docker的conda环境中安装mindspore（CPU版本？）

已于 2022-11-11 13:20:14 修改
阅读量834 收藏 1
点赞数
分类专栏： Linux python 文章标签： docker conda python
于 2022-11-11 13:19:02 首次发布
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/qq_39306047/article/details/127804406
版权
Linux 同时被 2 个专栏收录
78 篇文章 3 订阅
订阅专栏
python
58 篇文章 1 订阅
订阅专栏
该篇博客记录了在Docker环境下安装MindSpore CPU版本的过程，包括conda环境配置、清华源的使用以及MindSpore GPU版本安装失败后转为安装CPU版本的尝试。在安装过程中遇到CUDA版本不匹配的问题，最终成功安装CPU版本但导入时仍报错找不到libcudnn.so.7库文件。
摘要由CSDN通过智能技术生成

docker的conda环境中安装ms （CPU版本？）

export PATH=/root/miniconda3/envs/GLUE/bin:$PATH

conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free

conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/

conda config --set show_channel_urls yes
pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/1.9.0/MindSpore/gpu/x86_64/cuda-10.1/mindspore_gpu-1.9.0-cp38-cp38-linux_x86_64.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple

pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/1.9.0/MindSpore/gpu/x86_64/cuda-10.1/mindspore_gpu-1.9.0-cp38-cp38-linux_x86_64.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple

(GLUE) root@cd163c5c74c9:~/miniconda3/envs/GLUE/bin# conda install -c conda-forge -c bioconda scglue pytorch-gpu jupyterlab
Collecting package metadata (current_repodata.json): done
Solving environment: done

docker的conda环境中安装ms （CPU版本？）
(ms19) root@0b5003cae16c:/workspace# pip install scglue pytorch-gpu jupyterlab -i https://pypi.tuna.tsinghua.edu.cn/simple

pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/1.9.0/MindSpore/gpu/x86_64/cuda-10.1/mindspore_gpu-1.9.0-cp38-cp38-linux_x86_64.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple
pip 安装ms缺乏下面
Downloading and Extracting Packages
cudatoolkit-10.1.243 | 427.6 MB | 3
(ms19) root@0b5003cae16c:/workspace# conda install mindspore-gpu=1.9.0 cudatoolkit=10.1 -c mindspore -c conda-forge ####用conda比较好
No LSB modules are available.
CUDA 10.1 is not supported on Ubuntu 20.04

#####安装cuda11的
pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/1.9.0/MindSpore/gpu/x86_64/cuda-11.1/mindspore_gpu-1.9.0-cp38-cp38-linux_x86_64.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple
conda install mindspore-gpu=1.9.0 cudatoolkit=11.1 -c mindspore -c conda-forge

注意参考下方安装指南，添加运行所需的环境变量配置

(ms19) root@0b5003cae16c:/workspace# echo $LD_LIBRARY_PATH
/usr/local/nvidia/lib:/usr/local/nvidia/lib64
(ms19) root@0b5003cae16c:/workspace#

./opt/conda/pkgs/cudatoolkit-10.1.243-h8cb64d8_10/lib
./opt/conda/envs/ms19/lib/libcublas.so.10
./opt/conda/lib/libcublas.so
./opt/conda/lib/libcublas.so.11
./opt/conda/lib/libcublas.so.11.5.1.109
./opt/conda/lib/libcublas.so
./opt/conda/lib/libcublas.so.11
./opt/conda/lib/libcublas.so.11.5.1.109
./opt/conda/envs/ms19/lib/libcublas.so
./opt/conda/envs/ms19/lib/python3.8/site-packages/nvidia/cublas/lib/libcublas.so.11
./opt/conda/envs/ms19/lib/libcublas.so.10.2.1.243
./opt/conda/envs/ms19/lib/libcublas.so.10

[mindspore/run_check/_check_version.py:194] Cuda [‘10.1’, ‘11.1’, ‘11.6’] version(libcu*.so need by mindspore-gpu) is not found, please confirm that the path of cuda is set to the env LD_LIBRARY_PATH, or check whether the CUDA version in wheel package and the CUDA runtime in current device matches, please refer to the installation guidelines: https://www.mindspore.cn/install

