The Easiest Way to Download and Use Llama3

1. Download Ollama

On the Ollama website—https://ollama.com/library/llama3—you can see which model tags are available. For example, to use the 4-bit quantized base version of llama8b, simply run:

ollama run llama3

The finetuned quantized version is recommended for better Q&A quality.

ollama run llama3:instruct

It downloads to your local machine by default, about 4.7G in size; once running, you can interact with it directly in the terminal.

2. Download Chatbox

Then simply go to Settings, choose ollama as the backend, and you can use it directly inside Chatbox.