Skip to content

Latest commit

 

History

History
30 lines (23 loc) · 788 Bytes

File metadata and controls

30 lines (23 loc) · 788 Bytes

SPARK AI customizable (Local LLM Template)

This is a minimalist, terminal-based AI assistant powered by a local GGUF model using llama-cpp-python.

Features

Configurable system prompt and few-shot examples, Clean terminal interface, Logs conversations to /logs, Easily forked and modified for your own builds

Usage

pip install -r requirements.txt
python main.py

If perferred, use the included start.bat or start.sh launchers.

Config

Edit config/config.json to: Point to your .gguf model file, Adjust temp, top_p, output tokens, Enable/disable logging or rich output

Dependencies

llama-cpp-python rich (optional but looks good)

Exit

Built to be modified, enjoy.