MyungJoo Ham [Thu, 31 Jan 2019 01:47:35 +0000 (10:47 +0900)]
[Test] runTest.sh typo
If statements of bash uses ==, not =, for equal-comparison
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
MyungJoo Ham [Wed, 30 Jan 2019 10:36:02 +0000 (19:36 +0900)]
[Conf] Bugfix of array out of bound
The last element of vstr should be NULL.
Set NULL for such elements
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
MyungJoo Ham [Wed, 30 Jan 2019 07:23:56 +0000 (16:23 +0900)]
[Conf] Bugfix: if no path is given, use current dir.
If path is NULL, don't give it to glib path library.
Use "./" instead.
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
MyungJoo Ham [Tue, 29 Jan 2019 09:15:05 +0000 (18:15 +0900)]
[Filter:*] Cut build dependency of filter-subplugin to filter
Enable tensor-filter subplugins to be built without
dependencies on tensor_filter.
After this commit, we can build filter subplugins with
the exported "tensor_typedef.h".
This enables #786 and #761
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
MyungJoo Ham [Tue, 29 Jan 2019 09:06:17 +0000 (18:06 +0900)]
[Subplugin] Do not dlsym if it's already probed.
With dlopen, consturctor function is loaded, which
may call subplugin-probe.
If it has called subplugin-probe, registering the
subplugin to an element (filter/decoder) specific
internal list, we do not need to register it with
nnstreamer_subplugin structure.
- Add an internal API, "hold_register_subplugin", that
notified nnstremer_subplugin that it has probed already.
- Call "hold_register_subplugin" at probe function.
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
MyungJoo Ham [Tue, 29 Jan 2019 09:04:30 +0000 (18:04 +0900)]
Bugfix of nnstreamer_conf internal API
- Fix exported name of nnsconf_loadconf (typo)
- Fix incorrect enums of subplugin_prefixes (typo)
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
MyungJoo Ham [Tue, 29 Jan 2019 14:46:03 +0000 (23:46 +0900)]
[Codeowner] Add kparichay to default reviewers
And, we do not need special reviewers for packaging any more.
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
jijoong.moon [Wed, 16 Jan 2019 02:36:48 +0000 (11:36 +0900)]
[Repo] Add test case for repo dynamicity
In this PR, added repo test application. It constructs pipeline and
change slot id from 0 to 1 after spending time waiting for 200 micro
second.
repository
+------------+ +----------+ +-----------------+
| Multi-sink |<--| repo_src |<---+---| slot 0 |<---+--+
+------------+ +----------+ | +-----------------+ | |
+---| slot 1 |<---+ |
+-----------------+ |
|
+-----------+ +--------+ +-----------+ +-----------+ |
| Multi-src |-->| pngdec |-->| converter |-->| repo_sink |-----+
+-----------+ +--------+ +-----------+ +-----------+
**Self evaluation:**
1. Build test: [X]Passed [ ]Failed [ ]Skipped
2. Run test: [X]Passed [ ]Failed [ ]Skipped
Signed-off-by: jijoong.moon <jijoong.moon@samsung.com>
Hyoung Joo Ahn [Fri, 25 Jan 2019 06:17:32 +0000 (15:17 +0900)]
[Decoder/Boundingbox] make decoder possible to handle tensorflow ssd model
make it available to process the output tensors of tensorflow ssd model
Signed-off-by: Hyoung Joo Ahn <hello.ahn@samsung.com>
MyungJoo Ham [Mon, 28 Jan 2019 04:41:22 +0000 (13:41 +0900)]
[Filter/subplugin] Follow the subplugin naming rule.
Subplugin name rule (described at gst/nnstreamer/nnstreamer_subplugin.h)
states:
The filename should be subplugin_prefixes[type]${name}.so
The subplugins should be follow the rule to be recognized by runtime
plug-and-play mechanism of nnstreamer_subplugin.h
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
Wook Song [Mon, 28 Jan 2019 02:02:19 +0000 (11:02 +0900)]
[Common] Remove unused function, 'gst_tensor_media_type_from_caps'
This patch removes a unused function, gst_tensor_media_type_from_caps,
from tensor_common.
