 +0000 (15:36 +0900)]
[nnfw-api] Introduce resize input API (#8034)

* [nnfw-api] Introduce resize input API

Introduce API for resize input: nnfw_apply_tensorinfo
Internal implementation always return error - NYI
Update related API's comment based on doxygen format

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

4 years ago[nnapi-test] Use android 10 test generation (#8040)
오형석/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 14 Oct 2019 06:24:14 +0000 (15:24 +0900)]
[nnapi-test] Use android 10 test generation (#8040)

Use android 10 test generation
- Disable TestCompliance util
- Disable _Float16 feature
- Disable dynamic output shape test generation
- Fix to use c++11 feature
- Update generated code archive for tizen build

Signed-off-by: Hyeongseok Oh <hseok82.oh@samsung.com>
4 years ago[nnfw-api] explain nnfw_set_backend and nnfw_set_op_backend more (#8071)
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 06:13:34 +0000 (15:13 +0900)]
[nnfw-api] explain nnfw_set_backend and nnfw_set_op_backend more (#8071)

It explains more about nnfw_set_backend and nnfw_set_op_backend.

Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
4 years ago[nnpkg-tcs] Run nnpkg-tcs after tc exists (#8119)
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 06:12:10 +0000 (15:12 +0900)]
[nnpkg-tcs] Run nnpkg-tcs after tc exists (#8119)

Some nnpkg-tcs are not generated via `tf2tflite` test filter.
Thus, nnpkg-tcs checks whether test case exists before running.

Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
4 years ago[exo] Adding FuseConv2DAddSubPass into TFLOptimize.cpp (#8102)
윤현식/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 04:47:10 +0000 (13:47 +0900)]
[exo] Adding FuseConv2DAddSubPass into TFLOptimize.cpp (#8102)

Now FuseConv2DAddSubPass runs when UseFuseConv2DAddSubPass knob is on.

Signed-off-by: Hyun Sik Yoon <hyunsik.yoon@samsung.com>
4 years agoLet locomotiv be installable (#8089)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 14 Oct 2019 04:43:20 +0000 (13:43 +0900)]
Let locomotiv be installable (#8089)

From now on, install command will put locomotive library into the proper
location.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
4 years ago[moco-tf] Rename TestHelper filename (#8104)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 04:42:52 +0000 (13:42 +0900)]
[moco-tf] Rename TestHelper filename (#8104)

This will rename TestHelper to TestHelper.test so that it only exist for test executable not shared library

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-integration-test] Add renamed test case (#8098)
채성우/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 14 Oct 2019 04:41:03 +0000 (13:41 +0900)]
[moco-integration-test] Add renamed test case (#8098)

This commit adds renamed test case.

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
4 years agoLet make moco-tf installable (#8093)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 14 Oct 2019 04:34:56 +0000 (13:34 +0900)]
Let make moco-tf installable (#8093)

From now on, install command will put moco-tf library into a proper
location.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
4 years ago[exo] add copy of TFExporter as CircleExporter (#8094)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 04:34:43 +0000 (13:34 +0900)]
[exo] add copy of TFExporter as CircleExporter (#8094)

This will add a copy of TFExporter and related files as CircleExporter to export Circle file

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years agoLet make tf2tflite installable (#8096)
박종현/On-Device Lab(SR)/Staff Engineer/삼성전자 [Mon, 14 Oct 2019 04:25:41 +0000 (13:25 +0900)]
Let make tf2tflite installable (#8096)

From now on, install command will put tf2tflite binary into a proper
location.

Signed-off-by: Jonghyun Park <jh1302.park@samsung.com>
4 years ago[moco-tf] Remove ShapeInferenceData annotation (#8100)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 04:20:13 +0000 (13:20 +0900)]
[moco-tf] Remove ShapeInferenceData annotation (#8100)

This will remove ShapeInferenceData annotation that was replaced by loco NodeShape

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[moco-tf] TFShapeInferenceRule test (#8074)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 03:46:06 +0000 (12:46 +0900)]
[moco-tf] TFShapeInferenceRule test (#8074)

This will add TFShapeInferenceRule test from old FixShapeTransform test

Signed-off-by: SaeHie Park <saehie.park@samsung.com>
4 years ago[tflite2circle] Remove the limitation of running under nnfw root (#8068)
이상규/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 03:35:40 +0000 (12:35 +0900)]
[tflite2circle] Remove the limitation of running under nnfw root (#8068)

Now, it uses the pathname of the currently executing script to calculate the default path for other resources.

Signed-off-by: Sanggyu Lee <sg5.lee@samsung.com>
4 years ago[tf2tflite] Add renamed test case (#8088)
채성우/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 14 Oct 2019 03:17:21 +0000 (12:17 +0900)]
[tf2tflite] Add renamed test case (#8088)

This commit adds renamed test case to tf2tflite.

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
4 years ago[exo] Change target to exo (#8073)
박세희/On-Device Lab(SR)/Principal Engineer/삼성전자 [Mon, 14 Oct 2019 01:57:25 +0000 (10:57 +0900)]
[exo] Change target to exo (#8073)

* [exo] Change target to exo

This will update CMake target to exo
- add a temporary exo_tflite target for dependent targets

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

* add a dup instead of interface

4 years ago[moco-tf] Add Pad canonicalizer (#8030)
채성우/On-Device Lab(SR)/Engineer/삼성전자 [Mon, 14 Oct 2019 01:52:06 +0000 (10:52 +0900)]
[moco-tf] Add Pad canonicalizer (#8030)

* [moco-tf] Add Pad canonicalizer

This commit adds Pad canonicalizer to moco-tf.

Signed-off-by: seongwoo <sw4670.chae@samsung.com>
* apply commnets.

* apply comment.

* Simplify method to get rank of padding.

* fix wrong assert.

* add comments in canonicalization.

* make comment more clear.