RWKV-Runner/README_ZH.md

133 lines
4.2 KiB
Markdown
Raw Normal View History

2023-05-21 17:28:51 +08:00
<p align="center">
<img src="https://github.com/josStorer/RWKV-Runner/assets/13366013/d24834b0-265d-45f5-93c0-fac1e19562af">
</p>
<h1 align="center">RWKV Runner</h1>
<div align="center">
2023-05-22 11:48:56 +08:00
本项目旨在消除大语言模型的使用门槛全自动为你处理一切你只需要一个仅仅几MB的可执行程序。此外本项目提供了与OpenAI
API兼容的接口这意味着一切ChatGPT客户端都是RWKV客户端。
2023-05-21 17:28:51 +08:00
[![license][license-image]][license-url]
[![release][release-image]][release-url]
[English](README.md) | 简体中文
2023-06-06 23:50:52 +08:00
### 安装
[![Windows][Windows-image]][Windows-url]
[![MacOS][MacOS-image]][MacOS-url]
[![Linux][Linux-image]][Linux-url]
2023-06-03 20:28:21 +08:00
[视频演示](https://www.bilibili.com/video/BV1hM4y1v76R) | [疑难解答](https://www.bilibili.com/read/cv23921171) | [预览](#Preview) | [下载][download-url] | [懒人包](https://pan.baidu.com/s/1wchIUHgne3gncIiLIeKBEQ?pwd=1111)
2023-05-21 17:28:51 +08:00
[license-image]: http://img.shields.io/badge/license-MIT-blue.svg
[license-url]: https://github.com/josStorer/RWKV-Runner/blob/master/LICENSE
[release-image]: https://img.shields.io/github/release/josStorer/RWKV-Runner.svg
[release-url]: https://github.com/josStorer/RWKV-Runner/releases/latest
2023-05-25 10:02:29 +08:00
[download-url]: https://github.com/josStorer/RWKV-Runner/releases
2023-05-21 17:28:51 +08:00
2023-06-06 23:50:52 +08:00
[Windows-image]: https://img.shields.io/badge/-Windows-blue?logo=windows
[Windows-url]: https://github.com/josStorer/RWKV-Runner/blob/master/build/windows/Readme_Install.txt
[MacOS-image]: https://img.shields.io/badge/-MacOS-black?logo=apple
[MacOS-url]: https://github.com/josStorer/RWKV-Runner/blob/master/build/darwin/Readme_Install.txt
[Linux-image]: https://img.shields.io/badge/-Linux-black?logo=linux
[Linux-url]: https://github.com/josStorer/RWKV-Runner/blob/master/build/linux/Readme_Install.txt
2023-05-21 17:28:51 +08:00
</div>
2023-05-22 19:43:47 +08:00
#### 注意 目前RWKV中文模型质量一般推荐使用英文模型体验实际RWKV能力
2023-06-07 00:11:39 +08:00
#### 对于Mac和Linux用户请手动安装 Python3.10 (通常最新的系统已经内置了)
2023-06-06 23:57:28 +08:00
#### 预设配置已经开启自定义CUDA算子加速速度更快且显存消耗更少。如果你遇到可能的兼容性问题前往配置页面关闭`使用自定义CUDA算子加速`
2023-05-23 20:52:18 +08:00
2023-05-23 21:25:33 +08:00
#### 对于不同的任务调整API参数会获得更好的效果例如对于翻译任务你可以尝试设置Temperature为1Top_P为0.3
2023-05-21 17:28:51 +08:00
## 功能
- RWKV模型管理一键启动
2023-05-22 11:48:56 +08:00
- 与OpenAI API完全兼容一切ChatGPT客户端都是RWKV客户端。启动模型后打开 http://127.0.0.1:8000/docs 查看详细内容
2023-05-21 17:28:51 +08:00
- 全自动依赖安装,你只需要一个轻巧的可执行程序
2023-05-30 14:40:33 +08:00
- 预设了2G至32G显存的配置几乎在各种电脑上工作良好
2023-05-30 11:52:38 +08:00
- 自带用户友好的聊天和补全交互页面
2023-05-21 17:28:51 +08:00
- 易于理解和操作的参数配置
- 内置模型转换工具
- 内置下载管理和远程模型检视
- 多语言本地化
- 主题切换
- 自动更新
2023-05-27 00:38:39 +08:00
## API并发压力测试
```bash
ab -p body.json -T application/json -c 20 -n 100 -l http://127.0.0.1:8000/chat/completions
```
body.json:
```json
{
"messages": [
{
"role": "user",
"content": "Hello"
}
]
}
```
2023-05-21 17:28:51 +08:00
## Todo
2023-05-23 14:58:45 +08:00
- [ ] 模型训练功能
- [x] CUDA算子int8提速
2023-06-06 23:50:52 +08:00
- [x] macOS支持
- [x] linux支持
- [ ] 本地状态缓存数据库
2023-05-21 17:28:51 +08:00
## 相关仓库:
- RWKV-4-Raven: https://huggingface.co/BlinkDL/rwkv-4-raven/tree/main
- ChatRWKV: https://github.com/BlinkDL/ChatRWKV
- RWKV-LM: https://github.com/BlinkDL/RWKV-LM
## Preview
### 主页
![image](https://github.com/josStorer/RWKV-Runner/assets/13366013/9d25380a-a17b-443f-b823-86c754ebebf0)
### 聊天
![image](https://github.com/josStorer/RWKV-Runner/assets/13366013/0e66d5fa-f34a-409f-9cd4-d880815733f3)
2023-05-24 21:45:50 +08:00
### 补全
![image](https://github.com/josStorer/RWKV-Runner/assets/13366013/d4178ee9-a188-4878-9777-25c916872c29)
2023-05-21 17:28:51 +08:00
### 配置
![image](https://github.com/josStorer/RWKV-Runner/assets/13366013/ad9921fc-7248-40a3-9e18-03445b86e4bf)
### 模型管理
![image](https://github.com/josStorer/RWKV-Runner/assets/13366013/7c36f15f-3e77-49cd-a16d-99a29f870bdf)
### 下载管理
![image](https://github.com/josStorer/RWKV-Runner/assets/13366013/32fde30b-11dd-43b9-9667-ad6975be2106)
### 设置
![image](https://github.com/josStorer/RWKV-Runner/assets/13366013/e8a0f746-9da7-48e3-b3fc-e1453ac50de2)