Signed-off-by: Wook Song <wook16.song@samsung.com>
Wook Song [Fri, 25 Jan 2019 06:27:00 +0000 (15:27 +0900)]
[Dist/Tizen] Add plug-in API library to development package
This patch adds the plug-in API library to the development package for
Tizen.
Signed-off-by: Wook Song <wook16.song@samsung.com>
Wook Song [Fri, 25 Jan 2019 06:08:51 +0000 (15:08 +0900)]
[DIST/Debian] Add plug-in API library to development package
This patch adds the plug-in API library to the development pacakge for
Debian distros.
Signed-off-by: Wook Song <wook16.song@samsung.com>
Wook Song [Fri, 25 Jan 2019 06:02:02 +0000 (15:02 +0900)]
[PkgConfig] Update to use plug-in API lib
This patch updates the PkgConfig file for the development package so
that the plug-in developers can use newly added plug-in API library.
Signed-off-by: Wook Song <wook16.song@samsung.com>
Wook Song [Fri, 25 Jan 2019 01:57:33 +0000 (10:57 +0900)]
[Meson] Create shared and static libraries of plug-in APIs
This patch updates meson build script to create shared and static
libraries of plug-in APIs.
Signed-off-by: Wook Song <wook16.song@samsung.com>
Wook Song [Fri, 25 Jan 2019 01:27:55 +0000 (10:27 +0900)]
[Common] Remove duplicate contents from tensor_common
This patch removes contents duplicated in the plug-in APIs are removed
from tensor_common. Meson build script is also updated to avoid build
break.
Signed-off-by: Wook Song <wook16.song@samsung.com>
Wook Song [Fri, 25 Jan 2019 00:47:42 +0000 (09:47 +0900)]
[Common] Extract APIs for plug-in developers from tensor_common
Since we do not export tensor_common anymore, plug-in developers outside
our repository cannot use typedefed data types and mandatory functions
to implement NNStreamer extension plug-ins. In order to handle such
issue, this patch extracts types and functions can be regarded as APIs
from tensor_common. Note that only subset of all APIs which the plugin
developers would need has been currently extracted.
Signed-off-by: Wook Song <wook16.song@samsung.com>
Jaeyun Jung [Sat, 26 Jan 2019 03:15:27 +0000 (12:15 +0900)]
[Build/Deb] enable tensorflow
protobuf packaging done, enable tensorflow option.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Hyoung Joo Ahn [Fri, 25 Jan 2019 07:26:13 +0000 (16:26 +0900)]
[Filter/NNFW] add error message for user&debugging
add error message when errors have occurred
Signed-off-by: Hyoung Joo Ahn <hello.ahn@samsung.com>
Jaeyun [Fri, 25 Jan 2019 04:58:55 +0000 (13:58 +0900)]
[Filter/Test] flow-return to drop incoming buffer
check returned value from invoke vmethod and return flow-dropped to skip incoming buffer.
TODO:
define status code for NNFW and custom-filter
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
MyungJoo Ham [Fri, 25 Jan 2019 00:54:08 +0000 (09:54 +0900)]
Start of 0.1.1 Release Candidate Development
This is the start of 0.1.1 development.
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
MyungJoo Ham [Thu, 24 Jan 2019 06:35:55 +0000 (15:35 +0900)]
Release NNStreamer 0.1.0
- Build system migration cmake --> meson
- Support Tensorflow without input/output tensor memcpy
- other/tensor stream format updated
- From 0.1.0, a single property, "dimension", describes the whole dimension instead of "dim1", "dim2", ...
- Objective 1: in the future, we may support tensors with more than 4 dimensions without updating the protocol.
- Objective 2: it was just too ugly.
- Example applications migrated to other git repo to make this repo ready for upstreaming in the fugure and to ensure buildability for third party developers.
