A parameter-efficient fine-tuning method that adapts large models using quantization and low-rank updates.