Advanced Applied Deep Learning by Umberto Michelucci

Author:Umberto Michelucci
Language: eng
Format: epub
ISBN: 9781484249765
Publisher: Apress

We can define a loss function that depends on x1 with this code6 (what the loss is doing is not relevant):def custom_loss(layer):

def loss(y_true,y_pred):

return K.mean(K.square(y_pred - y_true) + K.square(layer), axis=-1)

return loss

Then we can simply use the loss function as before:model.compile(optimizer='adam',

loss=custom_loss(x1),

metrics=['accuracy'])

This is an easy way to develop and use custom losses. It is also sometimes useful to be able to train a model with multiple losses, as described in the inception networks. Keras is ready for this. Once you define the loss functions you can use the following syntaxmodel.compile(loss = [loss1,loss2], loss_weights = [l1,l2], ...)

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.