summary |
shortlog | log |
commit |
commitdiff |
tree
first ⋅ prev ⋅ next
Tae-Young Chung [Mon, 16 Jan 2023 06:09:15 +0000 (15:09 +0900)]
Add flex delegate
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
Hyunsoo Park [Thu, 7 Jul 2022 02:31:08 +0000 (11:31 +0900)]
Applying coding rule
Change-Id: Ieb020b720fa19cccd1826706d67ae7887332002b
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
heechul.jeon [Thu, 30 Jun 2022 07:15:52 +0000 (16:15 +0900)]
InferenceTFLite: Remove code redundancy
[Versin] 0.0.4
[Issue type] code cleanup
Change-Id: Ia6942cd730aedd74f5acbd98b75f6b4e1b7dabfa
Signed-off-by: heechul.jeon <heechul.jeon@samsung.com>
heechul.jeon [Thu, 30 Jun 2022 01:47:29 +0000 (10:47 +0900)]
InferenceTFLite: Remove code redundancy
[Versin] 0.0.3
[Issue type] code cleanup
Change-Id: Ifa72db34f18c9cd89ac8055a8dcc6c5529ec9a87
Signed-off-by: heechul.jeon <heechul.jeon@samsung.com>
Hyunsoo Park [Wed, 12 Jan 2022 03:52:53 +0000 (12:52 +0900)]
Adds info logs to 'SetInterpreterInfo()'
[Versin] 0.0.3-0
[Issue type] clean up
Change-Id: I6c4304b08028c92b3981099cb73fec7f0eadb80d
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Inki Dae [Wed, 22 Dec 2021 07:28:46 +0000 (16:28 +0900)]
fix seg. fault issue with GPU delegate
Fixed seg. fault issue with GPU delegate of Tensorflow lite 2.7.
With Tensorflow lite 2.7 GPU delegate creation needs a default option.
So this patch gets a default option for creating GPU delegate by calling
TfLiteGpuDelegateOptionsV2Default().
[Versin] 0.0.2-0
[Issue type] bug fix
Change-Id: I4a99377e393aacb415de62ac9004c3095623d923
Signed-off-by: Yongjoo Ahn <yongjoo1.ahn@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Hyunsoo Park [Thu, 13 May 2021 04:18:27 +0000 (13:18 +0900)]
Remove debug log of no use
Change-Id: I859dbdf688533834727913a25cc5fcd90d6c352f
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Hyunsoo Park [Wed, 12 May 2021 08:40:50 +0000 (17:40 +0900)]
Set OutputLayers when it is empty vector
Change-Id: I035159ce30cdc0edafdce72324a235693493d802
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Tae-Young Chung [Fri, 26 Mar 2021 00:45:37 +0000 (09:45 +0900)]
Remove duplicated code line
mInputLayerId and mOutputLayerId are swapped.
So they don't need to call clear() again and should be removed.
Change-Id: I9ef8df94e31baefc680837cc9d582095669fd399
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
Tae-Young Chung [Wed, 10 Mar 2021 09:09:52 +0000 (18:09 +0900)]
Change members of inference_engine_layer_property structure,
and change vector<inference_engine_tensor_buffer> to map<string, inference_engine_tensor_buffer>
This is based on
https://review.tizen.org/gerrit/#/c/platform/core/multimedia/inference-engine-interface/+/254892/
https://review.tizen.org/gerrit/#/c/platform/core/api/mediavision/+/254953/
Change-Id: I451e44dc31aeafe4a92f86baef34204daeed70a5
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
Inki Dae [Thu, 4 Feb 2021 00:48:54 +0000 (09:48 +0900)]
Implement SetCLTuner interface
Added SetCLTuner interface for CLTuner feature support of inference
engine interface framework which has a pure virtual function interface
of SetCLTUner function so the interface implementataion is required.
