VTC is a large-scale multimodal dataset containing video-caption pairs alongside comments that can be used for multimodal representation learning.
This dataset is a sample of a larger, unfiltered version of the original dataset that we have collected. From the initial version, we handpicked a list of ”safe” subreddits and removed posts if: 1) they had the ”NSFW” or ”over 18” tags; 2) the videos contained faces or the captions contained toxic or offensive text.
We are only publicly releasing urls such that if a user decides to remove a post, the link to the post will become invalid. This dataset should not be used for tasks that might disclose the identity of the users or directly or indirectly harm them.
The public csv containing post urls and comment ids can be downloaded from here. To download the media please see our code here.
Code to reproduce our experiments and saved model checkpoints can be found here.
@inproceedings{hanu2022vtc,
title={VTC: Improving Video-Text Retrieval with User Comments},
author={Laura Hanu and James Thewlis and Yuki M. Asano and Christian Rupprecht},
booktitle={ECCV},
year={2022}
}
For questions regarding the dataset or the paper please email [email protected].