Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix: 当用gpt-4的API时增加可用的Max Tokens #729

Merged
merged 4 commits into from
Mar 22, 2023

Conversation

simonwu53
Copy link
Contributor

已经有人收到gpt-4的邀请了(#706),我最近也收到了。当OPENAI_API_MODEL填gpt-4相关的模型时,可以自动更改api的选项。虽然不一定每次都用得到,但临时有需求的话可以总结长文章、文献、进行长对话了。

@goodwisdom
Copy link

目前OPENAI_API_MODEL的参数已经支持填写gpt-4并能生效吗?

@simonwu53
Copy link
Contributor Author

simonwu53 commented Mar 21, 2023

应该是生效的。不能直接问它你是不是GPT-4,因为训练数据和原来截止时间一样,所以它不知道有GPT-4这东西。附图问的问题:

F4F16F2F-6637-4EBD-A448-EC2364038DD8

从代码角度的话,后端用的上游gpt包支持用API Key的前提下自定义模型,理论上绝大多数api的参数也能改(比如自定义system message),目前前端还没支持。

@goodwisdom
Copy link

赞一个,MAX Tokens很有必要

@Chanzhaoyu Chanzhaoyu merged commit 3e509f6 into Chanzhaoyu:main Mar 22, 2023
@simonwu53 simonwu53 deleted the increase_max_token branch March 22, 2023 12:20
Equim-chan pushed a commit to Equim-chan/chatgpt-web that referenced this pull request Mar 22, 2023
* fix: 用GPT-4 API时增加max token

* style: eslint fix

---------

Co-authored-by: Redon <790348264@qq.com>
pzcn pushed a commit to pzcn/chatgpt-web2 that referenced this pull request Mar 22, 2023
* fix: 用GPT-4 API时增加max token

* style: eslint fix

---------

Co-authored-by: Redon <790348264@qq.com>
lizhengxiang87 added a commit to lizhengxiang87/chatgpt-web that referenced this pull request Mar 25, 2023
* chore: change docs

* feat: 添加 Prompt 模板和 Prompt 商店支持 (Chanzhaoyu#268)

* feat: 添加Prompt模板和Prompt商店支持

* feat: well done

---------

Co-authored-by: Redon <790348264@qq.com>

* feat: 设置添加关闭按钮

* chore: format

* chore: 2.10.4

* chore: cover

* fix: 更新JSON导入检查 (Chanzhaoyu#523)

* feat: Improve zh-TW locale (Chanzhaoyu#511)

* fix: cors in auth_secret_key model (Chanzhaoyu#499)

* fix: 修改最大行数与官方保持一致 (Chanzhaoyu#502)

* chore: rename environment variables files

* docs: update README.md about .env file

* feat: support long reply

* chore: upgrade chatgpt package and set long reply to false default

* chore: set long reply to false default

* fix: change maxRows to 8

* feat: mobile max row

---------

Co-authored-by: ChenZhaoYu <790348264@qq.com>

* feat: support linux command line proxy standard (Chanzhaoyu#308)

* feat: support linux command line proxy standard with env HTTPS_PROXY and ALL_PROXY

* chore: update docs for linux command linux proxy standard

* fix: Prompt商店在线导入可以导入两种recommend.json里提到的模板 (Chanzhaoyu#516) (Chanzhaoyu#521)

* fix(PromptStore): import prompt template using different dict keys

* chore(PromptStore): clear url field after downloading

* feat: 仓库添加翻译

* feat: 仓库查询筛选

* feat: 调整界面

* chore: update deps

* feat: improve zh-TW locale, translate new strings (Chanzhaoyu#542)

* feat: 语言设定改为下拉框选择

* feat: 修改 Prompt Store 移动端样式

* chore: docs change

* feat: 权限输入框改为密码类型

* fix: 清空上下文时不应该重置会话标题

* chore: docs change

* chore: optimize dockerfile to speed up building and reduce docker image size (Chanzhaoyu#520)

* chore: v2.10.5

* fix: 修复导入、导出错误

* fix: inverted parameter for before each function (Chanzhaoyu#585)

