PrivateGPT Step by Step: Chat with your private documents without internet connection
Private GPT Tool: https://github.com/imartinez/privateGPT
Download model from here: https://gpt4all.io/index.html
Download the embedding model names from here: https://www.sbert.net/docs/pretrained...
---
Welcome back, trailblazers of AI! In a world where artificial intelligence is becoming the norm, we are here to offer something exciting yet unique – a powerful tool that respects your privacy, meet 'privateGPT.' Leveraging the prowess of large language models, privateGPT empowers you to interact with your documents in ways you've never imagined before! But hold on, it gets better. Imagine doing all this without needing an internet connection. Yes, you heard it right! Ask questions, ingest documents, and make sense of your information - all offline. Stay tuned as we dive deep into how you can revolutionize the way you interact with your data, all with privateGPT. Remember, with privateGPT, your privacy is our priority. No data leaves your execution environment - Ever. Let's get started!
*Notes:*
internet is needed to download embedding model
*Findings:*
seems very useful right out of the bat because I can share my personal data and ask questions.
But it’s kinda buggy( shows the answer document, but can’t exactly say the answer)
Slow
Play around with models( both embedding and LLM) to get better results for your use case
*Steps:*
Step 1: git clone
Step 2: pip install requirements
Step 3: copy example env
Step 4: download model from gpt4all into models folder
Step 5: make changes to env if needed
Step 5: ingest
Step 6: run inference
00:00 Overview of PrivateGPT
02:54 Step by step begins