Dist/Tizen: Enable ncnn tensor_filter sub-plugin @open sesame 03/06 15:00

[wooksong]

As ncnn packages are released on the Tizen public repository,
this patch enables the ncnn sub-plugin for the tensor filter in Tizen.

Signed-off-by: Wook Song wook16.song@samsung.com

Self evaluation:

1. Build test: [*]Passed [ ]Failed [ ]Skipped
2. Run test: [ ]Passed [ ]Failed [*]Skipped

[taos-ci]

📝 TAOS-CI Version: 1.5.20200925. Thank you for submitting PR #4379. Please a submit 1commit/1PR (one commit per one PR) policy to get comments quickly from reviewers. Your PR must pass all verificiation processes of cibot before starting a review process from reviewers. If you are new member to join this project, please read manuals in documentation folder and wiki page. In order to monitor a progress status of your PR in more detail, visit http://ci.nnstreamer.ai/.

[wooksong]

Releasing of NCNN is currently WiP.

[taos-ci]

cibot: @wooksong, A builder checker could not be completed because one of the checkers is not completed. In order to find out a reason, please go to http://ci.nnstreamer.ai/nnstreamer/ci/repo-workers/pr-checker/4379-202401111318270.17357611656189-8575b45ff9b6a12267cb42fd6b7da37187bd2258/.

[songgot]

LGTM

[wooksong]

https://review.tizen.org/gerrit/gitweb?p=platform/upstream/ncnn.git;a=shortlog;h=refs/heads/sandbox/wooksong

NCNN is required to be published on Tizen.

[watermelon-copilot-for-code-review]

Watermelon AI Summary

This Pull Request integrates a newly released ncnn neural network component into Tizen's tensor filtering system, enhancing its machine learning capabilities, though it currently skips actual run testing.
Please create an account on Watermelon to get better results.

GitHub PRs

• #4379 - Dist/Tizen: Enable ncnn tensor_filter sub-plugin - By wooksong 1 months ago

• #1512 - [asan] Enable Address SANitizer for the debian build @open sesame 08/23 14:22 - By kparichay 2 years ago

• #1658 - [tensor] Power Tensor::apply() with c++17 parallel execution - By kparichay 2 years ago

nnstreamer is an open repo and Watermelon will serve it for free.
🍉🫶

[myungjoo]

[ 556s] [FAILED] nnstreamer_filter_ncnn (6 passed among 9 cases)

[myungjoo]

Issue 1: All positive unit TCs failed.
Issue 2: Negative TCs do not have the postfix, _n.

[myungjoo]

Issue 1 of TC 1/2 is caused by gstreamer version mismatch with your assumptions.
Older GStreamer (e.g., 1.14) does not have num-buffers property in the imagefreeze element.

[wooksong]

Issue 1: All positive unit TCs failed.
Issue 2: Negative TCs do not have the postfix, _n.
Issue 1 of TC 1/2 is caused by gstreamer version mismatch with your assumptions. Older GStreamer (e.g., 1.14) does not have num-buffers property in the imagefreeze element.

@myungjoo Well, NCNN has been published, and I almost forgot about this PR.
Even though I was only in charge of the packaging part (the implementation of the filter and TCs was the work of the SNU students), I will take a look as you commented. Thanks!

[wooksong]

Issue 2: Negative TCs do not have the postfix, _n.

This trivial issue has been resolved.

[wooksong]

Issue 1 of TC 1/2 is caused by gstreamer version mismatch with your assumptions. Older GStreamer (e.g., 1.14) does not have num-buffers property in the imagefreeze element.

The two failed TCs involve several sub-issues. I will summarize them and post them on the Issues. Merging them may take some time.

[myungjoo]

You may simple check gstreamer version (or property availability with gst-inspect) at the test script to skip this instance if it's old.