This project fine-tunes a large language model (LLM) locally using iMessage chat history to replicate your communication style. By keeping the process on your device, it ensures privacy, preventing data leakage to external servers. The goal is to create a personalized AI that generates organic, contextually relevant messages, matching how you typically text.
pip install -r requirements.txt
Run imessage_mlx.ipynb in your IDE.
- Incorperating RAG with your message history, attachments, calendar, notes, etc. to create a personal assistant and better messages per-person
- https://github.com/ShawhinT/YouTube-Blog/tree/main/LLMs/qlora-mlx: Local Fine-tuning on Mac (QLoRA with MLX)
- https://github.com/gavi/mlx-whatsapp: An mlx project to train a base model on your whatsapp chats using (Q)Lora finetuning
- https://github.com/ishan0102/iClone: Clone your friends with iMessage and MLX