(ms19) root@0b5003cae16c:/# echo $LD_LIBRARY_PATH
/usr/local/nvidia/lib:/usr/local/nvidia/lib64
(ms19) root@0b5003cae16c:/#

export LD_LIBRARY_PATH=/opt/conda/envs/ms19/lib/:$LD_LIBRARY_PATH

./opt/conda/envs/ms19/lib/python3.8/site-packages/nvidia_cudnn_cu11-8.5.0.96.dist-info
./opt/conda/envs/ms19/lib/python3.8/site-packages/nvidia_cuda_nvrtc_cu11-11.7.99.dist-info
./opt/conda/envs/ms19/lib/python3.8/site-packages/nvidia_cuda_runtime_cu11-11.7.99.dist-info
./opt/conda/envs/ms19/lib/python3.8/site-packages/nvidia
./opt/conda/envs/ms19/lib/python3.8/site-packages/nvidia_cublas_cu11-11.10.3.66.dist-info
export LD_LIBRARY_PATH=/opt/conda/envs/ms19/lib/python3.8/site-packages/:$LD_LIBRARY_PATH
这里安装ms是失败的！
(base) root@0b5003cae16c:/# conda activate ms19
(ms19) root@0b5003cae16c:/# uname -a
Linux 0b5003cae16c 5.10.16.3-microsoft-standard-WSL2 #1 SMP Fri Apr 2 22:23:49 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux
(ms19) root@0b5003cae16c:/#
因为docker 镜像 不能装 显卡的驱动 所以GPU 版本安装不了下面安装CPU
pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/1.9.0/MindSpore/cpu/x86_64/mindspore-1.9.0-cp38-cp38-linux_x86_64.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple
(ms19) root@0b5003cae16c:/# python
Python 3.8.13 (default, Oct 21 2022, 23:50:54)
[GCC 11.2.0] :: Anaconda, Inc. on linux
Type “help”, “copyright”, “credits” or “license” for more information.

import mindspore
[ERROR] ME(28009,7ff33eaaf6c0,python):2022-11-11-13:11:45.822.587 [mindspore/ccsrc/runtime/hardware/device_context_manager.cc:46] LoadDynamicLib] Load dynamic library libmindspore_gpu failed, returns [libcudnn.so.7: cannot open shared object file: No such file or directory].
quit()
(ms19) root@0b5003cae16c:/# pip uninstall mindspore
Found existing installation: mindspore 1.9.0
Uninstalling mindspore-1.9.0:
Would remove:
/opt/conda/envs/ms19/bin/cache_admin
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore-1.9.0.dist-info/*
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/*
Would not remove (might be manually added):
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/_ms_mpi.cpython-38-x86_64-linux-gnu.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libcuda_ops.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libgpu_collective.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libmpi_adapter.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libmpi_collective.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libnccl.so.2
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libnvidia_collective.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/plugin/libmindspore_gpu.so
Proceed (Y/n)? y
Successfully uninstalled mindspore-1.9.0
WARNING: Running pip as the ‘root’ user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
(ms19) root@0b5003cae16c:/# python
Python 3.8.13 (default, Oct 21 2022, 23:50:54)
[GCC 11.2.0] :: Anaconda, Inc. on linux
Type “help”, “copyright”, “credits” or “license” for more information.
import mindspore

