The video_to_text (VTT) task testing dataset in 2019 consisted of the following:

1- Flickr
  - A set of 1010 videos from Flickr were used and can be downloaded after submitting the data agreeemnt form. The files names are the Video Ids and range from 1 - 1010.
    The agreement form can be found here: https://www-nlpir.nist.gov/projects/tv2019/Flickr_Org.Form.txt
    Please send the signed form to asad.butt@nist.gov

2- Vines
   - A text file called Vines_URL_List. This file has 2 columns: the first column is the Id of the video, and the second column is the URL.
   - You will have to download all URLs to work on the task and submit results using the Video Ids. The Ids range from 1011 - 2054.

2- matching.ranking.files:
  
   - tv19.vtt.descriptions.A
   - tv19.vtt.descriptions.B
   - tv19.vtt.descriptions.C
   - tv19.vtt.descriptions.D
   - tv19.vtt.descriptions.E

   In each of the above files, the first column represents the Id of a text description, followed by the description.
   The task is to match the videos to their descriptions in each of the description files.