deepseek v3

Run DeepSeek V3 on Ubuntu: Step by Step Installation Guide

Anas Mohammad

Mar 25, 2025 • 3 min read

DeepSeek V3 is one of the most advanced AI models available today, designed to excel in natural language processing, reasoning, and decision-making tasks.

Running DeepSeek V3 on Ubuntu allows developers to harness its capabilities locally, avoiding reliance on cloud services while maintaining full control over their environment.

This article provides a detailed, step-by-step guide to installing and running DeepSeek V3 on Ubuntu, covering prerequisites, installation procedures, troubleshooting tips, and optimization strategies.

Overview of DeepSeek V3

DeepSeek V3 is a mixture-of-experts (MoE) language model with 671 billion parameters, 37 billion of which are activated per token. It is trained on 14.8 trillion high-quality tokens and incorporates innovative features such as:

Multi-Head Latent Attention (MLA): Enables the model to focus on multiple aspects of input simultaneously for improved inference efficiency.
Auxiliary-Loss-Free Load Balancing: Enhances training stability by evenly distributing computational load without relying on auxiliary loss functions.
Multi-Token Prediction: Allows the model to predict multiple tokens simultaneously, boosting performance and enabling speculative decoding for faster inference.

These features make DeepSeek V3 suitable for applications like code generation, data analysis, search engines, recommendation systems, and more.

System Requirements

To run DeepSeek V3 effectively on Ubuntu, your system must meet certain hardware and software specifications:

Minimum Requirements

Operating System: Ubuntu 18.04 or newer.
CPU: Multi-core processor (minimum 4 cores).
GPU: NVIDIA GPU with at least 8GB VRAM (CPU-only runs are possible but slower).
RAM: At least 8GB.
Storage: 4–50GB free space depending on the model size.
Python Version: Python 3.10 or higher.

Recommended Requirements

For smoother performance:

GPU: NVIDIA GPU with more VRAM for larger models.
RAM: 16GB or more.
Storage: Higher storage capacity for larger datasets.

Step-by-Step Installation Guide

Step 1: Update Your System

Ensure your system is up-to-date before proceeding with installations:

sudo apt update && sudo apt upgrade -y

Step 2: Install Python and Git

Python and Git are essential for managing dependencies and cloning repositories:

sudo apt install python3 python3-pip git -y

Verify installations:

python3 --version
pip3 --version
git --version

Step 3: Install Ollama

Ollama simplifies running large language models locally:

curl -fsSL https://ollama.com/install.sh | sh
sudo systemctl start ollama
sudo systemctl enable ollama

Step 4: Download the DeepSeek Model

Use Ollama to download the DeepSeek model:

ollama run deepseek-r1:7b

This process may take time depending on your internet speed. Verify installation:

ollama list

Step 5: Set Up Web UI

For an interactive interface, install Open WebUI:

Start the server:

open-webui serve

Install Open WebUI:

pip install open-webui

Create a virtual environment:

sudo apt install python3-venv -y
python3 -m venv ~/open-webui-venv
source ~/open-webui-venv/bin/activate

Access the Web UI at http://localhost:8080 in your browser.

Troubleshooting Common Issues

Issue 1: Insufficient Hardware Resources

If your system struggles to run DeepSeek V3 due to limited resources:

Consider using smaller model variants (e.g., 7B instead of 671B).
Upgrade your GPU or RAM.

Issue 2: Installation Errors

Ensure Python and pip versions are compatible with DeepSeek requirements. Reinstall dependencies if necessary.

Issue 3: Slow Performance

Running on CPU-only setups can be slow. Use an NVIDIA GPU for faster inference.

Optimizing Performance

Enable FP8 Mixed Precision

DeepSeek V3 supports FP8 mixed precision for reduced memory usage and faster computation.

Use Speculative Decoding

Leverage multi-token prediction to enhance inference speed.

Cloud-Based Alternatives

If local hardware limitations persist, consider cloud-based solutions like Novita AI.

Applications of DeepSeek V3

DeepSeek V3’s versatility makes it suitable for various tasks:

Code Analysis and Completion: Ideal for developers working on complex projects.
Search Engines: Improves accuracy and efficiency in retrieving relevant information.
Recommendation Systems: Enhances user experience by understanding query context.
Content Management Systems: Filters documents based on complex queries.

Conclusion

Running DeepSeek V3 locally on Ubuntu provides unparalleled access to its advanced AI capabilities. By following this guide, you can set up and optimize the model effectively while overcoming common challenges.

Whether you’re a researcher, developer, or AI enthusiast, DeepSeek V3 opens new possibilities for innovation in natural language processing and beyond.