what is the metrics of BlendedLoss()?
Ahandsomenaive opened this issue · comments
Thank you that you can share your codes with others.
then i do not understand a question in codes.
class BlendedLoss(object):
def init(self, main_loss_type, cross_entropy_flag):
super(BlendedLoss, self).init()
self.main_loss_type = main_loss_type
assert main_loss_type in MAIN_LOSS_CHOICES, "invalid main loss: %s" % main_loss_type
self.metrics = []
metrics is one set of some metric functions??
here you are not introduction it?
hope your reply! thank you
Before we make the codes public, there was online-triplet loss implementation which is from here.
Because self.metrics is only for the online-triplet loss, there is no meaning for now.
I'll remove that part soon.
Thank you.