Plot validation loss

I am using RetinaNet for object detection. I want to plot the validation loss. Can anyone assist me? and
Whenever I write learn.predict(), this is what I get

clip_remove_empty() missing 2 required positional arguments: ‘bbox’ and ‘label’