By default Ollama only uses your CPU for processing, but if you have a Nvidia or AMD GPU and want to also use them, folow the steps on the link: https://hub.docker.com/r/ollama/ollama
To download the model, run this on the terminal (take a while to download the model):
curl http://localhost:11434/api/pull -d '{
"name": "llama2"
}'
Or
curl http://localhost:11434/api/pull -d '{
"name": "mistral"
}'
List of the models: https://ollama.com/library