|
|
--- |
|
|
title: WHEC Chatbot |
|
|
emoji: 🧬 |
|
|
colorFrom: indigo |
|
|
colorTo: green |
|
|
sdk: docker |
|
|
pinned: false |
|
|
--- |
|
|
|
|
|
# WHEC RAG System |
|
|
|
|
|
This is a Multimodal Retrieval-Augmented Generation (RAG) system designed to answer questions based on the documents at WHEC (Warrior Heat- and Exertion-Related Events Collaborative) page: https://www.hprc-online.org/resources-partners/whec |
|
|
|
|
|
It allows users to query both text and images extracted from the documents. |
|
|
|
|
|
## How it Works |
|
|
The system uses a pre-built index (`multimodal_rag_index`) that contains embeddings for both text and images found in the research papers. When you ask a question, the system retrieves the most relevant text segments and images, then generates a consolidated answer using GPT-4o. |
|
|
|
|
|
## Features |
|
|
- **Multimodal Search**: Retrieves relevant text and images simultaneously. |
|
|
- **Visual Evidence**: Displays images extracted from the papers alongside the answer. |
|
|
- **Source Citations**: Provides references to the specific documents and pages used. |
|
|
|
|
|
## Usage |
|
|
Simply type your question in the search box. You can ask general questions or specifically request visual information (e.g., "Show me images of..."). |
