算法部署-使用TensorRT部署图像分割算法PIDNet-优质算法部署项目实战.zip

# PIDNet.TensorRT TensorRT project of [PIDNet](https://github.com/XuJiacong/PIDNet) ### 1. export weights as *.wts see custom.py. put custom.py into the pytorch project `tools/`, set `--w` to be `True`: ``` python tools/custom.py ```  to test png images and export weights as `PIDNet.wts`. #### 2. check TensorRT path in CMakeList.txt and wts path in pidnet.cpp #### 3. Convert wts to TensorRT engine ``` mkdir build cd build cmake .. make -j8 ./main -s ``` run above and convert trt engine `PIDNetS.engine`. #### 4. Demo put *.png images into folder samples and run: ``` ./main -d samples ```