Change-Id: I2cf2168c899f4762b1f22786acff85229ad26b83
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Inki Dae [Wed, 2 Dec 2020 05:10:55 +0000 (14:10 +0900)]
Add flatbuffer dependency and header cleanup
Change-Id: I673bc5f157b1f326d029c10bb22b5beeadc16c26
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Inki Dae [Mon, 23 Nov 2020 01:56:35 +0000 (10:56 +0900)]
Use offical gpu delegate
Change-Id: I5f419d65aaec5c12dd1810317e3718f5b6fce8a3
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Inki Dae [Tue, 10 Nov 2020 08:10:54 +0000 (17:10 +0900)]
Add GPU device capabality
Change-Id: I1c7ab126e2ef61c4c2b85e1c125f071629ec21e7
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Inki Dae [Tue, 3 Nov 2020 09:03:46 +0000 (18:03 +0900)]
Add tflite 2.3 gpu delegate support
Change-Id: I289316756e3d6a4c266810c57db68a0596b1733b
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Inki Dae [Tue, 3 Nov 2020 06:20:58 +0000 (15:20 +0900)]
Set thread number to -1
Refer to below reference document,
"num_threads should be >= -1. User may pass -1 to let the
TFLite interpreter set the no of threads avaiable to itself."
Change-Id: I9d9ce2c7f4a9df41fce0e6763b1b7eac1505f0b5
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Inki Dae [Fri, 28 Aug 2020 02:51:43 +0000 (11:51 +0900)]
Add int64 tensor type support
Change-Id: I57251340b8321940c66e4e8bd3d04a0373726088
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Inki Dae [Fri, 26 Jun 2020 00:22:24 +0000 (09:22 +0900)]
Fix coding style based on Tizen SE C++ Coding Rule
Tizen SE C++ Coding Rule:
https://wiki.tizen.org/Native_Platform_Coding_Idiom_and_Style_Guide#C.2B.2B_Coding_Style
Change-Id: I9462a7838c00bfd5c8c1ec18270b203db923eb87
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Inki Dae [Wed, 3 Jun 2020 09:12:04 +0000 (18:12 +0900)]
Add SetPrivateData function
This function is needed for inference engine interface to pass
tflite private data to this backend before loading a model file.
Change-Id: Ibd169b0ddc8c1894c68a8e3016523e47b3450b74
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Hyunsoo Park [Wed, 3 Jun 2020 10:07:10 +0000 (19:07 +0900)]
Add output layer info to GetOutputLayerProperty api
From this patch,
GetOutputLayerProperty api could be called without calling SetOutputLayerProperty api.
Change-Id: I2c7987d4923fb4f23ea3545d9f6861d8cebf10c6
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Tae-Young Chung [Tue, 2 Jun 2020 05:43:45 +0000 (14:43 +0900)]
Fix build error aarch64 and x86_64
Change-Id: I398542921f17fba7b80b4b057440be1b249645dd
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
Inki Dae [Tue, 12 May 2020 07:14:23 +0000 (07:14 +0000)]
Merge "Get tensor infos from interpreter" into tizen
Hyunsoo Park [Mon, 11 May 2020 09:25:03 +0000 (18:25 +0900)]
Get tensor infos from interpreter
After applying it,
GetInputLayerProperty/GetInputTensorBuffers could get infos from interpreter without setting.
Change-Id: I5263656a8ac73ea244582e7e98e1607809547d70
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Inki Dae [Tue, 12 May 2020 04:02:31 +0000 (13:02 +0900)]
Check if model file path is valid or not
Change-Id: Ib1b2f3f897d4ee0d3ff4646429308ded762651ea
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Hyunsoo Park [Thu, 7 May 2020 09:59:57 +0000 (18:59 +0900)]
Fix UNINIT.CTOR issue and remove useless call
Change-Id: I695f9a89e0d89e0ffb776c0f4ef2a77dc5a4afa1
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Hyunsoo Park [Wed, 6 May 2020 08:25:50 +0000 (17:25 +0900)]
tflite:add empty checking
For calling 'GetInputTensorBuffers', 'SetInputLayerProperty' should be called.
