How to Run Deepseek AI Locally in Intel AI Playground

Summarize This Article:

AI Chatbots have been a hot topic in the tech industry for some time now. While ChatGPT, Google Gemini, and Microsoft Copilot are some great options to choose from, however, a new entrant, Deepseek AI, shook the market with its advanced reasoning capabilities, open-source technology, performance capabilities, and more.

If you’re an avid YTECHB reader, you may already know that we have an Intel Arc PC at our office. We’ve tested several new things on our machine including running Deepseek R1 locally.

There are multiple ways to run Deepseek AI locally on Windows 11, but if you own an Intel Arc PC, one of the easiest methods is to use the Intel AI Playground app.

You can follow our step-by-step guide and install and run the Deepseek R1 AI model locally on your Intel Arc machine. Before diving into the tutorial, take a look at the prerequisites.

Requirements:

Intel Core Ultra H-series processor (Core Ultra 155H or 125H) or Intel Core Ultra 200v series processor
Intel Arc GPU Series A or Series B
16-20GB of data is required to download AI models and AI Playground
Windows OS (10 or 11)

Setting Up Deepseek R1 AI Model in Intel AI Playground

First things first, you need to install the Intel AI Playground app on your system. If the app is already installed, ensure that you have the latest version. You can navigate to GitHub’s AI Playground release page to check the latest version.

How to Run Deepseek AI Locally in Intel AI Playground

At the time of writing, v2.2.1 beta is the latest version and it allows you to easily download Deepseek R1. The model comes pre-loaded in the app, simply click the download option to install the model.

It can take a while to download and install the model on the machine.

Steps for older Intel AI Playground app:

If you are on the older version of the AI Playground app, you’ll need to manually download the AI model on your machine. Open the Playground app and navigate to the Answer tab. Now click the + icon placed on the top of the prompt box and paste the AI model name from the Hugging Face models.

Search for the DeepSeek R1 7B model on Hugging Face, copy the model name and paste it into the AI Playground app to download the model.

Time for Testing

Now you can ask anything you want to Deepseek. Simply type your query in the prompt box and click Generate. You will instantly get the response back and the best part is that everything runs locally on your machine.

So, I entered this prompt, “How many times ‘r’ appear in Raspberry?” Take a look at the reply.