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

ai-write-paper #1454

Open
wants to merge 5 commits into
base: main
Choose a base branch
from
Open

ai-write-paper #1454

wants to merge 5 commits into from

Conversation

wangwenju269
Copy link

@wangwenju269 wangwenju269 commented Aug 19, 2024

Features

  • ​ 在大模型的现有能力框架内,其对于处理长文档级别的智能写作任务存在一定的局限性。探讨多智能体在长文档写作这一应用场景,并在此基础上,提出了一套切实可行的技术解决方案。
  • 长文档智能写作是一个从简至繁的渐进过程,它包含四个明确的阶段:首先是人填机检阶段,其次是机填人检阶段,然后是机填机检阶段,最终达到机填不检的阶段。这四个阶段共同构成了智能写作的发展路径。

Feature Docs

Influence

写作方案

Result

Other

@codecov-commenter
Copy link

⚠️ Please install the 'codecov app svg image' to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 62.59%. Comparing base (ab846f6) to head (68bd5a6).

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files
@@           Coverage Diff           @@
##             main    #1454   +/-   ##
=======================================
  Coverage   62.59%   62.59%           
=======================================
  Files         287      287           
  Lines       17589    17589           
=======================================
  Hits        11009    11009           
  Misses       6580     6580           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

@geekan
Copy link
Owner

geekan commented Oct 20, 2024

Can you provide some running results and examples of what it can do?

@geekan geekan requested a review from better629 October 20, 2024 06:34
@wangwenju269
Copy link
Author

Can you provide some running results and examples of what it can do?

  1. 数据合成:类似Agentwriter(longwriter)的思想,构造长文本数据,解决模型输出短的限制;
  2. 应用领域: 标书、论文、小说、公文一键生成初稿;
  3. 输出示例:产业数字化对中国出口隐含碳的影响

@geekan
Copy link
Owner

geekan commented Oct 21, 2024

我的意思是,注意到 gradio 的实现,是否可以看一下运行过程和使用方法?
此外,有没有一些比较容易理解的例子,让我们可以看到它的能力边界?比如说让它写一个 林黛玉大战那美克星人?

@wangwenju269
Copy link
Author

大佬:

  1. 代码只应用在标书生成场景,且章节间是并行写的,没有引入长期记忆机制;因此在小说写作上,表现不佳;
  2. 写作的好坏受限于 rag检索的精准性 和 大纲任务规划子标题的细粒度;模型只能完成的初稿,不能直接满足用户需求,需要采用 润色插件 来完善初稿; 希望Metagp t团队能后续解决这难题;

@wangwenju269
Copy link
Author

@wangwenju269
Copy link
Author

大佬:后续借鉴了qwen_agent仓库里的方法,做检索;代码稍微修改,修改大纲目录树结构,能递归式保留更细粒度子标题;
个人仓库链接:
智能写作

@geekan
Copy link
Owner

geekan commented Oct 26, 2024

参考 https://github.com/geekan/MetaGPT/blob/main/.pre-commit-config.yaml 这里安装一下 pre-commit,格式化一下代码

# Add the list of Document objects as nodes to the retriever engine
self.store.retriever.add_nodes(doc_chunks)
self.working_memory.clear()
return
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  1. 使用 pre-commit
  2. 增加对应的测试用例
  3. 增加README.md,说明用法、设计思路、遗留问题等

@wangwenju269
Copy link
Author

安装pre-commit 存在问题,暂时未能解决,Readme已添加进仓库

@@ -1,4 +1,4 @@
default_stages: [ commit ]
default_stages: [ pre-commit ]
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do not modify this.


**功能区** : 根据 当前`chapter_id` 章节序号, 输入润色指令进行润色;**联网:**从网络上获取当前当前标题相关网络信息;


Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Don't leave so many blank lines, generally leave only one blank line.




.
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Incorrect input

@@ -0,0 +1,364 @@
import gradio as gr
import json,os ,shutil
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You need to use pre-commit. Or you can use the built-in function of vscode / pycharm to reformat. However, the submitted code needs to comply with most of the PEP8 standards, otherwise it will be difficult for us to accept it.

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I have been modified, please check again
thanks

Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

我查看了文件,根据以上 CR 意见,这些文件还需要进一步的修改

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

好的,我已对README文件和.pre-commit-config.yaml 文件进行修订

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