low-rank adaptation (LoRA)

In deep learning, fine-tuning is an approach to transfer learning in which the parameters of a pre-trained neural network model are trained on new data. (Wikipedia, CC BY-SA)