diff options
| -rw-r--r-- | packaging/ncnn.manifest | 5 | ||||
| -rw-r--r-- | packaging/ncnn.spec | 111 |
2 files changed, 116 insertions, 0 deletions
diff --git a/packaging/ncnn.manifest b/packaging/ncnn.manifest new file mode 100644 index 0000000..017d22d --- /dev/null +++ b/packaging/ncnn.manifest @@ -0,0 +1,5 @@ +<manifest> + <request> + <domain name="_"/> + </request> +</manifest> diff --git a/packaging/ncnn.spec b/packaging/ncnn.spec new file mode 100644 index 0000000..f921d3f --- /dev/null +++ b/packaging/ncnn.spec @@ -0,0 +1,111 @@ +%define builddir build +%define upstream_release_version 20240102 + +########################################################################### +# Package and sub-package definitions +Name: ncnn +Summary: A high-performance neural network inference framework optimized for the mobile platform +Version: %{upstream_release_version} +Release: 1e88fb8 +Group: Machine Learning/ML Framework +Packager: Wook Song <wook16.song@samsung.com> +License: BSD-3-Clause +Source0: %{name}-%{version}.tar +Source1001: %{name}.manifest + +## Define build requirements ## +BuildRequires: cmake +BuildRequires: ninja +BuildRequires: protobuf-lite-devel +BuildRequires: opencv-devel +BuildRequires: vulkan-loader-devel + +## Define Packages ## +%description +ncnn is a high-performance neural network inference computing framework +optimized for mobile platforms. ncnn is deeply considerate about deployment +and uses on mobile phones from the beginning of design. ncnn does not have +third party dependencies. It is cross-platform, and runs faster than all known +open source frameworks on mobile phone cpu. Developers can easily deploy deep +learning algorithm models to the mobile platform by using efficient ncnn +implementation, create intelligent APPs, and bring the artificial intelligence +to your fingertips. ncnn is currently being used in many Tencent applications, +such as QQ, Qzone, WeChat, Pitu and so on. + +%package devel +Summary: Development package for the ncnn framework +%description devel +Development package for the ncnn framework. +This contains corresponding header files and static archives. + +%package tools +Summary: Binary package for tools included in the ncnn framework +%description tools +This is a binary package that provides tools and model converters included in +the ncnn framework. + +%package examples +Summary: Binary package for native examples +%description examples +This is a binary package that contains native examples developed based on +the APIs of the ncnn framework. + +%prep +rm -rf ./%{builddir} +%setup -q +cp %{SOURCE1001} . + +%build +CXXFLAGS=`echo -std=c++11 -fno-builtin $CXXFLAGS` +%define cmake_common_options -DNCNN_VERSION=%{upstream_release_version}-%{release} -DCMAKE_BUILD_TYPE=release -DNCNN_SHARED_LIB=ON -G Ninja +%ifnarch aarch64 i586 x86_64 +%define cmake_arch_options -DNCNN_ENABLE_LTO=OFF +%else +%define cmake_arch_options -DNCNN_ENABLE_LTO=ON +%endif + +mkdir -p %{builddir} +pushd %{builddir} +cmake .. %{cmake_common_options} %{cmake_arch_options} +cmake --build . +popd + +%install +pushd %{builddir} +DESTDIR=%{buildroot} cmake --install . --prefix %{_prefix} +install -p -m 755 benchmark/benchncnn %{buildroot}%{_bindir} +mkdir -p %{buildroot}%{_bindir}/ncnn_examples +find examples -type f -executable -exec install -p -m 755 {} %{buildroot}%{_bindir}/ncnn_examples \; +popd + +%post -p /sbin/ldconfig + +%postun -p /sbin/ldconfig + +%files +%manifest %{name}.manifest +%defattr(-,root,root,-) +%license LICENSE.txt +%{_libdir}/*.so.* +%{_bindir}/benchncnn + +%files devel +%defattr(-,root,root,-) +%{_includedir}/ncnn/*.h +%{_libdir}/pkgconfig/*.pc +%{_libdir}/cmake/ncnn/*.cmake +%{_libdir}/*.so + +%files tools +%{_bindir}/ncnn2mem +%{_bindir}/ncnnmerge +%{_bindir}/ncnnoptimize +%{_bindir}/caffe2ncnn +%{_bindir}/darknet2ncnn +%{_bindir}/mxnet2ncnn +%{_bindir}/ncnn2int8 +%{_bindir}/ncnn2table +%{_bindir}/onnx2ncnn + +%files examples +%{_bindir}/ncnn_examples/* |