and chat in RAG style
Is an efficient, multi-language chat interface implemented using Apple’s MLX framework that supports Search Enhanced Generation (RAG).
You can easily and automatically integrate any HuggingFace and MLX-compatible open source model. Support chat interactions with Doc(x), PDF, txt files and YouTube videos via URLs.
Main functions:
1. Chat with data: Support chat interactions with doc(x), pdf, txt files and YouTube videos through URLs.
2. Multi-language support: Support Chinese, English, French, German, Hindi, Italian, Japanese, Korean, Spanish, Turkish and Vietnam.
3. Easy integration: You can easily integrate any HuggingFace and MLX-compatible open source model.
4. Supported models: Including Google Gemma, Mistral, OpenHermes, NousHermes, Mixtral, Quyen, StableLM2 Zephyr, Vistral, VBD-Llama2 and other models.
Project address:https://github.com/qnguyen3/chat-with-mlx
Video: