Keras Successive Regularization Wrapper

A wrapper that slows down the updates of trainable weights.

Install

pip install keras-succ-reg-wrapper

Usage

import keras
from keras_succ_reg_wrapper import SuccReg

input_layer = keras.layers.Input(shape=(1,), name='Input')
dense_layer = SuccReg(
    layer=keras.layers.Dense(units=1, name='Dense'),
    regularizer=keras.regularizers.L1L2(l2=1e-3),  # Any regularizer
    name='Output',
)(input_layer)
model = keras.models.Model(inputs=input_layer, outputs=dense_layer)
model.compile(optimizer='adam', loss='mse')
model.summary()