如何在Keras中对整个模型应用权重归一化？

Question

I have a huge model with lots of constitutional layers each having the W_regularizer=l2(0.01) parameter. I would like to remove it from each layer declaration an apply it to the model as a whole. Is it possible to do this on Keras?

Answer 1

Doesn't look like it from the docs

Regularizers allow to apply penalties on layer parameters or layer activity during optimization. These penalties are incorporated in the loss function that the network optimizes.

The penalties are applied on a per-layer basis. The exact API will depend on the layer, but the layers Dense, Conv1D, Conv2D and Conv3D have a unified API.