Implementere en KI fag chatbot

Abstract

In this paper, we look into the potential of using Petals to run large lan- guage models. Petals is a system for running large language models shared across several computers in a peer-to-peer fashion. We explore how to inte- grate Petals with Discord to create an AI course chatbot. We test Petals’s performance of this task through comparing it with Ollama, a framework for running large language models on your own hardware. We deploy the chatbot in the Discord server for web programming students and encourage everyone to test it and provide feedback on their preferred answer. Feed- back from students helps us evaluate how suitable the Petals system is for an interactive chatbot designed for the web programming course at the University of Stavanger. We conclude that Petals in its current state is not suited for use as a course Discord chatbot because of its slow speed, challenging framework, and in- ability to give it course specific information.

In this paper, we look into the potential of using Petals to run large lan- guage models. Petals is a system for running large language models shared across several computers in a peer-to-peer fashion. We explore how to inte- grate Petals with Discord to create an AI course chatbot. We test Petals’s performance of this task through comparing it with Ollama, a framework for running large language models on your own hardware. We deploy the chatbot in the Discord server for web programming students and encourage everyone to test it and provide feedback on their preferred answer. Feed- back from students helps us evaluate how suitable the Petals system is for an interactive chatbot designed for the web programming course at the University of Stavanger. We conclude that Petals in its current state is not suited for use as a course Discord chatbot because of its slow speed, challenging framework, and in- ability to give it course specific information.