YOLOv5 Lite PyTorch=>ONNX=>TensorRT

1.Export trt Model

python export.py ---weights weights/v5lite-g.pt --batch-size 1 --imgsz 640 --include onnx --simplify

trtexec --explicitBatch --onnx=./v5lite-g.onnx --saveEngine=v5lite-g.trt --fp16

2.Build yolov5 TensorRT Inference Project

mkdir build && cd build
cmake ..
make -j

3.Run yolov5_trt

• inference dir with v5lite-g

./yolov5_trt ../config.yaml ../samples

4.Results: