TS GPT Download - TS GPT Source code download

TS GPT

AI Source Code

1.0.0

Download

ts-gpt

Online demo address of GPT project built based on SpringCloud Alibaba + SpringBoot3

Function introduction

master为微服务版本(配置要求较高就算单机部署最低也要2c4m)

dec为单体架构版本(配置要求较低步步高点读机也能部署,俩着功能性无差异架构有差异)

任意以及自定义GPT模型对话GPT对话识图GPT对话生图GPT报错信息收集面板自定义对话参数预设ai写作小程序语音对话小程序对话DALL绘图(目前只有文生图,等官方完善还会继续) SD绘图(支持文生图图生图实时渲染出图进度展示局部绘图)预设词展示自定义外观自定义绘图参数(系统级)写的累了建议直接看本文档最下方的图片演示.

Project structure

vue web client

vue3
element plus

microservices backend service set

ts-auth处理登录用户信息商品交易支付
ts-gateway服务网关鉴权路由登录交易白名单交易路由负载均衡
ts-common公共工具类公共配置公共异常拦截常量枚举
ts-api存放公用DTO 远程调用接口feigb
ts-drawing处理绘图功能类服务
ts-chat处理对话类功能服务
SaToken
Redis
Maven
Spring Boot 3
Seata
Nacos
Spring Cloud Gateway
Feign
MySQL

wechat applet

uniapp
vue3

Deployment environment and preparation tools

Docker MySQL8 Redis7 Nacos2.x JDK17阿里OSS服务~~Seata~~如单机部署服务器最低要求2c4m SD绘图(非必须)百度翻译API(非必须)微信小程序APPID(目前非必须)

Front-end execution command

    # # 安装 yarn 命令
    npm install --global yarn

    # # 安装 项目所需依赖
    yarn install

    # # 运行本地环境
    yarn run serve

    # # 打包构建
    yarn run build

Backend required operations

.sql is the .sql directory required for the project, which is divided into data and table. Data is the table-building sql containing initial data, while table only has the table-building sql without any data. You can choose it by yourself.

.nacos is the configuration of this project. Please define your own configuration and import it into nacos.

Things to note

Every time you add a model to the project or perform an operation on the configuration class, you must click once to load the configuration. Otherwise, it cannot be applied unless you restart the project, because the configuration is loaded into the server memory (not redis) at startup and needs to be manually applied after modification.

The server needs at least 2 cores and 4G memory because nacos + mysql + redis takes up about 1.5 G. This does not include the capacity of 4 services. Of course, if you have enough servers, you can open multiple instances yourself for multi-instance distributed deployment, or you can do a single machine deploy , (2024-01-16) Update the dec branch single architecture version

This project does not support one SD being used by multiple services. For example, if you have an SD server, and then you share an SD with your friend's third-party project, serious problems will arise because this project is connected to real-time rendering and the SD only Supporting single QS will cause errors in real-time rendering effects. Of course, if you need to support Vme50?

For some functional failures, please log in to the system using a management account. You can go to the exception panel to view the specific error information (including specific priority).