- imagenet_classification(squeezenet, mobilenet, mobilenet_v2, resnet50, alexnet, googlenet, inception_v3, inception_v4, vgg16)
- mtcnn
- ssd
- mobilenet_ssd
- yolov2
- faster_rcnn
- lighten_cnn
- tengine_model
- classification(squeezenet, mobilenet, mobilenet_v2, resnet50, alexnet, googlenet, inception_v3, inception_v4, vgg16)
- convert(convert caffe model to tengine model)
Download the models from Tengine model zoo (psw: hhgc).
1. install Tengine
sudo apt-get install libopencv-dev
cd ~/tengine/examples
vim linux_build.sh
Set the correct Tengine path.
mkdir build
cd build
../linux_build.sh
make -j4
cd ~/tengine/examples
vim android_build_armv7.sh or vim android_build_armv8.sh
Set the correct NDK path, Tengine path, Opencv path and protobuf path. If you want to run Tengine with OpenBlas, please add the correct blas path in android_build_armv7.sh or android_build_armv8.sh.
-DBLAS_DIR=/home/usr/openbla020_android
mkdir build
cd build
../android_build_armv7.sh or ../android_build_armv8.sh
make -j4