2019最牛的梯度优化算法出炉,AdaBound实验对比代码

论文:Adaptive Gradient Methods with Dynamic Bound of Learning Rate 论文地址:https://openreview.net/pdf?id=Bkg3g2R9FX github地址:https://github.com/Luolc/AdaBound   AdaBound可以被视为一个优化器,随着训练步数的变大,它会从Adam动态转换为SGD
相关文章
相关标签/搜索