LlamaIndex RAG Chat

Perform RAG (Retrieval-Augmented Generation) from your PDFs using this Colab notebook!
Powered by Llama 2

Features

• Free, no API or Token required
• Fast inference on Colab's free T4 GPU
• Powered by Hugging Face quantized LLMs (llama-cpp-python)
• Powered by Hugging Face local text embedding models
• Set custom prompt templates
• Prepared Chat mode (not QA)
  
  

Getting started

1. Open in colab
2. Make sure the Colab's Runtime Type is set to T4 GPU (at least)
3. Edit preferences in Block 4
4. Upload your PDF into Files (Default name: rag_data.pdf)
5. Runtime > Run all