Thanks for your awesome work. I found that some scenes have multiple identical objects. How can I accurately locate the grasp poses of each instance?