Is your feature request related to a problem? Please describe.
Feature request from Clara users:

with hybrid loss we need to display each loss separately so that user can calibrate the weights

loses are on different scale to begin with dice range is 0-1 while cross entropy is 0-inf and depends on the size of the volume

need to have out put similar to the one in TF below
Epoch: 1/5, Iter: 1/11 [=                   ]  train_ce_loss: 0.3277  train_dice_loss: 0.9091  train_loss: 4.1865  time: 16.89s        
Epoch: 1/5, Iter: 2/11 [===                 ]  train_ce_loss: 0.2897  train_dice_loss: 0.8875  train_loss: 3.7847  time: 12.38s