docker里安装anaconda并手动配置环境

ISO：rocky8.9

Dockerfile配置文件：
 

#安装 Anaconda 到 Docker 容器
FROM ubuntu-desktop:22.04
# 设置代理（假设代理地址为 http://your-proxy-address:port）
ENV http_proxy=http://192.168.101.15:3128
ENV https_proxy=http://192.168.101.15:3128
# 安装必要的软件包
RUN apt-get update && apt-get install -y wget bzip2
# 下载 Anaconda 安装包
RUN wget https://repo.anaconda.com/archive/Anaconda3-2023.07-1-Linux-x86_64.sh
# 安装 Anaconda
RUN bash Anaconda3-2023.07-1-Linux-x86_64.sh -b -p /opt/anaconda
# 设置环境变量
ENV PATH="/opt/anaconda/bin:$PATH"

构建dockerfile

[root@asc2-gn01 ~]# docker build -t pytorch:v2.3.0 .

[root@asc2-gn01 ~]# docker images

[root@asc2-gn01 ~]# docker run -itd -v /gpfs/211840248:/home -v /home1/211840248:/mnt --gpus all pytorch:v2.3.0
e619eb5ce542f2a5f27a916ec50c543c66b72104373fb5bf78149a4e46b1e59b

[root@asc2-gn01 ~]# docker ps -a

[root@asc2-gn01 ~]# docker exec -it e619eb5ce542 /bin/bash

root@e619eb5ce542:/# export http_proxy=http://192.168.101.15:3128 && export https_proxy=http://192.168.101.15:3128

root@e619eb5ce542:~# conda create -n pytorch python=3.8

root@e619eb5ce542:~# conda info --env

root@e619eb5ce542:~# source activate pytorch

(pytorch) root@e619eb5ce542:~# conda install pytorch==2.3.0 torchvision torchaudio cudatoolkit=11.8 -c pytorch && conda install cudnn -c nvidia

测试：