L1

class dragon.vm.tensorflow.keras.regularizers.L1(l1=0.01)[source]

The L1 regularizer.

The L1 regularizer is defined as:

\[loss_{reg} = loss + \alpha|w| \]

__init__

L1.__init__(l1=0.01)[source]

Create a L1 regularizer.

Parameters:
  • l1 (float, optional, default=0.01) – The value to \(\alpha\).