Why HR and MRR are getting smaller and smaller when training with AttRec

cheungdaven / DeepRec

An Open-source Toolkit for Deep Learning based Recommendation with Tensorflow.

GNU General Public License v3.0

1.13k stars 294 forks source link

Closed linh47 closed 4 years ago

linh47 commented 4 years ago

The first output is

Epoch: 0005; HR@50:0.5005302226935313 MRR:0.08745163611856224 AUC:0.9075360408864436 2020-02-07 09:50:03

then

Epoch: 0123; HR@50:0.46977730646871685 MRR:0.08698753010475276 AUC:0.8941231973552161 2020-02-07 10:00:21

Aren't these two bigger the better????

linh47 commented 4 years ago

您好，我跑完了5000个epoch，我发现越训练这两个值越小，是不是损失函数或者训练过程有问题呢？希望能得到您的回复…

cheungdaven commented 4 years ago

@linh47 到了一定步骤会过拟合，一般不需要那么长时间的训练。