For Keras on TensorFlow, a layer class constructor comes with these:
For example, Dense layer:
The first one,
kernel_regularizer is easy to understand, it regularises weights, makes weights smaller to avoid overfitting on training data only.
kernel_regularizer enough? When should I use