It would be great to be able to see the training and validation loss on the same set of axes, so that they can be more easily compared without having to go back and forth between the two plots. This is especially true when the two y-axes are on different scales. It makes it very difficult to determine how close they are to one another.
It would be great to be able to see the training and validation loss on the same set of axes, so that they can be more easily compared without having to go back and forth between the two plots. This is especially true when the two y-axes are on different scales. It makes it very difficult to determine how close they are to one another.