CoolFace
Apppublic

ArthyP/technical-rag-assistant

sourceHugging Faceupdated 1y agoView on Hugging Face
0likes
App README

Technical Documentation RAG Assistant

Advanced RAG system with local Ollama LLM integration running in HuggingFace Spaces.

Features

  • โ€”๐Ÿฆ™ Local Ollama LLM - No API dependencies
  • โ€”๐Ÿ” Hybrid Search - Semantic + keyword retrieval
  • โ€”๐Ÿ“š Citation Support - Automatic source attribution
  • โ€”๐ŸŽฏ Confidence Scoring - Answer quality assessment
  • โ€”โšก Fast Performance - Local inference

Models Used

  • โ€”LLM: Llama 3.2 3B (via Ollama)
  • โ€”Embeddings: all-MiniLM-L6-v2
  • โ€”Fallback: HuggingFace API if Ollama fails

How it Works

  1. 1.Upload PDF documents
  2. 2.Documents are chunked and indexed
  3. 3.Ask questions in natural language
  4. 4.Get answers with citations and confidence scores

Built for technical documentation Q&A with Swiss engineering standards.