这是用户在 2024-9-21 13:58 为 https://github.com/NisaarAgharia/Advanced_RAG 保存的双语快照页面,由 沉浸式翻译 提供双语支持。了解如何保存?
Skip to content
Owner avatar Advanced_RAG Public

Advanced Retrieval-Augmented Generation (RAG) through practical notebooks, using the power of the Langchain, OpenAI GPTs ,META LLAMA3 ,Agents.
高级检索增强生成(RAG)通过实用笔记本,利用 Langchain、OpenAI GPTs、META LLAMA3、Agents 的力量。

Open in github.dev Open in a new github.dev tab Open in codespace

NisaarAgharia/Advanced_RAG

t

Add file

Add file

Folders and files

NameName
Last commit message
Last commit date

Latest commit

38956ae · Apr 26, 2024

History

43 Commits
Apr 9, 2024
Apr 10, 2024
Apr 11, 2024
Apr 12, 2024
Apr 14, 2024
Apr 15, 2024
Apr 16, 2024
Apr 17, 2024
Apr 22, 2024
Apr 24, 2024
Apr 26, 2024

Repository files navigation

Dive into the world of advanced language understanding with Advanced_RAG. These Python notebooks offer a guided tour of Retrieval-Augmented Generation (RAG) using the Langchain framework, perfect for enhancing Large Language Models (LLMs) with rich, contextual knowledge.
深入探索先进语言理解的世界,与 Advanced_RAG 一起。这些 Python 笔记本为您提供了使用 Langchain 框架的检索增强生成(RAG)的指南,非常适合增强大型语言模型(LLMs)的丰富,上下文知识。

Architecture Flows 架构流程

Basic RAG : 基本的 RAG:

Understand the journey of a query through RAG, from user input to the final generated response, all depicted in a clear, visual flow.
理解查询在 RAG 中的流程,从用户输入到最终生成的响应,所有这些都以清晰的视觉流程展示。
RAG_User_Flow

Advanced RAG Techniques :
高级 RAG 技术:

Explore the intricate components that make up an advanced RAG system, from query construction to generation.
探索构成先进 RAG 系统的复杂组件,从查询构造到生成。
Advanced RAG Components

02. Multi Query Retriever :
02. 多查询检索器:

Get to grips with the Multi Query Retriever structure, which enhances the retrieval process by selecting the best responses from multiple sources.
熟悉多查询检索器结构,该结构通过从多个源选择最佳响应来增强检索过程。
MQR

06. Self-Reflection-RAG :
06. 自我反思-RAG :

self-Rag

07. Agentic RAG : 07. 代理性 RAG:

download

08. Adaptive Agentic RAG :
08. 自适应代理 RAG:

adaptive_rag_agent

09. Corrective Agentic RAG :
09. 纠正性代理人 RAG:

correctiveRAG

10. LLAMA 3 Agentic RAG Local:
10. LLAMA 3 代理式 RAG 本地化:

LLAMA3_AGent

Notebooks Overview 笔记本概述

Below is a detailed overview of each notebook present in this repository:
以下是此存储库中每个笔记本的详细概述:

  • 01_Introduction_To_RAG.ipynb
    01_介绍_To_RAG.ipynb
    • Basic process of building RAG app(s)
      构建 RAG 应用程序的基本过程
  • 02_Query_Transformations.ipynb
    02_查询转换.ipynb
    • Techniques for Modifying Questions for Retrieval
      修改检索问题的技术
  • 03_Routing_To_Datasources.ipynb
    03_路由到数据源.ipynb
    • Create Routing Mechanism for LLM to select the correct data Source
      为 LLM 创建路由机制以选择正确的数据源
  • 04_Indexing_To_VectorDBs.ipynb
    04_索引到_VectorDBs.ipynb
    • Various Indexing Methods in the Vector DB
      向量数据库中的各种索引方法
  • 05_Retrieval_Mechanisms.ipynb
    05_检索机制.ipynb
    • Reranking, RaG Fusion, and other Techniques
      重新排名,RaG 融合,和其他技术
  • 06_Self_Reflection_Rag.ipynb
    06_自我反思_破布.ipynb
    • RAG that has self-reflection / self-grading on retrieved documents and generations.
      具有对检索文档和生成的自我反思/自我评级的 RAG。
  • 07_Agentic_Rag.ipynb
    • RAG that has agentic Flow on retrieved documents and generations.
      具有代理流程的 RAG 在检索的文档和生成上。
  • 08_Adaptive_Agentic_Rag.ipynb
    08_自适应_代理式_破旧.ipynb
    • RAG that has adaptive agentic Flow.
      具有自适应代理流的 RAG。
  • 09_Corrective_Agentic_Rag.ipynb
    09_修正_代理_破布.ipynb
    • RAG that has corrective agentic Flow on retrieved documents and generations.
      具有纠正性代理流程的 RAG 在检索的文档和生成中。
  • 10_LLAMA_3_Rag_Agent_Local.ipynb
    • LLAMA 3 8B Agent Rag that works Locally.
      翻译文本:LLAMA 3 8B 局部工作的代理抹布。

Enhance your LLMs with the powerful combination of RAG and Langchain for more informed and accurate natural language generation.
提升您的LLMs,通过强大的 RAG 和 Langchain 组合,以获得更多的信息和更准确的自然语言生成。

About

Advanced Retrieval-Augmented Generation (RAG) through practical notebooks, using the power of the Langchain, OpenAI GPTs ,META LLAMA3 ,Agents.
通过实用笔记本进行高级检索增强生成(RAG),利用 Langchain,OpenAI GPTs,META LLAMA3,Agents 的力量。

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published