In Chapter 5, you will head into generative Large Language Models (LLMs) and how to fine-tune them. With Retreival Augmented Generation (RAG) you create a specialized assistant that serves as a companion for robot programming.

For Entering Chapter five click here: Chapter 5!

Welcome to the fifth day of our hands-on course!

Today, you will head into generative Large Language Models (LLMs) and how to fine-tune them. With Retreival Augmented Generation (RAG) you create a specialized assistant that serves as a companion for robot programming. The software used is completely open-source and can be installed on your personal machines. For this course we offer the webservice RAGflow, utilizing models from Ollama to digest and extend its knowledge.

Goal: By the end of this session, you will know how to fine-tune an existing LLM with a knowledgebase in the fashion of RAG, and define the assistants behavior through (initial) prompt-engineering.

Prerequisites

Laptop with internet connection

(Optional) If run on your own machine:

16GB RAM
GPU recommended

Theoretical Background

Build your LLM knowledgebase with the content of the past weeks lectures.
Prompt-engineer to constraint and form an assistants behavior.
Refine your model to improve the assistent.

Step-by-Step Hands-On Exercises

Scope and Recap: We will get an overview of useful lecture material, presented over the past week, to refresh your memory and collect that as training data.
Introduction to RAGflow: What is Retreival Augmented Generation and how can you set it up for any kind of application you need?
Discuss first impressions: Gather our first ideas on strength and weakness of generative LLMs and RAG.
Refine your assistant: Exceed boundaries and try to break the system, explore creative ways of forming your assistant.
Share experiences: Condense the experience we made by sharing them with your peers.

Access to RAGFlow

TBD

Interactive Actions and/or Examples

Summary

By the end of this session you will have experience with the difficulties of configuring your own assistant, and in what ways fine-tuning can change the assistants behavior.