/llama-app

Run LLaMA inference on CPU, with Rust 🦀🚀🦙

Primary LanguageRustMIT LicenseMIT

LLaMa app

This is a simple app to use LLaMa language models on your computer, built with rust, llama-rs, tauri and vite.

Features

  • model selecting from your computer or download alpaca 7B from the app
  • starts a new thread with the model running when first message sent (or pressed start button)
  • remembers the context while the model is running (no context saving yet)
  • ends the thread when pressed stop
  • user can change prompt and parameters
  • saves conversations and settings to local storage

Installation

Download the latest release here.

Building and Developing

At first install dependencies with pnpm install from the root directory.

To build the app run pnpm tauri build from the root. To run the app in dev mode run pnpm tauri dev, but the text generation is very slow.

screenshot