Introduction

Running AI models like DeepSeek-R1 locally can be both an exciting learning experience and a powerful way to leverage AI capabilities without relying on cloud-based solutions. In this guide, we’ll explore two different methods for running DeepSeek-R1:

1. Using Ollama directly on your system – A simple, quick setup for running AI models locally.
2. Using Docker for portability and isolation – A containerized approach that ensures consistency across different environments.

By following these steps, you’ll be able to interact with DeepSeek-R1 in no time!

DeepSeek-R1 Model Requirements

Before setting up DeepSeek-R1, it’s essential to understand the model variations and their hardware requirements. DeepSeek offers multiple versions, ranging from lightweight models for personal use to enterprise-grade solutions.

Model Version	Minimum Hardware Requirements
DeepSeek-R1 1.5B	Basic CPU, minimal RAM (for chatbots & personal assistants)
DeepSeek-R1 7B/8B	Dedicated GPU (RTX 3060/3070), suitable for AI writing tools, summarization, and sentiment analysis
DeepSeek-R1 14B/32B	16GB – 32GB RAM, high-end GPU (RTX 4090), ideal for enterprise AI applications and real-time processing
DeepSeek-R1 70B	64GB+ RAM, multiple GPUs, designed for large-scale AI research and business automation
DeepSeek-R1 671B	Server-grade hardware (768GB RAM, NVIDIA A100 GPUs), used for large-scale AI research and commercial applications

Now that we’ve reviewed the hardware requirements, let’s dive into the setup process.

---

Method 1: Running DeepSeek-R1 Directly with Ollama

This method is ideal for those who want a straightforward installation without using Docker.

Step 1: Install Ollama

1. Visit the official Ollama website and download the appropriate version for your operating system.
2. Follow the installation instructions for your OS.

Step 2: Download DeepSeek-R1

Ollama provides different versions of DeepSeek-R1, ranging from 1.5B to 671B parameters. For this guide, we’ll use the 8B version.

Run the following command in your terminal to download DeepSeek-R1 8B:

ollama run deepseek-r1:8b

Step 3: Verify Installation

Check if the model is installed correctly by running:

ollama list

You should see deepseek-r1 listed as one of the available models.

Step 4: Run the Model

To start the model, execute:

ollama run deepseek-r1:8b

You can now interact with DeepSeek-R1 in the terminal by entering prompts. Congratulations! You’ve successfully set up DeepSeek-R1 locally.

---

Method 2: Running DeepSeek-R1 Inside a Docker Container with Ollama

If you prefer to run DeepSeek-R1 in an isolated environment, Docker is the best option. It ensures portability and avoids dependency conflicts.

Step 1: Run the Ollama Docker Container

Start the Ollama container using the following command:

docker run -d -v ollama:/root/.ollama -p 11434:11434 --name ollama ollama/ollama

• -d runs the container in detached mode.
• -v ollama:/root/.ollama mounts a volume for persistence.
• -p 11434:11434 exposes Ollama’s API on port 11434.

Step 2: Access the Container & Download the Model

Log into the running container:

docker exec -it ollama /bin/bash

Inside the container, pull the DeepSeek-R1 1.5B model (or any version you prefer):

ollama pull deepseek-r1:1.5b

Step 3: Run the Model

Once the model is downloaded, start it with:

ollama run deepseek-r1:1.5b

Now you can interact with the model inside the container.

Step 4: Enable Web-Based Interaction

For an improved experience, you can use a Web UI to interact with DeepSeek-R1. Run the following command to deploy Open-WebUI:

docker run -d -p 3000:8080 -e OLLAMA_BASE_URL=http://localhost:11434 -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main

• This command launches a web-based interface for interacting with DeepSeek-R1.
• After starting the container, navigate to http://localhost:3000 in your browser.

Step 5: Verify Running Containers

To check running containers, use:

docker ps

You should see both Ollama and Open-WebUI containers running.

---

Conclusion

Both methods allow you to run DeepSeek-R1 locally, each with its own advantages:

• Direct installation via Ollama is quick and easy.
• Using Docker ensures isolation and compatibility across different systems.