- Support run-time attaching subplugins (filter and decoder)
- Support "ini" and envvar configurations for subplugin locations
- Dynamic external recurrences
- Subplugin API sets (draft. do not expect backward compatibility)
- Bug fixes (memory leaks, incorrect logs, type checks, ...)
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
Hyoung Joo Ahn [Thu, 24 Jan 2019 12:39:51 +0000 (21:39 +0900)]
[Filter/NNFW] clean code of tensorflow & tf-lite
adjust the style and remove useless keywords
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Jaeyun [Thu, 24 Jan 2019 05:46:27 +0000 (14:46 +0900)]
[Common] common function to copy tensor info
add common function to copy tensor info
TODO:
1. seperate util-functions for developer (custom filter and plugins)
2. update custom examples and nnstreamer plugins with these functions
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
MyungJoo Ham [Thu, 24 Jan 2019 09:20:44 +0000 (18:20 +0900)]
[Doc] Add gitter badge
Let users find the nnstreamer gitter easily.
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
jijoong.moon [Fri, 18 Jan 2019 08:21:51 +0000 (17:21 +0900)]
[Repo] Implement Dynamic Property(slot_id) Change of Repo
Implement Dynamic slot_id property change of tensor_reposrc and
tensor_reposink. In order to check the status of current repo,
sink_changed(gboolean), sink_id(guint) and src_changed (gboolean),
src_id(guint) variables are added. While waiting buffer from
repo_sink, repo check if this variables are changed using
gst_tensor_repo_check_chagned call, and if it changed, it trys to get
buffer from new slot id. The changing status of repo is set when slot
id property changed using tensor_repo_set_changed.
**Changes proposed in this PR:**
- Added function calls to check and set the changing status of repo.
- Added variables to control flow in each repo.
**Self evaluation:**
1. Build test: [X]Passed [ ]Failed [ ]Skipped
2. Run test: [X]Passed [ ]Failed [ ]Skipped
Signed-off-by: jijoong.moon <jijoong.moon@samsung.com>
MyungJoo Ham [Thu, 24 Jan 2019 06:21:07 +0000 (15:21 +0900)]
[Doc] Add regular release status
Build/Text status and Test code coverage will be
available after https://github.com/nnsuite/TAOS-CI/issues/452
is resolved.
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
MyungJoo Ham [Thu, 24 Jan 2019 05:25:34 +0000 (14:25 +0900)]
[Doc] Update getting started.
Cleaning up the documentation, following up
3065d38fb8c56064ab695a8d46b2f9dbd3e1aad7
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
Jaeyun [Wed, 23 Jan 2019 08:41:30 +0000 (17:41 +0900)]
[Aggregator] condition to make output buffer writable
code review:
calls gst_buffer_make_writable when aggregator needs to concatenate output data
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Hyoung Joo Ahn [Wed, 23 Jan 2019 03:58:23 +0000 (12:58 +0900)]
[Filter] fix the memory leak at filter
this pr solves the memory leak problem about the `model_file` and the `name` of input/output tensor
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Mauro Carvalho Chehab [Thu, 17 Jan 2019 16:24:04 +0000 (14:24 -0200)]
[Documentation] Update install instructions
In order to install it on Debian, it is needed first to get
the ssat and tensorflow packages.
Add instructions about how to do that.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Hyoung Joo Ahn [Mon, 21 Jan 2019 11:39:46 +0000 (20:39 +0900)]
[Decoder/BoundingBox] rename parameters just for tf-lite
some parameters subordinated to the ssd model of TFLite should have their own name.
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Jaeyun [Tue, 22 Jan 2019 04:01:26 +0000 (13:01 +0900)]
[Build] update document and include path
1. update document (remove cmake guide)
2. change orc generated header path
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Tue, 22 Jan 2019 03:34:54 +0000 (12:34 +0900)]
[Build] remove cmake
transition to meson, remove all the cmake script.
relate issue : #927
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Tue, 15 Jan 2019 08:00:10 +0000 (17:00 +0900)]
[Common] change caps format for other/tensor
This PR changes caps format for other/tensor.
