@xinyulyu

A FastAPI application that uses a pre-chunked version of Romeo and Juliet consisting of 90 text chunks. It generates embeddings for each chunk using Azure OpenAI and stores them for semantic search. When a user asks a question, the application converts the question into an embedding, finds the most relevant chunks based on vector similarity, and provides them as context to an Azure OpenAI language model, which then generates an answer.

My own Webseite that tells about me. I Host it with my Raspberry Pi. And a AI Chat with my AI (Qwen 0.8b Fine turned by me). Its also Running on my Raspberry Pi