Hi there!

From me with love.

According to https://instantview.telegram.org/contest/vulture.com/template106/issue1/

First of all.
<div> with class "image-collection" is a slideshow
<div> with class "img-figure" contains slide from slideshow.
<div> with class "img-figure" contains two elements:
1) <div> with class "image-wrapper" is a set of images
2) <script> (next-sibling element) contains json structure where you can get:
- author
- caption
- content utl
- etc
The logic is simple - each image contains own caption and source. You can catch it in clear way. The original visual caption demonstrates it.

Secondly.
Sometimes some galleries has own caption that could be visually define before phrase "From left:". In this case you should leave gallery caption like a caption. Look at gallery "Hill, streetwear icon, in New York."

Thirdly.
How about user experience? What is the best looking thing - image with own caption and source or big gallery caption? Gallery with a couple of slides and the short descriptions