While it is helpful to have the torch optimizer return the loss as it iterates, I think it would be VERY helpful to provide a .loss() function to the registration class, so that the user can print out the current loss value once the registration is finished. Seems like it could be an easy call to the self._optimizer?
While it is helpful to have the torch optimizer return the loss as it iterates, I think it would be VERY helpful to provide a .loss() function to the registration class, so that the user can print out the current loss value once the registration is finished. Seems like it could be an easy call to the
self._optimizer
?