
The authors propose a novel parameter-efficient quasi-orthogonal fine-tuning method (qGOFT) that enhances the adaptation capability of pretrained language models to downstream tasks while maintaining parameter efficiency.


coremsg

parameter-efficient-quasi-orthogonal-fine-tuning-for-efficient-adaptation-of-pretrained-language-models


Parameter-Efficient Quasi-Orthogonal Fine-Tuning for Efficient Adaptation of Pretrained Language Models