Caps for tensor stream : other/tensor, framerate = (fraction) [0, max], type = (string) { uint8, .. }, dimension = (string) d1:d2:d3:d4
Related issue : #1027
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Tue, 22 Jan 2019 05:34:10 +0000 (14:34 +0900)]
[Document] update documents
update broken links and nnstreamer changes
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
jijoong.moon [Tue, 15 Jan 2019 01:31:26 +0000 (10:31 +0900)]
[Repo] Add test cases for negotiation check
In order to check negotiation error, added test cases for repo.
**Changes proposed in this PR:**
- Added Test cases for repo
Resolves: #1026
**Self evaluation:**
1. Build test: [X]Passed [ ]Failed [ ]Skipped
2. Run test: [X]Passed [ ]Failed [ ]Skipped
Signed-off-by: jijoong.moon <jijoong.moon@samsung.com>
Hyoung Joo Ahn [Tue, 22 Jan 2019 04:22:57 +0000 (13:22 +0900)]
[Filter] correct misspells and improper logic
there are some typos about in/out at a comment & exception handling
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
jijoong.moon [Mon, 14 Jan 2019 22:44:21 +0000 (07:44 +0900)]
[Repo] Add checking caps compatibility between repo_src & repo_sink
It is difficult to check caps compatibility during negotiation because
they are not pipelined each other. That's why reposrc needs caps
property by the way. However, we have to check compatibility somehow
during runtime. In order to do that, GstMetaRepo is introduced.
In this PR, functions and defines are included to manipulate the meta
for repo. During generating buffer in reposink, insert GstMetaRepo
data with GstCaps and push to repo. After that, reposrc extracts the
GstMetaRepo when pulling the gstbuffer from repo
and compare it with caps defined by user. If they are not compatible,
it shows error and return GST_FLOW_EOS.
**Changes proposed in this PR:**
- Add GstMetaRepo type and manipulation functions
- Add insert GstMetaRepo into buffer
- Add comparison routine during get gstbuffer from repo
- Add Error handling if they are not compatible
Resolves: Issue #1026
**Self evaluation:**
1. Build test: [X]Passed [ ]Failed [ ]Skipped
2. Run test: [X]Passed [ ]Failed [ ]Skipped
Signed-off-by: jijoong.moon <jijoong.moon@samsung.com>
MyungJoo Ham [Mon, 14 Jan 2019 01:28:23 +0000 (10:28 +0900)]
DIST: apply meson/ninja to debian build
Before remocing CMAKE infrastructure,
migrate build system to meson/ninja for debian builds.
This disabled tensorflow for Ubuntu build.
@hello-ahn reenable it after the update of protobuf
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
MyungJoo Ham [Sat, 19 Jan 2019 07:11:18 +0000 (16:11 +0900)]
[CI Fix] Temporarily Disable App Test
With recent updates on app test and build script,
we no longer have example applications in nnstreamer.git
It should be migrated to nnstreamer-example.git
We should build nnstreamer-example.git for app test, not
example directory in nnstreamer.git
This is an emergency fix for #1047
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
Hyoung Joo Ahn [Fri, 18 Jan 2019 06:34:44 +0000 (15:34 +0900)]
[Filter/Tensorflow] remove `memcpy` of the output tensor by implementing `destroyNotify()`
maintain the output tensors at the map for a while and erase it from the map when `destroyNotify()` is called
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Jaeyun [Fri, 18 Jan 2019 06:56:46 +0000 (15:56 +0900)]
[Meson] clean script for unittest
update build script for unittests and test-plugins
1. add script to copy exe for ssat
2. clean script to build test-plugins
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Mon, 21 Jan 2019 05:56:58 +0000 (14:56 +0900)]
[Example] remove all exec-examples
Now all the exe-examples are migrated to nnstreamer-example repository.
