I am looking to make Grounding Dino into a few-shot learner where lets say I give it a few images of german shephards as example image, It should be able to detect and box out germna shephards specifically from a test input image. Anything like this out there?