5.1 Inspection of the loss function#
N2V "ground-truth"
N2V is a very peculiar deep-learning algorithm. It is a so-called self-supervised method. What is used as "ground-truth" in N2V?
N2V loss
Given your answer to the previous question, how can we interpret the N2V loss function value?
Low number of validation
What could happen to the validation loss curve if you have too little validation examples?