Add 'ollama serve' between steps 3 and 4

[adolfont]

I got an error after command 3. I had to do 'ollama serve' first.

I am not sure it is necessary a new step.

3. Pull the model:

ollama pull llama3:8b

4. Run the model:

ollama run llama3:8b