# create_docker

**Repository Path**: dwithj/create_docker

## Basic Information

- **Project Name**: create_docker
- **Description**: create mindspore docker
- **Primary Language**: Python
- **License**: Apache-2.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 2
- **Created**: 2024-10-03
- **Last Updated**: 2024-12-03

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# 构建mindspore的docker环境


```bash
# 以下操作基于欧拉2.8操作系统, docker版本18.09


# 进入 cdreate_docker_ubuntu20.04/build_base目录
sudo docker build -f Dockerfile_Base -t mindspore_ubuntu20.04_base:v1 .

# 进入 cdreate_docker_ubuntu20.04/build_mindspore_py310
wget https://www.python.org/ftp/python/3.10.14/Python-3.10.14.tgz --no-check-certificate --no-check-certificate
# 到官网下载mindspore, CANN, MindIE软件包

# ascend-toolkit包
sudo docker build --build-arg CANN_MODE=TOOLKIT \
--build-arg PYTHONPATH_PREFIX=/usr/local/python3.10.14 \
--build-arg CANN_PREFIX=/usr/local/Ascend \
--build-arg MINDIE_PREFIX=/usr/local/Ascend \
-f Dockerfile_py310 -t mindspore_2.3.rc1:v1 .


# 镜像更新参考，进到docker_update目录
# 把需要更新的软件包复制过来，详情参考Dockerfile_update
sudo docker build -f Dockerfile_update --tag mindspore:v2 .

# modelarts镜像，把npu-smi命令也复制到当前目录
sudo docker build -f Dockerfile_update_modelarts --tag mindspore_modelarts:v2 .

# modelarts镜像没有装moxing包，训练任务拉起钱执行
pip install /home/ma-user/modelarts/package/moxing_framework*.whl











# 以下操作基于openeuler-20.03操作系统, docker版本18.09
1. yum install -y docker

2. 下载基础镜像
wget https://repo.huaweicloud.com/openeuler/openEuler-20.03-LTS-SP3/docker_img/update/2023-09-09/aarch64/openEuler-docker.aarch64.tar.xz --no-check-certificate
xz -d openEuler-docker.aarch64.tar.xz
docker load -i openEuler-docker.aarch64.tar

# 进入 create_docker_openeuler/build_base目录

wget https://github.com/git-lfs/git-lfs/releases/download/v3.1.2/git-lfs-linux-arm64-v3.1.2.tar.gz --no-check-certificate
wget https://github.com/Kitware/CMake/releases/download/v3.18.0/cmake-3.18.0.tar.gz --no-check-certificate

docker build -f Docker_Base --tag mindspore_base:v1 .

# 进入build_mindspore_py39目录

wget https://www.python.org/ftp/python/3.9.9/Python-3.9.9.tgz --no-check-certificate --no-check-certificate

# CANN包和mindspore, mindspore_lite包自行下载

docker build -f Dockerfile_py39 --tag mindspore_py39:v1 .


# 镜像更新参考，进到docker_update目录
# 把容器里面Python和ascend-toolkit复制到当前目录
docker build -f Dockerfile_update --tag mindspore:v2

# modelarts镜像，把npu-smi命令也复制到当前目录
docker build -f Dockerfile_update_modelarts --tag mindspore_modelarts:v2

# modelarts镜像没有装moxing包，训练代码目录新建ma-pre-start.sh，内容如下
pip install pip --ugrade && pip install /home/ma-user/modelarts/package/moxing_framework*.whl
```