Fine-tuning LLMs to 1.58bit: extreme quantization made easyBy Hugging Face - Blog / September 18, 2024