Agentic RAG with LangChain

Подписаться 9 тыс.

Просмотров 10 тыс.

50% 1

In this event, we’ll provide a brief history of Agents before spending time on the details of building what is becoming the industry-standard agentic RAG application, namely, one with access to backup internet searches. Per LangChain's recommended best practices, we’ll use OpenAI Function Calling to build an OpenAI Functions Agent.
Event page: lu.ma/agenticrag
Have a question for a speaker? Drop them here:
app.sli.do/event/8mC2FudkjQao...
Speakers:
Dr. Greg, Co-Founder & CEO
/ gregloughane
The Wiz, Co-Founder & CTO
/ csalexiuk
Join our community to start building, shipping, and sharing with us today!
/ discord
Apply for our new AI Engineering Bootcamp on Maven today!
bit.ly/aie1
How'd we do? Share your feedback and suggestions for future events.
forms.gle/eySGueVq29mTSbN19

Наука

Опубликовано:

19 мар 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 18

@AI-Makerspace 4 месяца назад

Google Colab Notebook: colab.research.google.com/drive/1_8r0Tj-f2UyRZ4SBuaP-RfhdUeXXxu0Z?usp=sharing Event Slides: www.canva.com/design/DAGADvu-BeU/GvKZrigVQMy_q4zFLSNapA/view?DAGADvu-BeU&

@fineline392 4 месяца назад

You guys are a Godsend, thank you for making these available on RU-vid!

@rajatc6527 2 месяца назад

So well stacked and explained! thank you!!

@krlospatrick 2 месяца назад

Excellent video! Thanks for sharing your knowledge

@AI-Makerspace 2 месяца назад

Thanks Patrick!

@MalTramp 4 месяца назад

Thanks for sharing the Google colab notebook. ✌️

@AI-Makerspace 4 месяца назад

Always!

@taylorfans1000 3 месяца назад

I have a usecase where I have to create different agents to work on different functionalities, eg: one will work on creating commands, one for scripts and the other for generating flows, Is this possible using agents? and do i need to create a common vector database for each or do i need to create separate databases for each agent. Can you guys point me in the right direction?

@AI-Makerspace 2 месяца назад

It's definitely "possible"! You can partition the data however you'd like - though I'd recommend keeping them separated for the most part, either by separate vector databases - or by metadata.

@taylorfans1000 2 месяца назад

@@AI-Makerspace Thank you so much ! I tried Different models like GPT3.5, GPT4 , gemini and llama 3, but GPT4 outperforms all of them, is there any particular reason why that happens because we are just making api calls to these models? or it does not depend on the model just on the rag pipeline but the pipeline was the same for all models?

@stanTrX 2 месяца назад

How to correctly define [system message?] Of these agents to make it work correctly? Most of the times it gets confused even i assign tools and additionally i tell it to use that skill/tool, it tries to do something else. In autogen studio

@AI-Makerspace 2 месяца назад

It depends on your model! Which model are you looking at?

@stanTrX 2 месяца назад

@@AI-Makerspace open source models, llama3 and autogen studio

@JohnSmith-he5xg 4 месяца назад

I'm increasingly torn on whether LangChain brings actual value to the problem space or is just constructing a mess of an abstraction that isn't worth it. EDIT: Maybe it'll grow on me. Regardless, nice video