Change-Id: I6e8695cf8be42c2cd04a43b31684323ec9ed8f85
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Inki Dae [Fri, 17 Apr 2020 07:52:48 +0000 (16:52 +0900)]
ange enumeration values to new ones
some enumeration values of inference-engine-interface have been
updated so change them.
Signed-off-by: Inki Dae <inki.dae@samsung.com>
Hyunsoo Park [Fri, 10 Apr 2020 05:54:39 +0000 (14:54 +0900)]
Change tensor shape type to size_t
Change-Id: I0077b7406de26620a31ecd0f7eee6bdab7838e3d
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Hyunsoo Park [Thu, 2 Apr 2020 09:16:04 +0000 (18:16 +0900)]
Support UINT_8
Change-Id: Idc7776c9123abd665f96204120864801657e2d54
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Hyunsoo Park [Wed, 1 Apr 2020 09:55:16 +0000 (18:55 +0900)]
Refatorying code according to refactoried inference-engine-interface
Change-Id: I73da801ac99a22ee5b0abb03752da76bdbc0117f
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
sangho park [Fri, 27 Sep 2019 02:28:52 +0000 (11:28 +0900)]
fix coverity issue
[1087465] Uninitialized pointer field
Change-Id: I9d9a7a94ac4d0cce7085268e1fa7f10058f0900a
Signed-off-by: sangho park <sangho.g.park@samsung.com>
Tae-Young Chung [Wed, 18 Sep 2019 08:36:11 +0000 (17:36 +0900)]
Remove apis related to IInferenceEngineVision
inference-engine-tflite is a plugin to provide inference only.
Thus, domain specific functions such as vision should be removed.
Instead, add apis GetInputLayerAttrType(), SetInputDataBuffer(), and GetInputDataPtr()
which can be used to access to memory.
Change-Id: I408a95c86bc2477465e5a08dab192bb6f3813ad1
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
Tae-Young Chung [Mon, 16 Sep 2019 01:54:37 +0000 (10:54 +0900)]
Remove BuildRequires: pkgconfig(tensorflow-lite) >= 1.13.1
There is no tensorflow-lite rpm, but only tensorflow-lite-devel.
That should be removed.
Change-Id: Ib41b40613dc53808024723a8ab759895824ada2e
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
Hyunsoo Park [Thu, 5 Sep 2019 05:03:24 +0000 (14:03 +0900)]
Add namespace.
Without declaring InferenceEngineInterface::Common, this class can't detect 'IInferenceEngineCommon'.
Change-Id: I443c6bd8e827ff3a6846c840738c255c2b7b9e72
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Hyunsoo Park [Thu, 22 Aug 2019 11:26:53 +0000 (20:26 +0900)]
Deletion of related methods as inherited class changing.
It is inherited by IInferenceEngineVision and IInferenceEngineCommon both.
But relationship of them is changed as inheritant. So i remove unneccessary methods.
Change-Id: I22303fbdea68b2944f9a64fc99fee77a8501d4e9
Signed-off-by: Hyunsoo Park <hance.park@samsung.com>
Tae-Young Chung [Mon, 12 Aug 2019 07:23:32 +0000 (16:23 +0900)]
Added missing libopencv_improc link and implemented face and facial landmark detection
Change-Id: I49546847d8f054bb026036e4a385be0cf17c5f0f
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
Tae-Young Chung [Wed, 24 Jul 2019 08:08:28 +0000 (17:08 +0900)]
Add INFERENCE_TARGET_CUSTOM, Support multiple output tensors and object detection
Note that Tizen supports tensorflow-lite version 1.13,
which supports the SSD including postprocess, i.e. NMS
Change-Id: I6542449a284727ab0e1eba8c9e9aa28fa9bc39ee
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
Tae-Young Chung [Thu, 4 Jul 2019 11:05:35 +0000 (20:05 +0900)]
Intial codes
Note that this initial code is under development and provides very limited functionality.
Signed-off-by: Tae-Young Chung <ty83.chung@samsung.com>
Tizen Infrastructure [Mon, 1 Jul 2019 02:10:31 +0000 (02:10 +0000)]
Initial empty repository