L2 regularization with only weight parameters

The code looks fine.
I’m not sure, if an L2 regularization of the bias terms leads to overfitting.
Do you have any references on it, as it’s quite interesting?

1 Like