TODO : we have to set up the app-test with the example repo.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Thu, 17 Jan 2019 10:38:22 +0000 (19:38 +0900)]
[Meson] change dependency for thread
change dependency for thread (guide in meson-build https://mesonbuild.com/howtox.html)
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Hyoung Joo Ahn [Fri, 18 Jan 2019 03:56:36 +0000 (12:56 +0900)]
[Example] use an existed function rather than custom
replace _strdup() to g_strdup() because it fulfills exactly same role
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Hyoung Joo Ahn [Fri, 18 Jan 2019 03:33:31 +0000 (12:33 +0900)]
[Filter/Example] apply destroyNotify() callback at example
add destroy_notify callback at the example using allocate_invoke.
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Jaeyun [Fri, 18 Jan 2019 09:10:04 +0000 (18:10 +0900)]
[Filter] remove definition of supported framework
FW definition is now unnecessary.
1. remove fw defines
2. update custom filter description
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Hyoung Joo Ahn [Fri, 18 Jan 2019 10:54:39 +0000 (19:54 +0900)]
[Documentation] update the doc readme&getting-started document
change the address of example and modify the guide about meson and ppa address
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Hyoung Joo Ahn [Tue, 15 Jan 2019 08:30:43 +0000 (17:30 +0900)]
[Filter] add a method destoryNotify() to handle data pointer
It will handle the data pointer after GstMemory being destoryed. If no specific method, it will be g_free()
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Jaeyun [Wed, 16 Jan 2019 04:52:13 +0000 (13:52 +0900)]
[Meson] add dependency for generated files
Add dependency for generated files.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Wed, 16 Jan 2019 10:30:23 +0000 (19:30 +0900)]
[Meson/Example] update script to install python example
update script to install python example in image-classification
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Wed, 16 Jan 2019 09:25:20 +0000 (18:25 +0900)]
[Example] remove unnecessary examples
Remove unnecessary examples and dependency.
All the exec-examples are moved to nnstreamer-example repository.
Remained examples are needed for unit-test and app-test.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Hyoung Joo Ahn [Thu, 17 Jan 2019 07:05:44 +0000 (16:05 +0900)]
[Filter/Tensorflow] fix bug at input tensor
fix broken logic at input tensor
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Jaeyun [Mon, 14 Jan 2019 07:58:43 +0000 (16:58 +0900)]
[Filter/Common] common functions for tensor info
Add common functions to handle the fomatted string of dimension, type, and name in tensors info.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Mon, 14 Jan 2019 10:08:10 +0000 (19:08 +0900)]
[Test] remove unnecesary file
testcase_mux.tar.gz is now unnecessary.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
jijoong.moon [Mon, 14 Jan 2019 01:04:12 +0000 (10:04 +0900)]
[Repo] Fix the test case dimension
The input dim of tensor is 3:16:16.
**Self evaluation:**
1. Build test: [X]Passed [ ]Failed [ ]Skipped
2. Run test: [X]Passed [ ]Failed [ ]Skipped
Signed-off-by: jijoong.moon <jijoong.moon@samsung.com>
Jaeyun [Mon, 14 Jan 2019 01:59:14 +0000 (10:59 +0900)]
[Build] remove unnecessary installation
installing libraries (.a) is unnecessary with meson.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
MyungJoo Ham [Fri, 11 Jan 2019 06:26:06 +0000 (15:26 +0900)]
Plug-and-play sub-plugins for Tensor_Filter
Allow tensor_filter sub-plugins to be independently built
and attached/detached in run-time with subplugin APIs.
This resolves one of "future work" in ATC 2019 submitted paper already :)
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
Jaeyun [Mon, 14 Jan 2019 01:28:58 +0000 (10:28 +0900)]
[Converter] remove code to update property
Updating tensor info property may cause an error when doing caps negotiation.
related issue : #941
error case:
1. run pipeviz (e.g., videotestsrc - capsfilter - tensor_converter - tensor_sink)
2. set caps video/x-raw,format=RGB,width=640,height=480 in capsfilter and run pipeline
3. stop pipeline, and change caps (width=320,height=240)
4. run pipeline again, then it will make caps-negotiation failure.
