YOLOX example

prepare: download MegCC release package

1. the MegCC release package can be get in MegCC Github repo https://github.com/MegEngine/MegCC
2. unzip your MegCC release package into your_release_MegCC_dir

step1: change directory to yolox_example

cd <your_release_MegCC_dir>/yolox_example

step2: get MegEngine yolox pretrained model

wget https://github.com/Megvii-BaseDetection/storage/releases/download/0.0.1/yolox_s.mge

step3: convert MegEngine model into MegCC mmodel

mkdir -p kernel_yolox_s_arm 

# build MegCC yolox kernel and MegCC lib for arm 
<your_release_MegCC_dir>/bin/mgb-to-tinynn --json="./yolox_arm.json" --arm64v7
python3 <your_release_MegCC_dir>/runtime/scripts/runtime_build.py --cross_build --kernel_dir ./kernel_yolox_s_arm/ --remove_old_build --cross_build_target_arch aarch64

step4: get the opencv andorid SDK

wget https://github.com/opencv/opencv/releases/download/4.6.0/opencv-4.6.0-android-sdk.zip
unzip opencv-4.6.0-android-sdk.zip 
mv  OpenCV-android-sdk OpenCV 

step5: build example test

mkdir -p build_arm64 && cd build_arm64 && mkdir -p install

export NDK_DIR=<your_NDK_DIR> 
cmake .. -DCMAKE_TOOLCHAIN_FILE="${NDK_DIR}/build/cmake/android.toolchain.cmake"  -DANDROID_NDK="$NDK_DIR" -DANDROID_ABI=arm64-v8a  -DANDROID_NATIVE_API_LEVEL=21 -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=$PWD/install -DRUNTIME_KERNEL_DIR=$PWD/../kernel_yolox_s_arm -DOpenCV_DIR=$PWD/../OpenCV/sdk/native/jni/abi-arm64-v8a
make install/strip -j32

step6: run the test

1. copy the build_arm64/install/yolox_test yolox_example/dog.png yolox_example/kernel_yolox_s_arm/yolox_s.tiny to your android device(scp over termux or adb push)
2. run test with cmdline:

./yolox_test yolox_s.tiny --input=dog.jpg --output=<your_output_image>

3. copy your_output_image back to your local machine, the result is shown in the image

example result

origin：

detection output：