DeepSeek-R1: Next-Generation Local Reasoning Model

Experience cutting-edge reasoning directly in your browser with DeepSeek-R1, a 1.5B parameter model powered by WebGPU. Everything runs locally with no data sent to servers, ensuring privacy and performance. Built with 🤗 Transformers.js and ONNX Runtime Web, it’s lightweight, offline-capable, and blazing-fast at 60 tokens per second. More information can be found here.

Why DeepSeek-R1?

DeepSeek-R1 is designed for developers, researchers, and enthusiasts who need a fast, secure, and privacy-focused reasoning model. By running locally in your browser, DeepSeek-R1 eliminates server latency and ensures complete control over your data.

Features

• Local Execution: Runs entirely in your browser—no external dependencies or server-side processing using php.
• WebGPU Acceleration: Leverages WebGPU for fast, efficient computation.
• Offline Support: Once loaded, you can use the model without an internet connection.
• Privacy First: All processing stays on your device.
• Open Source: Fully transparent, with source code available on GitHub.

Note: Code is tested on MacOS and Linux (Ubuntu).

Getting Started

Follow these steps to set up and run DeepSeek-R1 locally.

1. Clone the Repository

Clone the project to your local machine:

git clone http://localhost:8000/

2️. Navigate to the Project Directory

Move into the project directory:

cd deepseek-r1

3. Start the Application Server

Start the local development server with:

bash run.sh

4. Example

example.mov

The application will open automatically in your default web browser at http://localhost:8000.

🔗 Base Code Attribution

This project builds upon the excellent work available in the Hugging Face Examples.