If set tensor_info (line 1199), stop and run pipeline again,
- It resets the flag tensor_configured false. (in gst_tensor_converter_change_state())
- After running pipeline again, gst_tensor_info_validate (line 1189) returned true
and in-caps is different with saved tensor_info. This makes caps-negotiation failure.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
MyungJoo Ham [Mon, 14 Jan 2019 01:34:32 +0000 (10:34 +0900)]
Reviewer List Updated
The two reviewers have left the team.
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
Jaeyun [Fri, 11 Jan 2019 04:12:47 +0000 (13:12 +0900)]
[Meson] change option name
change option to avoid duplicated feature name
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Tue, 8 Jan 2019 09:45:13 +0000 (18:45 +0900)]
[Test] remove warning in testcases
remove unnecessary path (ld preload)
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Tue, 8 Jan 2019 05:12:40 +0000 (14:12 +0900)]
[Build] change dir structure
1. update meson/cmake to change directory structure (move common files and plugins to gst/nnstreamer)
2. move nnstreamer meson-build to gst/nnstreamer
3. set project args to build nnstreamer
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Mon, 7 Jan 2019 10:19:39 +0000 (19:19 +0900)]
[Example] remove script to build test-app
1. remove .spec and cmake to build test-app
2. remove unnecessary include in example obj-detection
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Hyoung Joo Ahn [Tue, 8 Jan 2019 02:24:22 +0000 (11:24 +0900)]
[Example] Upload the tensorflow example
this example detects objects with ssd model. it works similar with the tf-lite example
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Hyoung Joo Ahn [Mon, 7 Jan 2019 06:54:33 +0000 (15:54 +0900)]
[Unittest] Tensorflow unittest with MNIST
the basic unittest of tensorflow with MNIST model. it resolves #994
Signed-off-by: Hyoung Joo Ahn <hello.ahnn@gmail.com>
Jaeyun [Fri, 4 Jan 2019 10:40:31 +0000 (19:40 +0900)]
[Filter] reset prop values when closing nnfw
1. init properties when closing nnfw
2. change logger format and code clean
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Fri, 4 Jan 2019 07:12:24 +0000 (16:12 +0900)]
[Test] add testcases for tensor-aggregator
testcases for testing data concat in tensor-aggregator
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Fri, 4 Jan 2019 03:46:56 +0000 (12:46 +0900)]
[Split] use common functions
1. use nns_memcpy to copy tensor data
2. use common function to get tensor elements count
3. remove unnecessary local val
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Wed, 19 Dec 2018 11:29:34 +0000 (20:29 +0900)]
[Meson] gbs build using meson
update .spec to build using meson
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Thu, 20 Dec 2018 07:25:19 +0000 (16:25 +0900)]
[Meson] set dependency with include dir
declare nnstreamer-lib dependency with include directory
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Thu, 20 Dec 2018 06:14:19 +0000 (15:14 +0900)]
[Filter/TF] change macro to copy in/out tensor
check data type in outside the loop and code clean
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Wed, 19 Dec 2018 06:57:57 +0000 (15:57 +0900)]
[Filter/Cmake] link proper modules
set pkgs for tensorflow and link with necessary modules
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
MyungJoo Ham [Thu, 20 Dec 2018 06:10:49 +0000 (15:10 +0900)]
Fixed typo in the comments
Fixed trivial typos in the comments.
HyoungjooAhn [Wed, 19 Dec 2018 04:26:28 +0000 (13:26 +0900)]
[Filter/Tensorflow] add an exception case for tensorflow
checking the status of tensorflow after running model
Signed-off-by: HyoungjooAhn <hello.ahnn@gmail.com>
Jaeyun [Thu, 20 Dec 2018 04:16:03 +0000 (13:16 +0900)]
[Filter/TF] fix warning
To fix coverity issue, init variables and code clean.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Tue, 18 Dec 2018 04:44:07 +0000 (13:44 +0900)]
[Filter/TF/TF-lite] remove link of tensor-common
remove dependency of tensor-common and unnecessary headers
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
HyoungjooAhn [Mon, 17 Dec 2018 10:26:01 +0000 (19:26 +0900)]
[Filter/Tensorflow] take the original rank of model internally
to run model, hold the original rank internally
Signed-off-by: HyoungjooAhn <hello.ahnn@gmail.com>
Wook Song [Mon, 17 Dec 2018 02:22:06 +0000 (11:22 +0900)]
[Test/TensorMux] Add test cases for multi-tensors multiplexing
This provides unit test cases for multiplexing of multi-tensors.
