Regularization of activations once the model is initialized

Question

I would like to add regularization of activations in Tensorflow.keras on a pretrained network, using a loop over layers.

If I want to regularize weights or biases, I can do:

l1=0.001
l2=0.001
for layer in model.layers:
    if isinstance(layer, DepthwiseConv2D):
        layer.add_loss(regularizers.l1_l2(l1,l2)(layer.depthwise_kernel))
    elif isinstance(layer, layers.Conv2D) or isinstance(layer, layers.Dense):
        layer.add_loss(regularizers.l1_l2(l1,l2)(layer.kernel))
    if hasattr(layer, 'bias_regularizer') and layer.use_bias: 
        layer.add_loss(regularizers.l1_l2(l1,l2)(layer.bias))

As far as I understood and tested: this is working.

However I do not see clearly how to do this for activations regularization. Specifically, I want to add to the loss the OUTPUT of the Activation layer.

I guess I should do something like:

for layer in model.layers:
    if isinstance(layer, Activation):
        layer.add_loss(regularizers.l1_l2(l1,l2)(layer.XXX))

But it is not clear to me what should be replacing XXX in the above.

Thanks in advance for your help!

Answer 1

But it is not clear to me what should be replacing XXX in the above.

layer.output