RAG with Llama-Index: Vector Stores

No video

RAG with Llama-Index: Vector Stores

Рет қаралды 12,982

Күн бұрын

In this third video of our series on Llama-index, we will explore how to use different vector stores in llama-index while building RAG applications. We will look at self-hosted solution (chroma-DB) and cloud-based solution (Pinecone).
CONNECT:
☕ Buy me a Coffee: ko-fi.com/prom...
|🔴 Support my work on Patreon: Patreon.com/PromptEngineering
🦾 Discord: / discord
▶️️ Subscribe: www.youtube.co...
📧 Business Contact: engineerprompt@gmail.com
💼Consulting: calendly.com/e...
Links:
Vector stores in LlamaIndex: tinyurl.com/2p...
Google Colab: tinyurl.com/2s...
llamaIndex playlist: • Llama-Index
Timestamps:
[00:00] Intro
[00:21] Vector stores in llamaIndex
[01:24] Basic Setup
[02:45] Upload files
[03:45] Self-Hosted Vector Store
[09:00] Cloud based Vector Store

Пікірлер: 25

@engineerprompt 10 ай бұрын

Want to connect? 💼Consulting: calendly.com/engineerprompt/consulting-call 🦾 Discord: discord.com/invite/t4eYQRUcXB ☕ Buy me a Coffee: ko-fi.com/promptengineering |🔴 Join Patreon: Patreon.com/PromptEngineering

@user-em4ld3zc9y 8 ай бұрын

This tutorial series is great ! Best one I found so far. Thank you for sharing this.

@gregorykarsten7350 10 ай бұрын

Great work. Excellent topic. Llama index opens up so much more possibility for RAG. Im very much interested in building a knowledge base. That gets added to on a daily basis. What do think of knowledge graph in this context

@anilshinde8025 9 ай бұрын

great video. Thanks. waiting for addition of Local LLM in the same code

@vitalis 10 ай бұрын

Super interesting, looking forward to the video

@engineerprompt 10 ай бұрын

Thank you 🙏

@hassentangier3891 10 ай бұрын

Awesome Work, Like Always. Can you refer to documentation or video "on how to update the chromadb in this context"

@fuba44 10 ай бұрын

This was great, love this kind of content! ❤❤❤

@engineerprompt 10 ай бұрын

Thank you 🙏

@kdlin1 6 ай бұрын

Why is OpenAI API Key needed when it does not use OpenAI? Thanks!

@arkodeepchatterjee 10 ай бұрын

please make the video comparing different embedding models

@Kishorekkube 10 ай бұрын

Self hosting? Seems interesting

@smoq20 10 ай бұрын

I always seem to run into the problem of exclusions when using vector similarity search for RAG. I.e. when you run a query for "Tell me everything you know about dogs other then Labradors." guess which documents will be returned as first 10 (assuming you have a lot of chunks)? Yes, about Labradors. Has anyone figured a way around that yet? I've been attempting to filter out results if queries include exclusions with additional LLM passes, but only GPT4 seems to have enough brains to do it correctly. PaLM 2 gets it right in 50% of cases.

@saikashyapcheruku6103 6 ай бұрын

Is there a way to bypass the rate limit error for openai api? Additionally, why is the openai being used even after specifically mentioning the service context?

@scorpionrevenge 6 ай бұрын

I keep receiving this error : cannot import name 'Doc' from 'typing_extensions' I am trying to run your codes on jupyter notebook environment. Can you please help and let me know how to create a vector db?

@Rahul-zq8ep 8 ай бұрын

Great I understood most of the explanation in video but Where is the RAG implementation in it ? I have also created a vector_store, storage_context, index etc when I was implementing chatBot with my data, but I am confused on how to implement RAG as an added functionality ?

@srikanth1107262 7 ай бұрын

Would like to have a video on local download model ( llama2 ggml/gguf ) using llamaindex to build rag pipeline with chormadb. Thank you for videos its helps a lot.

@user-hq6or3fh9d 6 ай бұрын

hey i have question as we have injested our data to the vector db how do retrive answer without runnin the injestion code all the time

@chrisksjdvs603 4 ай бұрын

setting up the vector store as persistent should help like he says in the video. Once you have your data stored you just need to load the vector store to communicate with the data if I understand it correctly