Experience with Deepseek R1

A Hands-On Look at Deepseek R1

I'm incredibly impressed with Deepseek! I'm eager to share my experience so you can explore its capabilities firsthand. This guide provides a practical walkthrough:

Step 1:

1. Navigate to ollama.com.
2. Download the Ollama application.
3. Select a language model compatible with your computer's resources.

Step 2: Open your terminal or command prompt and execute the following commands:

Bash

ollama run llama3.2
ollama run deepseek-r (select the 7b version)

Step 3:

1. Visit docker.com.
2. Download and install Docker Desktop, choosing the appropriate version for your operating system.
3. Minimize the Docker Desktop application.
4. In your terminal, run:

Bash

ollama run OpenWebUI

Step 4:

1. Go to openwebui.com and register for an account.
2. Consult the documentation at docs.openwebui.com and copy the provided Docker installation script (shown below).
3. In your terminal, execute the copied script, which will resemble this:

Bash

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

4. Verify the container's ID within the Docker Desktop interface.

Step 5:

1. Log in to OpenWebUI at https://openwebui.com/auth.
2. You should see a screen similar to the one below.

In the "Select a model" area, choose Deepseek. Use the "+" button to add a second model, such as Llama3.2 (which we previously launched/downloaded).

Now, let's compare Llama3.2 and Deepseek-r1 by engaging in some conversational interactions.

For instance, you might ask:

"Provide a code example for a website's fixed header using CSS and JavaScript."

The system's response will be similar to the example shown below. (Include a screenshot or example response here)

Feel free to experiment with different settings and parameters! You can even combine the outputs from the two models by clicking the designated icon at the bottom right.

You can adjust setting and parameters thro chat controls which is availabel at right top icon.

Deepseek R1 offers a broad spectrum of natural language processing abilities. Try baby Try!

For additional guidance on using these models, please refer to the following resources:

• https://www.youtube.com/watch?v=5kFV20LatL8&t=606s (With Credits and Thanks)

• https://www.youtube.com/watch?v=syR0fT0rkgY&t=3s for deployment