Running DeepSeek-R1 with Ollama using Only Docker: A Complete Guide

Introduction

DeepSeek-R1 is a powerful open-source LLM (Large Language Model) that can be easily run using Ollama inside Docker. This guide will walk you through setting up DeepSeek-R1 on a normal laptop with just Docker. If you have an NVIDIA GPU, an optional section will cover GPU acceleration.

By the end of this guide, you will be able to:

• Run Ollama in Docker with just a normal laptop.
• Pull and run DeepSeek-R1 using only CPU (no need for GPU).
• Enable GPU acceleration if your system has an NVIDIA GPU.
• Run the entire setup with a single command for ease of execution.
• Optionally use a Web UI for a better experience instead of CLI.

Prerequisites (CPU Execution - Recommended for Most Users)

This guide is structured to prioritize CPU usage, ensuring that any normal laptop with Docker installed can run DeepSeek-R1 efficiently.

• Only Docker is required (Install Docker).
• No special hardware is needed—your normal laptop will work!
• 16GB+ RAM recommended (for smooth performance).

Step 1: Pull and Run Ollama in Docker (CPU Only)

Ollama provides a convenient runtime for models like DeepSeek-R1. We will first run Ollama inside a Docker container.

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

This will:

• Start Ollama in a Docker container.
• Expose it on port 11434.
• Persist downloaded models using a volume (ollama:/root/.ollama).

To verify the container is running:

docker ps

Step 2: Pull and Run DeepSeek-R1 Model (CPU Only)

Now that Ollama is running, we can pull and execute DeepSeek-R1.

Pull DeepSeek-R1 Model

docker exec -it ollama ollama pull deepseek-r1:8b

Run DeepSeek-R1 (CPU Mode)

docker exec -it ollama ollama run deepseek-r1:8b

Step 3: Running Everything in One Command (CPU Only)

docker run -d -v ollama:/root/.ollama -p 11434:11434 --name ollama ollama/ollama && \
docker exec -it ollama ollama pull deepseek-r1:8b && \
docker exec -it ollama ollama run deepseek-r1:8b

---

Optional: Running DeepSeek-R1 with Web UI

If you prefer a graphical interface instead of using the command line, you can set up a Web UI for Ollama and DeepSeek-R1.

Step 1: Run Open WebUI with Ollama

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

This will:

• Start Open WebUI, which provides a web-based chat interface for DeepSeek-R1.
• Expose it on http://localhost:3000.
• Connect it to the running Ollama container.

Now, open your browser and navigate to http://localhost:3000 to chat with DeepSeek-R1 using an easy-to-use UI.

---

Optional: Running DeepSeek-R1 with GPU Acceleration

If you have an NVIDIA GPU, you can enable GPU acceleration for improved performance.

Prerequisites (GPU Execution)

• NVIDIA GPU (with CUDA support).
• NVIDIA Drivers installed (Check GPU Compatibility).
• Docker with NVIDIA Container Toolkit installed.

Step 1: Run Ollama in Docker (With GPU Support)

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

Step 2: Run DeepSeek-R1 with GPU

docker exec -it ollama ollama run deepseek-r1:8b --gpu

Step 3: Running Everything in One Command (GPU Enabled)

docker run -d --gpus=all -v ollama:/root/.ollama -p 11434:11434 --name ollama ollama/ollama && \
docker exec -it ollama ollama pull deepseek-r1:8b && \
docker exec -it ollama ollama run deepseek-r1:8b --gpu

Step 4: Verify GPU Utilization

To ensure DeepSeek-R1 is using your GPU, check NVIDIA System Management Interface (nvidia-smi):

docker exec -it ollama nvidia-smi

You should see processes running under the GPU Memory Usage section.

---

Step 5: Stop and Remove Ollama Docker Container

If you ever need to stop and remove the container, use:

docker stop ollama && docker rm ollama

This will:

• Stop the running Ollama container.
• Remove it from the system (but the model files will persist in the Docker volume).

---

Conclusion

In this guide, we covered how to:
✅ Set up Ollama in Docker.
✅ Pull and run DeepSeek-R1 using just a normal laptop (CPU only).
✅ Enable GPU acceleration if needed.
✅ Use a Web UI for a better experience.
✅ Execute everything in a single command.
✅ Verify GPU utilization (if applicable).

By following these steps, you can easily deploy DeepSeek-R1 in a Dockerized environment with minimal setup. 🚀