A New Comprehensive Parameter Adjustment Method for Learning Rate Attenuation Factor in Transfer Learning Scenario
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.
transfer learninglearning rate decay factordeep learning
万方数据
维普