Signed-off-by: Wook Song <wook16.song@samsung.com>
Wook Song [Mon, 17 Dec 2018 02:14:03 +0000 (11:14 +0900)]
[TensorMux] Support multi-tensors for input
This patch makes the TensorMux element have multi-tensors as its input
stream.
Signed-off-by: Wook Song <wook16.song@samsung.com>
Jaeyun [Wed, 19 Dec 2018 03:44:24 +0000 (12:44 +0900)]
[Decoder] remove build warning
rename internal functions and sub-plugin callback
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Mon, 17 Dec 2018 04:29:04 +0000 (13:29 +0900)]
[Meson/Submodule] install both static/shared libraries
build and install both static/shared libraries (filter-submodules and common libs)
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Wed, 12 Dec 2018 11:30:23 +0000 (20:30 +0900)]
[Meson] add installation
1. add option to install examples
2. set install path and add configuration
3. install libraries, headers, and examples
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Wed, 12 Dec 2018 07:45:18 +0000 (16:45 +0900)]
[Meson/Test] add code to make bmp2png for test
1. make bmp2png for the unittest
2. set gtest dependency false
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Wed, 12 Dec 2018 07:28:18 +0000 (16:28 +0900)]
[Meson/Example] dependency of examples
set proper dependencies for each examples.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Wed, 12 Dec 2018 06:24:58 +0000 (15:24 +0900)]
[Meson] code clean
1. add flags for tensorflow and tensorflow-lite
2. set objs with the flags and dependency of libraries
3. build with each plugin files
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Wook Song [Mon, 17 Dec 2018 08:05:29 +0000 (17:05 +0900)]
[Test] Set LD_PRELOAD to the nnstreamer library
Fixes #919
This patch set LD_PRELOAD to libnnstreamer.so in runTest.sh for testing
custom filter and arithmetic/dimchg/stand/transpose/typcast transforms
so that fixing the issue related to the warning messages about
libtensor_filter_tflitecore.so load failure.
Signed-off-by: Wook Song <wook16.song@samsung.com>
HyoungjooAhn [Fri, 14 Dec 2018 02:55:34 +0000 (11:55 +0900)]
[bugfix] fix checking the dimension of input tensor
logic is fixed to make it work properly for the any type of input dimension
Signed-off-by: HyoungjooAhn <hello.ahnn@gmail.com>
Jaeyun [Wed, 12 Dec 2018 04:25:08 +0000 (13:25 +0900)]
[Example/Cmake] remove unnecessary link
Examples don't need to link nnstreamer libs.
Remove this link and add modules for each examples.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Wed, 12 Dec 2018 03:57:27 +0000 (12:57 +0900)]
[Example] remove common header
In custom filter examples, remove tensor-common header.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Wed, 12 Dec 2018 07:34:29 +0000 (16:34 +0900)]
[Test] add plugin path
set plugin path for the unittest using nnstreamer elements.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
HyoungjooAhn [Tue, 11 Dec 2018 04:05:53 +0000 (13:05 +0900)]
[Filter] add condition to check input/output name
check the existance of the value before free for input/output name properties
Signed-off-by: HyoungjooAhn <hello.ahnn@gmail.com>
Jaeyun Jung [Mon, 10 Dec 2018 14:50:12 +0000 (23:50 +0900)]
[Transform/Orc] remove unnecessary memcpy
Related issue: #975
In arith mode, copying input array to output is unnecessary.
