114 lines
2.5 KiB
Markdown
114 lines
2.5 KiB
Markdown
|
# ParamHub - 参数百科
|
||
|
|
|
||
|
|
> 一站式AI大模型参数与硬件参数速查与对比平台
|
||
|
|
|
||
|
|
## 功能特点
|
||
|
|
|
||
|
|
- **模型数据库**: GPT-4、Llama、Claude等100+大模型参数规格
|
||
|
|
- **GPU数据库**: H100、A100、RTX 4090等GPU详细规格对比
|
||
|
|
- **CPU数据库**: AMD EPYC、Intel Xeon等服务器级CPU参数
|
||
|
|
- **实用工具**: 显存计算器、成本估算器、参数对比工具
|
||
|
|
- **知识库**: 参数解释、选型指南、最佳实践
|
||
|
|
|
||
|
|
## 技术栈
|
||
|
|
|
||
|
|
- **框架**: Next.js 14 (React)
|
||
|
|
- **样式**: Tailwind CSS
|
||
|
|
- **数据库**: PostgreSQL (Prisma ORM)
|
||
|
|
- **图表**: Recharts
|
||
|
|
|
||
|
|
## 快速开始
|
||
|
|
|
||
|
|
### 1. 安装依赖
|
||
|
|
|
||
|
|
```bash
|
||
|
|
npm install
|
||
|
|
```
|
||
|
|
|
||
|
|
### 2. 配置数据库
|
||
|
|
|
||
|
|
```bash
|
||
|
|
# 复制环境变量
|
||
|
|
cp .env.example .env
|
||
|
|
|
||
|
|
# 编辑 .env 文件，设置 DATABASE_URL
|
||
|
|
```
|
||
|
|
|
||
|
|
### 3. 初始化数据库
|
||
|
|
|
||
|
|
```bash
|
||
|
|
npm run db:generate
|
||
|
|
npm run db:push
|
||
|
|
npm run db:seed # 可选，填充初始数据
|
||
|
|
```
|
||
|
|
|
||
|
|
### 4. 启动开发服务器
|
||
|
|
|
||
|
|
```bash
|
||
|
|
npm run dev
|
||
|
|
```
|
||
|
|
|
||
|
|
访问 http://localhost:3000
|
||
|
|
|
||
|
|
## 项目结构
|
||
|
|
|
||
|
|
```
|
||
|
|
param-hub/
|
||
|
|
├── src/
|
||
|
|
│ ├── app/ # Next.js 页面
|
||
|
|
│ │ ├── models/ # 模型数据库页面
|
||
|
|
│ │ ├── gpus/ # GPU数据库页面
|
||
|
|
│ │ ├── cpus/ # CPU数据库页面
|
||
|
|
│ │ ├── tools/ # 计算工具页面
|
||
|
|
│ │ ├── compare/ # 对比工具页面
|
||
|
|
│ │ └── knowledge/ # 知识库页面
|
||
|
|
│ ├── components/ # React组件
|
||
|
|
│ ├── lib/ # 工具函数
|
||
|
|
│ └── data/ # 初始数据
|
||
|
|
├── prisma/
|
||
|
|
│ ├── schema.prisma # 数据库schema
|
||
|
|
│ └── seed.ts # 数据填充脚本
|
||
|
|
├── public/ # 静态资源
|
||
|
|
└── README.md
|
||
|
|
```
|
||
|
|
|
||
|
|
## 数据库表结构
|
||
|
|
|
||
|
|
### models (模型数据)
|
||
|
|
- name, organization, parametersCount, architecture
|
||
|
|
- layers, hiddenSize, attentionHeads, contextLength
|
||
|
|
- benchmarkMmlu, benchmarkHumaneval
|
||
|
|
- isOpenSource, license
|
||
|
|
|
||
|
|
### gpus (GPU数据)
|
||
|
|
- name, manufacturer, architecture
|
||
|
|
- cudaCores, tensorCores, memoryGb
|
||
|
|
- fp32Tflops, fp16Tflops, priceUsd
|
||
|
|
|
||
|
|
### cpus (CPU数据)
|
||
|
|
- name, manufacturer, architecture
|
||
|
|
- cores, threads, baseClockGhz, boostClockGhz
|
||
|
|
- l3CacheMb, tdpWatts, priceUsd
|
||
|
|
|
||
|
|
## 部署
|
||
|
|
|
||
|
|
### Vercel (推荐)
|
||
|
|
|
||
|
|
```bash
|
||
|
|
vercel deploy
|
||
|
|
```
|
||
|
|
|
||
|
|
### Docker
|
||
|
|
|
||
|
|
```bash
|
||
|
|
docker build -t param-hub .
|
||
|
|
docker run -p 3000:3000 param-hub
|
||
|
|
```
|
||
|
|
|
||
|
|
## 贡献数据
|
||
|
|
|
||
|
|
欢迎贡献新的模型/GPU数据！请提交PR或Issue。
|
||
|
|
|
||
|
|
## License
|
||
|
|
|
||
|
|
MIT
|