参考链接

EdgeAndRobotics

进入YOLOV5MultiInput代码目录

cd EdgeAndRobotics/Samples/YOLOV5MultiInput/

安装依赖

略

获取PyTorch框架的YoloV5s模型(*.onnx)，并转换为昇腾AI处理器能识别的模型(*.om)

cd model
wget https://obs-9be7.obs.cn-east-2.myhuaweicloud.com/003_Atc_Models/yolov5s/yolov5s_nms.onnx --no-check-certificate
wget https://obs-9be7.obs.cn-east-2.myhuaweicloud.com/003_Atc_Models/yolov5s/aipp.cfg --no-check-certificate

模型转换

# 设置如下两个环境变量减少atc模型转换过程中使用的进程数，减小内存占用
export TE_PARALLEL_COMPILER=1
export MAX_COMPILE_CORE_NUMBER=1
atc --model=yolov5s_nms.onnx --framework=5 --output=yolov5s_nms --input_shape="images:1,3,640,640;img_info:1,4"  --soc_version=Ascend310B4 --insert_op_conf=aipp.cfg

准备测试视频

cd ../data/
wget https://obs-9be7.obs.cn-east-2.myhuaweicloud.com/003_Atc_Models/yolov5s/test.mp4 --no-check-certificate

安装opencv

apt update
apt install libopencv-dev

模型推理

编译样例源码

cd ../scripts
bash sample_build.sh

运行样例

bash sample_run.sh stdout
bash sample_run.sh imshow