01.04.01.DeepSeek：环境搭建篇（轻量化本地部署工具Ollama：安装配置）

操作流程1、下载安装2、配置3、测试下载安装准备期susetenforce0mkdir-p/usr/local/software/ollamamkdir-p/usr/local/software/ollama/modelschmod0777-R/usr/local/software/ollama /usr/local/software/ollama/modelschown$USER:$USER-R/usr/local/software/ollama /usr/local/software/ollama/modelsOllamacd/usr/local/software# 官网https://ollama.com/download/linux# ollama在线安装: curl -fsSL https://ollama.com/install.sh | sh# ollama离线安装mkdir-pollamachmod0777-Rollamachown$USER:$USER-Rollamawgethttps://github.com/ollama/ollama/releases/download/v0.6.5/ollama-linux-amd64.tgztar-xzvfollama-linux-amd64.tgz-C/usr/local/software/ollamarm-rfollama-linux-amd64.tgz配置# 创建ollama开机自启cat/etc/systemd/system/ollama.serviceEOF [Unit] DescriptionOllama Service Afternetwork-online.target [Service] Typeollama Userroot Grouproot Restartalways RestartSec3 ExecStart/usr/local/software/ollama/bin/ollama serve EnvironmentOLLAMA_HOST0.0.0.0 EnvironmentOLLAMA_PORT11434 EnvironmentOLLAMA_GPU_LAYERcuda EnvironmentOLLAMA_ORIGINS* EnvironmentOLLAMA_DEBUG1 EnvironmentOLLAMA_MODELS/usr/local/software/ollama/models EnvironmentCUDA_VISIBLE_DEVICESGPU-e1d26ec1-7ba8-8d54-49a8-d1f0b9d865c2 [Install] WantedBymulti-user.target EOFgroupaddollamauseradd-gollama-s/usr/sbin/nologin ollamachmod0777-R/usr/local/software/ollama /etc/systemd/system/ollama.servicechown$USER:$USER-R/usr/local/software/ollama /etc/systemd/system/ollama.servicechownollama:ollama-R/usr/local/software/ollama /etc/systemd/system/ollama.service#查看GPU的UUID#命令nvidia-smi -Lcat/etc/profileEOF #set ollama environment OLLAMA_HOME/usr/local/software/ollama OLLAMA_HOST0.0.0.0 OLLAMA_PORT11434 OLLAMA_GPU_LAYERcuda OLLAMA_MODELS/usr/local/software/ollama/models OLLAMA_ORIGINS* OLLAMA_DEBUG1 PATH$OLLAMA_HOME:$OLLAMA_HOME/bin:$OLLAMA_GPU_LAYER:$OLLAMA_MODELS:$PATH:$OLLAMA_HOST:$OLLAMA_PORT:$OLLAMA_ORIGINS:$OLLAMA_DEBUG export OLLAMA_HOME OLLAMA_GPU_LAYER OLLAMA_MODELS PATH OLLAMA_HOST OLLAMA_PORT OLLAMA_ORIGINS OLLAMA_DEBUG EOFsource/etc/profile# 启动另外一个终端systemctl daemon-reload systemctlenableollama systemctl stop ollama systemctl start ollama systemctl status ollamareboot测试ollama-v# 提问 查看PROCESSOR列比例值是否有使用GPUollamapsapihttps://docs.ollama.com/api