Contents:
How are the scores distributed? How do the different algorithms perform?
How does Google’s Universal Sentence Encoder perform on these captions?