How to Set Up and RunDeepSeek-R1 671B with ollama + docker + webui on windows

Setting up and running DeepSeek-R1 671B with Ollama, Docker, and WebUI on Windows requires a robust hardware setup, proper software configuration, and careful optimization. Follow this step-by-step guide to deploy it effectively.

---

Prerequisites

1. Hardware Requirements

Ensure your system meets the minimum hardware requirements for DeepSeek-R1 671B:

• CPU: High-end multi-core processor (AMD Threadripper / Intel Xeon)
• RAM: 1TB+ DDR5 ECC
• GPU: Multi-GPU setup, such as:
• 20x Nvidia RTX 3090 (24GB each)
• 10x Nvidia RTX A6000 (48GB each)
• Storage: At least 10TB+ NVMe SSD
• Power & Cooling: Sufficient power supply and cooling solutions for multi-GPU operation

---

2. Install Required Software

Step 1: Install Docker Desktop on Windows

1. Download Docker Desktop from Docker’s official site.
2. Enable WSL 2 integration during installation.
3. Verify installation by running:

   docker --version

4. Make sure your system meets the Docker resource requirements (allocate at least 512GB RAM & multiple GPUs in Docker settings).

---

Step 2: Install Ollama

1. Download and install Ollama from Ollama’s official site.
2. Verify the installation:

   ollama --version

---

Step 3: Pull the DeepSeek-R1 671B Model

1. Pull the DeepSeek-R1 model using Ollama:

   ollama pull deepseek-r1:671b

---

Step 4: Install WebUI for DeepSeek-R1

1. Clone the DeepSeek WebUI repository:

   git clone https://github.com/deepseek-ai/webui.git
   cd webui

2. Build and run WebUI inside Docker:

   docker build -t deepseek-webui .
   docker run --gpus all -p 7860:7860 -v deepseek_cache:/root/.cache deepseek-webui

---

3. Running DeepSeek-R1 671B

Option 1: Run Using Ollama

ollama run deepseek/deepseek-r1-671b

Option 2: Run with WebUI

1. Open your browser and navigate to:

   http://localhost:7860

2. Start interacting with DeepSeek-R1 671B via the web interface.

---

4. Optimizations for Performance

• Enable CUDA & TensorRT: Ensure NVIDIA CUDA & TensorRT are installed for better GPU acceleration.
• Allocate More RAM to Docker: Modify Docker settings to allocate at least 512GB RAM.
• Use Multi-GPU Scaling: Leverage NVIDIA NCCL for better scaling across multiple GPUs.
• Enable Shared Memory in Docker: Use --shm-size=512g to increase shared memory:

  docker run --gpus all --shm-size=512g -p 7860:7860 deepseek-webui

---

Conclusion

You have now set up DeepSeek-R1 671B on Windows using Ollama, Docker, and WebUI. This setup ensures a seamless AI model deployment while leveraging your hardware efficiently.