RAG with LlamaParse, Qdrant and Groq | Step By Step

In this video, I will show you how to create a effective RAG with LlamaParse, Qdrant and Groq. I will explain what LlamaParse is and briefly walk you through the recent blog post they provided. LlamaParse is a state-of-the-art parser designed to specifically unlock RAG over complex documents. I will explain the following in this video.
- How to parse the document using LlamaParse
- Create embeddings from that parsed document and store it in Qdrant Cloud
- Use models from Groq for super fast Inference
- Also show you how to use OllamaEmbeddings and models from Ollama
- Some extra tips to make RAG better.

👉🏼 Links:

NOTE: Readme file in Github has additional links.

------------------------------------------------------------------------------------------

⏰ Timestamps
00:00 Introduction
01:15 LlamaParse blogpost walkthrough
04:16 Getting Started ( GitHub code )
06:13 Llamaparse
10:00 Qdrant Cloud
11:28 FastEmbedEmbeddings
12:44 Models from Groq
17:16 Asking Questions
24:57 Make RAG better ( tips )
26:23 Conclusion
------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------
🔗 🎥 Other videos you might find helpful:

------------------------------------------------------------------------------------------
🤝 Connect with me:

#llamaindex #llamaparser #llamacloud #qdrant #groq #datasciencebasics

[NirantK]

Hey! FastEmbed creator here, thanks for the shout out!

[AdarshMamidpelliwar]

How we can show relevant images as well along with text response in output, pls provide some information or implementation process.

[THE-AI_INSIDER]

Grt video! can you please make a UI around this implementation? using chainlit or something similar?

[THE-AI_INSIDER]

Around 16 min mark, you have suggested to comment vector_store, storage_context, index and uncomment the


# storage_context =
# index =

once the storage is made

I observed that we get the Error - ImportError: cannot import name 'load_index_from_storage' from 'llama_index' (unknown location) for the latest version on llama-index.

hence it is problematic. I found that for version 0.6 to almost 0.8 llama index recognizes load_index_from_storage, but for the latest versions it doesn't, you might want to look into this.

If you find a solution and the correct version please let me know

[AmanKumar-uzri]

how can i add llama parse in open web ui so that i can parse documents and then fed them to ollama models

[Jeganbaskaran]

Lamaparse support image retrieval as well??ex I need to store all the image embedding and retrieve as image based on user questions??

[venkatkumar]

its fantastic video, llamaparser. its possible to extract the table