DeepSeek WebUI Installation using Ollama in Nvidia Jetson Nano

This blog shows you how to set up DeepSeek-R1, a powerful local reasoning model, using Ollama and Open WebUI interface simillar to ChatGPT.

Prerequisites

Python 3.11 or higher
Conda package manager
Admin/sudo privileges for installation
8GB RAM minimum (16GB+ recommended)
Note: The demo is tested in Linux/Ubuntu OS

Step-by-Step Installation Guide

1. Create and Activate Conda Environment

First, set up a dedicated virtual environment using anaconda:

conda create -n webui python=3.11 -y && conda activate webui 

2. Install Ollama

Ollama is a lightweight model server that manages and runs AI models locally. Choose your operating system below for installation instructions:

Linux

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

# Verify installation
ollama --version

macOS

Download the latest version from Ollama for macOS
Open the downloaded .dmg file
Drag Ollama to your Applications folder
Launch Ollama from Applications

Windows

Download the installer from Ollama for Windows
Run the downloaded .exe file
Follow the installation wizard
Launch Ollama from the Start menu

3. Install Open WebUI

Set up the web interface using open-webui:

pip install open-webui

4. Install DeepSeek Models using Ollama

Choose your preferred model size:

# DeepSeek-R1-Distill-Qwen-1.5B
ollama run deepseek-r1:1.5b

You can install other DeepSeek models using Ollama.

# DeepSeek-R1
ollama run deepseek-r1:671b

# DeepSeek-R1-Distill-Qwen-7B
ollama run deepseek-r1:7b

# DeepSeek-R1-Distill-Llama-8B
ollama run deepseek-r1:8b

# DeepSeek-R1-Distill-Qwen-14B
ollama run deepseek-r1:14b

# DeepSeek-R1-Distill-Qwen-32B
ollama run deepseek-r1:32b

# DeepSeek-R1-Distill-Llama-70B
ollama run deepseek-r1:70b

5. Launch the Web Interface

Start the WebUI server:

open-webui serve

6. Access the Interface

Open your browser
Navigate to http://localhost:8080

Demo

After installation, access the WebUI through your browser and start interacting with DeepSeek-R1.

Nvidia Jetson Nano

DeepSeek-R1 works on edge devices like the Nvidia Jetson Nano, using only 8 GB of RAM. I tested it with 1.5b parameters, and it runs smoothly.