* fix: 🐛 fix scroll bar style (Chanzhaoyu#617)

* fix: typo (Chanzhaoyu#636)

* perf: 优化移动端 prompt 商店样式 (Chanzhaoyu#608)

* feat: 修改移动端 prompt 商店样式

* add: 修改移动端 prompt 商店样式

* perf: 清理代码

---------

Co-authored-by: ChenZhaoYu <790348264@qq.com>

* chore: extract proxy setup code into reusable function (Chanzhaoyu#583)

* perf: 记录上下文设定保存到本地

* perf: 优化部份判断

* docs: 移除 docker-compose.yml 省略

* feat: show api balance (Chanzhaoyu#582)

* feat: show api balance

* Update index.ts

* 保留小数点后五位

* perf: 判断优化

---------

Co-authored-by: Redon <790348264@qq.com>

* chore: update deps and lint fix

* feat: session 返回接口方式做显示判断

* fix: highlight.js 控制台兼容性警告

* chore: v2.10.6

* fix: 回退 chatgpt 版本导致的代理失效

* fix: 添加 usingContext 缺省

* chore: v2.10.7

* feat: 查询余额使用 API_BASE_URL 参数

* feat: 渲染的链接新窗口打开

* perf: config 接口添加权限验证

* fix: 手动重置异常的 loading

* feat: 添加用于显示回复消息原文的选项 (Chanzhaoyu#672)

* feat: 添加显示用于原文的选项

* chore: 修复暗色主题下文本颜色问题

给输入和输出气泡添加了 css 类,用来处理在暗色主题下聊天气泡的文本颜色

* feat: 用户输入不应该被渲染,防止 xss

---------

Co-authored-by: ChenZhaoYu <790348264@qq.com>

* feat: 新增限流功能 (Chanzhaoyu#718)

* 请求速率限制

* perf: 优化代码

---------

Co-authored-by: ChenZhaoYu <790348264@qq.com>

* feat: 添加角色设定预留API 设定页(Chanzhaoyu#768)

* add systemMessage

* perf: 优化代码和类型

* perf: 补全翻译和为以后做准备

---------

Co-authored-by: ChenZhaoYu <790348264@qq.com>

* perf: 重置回答时滚动定位至该回答 (Chanzhaoyu#781)

* chore: 重置回答时滚动定位至该回答

* perf: format code

---------

Co-authored-by: ChenZhaoYu <790348264@qq.com>

* perf: scrollToBottom 改成 scrollToBottomIfAtBottom (Chanzhaoyu#771)

* feat: 更新依赖并兼容已存在的代理

* feat: 当用gpt-4的API 时增加可用的 Max Tokens (Chanzhaoyu#729)

* fix: 用GPT-4 API时增加max token

* style: eslint fix

---------

Co-authored-by: Redon <790348264@qq.com>

* feat: Create start.cmd (Chanzhaoyu#656)

在windows下也可以运行

* chore(vsc-debug): 增加 visual-studio-code 的 debug 启动文件 (Chanzhaoyu#296)

* fix line end with new-line in some backend (Chanzhaoyu#474)

* fix line end with new-line in some backend

In backend with golang (like https://github.com/sashabaranov/go-openai), the responseText always has '\n'.

* feat: 补全遗漏

---------

Co-authored-by: ChenZhaoYu <790348264@qq.com>

* feat: 增加 ChatGPTUnofficialProxyAPI 的模型自定义配置 (Chanzhaoyu#632)

* feat: 增加 ChatGPTUnofficialProxyAPI 的模型自定义配置(Chanzhaoyu#618)

* perf: 修改判断

---------

Co-authored-by: ChenZhaoYu <790348264@qq.com>

* chore: 显示原文时使用 .whitespace-pre-wrap (Chanzhaoyu#779)

显示 ChatGPT 输出的原文时保留空格和换行

* feat:  PWA 默认不开启

* feat:  切换 history 时自动聚焦输入框 (Chanzhaoyu#735)

