This paper investigates a parameter tuning method for the learning rate decay factor of the learning rate schedulers applied in deep learning.Learning rate scheduler is a technique used to adap-tively adjust the model's learning rate,which can help the model converge and generalize better.This paper proposes a method to adjust the learning rate decay factor for the learning rate schedulers in trans-fer learning,which can quickly and accurately improve the model's fitting speed and accuracy,thereby enhancing the model's performance on the target task.
关键词
迁移学习/学习率衰减因子/深度学习
Key words
transfer learning/learning rate decay factor/deep learning
维普
万方数据