Sign inSign up

tunmx/solexcv_ai_develop_cuda

By tunmx

Updated almost 3 years ago

SolexCV-AI Development environment with opencv4.1, cuda11.2, mnn2.2.0, protobuf3.6, python3.8.

Image
0

1.7K

tunmx/solexcv_ai_develop_cuda repository overview

SolexCV-AI-Develop

A dedicated development/deployment environment image for the SolexCV-AI project.

   _____       __          _______    __      ___    ____
  / ___/____  / /__  _  __/ ____/ |  / /     /   |  /  _/
  \__ \/ __ \/ / _ \| |/_/ /    | | / /_____/ /| |  / /
 ___/ / /_/ / /  __/>  </ /___  | |/ /_____/ ___ |_/ /
/____/\____/_/\___/_/|_|\____/  |___/     /_/  |_/___/
                  by tunmx, solexcv_ai_develop_cuda/1.0a

System

Software version of the built-in environment

  • System: Ubuntu18.04
  • Core: 5.10.124-linuxkit
  • gcc/g++: 7.5.0
  • CUDA: 11.2
  • cudnn: 8
  • CMake: 3.10.2
  • Protobuf: 3.6
  • OpenCV: 4.5.1(with contrib)、3.4.1(armv7)
  • MNN: 2.2.0
  • SolexCV: 0.0.2
  • Python: 3.8(runtime)、3.6、2.7
  • arm-linux-gnueabihf: gcc-arm-8.3-2019.03-x86_64
  • xmake: v2.7.3

Develop & Building

Pull this image.

docker pull tunmx/solexcv_ai_develop_cuda:latest

You can use the image as an operating system for software development by using the -it command to access the interactive page:

docker run -it IMAGE_ID

In addition, you can use it to build the Dockerfile file and compile the SolexCV-AI-related projects and dependencies that you need to compile. We put the tripartite dependencies in the /3rdparty location:

FROM tunmx/solexcv_ai_develop_cuda:latest

MAINTAINER jingyuyan<[email protected]>

WORKDIR /workspace

COPY . /workspace

RUN rm -rf 3rdparty

RUN ln -sf /usr/share/zoneinfo/Asia/Shanghai /etc/localtime
RUN echo 'Asia/Shanghai' >/etc/timezone

RUN ln -s /3rdparty .

# push you c/c++ project build script

The advantage of using docker-compose to build and deploy projects is that you can build multiple tasks using the services configuration in a single description file, and you can flexibly compose and compose multiple containers for compiling, building, testing, and deploying projects. If you use docker-compose, you need to create docker-compose.yml file for description file writing. Currently, it is recommended to use the compile, test, and runtime deployment styles as shown in the following example:


version: "3"
services:
    build_cuda:		# Compile backend for cuda
        container_name: YOUR_CONTAINER_NAME		# The container name of the compilation work
        image: tunmx/solexcv_ai_develop_cuda:latest
        volumes:
            - .:/work
        working_dir: /work
        command: bash command/build_linux_cuda.sh /3rdparty		# Custom compiled bash scripts
        tty: true
    
    run_test_cuda: # Execute the cuda version of the test case
        container_name: YOUR_CONTAINER_NAME
        image: tunmx/solexcv_ai_develop_cuda:latest
        volumes:
            - .:/work
        working_dir: /work
        command: bash command/test.sh cuda
        tty: true
    
    server_runtime_cuda: # Run CUDA-based runtime services
        container_name: YOUR_CONTAINER_NAME
        image: tunmx/solexcv_ai_develop_cuda:latest
        volumes:
            - .:/work
        working_dir: /work
        command: bash command/run_server.sh cuda
        tty: true

After completing the description file, you can execute the build script to perform the build:

# Compile backend for cuda
docker-compose up build_cuda

# The container name of the compilation work
docker-compose up run_test_cuda

# Execute the cuda version of the test case
docker-compose up server_runtime_cuda

Deployment & Serving

In addition to development, you can use the image for rapid deployment of SolexCV-AI-related projects. The image provides C/C++ and Python3.8 development environments. You can build your software deployment requirements in the Dockerfile and quickly build deployment services. The advantage of using this mirroring deployment is that the configuration is consistent with the development environment, reducing the error rate during deployment.

Tag summary

Content type

Image

Digest

sha256:86c2570e3

Size

6.4 GB

Last updated

almost 3 years ago

docker pull tunmx/solexcv_ai_develop_cuda