Closed
Description
It might be possible to quickly visualize at a glance roughly how much training is left by reporting the number of epochs since the Best Model, and/or as a fraction relative to Training Patience. The Best Model could also be marked on the plot with a red dot. This way, one could quickly get a feel for if it makes sense to stop training early or not.
Plotting the Validation Loss with black markers might also help with discriminability, I've illustrated this idea below.
.
Thanks!