Adam Dawud

Home / Projects / Active / Computer Science / Custom LLM

0.1 - Custom LLM

project 2025-02-08 2 backlinks
#type/project#topic/ai#topic/computer-science#topic/llm#deep-learning#WSL2#anaconda#area/projects#project/custom-llm#status/active#topic/machine-learning

Project Plan: Custom Fine-Tuned LLM

1. Overview

The purpose of this project is to simply understand how to use, fine tune and apply LLMs in real world scenarios.

Build up a portfoio of LLMs so that iccan show lecturers to and a researcj project in uq

Goal / Desired Outcome:

  • First open source project on github.

2. Objectives & Key Results (OKRs)

List specific objectives and measurable key results for this project.

  1. Objective 1
    • KR1: Understand fundamentally how to use and fine tune LLMs
  2. Objective 2
    • KR1: Create your first git project to get a good portfolio

3. Stakeholders & Areas

  • Stakeholders: List the people or teams involved.
  • Related Area(s): Link to relevant area notes only when they exist.

Current

So i’ve downloaded WSL2 linux. Which stands for Windows Subsystem for Linux. Allows you to run a linux environment without the need of a seperate virtual machine.

I finished downloading Anaconda, which is an Open Source AI and data sceince distribution in python. In a way, its a great tool for Scientific Computing just bcause of how convenient it is, it downloads a bunch of libraries which are particularly useful for data science.

Because of hardware limitations, we must run a 1.5 Quantization of DeepSeekR1.

DeepSeek is a 1.1 Large Language Model