quit()
(ms19) root@0b5003cae16c:/# pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/1.9.0/MindSpore/cpu/x86_64/mindspore-1.9.0-cp38-cp38-linux_x86_64.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple
Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple
Collecting mindspore1.9.0
Using cached https://ms-release.obs.cn-north-4.myhuaweicloud.com/1.9.0/MindSpore/cpu/x86_64/mindspore-1.9.0-cp38-cp38-linux_x86_64.whl (158.7 MB)
Requirement already satisfied: packaging>=20.0 in /opt/conda/envs/ms19/lib/python3.8/site-packages (from mindspore1.9.0) (21.3)
Requirement already satisfied: psutil>=5.6.1 in /opt/conda/envs/ms19/lib/python3.8/site-packages (from mindspore1.9.0) (5.9.4)
Requirement already satisfied: pillow>=6.2.0 in /opt/conda/envs/ms19/lib/python3.8/site-packages (from mindspore1.9.0) (9.3.0)
Requirement already satisfied: numpy>=1.17.0 in /opt/conda/envs/ms19/lib/python3.8/site-packages (from mindspore1.9.0) (1.23.4)
Requirement already satisfied: astunparse>=1.6.3 in /opt/conda/envs/ms19/lib/python3.8/site-packages (from mindspore1.9.0) (1.6.3)
Requirement already satisfied: scipy>=1.5.2 in /opt/conda/envs/ms19/lib/python3.8/site-packages (from mindspore1.9.0) (1.9.3)
Requirement already satisfied: protobuf>=3.13.0 in /opt/conda/envs/ms19/lib/python3.8/site-packages (from mindspore1.9.0) (3.20.1)
Requirement already satisfied: asttokens>=2.0.4 in /opt/conda/envs/ms19/lib/python3.8/site-packages (from mindspore1.9.0) (2.1.0)
Requirement already satisfied: six in /opt/conda/envs/ms19/lib/python3.8/site-packages (from asttokens>=2.0.4->mindspore1.9.0) (1.16.0)
Requirement already satisfied: wheel<1.0,>=0.23.0 in /opt/conda/envs/ms19/lib/python3.8/site-packages (from astunparse>=1.6.3->mindspore1.9.0) (0.37.1)
Requirement already satisfied: pyparsing!=3.0.5,>=2.0.2 in /opt/conda/envs/ms19/lib/python3.8/site-packages (from packaging>=20.0->mindspore1.9.0) (3.0.9)
Installing collected packages: mindspore
Successfully installed mindspore-1.9.0
WARNING: Running pip as the ‘root’ user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
(ms19) root@0b5003cae16c:/#
[GCC 11.2.0] :: Anaconda, Inc. on linux
Type “help”, “copyright”, “credits” or “license” for more information.
import mindspore
[ERROR] ME(26924,7f33a41ca6c0,python):2022-11-11-13:14:25.916.924 [mindspore/ccsrc/runtime/hardware/device_context_manager.cc:46] LoadDynamicLib] Load dynamic library libmindspore_gpu failed, returns [libcudnn.so.7: cannot open shared object file: No such file or directory].
quit()
(ms19) root@0b5003cae16c:/# pip uninstall mindspore
Found existing installation: mindspore 1.9.0
Uninstalling mindspore-1.9.0:
Would remove:
/opt/conda/envs/ms19/bin/cache_admin
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore-1.9.0.dist-info/*
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/*
Would not remove (might be manually added):
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/_ms_mpi.cpython-38-x86_64-linux-gnu.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libcuda_ops.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libgpu_collective.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libmpi_adapter.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libmpi_collective.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libnccl.so.2
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/libnvidia_collective.so
/opt/conda/envs/ms19/lib/python3.8/site-packages/mindspore/lib/plugin/libmindspore_gpu.so
Proceed (Y/n)? Y
Successfully uninstalled mindspore-1.9.0
WARNING: Running pip as the ‘root’ user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv
(ms19) root@0b5003cae16c:/#
import mindspore
(ms19) root@0b5003cae16c:/# python
Python 3.8.13 (default, Oct 21 2022, 23:50:54)
[GCC 11.2.0] :: Anaconda, Inc. on linux
Type “help”, “copyright”, “credits” or “license” for more information.
import mindspore
import torch
import scglue

