This document mainly introduces how to compile and use NPU applications in Edge2.

Clone the source code form our Github.

$ git clone https://github.com/khadas/edge2-npu

There are two types of application source code in C++ and Python.

Enter Python directory,

$ cd Python

Install dependences,

$ sudo cp ../C++/runtime/librknn_api/aarch64/librknnrt.so /usr/lib
$ sudo apt-get install -y python3-dev python3-pip
$ sudo apt-get install -y python3-opencv python3-numpy
$ pip3 install ./wheel/rknn_toolkit_lite2-1.3.0-cp310-cp310-linux_aarch64.whl

Take yolov5 as an example, other demos are similar.

$ cd resnet18
$ python3 resnet18.py 
--> Load RKNN model
done
--> Init runtime environment
I RKNN: [17:07:10.282] RKNN Runtime Information: librknnrt version: 1.3.0 (c193be371@2022-05-04T20:16:33)
I RKNN: [17:07:10.282] RKNN Driver Information: version: 0.7.2
I RKNN: [17:07:10.282] RKNN Model Information: version: 1, toolkit version: 1.3.0-11912b58(compiler version: 1.3.0 (c193be371@2022-05-04T20:23:58)), target: RKNPU v2, target platform: rk3588, framework name: PyTorch, framework layout: NCHW
done
--> Running model
resnet18
-----TOP 5-----
[812]: 0.9996383190155029
[404]: 0.00028062646742910147
[657]: 1.632110434002243e-05
[833 895]: 1.015904672385659e-05
[833 895]: 1.015904672385659e-05
 
done

Enter C++ directory,take yolov5 as an example, other demos are similar.

$ cd C++

Install dependences,

$ sudo apt install cmake

Compile,

$ cd yolov5
$ ./build.sh

Run.

$ cd install/yolov5
$ ./yolov5 data/model/yolov5s-640-640.rknn data/img/bus.jpg 
post process config: box_conf_threshold = 0.50, nms_threshold = 0.60
Read data/img/bus.jpg ...
img width = 640, img height = 640
Loading mode...
sdk version: 1.3.0 (c193be371@2022-05-04T20:16:33) driver version: 0.7.2
model input num: 1, output num: 3
  index=0, name=images, n_dims=4, dims=[1, 640, 640, 3], n_elems=1228800, size=1228800, fmt=NHWC, type=INT8, qnt_type=AFFINE, zp=-128, scale=0.003922
  index=0, name=output, n_dims=5, dims=[1, 3, 85, 80], n_elems=1632000, size=1632000, fmt=UNDEFINED, type=INT8, qnt_type=AFFINE, zp=77, scale=0.080445
  index=1, name=371, n_dims=5, dims=[1, 3, 85, 40], n_elems=408000, size=408000, fmt=UNDEFINED, type=INT8, qnt_type=AFFINE, zp=56, scale=0.080794
  index=2, name=390, n_dims=5, dims=[1, 3, 85, 20], n_elems=102000, size=102000, fmt=UNDEFINED, type=INT8, qnt_type=AFFINE, zp=69, scale=0.081305
model is NHWC input fmt
model input height=640, width=640, channel=3
once run use 32.872000 ms
loadLabelName ./data/coco_80_labels_list.txt
person @ (474 250 559 523) 0.996784
person @ (112 238 208 521) 0.992214
bus @ (99 141 557 445) 0.976798
person @ (211 242 285 509) 0.976798
loop count = 10 , average run  26.577900 ms