Sign inSign up

jungin500/nni

By jungin500

•Updated about 3 years ago

Tencent/NNI master branch montly build with Raspberry Pi 3B/4B (armv7l, arm64 respectively) + amd64

Image
0

238

jungin500/nni repository overview

⁠Usage

  • docker run -it --rm jungin500/nni:22.05
  • jungin500/nni:22.05 contains all architecture and will be automatically choosen.
  • Check out Workspace location for built files location and space to begin with.

⁠Tags by architecture

  • AMD64 (x86_64): jungin500/nni:22.05-amd64
  • ARM64 (arm64, aarch64): jungin500/nni:22.05-arm64v8
  • ARM32 (arm32, armv7l): jungin500/nni:22.05-arm32v7

⁠Workspace location

  • Code: /workspace/ncnn
  • Build artifacts: /workspace/ncnn_build
  • Binary: /usr/local

⁠Miscelloneous

⁠Dockerfile
# Assume Raspberry Pi OS 11 (Bullseye)
FROM debian:11

RUN set -ex &&\
    \
    sed -i "s/deb.debian.org/ftp.harukasan.org/g" /etc/apt/sources.list &&\
    apt-get -qq update &&\
    apt-get -qq install git wget curl cmake build-essential ninja-build libprotobuf-dev protobuf-compiler &&\
    rm -rf /var/lib/apt/lists

WORKDIR /workspace

RUN set -ex &&\
    \
    cd /workspace &&\
    git clone --recursive https://github.com/Tencent/ncnn &&\
    mkdir ncnn_build &&\
    cd ncnn_build &&\
    cmake \
        -D CMAKE_BUILD_TYPE=Release \
        -D CMAKE_INSTALL_PREFIX=/usr/local \
        -D NCNN_DISABLE_RTTI=OFF \
        -D NCNN_BUILD_EXAMPLES=ON \
        -D NCNN_BUILD_TOOLS=ON \
        -D CMAKE_TOOLCHAIN_FILE=../ncnn/toolchains/aarch64-linux-gnu.toolchain.cmake \
        ../ncnn &&\
    make -j4 &&\
    make install
⁠Binary list
/usr/local/lib/libncnn.a
/usr/local/include/ncnn/allocator.h
/usr/local/include/ncnn/benchmark.h
/usr/local/include/ncnn/blob.h
/usr/local/include/ncnn/c_api.h
/usr/local/include/ncnn/command.h
/usr/local/include/ncnn/cpu.h
/usr/local/include/ncnn/datareader.h
/usr/local/include/ncnn/gpu.h
/usr/local/include/ncnn/layer.h
/usr/local/include/ncnn/layer_shader_type.h
/usr/local/include/ncnn/layer_type.h
/usr/local/include/ncnn/mat.h
/usr/local/include/ncnn/modelbin.h
/usr/local/include/ncnn/net.h
/usr/local/include/ncnn/option.h
/usr/local/include/ncnn/paramdict.h
/usr/local/include/ncnn/pipeline.h
/usr/local/include/ncnn/pipelinecache.h
/usr/local/include/ncnn/simpleocv.h
/usr/local/include/ncnn/simpleomp.h
/usr/local/include/ncnn/simplestl.h
/usr/local/include/ncnn/vulkan_header_fix.h
/usr/local/include/ncnn/ncnn_export.h
/usr/local/include/ncnn/layer_shader_type_enum.h
/usr/local/include/ncnn/layer_type_enum.h
/usr/local/include/ncnn/platform.h
/usr/local/lib/cmake/ncnn/ncnn.cmake
/usr/local/lib/cmake/ncnn/ncnn-release.cmake
/usr/local/lib/cmake/ncnn/ncnnConfig.cmake
/usr/local/bin/ncnn2mem
/usr/local/bin/ncnnmerge
/usr/local/bin/ncnnoptimize
/usr/local/bin/caffe2ncnn
/usr/local/bin/mxnet2ncnn
/usr/local/bin/onnx2ncnn
/usr/local/bin/darknet2ncnn
/usr/local/bin/ncnn2table
/usr/local/bin/ncnn2int8

Tag summary

Content type

Image

Digest

sha256:61021cb4d…

Size

3.1 GB

Last updated

about 3 years ago

docker pull jungin500/nni:2.10.1