@geopavlakos noticed that the memory usage was extremely high when using the function visibility_compute inside of a training loop. I traced it back to visibility_destructor never being called to clean up the search tree that is created when it isn't supplied as an argument.
The output of the function doesn't seem to be dependent on the search tree, so I simply delete it right before the return if it wasn't supplied as an argument. If there's a different approach that you would like, I'd be more than happy to make any changes.
@geopavlakos has been testing this fix for a while now and it has been stable without any side effects.
@geopavlakos noticed that the memory usage was extremely high when using the function visibility_compute inside of a training loop. I traced it back to visibility_destructor never being called to clean up the search tree that is created when it isn't supplied as an argument.
The output of the function doesn't seem to be dependent on the search tree, so I simply delete it right before the return if it wasn't supplied as an argument. If there's a different approach that you would like, I'd be more than happy to make any changes.
@geopavlakos has been testing this fix for a while now and it has been stable without any side effects.