I wrote custom loss function in TF 2.3 where the loss is comprised of several sub-losses, since I want to track the sub-losses I\'ve printed them using tf.print:
tf.print
