Jetson Nano使用TensorRT加速yolov3-tiny目标识别

文章目录

• ​​环境配置​​
• ​​运行TRT-yolov3​​
• ​​测试(识别)​​
• ​​模型替换(详细)​​

• ​​yolov3-tiny -> onnx​​
• ​​onnx -> trt​​

• ​​Resource​​

环境配置

• 安装protobuf

sudo pip3 install -i https://pypi.tuna.tsinghua.edu.cn/simple protobuf==3.8.0

• 安装onnx

sudo apt-get install protobuf-compiler libprotoc-dev

• 安装pillow

sudo pip3 install Pillow

• 安装pycuda：如果下述脚本无法安装，请看[[Jetson Nano安装pycuda(踩坑传)]]

export PATH=/usr/local/cuda/bin:\${PATH}
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:\${LD_LIBRARY_PATH}
sudo pip3 install pycuda

• 安装numpy

sudo pip3 install numpy

运行TRT-yolov3

标签：yolov3,Nano,onnx,py,TensorRT,tiny,416,cfg
From： https://blog.51cto.com/doubleZ/5986621