I see that the detection model can be slow if an input image has a large size.
Are there any guidelines on an "optimal" image size to pass to the model?
Thanks!