Download a model.
Try a conversation.
By Alvin
In the workshop Terminal window, run these lines one at a time. pull downloads the model files when needed, list shows installed names, and run starts a conversation. Each download needs internet access and disk space; replies use the installed local model.
ollama pull qwen3.5:4b ollama pull llama3.2:latest ollama list ollama run qwen3.5:4b
Recognize the chat prompt
When >>> appears, you are talking to the model. Ask a short question, such as “What is a Python function?” Wait for the reply. Then type /bye and press Return to get back to Terminal. In our installed Ollama 0.32.0, /? listed /bye as Exit; we exercised that route. [41]
Know what the reply gives you
This chat produces text. A coding agent can also write project files using its tools. The next page’s driver sends a complete brief, saves the returned Python function and checks its behavior on invented notes. It supplies the file-writing and testing steps for this exercise.