论文与标准

LoRA: Low-Rank Adaptation of Large Language Models

冻结预训练权重，用低秩矩阵学习参数更新，降低模型微调的参数与存储开销。

Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen · 2021