Run a model on your laptop

1. 01Install Ollama and pull a small model

   Everything in this course can start on a model this size. Bigger models come later, when you can measure whether they are worth it.

   ollama pull llama3.2 && ollama list

   You pass when: The list shows llama3.2 with a size of about 2 GB.

   Check step 1

2. 02Ask it one question from the terminal

   ollama run llama3.2 "In one sentence, what is a token?"

   You pass when: One sentence prints and you get your prompt back.

   Check step 2

3. 03Call the same model over HTTP

   This is the same request shape most hosted models accept. Every later lab is this call with more around it.

   curl -s http://localhost:11434/v1/chat/completions -H 'Content-Type: application/json' -d '{"model":"llama3.2","messages":[{"role":"user","content":"Say hi in three words."}]}'

   You pass when: JSON comes back with the answer under choices[0].message.content and a usage block that counts prompt and completion tokens.

   Check step 3

The check runs on the computer, not on your word: a step passes when its command exits with 0 there.