Ollama

Install Ollama

curl -fsSL https://ollama.com/install.sh | sh

Run Ollama

Pick a model:

https://ollama.com/library

The smallest distilled model:

ollama run deepseek-r1:1.5b

Let's use a bigger model, DeepSeek-R1-Distill-Llama-8B

ollama run deepseek-r1:8b

Talk to Ollama in the Terminal

Once the program starts running, you can access it from the terminal.

Talk to Ollama in a Jupyter Notebook

Open a Notebook using the Pytorch kernel

import ollama