Machine Learning in Linux: Ollama – self-hosted Llama 2

Installation

Currently, you’ll have to build from source to run Ollama under Linux. Fortunately, the process is straighforward.

First, clone the project’s GitHub repository with the command:

$ git clone https://github.com/jmorganca/ollama

Change into the newly created directory:

$ cd ollama

Build the software:

$ go build .

We see an error message, but the software builds fine.

Start the server:

$ ./ollama serve &

If you want to run Ollama without needing to use ./ollama each time, add the ollama directory to your $PATH environment variable. We’ll leave this as an exercise for the reader 🙂

The server listens on http://127.0.0.1:11434. If you point your web browser at that address, it’ll confirm Ollama is running.

Let’s test the Llama 2 model. Run the command:

$ ollama run llama2

Ollama proceeds to download the Llama 2 model. You’ll see output like the image below. It’s a 3.8GB download.

The final line displays the prompt >>>

We’re now ready to test the Llama 2 model.

Next page: Page 3 – In Operation

Pages in this article:
Page 1 – Introduction
Page 2 – Installation
Page 3 – In Operation
Page 4 – Summary