Finetuning Large Language Models on One Consumer GPU in 2 Bits - View it on GitHub