What is PEFT?

PEFT (Parameter-Efficient Fine-Tuning) is a set of techniques that enable the efficient adaptation of pre-trained language models (PLMs) to various downstream tasks. Instead of fine-tuning all the parameters of a large PLM, PEFT methods only fine-tune a small number of (extra) model parameters, thereby greatly decreasing the computational and storage costs. Common PEFT methods include LoRA (Low-Rank Adaptation), Prefix-Tuning, Prompt Tuning, and Adapter Tuning. These techniques are commonly used to adapt large language models to specific tasks while minimizing resource consumption.