Llama

Llama 3.2 running on a Web browser. Designed in JavaScript and WebAssembly. The model may take several minutes to download (708 MB) and your Web browser must support WebGPU because it's necessary to run the WebLLM engine. The model has 1.23 billion parameters and supports English, German, French, Italian, Portuguese, Hindi, Spanish and Thai.

Web:

https://lrusso.github.io/Llama/Llama.htm

How to run Llama 3.2 locally:

• Install Node.js.
• Open the Terminal.
• Go to the project folder.
• Run node server.js
• Browse to http://localhost

How to run the server in the background:

1. Open the Terminal.
2. Install Forever: npm install -g forever
3. Go to the project folder.
4. Start the server: forever start -a -l /dev/null -c node server.js > /dev/null 2>&1
5. Stop the server: forever stop -a -l /dev/null -c node server.js > /dev/null 2>&1

Based on the work of:

https://github.com/mlc-ai/web-llm