Lab 1: Running Your First Model with Ollama

In this lab, you’ll learn how to run your first language model using Ollama and interact with it through the command line.

Prerequisites

Ollama installed on your system (Installation Guide)
Terminal or command prompt access
At least 5GB of free disk space

Lab Objectives

By the end of this lab, you will:

Download and run a language model
Understand the basic Ollama commands
Interact with the model through the command line
Use system prompts to customize model behavior

Step 1: Verify Ollama Installation

Before we begin, let’s make sure Ollama is correctly installed. Open your terminal and run:

ollama --version

You should see the version number of your Ollama installation. If you get a “command not found” error, please follow the installation guide for your platform.

Step 2: List Available Models

Let’s see what models are available in the Ollama library:

ollama list

If you’ve just installed Ollama, this command will likely return an empty list since you haven’t downloaded any models yet.

Step 3: Download Your First Model

Let’s download the Llama 2 model, which is a good starting point for most users:

ollama pull llama2

This will download the Llama 2 model, which is approximately 3.8GB. The download might take a few minutes depending on your internet connection.

Step 4: Run the Model

Now let’s start a chat session with the model:

ollama run llama2

You’ll see a welcome message from the model, and the cursor will be waiting for your input.

Step 5: Interact with the Model

Try asking the model some questions or giving it tasks. Here are some examples:

“Explain the concept of machine learning in simple terms.”
“Write a short poem about technology.”
“What are three ways to improve productivity?”
“Create a simple Python function that calculates the factorial of a number.”

After each response, you’ll be prompted to enter a new message. To exit the chat session, press Ctrl+C or type /exit.

Step 6: Use System Prompts

System prompts allow you to define the behavior and context for the model. Let’s try using a system prompt:

ollama run llama2 "You are a helpful math tutor. Explain concepts clearly and provide step-by-step solutions."

Now the model will behave as a math tutor. Try asking it math-related questions like:

“Explain the Pythagorean theorem.”
“How do I solve quadratic equations?”
“What’s the derivative of x²?”

Step 7: Customize Model Parameters

You can customize various model parameters using environment variables. For example, to adjust the temperature (randomness) of the responses:

OLLAMA_TEMPERATURE=0.5 ollama run llama2

A lower temperature (like 0.2) will make responses more deterministic and focused, while a higher temperature (like 0.8) will make them more creative and varied.

Step 8: Explore Different Models

Ollama supports many different models. Let’s try a few more:

# Pull and run the Mistral model
ollama pull mistral
ollama run mistral

# Pull and run a smaller model for faster responses
ollama pull tinyllama
ollama run tinyllama

Compare the responses from different models to see how they perform on the same prompts.

Conclusion

Congratulations! You’ve successfully run your first language model with Ollama. You’ve learned how to:

Download and run models
Interact with models through the command line
Use system prompts to customize behavior
Adjust model parameters
Explore different models

Next Steps

Now that you’re familiar with the basics of Ollama, you can proceed to Lab 3: Python Integration to learn how to use Ollama in your Python applications.