Solution to model training not converging or model overfitting

Some very useful blogs:

1、

mm​​​​​​​​​​The training set loss does not decrease_Nie Xiaoxian's Blog-CSDN Blog_The training set loss does not decrease

2、

Deep learning network not converging? Network output all zeros? How should it be checked? _Xinyu_cheng's blog-CSDN blog_Deep learning output all 0

3、

[Deep Learning] - Over-fitting processing method