`Callback` that adds AR and TAR regularization in RNN training

RNNRegularizer(alpha = 0, beta = 0)

Arguments

alpha

alpha

beta

beta

Value

None