General question#
Diverging training and validation losses
What is happening when the training loss becomes very small, while the validation loss does not change or even increases?
What is happening when the training loss becomes very small, while the validation loss does not change or even increases?