I would like to know if the class weighting is also used in evaluating the loss in the validation data during the training. If not, is there a way to adjust the fit() function so that it takes into account the class weighting during the validation loss calculation.
$\begingroup$
$\endgroup$
1
-
$\begingroup$ I wonder why you want to add weights to the class during validation!! $\endgroup$– enterMLSep 7, 2017 at 5:52
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
2
As described in Keras.models.py[1], validation_data can be a tuple (inputs, targets, sample_weights). So, you can use this parameter for class weighting during the validation loss calculation.
[1] https://github.com/fchollet/keras/blob/master/keras/models.py
-
1$\begingroup$ This is sample weight, which means that I have to check for each class sample and make a vector that has the right class weight for each sample $\endgroup$– HatemBSep 4, 2017 at 12:49
-
1$\begingroup$ @moh Any idea how to supply class weight for validation data when using a generator? $\endgroup$– EranSep 26, 2018 at 20:09