Ollama Essentials — Familiarization Course

0Orientation & objectives

This short course familiarizes a new operator with the day-to-day essentials of running local models through Ollama: the core command-line interface, the context-window setting that most often causes silent failures, and how local, cloud, and coding-tool execution relate. It is a demonstration course built from this repository's own wiki pages.

Learning objectives. On completion a participant will be able to run and manage models from the command line, set the context window at the correct level for a given task, and distinguish local, cloud, and integrated-tool execution.

Element	Detail
Format	Self-paced; approximately 20 minutes
Formative questions	Multiple-choice, per section (feedback, not scored)
Final assessment	8 multiple-choice questions; pass mark 75%
Sources	This repository's wiki, built from official documentation (2026-07-07)

1Command-line basics

Learning objective. Run, retrieve, list, and unload models from the command line, and identify which command reports models loaded in memory versus models on disk.

A model chat starts with ollama run <model>; /bye exits an interactive session. Models are retrieved and removed with ollama pull <model> and ollama rm <model>. A custom model is built from a Modelfile with ollama create -f Modelfile, and the server itself is started with ollama serve. (source: ollama-cli-reference)

Two listing commands are easy to confuse: ollama ls lists the models downloaded to disk, while ollama ps lists the models currently loaded and running in memory. ollama stop <model> unloads a running model immediately. (source: ollama-cli-reference)

Check your understanding

1.1 Which command lists the models currently loaded and running in memory?

1.2 How do you unload a running model from memory immediately?