Hey everybody 👋 I'm Alex, and over the past few weeks I've been building LlamaChat as a native, open-source macOS app to allow you to easily chat with LLaMA models locally. 🔮 Alongside the ChatGPT boom of the last few months, there has been a flurry of projects building on top of Meta's LLaMA LLM to make it more amenable to conversations. Projects include Stanford's Alpaca, Nomic AI's GPT4All and Vicuna. 🚀 Thanks to the excellent llama.cpp kickstarted by Georgi Gerganov, it is now easier than ever to run these models locally on your device. LlamaChat wraps this up in a native macOS interface, with a simple chat UI and support for importing and converting your raw PyTorch model checkpoints. 💬 Up next, I'd like to add support for other LLaMA-based projects, like Vicuna and Koala, as well as non-English language models like Chinese LLaMA/Alpaca and Vigogne. For the ML dabblers I'd like to add to the existing simple debugging and tweaking support (currently LLamaChat allows you to inspect the existing model context including raw tokens).

LlamaChat

Chat with your favourite LLaMA models right on your Mac

Chat with your favourite LLaMA models right on your Mac