构建镜像

docker build -t ultralytics_yolov8 -f docker/Dockerfile-base-ultralytics .
docker tag ultralytics_yolov8 ${CNB_DOCKER_REGISTRY}/${CNB_REPO_SLUG_LOWERCASE}/ultralytics_yolov8:latest
docker push ${CNB_DOCKER_REGISTRY}/${CNB_REPO_SLUG_LOWERCASE}/ultralytics_yolov8:latest

拉取镜像运行

docker pull cnb.cool/rzhangsan/yolov8_on_rv1126b/ultralytics_yolov8/ultralytics_yolov8:latest
docker tag cnb.cool/rzhangsan/yolov8_on_rv1126b/ultralytics_yolov8/ultralytics_yolov8:latest ultralytics_yolov8:latest

# pt -> onnx
# 使用 ultralytics_yolov8 镜像，避免和开发环境的 py 库冲突
docker run -it --rm -v /workspace:/workspace -w /workspace ultralytics_yolov8:latest /bin/bash

# 检查版本是否是所期望的版本
yolo 

yolo export model=yolov8n.pt format=onnx opset=12
# 导出成功后会生成：yolov8n.onnx