Skip to content

Commit

Permalink
update doc
Browse files Browse the repository at this point in the history
  • Loading branch information
c121914yu committed Jan 14, 2025
1 parent 5579aff commit bc01b65
Show file tree
Hide file tree
Showing 7 changed files with 27 additions and 5 deletions.
Binary file added docSite/assets/imgs/image-83.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added docSite/assets/imgs/image-84.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added docSite/assets/imgs/image-85.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added docSite/assets/imgs/image-86.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added docSite/assets/imgs/image-87.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
8 changes: 7 additions & 1 deletion docSite/content/zh-cn/docs/development/upgrading/4818.md
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
---
title: 'V4.8.18(进行中)'
title: 'V4.8.18'
description: 'FastGPT V4.8.18 更新说明'
icon: 'upgrade'
draft: false
Expand All @@ -9,6 +9,12 @@ weight: 806

## 更新指南

### 1. 更新镜像:

- 更新 fastgpt 镜像 tag: v4.8.18
- 更新 fastgpt-pro 商业版镜像 tag: v4.8.18
- Sandbox 镜像无需更新

### 2. 运行升级脚本

从任意终端,发起 1 个 HTTP 请求。其中 {{rootkey}} 替换成环境变量里的 `rootkey`;{{host}} 替换成**FastGPT 域名**
Expand Down
24 changes: 20 additions & 4 deletions docSite/content/zh-cn/docs/faq/app.md
Original file line number Diff line number Diff line change
Expand Up @@ -7,18 +7,34 @@ toc: true
weight: 908
---

## 工作流中多轮对话场景中如何使连续问题被问题分类节点正确的归类
## 多轮对话中如何使连续问题被问题分类节点正确的归类

问题分类节点具有获取上下文信息的能力,当处理两个关联性较大的问题时,模型的判断准确性往往依赖于这两个问题之间的联系和模型的能力。例如,当用户先问“我该如何使用这个功能?”接着又询问“这个功能有什么限制?”时,模型借助上下文信息,就能够更精准地理解并响应。

但是,当连续问题之间的关联性较小,模型判断的准确度可能会受到限制。在这种情况下,我们可以引入全局变量的概念来记录分类结果。在后续的问题分类阶段,首先检查全局变量是否存有分类结果。如果有,那么直接沿用该结果;若没有,则让模型自行判断。

建议:构建批量运行脚本进行测试,评估问题分类的准确性。

## 系统编排配置中的定时执行,如果用户打开分享的连接,停留在那个页面,定时执行触发问题
## 定时执行的时机问题

发布后,后台生效。
系统编排配置中的定时执行,如果用户打开分享的连接,停留在那个页面,定时执行触发问题:

定时执行会在应用发布后生效,会在后台生效。

## AI对话回答要求中的Markdown语法取消

在针对知识库的回答要求里有, 要给它配置提示词,不然他就是默认的,默认的里面就有该语法。
修改知识库默认提示词, 默认用的是标准模板提示词,会要求按 Markdown 输出,可以去除该要求:

| | |
| --- | --- |
| ![](/imgs/image-83.png) | ![](/imgs/image-84.png) |

## 应用在不同来源效果不一致

Q: 应用在调试和正式发布后,效果不一致;在 API 调用时,效果不一致。

A: 通常是由于上下文不一致导致,可以在对话日志中,找到对应的记录,并查看运行详情来进行比对。

| | | |
| --- | --- | --- |
| ![](/imgs/image-85.png) | ![](/imgs/image-86.png) | ![](/imgs/image-87.png) |

0 comments on commit bc01b65

Please sign in to comment.