* feat: 修改 scrollToBottom 行为(当滚动条在底部或发送问题时才会进入自动滚动,滚轮向上滚动会中止自动滚动,方便复制和查看之前的代码; 切换 history 时自动聚焦输入框,方便直接输入文字,不用再次去点击输入框;)

* chore: 通过eslint检查

* perf: 清理重复功能

---------

Co-authored-by: Redon <790348264@qq.com>

* perf: 网络连接时 500 自动跳转到主页

* docs: fix docker port to private (Chanzhaoyu#802)

* feat: add zh-TW prompts in the Promt Store (Chanzhaoyu#796)

* fix: prompt store 样式问题

* chore: v2.10.8

* fix: 重置时应该移除本地储存

* fix: PWA 未添加的问题 (Chanzhaoyu#807)

Co-authored-by: bingo235 <>

* fix: 移动端新建会话关闭侧边栏 (Chanzhaoyu#813)

* perf: 移动端删除 chat 时,侧边栏应该收起

* fix: ESM error (Chanzhaoyu#826)

* perf: 自动焦点时移动端上的不便

* fix: 移动端焦点不触发的问题

* perf: 本地地址

* fix: 反向代理限流失效 (Chanzhaoyu#863)

https://docs.colyseus.io/zh_hk/colyseus/how-to/rate-limit/

使用nginx限流会只识别为服务器ip,需启用trust proxy

---------

Co-authored-by: ChenZhaoYu <790348264@qq.com>
Co-authored-by: Nothing1024 <78358913+Nothing1024@users.noreply.github.com>
Co-authored-by: Peter Dave Hello <hsu@peterdavehello.org>
Co-authored-by: Yige <a@wyr.me>
Co-authored-by: Eason <tobewhatwewant@gmail.com>
Co-authored-by: 吴杉(Shan Wu) <Shan.Wu.1995@gmail.com>
Co-authored-by: Wisdom <pdsu.wwz@foxmail.com>
Co-authored-by: Hayden Chen <hayden.chen.b612@gmail.com>
Co-authored-by: 碎花爱洋裙 <44236581+lizhongyuan3@users.noreply.github.com>
Co-authored-by: CornerSkyless <573196853@qq.com>
Co-authored-by: Kirk Lin <linkirk@163.com>
Co-authored-by: pzcn <sdustpedro@gmail.com>
Co-authored-by: Yi <32430186+yilozt@users.noreply.github.com>
Co-authored-by: xiaozhu <501801307@qq.com>
Co-authored-by: quzard <1191890118@qq.com>
Co-authored-by: 舜岳 <1277952981@qq.com>
Co-authored-by: 吴杉(Shan Wu) <Shan.Wu@ut.ee>
Co-authored-by: vulgate <vulgate@qq.com>
Co-authored-by: KeepRunning <ChandlerVer5@gmail.com>
Co-authored-by: Liut <Eagle.L@gmail.com>
Co-authored-by: hncboy <619452863@qq.com>
Co-authored-by: Allen <987425154@qq.com>
Co-authored-by: 东子 <16423601+kilvn@users.noreply.github.com>
Co-authored-by: Bing <bingo235@users.noreply.github.com>
Co-authored-by: Kid <44045911+kidonng@users.noreply.github.com>
Co-authored-by: gitgitgogogo <gitgitgogogo@users.noreply.github.com>
jingChen55 pushed a commit to jingChen55/chatgpt-web that referenced this pull request Mar 27, 2023
* fix: 用GPT-4 API时增加max token

* style: eslint fix

---------

Co-authored-by: Redon <790348264@qq.com>
jingChen55 pushed a commit to jingChen55/chatgpt-web that referenced this pull request Mar 27, 2023
* fix: 用GPT-4 API时增加max token

* style: eslint fix

---------

Co-authored-by: Redon <790348264@qq.com>
suikodev referenced this pull request in AstraSurge/gpteams Mar 27, 2023
* fix: 用GPT-4 API时增加max token

* style: eslint fix

---------

Co-authored-by: Redon <790348264@qq.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants