ffmpeg_openvino/decode_resize_inference build error

[KiwiHana]

dlstreamer/samples/ffmpeg_openvino/cpp/decode_resize_inference$ ./build_and_run.sh /home/test/kiwi/test_1080p_15.mp4 /home/test/kiwi/yolov5s_openvino_model/yolov5s.xml
CMake Warning (dev) in CMakeLists.txt:
  No project() command is present.  The top-level CMakeLists.txt file must
  contain a literal, direct call to the project() command.  Add a line of
  code such as

    project(ProjectName)

  near the top of the file, but after cmake_minimum_required().

  CMake is pretending there is a "project(Project)" command on the first
  line.
This warning is for project developers.  Use -Wno-dev to suppress it.

-- The C compiler identification is GNU 11.4.0
-- The CXX compiler identification is GNU 11.4.0
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found PkgConfig: /usr/bin/pkg-config (found version "0.29.2")
-- Checking for modules 'libavformat;libavcodec;libswscale;libavutil'
--   Found libavformat, version 60.16.100
--   Found libavcodec, version 60.31.102
--   Found libswscale, version 7.5.100
--   Found libavutil, version 58.29.100
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE
-- Looking for CL_VERSION_2_2
-- Looking for CL_VERSION_2_2 - found
-- Found OpenCL: /usr/lib/x86_64-linux-gnu/libOpenCL.so (found version "2.2")
-- Checking for module 'dl-streamer'
--   Found dl-streamer, version 2.0.0
-- Configuring done
-- Generating done
-- Build files have been written to: /home/test/intel/dl_streamer/samples/ffmpeg_openvino_decode_resize_inference/build
[ 50%] Building CXX object CMakeFiles/ffmpeg_openvino_decode_resize_inference.dir/decode_resize_inference.cpp.o
In file included from /opt/intel/openvino_2024.0/runtime/include/openvino/runtime/intel_gpu/ocl/ocl_wrapper.hpp:50,
                 from /opt/intel/openvino_2024.0/runtime/include/openvino/runtime/intel_gpu/ocl/ocl.hpp:17,
                 from /opt/intel/openvino_2024.0/runtime/include/openvino/runtime/intel_gpu/ocl/va.hpp:23,
                 from /home/test/kiwi/dlstreamer/samples/ffmpeg_openvino/cpp/decode_resize_inference/decode_resize_inference.cpp:20:
/usr/include/CL/cl2.hpp:18:151: note: ‘#pragma message: cl2.hpp has been renamed to opencl.hpp to make it clear that it supports all versions of OpenCL. Please include opencl.hpp directly.’
   18 | encl.hpp to make it clear that it supports all versions of OpenCL. Please include opencl.hpp directly.")
      |                                                                                                        ^

[100%] Linking CXX executable ffmpeg_openvino_decode_resize_inference
[100%] Built target ffmpeg_openvino_decode_resize_inference
terminate called after throwing an instance of 'ov::AssertFailure'
  what():  Check 'it != m_names.end()' failed at src/core/src/layout.cpp:203:
W dimension index is not defined

./build_and_run.sh: line 32: 94029 Aborted                 (core dumped) "${BUILD_DIR}"/${EXE_NAME} -i "${VIDEO_FILE}" -m "${MODEL}"

[brmarkus]

Just follow "./build_and_run.sh INPUT_VIDEO [MODEL]".

Can you try like this, please (without -i, without -m)?

$ ./build_and_run.sh /home/test/kiwi/test_1080p_15.mp4 /home/test/kiwi/yolov5s_openvino_model/yolov5s.xml

(-i and -m get added in the script)

[KiwiHana]

$/home/test/intel/dl_streamer/samples/ffmpeg_openvino_decode_resize_inference/build/ffmpeg_openvino_decode_resize_inference -i /home/test/kiwi/test_1080p_15.mp4 -m /home/test/kiwi/yolov5s_openvino_model/yolov5s.xml
terminate called after throwing an instance of 'ov::AssertFailure'
  what():  Check 'it != m_names.end()' failed at src/core/src/layout.cpp:203:
W dimension index is not defined

Aborted (core dumped)

[brmarkus]

Does it work with another model, like the default model face-detection-adas-0001 from the build-script, "https://github.com/dlstreamer/dlstreamer/blob/master/samples/ffmpeg_openvino/cpp/decode_resize_inference/build_and_run.sh#L9C33-L9C57"?

[KiwiHana]

Does it work with another model, like the default model face-detection-adas-0001 from the build-script, "https://github.com/dlstreamer/dlstreamer/blob/master/samples/ffmpeg_openvino/cpp/decode_resize_inference/build_and_run.sh#L9C33-L9C57"?

:~/intel/dl_streamer/samples/ffmpeg_openvino_decode_resize_inference/build$ ./ffmpeg_openvino_decode_resize_inference -i /home/test/kiwi/test_1080p_15.mp4 -m /home/test/kiwi/face-detection-adas-0001/FP32/face-detection-adas-0001.xml
model /home/test/kiwi/face-detection-adas-0001/FP32/face-detection-adas-0001.xml
read model
ffmpeg_openvino_decode_resize_inference: /home/dlstreamer/dlstreamer/src/vaapi/vaapi_batch_proc/vaapi_batch_proc.cpp:65: virtual dlstreamer::FramePtr dlstreamer::VaapiBatchProc::process(dlstreamer::FramePtr): Assertion `false' failed.
Aborted (core dumped)

[brmarkus]

Hmm, this looks bad...

The code at this assertion is this, see here: "https://github.com/dlstreamer/dlstreamer/blame/master/src/vaapi/vaapi_batch_proc/vaapi_batch_proc.cpp#L65"

// This code path is work in progress and not active in DLStreamer architecture 1.0
assert(false);

It got added in the last release "Release v2024.0"

Sounds like the port to a newer architecture is work-in-progress...?

Could you maybe use an older version...?

[pmalatyn]

hello, @brmarkus you are right this part is part of Architecture 2.0 which is under development now.
@KiwiHana we can fix your issue in release planned at the end of March. Will this help? Plus 2nd question - anything else from Architecture 2.0 that you are interested in?

[KiwiHana]

hello, @brmarkus you are right this part is part of Architecture 2.0 which is under development now. @KiwiHana we can fix your issue in release planned at the end of March. Will this help? Plus 2nd question - anything else from Architecture 2.0 that you are interested in?

Thanks @brmarkus and @pmalatyn ! March goes by, do you mean fix this issue on May?
Also, we are interested in more image preprocess in ffmpeg pipeline, like color conversion, clip, fill etc.

[pmalatyn]

yeah May not March ;)
About your use cases are you able to share user stories how you would like to use it? Preferably in a form like: "As a user I want to ..." ?