This project enables running the DeepSeek-R1-Distill-Qwen-1.5B-ONNX model entirely client-side within a web browser. It targets developers and users seeking to leverage powerful language models without server-side infrastructure, offering a convenient and accessible way to interact with AI.

How It Works

The project utilizes WebGPU to execute the ONNX-formatted DeepSeek model directly in the browser. This client-side approach avoids the need for backend servers, reducing infrastructure costs and latency. By leveraging WebGPU, it aims to provide a performant inference experience on compatible hardware.

Quick Start & Requirements

• Install: npm install
• Run: npm run dev
• Requirements: Node.js >= 18.18.0, npm or yarn, a browser with WebGPU support.

Highlighted Details

• Runs DeepSeek-R1-Distill-Qwen-1.5B-ONNX model.
• Entirely client-side execution.
• Leverages WebGPU for inference.

Maintenance & Community

No specific details on contributors, sponsorships, or community channels are provided in the README.

Licensing & Compatibility

The license is not specified in the provided README.

Limitations & Caveats

Performance is dependent on the user's browser and hardware capabilities for WebGPU. The README does not specify compatibility with older browsers or devices lacking WebGPU support.