Lit-LLaMA is an implementation of the nanoGPT-based LLaMA language model that supports quantization, LoRA fine-tuning, and pre-training.

Design Principles

• Simple: single file implementation, no boilerplate code
• Correct: Numerically equivalent to the original model
• Optimized: run on consumer hardware or at scale
• Open source: no strings attached

set up

Clone repository

git clone https://github.com/Lightning-AI/lit-llama
cd lit-llama

install dependencies

pip install -r requirements.txt