Low-Rank Adaptation (LoRA)

A parameter-efficient fine-tuning technique that adapts pre-trained models using a small number of additional trainable parameters.