jwnt4/qvhighlights-25frames download history
jwnt4/qvhighlights-25frames is a video text to text dataset on the Hugging Face Hub. In the last 30 days it was downloaded 207 times (13 in the last 7 days), and 916 times in total. It ranks #65,726 among datasets by monthly downloads.
QVHighlights Sets: train and val. Test set will be available soon. Number of frames per video: 25 frames. Frame variations: 3 Answer format: 25 frame segmentation mask (e.g. 1010100101100101011110011) Images are in videos.tar.gz (no inside folders). Each image is noted by {video name}_fr