[英]Ensemble (Combine) multiple deep learning regression models which already have dropout layers
Currently I have multiple trained models for regression task, each model is of the same architecture but while training, I have dropout layer, to improve the performance, is that still possible for me to combine those trained models and calculate the mean of the weights as the combined, new model? 目前,我有多个训练有素的模型用于回归任务,每个模型都具有相同的体系结构,但是在训练时,我具有辍学层以提高性能,我仍然可以将这些训练后的模型结合起来并计算权重的平均值合并的新模型? I just heard that there is an ensemble prediction method which allow us to do, but I am not sure whether I can still do this because I already have random dropout layer.
我刚刚听说有一种整体预测方法可以使我们做到这一点,但是我不确定是否仍然可以做到这一点,因为我已经有了随机的辍学层。
Any hint is much appreciated! 任何提示,不胜感激!
I think the presence of dropout is irrelevant to what you want to do. 我认为辍学与您要执行的操作无关。 Ensembling should work just fine with dropout.
集成应该可以很好地与辍学一起工作。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.