Overview
A model fine-tuned for Apple M1 hardware with the MLX framework. It builds on SmolLM-135M-Instruct-4bit using LoRA over 50 iterations on Tiny Shakespeare, and is tuned for an M1 with 8GB of RAM. It loads through mlx-lm and huggingface_hub.
Installation
git clone https://github.com/coccinella-labs/mlxlm