Deepseek本地部署指南

Deepseek服务器部署指南

• Docker

• CUDA

• Ollama

• Deepseek

• Dify

• 注意: 不是面向零linux基础人群

Docker

如果有Docker就跳过

1.下载安装包

Index of linux/static/stable/x86_64/

选择26.1.1为例下载

上传至服务器后 cd 到文件位置执行解压

注: 仔细检查文件名是否拼写正确

tar -zxvf docker-26.1.3-ce.tgz

解压好之后移动文件位置到/usr/bin

cp ./docker/* /usr/bin

2.配置docker

首先创建docker.service文件

cd /etc/systemd/system/

touch docker.service

编辑文件内容

vim docker.service

内容如下:

注意:insecure-registry此处为服务器的ip

[Unit]
Description=Docker Application Container Engine
Documentation=https://docs.docker.com
After=network-online.target firewalld.service
Wants=network-online.target
 
[Service]
Type=notify
# the default is not to use systemd for cgroups because the delegate issues still
# exists and systemd currently does not support the cgroup feature set required
# for containers run by docker
ExecStart=/usr/bin/dockerd --selinux-enabled=false --insecure-registry=192.168.205.230
ExecReload=/bin/kill -s HUP $MAINPID
# Having non-zero Limit*s causes performance problems due to accounting overhead
# in the kernel. We recommend using cgroups to do container-local accounting.
LimitNOFILE=infinity
LimitNPROC=infinity
LimitCORE=infinity
# Uncomment TasksMax if your systemd version supports it.
# Only systemd 226 and above support this version.
#TasksMax=infinity
TimeoutStartSec=0
# set delegate yes so that systemd does not reset the cgroups of docker containers
Delegate=yes
# kill only the docker process, not all processes in the cgroup
KillMode=process
# restart the docker process if it exits prematurely
Restart=on-failure
StartLimitBurst=3
StartLimitInterval=60s
 
[Install]
WantedBy=multi-user.target

添加权限 , 注意要cd到目录下

chmod +x docker.service

重新加载

systemctl daemon-reload

启动

systemctl start docker

查看运行状态

systemctl status docker

给Docker设置上开机自启, 注意要cd到目录下

systemctl enable docker.service

CUDA

先安装一下conda

wget -c https://repo.anaconda.com/archive/Anaconda3-2023.03-1-Linux-x86_64.sh

一直输入yes按回车

bash Anaconda3-2023.03-1-Linux-x86_64.sh

设置环境变量 在profile最后加一行 export PATH=~/anaconda3/bin:$PATH

vim /etc/profile

重载

source /etc/profile
source ~/.bashrc

查看CUDA要下载的版本 , 小方块右上角有一个 CUDA version就是

nvidia-smi

使用conda安装cuda, version 这边写版本

conda install -c conda-forge cupy cudnn cutensor nccl cuda-version=XX.X

Ollama

通过一键下载文件直接安装

curl -fsSL https://ollama.com/install.sh | bash

启动ollama

ollama serve	

查看已安装模型

ollama list

安装Deepseek-r1模型, 安装完成后再次输入可进行对话, 按Ctrl+d可退出

ollama run deepseek-r1:32b

Dify

首先克隆dify的源码

git clone --depth 1 https://github.com/langgenius/dify.git

然后切换到dify/docker下边

cd dify/docker

更改配置文件

cp .env.example .env

用docker拉取

docker compose up -d   如报错找不到命令则用 docker-compose up -d

dify代理的是80端口

打开后设置管理员账号, 在右上角有用户, 点设置增加大模型的供应商 , 然后选择ollama,

下边的基础URL写服务器ip, 如果报错就写http://host.docker.internal:11434, 在自己电脑部署的访问的时候是写后者