|
所在平台: Coursera |
课程主页: https://www.coursera.org/learn/project-generative-ai-applications-with-rag-and-langchain
课程评论：没有评论
课程名称：项目：使用RAG和LangChain的生成式AI应用 课程概述：准备好将您的生成式AI工程技能付诸实践！在这个动手指导的项目中，您将应用在程序之前课程中获得的知识和技术，构建自己的真实世界生成式AI应用。 您将首先补充关键知识空白，例如使用LangChain的文档加载器从各种来源导入文档。接着，您将探索并应用文本拆分策略，以提高模型响应能力，并使用IBM watsonx进行嵌入处理。 课程大纲： 1. **文档加载器使用LangChain** - 描述：在本模块中，您将探索用于加载、准备和结构化文档的基本技术，以构建有效的增强检索生成（RAG）应用。您将学习如何使用LangChain的文档加载器导入来自不同来源的内容，应用文档导入的最佳实践，并实施文本拆分策略以增强模型的响应能力。此外，您还将学习在何时以及如何将整个文档纳入提示，以获得最佳输出。通过动手实验，您将获得在现实场景中加载文档和应用文本拆分技术的实践经验。 2. **使用LangChain的RAG** - 描述：在本模块中，您将学习如何使用watsonx的嵌入模型嵌入文档，并使用向量数据库（如Chroma DB和FAISS）存储这些嵌入。您将探索嵌入在RAG管道中的作用，配置向量存储以管理这些嵌入，并使用LangChain进行文档的预处理。此外，您将获得使用LangChain的高级检索器（如基于向量存储的、多个查询、自查询和父文档检索器）高效提取相关信息的动手经验。最后，您将比较基于RAG的方法与使用InstructLab的微调，以评估它们的权衡和适用性。 3. **创建一个能阅读文档的QA机器人** - 描述：在本模块中，您将结合所学的所有组件，使用LangChain和RAG构建一个完整的生成式AI应用。您将学习如何实施RAG以改善信息检索，使用Gradio设置用户界面，并构建一个利用大型语言模型（LLMs）和LangChain的问答机器人，以响应从加载文档中提问的查询。通过动手实验，您将练习构建Gradio界面和开发自己的QA机器人。在最终项目中，您将使用RAG和LangChain构建一个AI应用。辅导材料如备忘单和词汇表将增强您的理解，建立对实施技能的信心，并通过分级测验评估您的学习成果。您将带着一个可部署的AI助手离开这个模块，并明确下一步提升技能的方向。
Name: Document Loader Using LangChain
Description:In this module, you will explore essential techniques for loading, preparing, and structuring documents to build effective retrieval-augmented generation (RAG) applications using LangChain. You will learn how to use LangChain’s document loaders to import content from various sources, apply best practices for document ingestion, and implement text-splitting strategies to enhance model responsiveness. You will also examine when and how to incorporate entire documents into prompts for optimal output. Through hands-on labs, you’ll gain practical experience by loading documents and applying text-splitting techniques in real-world scenarios.
Name:RAG Using LangChain
Description:In this module, you will learn how to embed documents using watsonx’s embedding model and store these embeddings using vector databases, such as Chroma DB and FAISS. You will explore the role of embeddings in RAG pipelines, configure vector stores to manage these embeddings, and use LangChain to preprocess documents for embedding. Additionally, you will gain hands-on experience with advanced retrievers in LangChain, such as Vector Store-Based, Multi-Query, Self-Query, and Parent Document retrievers, to extract relevant information from documents efficiently. Finally, you’ll compare RAG-based approaches with fine-tuning using InstructLab to evaluate their trade-offs and applicability.
Name:Create a QA Bot to Read Your Document
Description:In this module, you will combine all the components you’ve learned to build a complete generative AI application using LangChain and RAG. You’ll learn how to implement RAG to improve information retrieval, set up user interfaces using Gradio, and construct a question-answering bot that leverages LLMs and LangChain to respond to queries from loaded documents. Through hands-on labs, you’ll practice building a Gradio interface and developing your own QA bot. In the final project, you will build an AI application using RAG and LangChain. The supporting materials, like a cheat sheet and glossary, will reinforce your understanding, build confidence in your implementation skills, and assess your learning through a graded quiz. You'll leave this module with a deployable AI-powered assistant and clear the next steps for advancing your skills.
Get ready to put your generative AI engineering skills into practice! In this hands-on guided project, you’ll apply the knowledge and techniques gained throughout the previous courses in the program to build your own real-world generative AI application. You’ll begin by filling in key knowledge gaps, such as using LangChain’s document loaders to ingest documents from various sources. You’ll then explore and apply text-splitting strategies to improve model responsiveness and use IBM watsonx to e
