Running Deepseek-R1 on Windows with Ollama
If you’re looking for a light, straightforward way to explore large language models on Windows, Ollama is a great place to start. Rather than wrestling with dual-boot Linux or WSL drivers, you can take advantage of Ollama’s native Windows support—it’s easier to set up and still makes the most of your consumer-grade GPU.
On a system powered by an AMD Radeon RX 7800 XT, for instance, Ollama can tap into the GPU’s horsepower to speed up model inference—often beating what you’d see under WSL. In this guide, we’ll show you how to install Ollama on Windows, run the deepseek-r1:7b model, and enable local network access so you can quickly share your AI experiments with others.
System Requirements
While Ollama can run on plenty of Windows machines, here’s one example that shows off its capabilities:
OS: Windows 11 Pro (Build 26100)
CPU: AMD64 Family 25 Model 97 (~4701 MHz)
RAM: 32GB Physical Memory
GPU: AMD Radeon RX 7800 XT (4GB VRAM)
The actual requirements may be lower, but having a dedicated GPU and sufficient RAM will significantly improve performance when running large language models.
Installing Ollama on Windows
Installation is straightforward using Windows Package Manager (winget):
winget install Ollama.Ollama
Running the deepseek-r1:7b Model
Once Ollama is installed, you can download and run the deepseek-r1:7b model with a simple command:
ollama run deepseek-r1:7b
Useful Ollama Commands
Here are some helpful commands for managing your Ollama installation:
# Start the Ollama server
ollama serve
# List available models
ollama list
# If you need to force-quit Ollama
taskkill /F /IM ollama.exe
Testing Local Access
You can test your Ollama installation locally using either WSL Ubuntu or your browser. Here’s a simple curl command to verify it’s working:
curl http://localhost:11434/api/tags
Enabling Network Access
If you want to access the LLM from a remote machine on your local network, you have a couple of options:
Option 1: Setting Environment Variables
On Windows, Ollama can use your system environment variables to allow network access:
- Quit Ollama from the taskbar.
- Open Settings (Windows 11) or Control Panel (Windows 10) and search for environment variables.
- Edit or create a new variable for
OLLAMA_HOST
and set it to0.0.0.0:11434
. - Restart Ollama from the Start menu.
Option 2: Using ncat
Alternatively, you can use ncat to expose Ollama:
- Download ncat from the Nmap website.
- Use the command
ncat -lk 0.0.0.0 11434 --sh-exec "ncat 127.0.0.1 11434"
to allow network access.
Testing Remote Access
To test remote access, other devices on your network can use this curl command (replace IP address with your Windows machine’s IP):
curl -X POST http://192.168.4.175:11434/api/generate -H "Content-Type: application/json" -d '{
"model": "deepseek-r1:7b",
"prompt": "Hello!",
"stream": false
}'
A successful response will look something like this:
{
"model": "deepseek-r1:7b",
"created_at": "2025-02-21T15:52:58.4296691Z",
"response": "<think>\n\n</think>\n\nHello! How can I assist you today? 😊",
"done": true,
"done_reason": "stop",
"context": [151644,9707,0,151645,151648,271,151649,271,9707,0,2585,646,358,7789,498,3351,30,26525,232],
"total_duration": 221054700,
"load_duration": 9989100,
"prompt_eval_count": 5,
"prompt_eval_duration": 2000000,
"eval_count": 16,
"eval_duration": 207000000
}
Security Considerations
When exposing Ollama to your local network:
- Only do this on trusted networks
- Be aware that any device on your network can send requests to your Ollama instance
- Consider implementing additional security measures like a reverse proxy with authentication if needed
Troubleshooting
If you encounter issues:
- Ensure Ollama is running (
ollama serve
) - Check if the port 11434 is not being used by another application
- Verify your Windows firewall settings allow traffic on port 11434
- Make sure ncat is properly installed and in your system PATH