Simple way to run deepseek-r1 locally

Before we begin, let's first understand what DeepSeek is. It is a Chinese-origin company established in 2023. Initially focused on building large language models (LLMs), DeepSeek gained prominence when it launched its open-source, low-cost R1 model. This model competed with OpenAI-o1 and became a game-changer in the AI industry by enabling anyone to run GPT-like LLMs on their own servers.

DeepSeek-R1 features six dense models along with a base model. In this guide, we will run the most distilled version, which is the 1.5B model (approximately 1.1GB). This model is optimized to run smoothly on most devices.

Installation

To run DeepSeek-R1 locally, you will need Ollama, an open-source tool that enables running large language models (LLMs) on your local machine.

• Visit the Ollama download page and download the application compatible with your system.

• Launch the application. Then, open your terminal and execute the following command: ollama run deepseek-r1:1.5b This command will pull the model and start it. Once running, you will see the prompt >>>. You can now enter your text prompts directly.

Usages

Here are some basic commands to get started:

• To view all available models, type: /show
• To exit the model, type: /bye

Wrap up

If you found this guide helpful, consider subscribing to my newsletter on jyotirmoy.dev/blogs , You can also follow me on Twitter jyotirmoydotdev for updates and more content.