reidmen's picture
using docker to trigger building
4ec36ec
---
title: DeepSeek-R1 WebGPU
emoji: 🧠🐳
colorFrom: red
colorTo: blue
sdk: docker
pinned: false
app_port: 7860
license: apache-2.0
short_description: Small reasoning model running in browser
thumbnail: >-
https://huggingface.co/spaces/webml-community/deepseek-r1-webgpu/resolve/main/banner.png
---
# DeepSeek-R1 WebGPU
Next-generation reasoning model running entirely in your browser using WebGPU acceleration.
## System Requirements
- WebGPU-compatible browser
- 6GB+ RAM recommended
## Features
- πŸš€ Runs entirely in browser using WebGPU acceleration
- 🧠 1.5B parameter reasoning model
- πŸ’­ Shows step-by-step reasoning process
- πŸ”’ LaTeX math support
- πŸŒ™ Dark mode support
## Getting Started
Follow the steps below to set up and run the application.
### 1. Clone the Repository
Clone the examples repository from GitHub:
```sh
git clone https://github.com/huggingface/transformers.js-examples.git
```
### 2. Navigate to the Project Directory
Change your working directory to the `deepseek-r1-webgpu` folder:
```sh
cd transformers.js-examples/deepseek-r1-webgpu
```
### 3. Install Dependencies
Install the necessary dependencies using npm:
```sh
npm i
```
### 4. Run the Development Server
Start the development server:
```sh
npm run dev
```
The application should now be running locally. Open your browser and go to `http://localhost:5173` to see it in action.
## Usage
1. Click "Load model" to download and initialize the model
2. Wait for model initialization to complete (~1-2 minutes)
3. Type your question or select an example
4. View the model's step-by-step reasoning
## Browser Compatibility
- βœ… Chrome
## Troubleshooting
- **WebGPU Not Available**: Ensure you're using a compatible browser with WebGPU enabled
- **Out of Memory**: Try closing other browser tabs to free up GPU memory
- **Model Loading Fails**: Check your internet connection and try reloading
## License & Attribution
- Model: [DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B) by DeepSeek AI
- Framework: [πŸ€— Transformers.js](https://github.com/huggingface/transformers.js)
- License: Apache 2.0