第四十篇-TeslaP40+Ollama+Ollama-WebUI(自编译)

木卫二号Coding

于 2024-08-28 23:37:29 发布

阅读量63

点赞数 4

分类专栏： AI-LLM-实战文章标签： chatgpt 人工智能 llama python 开源

本文链接：https://blog.csdn.net/hai4321/article/details/141652456

版权

AI-LLM-实战专栏收录该内容

40 篇文章 22 订阅

订阅专栏

本文介绍用自己编辑ollama-webui,链接本地ollama

环境

系统：CentOS-7
CPU: 14C28T
内存：32G
显卡：Tesla P40 24G
驱动: 535
CUDA: 12.2
Ollama: 0.3.0

本地ollama

参考
[第二十四篇-Ollama-在线安装](https://blog.csdn.net/hai4321/article/details/138241623)

验证

http://192.168.31.222:11434/
Ollama is running

安装nodejs

https://registry.npmmirror.com/binary.html?path=node/v18.20.2/
	node-v18.20.2-linux-x64.tar.gz

	tar -zxf node-v18.20.2-linux-x64.tar.gz

	vim /etc/profile
 
	export NODE_HOME=/opt/soft/node-v18.20.2-linux-x64/bin
	export PATH=$PATH:$NODE_HOME:/usr/local/bin/
 
	使变量文件生效
	source /etc/profile

	验证
		node -v
		npm -v

	配置镜像源
		# 设置为淘宝源地址
		npm config set registry https://registry.npmmirror.com
		# 查看是否设置成功
		npm config get registry

编译

下载代码：
git clone https://github.com/ollama-webui/ollama-webui-lite ollama-webui

安装依赖的工具包
npm install --registry=http://mirrors.cloud.tencent.com/npm/


启动Web可视化
npm run dev

ollama-webui-lite@0.0.1 dev
vite dev --host --port 3000
VITE v4.5.2 ready in 1068 ms
➜ Local: http://localhost:3000/
➜ Network: http://192.168.31.222:3000/
➜ Network: http://172.17.0.1:3000/

访问

http://192.168.31.222:3000
配置ollama的本地api

在这里插入图片描述
保存，选择你ollama中有的模型

[root@ai-server soft]# ollama list
NAME                                    ID              SIZE    MODIFIED     
tinyllama:latest                        2644915ede35    637 MB  4 days ago  
llama-3-8b-u:latest                     a4f9abe93500    4.9 GB  11 days ago 
yi-1.5-34b:latest                       28fcd739309f    20 GB   2 weeks ago 
internlm2_5-20b:latest                  eddbf49c317b    11 GB   3 weeks ago 
mistral-nemo:12b-instruct-2407-q4_K_M   ad2adccd15ad    7.5 GB  4 weeks ago 
llama3.1:8b-instruct-q4_K_M             f8f83c5f9af1    4.9 GB  4 weeks ago 
yi:9b-chat-v1.5-q4_K_M                  70f792429c07    5.3 GB  4 weeks ago 
llama3.1:latest                         a340353013fd    4.7 GB  4 weeks ago 
qwen2:0.5b                              6f48b936a09f    352 MB  4 weeks ago 
qwen2:1.5b                              f6daf2b25194    934 MB  5 weeks ago 
athene-70b-q2:latest                    03ce01bc2c99    26 GB   5 weeks ago 
mofanke/acge_text_embedding:latest      dd257071cdac    651 MB  5 weeks ago 
quentinz/bge-base-zh-v1.5:latest        cd232613fa6f    204 MB  5 weeks ago 
quentinz/bge-large-zh-v1.5:latest       bc8ca0995fcd    651 MB  5 weeks ago 
glm4:latest                             5b699761eca5    5.5 GB  6 weeks ago 
codegeex4:latest                        867b8e81d038    5.5 GB  6 weeks ago 
internlm2_5-7b:latest                   357634f5a926    4.7 GB  7 weeks ago 
gemma-2-27b:latest                      4405ce42718c    16 GB   8 weeks ago 
deepseek-coder-v2-lite-8:latest         20c0b5b7c063    16 GB   2 months ago
qwen2-57b:latest                        47449de09cb8    32 GB   2 months ago
codeqwen1.5-7b:latest                   39a6a464a2d2    4.7 GB  3 months ago
mixtral-7_8b:latest                     40a7cfe6dbf1    26 GB   4 months ago
llama-3-8b-instruct:latest              2143c3e3f672    4.9 GB  4 months ago
qwen1.5-32b:latest                      96743882fd6d    19 GB   4 months ago
qwen1.5-14b:latest                      3cc8bb79440c    9.2 GB  4 months ago
qwen:0.5b-chat-v1.5-q4_K_M              e1c9c6192a7e    407 MB  4 months ago
qwen1.5-moe:latest                      b3247e5069ed    10 GB   4 months ago