qq_39306047
关注
专栏目录
docker安装conda
ppp092
09-10 4082
记录docker安装conda遇到的问题 镜像 https://hub.docker.com/r/continuumio/anaconda3 docker-anaconda 带有可立即使用的Anaconda（基于Python 3.5）的自举式安装的Docker容器。 Anaconda发行版已安装到该/opt/conda文件夹中，并确保默认用户conda在其路径中具有命令。 Anaconda是由Python支持的领先的开放数据科学平台。Anaconda的开源版本是一个高性能发行版，其中包括100多个用于数据
【MindSpore】Docker上成功使用MindSpore1.0.0的GPU版本
sun_5flower的博客
09-27 505
本文是在宿主机Ubuntu16.04上安装Docker，并成功进行MindSpore1.0.0的GPU训练； Ubuntu 16.04 Docker Nvidia-docker (cuda:10.1-base) mindspore-gpu:1.0.0 Dokcer hub上配置好的mindspore-gpu镜像 一、 先讲述成功的流程 【1】mindspore官网GPU安装教程查看具体流程 系统需要ubuntu18.04，但是我的ubuntu系统版本是16.04，因为是多人共用系统，不能直接升级系统
就是越来越好用，MindSpore自动安装！Conda方式安装MindSpore CPU版本
BaldheadedM的博客
07-15 510
win10x641.先安装conda环境2.创建并进入Conda虚拟环境通过condaconfig--addchannels添加国内镜像源修改后的.condarc文件，路径C\Users\xxx正在上传…重新上传取消3.确认处于Conda虚拟环境中，并执行如下命令安装最新版本的MindSpore。4.验证是否成功安装。...
Conda方式安装MindSpore CPU版本（Windows）
qiang______zi的博客
03-24 2551
Conda方式安装MindSpore CPU版本（Windows）
Windows 11安装NVIDIA+CUDA(带cuDNN)
taoshiwu的博客
09-29 1557
介绍了Win11下安装NVIDIA驱动和配置CUDA+cuDNN环境教程。
Conda方式安装MindSpore CPU版本-Windows
John_Snowww的博客
06-21 643
Conda方式安装MindSpore CPU版本-Windows
conda指定路径安装mindspore
m08520a的博客
11-25 879
mindspore安装和使用。
MindSpore保姆级安装教程
热门推荐
m0_51662391的博客
11-24 1万+
MindSpore保姆级安装教程
Anacoda3环境华为MindSpore1.2深度学习框架安装及应用
袁易学的博客(Yuan Yixue's Blog)
05-22 3707
在人工智能与机器学习研究与应用领域，目前较流行的深度学习框架主要有Facebook的PyTorch、Google的TensorFlow/Keras、华为的MindSpore 、百度的PaddlePaddle等，Python/Anaconda语言常与这些深度学习框架联合使用。本文以Windows10 Anaconda3(Python3.7.0)和华为公司2021年3月底发布的MindSpore1.2开源深度学习框架为例，介绍Anaconda、MindSpore及Jupyter Notebook IDE Pyt
docker-alpine-miniconda3:带有Miniconda3（Python 3.7）的最小Docker映像（〜143MB）
05-25
Miniconda Python 3.7 Docker映像 该映像基于Alpine Linux映像，该映像仅为5MB映像，并包含Continuum与Conda软件包管理器一起打包的 。 该图像的下载大小仅为： 注意：Conda存储库仅包含Linux的Glibc链接打包二...
cudnn-windows-x86-64-8.5.0.96-cuda11-archive.zip
11-17
cudnn windows适合cuda11.x cudnn为8.5.0
cudnn-linux-x86-64-8.5.0.96-cuda11-archive.zip
11-17
适合linux cuda11.x为cudnn8.5.0版本
保存conda的虚拟环境到docker hub
Louie_worth的博客
03-11 1520
由于自己最近mentor要求保存conda环境到docker hub，以下记录自己的踩坑过程和部署过程。 Docker 安装和使用 此部分主要参考Docker官方文档：https://docs.docker.com/engine/install/ubuntu/ Ubuntu Docker安装 bug: Get Permission Denied 原因 摘自docker mannual上的一段话 Manage Docker as a non-root user The docker daemon bin.
MindSpore: MindSpore GPU版本安装问题
BaldheadedM的博客
08-03 445
方式2 cuda10.1可以只安装lib库。驱动可以和cuda10.2共用高版本，这样你的服务器可以有多个cuda版本，需要你设置cuda环境变量区分。方式1 可以尝试升级高版本驱动，比如440.33，用于匹配你cuda10.2。如下，是可以正常运行的。根据官网信息，要求CUDA 10.1，是不是CUDA版本不匹配呢？这个问题的原因可能是驱动和版本不匹配导致的。...
【昇思MindSpore】MindSpore的安装
周周の博客小屋
09-25 1502
【昇思MindSpore】MindSpore的安装
Mindspore环境配置
m0_53328738的博客
04-20 3435
Mindspore环境配置 打开mind pore官网： MindSpore官网 选择配置 以如下为例 检查是否已经安装anaconda3 Win+R输入cmd打开终端输入conda -V 这里已经安装了， 进入Anaconda虚拟环境下载mindspore_py39新的虚拟环境 Win+R输入cmd打开终端输入conda activate进如conda安装时默认的虚拟环境，一般是base ...
MindSpore v1.0 安装初体验
skytttttt9394的博客
07-27 300
转载地址：https://bbs.huaweicloud.com/forum/thread-80568-1-1.html 作者：franco52576 MindSpore v1.0 安装初体验 MindSpore 自从开源以来，备受广大开发者的关注。 在全球开发者的测试验证下，经过几个测试版本的不停改进，在前不久的华为全联接大会上，MindSpore正式发布v1.0.0版本！！！ 大家可以快去官网围观一下，MindSpore官网：https://www.mindspore.cn .
Anconda换源+虚拟环境创建及使用（界面操作）
lwbCUMT的博客
06-16 1438
anconda简单使用
在docker中创建conda环境
最新发布
10-11
在Docker中创建Conda环境通常是为了隔离项目依赖，这样每个容器都有自己的独立环境。以下是步骤： 1. **安装Docker**: 首先确保你的系统上已经安装了Docker。如果没有，你可以从Docker官网下载并安装。 2. **获取...
评论
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值