Skip to content

Latest commit

 

History

History
31 lines (24 loc) · 1.35 KB

README.md

File metadata and controls

31 lines (24 loc) · 1.35 KB
Raw

Multi-Model Llama Chat

Welcome to the Multi-Model Llama Chat! This is a versatile chatbot that supports various Llama quantized models, allowing you to chat with different AI models of your choice.

Features

  • Apple Silicon Support: The chatbot is optimized for Apple Silicon.
  • Intel CPUs Compatibility: For Intel CPUs, ensure to install the non-metal binding. Run npx --no node-llama-cpp download --no-metal in the terminal.
  • Model Selection: Choose from multiple Llama models stored in the models folder at startup.
  • Flexible Chat Experience: Interact with various models through a simple command-line interface.

Getting Started

  1. Clone the Repository:

    git clone https://github.com/shatfield4/multi-model-llama-chat

  2. Install Dependencies: Navigate to the project directory and run yarn or npm install to install the necessary dependencies.

  3. Model Placement: Place your .gguf format Llama models in the models directory.

  4. Run the Chatbot:

    node index.js

Note for Intel CPU Users

If you are on an Intel-based MacOS, it's crucial to install the no-metal binding of node-llama-cpp. Run the following command before starting the chatbot:

npx --no node-llama-cpp download --no-metal

Enjoy conversing with a variety of AI models through the Multi-Model Llama Chat!