Plot loss history

[mostaf...@gmail.com]

I am trying to plot the loss history. However, when I do:

    net1_log= net1.fit(X_train, Y_train, nb_epoch=3, batch_size=16, show_accuracy=True, callbacks=[History])

I get: 

   NameError: name 'History' is not defined

What should I do?

[Al]

Do you have:

from keras.callbacks import Callback

and:

history = LossHistory()

[Al]

I forgot one part:

class LossHistory(Callback):

    def on_train_begin(self, logs={}):

        self.losses = []

    def on_batch_end(self, batch, logs={}):

        self.losses.append(float(logs.get('loss')))

On Monday, August 24, 2015 at 12:05:11 PM UTC-7, Mostapha Benhenda wrote:

[Mostapha Benhenda]

is this part necessary?

 because if I look into https://github.com/fchollet/keras/blob/master/keras/callbacks.py

there is already:

class History(Callback):

how can I use it?

[Eric Nichols]

I also was using a locally-defined LossHistory() class in my own work. But the History class looks great; you still need to instantiate an object of that class though. I believe something like this will work:

history = History()

model.fit(...., callbacks=[history])

# And then something like

print history.history

print history.totals

--
You received this message because you are subscribed to the Google Groups "Keras-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to keras-users...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/keras-users/2d04be81-5847-4bd6-a7c7-b2e1c8dfce0d%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Mostapha Benhenda]

thanks, this worked:

from keras.callbacks import History

history = History()

model.fit(X_train, Y_train, nb_epoch=5, batch_size=16, callbacks=[history])

print history.history