Llamaindex Local Models

TL;DR: Build a RAG application using llamaindex and local models (embedding + LLM), with weave for LLM observability

📦 Packages and Basic Setup

💿 The Dataset

In this example, we'll use the original Jurassic Park screenplay to act as our dataset.

✍️ Model Architecture & Training

Since we're using all local models in this example, we'll have to our own Embedding model and llm. In this particular example we'll use "BAAI/bge-small-en-v1.5" as our local embedding model and "Writer/camel-5b-hf" as the local LLM.

🗂 Creating a Index

Based on the value you set for config.fetch_index_from_wandb we can either create our own index, or simply download the index stored as an artifact.