Choose with understanding
Read a model card, distinguish open weights from open source, and choose a model that fits your task and computer.
CyberCorps.uk · Windows course
Artificial Intelligence: Open Weight Models
Run a language model on your own computer. Learn what open weights give you, get started with Ollama, and build a small assistant whose answers you can test.
- Guided lessons
- 8
- Suggested study time
- 3–4 hours
- Starting level
- Beginner
Follow the native Windows route using the Ollama desktop application, PowerShell, and Python 3. Keep your prompts, Modelfile, Python clients, and test evidence together in a CyberCorps-Ollama folder under your Windows user profile.
What you will be able to do
Run models locally
Install Ollama, pull a model, chat in the terminal, and manage downloaded and running models.
Build a repeatable workflow
Use the local API and a Modelfile to turn a useful prompt into a reusable assistant.
Evaluate the result
Test factual accuracy, uncertainty, and instruction following with a small, documented evaluation set.
Before you begin
- A Windows release and hardware supported by Ollama's current Windows guide; check available disk space and memory before downloading models.
- Windows PowerShell 5.1 or PowerShell 7, plus a plain-text editor such as Notepad.
- Permission to install the native Ollama app and, for the API lessons, Python 3 from the official Python distribution.
- Network access for software and model downloads. The course's local inference requests use gemma3:1b and require no cloud API key.
Each lesson includes a practical task, a success checklist, and a knowledge check. Keep your prompts, settings, and observations in a learning log as you work.
Your learning path
From first model to working assistant
Work through the lessons in order, or return to a topic when you need a refresher.
Lesson 1 · 20 minutes
Understand open weight models
Understand weights, inference, licences, and the difference between local and cloud execution.
Open lesson →Lesson 2 · 25 minutes
Install Ollama and check your setup
Prepare your computer and verify that your local Ollama service is ready.
Open lesson →Lesson 3 · 25 minutes
Choose and run your first model
Download a small model, start a chat, and inspect the model running on your computer.
Open lesson →Lesson 4 · 25 minutes
Prompt clearly and evaluate answers
Ground answers in supplied notes and test accuracy, format, and uncertainty with repeatable prompts.
Open lesson →Lesson 5 · 30 minutes
Use the local Ollama API
Send chat requests from your terminal and Python, then read and validate the response.
Open lesson →Lesson 6 · 25 minutes
Create a reusable model with a Modelfile
Package a system instruction and generation settings into a named local assistant.
Open lesson →Lesson 7 · 25 minutes
Manage models, memory, and privacy
Inspect resource use, troubleshoot common problems, and keep local inference under your control.
Open lesson →Lesson 8 · 35 minutes
Build and assess a local study assistant
Bring the course together in a small project with evidence, evaluation, and a clear handover.
Open lesson →
Keep the essentials close
Find the main Ollama commands, a local API request, and the terms used throughout the course.