To fix this, add source in orc functions and set the data ptr for each operator.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Mon, 10 Dec 2018 10:27:33 +0000 (19:27 +0900)]
[Meson/CustomFilter] link custom-filter obj
In example custom-filter, link custom-filter.
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Mon, 10 Dec 2018 04:53:48 +0000 (13:53 +0900)]
[Filter/TF] fix warning and check dependency
1. check dependency of protobuf, tensorflow
2. fix build warning and code indentation
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
MyungJoo Ham [Sat, 8 Dec 2018 08:58:05 +0000 (17:58 +0900)]
[Filter] Build Error Fix (gcc > 6)
Recent code commit has introduced the following errors in
modern gcc:
```
[ 20%] Building C object gst/tensor_filter/CMakeFiles/tensor_filterOBJ.dir/tensor_filter.c.o
cd /<<PKGBUILDDIR>>/build/gst/tensor_filter && /usr/bin/cc -DENABLE_TENSORFLOW_LITE -DHAVE_ORC=1 -DVERSION=\"0.1.0\" -I/<<PKGBUILDDIR>>/gst/nnstreamer -isystem /usr/include/gstreamer-1.0 -isystem /usr/include/glib-2.0 -isystem /usr/lib/x86_64-linux-gnu/glib-2.0/include -isystem /usr/include/orc-0.4 -g -O2 -fdebug-prefix-map=/<<PKGBUILDDIR>>=. -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -Wall -Werror -fPIC -g -std=c89 -Wno-sign-compare -o CMakeFiles/tensor_filterOBJ.dir/tensor_filter.c.o -c /<<PKGBUILDDIR>>/gst/tensor_filter/tensor_filter.c
In file included from /usr/lib/x86_64-linux-gnu/glib-2.0/include/glibconfig.h:9:0,
from /usr/include/glib-2.0/glib/gtypes.h:32,
from /usr/include/glib-2.0/glib/galloca.h:32,
from /usr/include/glib-2.0/glib.h:30,
from /usr/include/gstreamer-1.0/gst/gstinfo.h:27,
from /<<PKGBUILDDIR>>/gst/tensor_filter/tensor_filter.c:57:
/<<PKGBUILDDIR>>/gst/tensor_filter/tensor_filter.c: In function ‘gst_tensor_filter_set_property’:
/<<PKGBUILDDIR>>/gst/tensor_filter/tensor_filter.c:560:51: error: comparison between pointer and zero character constant [-Werror=pointer-compare]
g_assert (prop->input_meta.info[i].name != '\0');
^
/<<PKGBUILDDIR>>/gst/tensor_filter/tensor_filter.c:560:21: note: did you mean to dereference the pointer?
g_assert (prop->input_meta.info[i].name != '\0');
^
/<<PKGBUILDDIR>>/gst/tensor_filter/tensor_filter.c:581:52: error: comparison between pointer and zero character constant [-Werror=pointer-compare]
g_assert (prop->output_meta.info[i].name != '\0');
^
/<<PKGBUILDDIR>>/gst/tensor_filter/tensor_filter.c:581:21: note: did you mean to dereference the pointer?
g_assert (prop->output_meta.info[i].name != '\0');
^
```
Signed-off-by: MyungJoo Ham <myungjoo.ham@samsung.com>
Jaeyun [Fri, 7 Dec 2018 10:40:57 +0000 (19:40 +0900)]
[Meson] add unittest
check dependency of gtest and add exe to run unittest
(command '$ ninja test' will build and start gtest)
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Fri, 7 Dec 2018 10:34:42 +0000 (19:34 +0900)]
[Meson] code clean
1. rename nnstreamer include path and dependencies
2. link proper objs
3. fix indentation
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>
Jaeyun [Fri, 7 Dec 2018 07:55:03 +0000 (16:55 +0900)]
[Meson] rename gst/glib dependency
rename dependencies of gstreamer, glib (follows gst-base)
Signed-off-by: Jaeyun Jung <jy1210.jung@samsung.com>