Query-Reference Pairs not clear in the subset json files and other minor issues

[uday-me]

1. In the train/test/val.json files, each element is as follows : "{"location": [0, 197], "label": 280, "id": "JDg--pjY5gg", "groundtruth": [95, 101]}".
   It is unclear what the location is specifying. If it is the location of the segment in the query video, how to determine the query video id. If you could please provide insights into this.

2. How are query and reference videos being mapped ?

3. Moreover, there are 9400 total segments/elements after adding elements in the train/test/val.json files, which is slightly less than the number of segments mentioned in the paper (9530 segments).