How to build a local PDF RAG solution with zero code
In this video, we learn how to run open-source LLMs locally using Ollama!
🚀🤖🌐
We will explore the user-friendly Open WebUI for model access, and build a local RAG solution with ZERO code. You can index all your financial PDFs, ask questions and predict the next market crash😊.
We will also build a simple chatbot app with Ollama’s local API.
Hope you will learn something new and enjoy the video🎥.
Timestamp:
0:00 Intro
1:12 Ollama setup
7:18 Open WebUI setup
11:15 Open WebUI Login
15:15 Chat with multiple LLMs
18:20 Local RAG solution with Open WebUI
24:36 How to use Ollama local API
28:45 Use API in a Python app
38:10 Demo time
Code to build chatbot with local Ollama API:
github.com/nhtkid/chatwithLLM...
![[Best LLM] Best Image Analysis Multimodal Review - Jul 2024 Update](/img/d.gif)
Пікірлер: 11
Amazing video, thank you so much.
Gold! Great video
Great vid
How do you set threshold/similarity scores for QnA with Documents and also if it can reference the page number/file?
@gptOdyssey
21 күн бұрын
No you can't. It's just an easy-to-use RAG solution with Open WebUI, off-the-shelf with no customisation. Have a look at their RAG code if you are interested. github.com/open-webui/open-webui/blob/2fa94956f4e500bf5c42263124c758d8613ee05e/backend/apps/rag/main.py#L328
Hey your computer can't handle 3 models at the same time unless you have more RAM to run them all.
@gptOdyssey
Ай бұрын
Yeah seems like the Mac is not powerful enough.
"In this short video....." Dude, it is a 41min video.
@gptOdyssey
Ай бұрын
Lol you are right. I was hoping to make a short video. Hopefully the timestamp helped!
this does not have rag !!!
@gptOdyssey
21 күн бұрын
docs.openwebui.com/tutorial/rag